Solid-State Ion Exchange in Microporous

and Mesoporous Materials

Hellmut G. Karge 1 and Hermann K. Beyer 2
1 Fritz Haber Institute of the Max Planck Society, Faradayweg 4–6, 14195 Berlin, Germany;
e-mail: [email protected]
2 Chemical Research Center, Institute of Chemistry, Hungarian Academy of Sciences,
Pusztaszeri út 59–67, 1025 Budapest, Hungary; e-mail: [email protected]

Dedicated to Professor Gerhard Ertl on the occasion of his 65th birthday

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2 Concept of Solid-State Ion Exchange (SSIE) . . . . . . . . . . . . . 49
3 Experimental Procedures for SSIE . . . . . . . . . . . . . . . . . . 50
4 Techniques for Monitoring SSIE . . . . . . . . . . . . . . . . . . . 51
4.1 Introductory Remarks . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.2 Chemical Analysis (CA) . . . . . . . . . . . . . . . . . . . . . . . . 51
4.3 Thermogravimetric Analysis (TGA) . . . . . . . . . . . . . . . . . 51
4.4 Temperature-Programmed Evolution of Gases (TPE) . . . . . . . . 52
4.5 Combination of TGA and TPE . . . . . . . . . . . . . . . . . . . . . 53
4.6 X-ray Diffraction (XRD) . . . . . . . . . . . . . . . . . . . . . . . . 53
4.7 Infrared Spectroscopy (IR) and Fourier Transform (FT)
Raman Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4.8 Electron Spin Resonance Spectroscopy (ESR) . . . . . . . . . . . . 54
4.9 Magic Angle Spinning Nuclear Magnetic Resonance
Spectroscopy (MAS NMR) . . . . . . . . . . . . . . . . . . . . . . . 55
4.10 Mössbauer Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . 55
4.11 X-ray Photoelectron Spectroscopy (XPS) . . . . . . . . . . . . . . . 55
4.12 X-ray Spectroscopy: EXAFS, XANES . . . . . . . . . . . . . . . . . 56
5 Systems Investigated . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.1 SSIE of Alkaline (M+) and Alkaline Earth (M2+) Metal Cations . . . 60
5.1.1 Introductory Remarks . . . . . . . . . . . . . . . . . . . . . . . . . 60
5.1.2 Application of TPE . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
5.1.3 SSIE and Lattice Energy . . . . . . . . . . . . . . . . . . . . . . . . 60
5.1.4 Application of a Combination of TGA and TPE . . . . . . . . . . . 65
5.1.5 Stoichiometry of SSIE of M+ and M2+ Halides with H-Zeolites . . . 67
5.1.6 Preservation of Crystallinity upon SSIE . . . . . . . . . . . . . . . 69
5.1.7 Role of the Nature of the Anions . . . . . . . . . . . . . . . . . . . 70
5.1.8 SSIE of M+ and M2+ Halides with H-Zeolites Investigated by IR . . 70
5.1.9 SSIE of M+ and M2+ Halides with Na-Zeolites Investigated by IR 75
5.1.10 SSIE of M+ and M2+ Halides with Na-Zeolites Investigated
by MAS NMR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

Molecular Sieves, Vol. 3

© Springer-Verlag Berlin Heidelberg 2002
44 H.G. Karge · H.K. Beyer

5.1.11 Catalytic Activity of M2+-Zeolites Prepared via SSIE . . . . . . . . 79

5.2 SSIE of Lanthanum (La3+) Cations . . . . . . . . . . . . . . . . . . 79
5.2.1 Introductory Remarks . . . . . . . . . . . . . . . . . . . . . . . . . 79
5.2.2 SSIE of La3+ Chloride with H-Zeolites Investigated by TPE . . . . . 80
5.2.3 Stoichiometry of SSIE of La3+Chloride with H-Zeolites . . . . . . . 82
5.2.4 SSIE of La3+ Chloride with H-Zeolites Investigated by IR . . . . . . 83
5.2.5 SSIE of La3+ Chloride with Na-Zeolites Investigated by XRD . . . . 85
5.2.6 SSIE of La3+ Chloride with Na-Zeolites Investigated by IR . . . . . 87
5.2.7 SSIE of La3+ Chloride with Na-Zeolites Investigated
by 23Na MAS NMR . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
5.2.8 Catalytic Activity of La3+-Zeolites Prepared via SSIE . . . . . . . . 89
5.3 SSIE of Other Transition Metal Cations . . . . . . . . . . . . . . . . 94
5.3.1 Introductory Remarks . . . . . . . . . . . . . . . . . . . . . . . . . 94
5.3.2 SSIE of Copper, Silver and Gold Compounds with Zeolites . . . . . 94
5.3.2.1 Introduction of Copper . . . . . . . . . . . . . . . . . . . . . . . . 94
5.3.2.2 Introduction of Silver . . . . . . . . . . . . . . . . . . . . . . . . . 108
5.3.2.3 Introduction of Gold . . . . . . . . . . . . . . . . . . . . . . . . . . 110
5.3.3 SSIE of Zinc, Cadmium and Mercury Compounds with Zeolites . . 110
5.3.3.1 Introduction of Zinc . . . . . . . . . . . . . . . . . . . . . . . . . . 110
5.3.3.2 Introduction of Cadmium . . . . . . . . . . . . . . . . . . . . . . . 113
5.3.3.3 Introduction of Mercury . . . . . . . . . . . . . . . . . . . . . . . . 113
5.3.4 SSIE of Iron, Cobalt, Nickel and Manganese Compounds
with Zeolites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
5.3.4.1 Introduction of Iron . . . . . . . . . . . . . . . . . . . . . . . . . . 114
5.3.4.2 Introduction of Cobalt . . . . . . . . . . . . . . . . . . . . . . . . . 124
5.3.4.3 Introduction of Nickel . . . . . . . . . . . . . . . . . . . . . . . . . 127
5.3.4.4 Introduction of Manganese . . . . . . . . . . . . . . . . . . . . . . 131
5.3.5 SSIE of Vanadium, Niobium, Antimony, Chromium,
Molybdenum and Tungsten Compounds with Zeolites . . . . . . . 134
5.3.5.1 Introductory Remarks . . . . . . . . . . . . . . . . . . . . . . . . . 134
5.3.5.2 Introduction of Vanadium . . . . . . . . . . . . . . . . . . . . . . . 135
5.3.5.3 Introduction of Niobium and Antimony . . . . . . . . . . . . . . . 140
5.3.5.4 Introduction of Chromium . . . . . . . . . . . . . . . . . . . . . . 141
5.3.5.5 Introduction of Molybdenum . . . . . . . . . . . . . . . . . . . . . 147
5.3.5.6 Introduction of Tungsten . . . . . . . . . . . . . . . . . . . . . . . 153
5.4 SSIE of Noble Metal Cations with Zeolites . . . . . . . . . . . . . . 153
5.4.1 Introductory Remarks . . . . . . . . . . . . . . . . . . . . . . . . . 153
5.4.2 Preparation of Noble-Metal-Containing Large-Pore Zeolite
Catalysts by SSIE . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
5.4.3 Preparation of Noble-Metal-Containing Narrow-Pore Zeolites
by SSIE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
5.4.4 Preparation of Bifunctional Zeolite Catalysts by SSIE . . . . . . . . 158

6 Modified SSIE and Related Processes . . . . . . . . . . . . . . . . . 162

6.1 Introductory Remarks . . . . . . . . . . . . . . . . . . . . . . . . . 162
6.2 Contact-Induced Ion Exchange . . . . . . . . . . . . . . . . . . . . 163
Solid-State Ion Exchange in Microporous and Mesoporous Materials 45

6.3 Gas-Phase-Mediated Processes Related to SSIE . . . . . . . . . . . 165

6.3.1 Introduction of Cations into Zeolites Through a Vapor Phase
Containing the In-Going Species . . . . . . . . . . . . . . . . . . . 165
6.3.2 Effect of Additional Molecules in the Vapor Phase on SSIE
at Elevated Temperatures . . . . . . . . . . . . . . . . . . . . . . . 166
6.3.3 Oxidative and Reductive SSIE . . . . . . . . . . . . . . . . . . . . . 168
6.3.3.1 Oxidative SSIE of Ag0, Cu0 and Pd0 in the Presence of O2 or Cl2 . . 168
6.3.3.2 Reductive SSIE of Ga2O3 . . . . . . . . . . . . . . . . . . . . . . . . 168
6.3.3.3 Reductive SSIE of In2O3 . . . . . . . . . . . . . . . . . . . . . . . . 171

7 The Role of Water in and Mechanisms of SSIE . . . . . . . . . . . . 181

7.1 Role of Water in SSIE . . . . . . . . . . . . . . . . . . . . . . . . . . 181
7.2 Possible Mechanisms of SSIE . . . . . . . . . . . . . . . . . . . . . 184

8 Kinetics of SSIE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186

9 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . 191

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192

Abbreviations
A zeolite structure (cf. [180])
A absorbance, e.g., of OH groups in IR, A(OH), etc.
AES Auger electron spectroscopy
ALPHA zeolite structure (cf. [180])
AlPO4-5 microporous aluminophosphate structure (cf. [180])
At absorbance of a typical band at time t
At = • , A• absorbance after infinite time t
ATS microporous aluminophosphate structure (cf. [180])
B Brønsted (e.g., acid sites, acidity)
BET Brunauer-Emmet-Teller (method for determination of surfaces)
BETA zeolite structure (cf. [180])
CA chemical analysis
CE conventional ion exchange
CLIN clinoptilolite, zeolite structure (cf. [180])
DENOX process for removal of nitrogen oxides
DEXAFS dispersive extended X-ray absorption fine structure
DPPH 2,2-diphenyl-1-picrylhydrazyl, ESR standard
DRS diffuse reflectance spectroscopy (in IR or UV-Vis range)
DTG differential thermogravimetry
DH
_ ad (differential) heat of adsorption
E most frequent energy of desorption in an energy distribution
EL lattice energy
EDAX energy dispersive X-ray (spectroscopy)
EDS energy dispersive X-ray (spectroscopy)
46 H.G. Karge · H.K. Beyer

EDX energy dispersive X-ray (spectroscopy)

ESEM electron spin echo modulation (spectroscopy)
ESR electron spin resonance (spectroscopy)
EXAFS extended X-ray absorption fine structure
EU-1 zeolite structure (cf. [180])
FER ferrierite, zeolite structure (cf. [180])
FTIR Fourier transform infrared (spectroscopy)
GHSV gaseous hourly space velocity
H magnetic induction (related to magnetic field strength), in tesla
HFS hyperfine splitting (of ESR signals)
HT high-temperature (peak, etc.)
I nuclear spin quantum number
IR infrared (spectroscopy)
IS isomer shift
L Lewis (e.g., acid sites, acidity)
LT low-temperature (peak, etc.)
M metal
MAPO-36 microporous Mg aluminophosphate structure, (cf. [180])
MAS NMR magic angle spinning nuclear magnetic resonance (spectroscopy)
MCM-41 mesoporous silicate structure (cf. Vol. 1, Chap. 4)
MFI zeolite structure (cf. [180])
MM montmorillonite
MOR mordenite, zeolite structure (cf. [180])
MS mass spectrometry
nM /nAl ratio of metal to aluminum atoms
nSi /nAl ratio of silicon to aluminum atoms in the framework
OFF offretite, zeolite structure (cf. [180])
Py pyridine
QS quadrupole splitting
RSSIE reductive solid-state ion exchange
SAPO-n microporous silicoaluminophosphates, n = 5, 34, etc. (cf. [180])
SEM scanning electron microscopy
SCR selective catalytic reduction
SHFS super hyperfine splitting (of ESR signals)
SSIE solid-state ion exchange
STEM scanning transmission electron microscopy
T tesla, unit of magnetic induction (related to magnetic field
strength)
TCD thermal conductivity detector
TEA+ tetraethylammonium cation, derived from TEAOH (template)
TEM transmission electron microscopy
TG thermogravimetry
TGA thermogravimetric analysis
Tp peak temperature
TPE temperature-programmed evolution (of gases)
TPEH temperature-programmed evolution of hydrogen
TPD temperature-programmed desorption
Solid-State Ion Exchange in Microporous and Mesoporous Materials 47

TPDA temperature-programmed desorption of ammonia

TPO temperature-programmed oxidation
TPR temperature-programmed reduction
USY ultra stable Y-zeolite
UV-Vis ultraviolet/visible (spectroscopy)
X zeolite structure (cf. [180])
X conversion of a compound, in percent upon reaction, X(%)
XAES X-ray induced Auger electron spectroscopy
XAS X-ray absorption spectroscopy
XANES X-ray absorption near edge structure
XPS X-ray photoelectron spectroscopy
XRD X-ray diffraction
Y zeolite structure (cf. [180])
Y yield of a product upon reaction, in percent, Y(%)
ZK-5 zeolite structure (cf. [180])
ZSM-5 zeolite structure (cf. [180])
ZSM-35 zeolite structure (cf. [180])
ZSM-48 zeolite structure (cf. [180])

1
Introduction
In zeolites, i.e., crystalline microporous aluminosilicates [1–3], the framework
exhibits negative charges as a consequence of the incorporation of trivalent alu-
minum atoms instead of tetravalent silicon. The same situation is encountered in
most cases of related crystalline microporous (cf. [4, 5]) and crystalline-like meso-
porous materials, e.g., in isomorphously (by Fe3+, Ga3+, B3+, etc.) substituted zeolite
structures and Al-containing M41S materials, respectively (cf., e.g., [6, 7]). These
negative charges of the frameworks must be compensated by the positive charges
of extra-framework cations or via the interaction of the framework oxygen atoms
with protons under formation of acid hydroxyls, i.e., so-called Brønsted centers
(cf., e.g., [8]). According to the synthetic procedures (see, e.g., [9, 10]), usually Na+,
K+ or template cations are present in as-synthesized micro- or mesoporous mate-
rials and play the role of charge-compensating species; proton attack of the oxygen
atoms of the framework occurs,e.g.,on removal of the organic template molecules.
However, the charge-compensating entities (alkali metal cations, protons, etc.) can
be replaced by other cations, and this makes zeolites inorganic cation exchangers
(cf., e.g., [11, 12]). The ion-exchange capacity of microporous materials, especially
of zeolites such as LTA- and P-type zeolites (see [13,14]),is the basis for their world-
wide application as detergent builders [15]. However, ion exchange is also one of
the most important processes for post-synthesis modification of microporous
materials, for instance, in order to tailor adsorbents and catalysts. Ion exchange is,
therefore, carried out not only in research laboratories, but also on an industrial
scale. In what follows, we refer to zeolites as the most important representatives of
microporous materials, in particular in ion exchange. However, related micro-
porous and mesoporous materials may be treated in an analogous manner.
48 H.G. Karge · H.K. Beyer

Conventionally, ion exchange is carried out by suspending the zeolite powder

in aqueous solutions of salts which contain the desired in-going cation, or in
mineral acids to introduce protons into acid-resistant zeolites. After stirring the
suspension for a period of time (frequently at temperatures higher than ambi-
ent), the zeolite powder and the solution must be separated. Generally, the pro-
cedure must be repeated several times to achieve a sufficiently high degree of
exchange. Ion exchange in aqueous solution is, thus, a time-consuming process
and, in any case, it produces large amounts of waste solutions that must be
regenerated or discarded in an environmentally friendly manner. Moreover, in
several instances, it is very difficult or even totally impossible to introduce par-
ticular cations into a given zeolite by means of conventional ion exchange in
aqueous solution. Examples are encountered where the cations are available
only in complexes which are too bulky to enter the narrow pores of the respec-
tive zeolites (vide infra, e.g., Sect. 5.4.3).
The situation is significantly different in solid-state ion exchange, where dry
powders of zeolites and salts or oxides containing the in-going cation are react-
ed. In favorable cases, very high degrees of exchange can be achieved in only one
step and, also, certain metal cations can be incorporated into narrow-pore
zeolites where exchange attempts in aqueous solutions fail (vide infra). Thus,
solid-state ion exchange appears to be in many cases an interesting alternative
to conventional exchange and has indeed attracted ever-growing attention and
experienced increasing application during the past two decades.
Scattered in the literature, some early work was reported on solid-state ion
exchange as early as the beginning of the 1970s. For instance, Rabo et al. [16, 17]
used the reaction between KCl and an acidic form of a faujasite-type catalyst,
Ca,H-Y, to remove residual Brønsted acid sites from the microporous solid and,
thus, eliminate any catalytic activity for acid-catalyzed reactions. Infrared (IR)
spectroscopy was used to confirm the elimination of the acidic OH groups.
Another group, viz., Clearfield and co-workers, published a two-page “Recent
Research Report” [18] on the solid-state reaction between A-, X-, and Y-type zeo-
lite samples and salts of transition metals such as Ni and Co, monitoring the
process by electron spin resonance spectroscopy (ESR). The ion exchange
between two physically contacted forms of a zeolite, viz., one sample of zeolite A
loaded with Li+ and another sample containing Na+, was studied by Kokotailo et
al. [19] and Fyfe et al. [20]; these authors employed X-ray diffraction (XRD) and
nuclear magnetic resonance (NMR) to prove the exchange in the solid state. Simi-
larly, it was suspected that, on applying the KBr technique in IR spectroscopy of
zeolites, an exchange of K+ of the KBr matrix and the zeolite cations may occur.
Systematic work on solid-state ion exchange in zeolites started, however, only
in the mid-1980s in the groups of Slinkin in Moscow [21] and Karge in Berlin
[22]. Since then, a large body of systems has been investigated, numerous tech-
niques for monitoring solid-state ion exchange in zeolites have been developed,
and specific problems related to solid-solid reactions in these systems tackled.
Solid-state ion exchange has, thus, become a well-established method for post-
synthesis modification of zeolites and related materials. This technique is, there-
fore, frequently employed in solving special problems of post-synthesis modifi-
cation of zeolites both in research and industrial processes.
Solid-State Ion Exchange in Microporous and Mesoporous Materials 49

Earlier reviews of the field were somewhat limited and focused on the intro-
duction of transition metal cations [23] or alkali metal and rare earth cations
into zeolites [24, 25]. The present contribution intends to provide a more extend-
ed and almost complete overview of the state of the art and, also, to address more
recent developments such as reductive solid-state ion exchange, related process-
es and questions concerning the role of water in contact-induced ion exchange,
comparison of conventional and solid-state exchange, and incorporation of
noble metals into narrow-pore zeolites.

2
Concept of Solid-State Ion Exchange (SSIE)
Scheme 1 shows the general chemistry of ion exchange in aqueous solution
(conventional exchange, CE) and, in contrast, of exchange in the solid state
(solid-state ion exchange, SSIE).
In the case of SSIE, the main feature is that, usually at elevated temperatures,
the pure solids are brought to reaction, whereas in CE the solvent (water) is
involved as a third component, and the in-going and out-going cations are sol-
vated. Solvation of the cations in many cases impedes ion exchange; this cannot
occur in SSIE. In both SSIE and CE, the exchange process almost always leads to
an equilibrium (cf. Schemes 1a and 1b). However, in solid-state ion exchange,

a Conventional exchange in aqueous solutions (CE)

p M1n1+ (Z–)n1 + q [M2(H2O)w2]n2+ + (n2/m)q [A(H2O)wA]m– + wH2O
¤ x M2n2+ (Z–)n2 + (p – [n1/n2]x) M1n1+ (Z–)n1 + (n1/n2)x [M1(H2O)w1]n1+ (1)
+(n2/n1)x [A(H2O)w]m– + (q – x) [M2(H2O)w2 ]n2+ + (q – x) n2/m [A(H2O)wA]m– + w¢H2O

b Solid-state ion exchange (SSIE)

p M1n1+ (Z–)n1 + q M2n2+ (Am–)n2/m ¤ xM2n2+ (Z–)n2 + (p – [n1/n2]x) M1n1+(Z–)n1
+ (n1/n2)x M1n1+(Am–)n1/m + (q – x) Mn2+ m–
2 (A )n2/m (2)

c Solid-state ion exchange under removal of one (volatile) component of the product
M1 = H; n1 = 1; M2 = Ca; n2 = 2; A = Cl; m = 1
p H-Z+q CaCl2 Æ x Ca-Z2 + (p-2x) H-Z + 2x HCl≠ + (q-x) CaCl2 (3)

d Solid-state ion exchange with a degree of 100% with respect to H-Z ; p = 2x

p H-Z + q CaCl2 Æ p/2 Ca-Z2 + p HCl≠+ (q – p/2) CaCl2 (4)

e Stoichiometric solid-state ion exchange; q = p/2

p H-Z + p/2 CaCl2 Æ p/2 Ca-Z2 + p HCl≠ (5)

Scheme 1. General formulation of ion exchange in zeolites. M1 out-going cation, originally

present in the zeolite to be modified; n1 valency of M1; M2 in-going cation to be introduced
into the zeolite; n2 valency of M2; A anion of the M2-containing compound (salt, oxide);
m valency of A; Z– monovalent fragment of the zeolite framework; p, q, x, w1, w2, w, w¢:
stoichiometric coefficients
50 H.G. Karge · H.K. Beyer

the conditions can be chosen such that one component of the exchange product
is continuously removed and the equilibrium shifted to the right side, so that an
exchange degree of 100% may be achieved. This is illustrated in Scheme 1c for
the hydrogen form of a zeolite, H-Z (M1 = H, n1 = 1), where, for sake of simplic-
ity, M2 = Ca, n2 = 2 and A = Cl, m = 1. The solid-state reaction may be conduct-
ed in a mixture with the salt being present in a sub-stoichiometric (q<p/2), a
stoichiometric (q = p/2) or an excess (q>p/2) amount. This gives the possibility
of achieving a certain desired degree of exchange.

3
Experimental Procedures for SSIE
The procedures for solid-state ion exchange, both in laboratory experiments and
on an industrial scale, are relatively simple. The main requirement is to achieve
a mixing as intimate as possible of the zeolite powder and the component con-
taining the in-going cation. This can be brought about by careful grinding in a
mortar or by milling. However, there are cases where the zeolite lattice is sensi-
tive to mechanical stress (cf. [26, 27]). In such a situation a different way of mix-
ing may be chosen: the zeolite powder and the carefully ground salt or oxide of
the in-going cation are suspended together in an inert vaporizable liquid such as
n-hexane. After sufficient mixing by stirring or shaking, the liquid is evaporat-
ed, which yields a very intimate mixture of the crystallites of the zeolite and the
respective salt or oxide (cf. [22]). In a further alternative procedure, the two com-
ponents, i.e., the salt or oxide and the zeolite, are separately pretreated and
ground and subsequently mechanically mixed under application of an ultrason-
ic treatment [28].
The dried intimate mixture of the zeolite and salt or oxide is then heated at a
low rate (5–10 K min–1) to the reaction temperature which is usually between
675 and 875 K. Heating is carried out in a stream of an inert gas (e.g., air, N2 , He)
or in high vacuum. In most cases it is sufficient to keep the mixture at the reac-
tion temperature for 2–4 h; in particular cases it might be advisable to extend
the reaction time up to 12–24 h. Long reaction times are especially required for
systems in which the migration of the in-going and/or out-going species is slow.
(cf. Sect. 8). In a number of instances, the first stage of SSIE proceeds easily and
starts at relatively low temperatures (followed by a high-temperature process),
or it happens that a fraction of the cations is already exchanged during grinding
or milling of the mixtures.
When the reaction is completed, the mixture is cooled to ambient tempera-
ture and analyzed (cf. Sect. 4). If an excess of the compound of the in-going
cation has been applied, non-reacted salt can be removed by brief washing with
water provided that the respective compound is soluble. Similarly, if the zeolite
employed is not in the hydrogen form and an equilibrium has been reached
(cf. Scheme 1b), the residual salt can be removed by brief washing and, if
required, the procedure may be repeated. However, one has to be aware of the
possibility that salts remained entrapped in the zeolite matrix (cf. [16, 17]).
Solid-State Ion Exchange in Microporous and Mesoporous Materials 51

4
Techniques for Monitoring SSIE

4.1
Introductory Remarks

An extremely large armory of techniques has been developed and applied to

identify and monitor the process of solid-state ion exchange (SSIE). The respec-
tive methods can be distinguished as spectroscopic and non-spectroscopic tech-
niques. Both kinds of experimental tools are usually applicable for both simply
confirming the SSIE after completion of the reaction and monitoring the SSIE
during the reaction by in situ measurements. Methods described in the follow-
ing sections will be illustrated by a large number of pertinent examples, provid-
ed especially in Sect. 5.

4.2
Chemical Analysis (CA)

In those cases where a stoichiometric conversion has occurred or a removal of

an excess of the in-going cations was possible, simple chemical analysis of the
exchanged zeolite will be suitable to verify the fact of solid-state ion exchange
and will also provide the ratio of nM2/nM1 (M1 = original cation, M2 = in-going,
M1 replacing cation) in the product, i.e., the degree of exchange (for methods of
analysis, cf. Vol. 5, Chap. 1, this series).

4.3
Thermogravimetric Analysis (TGA)

Since the solid-state ion reaction between hydrogen forms of zeolites and salts
(fluorides, chlorides, bromides) or oxides implies the release of volatile products
(HF, HCl, HBr, H2O), the exchange can be followed by measuring the weight loss
of the mixture. Prior to the reaction, the salt/zeolite or oxide/zeolite mixture
must be kept at about 400 K in a sensitive microbalance under dynamic vacuum
or an inert gas stream in order to remove physically adsorbed water. Subse-
quently, the mixture is heated at a rate of 5–10 K min–1 until a constant weight is
reached which usually occurs at about 900 K. From the weight loss, the degree of
ion exchange can be calculated according to the stoichiometry (cf. Scheme 1). It
should be noted, however, that not only the Brønsted acid centers but to some
extent also silanol groups may react, so that the gravimetrically determined
degree of exchange in some cases may exceed the value corresponding to the
amount of ∫Al-(OH)-Si∫ exchange sites. In such a situation at least one addi-
tional technique listed in this section must be employed.
52 H.G. Karge · H.K. Beyer

4.4
Temperature-Programmed Evolution of Gases (TPE)

The thermal treatment in this technique is similar to that applied in TGA (vide
supra). However, the solid-state reaction may be carried out in a quartz glass
oven or in a stainless-steel device. The gases evolved, e.g., HCl or H2O, are deter-
mined by mass spectrometry (MS) or with a thermal conductivity detector
(TCD). A suitable device including MS is seen in Fig. 1 and is described in more
detail elsewhere [29].

Fig. 1. A Scheme of the apparatus for temperature-programmed desorption or temperature-

programmed evolution of gases: 1, roughing pump; 2, bellows; 3, turbo molecular pump;
4, swing gate valves; 5, ionization gauge and control unit; 6, sample holder and heating device
(pan, see B); 7, temperature programmer; 8, ion source and mass filter; 9, secondary electron
multiplier (SEM/MS); 10, dosing valves; 11, Baratron gauge, Baratron control unit; 13, pyridine
reservoir; 14, cylinder with probe (e.g., ammonia); 15, calibrated volume; 16, cooling finger.
B Sample holder and heater (stainless steel): 1, feed through; 2, heating wire; 3, heating block
(copper); 4, thermocouple; 5, CF 35 flange (after [29], with permission)
Solid-State Ion Exchange in Microporous and Mesoporous Materials 53

Application of MS to monitor SSIE has the advantage that several simultane-

ously evolved gases can be separately analyzed, for instance, NH3 , HCl and H2O.
In the case of evolved anhydrous acids such as HCl, the vapor can be transferred
by a stream of N2 or He into a trap containing an aqueous solution of a base (e.g.,
NaOH). When the reaction is complete, the degree of exchange can be deter-
mined via back-titration of the residual base. Also, the procedure can be
designed such that a continuous titration of the HCl evolved and trapped in
water is possible.

4.5
Combination of TGA and TPE

A combination of gravimetric analysis (cf. Sect. 4.3) and temperature-pro-

grammed evolution of volatile product components (cf. Sect. 4.4) is especially
advantageous to monitor SSIE in zeolites. Analysis and determination of the
evolved gases by MS, TCD or titration enables us to interpret more specifically
the change in weight of the reacting salt (oxide)/zeolite mixture. An instructive
example will be discussed in Sect. 5.1.4. Temperature-programmed evolution of
gases may also be advantageously combined with XRD, IR, magic-angle spin-
ning (MAS) NMR, and Mössbauer spectroscopy, even though only a few exam-
ples have been reported so far.

4.6
X-ray Diffraction (XRD)

X-ray diffraction may be applied to the starting mixture of a salt or oxide and
the zeolite and, after completion of the solid-state reaction at a given constant
temperature, to the product mixture. A comparison of the respective XRD pat-
terns will then show whether (1) the intensities of reflections typical of the com-
pound containing the in-going cation have decreased or disappeared and (2) the
reflections of the zeolitic phase have changed accordingly. In principle, the
intensities of the reflections of the zeolitic phase are dependent on the nature of
the charge-compensating extra-framework cations. In many cases these intensi-
ties are remarkably sensitive to an exchange of the cations. This type of XRD
experiment can be carried out at different constant reaction temperatures,
which will provide insight into the temperature dependence of the achievable
exchange degree. However, if the diffractometer is equipped with a continuously
heatable XRD chamber, it is possible to monitor in situ the progress of the sol-
id-state reaction. In those cases where the hydrogen form of a zeolite is reacted,
the XRD analysis can be combined with the measurement of the temperature-
programmed evolution of gases (TPE). Application of XRD in SSIE is illustrated
in Sect. 5.2.5 using the introduction of La3+ cations into an Na-Y zeolite as an
example.
54 H.G. Karge · H.K. Beyer

4.7
Infrared Spectroscopy (IR) and Fourier Transform (FT) Raman Spectroscopy

IR spectroscopy can be employed to study solid-state ion exchange with and

without probe molecules. The application of IR is most straightforward with
respect to the exchange of protons of the hydrogen forms of zeolites. Upon reac-
tion with a salt or oxide, the intensity of the band(s) originating from the Brøn-
sted acid hydroxyl groups will decrease. The loss of intensity provides a measure
of the degree of exchange. The IR experiments must be carried out with salt
(oxide)/zeolite mixtures pressed into self-supporting thin wafers (cf., e.g., [30,
31]) in an appropriate heatable cell (cf., e.g., [31]). If the IR cell is continuously
heatable, the progress of exchange can be monitored in situ under high vacuum
or in a flow of inert gas passed through the cell; this may be combined with a
TPE measurement.
Another approach is to employ probe molecules such as pyridine. Pyridine
attached to different cations gives rise to IR bands in the region from 1430 to
1460 cm–1 that are indicative of the nature of the cations [32]. This enables IR
spectroscopy to be employed not only when the exchange with hydrogen forms
of zeolites is investigated, but also when the starting zeolite material contains
one kind of cations which should be replaced by another. This is dealt with in
Sects. 5.2.6 and 5.3.2 where the exchange of Na+ by La3+ and Cu+, respectively, is
discussed. Historically, the IR method was used in the early experiments by
Rabo et al. [16, 17] mentioned in Sect. 1.
Fourier transform (FT) Raman spectroscopy was applied by Huang et al. [33]
to study the solid-state ion exchange between LiCl or CaCl2 and NH4-Y or Na-Y
as well as contact-induced ion exchange in the systems Li-A/Na-Y, Li-A/Na-X
and Li-A/Ca-A (cf. Sect. 6.2).

4.8
Electron Spin Resonance Spectroscopy (ESR)

The incorporation of transition metal cations both into hydrogen and cationic
forms of zeolites (H-Z, M-Z, with M = Na, K, etc.) via reaction with the com-
pounds containing the respective transition metal is, in many cases, easily seen
by electron spin resonance (ESR) spectroscopy. In fact, the early studies by
Clearfield et al. [18] were conducted using ESR (cf. Sect. 1). Since ESR measure-
ments can also be carried out at elevated temperatures, in situ observations are
possible as well.A suitable high-temperature ESR cell was developed by Karge et
al. [34]. Using ESR, it can be confirmed that solid-state ion exchange has taken
place in a particular system containing transition metal cations and, moreover,
the coordination of the incorporated cations can be determined. However, the
ESR technique is frequently not capable of providing satisfactory quantitative
results. Examples of ESR investigations of SSIE are reported in Sects. 5.3.2, 5.3.4
and 5.3.5.
Solid-State Ion Exchange in Microporous and Mesoporous Materials 55

4.9
Magic Angle Spinning Nuclear Magnetic Resonance Spectroscopy (MAS NMR)

Magic angle spinning nuclear magnetic resonance (MAS NMR) spectroscopy

was first applied to problems of solid-state ion exchange when the system
LaCl3 ◊7H2O/Na-Y was studied [35]. With the help of 23Na MAS NMR, the fact
that exchange had taken place was impressively demonstrated by the appear-
ance of a signal at 0 ppm when the spectrum was referenced to NaCl (cf.
Sect. 5.2.7). Moreover, the change in the intensities of the signals indicating Na+
in super- and b-cages was monitored as a function of the reaction temperature
and related to the exchange features. However, the behavior of many other
cations during SSIE can also be studied by MAS NMR as well (cf. Sect. 5.1.10).
Particularly interesting examples are the introduction of Cs+ into H-Y investi-
gated by Weitkamp et al. with the help of 133Cs MAS NMR [36] and the study of
the behavior of hydrogen zeolites in SSIE by 1H MAS NMR. Usually, these MAS
NMR studies are conducted at room temperature after ex situ preparation of the
samples. However, recent developments of the experimental MAS NMR tech-
nique have rendered high-temperature investigations possible [37–39]. Thus,
this should enable SSIE in zeolites to be monitored by in situ measurements.

4.10
Mössbauer Spectroscopy

There are only a few “Mössbauer nuclei” which are interesting in zeolite chem-
istry and, thus, candidates for application of Mössbauer spectroscopy in solid-
state ion exchange. However, among them is one of the most important ele-
ments, viz., iron, which has also attracted much attention in zeolite chemistry as
a key component of possible catalyst formulations. Mössbauer spectroscopy
proved to be exceptionally successful in discriminating Fe2+ and Fe3+ cations
residing on extra-framework sites after introduction of iron via solid-state ion
exchange. Moreover, Mössbauer spectroscopy provides information about the
various coordinations of Fe2+ and Fe3+ in zeolite lattices (cf. Sect. 5.3.4).

4.11
X-ray Photoelectron Spectroscopy (XPS)

Only a few experiments have been reported where X-ray photoelectron spec-
troscopy (XPS) has been used to study solid-state reactions between salts and
zeolites. XPS enables us to determine changes in the surface composition of a
zeolite sample before and after it has been subjected to solid-state ion exchange.
This technique is suitable to monitor, for instance, variations in the surface
ratios nM/nAl and nM/nSi of the zeolite upon solid-state reaction as a function of
temperature, the ratio salt/zeolite and the reaction time. Examples will be pro-
vided in, e.g., Sects. 5.3.2.1 and 5.3.4.4.
56 H.G. Karge · H.K. Beyer

4.12
X-ray Spectroscopy: EXAFS and XANES

Analysis of the extended X-ray absorption fine structure (EXAFS) and X-ray
absorption near edge structure (XANES) of mixtures of salts and zeolites before
and after heat-treatment enables us to prove or disprove that a solid-state reac-
tion has occurred. Moreover, EXAFS is a tool to determine the coordination of
cations introduced via SSIE into the zeolite matrix. If these cations are reducible
to the zero-valent state as, e.g., Pt2+ or Pd2+ cations are, EXAFS also provides
information about the size of the metal particles generated by reduction. To
date, unfortunately, only a few pertinent studies have been reported (cf., e.g.,
[40] and Sect. 5.4.2).

5
Systems Investigated
Even though the subsequent list may, in fact, not be exhaustive, Table 1 compris-
es at least most of the systems subjected up to date to solid-state ion exchange.
This, however, does not mean that all the systems enumerated in Table 1 were
investigated in great detail. With respect to many of them, several important
questions still remained unanswered. Certainly, more detailed research on sev-
eral particular systems will be carried out only if the interest in their solid-state
ion exchange is stimulated by the need to solve special problems related to these
systems and if there appears to be a chance to successfully solve the problems via
SSIE as, for instance, in catalyst preparation (cf. Sects. 5.3 and 5.4).
With respect to the types of cations desired to be introduced by SSIE, there is,
nevertheless, practically no important group which has not been dealt with so
far. Thus, incorporation of cations via solid-state ion exchange has been studied
not only for alkaline, alkaline earth and rare earth, but also for transition and
noble metal cations. However, there is still not much general insight into why, for
instance, the achievable maximum exchange degree strongly depends on the
nature of a respective in-going cation or why the nature of the anion of the
compound with the in-going ion plays a significant role. In particular, to date, no
systematic investigation on the latter problem has been carried out. There is,
therefore, still ample room for research in the field of solid-state ion exchange
(see also Concluding Remarks, Sect. 9) and only a few of these questions con-
cerning the systems studied to date can be addressed in the context of the sub-
sections that follow.
Solid-State Ion Exchange in Microporous and Mesoporous Materials 57

Table 1. Overview of systems investigated for solid-state ion exchange

Compound of the introduced cation Zeolite References

LiCl, NaCl, KCl, RbCl, CsCl H-ZSM-5, NH4-ZSM-5, NH4-Y, [16, 17, 22, 24, 33,
H-MOR, H-BETA, NH4-BETA, 43, 47–49, 58,
H-EMT, NH4-EMT, Na-Y 60–63, 73]
Na-X [67–69]
NH4-X, NH4-Y, K-Y, K-L [56, 58, 59]
RbCl, CsOH NH4-Y [36]
NaCl H-MOR, MAPO-36 [70, 130]
CsCl NH4-BETA [49, 50]
Cs[PW12O40] H-ZSM-5 [43]
NH4Cl H-BETA [51]
Na-BETA, Cs-BETA, [51]
NR4Cl (NR4: organoammonium) Na-MM (MM: montmorillonite) [28, 52]
Na-MM (MM: montmorillonite) H-ZSM-5, H-MOR [72]
Li-A, Na-A Na-A, Li-A [19, 20]
Li-A/Na-Y Li-A/Na-X Li-A/Ca-X Na-A/Li-Y, Na-A/Li-X, Ca-A/Li-X [32]

CaCl2 · 2H2O, MgCl2 · 2H2O H-MOR, NH4-MOR [41, 66]

CaCl2 NH4-Y [33, 139]
BeCl2 Na-Y [73]

LaCl3 ◊ 7H2O NH4-Y, H-ZSM-5, H-BETA [78, 83]

LaCl3 (water-free) NH4-Y, H-MOR, H-L, H-FER [287]
LaCl3 ◊ 7H2O Na-Y [79, 88]
CeCl3 , NdCl3, SmCl3 , EuCl3 , YbCl3 NH4-Y [83]
EuCl3 NH4-Y [85]
CeCl3 H-ZSM-5 [84]

Cu0 (+O2) H-MOR, H-ZSM-5 [21, 141, 142, 253,

254]
CuCl2 ◊ 2H2O, CuF2 , Cu3[(OH)CO3]2 , NH4-A, NH4-X, NH4-Y, H-MOR [18, 21, 98, 99, 104,
Cu3(PO4)2 105]
CuCl2 H-ZSM-5 [106, 114–126]
CuCl H-ZSM-5 [98]
CuCl NH4-Y, H-MOR, H-L [101–103, 112, 113]
CuCl H-BETA, H-CLIN [107–109]
CuCl Na-Y, Na-MOR [289, 290]
CuO H-[Ga]ZSM-5, Fe, H-ZSM-5 [128, 129, 148]
H-MOR, H-ZSM-5 [21, 93, 98, 110,
111, 127]
Cu2O NH4-Y [97, 98, 113]
Cu2O, Cu2S H-ZSM-5 [21, 98]
CuCl, Cu(NO3)2 , Cu(CH3COO)2 Na-ZSM-5, Na-ZSM-5/H-ZSM-5 [21, 98, 126]
CuCl2 MCM-41 [125]
CuCrO4 , CuO+CrO3 H-ZSM-5 [93, 198]

Ag0 (+O2) H, Ag-MOR [141, 253, 254]

AgCl H-ZSM-5 [130]

AuCl3 Na-Y [131]

Zn0 NH4-Y, NH4-USY, H-MOR [136, 140, 247–250]

ZnCl2 NH4-Y [104, 105, 136, 137]
ZnCl2 H-Y [139]
58 H.G. Karge · H.K. Beyer

Table 1 (continued)

Compound of the introduced cation Zeolite References

ZnCl2 NH4-Y, Na-Y [139, 144–146]

ZnCl2 H-ZSM-5 [244]
ZnO NH4-Y [143]
ZnO H-ZSM-5 [132–135]

Cd0 NH4-Y [140, 247–249]

Cd(NO3)2 , CdCl2 , CdO, CdS, CdSO4 NH4-X, NH4-Y, NH4-MOR [139]

Hg2Cl2 H-ZSM-5 [130]

FeCl2 ◊ 4H2O NH4-A, NH4-X, NH4-Y [18, 114, 148–150,

159]
FeCl2 ◊ 4H2O H-MFI, H-FER [122, 156–160]
FeCl3 H-ZSM-5 [118, 155]
FeCl3 H-ZSM-5, Na-ZSM-5, [129, 148]
H-[Ga]ZSM-5
FeO, Fe3O4 H-ZSM-5 [148]
Fe2O3 H-ZSM-5, Na-ZSM-5 [147]
Fe(NO3)3 · 9H2O Na-MM (MM: montmorillonite) [28]

CoCl2, CoCl2 · 6H2O NH4-Y, H-ZSM-5 [106, 114, 121, 139,

162, 164–170,
173–175]
CoCl2 , Co(NO3)2 NH4-CLIN, H-CLIN [109, 163]
CoCl2 H-BETA [173]
H-FER [175]
Co2(CO)8 H-Y [251]

NiCl2 NH4-Y [105, 177]

NiCl2 , NiSO4 , Ni(CH3COO)2 , NiO H-ZSM-5 [114, 121, 176]
NiCl2 H-ZSM-5 [164, 165, 167]
NiCl2 , NiCl2 · 6H2O, NiO SAPO-42 [44, 179]
NiCl2 SAPO-n (n = 5, 8, 11, 34) [178, 181–185]
NiCl2 H-MOR, H-BETA [169]
MCM-41 [186]
Raney nickel Cu-ZSM-5 [187]

MnCl2 , MnSO4 , Mn(NO3)2, Mn3O4 , H-ZSM-5 [147, 188]

Mn(CH3COO)2
MnCl2 NH4-Y [139]

V2O5 H-MOR, H,Na-MOR, H-ZSM-5, [92, 190–194]

H,Na-ZSM-5
V2O5 H,Na-X, H-Y, H-ZSM-5, H-MOR [208, 209]
V2O5 Na-Y [195–197]
V2O5 H-[Ga]ZSM-5 [128, 129]
VO(NO3)2 AlPO4-5 [206]
V2O5 NH4-ZSM-5 [205]
VOCl3 H-ZSM-5 [246]
V2O5 + CuO H-ZSM-5 [93, 198]

Nb2O5 NH4-Y, NH4 , Na-Y [202]

Solid-State Ion Exchange in Microporous and Mesoporous Materials 59

Table 1 (continued)

Compound of the introduced cation Zeolite References

Sb2O3 Na-Y, La, Na-Y [210]

CrCl3 H-Y [100, 199]

CrO3 H-MOR, H-ZSM-5 [92, 190, 199]
CuCrO4 H-MOR, H-ZSM-5 [92, 93, 190, 199]
CrO3 H-[Ga]ZSM-5, H-[Fe]ZSM-5 [128, 129, 148, 151]
Cr2O3 H-MOR, H-ZSM-5 [92, 190, 213, 214]
CrO2Cl2 MFI [246]
CrO3 + CuO H-ZSM-5 [198]
CrO3 SAPO-11 [215]

MoCl5 (MoOCl4) NH4-Y, NH4-DAY, Co, H-Y [200, 201, 216]

MoCl3 H-ZSM-5, H-MOR, H-ZSM-35, [23, 44, 179]
H-EU, H-FER, H-ZSM-48, H-L
MoO3 H-ZSM-5, H-USY, H-FER, [92, 190, 201,
H-BETA 216–220]
MoO3 Na-Y, Na-ZSM-5 [197, 217]
WO3 Na-Y [197]
WO3 (+CCl4) H-USY, H-ZSM-5, H-FER [252]

Pt0 K-L [226]

Pd0 (+Cl2) H-Y [256]
PdCl2 , Pd(NO3)2 , PdO, PtCl2 , NH4-Y, H-ZSM-5 [221–224]
PtCl4 , PtO2
PdCl2 H-ALPHA, H-RHO, H-ZK-5, [44, 45, 231]
H-ZSM-58, H-SAPO-42
PdO H-SAPO-42 [44]
PtCl2 H-RHO, H-ZK-5, H-ALPHA, [44, 45, 231]
H-ZSM-58, H-SAPO-42
PtCl2 , PtBr2 , PtO2 NH4-Y [40, 225]
RhCl3 H-ALPHA, H-SAPO-42, H-ZK-5 [44, 45, 231]
RhCl3 SAPO-11 [228]
RhCl3 DAY [227]
PdCl2 + CaCl2 , PdCl2 + LaCl3 H-ZSM-5 [221–224]

Ga0 H-ZSM-5 [140]

Ga2O3 H-ZSM-5 [111, 257–262, 265,
266]
Ga2O3 H-MOR, H-ZSM-5, H-BETA, H-Y [267–269]
Ga2O3 H-[Ga]ZSM-5 [258]
Ga2O3 H-SAPO-n (n = 5, 34, 37) [282]
(+ template)
In0 H-ZSM-5 [140]
In2O3 H-ZSM-5, NH4-Y, H-MOR, H-OFF [111, 260, 261,
267–269]
In2O3 H-ZSM-5, H, Na-Y, NH4-MOR, [271–273, 277, 278]
NH4-Y
In2O3 H-BETA (TEA-BETA) [274, 275]
In2O3 H-SAPO-n (n = 5, 34, 37) [282–284]
(+ template)
In2O3 MCM-41 [285]
60 H.G. Karge · H.K. Beyer

5.1
SSIE of Alkaline (M+) and Alkaline Earth (M2+) Metal Cations

5.1.1
Introductory Remarks

As already mentioned in Sect. 1 (Introduction), historically, the first experiment

on solid-state ion exchange between alkaline metal cations and protons of an
acidic zeolite was that described by Rabo et al. [16, 17] for the systems NaCl/Ca,
H-Y; NaCl/Ba, H-Y and NaCl/Zn, H-Y, where the authors attempted to remove
residual activity from the acidic zeolite. Without any knowledge of this rather
obscure report at the beginning of their work, Karge and Beyer started in the
mid-1980s systematic investigations on solid-state reactions between alkaline
and alkaline earth salts and zeolites [22, 41]. The experimental techniques
employed were TPE, IR, CA, XRD, and MAS NMR (cf. Sect. 4).

5.1.2
Application of TPE

The complete series of alkaline chlorides, MCl (M = Li, Na, K, Rb, Cs), was mixed
with NH4-ZSM-5 or NH4-Y powder as described in Sect. 3 and subsequently
heated. The progress of solid-state ion exchange was monitored via TPE of NH3 ,
HCl and H2O. A mass spectrometer was used to measure the relative amounts of
volatile products evolved (cf. also Fig. 1).
Figure 2 displays TPE profiles obtained during solid-state reaction between
alkaline chlorides and NH4-ZSM-5. The release of HCl (mass 36) and NH3 (mass
16, NH +2 ) allowed a distinction to be made between a low-temperature (LT) and
a high-temperature (HT) regime. The peak temperatures (Tp) of both reaction
regimes decreased in the regular sequence Tp (Na)>Tp (K)>Tp (Rb)>Tp (Cs). The
amount of cations involved in the HT reaction decreased in a similar sequence,
lithium being an exception. This element, however, frequently behaves in an
irregular manner in zeolite chemistry (see, e.g., [42]), probably because of the
small size of the Li+ cation. Thus, the LT reaction of LiCl predominates and only
a small peak around 580 K appears upon solid-state ion exchange of LiCl and
NH4-ZSM-5. Very similar results were obtained with the systems MCl/H-ZSM-5
[22] and MCl/NH4-Y ([43]; Fig. 3).

5.1.3
SSIE and Lattice Energy

The ranking of the reactivity of NaCl, KCl, RbCl, and CsCl in the HT regime of
SSIE with NH4-ZSM-5, H-ZSM-5 and NH4-Y indicated above can be related to
the lattice energy (EL) of those salts. The lattice energies decrease in the same
sequence as the peak temperatures. The following explanation may apply: the
lower the lattice energy, the lower the energy required to separate salt entities
(most likely MCl molecules) from the salt crystallites (cf. Sect. 7.2). This, in turn,
reduces the temperature that must be applied to make the solid-state reaction
Solid-State Ion Exchange in Microporous and Mesoporous Materials 61

Fig. 2. Temperature-programmed evolution (TPE) of hydrogen chloride: evolution curves

of mixtures containing NH4-ZSM-5 and LiCl (–æ), NaCl (– – –) KCl (–·–·–), RbCl (–··–··–),
CsCl (–···–···–); nMCl/nAl = 1.89 (after [22], with permission)

Fig. 3. Temperature-programmed evolution of gases (m/e = 36, HCl, æ; m/e = 18, H2O, ••••;
m/e = 17, NH3, ● ● ● ● or – - – -) from pure NH4-Y and MCl/NH4-Y mixtures upon solid-state
reaction (after [43], with permission)
62 H.G. Karge · H.K. Beyer

Table 2. Lattice energies and exchange temperatures (high-temperature peak)

Salt Lattice energy (kJ mol–1) Exchange temperature (K)

LiCl 834 –
NaCl 769 895
KCl 701 855
RbCl 680 840
CsCl 657 815

occur. In contrast, the lower degree of exchange reached through temperature-

programmed heating in the high-temperature regime when going from NaCl to
CsCl is due to the increasing size of the cations (MCl molecules), i.e., their
decreasing ability to migrate into the pore system and, in the particular case of
Y-type zeolites, to penetrate the six-membered rings to enter the b-cages (vide
infra). (With respect to introduction of alkali metal cations via SSIE in static
experiments at constant temperatures, where a 100% degree of exchange was
achieved, cf. Sect. 5.1.8.) The above-mentioned relationship between EL and Tp is
substantiated by the data of Table 2.
The relationship between the lattice energy and the reactivity for SSIE of
metal compounds in mixtures with zeolites was generally confirmed by the
work of Weitkamp and co-workers (cf. [44, 45] and Sect. 5.3) concerning the sys-
tems of noble metal chlorides/hydrogen forms of zeolites. It should be noted,
however, that such a relationship was not found in SSIE experiments with Mn
oxides as compounds of the in-going cation [46].
The studies of SSIE with alkaline salts discussed so far employed A-, X-,Y- and
ZSM-5-type zeolites. Interesting investigations were, however, carried out by
Barthomeuf and co-workers (cf. [47, 48]) and Mavrodinova (cf. [49]), who used
samples of zeolite EMT and BETA. These zeolites were synthesized with a tem-
plate, e.g., TEA in the case of BETA [49]. Therefore, post-synthesis removal of the
template by oxidation resulted in the formation of protons, and a mixed hydro-
gen-sodium form was obtained. Barthomeuf and co-workers [47, 48] studied
solid-state ion exchange of M+ (M+ = Li+, Na+, K+, Rb+, Cs+) into H-BETA and H-
EMT zeolites under static conditions, i.e., without vacuum or a flow of inert gas.
The results were compared with those of conventional exchange in aqueous
solutions. It turned out to be difficult to replace all the protons by alkali metal
cations via solid-state ion exchange. In contrast, an almost complete replace-
ment was achieved in one step through (conventional) exchange for NH4+ and
subsequent SSIE with CsCl, where the solid-state reaction was monitored by TPE
and TGA (see Fig. 4 and [47]). The disappearance of the CsCl reflections upon
heating the mixture CsCl/H,Na-BETA at 823 K was a further confirmation of the
occurrence of solid-state ion exchange.
The difference in the results of TPD of 1-propylamine from the BETA sample
prior to and after SSIE indicated a 92.7% consumption of the Brønsted acid
sites.
When NaCl was employed instead of CsCl, in fact only 50% of the protons
were replaced by Na+. The reason for this different behavior of the systems
Solid-State Ion Exchange in Microporous and Mesoporous Materials 63

Fig. 4. A Temperature-programmed evolution of 1 NH3 and 2 HCl from a CsCl/NH4-BETA

mixture and 3 NH3 from pure NH4-BETA. B DTG curves of CsCl/NH4-BETA mixtures
1 CsCl/NH4-BETA (I) with nSi /nAl = 25.1, 2 CsCl/NH4-BETA (II) with nSi /nAl = 17.06, 3 pure
NH4-BETA (I), and 4 pure NH4-BETA (II) (after [49], with permission)

CsCl/H,Na-BETA and NaCl/H,Na-BETA is not yet clear. However, Mavro-

dinova observed that the rate and degree of SSIE were higher the larger the
amount of admixed NaCl (compare the analogous behavior of the system
CuCl/H-ZSM-5; Sect. 5.3.2). The XRD pattern showed that the crystal integrity
was preserved.
Solid-state ion exchange between NH4-BETA and CsCl was also investigated
by Xu et al. [50]. These authors monitored SSIE through XRD and TGA/DTA and
examined the products with respect to their acid-base properties by IR with
pyridine as a probe and by determination of their activity in decomposition of
isopropanol. The results showed that it is easy to exchange Cs+ cations into BETA
zeolite with a very high degree of exchange. Furthermore, the products of this
SSIE process had a higher catalytic activity in base-catalyzed dehydrogenation
of isopropanol than conventionally prepared Cs-BETA formulations.
64 H.G. Karge · H.K. Beyer

In the processes described in the preceding paragraphs, solid-state ion

exchange was carried out in mixtures of MCl and H-zeolites or NH4-zeolites,
where M = Li, Na, K, Rb or Cs. Mavrodinova [51] extended these investigations
to solid-state reaction of NH4Cl with M-BETA (M = Na, Cs) or H-BETA. For
instance, a low-temperature reaction was shown to occur according to Eq. (6):
x NH4Cl + CsyHz – BETA Æ x NH3≠ + Cs(y–x)H(z+x)-BETA + x CsCl (6)
This was evidenced by TPE/MS which showed a low-temperature evolution of
NH3 but a lack of HCl formation in the same temperature range (Fig. 5). Only at
higher temperatures (805 K) was HCl evolved [see Eq. (7)] due to the reaction

Fig. 5. A Temperature-programmed evolution of NH3 from 1 NH4Cl/Cs,H-BETA and 2 NH4Cl/

Na,H-BETA mixtures. B Temperature-programmed evolution of HCl from 1 NH4Cl/Cs,H-
BETA and 2 NH4Cl/Na,H-BETA mixtures; heating rate: dT/dt = 10 K min–1; isothermal treat-
ment at 823 K for 1 h (after [51], with permission)
Solid-State Ion Exchange in Microporous and Mesoporous Materials 65

between CsCl and Cs(y–x)H(z+x)-BETA generated according to Eq. (6) at low tem-
perature:
x CsCl + Cs(y–x)H(z+x)-BETA Æ CsyHz-BETA + xHCl ≠ (7)
In the case of NH4Cl/H,Na-BETA, solid-state ion exchange in analogy to Eq. (6)
was indicated by an increase in the intensity of the NaCl reflections observed by
XRD upon heating the mixture NH4Cl/H,Na-BETA. The resulting composition
of the modified zeolite material, i.e., the extent of reactions according to Eqs. (6)
and (7), is controlled by both the reaction temperature and the concentrations
of the competing cations.
Reports on solid-state ion exchange with clays instead of zeolites are still
rather rare. In 1990 and 1992, however, two studies were published concerning
incorporation of organoammonium (NR+4 ), iron (Fe3+) and aluminum (Al3+)
cations into montmorillonite (MM; cf. [28, 52] and also Sect. 5.3.4). Thus, Ogawa
et al. [52] observed formation of organoammonium montmorillonite upon
reaction of solid organoammonium halides and dehydrated sodium montmo-
rillonite (MM):
Na-MM + NR4Cl Æ NR4-MM + NaCl (8)

5.1.4
Application of a Combination of TGA and TPE

Solid-state ion exchange in the system alkaline chloride/H-ZSM-5 was also

monitored via a combination of TPE and thermogravimetric analysis (TGA)
[22]. Figure 6 illustrates as an example the results obtained with the system
NaCl/H-ZSM-5.

Fig. 6. a TGA and b thermal gas titration curves of a NaCl/H-ZSM-5 mixture (nNaCl/nAl = 1.89)
at a heating rate of 2.5 K min–1 (after [22], with permission)
66 H.G. Karge · H.K. Beyer

The thermogravimetric curve consists of three distinct steps. The first step
can be attributed to the removal of physically adsorbed water and the concomi-
tant evolution of small amounts of HCl (LT process). The second step is not
observed with the pure materials and obviously indicates a solid-state reaction
between the two components, i.e., NaCl and H-ZSM-5 (HT process). The last step
in the TGA curve probably originates from a volatilization and/or decomposi-
tion of NaCl at temperatures above 1100 K. In parallel with TGA, the tempera-
ture-programmed evolution of HCl was followed by continuous and automa-
tized titration. The TPE curve also exhibits three steps, a small LT step at about
470 K, a large HT step between ca. 770 and 900 K and a third less distinct one at
even higher temperatures. The assignments confirm those given for the TGA
curve: the first TPE peak results from the small contribution of the HCl release
in the LT regime of the solid-state reaction and proves that most of the first step
in the weight loss curve is caused by the desorption of physically adsorbed
water; the middle and most pronounced TPE peak gives evidence of the large
contribution by the HT regime of SSIE; the last TPE step above 1100 K is indica-
tive of decomposition products (chlorine) of the salt.
Analogous to the observations described above, in the system CaCl2 ◊ 2H2O/H-
ZSM-5 the combined TGA/TPE experiment yielded also two steps in the curve of
temperature-programmed evolution of HCl [41] (cf. Fig. 7). One (at 125–300°C)
was ascribed to the LT regime, the other one (at about 500°C) to the HT regime

Fig. 7. a TGA and b thermal gas titration curves of a CaCl2 · 2H2O/H-MOR mixture (after [41],
with permission)
Solid-State Ion Exchange in Microporous and Mesoporous Materials 67

Table 3. Starting materials, mixtures and results of solid-state ion exchange in the system
NaCl/H-ZSM-5a. 1, parent zeolite; 2, nSi/nAl ratio of the parent zeolite; 3, Al content of the
parent zeolite; 4, NaCl admixed to 1 g (dry) zeolite; 5, HCl evolved on solid-state reaction;
6, Cl– extracted after solid-state reaction; 7, Na+ extracted after solid-state reaction; 8, Na+ irre-
versibly held in zeolite after extraction; 9, degree of exchange (%) (data in column 8 divided
by data in column 3)

1 2 3 4 5 6 7 8 9
Zeolite a nSi/nAl Al NaCl HCl Cl Na Na d
(employed) (evolved) (extracted) (extracted) (irrev. held) (%)

H-ZSM-5 (I) 155 0.107 0.808 0.549 0.260 0.707 0.101 94

H-ZSM-5 (II) 23 0.691 1.306 0.838 0.478 0.670 0.636 92
a All data in mmol g–1 zeolite fired at 1273 K.

of the solid-state reaction. It was difficult to reach completion of the high-tem-

perature exchange. Only after repeated heating and an isothermal reaction peri-
od at 600°C did the reaction cease. Titration of the evolved HCl showed that a
total of 2.52 mmol HCl per gram dry zeolite was released (cf. Sect. 5.1.5, Table 4).
At variance with the TPE results, the TGA curve of the system CaCl2 ◊ 2H2O/H-
ZSM-5 exhibited only one very steep step of weight loss of the sample, viz., in the
LT region. This resulted from the evolution of both water and HCl, as described
for NaCl/H-ZSM-5 (vide supra). However, above 200°C, the TGA curve continu-
ously declined and the HT regime of the SSIE could not be clearly discriminat-
ed. Obviously, the TGA measurement was markedly disturbed by the presence
and release of crystal water.

5.1.5
Stoichiometry of SSIE of M+ and M2+ Halides with H-Zeolites

From the analysis of (1) the starting zeolite material, (2) the gases evolved, (3)
the aqueous extracts obtained from the salt/zeolite mixtures after reaction, and
(4) the exchanged zeolite, one can determine the stoichiometry of the solid-state
ion exchange. This is exemplified by Tables 3 and 4, which provide data for the
systems NaCl/H-ZSM-5, CaCl2/H-MOR and CaCl2/NH4-MOR, given in mmol
per gram zeolite fired at 1273 K.
In the respective experiments, excess amounts of NaCl or CaCl2 were applied,
i.e., the employed millimoles of NaCl or milliequivalents of CaCl2 considerably
exceeded the Al-content of the starting zeolites (cf. columns 4 and 3 of Tables 3
and 4, respectively). However, the amounts of HCl (NH4Cl) evolved were also
markedly higher than the corresponding Al-contents (compare columns 5 and
3) and, consequently, than the amount of Brønsted acid sites per gram. This
means that a fraction of the salts has reacted with silanol groups. However, due
to hydrolysis of Na bound to silanol groups, i.e., Na+(O–Si∫)– species, and dis-
solution of the non-converted salt, the excess of NaCl could be extracted. From
the results of analyses presented in Table 3 it was derived that, within the limits
of error, the amounts of employed NaCl were equal to the sum of (1) evolved HCl
68 H.G. Karge · H.K. Beyer

Table 4. Starting materials, mixtures and results of solid-state ion exchange in the systems
CaCl2/H-MOR and CaCl2/NH4-MORa. 1, parent zeolite; 2, nSi/nAl ratio of the parent zeolite;
3, Al content of the parent zeolite; 4, CaCl2 admixed to 1 g (dry) zeolite; 5, HCl evolved on
solid-state reaction; 6, Cl– extracted after solid-state reaction; 7, Ca2+ extracted after solid-
state reaction; 8, Ca2+ irreversibly held in zeolite after extraction; 9, CaCl2 occluded upon
solid-state reaction; 10, degree of exchange (%) (data in column 8 (multiplied by 2 to give
mequiv) divided by data in column 3)
1 2 3 4 5 6 7 8 9 10
Zeolite a nSi/nAl Al CaCl2 HCl; NH4Cl Cl Ca Ca CaCl2 d
(employed) (evolved) (extracted) (extracted) (irrev. held) (occluded) (%)

H-MOR 6.7 2.18 1.98 2.52 1.44 0.89 1.09 – 100

NH4-MOR 5.4 2.50 2.48 2.54 1.88 0.95 1.26 0.27 100

a All data in mmol g–1 zeolite fired at 1273 K.

and extracted Cl– in the zeolite matrix (e.g., for H-ZSM-5(I): 0.549 + 0.260 =
0.809 mmol; cf. columns 4, 5, and 6) or (2) extracted NaCl and Na+ irreversibly
held in the zeolite matrix (e.g., for H-ZSM-5 (I): 0.707 + 0.101 = 0.808 mmol; cf.
columns 4, 7 and 8). Thus, the resulting content of (irreversibly held) Na+ in the
exchanged zeolite (0.101 mmol) was close to the amount of Al in the framework
(0.107 mmol; cf. columns 8 and 3 of Table 3). In other words, the degree of
exchange was almost 100% (cf. column 9). Similar results were obtained with the
sample H-ZSM-5 (II), as can be seen from Table 3.
The evaluation of the data for mordenites presented in Table 4 was similar to
that of the H-ZSM-5 samples; however, with the systems CaCl2/H-MOR and
CaCl2/NH4-MOR, the situation was somewhat more complicated. The H-MOR
sample contained considerable amounts of extra-framework Al (0.40 mmol per
gram dry zeolite), as was proven by IR after adsorption of pyridine and deter-
mined by 27Al MAS NMR. From a comparison of columns 3 and 8 it becomes
obvious that CaCl2 has reacted with the total Al (framework and non-framework
Al), since 1.09 mmol or 2.18 mequiv Ca2+ were irreversibly held. Most likely, the
extra-framework Al was connected to OH groups which possibly had been con-
verted according to Eq. (9):
2 AlOOH + CaCl2 Æ Ca(AlO2)2 + 2 HCl≠ (9)
Calcium aluminates are not readily soluble in water and were, therefore, not
extracted when the mixture was washed with water after the reaction had been
completed.
The NH4-MOR sample was essentially free of extra-framework Al. Thus, the
Al content given in column 3 of Table 4 exclusively corresponded to tetrahedral-
ly coordinated Al, as confirmed by 27Al MAS NMR. This amount (2.50 mmol Al)
was found to equal the amount of Ca2+ introduced via solid-state ion exchange,
i.e., 1.26 mmol or 2.52 mequiv of Ca2+ (cf. columns 3 and 8, Table 4). However,
careful analysis showed that a certain fraction of the admixed CaCl2 remained
occluded after completion of the reaction and subsequent washing with water
(cf. column 9). The amount of 0.27 mmol CaCl2 per gram dry zeolite corre-
sponded to about 50% of the side-pockets of the MOR structure where the CaCl2
Solid-State Ion Exchange in Microporous and Mesoporous Materials 69

molecules were most probably trapped. No such salt occlusion was observed
with H-ZSM-5 samples. Similar observations were made, however, when NaCl
was reacted with the NH4-MOR sample. Salt occlusion in zeolites is a well-
known phenomenon [16, 17, 53]. Rabo et al. [53] have shown that the thermal
stability of zeolite structures is frequently improved by occlusion of salts. Thus,
in many cases, it should be advantageous if ion exchange via solid-state reaction
is accompanied by occlusion of molecules of the salt present in the reaction mix-
tures (cf. also Sect. 5.2.3, incorporation of La3+).

5.1.6
Preservation of Crystallinity upon SSIE

Solid-state ion exchange is usually conducted at higher temperatures and, if

metal halides and hydrogen forms of zeolites are reacted, occurs in the presence
of hydrogen halides. Therefore, the question may arise as to whether the integri-
ty of the zeolite structure has been preserved when solid-state ion exchange was
carried out. There are several ways to check whether the crystallinity of the zeo-
lite material has deteriorated during the solid-state reaction.
A most suitable test is provided by X-ray diffraction (XRD). Generally, a
number of reflections are selected and the sum of their intensities prior to and
after the solid-state reaction compared. However, the absolute intensities fre-
quently change; they may increase or decrease depending on the contributions
of charge-compensating, in-going and out-going cations to the structure factors
of the reflections. Nevertheless, this method provides reliable results provided
these effects are adequately considered (cf. [54]). The XRD tests showed that in
the cases described above (Sect. 5.1) no loss of crystallinity had occurred.
Another convenient method is that of cation re-exchange. After introduction
of the desired metal cations, one can, for instance, in a small representative test
sample of the exchanged zeolite, re-exchange the metal cations by NH +4 through
repeated treatment in aqueous NH4-salt solution, and subsequently deammoni-
ate the sample, whereby the evolved NH3 may be determined via titration;
alternatively, the OH groups generated by deammoniation may be determined
via IR or 1H MAS NMR (cf. Sects. 5.1.8 and 5.1.10). If the exchange capacity is
the same after solid-state reaction as it was before, one can conclude that the
integrity of the structure has remained unaffected. However, the interpretation
is difficult if the exchange capacity has decreased, since a fraction of the cations
incorporated by SSIE may be irreversibly trapped and, at least under the con-
ditions of treatment with an aqueous solution, hindered to re-exchange. In such
a case, the solid zeolite must be chemically analyzed after the re-exchange ex-
periment.
29Si MAS NMR is also an appropriate tool to decide whether or not SSIE has

caused a (partial) collapse of the zeolite structure.If such a collapse has happened,
it would be accompanied by a loss of framework T-atoms. This would mean that,
with T = Si, Al, the framework nSi/nAl ratios [55] of the parent zeolite and the zeo-
lite after SSIE are likely to be different.Therefore,the unchanged nSi/nAl ratio of the
Y-type sample after introduction of, e.g., Cs+, confirmed the finding obtained by
XRD that no loss of crystallinity had occurred (cf. Sect. 5.1.10).
70 H.G. Karge · H.K. Beyer

If the structure of the zeolite is sensitive to mechanical stress imposed by

grinding or milling during the preparation of the salt (oxide)/zeolite mixture,
the crystallinity can be preserved by suspending the fine powders of salt (oxide)
and zeolite in an inert volatile liquid such as n-hexane (cf. Sect. 3).

5.1.7
Role of the Nature of the Anions

To date, no systematic study of the effect of the anion’s nature in the compound
of the in-going cation has been undertaken. Only in the system M+m (Am–)/H-
ZSM-5 (A = anion) has a series of experiments with various anions (A = F, Cl, Br,
J, OH, NO3, CO3) been carried out [22]. It turned out that solid-state ion exchange
proceeded most easily with the chlorides, whereas with fluorides and bromides
the reaction was slower and frequently incomplete. Mixtures of iodides of alka-
line metals and H-ZSM-5 turned yellow simply on grinding; obviously, the
iodides decomposed in the presence of the hydrogen form of the zeolite. Cesium
hydroxide was employed for introduction of Cs+ into Y-type zeolites [36]. Con-
cerning other salts, only a limited number of observations have been reported.
For instance, MgF2 mixed with H-MOR reacted only to an exchange degree of
about 40%, whereas with MgCl2 an almost 100% exchange was achieved [41].
Reactions of salts with complex anions such as NO –3 and SO42– are more compli-
cated. The anions were more or less decomposed (cf. Sect. 5.3.4). In the case of
carbonates no solid-state reaction was observed [22]. Most probably, the car-
bonates decomposed at the elevated temperatures applied (M n+ n+
2/nCO3 Æ M 2/nO +
CO2 , n = 1, 2), and the oxide was not or only slightly reactive. Solid-state reac-
tions of complex oxo- or chloro-anions (e.g., chromates, vanadates, chloro-
molybdates) with zeolites will be discussed in subsequent sections (cf., e.g.,
Sect. 5.3.5).

5.1.8
SSIE of M+ and M2+ Halides with H-Zeolites Investigated by IR

Solid-state ion exchange between alkali halides and hydroxyl groups of the
hydrogen form of zeolites was also monitored by infrared spectroscopy.An early
example is taken from the study by Rabo et al. [16, 17] mentioned earlier and is
reproduced in Fig. 8.
A mixture of Ca,H-Y and NaCl was calcined in air. IR spectra of the mixture
prior to and after thermal treatment proved the elimination of the hydroxyl
groups (disappearance of the OH stretching bands at 3745, 3690 and 3640 cm–1)
under evolution of HCl. This resulted in a complete elimination of the catalytic
activity of the zeolite sample in the acid-catalyzed isomerization of 1-butene to
2-butene.
Similar to the approach of Rabo, Jiang and Tatsumi [56] used SSIE with KCl
to eliminate acid centers in K-Y and K-L loaded with Mo3S4 clusters, which had
been generated upon activation of the zeolite/Mo3S4 catalysts.After SSIE, the cat-
alysts produced greater amounts of alcohols in the hydrogenation of CO, due to
the decreased acidity.
Solid-State Ion Exchange in Microporous and Mesoporous Materials 71

Fig. 8. Solid-state ion exchange between NaCl and Ca,H-Y. OH spectra of Ca,H-Y (A, solid
line) and NaCl/Ca,H-Y (B, broken line), both after calcination in air at 825 K for 48 h (after
[16], with permission)

In systematic investigations of solid-state ion exchange in the systems MCl/

H-ZSM-5, NH4-ZSM-5 and MCl/NH4-Y (cf. [22, 43]), the decrease in the intensi-
ty of the IR bands that are indicative of hydroxyl groups was also used to deter-
mine solid-state ion exchange. From Fig. 9 it can be seen that not only the acidic
OH groups (Brønsted acid sites, as indicated by the band at 3605 cm–1) have
reacted with NaCl but that, to some extent, the weakly acidic silanol groups
(band at 3740 cm–1) were also involved as well (cf. also Fig. 8). However, since the
silanol groups possess only a low strength of acidity [57], brief washing with
deionized water and subsequent drying restored the original intensity at
3740 cm–1. The intensity at 3605 cm–1 re-appeared only after re-exchange with
NH4Cl solution, subsequent dehydration and deammoniation in the IR cell.
Simultaneously, this was a test for the integrity of the zeolite framework: the re-
appearance of the bands of the Brønsted acid sites showed that no (partial) col-
lapse of the lattice had occurred during the solid-state reaction. Otherwise, at
least part of the tetrahedrally coordinated aluminum would have left the frame-
work and the density of the Brønsted centers would have decreased accordingly.
Usually, reaction of MCl with H-ZSM-5, NH4-ZSM-5 or NH4-Y leads to a
complete disappearance of the bands of the Brønsted acid sites, thus proving
an exchange degree of 100%. An exception was observed with the system
CsCl/NH4-Y.However,application of a slight excess of CsCl and a second exchange
step also produced a 100% replacement of (acidic) protons by Cs [58] (Fig. 10).
In aqueous solution only less than 70% of the original Na+ cations could be
exchanged. Thus, the 100% degree of exchange achieved by solid-state reaction
is an interesting result in that a high degree of exchange is required in order to
increase the basicity of zeolites as catalysts for base-catalyzed reactions.
Weitkamp et al. [36], therefore, employed SSIE of cesium salts with Y-type zeo-
lites for the preparation of base zeolite catalysts. As mentioned above, conven-
72 H.G. Karge · H.K. Beyer

Fig. 9. IR spectra of the hydroxyl stretching frequency region after degassing at 723 K for 2 h
(final vacuum about 10–3 Pa). a H-ZSM-5; b NaCl/H-ZSM-5 mixture (nNaCl/nAl = 1.89) calcined
at 900 K for 1 h; c (b) washed with water; d (b) twice exchanged with 1 N NH4Cl solution (after
[22], with permission)

Fig. 10. IR spectra of the hydroxyl stretching frequency region after degassing of a NH4-Y at
725 K for 12 h (final vacuum about 10–5 Pa) and b CsCl/NH4-Y (nCs/nAl = 1.1) at 725 K for 20 h
(ex situ SSIE) and subsequently for 2 h in high vacuum (in situ SSIE); degree of exchange in
NH4-Y was 98%
Solid-State Ion Exchange in Microporous and Mesoporous Materials 73

tional exchange of cesium into Y-type zeolites does not yield a very high degree
of exchange. This is due to the large diameter of the Cs+ cation (0.37 nm) that
does not allow penetration of the six-membered rings (0.22 nm) leading to the
small cavities of the structure. Solid-state ion exchange, however, is possible
because of the higher temperatures that can be applied during the process and
at which the six-membered rings become more flexible (cf. also [53]). The prod-
ucts prepared by Weitkamp et al. via SSIE with KCl and RbCl were characterized
by TGA, IR, XPS, 27Al MAS NMR, 29Si MAS NMR, and, particularly for Cs, by 133Cs
MAS NMR, as well as by conversion of isopropanol and methanol. A 100%
exchange in the case of Rb-Y and an almost 100% exchange in the case of Cs-Y
could only be achieved by SSIE. The catalytic behavior of the materials prepared
in this way was tentatively explained by an increasing basicity of the oxygen
atoms of the framework related to the decreasing electronegativity of the alka-
line metal cations.
In a study by Xu et al. [50], introduction of Cs+ via solid-state ion exchange
was monitored by XRD and TGA/DTA. The Cs-BETA catalyst easily obtained by
SSIE was found to have a rather high degree of exchange, viz., 94%, and possess
a high activity in base-catalyzed dehydrogenation of isopropanol (vide supra).
Systematic studies of the solid-state ion exchange of alkaline halides with
NH4-Y were more recently resumed by Jiang et al. [59]. These authors also
observed a significantly higher degree of exchange with Rb+ and Cs+ via SSIE
than via exchange in aqueous solution. The degree of exchange (by 1-fold solid-
state reaction) decreased, however, in the sequence K+ >Rb+ >Cs due to geomet-
ric constraints.
FT-Raman spectroscopy was employed and proved to be a useful tool in
investigating solid-state ion exchange when Huang et al. [33] reacted LiCl or
CaCl2 with NH4-Y or Na-Y to produce Li-Y or Ca-Y. The results were verified by
XRD.
Series of alkaline-metal-containing M+-X and M+-Y zeolites (M+ = Na+, K+,
Rb+, Cs+) for detailed IR spectroscopic investigations were prepared by Ese-
mann and Förster [60, 61], Esemann et al. [62] and Geidel [63]. Solid-state reac-
tion between the ammonium forms of the zeolites and MCl proved to be most
efficient and provided a high degree of exchange. From the experiments it was
concluded that the solid-state reaction occurred to a large extent prior to deam-
moniation. Since the introduction of cations with a larger diameter caused an
expansion of the lattice and a lowering of the bond strength, above 1000 cm–1 a
frequency shift of the valence vibrations to lower values was observed (cf.
Fig. 11). Also, spectra in the far-infrared region (50–300 cm–1) characteristic of
M-X and M-Y zeolites were reported [63].
IR studies were also carried out with mixtures of alkaline earth salts and
hydrogen forms of zeolites. As an example, IR spectra of H-MOR prior to and
after solid-state reaction with CaCl2 are shown in Fig. 12 (cf. [41]).
Comparison of spectra 1a and 2a provided evidence for the complete
exchange of the protons of the acidic OH groups as indicated by the disappear-
ance of the band at 3610 cm–1. The decrease in the intensity of the band at
3750 cm–1 shows also that a fraction of the silanol groups have reacted with
CaCl2 . Pyridine adsorption confirmed that, after solid-state ion exchange of
74 H.G. Karge · H.K. Beyer

Fig. 11. Spectra in far-infrared of M-Y zeolites with M = Na (a), K (b), Rb (c), and Cs (d)
obtained by solid-state ion exchange in the IR cell in a flow of N2 at 600 K (12 h) from mixtures
of the respective salts and NH4-Y (after [63], with permission)

CaCl2 and H-MOR, there were no longer any acidic Brønsted sites present, since
no band around 1540 cm–1 (which is typical of pyridinium ions and would form
in the presence of such acid sites) is seen. Instead, a band at 1446 cm–1 appeared
in spectrum 2b, indicating pyridine coordinatively attached to Ca2+ and thus
proving the incorporation of these cations into the mordenite structure.
A second sample (wafer of a mixture of CaCl2 ◊ 2H2O and H-MOR) was heat-
ed in the IR cell at 775 K to obtain a spectrum like 2a in Fig. 12. Subsequently, the
wafer was briefly contacted with 1.3 kPa H2O vapor at 400 K. After pumping off
the excess water, spectrum 3a was registered. It exhibited an intense band at
3618 cm–1 which originated from acidic OH groups similar to but not identical
with those of the parent sample of H-MOR (spectrum 1a). After pyridine
adsorption and degassing, spectrum 3b was observed which contained an
intense band typical of pyridinium ions (1540 cm–1). It resulted from the reac-
tion of pyridine with acidic Brønsted centers that were generated through the
Hirschler-Plank mechanism [64, 65]:
Ca2+Z–2 + H2O Æ H+Z– + Ca(OH)+Z– (10)
where Z– is a monovalent negatively charged framework fragment of the zeolite
structure.
Additionally, spectrum 3b in Fig. 12 displayed an intense band at 1446 cm–1
due to pyridine interacting with Ca(OH)+, resulting in a coordination complex
with pyridine, and a band at 1455 cm–1, typical of pyridine bound to ‘true’ Lewis
sites (Al-containing extra-framework species).
Very similar results were reported for the system MgCl2/H-MOR [41]. Elimi-
nation of (most of) the band arising from OH groups of H-MOR, indication
of Mg2+ incorporation via pyridine adsorption (resulting in a band around
1448 cm–1), and rehydroxylation (re-appearance of OH groups and formation of
Solid-State Ion Exchange in Microporous and Mesoporous Materials 75

Fig. 12. IR spectra of H-MOR and the system CaCl2 ◊ 2H2O/H-MOR with nCa/nAl = 0.5, after
thermal treatment at 775 K and 10–5 Pa. Spectra 1a-3a in the OH stretching frequency region,
1b-3b in the region of pyridine ring deformation frequencies. 1a, 1b without and 2a, 2b with
admixed CaCl2 ◊ 2H2O; 2b after pyridine adsorption and 3a after rehydroxylation, 3b after
rehydroxylation and pyridine adsorption (after [41], with permission)

pyridinium ions) were found. It should be mentioned, however, that solid-state

reaction in the system MgCl2/H-ZSM-5 did not proceed as rapidly as in the case
of MgCl2/H-MOR. Moreover, in the former system, the replacement of protons
by Mg2+ cations was frequently incomplete.
Li et al. [66] reported on Mg2+ incorporation into H-ZSM-5 through SSIE as
an effective and convenient route for modifying the zeolite. Samples with vari-
ous Mg2+ contents were prepared. IR spectroscopy as well as XRD and TPD in-
dicated that their structure and crystallinity were not changed in comparison
with the parent H-ZSM-5. However, the concentration of Brønsted acid sites
decreased with increasing Mg2+ content, whereas the density of Lewis sites was
slightly enhanced and the catalytic behavior in alkylation with methanol signif-
icantly improved.

5.1.9
SSIE of M+ and M2+ Halides with Na-Zeolites Investigated by IR

Solid-state ion exchange with the sodium form of zeolites can be investigated by
IR spectroscopy using probe molecules such as pyridine. In principle, this type
of SSIE leads to an equilibrium (cf. Scheme 1b).
Solid-state ion exchange between alkaline metal salts and a sodium form of
zeolite X (13X) was investigated by Yang and Xu [67–69] using XRD and IR. The
76 H.G. Karge · H.K. Beyer

behavior of the products was compared with that of samples obtained via
exchange in aqueous solution: In contrast to the case of MAPO-36 (cf. [70], vide
infra), the surface area, pore volume and catalytic activity in isopropanol con-
version were found to be independent of the exchange method used for prepa-
ration of the samples. The properties of the materials were only determined by
the degree of exchange. Samples exchanged with Cs+ exhibited higher basicities
and dehydrogenation activities than K+-exchanged catalysts with the same
degree of exchange. However, it was more difficult to replace Na+ by the bulky
Cs+ than by K+ cations.
Figure 13 shows, as an example for the solid-state reaction with a sodium
form of a zeolite, the spectrum of a wafer made from a BeCl2/Na-Y mixture,
heated in an ultra-high vacuum (p = 10–5 Pa) at 400 K and subsequently con-
tacted with pyridine vapor (spectrum b).
An intense band at 1453 cm–1 indicated Be2+ cations on extra-framework sites
[71]. No band at 1444 cm–1, typical of pyridine attached to Na+ [32], was
observed. However, when the mixture of BeCl2 and Na-Y was heated to 725 K, the
band at 1453 cm–1 decreased in intensity and the Py Æ Na+ band at 1444 cm–1
developed. At higher temperatures, the equilibrium (cf. Scheme 1b) is most like-
ly shifted: a fraction of Na+ remigrates from the NaCl crystallites formed (vide
infra) (cf. Sect. 5.1.10) to cation sites, thereby replacing Be2+ cations. Similar
observations were made in the system LaCl3/Na-Y (see Sects. 5.2.6 and 5.2.7).
Akolekar and Bhargava [70] succeeded in modifying MAPO-36 by introduc-
tion of Na+ via (1) conventional and (2) solid-state ion exchange and character-

Fig. 13. IR spectra after degassing at 400 K and subsequent pyridine adsorption: a Na-Y; b the
mixture BeCl2/Na-Y; c sample of spectrum (b) heated at 725 K (after [73], with permission)
Solid-State Ion Exchange in Microporous and Mesoporous Materials 77

ized the products by XRD, SEM, FTIR, TGA, BET, 27Al MAS NMR and 31P MAS
NMR. Samples obtained by method (1) exhibited the same crystallite morphol-
ogy and surface area as the parent MAPO-36, whereas these properties were sig-
nificantly affected by method (2). Moreover, the products of both preparative
procedures differed in the 27Al MAS NMR and 31P MAS NMR spectra. Neverthe-
less, the catalytic activities of the Na+-exchanged materials in o-xylene conver-
sion were similar.
Finally, the importance of SSIE with respect to an interaction of zeolite
matrices and clays used as binders in catalyst formulations has been stressed
several times. Thus, Canizares et al. [72] showed that SSIE between Na-mont-
morillonite (binder material) and H-ZSM-5 or H-mordenite is the reason for
decreased Brønsted acidity in the zeolite matrices as compared to the unbound
zeolite.

5.1.10
SSIE of M+ and M2+ Halides with Na-Zeolites Investigated by MAS NMR

Ground mixtures of LiCl, KCl, BeCl2 or CaCl2 with Na-Y were investigated
by 23Na MAS NMR [73]. The signals were referenced to crystalline NaCl.
Na+ cations residing in the supercages of the Y-structure give rise to a signal
at about –9 ppm, whereas a signal at about –13 ppm is assigned to Na+ cations
in the small cavities [74]. (This assignment is at some variance to the assign-
ment suggested earlier [73].) Regarding solid-state ion exchange at low
temperature (LT regime) or upon mere grinding (“contact-induced ion ex-
change”, cf. Sect. 6.2) of the mixtures mentioned above, the most striking
feature is, however, that the signal at ca. –8 ppm of the parent zeolite disap-
peared and a sharp signal at 0 ppm developed instead (Fig. 14). The expla-
nation is that the Na+ cations in the supercages were, to a large extent, replaced
by Li+, K+, Be2+ or Ca2+ and a fraction of the expelled sodium cations formed
with Cl– anions of the admixed salts tiny NaCl crystallites outside the zeolite
grains.
A broad signal around –13 ppm was left in the spectrum, mainly due to Na+
cations in the small cavities. It cannot be excluded, however, that a small fraction
of the Na+ cations are still located in the supercages contributing to the intensi-
ty of the –13 ppm signal at its high-field wing. In any event, the intensity of this
signal appears enhanced which is most likely caused by a migration during LT
ion exchange of a fraction of Na+ cations from the supercages into the small cav-
ities. Thus, these experiments have at least qualitatively proven that solid-state
ion exchange has taken place in the above systems. The respective measure-
ments can probably be improved to become suitable for quantitative determina-
tions as well.
Weitkamp et al. [36], who prepared a series of alkaline metal zeolites (K-Y, Rb-
Y, Cs-Y) via solid-state ion exchange in order to study the effect of the basicity
of such zeolites on their catalytic behavior, used 133Cs MAS NMR for the charac-
terization of Cs-Y. Figure 15 shows (a) the spectrum of Cs-Y, (b) the simulated
spectrum and (c) the individual components derived from a decomposition of
spectrum (b).
78 H.G. Karge · H.K. Beyer

Fig. 14. 23Na MAS NMR spectra of Na-Y and mixtures of halides and Na-Y (see text; after [73],
with permission)

Fig. 15. 133Cs MAS NMR spectrum a of Cs-Y prepared by solid-state ion exchange in a
CsOH/NH4-Y mixture, b after simulation and c after decomposition of (b) into individual
components (after [36], with permission)
Solid-State Ion Exchange in Microporous and Mesoporous Materials 79

Table 5. Cs+ population of the cation sites in zeolite Cs-Y prepared by SSIE; results of 133Cs

MAS NMR measurements (after [36])

Site I¢ III II Cs(OH)/b-cage

Chemical shift a (ppm) –78 –66 –61 +57

nCs+/u.c. 9 6 17 21
a Referenced vs. 1 M aqueous CsCl solution.

Four signals can be distinguished (referenced to 1 M aqueous CsCl solution),

viz., at –78, –66, –61 and +57 ppm. The authors were able to assign, on the basis
of earlier work [75], the lines at –78, –66, and –61 ppm to cesium cations on sites
I¢ (in the sodalite cages), sites III and sites II (both in the supercages), respec-
tively. The signal at +57 ppm was tentatively attributed to Cs(OH) ◊ H2O, since
the solid-state ion exchange was carried out with Cs(OH) · H2O and NH4-Y. The
authors were able to derive from the 133Cs MAS NMR results the population of
the respective sites by Cs+ and Cs(OH) · H2O. The data are presented in Table 5.
Evaluation of the 29Si MAS NMR and 27Al MAS NMR spectra proved that the
integrity of the zeolite framework had been preserved.

5.1.11
Catalytic Activity of M2+-Zeolites Prepared via SSIE
The catalytic activity of Ca,H-MOR and Mg,H-MOR obtained via solid-state ion
exchange was tested using the disproportionation of ethylbenzene as a test reac-
tion [41, 76]. Expectedly, the modification of H-MOR by incorporation of Ca2+
or Mg2+ (nCa2+/nAl = 0.50 and nMg2+/nAl = 0.65 at 675 and 875 K, respectively; cf.
Sects. 5.1.5 and 5.1.8) considerably decreased the activity (by ca. 90 and 77%,
respectively). The activity was not completely eliminated, since residual acidic
OH groups were, after calcination of the MCl2/H-MOR mixtures, still present
(vide supra). For example, as evidenced by in situ IR, 30% of the original Brøn-
sted acid OH groups survived the reaction with MgCl2 . Moreover, interaction
with H2O vapor of the ambient air was not excluded in the catalytic experiments
following the solid-state ion exchange. However, subsequent rehydroxylation
by brief contact with admitted H2O vapor (vide supra) increased the catalytic
activity of Ca,H-MOR and Mg,H-MOR obtained via SSIE by about 50%. Thus,
in principle, solid-state ion exchange offers an interesting route for preparation
of catalytically active zeolites (see also Sects. 5.2.8, 5.4.3 and 5.4.4).

5.2
SSIE of Lanthanum (La3+) Cations

5.2.1
Introductory Remarks
Lanthanum, or virtually all rare earth cations, are important constituents of cat-
alysts employed in cracking of vacuum distillates from petroleum [77]. Thus, it
seemed interesting to explore the possibility of introducing La3+ cations into
80 H.G. Karge · H.K. Beyer

zeolites by solid-state ion exchange. Studies were preferentially conducted with

hydrogen (ammonium) or sodium forms of Y-type zeolites, since this zeolite is
widely employed in cracking catalysis. The solid-state reaction was monitored
by chemical analysis (CA), temperature-programmed evolution of gases (TPE),
infrared (IR) spectroscopy, X-ray diffraction (XRD), and 23Na MAS NMR [78,
79]. More recently, Jia et al. [80] compared in an extended study the incorpora-
tion of lanthanum via conventional and solid-state ion exchange into H-BETA
zeolite and investigated the effect of many parameters on this example of SSIE.
These authors employed XRD, BET, FTIR TEM, EDS (energy dispersive spec-
troscopy) and chemical analysis for characterization of the products of ion
exchange. The effect of exclusion of water vapor and deliberate hydroxylation
was particularly investigated with mixtures of NH4-Y, H-MOR, H-FER with
LaCl3 ◊ 7H2O or water-free LaCl3 (cf. also Sect. 7). The catalytic activity of La,
H-Y obtained by solid-state reaction was tested in the disproportionation of eth-
ylbenzene and cracking of n-decane.

5.2.2
SSIE of La3+ Chloride with H-Zeolites Investigated by TPE

Figure 16 displays the TPE profiles [78] obtained during temperature-pro-

grammed heating of a finely dispersed sample of the parent zeolite, NH4-Y, (for
comparison; m/e = 16, deammoniation; m/e = 18, dehydroxylation) and a mix-
ture of LaCl3 · 7H2O with the ammonium (hydrogen) form of zeolite Y (for solid-
state reaction; m/e = 16, deammoniation; m/e = 18, dehydration; m/e = 36, evo-
lution of HCl).

Fig. 16. Temperature-programmed evolution of H2O (m/e = 18) from NH4-Y and NH2+ (m/e =
16), HCl (m/e = 36) and H2O (m/e = 18) from a LaCl3/NH4-Y mixture (after [78], with per-
mission)
Solid-State Ion Exchange in Microporous and Mesoporous Materials 81

Similar to the case of the systems where the exchange occurred with alkaline
salts (cf. Sect. 5.1), one can also distinguish between a low-temperature (LT) and
a high-temperature (HT) regime (see the curve describing the evolution of HCl,
m/e = 36). Comparison of the profiles obtained with NH4-Y (broken line) and
LaCl3/NH4-Y (solid lines) shows that, in the latter case, no peak occurs in the
region around 950 K [78]. With pure NH4-Y, this well-developed peak originates
from the dehydroxylation of NH4-Y or, more precisely, of H-Y, since the NH4-Y
sample is deammoniated already at lower temperatures (cf. broken line; NH4-Y,
m/e = 16). In the case of the system LaCl3/NH4-Y, such a dehydroxylation cannot
occur because the OH groups generated via deammoniation in the LT region are
consumed by the solid-state reaction at temperatures below 950 K, i.e., at about
650 to 900 K and are no longer available for dehydroxylation. Both these find-
ings, i.e., the evolution of ammonia and hydrogen chloride as well as the absence
of the dehydroxylation peak, proved that NH+4 cations from the NH+4 -zeolite and,
at higher temperatures, H+ ions from the deammoniated NH+4 -zeolite, were
exchanged for La3+ cations of solid LaCl3 . NH4Cl, which may have formed inter-
mittently, would be thermally decomposed into NH3 and HCl. Combined TPE
and TGA experiments (cf. Sect. 5.1.4) could not be carried out because of the
simultaneous release of HCl and NH3 .
However, titration was possible with the system LaCl3 ◊ 7H2O/H-ZSM-5 (cf.
Fig. 17). The curves a (nLa/nAl = 0.33) and b (nLa/nAl = 0.67) reveal that in a low-

Fig. 17. Temperature-programmed evolution of HCl monitored via continuous titration dur-
ing solid-state ion exchange between LaCl3 · 7H2O and H-ZSM-5 with nLa/nAl ratios of 0.33
(curve a) and 0.67 (curve b) and isothermal steps at 945 K for 2 and 0.5 h, respectively; curve
c represents the behavior of similarly treated LaCl3 · 7H2O
82 H.G. Karge · H.K. Beyer

temperature process up to ca. 570 K only a minor fraction of the hydrogen form
of H-ZSM-5 was converted and the major part of exchange took place in a high-
temperature process between 675 and 950 K, where the sample was kept for 2 h.
Curve c describes the behavior of pure LaCl3 ◊ 7H2O. Comparison of curves a, b
and c proves that the evolution of HCl is not simply caused by decomposition of
the lanthanum salt, but is indeed the result of an interaction of the two solid
components of the mixture, i.e., the salt and the hydrogen form of ZSM-5. The
exchange was more rapid and led to a higher degree of exchange when an excess
of LaCl3 ◊ 7H2O was employed. However, in both cases (a and b), the reaction pro-
duced on isothermal treatment an exchange degree of less than 100%. This was
at variance with the observations on LaCl3 ◊ 7H2O/NH4-Y (cf. Sect. 5.2.3). Most
likely, a 100% exchange is not achieved with H-ZSM-5 because of the difficulty
to electrically neutralize the more distant exchange sites in ZSM-5 by a trivalent
(naked) cation such as La3+ compared to Y-type zeolites where the exchange
sites are closer to each other due to the higher Al content. The difference in the
behavior of ZSM-5-type and Y-type zeolites in solid-state ion exchange with
La3+ chloride was confirmed by IR investigations (cf. Sect. 5.2.4).

5.2.3
Stoichiometry of SSIE in La3+ Chloride with H-Zeolites

The stoichiometric measurements [79] in the system LaCl3 · 7H2O/NH4,Na-Y via

chemical analyses prior to and after solid-state reaction were very instructive
(cf. Table 6). The data in Table 6 are the results of chemical analyses obtained for
a NH4,Na-Y zeolite with a 89% replacement of Na+ by NH+4 cations. Thus, the
first mixture prepared for SSIE of the NH4+ cations contained a small excess
of La3+, since its composition corresponded to one La3+ per three Al (nAl/nLa =
0.33), but 11% of the charges generated by the framework Al were compensated
by Na+. The experimental results were obtained by titration of the evolved
hydrogen chloride and chemical analysis of both the product after extraction
with water and the extract solution. The data presented in Table 6 proved an
excellent stoichiometry: within the limits of error, the amount of La3+ intro-
duced via solid-state reaction and irreversibly held on washing with water
(4.80 mequiv per gram) corresponded exactly to the amount of framework alu-
minum (4.83 mequiv per gram) or the maximum of bridging OH groups. How-
ever, even half of the Na+ cations of the starting material (1.61 mequiv per gram)
were removed from the zeolite. Only about 0.7 mequiv Na+ and 0.8 mequiv Cl–
per gram remained in the structure. This would correspond to about one NaCl
molecule per b-cage. It seems likely that this amount of NaCl is occluded in the
structure and would, according to Rabo’s work [16, 17, 53], enhance the thermal
stability of the exchanged zeolite (cf. also Sect. 5.1.5).
Essentially the same results were obtained when a higher excess of La3+ was
applied. Table 7 presents the data for a ratio nLa/nAl = 0.67. The only difference
was that, after completion of the reaction, a higher amount of LaCl3 was found
in the washing water, viz., the total excess was extracted by the washing water
and again only an amount of La3+ corresponding to the Al content of the frame-
work was irreversibly held.
Solid-State Ion Exchange in Microporous and Mesoporous Materials 83

Table 6. Stoichiometry of solid-state ion exchange in the system LaCl3/NH4-Ya; nLa/nAl = 0.33;
heat-treatment at 850 K

La3+ Cl– Na+ NH+4 Al

Parent zeolite – – 1.61 3.29 4.83

Admixed 1.61 4.83 – – –
Evolved as NH4Cl/HCl – 3.29 – 3.29 –
Extracted with water 0.06 0.72 0.94 – –
Irreversibly held
(mmol g–1) 1.60 0.82 0.67 – 4.83
(mequiv g–1) 4.80 0.82 0.67 – 4.83
a All data given in mmol per gram, except data of last line (mequiv per gram).

Table 7. Stoichiometry of solid-state ion exchange in the system LaCl3/NH4-Y a; nLa/nAl = 0.67;
heat-treatment at 850 K

La3+ Cl– Na+ NH+4 Al

Parent zeolite – – 1.61 3.29 4.83

Admixed 3.22 9.65 – – –
Evolved as NH4Cl/HCl – 3.29 – 3.29 –
Extracted with water 1.57 5.51 0.94 – –
Irreversibly held
(mmol g–1) 1.65 0.85 0.67 – 4.83
(mequiv g–1) 4.95 0.85 0.67 – 4.83
a All data given in mmol per gram, except data of last line (mequiv per gram).

With SSIE in the system LaCl3 ◊ 7H2O/H-BETA, however, it appeared that not
all of the incorporated lanthanum cations were involved in balancing the charge
of the framework [80]. This was mainly inferred from the FTIR data (vide infra)
even for those exchange experiments at 773 K, where 100% of the admixed La-
salt was consumed and a ratio of nLa/nAl = 0.33 measured. This finding was large-
ly ascribed to the reaction of LaCl3 with part of the silanol groups and extra-
framework Al(OH) species. Moreover, it was pointed out that, due to the high
atomic ratio nSi/nAl = 15, the distances between the negatively charged exchange
sites were generally rather large. Thus, usually one La3+ will not be able to simul-
taneously balance the negative charges of three exchange sites close to three
framework Al atoms (cf. Sect. 5.2.2, SSIE of LaCl3 and H-ZSM-5). However, it was
found that, at variance with conventional ion exchange, solid-state reaction
between lanthanum chloride and H-BETA led to a higher degree of exchange
and a much more homogeneous distribution of the La cations.

5.2.4
SSIE of La3+ Chloride with H-Zeolites Investigated by IR
Further evidence for the solid-state ion exchange with La3+ was provided by IR
[78]. An IR transmittent wafer made from a stoichiometric mixture of NH4-Y
84 H.G. Karge · H.K. Beyer

Fig. 18. IR spectra of La,Na-Y obtained by solid-state ion exchange (set S1) and conventional
ion exchange in aqueous solution (set S2) (see text; after [78], with permission)

(more precisely, a NH4,Na-Y sample with 75% of the original Na+ cations
exchanged by NH4+) and LaCl3 ◊ 7H2O was heated at 673 K. The resulting spec-
trum is shown as spectrum a in Fig. 18.
Essentially none of the OH bands typical of H-Y at 3640 and 3550 cm–1 [8]
occurred which showed that all of the available protons of the Brønsted acid sites
were replaced by La3+. However, contact of the heat-treated wafer with water
vapor (0.65 Pa at 643 K) and subsequent degassing at temperatures increasing
from 523 to 773 K provided spectra b (set 1) in Fig. 18. These exhibited two
prominent bands in the OH stretching region and were very similar to those of
conventionally exchanged samples, the spectra of which are displayed in series
c (set 2). Also, the ratios of the intensities of the low-frequency bands and high-
frequency bands at 3645 and 3535 cm–1, respectively, were almost identical for
sets 1 and 2. The absolute intensities did, in fact, differ. This, however, can be
ascribed to the difference in the conditions of dehydration: the spectra of set 1
were run after treatment in high vacuum (10–5 Pa), whereas those of set 2 were
run in a flow of dry nitrogen.
Later, exchange experiments were carried out starting with NH4-Y materials
with a degree of exchange of almost 100% which can be achieved by repeated
exchange in aqueous ammonium salt solutions. When these starting materials
were subjected to solid-state ion exchange, a 100% replacement of NH4+ (or H+)
by La3+ was achieved in one step (cf. Sect. 5.2.8).
In contrast to the above findings with the ammonium form of faujasite-type
zeolites with a regular low nSi/nAl ratio of 2.5, exchange experiments with H-
ZSM-5 or NH4-ZSM-5 (nSi/nAl >15) and LaCl3 ◊ 7H2O led to incomplete removal
of the IR band at 3605 cm–1 which indicates the acidic OH groups. Thus, the IR
investigations confirmed the observations described in Sect. 5.2.2.
Solid-State Ion Exchange in Microporous and Mesoporous Materials 85

Lanthanum introduction into H-BETA by SSIE was further studied by Jia et

al. [80] using FTIR as well as XRD, BET, TEM, EDS and AAS. The results were
compared with those obtained by conventional exchange in aqueous solution. In
contrast to conventional exchange, a 100% incorporation of La3+ was achieved
by SSIE. However, on the one hand, the acidic OH groups (band at 3610 cm–1)
were not completely removed, not even in the case of a salt excess. On the other
hand, it was seen that LaCl3 also reacted with hydrogen bonded and isolated
silanol groups (weakening of the bands at 3700–3700 and 3743 cm–1, respective-
ly) and with OH groups on Al totally or partly disconnected from the framework
(disappearance of the bands at 3782 and 3660 cm–1). Moreover, the resulting
materials were highly homogeneous and did not show any loss of crystallinity.
Far infrared spectroscopy was employed by Esemann and Förster [81] as a pow-
erful tool to monitor the incorporation of cations such as La3+ into Y-type zeolites
upon solid-state reaction of LaCl3 with H-Y or NH4-Y. This method provided
insight into the distribution of the La3+ cations on cation sites of the structure and
was also applied on SSIE with alkaline, alkali metal and zinc chlorides (vide infra).
Difallah and Ginoux [82] demonstrated that La-Y prepared from NH4-Y via
SSIE with LaCl3 exhibited an increased capacity for CO adsorption.
Solid rare earth chlorides (CeCl3 , NdCl3 , SmCl3 , EuCl3 , or YbCl3) were react-
ed with NH4-Y (H-Y) with nSi/nAl ratios of 2.6, 12.5 and 28 [83]. The activity of
the exchanged materials for catalytic liquid-phase oxidation of cyclohexane
depended on the nSi/nAl ratio, type of cation (CeªYb>Sm>Eu>Nd) and the
reaction temperature.
A remarkably active Ce-ZSM-5 catalyst for reduction of NOx in diesel exhaust
was produced by van Kooten et al. [84] via SSIE. Ce-ZSM-5 did not convert nitric
oxide or ammonia into the greenhouse gas nitrous oxide. At 770 K and a GHSV
(gaseous hourly space velocity) of 50,000 h–1 a conversion degree of 70% NOx
could be reached.
The rare earth cation Eu3+ was incorporated into Y-zeolite by solid-state reac-
tion between EuCl3 and NH4-Y [85]. At constant temperature, the degree of
exchange depended on the reaction time. A 94% exchange was easily achieved
after 4 h. The crystallinity as determined by XRD was little affected. Migration
and location of the Eu3+ cations in the structure were discussed on the basis of
spectroscopic data.

5.2.5
SSIE of La3+ Chloride with Na-Zeolites Investigated by XRD

Since most zeolites are synthesized in the sodium form, it was interesting to
study also the solid-state reaction between LaCl3 and the sodium form of, e.g.,
Y-type zeolite [35, 74, 79]. It was expected that such a reaction should occur
because it was observed that residual Na+ in NH4 ,Na-Y had indeed been
replaced by La3+ (cf. Sect. 5.2.3). However, in view of Scheme 1b, it appeared like-
ly that such a reaction would lead to an equilibrium (cf. Sects. 5.1.9 and 5.1.10)
rather than to a 100% exchange, because it would not be possible to remove con-
tinuously one product component from the reacting mixture, as had been
accomplished in the case of M+, M2+, or M3+ halides admixed to H- or NH4-zeo-
86 H.G. Karge · H.K. Beyer

Fig. 19. Schematic representation of XRD patterns of a Na-Y, b LaCl3/Na-Y after SSIE, c La,
Na-Y obtained by CE, d (b) heated at 850 K, and e (c) heated at 850 K (after [79], with permission)

lites. Indeed, it was proven by IR, XRD and 23Na MAS NMR that it is possible to
replace Na+ in Na-Y by La3+ via solid-state reaction, but that only a fraction of
the Na+ cations react until an equilibrium is reached.
In Fig. 19 sets of schematized XRD patterns are shown of a (1) the parent
Na-Y, (2) a ground mixture of LaCl3 ◊ 7H2O/Na-Y, (3) conventionally exchanged
La-Y(74) with an exchange degree of 74%, (4) a ground mixture of LaCl3 ◊ 7H2O/
Na-Y after heat treatment at 850 K, and (5) conventionally exchanged La-Y(74)
after heat treatment at 850 K.
Figure 19 demonstrates that upon solid-state cation exchange between LaCl3
and Na-Y the reflections of the product NaCl appear (cf. part b of the figure).
Moreover, the intensities of other reflections [e.g., those of the reflections (222),
(400), (511), and (551) of the faujasite phase] change towards those of La,Na-
Y(74) prepared via conventional exchange in aqueous solution (compare parts
a-c of Fig. 19). When, however, the samples were treated at higher temperatures,
the intensities of the NaCl reflections markedly decreased (compare parts b and
d of Fig. 19) and, also, the reflections of the faujasite phase indicated above
changed back toward those shown in Fig. 19, part a. The latter observation was
similar to that made with the sample La-Y (74) upon heating to 850 K (compare
parts c, e, and a of Fig. 19).
Solid-State Ion Exchange in Microporous and Mesoporous Materials 87

5.2.6
SSIE of La3+ Chloride with Na-Zeolites Investigated by IR
Qualitatively, the formation of OH groups typical of La-Y as a result of SSIE of
LaCl3 ◊ 7H2O with the sodium form of Y-type zeolites was substantiated by IR.
This is demonstrated by Fig. 20, where the appearance of bands in the OH
stretching region at ca. 3650 cm–1 and, after hydration, at 3540 cm–1 (assigned to
OH groups associated with La cations) on reaction of LaCl3 ◊ 7H2O and Na-Y is
shown (column B) and compared with the features of a conventionally
exchanged La-Y(76) (column A) [35]. After pyridine adsorption, the band at
3650 cm–1 was completely eliminated and the intensity of the band around
3540 cm–1 significantly weakened. Instead, the corresponding bands appeared in
the deformation region (vide infra).
Similar to what was reported for, e.g., alkaline earth/Na-Y systems (cf. Sect.
5.1.9), IR spectroscopy is also suitable to monitor the solid-state ion exchange in
LaCl3 ◊ 7H2O/Na-Y mixtures when a probe such as pyridine is employed. On acti-
vation of a wafer of a LaCl3 ◊ 7H2O/Na-Y mixture in the IR cell (10–5 Pa, 625 K)
and subsequent contact with pyridine (0.6 Pa, 375 K, followed by degassing), a
band at 1447–1448 cm–1 developed besides a second one at 1439 cm–1 being
indicative of pyridine coordinatively bound to La3+ and Na+ on cation sites, re-
spectively (cf. Fig. 21).
TRANSMITTANCE

WAVENUMBER [cm–1]
Fig. 20. IR spectra of (A), La,NaY [76] conventionally exchanged and (B) ground mixture of
hydrated Na-Y with crystalline LaCl3 after heat treatment at 850 K and washing with water: (a)
pretreated at 725 K in high vacuum for 3 h; (b) Sample (a) rehydrated with water (6.5 mbar) at
525 K for 0.5 h and subsequently degassed in 525 K, high vacuum for 1 h; (c) Sample (b)
exposed to pyridine (5.7 mbar) at 425 K for 2 h and again evacuated at 475 K for 1 h (after [35],
with permission)
88 H.G. Karge · H.K. Beyer

Fig. 21. IR spectra of the pyridine ring deformation frequency region after pyridine adsorp-
tion and subsequent degassing at 475 K and 10–5 Pa of La,Na-Y obtained A by conventional
and B contact-induced ion exchange (after [35], with permission)

Fig. 22. 23Na MAS NMR spectra of a parent Na-Y zeolite, b mixture of LaCl3/Na-Y, c sample (b)
after heat treatment at 850 K
Solid-State Ion Exchange in Microporous and Mesoporous Materials 89

Similar to the case of BeCl2/Na-Y (cf. Sect. 5.1.9), only at lower reaction tem-
peratures did the Py Æ Na+ band at 1439 cm–1 disappear completely, but at the
high temperature of 775 K the equilibrium (Scheme 1b) was shifted toward a
partial replacement of Na+ by La3+ only, as can be realized from the partial reap-
pearance of the 1439 cm–1 band. This is in agreement with the XRD and 23Na
MAS NMR findings described in Sects. 5.2.5 and 5.2.7, respectively.

5.2.7
SSIE of La3+ Chloride with Na-Zeolites Investigated by 23Na MAS NMR

As described in Sect. 5.1.10, 23Na MAS NMR is a suitable tool for detection and
determination of cation migration upon solid-state reaction between M+Cl or
M2+Cl2 and sodium forms of zeolites. Similar results obtained for the system
LaCl3 ◊ 7H2O /Na-Y are illustrated by Fig. 22 [35, 74].
A sharp signal at about –9 ppm (referenced to NaCl) is indicative of Na+ in
the large cavities, and Na+ cations in the small cavities are responsible for a sig-
nal at about –13 ppm (cf. [74, 86, 87]). In Na-Y containing sorbed water, both
bands overlapped, and the resulting band was found at –8.2 ppm (Fig. 21a).After
intense grinding of a mixture of LaCl3 ◊ 7H2O and Na-Y the spectrum b shown
in Fig. 22 was observed. The most striking feature was the appearance of a sharp
signal at 0 ppm that indicated that Na+ had been replaced by La3+ and tiny crys-
tallites of NaCl had formed. As can be derived from the second signal at
–13.1 ppm, all the remaining sodium cations resided in the small cavities. When
the ground mixture was subsequently heated to 850 K, the exchange was par-
tially reversed. This was evidenced by a decrease of the NaCl signal at 0 ppm and
the reappearance of the signal of Na+ in the supercages (around –9 ppm). The
23Na MAS NMR results, therefore, confirmed the observations made by XRD and

IR (cf. Sects. 5.2.5 and 5.2.6).

In comparison to conventional exchange, solid-state ion exchange between
LaCl3 ◊ 7H2O and the sodium form of Na-Y was also investigated by Hunger et
al. [88]. These authors employed 139La MAS NMR,27Al MAS NMR and 29Si MAS
NMR and suggested that La3+ migration into the sodalite cages was hindered
due to a blockage of the six-membered ring windows by formation of NaCl.
Sections 5.1 and 5.2 dealt with solid-state ion exchange with M2+ salts (M =
Mg or Ca) and La3+ chloride, respectively. However, in these cases, the salts were
initially not water-free but contained physically adsorbed and crystal water. The
role of water in SSIE will be discussed in detail in Sect. 7.1.

5.2.8
Catalytic Activity of La3+-Zeolites Prepared via SSIE

As mentioned in the introductory remarks of Sect. 5.2, a particularly interesting

question arose as to whether solid-state ion exchange would enable the conve-
nient preparation of catalysts for acid-catalyzed reactions. Various experiments
have indeed shown that solid-state ion exchange offers a new route for catalyst
production. Some of the pertinent results are described below (cf. also
Sects. 5.1.11, 5.3 and 5.4).
90 H.G. Karge · H.K. Beyer

Fig. 23. Disproportionation of ethylbenzene in a flow-reactor at 425 K after catalyst activation

at 725 K in a flow of dry helium, subsequent hydration (1.2 · 103 Pa water vapor followed by
degassing at 625 K) over La,Na-Y catalysts prepared by A solid-state ion exchange and B con-
ventional exchange; degree of exchange in La,Na-Y about 98% (after [78], with permission)

An almost completely exchanged La-Y sample (0.25 g), viz., La-Y(98) with
only 2% of residual Na+, was prepared via solid-state reaction of NH4-Y(98) with
LaCl3 ◊ 7H2O at 725 K. This process was carried out in a conventional micro-flow
reactor [78], followed by cooling to 455 K, a hydroxylation step (1.2 kPa H2O
vapor, 455 K, 2 h), degassing in a helium flow at 625 K and cooling to 425 K (reac-
tion temperature). Upon admission of the feed stream of 1.3 vol.% ethylbenzene
in dry helium (30 ml min–1), disproportionation to benzene and diethylben-
zenes was observed. The results of the conversion measurements are shown
in Fig. 23 and compared with those obtained over a conventionally prepared
La-Y(96) catalyst.
As can be seen from Fig. 23, the catalytic performance of catalyst A obtained
via solid-state ion exchange was even superior to that of the conventionally
exchanged material B, in that the conversion was somewhat higher under equal
conditions. The selectivity after completion of the induction period was, how-
ever, the same for catalysts A and B, i.e., in both cases benzene and diethylben-
zenes formed in the ratio 1:1.
A very interesting result was obtained when the SSIE in the system LaCl3
7H2O/NH4-Y(98) was conducted under careful exclusion of water vapor and the
zeolite produced tested for an acid-catalyzed reaction. Solid-state ion exchange
was carried out in situ in a high-vacuum-tight IR flow-reactor cell connected to
a gas chromatograph [78] in an extremely anhydrous flow of helium at 455, 575
and 675 K.When the solid-state reaction in the IR transmittant wafer containing
the mixture of the salt and the zeolite was completed, a feed stream of ethylben-
zene in carefully dried helium was passed through the cell.
It is clear from Fig. 24 that, at 455 K, the deammoniation of the NH4-Y compo-
nent was not yet complete: bands at 3630 and 3549 cm–1, indicative of acidic OH
Solid-State Ion Exchange in Microporous and Mesoporous Materials 91

Fig. 24. In situ IR spectra of and ethylbenzene conversion over La,Na-Y obtained by heat-
treatment of a LaCl3/NH4-Y mixture in a flow-reactor IR cell A prior to and B after brief con-
tact with 10 Pa water vapor; a, b and c pretreatment at 455, 575 and 675 K, respectively; d after
short contact with 10 Pa water vapor; e after admission of the ethylbenzene/helium feed
stream (for details see text; after [78], with permission)

groups, had already developed, but in the NH stretching region a broad peak
around 3200 cm–1 with shoulders at 3350 and 3080 cm–1 could still be seen indi-
cating that residual NH+4 was present. Between 455 and 575 K (spectra a and b,
respectively), the residual ammonium disappeared, but the OH bands did not
develop further. On the contrary, the intensities of the OH bands were diminished
because the solid-state reaction between the hydroxyls and LaCl3 has already
started to consume the OH groups generated via deammoniation. This solid-state
reaction was completed at 675 K (spectrum c). After admission of the ethylben-
zene feed (spectrum d) no conversion was measured (curve A).In the IR spectrum
only the bands of ethylbenzene appeared (not shown). Only after an extended
time on stream (t>1 h) did a very small conversion take place (vide infra). When,
however, after completion of the solid-state reaction first a brief (2 min) contact
with H2O vapor was allowed (0.1 kPa water vapor injected into the He-stream),fol-
lowed by degassing, OH bands at 3616 and 3518 cm–1 typical of an acidic La3+-con-
taining Y-type zeolite appeared (cf. Fig. 24, spectrum d). Upon passing the feed
stream over this hydroxylated wafer, the ethylbenzene interacted with the acidic
OH groups in the large cavities, since the band at 3616 cm–1 disappeared and the
bands of ethylbenzene adsorbed onto the catalyst (at 3074, 3032, 2972, 2936 and
2887 cm–1) developed (spectrum e). Moreover, an immediate onset of ethylben-
zene conversion was detected by gas chromatography (GC). After a steep increase
in the conversion, a steady state was reached at a conversion of 15% (curve B in
Fig. 24). Thus, the following conclusions can be drawn:
1. the material produced by solid-state ion exchange in the absence of water is
inactive in acid-catalyzed reactions such as ethylbenzene disproportionation;
the naked La3+ cations introduced by solid-state ion exchange are not able to
generate the necessary intermediates (carbenium ions);
92 H.G. Karge · H.K. Beyer

2. to prepare an active La-Y catalyst, contact with the co-catalyst H2O is a pre-
requisite for generation of acid Brønsted sites which are able to interact with
the hydrocarbon to give the carbenium ions necessary for the catalytic mech-
anism [76, 89, 90]. The formation of Brønsted sites obviously occurs accord-
ing to the Hirschler-Plank mechanism (cf. Eq. (10) and [64, 65], vide supra),
which assumes that a splitting of H2O molecules in the Coulomb field of mul-
tivalent cations such as La3+ takes place:
La3+Z–3 + 2 H2O Æ 2 H+Z– + La(OH)+2 Z– (10)
where Z– denotes a negatively charged monovalent fragment of the zeolite
framework.
This also explains the slow increase in the conversion from zero to about 0.5%
after 2 h on stream in the case without a deliberate hydroxylation step (see curve
A in Fig. 24). Here, traces of H2O introduced by the feed stream cause a minor acti-
vation of the La-Y wafer, the spectrum of which is shown in Fig. 24, spectrum c.
A similar comparison to that discussed above for a catalyst derived from
LaCl3 ◊ 7H2O/NH4-Y(98) was made between a conventionally exchanged La,
Na-Y(76) catalyst and a catalyst produced by SSIE in LaCl3 ◊ 7H2O /Na-Y, i.e., a
system containing the sodium form of Y-zeolite instead of the hydrogen form [79]:
After solid-state reaction in a stoichiometric mixture of LaCl3 ◊ 7H2O/Na-Y, the
material was briefly washed twice with a few milliliters of water to remove the
NaCl crystallites, which had formed (cf. Sects. 5.2.5 and 5.2.7), and non-reacted
LaCl3 . Subsequently, the sample (0.25 g) was activated at 625 K in high vacuum
(10–4 Pa, 3 h), cooled to reaction temperature (425 K) and contacted with the feed
stream (1.3 vol.% ethylbenzene in dry helium, 30 ml min–1). Figure 25 demon-
strates that the zeolite prepared from the sodium form of the Y-type zeolite was

Fig. 25. Conversion of ethylbenzene upon disproportionation over La,Na-Y (degree of

exchange: 76% of Na+) obtained by A solid-state ion exchange in a LaCl3/Na-Y mixture on
heat treatment for 3 h in high vacuum at 625 K and B conventional exchange; reaction tem-
perature: 455 K (after [79], with permission)
Solid-State Ion Exchange in Microporous and Mesoporous Materials 93

Fig. 26. In situ IR spectra of and decane cracking over highly exchanged La,Na-Y catalysts
obtained A, a by solid-state ion exchange and B, b conventional ion exchange; degree of
exchange: 98 and 96%, respectively; for details, see text (after [91], with permission)

indeed an active catalyst in an acid-catalyzed reaction such as ethylbenzene dis-

proportionation.
In fact, the conventionally produced catalyst, La,Na-Y(76), exhibited a short-
er induction period and a higher initial conversion than the catalyst obtained
through SSIE. However, both catalysts approached the same steady state conver-
sion and produced benzene and diethylbenzenes in exactly the ratio of 1:1 (not
shown). This corresponded to 100% selective disproportionation and excluded
concomitant dealkylation that would have resulted in catalyst deactivation [76,
89, 90].
Results similar to those discussed with respect to the disproportionation of
ethylbenzene over La-Y catalysts prepared by solid-state ion exchange were
obtained when cracking of n-decane was used as a test reaction [91]. Figure 26
displays the in situ IR spectrum of highly exchanged La-Y formed upon solid-
state reaction as described before and designated as La-Y(98; SE) and, for the
sake of comparison, the IR signals of La-Y produced by conventional exchange
with almost the same degree of exchange, designated as La-Y(96; CE).
Interestingly, the band at 3520 cm–1 is missing in the La-Y(98; SE) sample.
This indicates that the exchange has (preferentially) occurred in the large cavi-
ties (cf. Sect. 5.2.7). However, only the acid sites created in the large cavities and
seen through the band at 3630 cm–1 play a role in the catalytic process, because
only these are accessible for the feed, n-decane. After equal treatment both cata-
lysts, i.e., La-Y(98; SE) and La-Y(96; CE), exhibited the same steady-state activi-
ty in n-decane conversion (cf. Fig. 26, curves A and B). The main differences were
found in the product distribution: a higher yield of C3 to C5 over La-Y(98; SE)
and formation of C1, 1-hexene and a higher yield of hexane over La-Y(96; CE).
94 H.G. Karge · H.K. Beyer

5.3
SSIE of Other Transition Metal Cations

5.3.1
Introductory Remarks

Investigation of the exchange of transition metal cations into zeolites through

solid-state reactions is, to a larger extent, the domain of ESR spectroscopy. As
already mentioned in the Introduction (Sect.1),Clearfield and co-workers applied
this technique in an early study to prove solid-state ion exchange of, e.g., Cu2+ into
zeolite Y. Since the respective report [18] is not easily available, some of those ear-
ly results by Clearfield et al. will be described in more detail (see below). Resum-
ing Clearfield’s pioneering work, Slinkin and his group started in the mid-1980s a
systematic investigation of the introduction of copper cations from various cop-
per compounds into zeolites, particularly into ZSM-5 and mordenite-type zeolites
[21]. In addition, they carried out the first studies of the reaction of oxides such as
CuO, CrO3,V2O5 and Na2CrO4 with zeolites [21, 23, 92, 93].
However, it is not only ESR that can be employed to study transition metal
introduction into zeolites via solid-state reaction; as discussed in Sects. 4.2–4.7,
chemical analysis, temperature-programmed evolution of product gases, X-ray
diffraction and infrared spectroscopy are also generally applicable. In a few par-
ticular cases, other methods such as Mössbauer spectroscopy or UV-Vis spec-
troscopy, etc., have been equally helpful.
In fact, the incorporation of transition metal cations into zeolites became
very important in view of the generation of redox catalysts, particularly for
removal of NOx from exhausts (DENOX processes). Extensive studies were car-
ried out on the modification of zeolites, particularly of H-ZSM-5, by solid-state
ion exchange with copper, iron, cobalt, and nickel. Catalysis-related studies of
solid-state ion exchange with transition metal cations are reported at the end of
Sects. 5.3.2.1, and also in Sects. 5.3.4.1–5.3.4.3.
Therefore, the subsequent examples are not categorized with respect to the
methods used for monitoring solid-state reactions of compounds containing tran-
sition metal cations with zeolites. Rather, a classification of the examples is adopt-
ed according to the type of metal cations to be introduced into the zeolite matrix.

5.3.2
SSIE of Copper, Silver and Gold Compounds with Zeolites

5.3.2.1
Introduction of Copper

In their work, Clearfield et al. [18] first converted the sodium forms of zeolites
A, X and Y partially (exchange degree of 36%–58%) into the ammonium
(hydrogen) forms using the conventional method of exchange in aqueous solu-
tions of ammonium salts. The ammonium forms were deammoniated and react-
ed with the respective chlorides at 625–725 K in a helium flow or, for the ESR
measurements, under vacuum. Figure 27 (curve A) shows the ESR spectrum of
Solid-State Ion Exchange in Microporous and Mesoporous Materials 95

Fig. 27. ESR spectra of Cu,Na-Y obtained by A solid-state ion exchange in the mixture
CuCl2/H,Na-Y and B conventional exchange; C ESR spectrum of the parent Na-Y; degree of
exchange for A and B: 15% of the original Na+ cations were replaced by Cu2+(after [18, 24],
with permission)

zeolite Y exchanged in the solid state: 15% of the exchangable ions in H,Na-Y
were replaced by Cu2+; the sample, however, still contained appreciable amounts
of protons due to the preceding deammoniation. For the sake of comparison,
curve B presents the ESR spectrum of a Cu,Na-Y sample with the same content
of Cu2+ but prepared by conventional exchange in an aqueous solution of cop-
per acetate. Finally, curve C was obtained from the parent Na-Y.
Spectra A and B are in good agreement. Both exhibit two sets of lines with four
lines belonging to each set. This is indicative of two different environments of the
Cu2+ cations. The g-values measured for the first set of lines were g = 2.35 and g^
= 2.06, and for the second one g = 2.30 and g^ = 2.06 (vide infra). These almost
coincide with those reported by Krüerke and Jung [94] for conventionally pre-
pared Cu,Na-Y.Very similar spectra were produced for a significantly lower load-
ing of 5%. Solid-state ion exchange with CuCl2 as well as with other chlorides
(CoCl2, NiCl2; cf. [18] and Sect. 5.3.4) was also monitored by back-titration of the
HCl evolved and trapped in NaOH solution. These experiments showed, inter
alia, that the degree of solid-state exchange was increased when the ratio n M 2+/n +
H
2+ +
was enhanced. Thus, n Cu /n H = 0.4 and 2.5 led to a replacement of 40 and 70% of
the protons in H,Na-Y by Cu2+, respectively (vide infra). It was further shown that
zeolite A was least resistant against the attack of HCl upon solid-state reaction,
whereas Y zeolite was essentially stable and did not measurably lose crystallinity.
ESR spectra similar to those reproduced in Fig. 27 were obtained by Slinkin
et al. [23] when these authors studied the reactions between copper compounds
and H-ZSM-5 or hydrogen mordenite, H-MOR. In reactions with oxides, H2O is
formed instead of HCl, as was discussed above for the reactions with chlorides
(see below and Scheme 1c–e). The ESR spectrum resulting from a reaction
96 H.G. Karge · H.K. Beyer

Fig. 28. Comparison of ESR spectra of Cu2+-containing ZSM-5 samples. a Sample prepared by
solid-state ion exchange in a CuO/H-ZSM-5 mixture in vacuum at 1073 K, b sample obtained
by conventional exchange, calcined in air at 1073 K and evacuated at 300 K, c spectrum
obtained after contacting sample (a) with air (after [21, 95], with permission)

between CuO and H-ZSM-5 and subsequent evacuation is displayed in Fig. 28

(spectrum b).
Obviously, solid-state reactions in systems such as CuO/H-ZSM-5 with oxides
of the in-going cation require relatively high reaction temperatures: spectrum b
of Fig. 28 was recorded after reaction at 1073 K, and it was reported that increas-
ing the temperature from 793 to 1073 K considerably enhanced the degree of
exchange. Spectrum b of Fig. 28 is in complete agreement with that generated by
the same treatment of a conventionally exchanged Cu,H-ZSM-5 sample (see
spectrum a of Fig. 28). The g-values and the hyperfine splitting (HFS) constants
fully coincided. Similar to the system CuCl2/Y-zeolite (vide supra), two sets of g-
values and HFS constants were observed in both spectra a and b: g 1 = 2.29, g 1^ =
2.05, A1 = 15.6 mT, A^1 = 2.3 mT and g 2 = 2.31, g^2 = 2.06, A2 = 15.3 mT, A^2 =
2.25 mT. The computed g-values were found to be in good agreement with
experimentally determined ones. Therefore, Kucherov and Slinkin [21] con-
cluded like Clearfield et al. [18] that the Cu2+ cations resided on two differently
coordinated sites, with set (1) corresponding to a square planar and set (2) to a
pyramidal, fivefold coordinated state (cf. [95, 96]). The first site was visualized as
being close to the walls of the straight channels of the ZSM-5 structure, where
the Cu2+ cations would be coordinated to three oxygen atoms of the framework
and one extra-framework ligand. It was assumed that the second type of Cu2+
cation was surrounded by six oxygen atoms of the framework and at a greater
distance from the channel wall.
When the authors carried out experiments to contact Cu2+, which had been
introduced by solid-state reaction, with adsorbates such as oxygen, they
Solid-State Ion Exchange in Microporous and Mesoporous Materials 97

Fig. 29. Linear relationship between the maximum ESR intensity of Cu2+ in Cu,H-ZSM-5
(obtained by solid-state ion exchange in a CuO/H-ZSM-5 mixture) and the Al content of the
ZSM-5 framework (after [21], with permission)

observed significant modifications of the ESR signals (cf. spectrum c in Fig. 28).
The hyperfine splitting (HFS) completely disappeared. This showed that the
introduced Cu2+ cations were (i) coordinatively unsaturated and (ii) accessible
to the adsorbate molecules. Upon desorption of the adsorbate, the original spec-
trum was restored, i.e., the adsorption process was entirely reversible.
When H-ZSM-5 samples with increasing nSi/nAl ratios were used, a linear rela-
tionship between the integrated ESR signal intensity and the Al content of the
zeolite matrix was found (Fig. 29). Since an increasing Al content corresponded
to an increasing number of Brønsted acid sites (cf. Introduction), this finding
proved the important role of those acid centers in the solid-state ion exchange
taking place in the system CuO/H-ZSM-5.
Interestingly, upon interactions of CuO with hydrogen forms of ZSM-5,
unsaturated isolated Cu2+ cations were observed to a much lesser extent when
H,Na-ZSM-5 was employed instead of H-ZSM-5. With ZSM-5, which contained
significant amounts of Na+ cations, Cu2+ was incorporated in octahedral coordi-
nation. No ESR signal at all was obtained after solid-state interaction of CuO
with sodium forms of the zeolites, i.e., Na-ZSM-5 or Na-MOR.
Slinkin and co-workers reported results similar to those found with CuO for
solid-state reactions of CuCl2 , CuF2 , Cu3[(OH)CO3]2 , Cu3(PO4)2 and Cu0 (in the
presence of air). However, the reaction proceeded most easily with CuCl2 , where-
as it was markedly slower in the other cases and slowest with Cu0 in air. It could
be shown by the authors via comparison with ESR measurements of conven-
tionally exchanged Cu-zeolites that Cu2+ was not solely coordinated to oxygen
atoms of the framework but also to one negatively charged extra-framework
species such as OH–, Cl–, F–, or PO43–. Furthermore, evidence was provided that
monovalent Cu+ was incorporated into the zeolite matrix upon reaction of Cu2S
with H-ZSM-5. Even when this was carried out in a mildly oxidative atmosphere
and CuO formation excluded, the ESR lines typical of isolated Cu2+ on the above-
mentioned sites were observed.
98 H.G. Karge · H.K. Beyer

Solid-state ion exchange after SSIE at 1073 K in, e.g., CuF2/H-ZSM-5 was
checked by XRD: no sign of lattice destruction was detected.
Calcination of H-MOR with copper compounds such as CuO, Cu3[(OH)CO3]2 ,
or with Cu0 (in air) at 823 K yielded ESR spectra (g = 2.325, g^ = 2.055, A =
14.4 mT, A^ = 1.9 mT) very similar to those obtained with H-ZSM-5 [21].
According to Jirka et al. [97], grinding of a mixture of Cu2O and NH4-Y fol-
lowed by heat-treatment at temperatures up to 770 K caused an exchange of cop-
per cations into the zeolite matrix. The incorporation started around room tem-
perature, but was markedly enhanced on heating between 420 and 620 K. The
authors derived from their XPS and XAES results that first the Cu2O particles
disaggregated and were oxidized to Cu(OH)2. Subsequently, Cu2+ species migrat-
ed into the zeolite channels. The process was facilitated by a pre-exposure of the
mixture Cu2O/NH4-Y to water vapor.
High-temperature ion exchange between solid, Cu-containing phases and H-
ZSM-5 or NH4-ZSM-5 zeolites was also studied by Karge et al. [98]. CuCl, CuCl2 ,
Cu2O and CuO were employed and the solid-state reactions monitored by ESR,
XRD, TPE and IR. ESR investigations were carried out after oxidation (1.3 kPa
O2 , 570 K, 1 h; followed by evacuation). The features of all of the ESR spectra
were very similar. The signals were compared with those of conventionally ex-
changed, equally treated Cu,H-ZSM-5 samples and could be assigned to isolated
Cu2+ cations introduced into the zeolite matrix, thus qualitatively confirming
that SSIE has occurred. The spectra were essentially in agreement with those
obtained by Slinkin et al. [21] and discussed above. Parameters for a square-pla-
nar coordination determined were: g = 2.33, g^ = 2.07, A = 12.5–14.0 mT (cf.
[95], where data for conventionally exchanged Cu-ZSM-5 samples were report-
ed as g = 2.31, g^ = 2.06, A = 15.3 mT, A^ = 2.25 mT). Two sets of parameters
were probably superimposed corresponding to the square-planar and fivefold
pyramidal coordinations as reported [21, 95] (vide supra). However, the resolu-
tion was not as good as achieved by Slinkin and co-workers: In particular, the
splitting at g^ was poorly defined. The intensities of the ESR signals after solid-
state reaction with the Cu oxides were, in fact, distinctly lower (by about 50%)
compared to those obtained from the heat-treated mixtures of H-ZSM-5 with Cu
chlorides. This agrees with the IR results (vide infra).
An XRD experiment on solid-state reaction of CuCl2 and CuCl with H-ZSM-
5 is illustrated by Fig. 30. It shows the XRD patterns of CuCl2/H-ZSM-5 and
CuCl/H-ZSM-5 before and after solid-state reaction at 775 K [98]. No sign of
damage of the crystal lattice was observed. However, the fact of solid-state ion
exchange was clearly indicated by the disappearance of the reflections of the salt
components, i.e., of CuCl2 ◊ 2H2O and CuCl. Thus, the XRD experiment con-
firmed the results of the investigations by ESR (see above), TPE and IR (see
below). Similarly, no deterioration of the crystallinity was detected by XRD
when, instead of H-ZSM-5, samples of NH4-Y (H-Y) were employed in SSIE with
copper compounds (vide infra).
A careful and extensive structural analysis of Cu-Y obtained via SSIE (Treact =
698 K, treact = 18 h) between CuCl2 ◊ 2H2O and highly exchanged NH4-Y was con-
ducted by Haniffa and Seff, using pulsed-neutron diffraction [99]. The samples
prepared in this manner had the composition Cu24Na5H17Cl15Al55Si137O384 exclu-
Solid-State Ion Exchange in Microporous and Mesoporous Materials 99

Fig. 30. XRD patterns of CuCl2 · 2H2O/H-ZSM-5 and CuCl/H-ZSM-5 mixtures before and after
solid-state reaction at 775 K (after [98], with permission)

sive of water (designated as Cu24-Y and, after interaction with D2O, Cu24-Y◊ D2O).
The authors assumed that most or all of the protons stemmed from the decom-
position of NH+4 and were retained as H3O+ together with Cl–. Significant salt
imbibition was detected (cf. Sects. 5.1.5, 5.2.3 and [53]). From the analysis of the
XRD data, the population of sites III¢, I¢ and I by Cu2+ was derived. The Cu2+
cations on site III¢ were found to coordinate to four framework oxygen atoms in
a (distorted) square-planar manner and, perhaps, also to one extra-framework
species such as H2O, OH– or Cl– to give a distorted square pyramid. This is essen-
tially in agreement with the ESR and UV-Vis results reported by Slinkin et al.
[93, 95, 96] and Weckhuysen et al. [100], respectively.
Temperature-programmed heating of both CuCl/H-ZSM-5 and CuCl2/H-
ZSM-5 mixtures revealed that above 570 K a substantial amount of HCl was
evolved. The solid-state reaction in CuCl/H-ZSM-5 mixtures was strongly con-
trolled by the temperature. It was rather fast in the initial stage but then
(T>650 K) proceeded very slowly (cf., e.g., Figs. 31 and 32, Table 8). The TPE
profile for the system CuCl2/H-ZSM-5 looked almost identical to that of CuCl/
H-ZSM-5 [98].
Similar features are seen in Figs. 31 and 32: an increasing rate of solid-state
reaction between ca. 550 and 650 K followed by slow subsequent conversion.
From Fig. 32 it can be seen that the rate (measured through the slope of the
ascending part of the curves) and degree of solid-state ion exchange are
enhanced by an increase in the content of the Cu-containing compound in the
mixture in agreement with the finding by Clearfield et al. [18] (vide supra).
When an excess of CuCl was applied (nCu + /n
OH = 2.25), 86% of the bridging OH
groups were consumed, the protons being replaced by Cu+.
Hartmann and Boddenberg [101, 102] used SSIE of CuCl and NH4-Y(70) with
an exchange degree of 70% to prepare a Cu-Y sample for a study of CO and Xe
adsorption monitored by 13C and 129Xe MAS NMR, respectively. Introduction of
100 H.G. Karge · H.K. Beyer

Table 8. Chemical composition of the original mixtures of H-ZSM-5 with CuCl, CuCl2 , Cu2O,
CuO, and their characteristics after heat-treatment in high vacuum (SSIE)

Composition of the mixture a,b Heat-treatment c ESR signals d IR results

Intensity
Salt/oxide Cu m · nCum+/nOH T (K) t (h) (arb. units) OH con- Absorbance f
(mmol g–1) sumed e (arb. units)
(%)
PyH+ PyL

CuCl 0.29 0.32 770 12 – 30 – –

0.92 1.00 670 12 70 38 – –
0.92 1.00 770 0.5 – 47 – –
0.92 1.00 770 12 120 53 0.26 1.04
CuCl2 0.46 1.00 770 12 135 43 – –
Cu2O 0.46 1.00 770 12 66 17 0.37 0.47
CuO 0.46 1.00 770 12 50 25 0.39 0.46
a Cu content per gram dry zeolite.
b Number of Cu cations related to the original number of the zeolite OH groups (0.91 mmol
per gram dry zeolite); m, cation valency.
c Under vacuum at 10–4 Pa.
d After heat-treatment the mixture was oxidized in 103 mbar oxygen at 570 K for 1 h and evac-
uated at 420 K for 10 min.
e Calculated from the intensity of the IR band at 3610 cm–1.
f Pyridine (6.3 · 102 Pa vapor pressure) adsorption at 470 K for 2 h, desorption at 470 h for
10 min. PyH+ from the intensity of the band at 1545 cm–1 and PyL from that at 1450 cm–1; PyH+
and PyL intensities for the parent H-ZSM-5 were 0.51 and 0.31 (arb. units), respectively.

Cu+ via SSIE proved to be a reliable method. Using SSIE an exchange degree of
70% was achieved (Cu-Y(70; SSIE)). The crystallinity was preserved upon SSIE,
as confirmed by XRD, 29Si and 27Al MAS NMR. The properties of a sample pre-
pared by introduction of Cu+ via SSIE, i.e., Cu-Y(70; SSIE), were compared with
those of samples obtained by conventional exchange with Cu2+ in aqueous
Cu(NO3)2 solution after dehydration, oxidation and subsequent reduction, i.e.,
Cu-Y(75; CE). Cu-Y(70; SSIE) exhibited by far the highest concentration of Cu+
(30 Cu+/u.c.). About 70% of the Cu+ cations (27 of 29/u.c.) were shown to reside
in the supercages, where they quantitatively replaced the Na+ ions. Cu-Y(70;
SSIE) also exhibited the highest adsorption capacities for CO (strongly bound to
Cu+) and Xe compared with that of a mildly reduced Cu-Y(75; CE) which con-
tained ca. 11Cu+/u.c. More severely reduced Cu-Y(75; CE) contained no Cu+ at
all but significant amounts of metallic Cu0. Furthermore, it was shown by 13C
and 129Xe MAS NMR that xenon was a more advantageous probe for Cu+ in zeo-
lites than the widely used CO since Xe, in contrast to CO, does not (through
adsorption) change the distribution of the cations.
Solid-state ion exchange was confirmed by TPE in the systems Cu(I) oxide/
H-ZSM-5 as well [98]. The evolution of H2O was monitored by MS and profiles
such as those displayed in Fig. 33 were observed. Interestingly, solid-state reac-
tions with oxides extended to markedly higher temperatures compared to reac-
tions with chlorides. This is in agreement with IR investigations, where up to a
temperature of 670 K, which is far above the onset of SSIE with Cu chlorides, no
Solid-State Ion Exchange in Microporous and Mesoporous Materials 101

Fig. 31. Temperature-programmed (10 K min–1) evolution of HCl, m/e = 36, from a CuCl2/
H-ZSM-5 mixture (2nCu2+/nOH = 1.0) pretreated at 390 K

Fig. 32. Temperature-programmed (10 K min–1) evolution of HCl, m/e = 36, from CuCl/
H-ZSM-5 mixtures as a function of the amount of admixed CuCl with nCu+/nOH = 0.32, 1.00,
1.60, and 2.25 (after [98], with permission)
102 H.G. Karge · H.K. Beyer

Fig. 33. Temperature-programmed (10 K min–1) evolution of H2O, m/e = 18, from Cu2O/
H-ZSM-5 mixtures (nCu+/nOH = 1.0) pretreated at 390 K for 1 h (after [98], with permission)

measurable decrease in the number of OH groups was observed. The profile in

Fig. 33 exhibits (i) a maximum at 475 K (removal of strongly held H2O), (ii) a step
at 530 K and a shoulder above 550 K (H2O removal overlapping with a low-tem-
perature reaction), and (iii) a feature around 775 K (due to a high-temperature
reaction), analogously to what was observed in the systems MCl/hydrogen zeo-
lites or MCl2/hydrogen zeolites (cf. Sects. 5.1.2 and 5.1.4).
The introduction of Cun+ into H-ZSM-5 or H-Y was also monitored by the
consumption of the acid OH groups indicated by the decreasing intensity of the
IR bands at 3610 cm–1 (H-ZSM-5) or 3640 (HF) and 3555 cm–1 (LF) (cf. Fig. 34
and Table 8). Also, to some extent, silanol groups (3745 cm–1) were affected. (For
the effect of pyridine adsorption after SSIE see below.)
It could be excluded that the changes in the content of bridging hydroxyls was
caused by dehydroxylation, since no evolution of H2O was detected by MS dur-
ing heat-treatment. The quantitative data of Table 8 confirm the above state-
ments derived from TPE. For instance, an increase in the reaction temperature
from 670 to 770 K enhanced the degree of exchange (as measured by the con-
sumption of OH groups) from 38 to 53%, respectively (Table 8, rows 2 and 4).
Similarly, an extension of the reaction time from 0.5 to 12 h brought about an
increase of the degree of exchange from 47 to 53%, respectively (Table 8, rows 3
and 4). Further prolonged heating, however, did not result in further consump-
tion of OH groups. Finally, an elevation of the CuCl content in the mixture from
nCu+/nOH = 0.32 to nCu+/nOH = 1.0 raised the degree of exchange from 30 to 53%
(cf. Table 8, rows 1 and 4). Analogous effects of the exchange temperature and
time on the solid-state reaction between CuCl, CuCl2 , Cu2O and H-Y can be rec-
ognized from Table 9. It is worthy to note that after SSIE in the system CuCl2/H-
Y the consumption of OH groups in the large and small cavities was found to be
similar, viz., 67%–75% and 69%–75%, respectively.When, however, the heating
was stopped at 620 K, both TPE and IR spectroscopy evidenced that only the
Solid-State Ion Exchange in Microporous and Mesoporous Materials 103

Fig. 34. IR spectra of OH groups and adsorbed pyridine: a parent H-ZSM-5 zeolite heated in
vacuum at 770 K for 12 h; b and c CuCl/H-ZSM-5 (nCu +/nOH = 1.0) heated in high vacuum at
670 and 770 K, respectively, for 12 h; d after adsorption of pyridine at 470 K subsequent to (c)
(after [98], with permission)

Table 9. Chemical composition of the original mixtures of NH4-Y with CuCl, CuCl2 , Cu2O, and
their characteristics after heat-treatment in high vacuum (SSIE)

Composition of the mixture a,b Heat-treatment c IR results

Salt/oxide Cu m · nCum+/nOH T (K) t (h) OH consumed d Absorbance e

(mmol g–1) (%) (arb. units)

HF LF PyH+ PyL

CuCl 0.92 1.00 670 2 60 61 – –

0.92 1.00 670 12 62 64 0.46 2.12
CuCl2 0.46 1.00 670 12 75 75 – –
Cu2O 0.92 1.00 670 12 23 15 0.81 0.54
a Cu content per gram dry zeolite.
b Number of Cu cations related to the number of zeolite bridging OH groups; m, cation valency.
c Under vacuum at 10–4 Pa.
d Calculated from the intensity of the IR bands at 3640 cm–1 (HFband) and 3555 cm–1 (LFband).
e Pyridine (6.3 · 102 Pa vapor pressure) adsorption at 470 K for 2 h and desorption at 470 K
for 1 h. PyH+ from the intensity of the band at 1540 cm–1 and PyL from that at 1452 cm–1;
PyH+ and PyL (true Lewis sites, [171–172]) intensities for the parent H-Y after similar treat-
ment were 0.81 and 0.45 (arb. units), respectively.

number of OH groups located in the large cavities (supercages) had decreased

while those in the small cages were almost unaffected (stage 1). These findings
suggest that in the first stage of SSIE the protons located in the supercages, which
are more easily accessible, predominantly interact with the salt. Only at higher
temperatures are the in-going species (possibly CuCl2 molecules, cf. Sect. 7) able
to penetrate the six-membered rings into the b-cages, and/or a rapid migration
104 H.G. Karge · H.K. Beyer

Fig. 35. Temperature-programmed (10 K min–1) evolution of HCl, m/e = 36, from CuCl2/
NH4-Y (nCu2+/nOH= 0.5) pretreated at 390 K for 1 h (after [98], with permission)

of Cu cations from the large supercages into the small b-cages becomes possible
(stage 2). This seemed to be reflected in the appearance of two maxima in the
TPE profile of the system CuCl2/NH4-Y (cf. Fig. 35). The LT-peak would then cor-
respond to stage 1, whereas the HT-peak would be due to stage 2. From a com-
parison of Tables 8 and 9 it can also be seen that the exchange of protons of
bridging OH groups for Cun+ was, for a given temperature, considerably higher
in the case of mixtures of Cu chlorides or oxides with H-Y than with H-ZSM-5.
Thus, SSIE seemed to be influenced by the structure type of the zeolite em-
ployed.
IR investigations of the solid-state reaction of Cu compounds with H-ZSM-5
or H-Y were also conducted using IR spectroscopy and pyridine as a probe (cf.
also Sect. 8, where this method was employed for investigations of the kinetics
of SSIE). For instance, Figs. 34 and 36 clearly show that introduction of Cun+ is
indicated by a band at 1452–1453 cm–1 due to pyridine coordinated to the cation
acting as a Lewis acid site (Py Æ L). Correspondingly, the extent of SSIE may be
derived from a comparison of the spectra of the parent material and the
exchanged zeolite after pyridine adsorption: the intensity decrease of the band
at 1540 cm–1, which originated from pyridinium ions (PyH+), is a measure of the
degree of exchange. Some quantitative results are included in Tables 8 and 9.
From a comparison of the IR data presented in Tables 8 and 9 (consumption
of OH groups, formation of Py Æ L, decrease in the intensity of the PyH+ band),
it can be concluded that the degree of exchange with H-ZSM-5 decreased, under
otherwise identical conditions, in the sequence CuCl>CuCl2 >CuO: for example,
for T = 770 K, t = 12 h, m · nCum+/nOH = 1.0, the percentage of OH groups consumed
follows the sequence 53 (CuCl), 43 (CuCl2), 25 (CuO). In contrast, with H-Y, the
degree of exchange measured via consumption of the OH groups seemed to be
slightly higher when the zeolite was reacted with CuCl2 instead of CuCl: 75%
Solid-State Ion Exchange in Microporous and Mesoporous Materials 105

Fig. 36. IR spectra of OH groups and adsorbed pyridine: a parent NH4-Y zeolite heated in vac-
uum at 670 K for 12 h; b CuCl2/NH4-Y (nCu2+/nOH = 0.5) heated in high vacuum at 670 K for 12 h;
c after pyridine adsorption subsequent to (b) (after [98], with permission)

(CuCl2)>62%–64% (CuCl). Most likely, this was due to the fact that in H-Y
there is a considerably lower nSi/nAl ratio than in H-ZSM-5, viz., 2.6 vs. 13.5,
which facilitated the incorporation of bivalent cations in H-Y. The PyH+ and
Py Æ L results indicated for both zeolite systems a higher degree of exchange in
the case of chlorides than oxides. However, after reaction with H-ZSM-5, similar
degrees of exchange were observed for both Cu oxides.
Another spectroscopic investigation was carried out by Borovkov et al. [103],
who investigated by diffuse reflectance IR spectroscopy the vibration modes of
CO adsorbed on Cu(I)-Y prepared via SSIE of NH4-Y and CuCl. The various fun-
damental and 1st overtone bands were compared with those observed with con-
ventionally ion-exchanged Cu-Y, Cu-MOR and Cu-L, which were subsequently
auto-reduced at 673 K in vacuum.
Esemann and Förster [104] employed far-infrared and X-ray absorption
spectroscopy, assisted by computer modeling, to study copper exchange into
ZSM-5 by solid-state reaction, the siting of the introduced cations, NO adsorp-
tion and decomposition on and redox behavior of the obtained Cu-ZSM-5.
Förster and Hatje [105] applied EXAFS techniques to study the incorporation of
Cu+ as well as of Zn2+ and Ni2+ (vide infra) into Y-type zeolite. The oxidation
state of introduced Cu+ remained unchanged upon SSIE: The determined Cu+-
oxygen distances were too small as to suggest a coordination to the oxygen
106 H.G. Karge · H.K. Beyer

atoms of the six-membered rings. Therefore, it was tentatively assumed that

asymmetric cation positions occur or, more likely, Cu-O-Cu complexes were
formed.
Introduction of Cu into H-ZSM-5 through solid-state and conventional ion
exchange was carried out, and the resulting materials were compared in a
study by Auroux et al. [106] using microcalorimetry and XPS. A combination of
microcalorimetric determination of the acidity by adsorption of NH3 and XPS
measurements of the N1s lines of the adsorbed ammonia allowed these authors
to discriminate between the Brønsted (B) and Lewis (L) acidity. The experiments
revealed an extreme increase of the density of L-sites as a consequence of Cu
incorporation.
Besides Y- and MFI-type zeolites, other zeolites such as BETA [107, 108] and
clinoptilolite [109] were also loaded with Cu via SSIE. H-BETA was reacted with
increasing amounts of CuCl, finally resulting in a complete disappearance of the
acid OH groups as evidenced by IR. It was shown that two kinds of Cu(I) sites
exist. The equilibria for CO adsorption on these sites were involved in the car-
bonylation of alcohols which, however, required the simultaneous presence of
residual adjacent acid OH groups. Copper could be introduced by SSIE into nat-
ural clinoptilolite to an extent which was comparable to that of conventional
exchange [109]. The zeolite obtained by solid-state reaction was successfully
employed in the catalytic 3-methyl-3-butyn-2-ol decomposition.
Cu-ZSM-5 catalysts for the aromatization of light paraffins and conversion of
1-propylamine were prepared via SSIE by Kanazirev and Price [110, 111] (cf. also
Sects. 6.3.3.2 and 6.3.3.3).
King [112, 113] found that Cu(I)-Y prepared by SSIE was, in contrast to con-
ventionally produced Cu(II)-Y, an active catalyst for oxidative carbonylation of
CH3OH to dimethyl carbonate. The presence of NH3 facilitated the migration of
copper ions into the zeolite structure.
As mentioned in Sect. 5.3.1, Cu-containing zeolites prepared by SSIE have
been employed as DENOX catalysts (cf. [114–128]). However, the various stud-
ies did not lead to an entirely consistent picture of the catalytic behavior of the
resulting materials. Most probably this is due to important but not well-docu-
mented differences in preparation and application.
For example, Varga et al. [114] compared, in the context of the catalytic
removal of NOx from exhausts, the behavior of Cu-ZSM-5 produced by conven-
tional exchange, Cu-ZSM-5(CE), with that of Cu-ZSM-5(SSIE) as prepared via
solid-state ion exchange. The latter procedure was carried out with CuCl2/H-
ZSM-5 mixtures which were calcined at 873 K in air for 8 h. The modified zeo-
lite samples were characterized by TGA, XRD, IR and BET measurements, NO
adsorption and decomposition. XRD, IR (KBr technique) and BET measure-
ments proved that the crystallinity of the samples was maintained during SSIE.
In contrast to Cu-ZSM-5(CE), the samples obtained via SSIE did not exhibit
bands due to Brønsted acid sites in the OH stretching region or, after pyridine
adsorption, in the pyridine ring deformation region (no pyridinium ions were
formed). On adsorption of NO, formation of NO2 was observed with both types
of samples, even though it was more pronounced with Cu-ZSM-5(CE). The dif-
ferences in the behavior (adsorption and reaction of NO) of the samples were
Solid-State Ion Exchange in Microporous and Mesoporous Materials 107

ascribed to differences in concentration and positions of the introduced copper

cations. Similar experiments were carried out with FeCl2, CoCl2, and NiCl2 (vide
infra, Sect. 5.3.4). In subsequent work, Varga et al. found that Cu-ZSM-5 pro-
duced via SSIE and characterized by ESR [115–116] or IR, BET, derivatographic
and acidity measurements [117] exhibited NO adsorption and catalytic behav-
ior different from that of conventionally exchanged Cu-ZSM-5. In particular, the
samples obtained by SSIE were reported to be less active in NO decomposition
[118]. In contrast, Schay et al. [119] found no difference in catalysis as long as
catalysts with the same Cu-loadings were employed.When, however, Cu,H-ZSM-
5 was prepared with an excess of CuCl (CnCu/nOH = 1.5) and reduced by CO at
773 K, the resulting material was highly active in NO decomposition even at tem-
peratures below 573 K [120]. Similarly, Varga et al. [121] described the prepara-
tion of good catalysts for NO decomposition via SSIE of CuCl2 with H-ZSM-5.
Selective catalytic reduction of NO by propane in the presence of oxygen over
Cu,H-ZSM-5 prepared via SSIE was studied by Halasz and Brenner [122].
Despite its relatively low degree of exchange (63%), the catalyst was found to be
rather active and produce more N2 than analogously prepared Ag,H-ZSM-5 and
Li,H-ZSM-5 samples. However, the selectivity for N2 was reported to be lower
over Cu,H-ZSM-5 since NO2 was also formed. Similarly, Halasz et al. [123]
showed that such Cu-ZSM-5 catalysts exhibited a high activity in the selective
reduction of NO by ammonia, propane or propene. Another comparative study
was conducted by Setzer et al. [124]. These authors investigated Cu-ZSM-5 sam-
ples, produced via (i) conventional exchange, (ii) SSIE and (iii) direct crystal-
lization from Cu cations-containing synthesis gels, with respect to their catalyt-
ic activity in NO decomposition in diesel exhaust gases. The lower activity of
materials produced through method (ii) or (iii) was ascribed to the fact that they
contain Cu predominantly in a distorted octahedral coordination, whereas Cu
introduced according to (i) exhibits a square-planar or quadratic pyramidal
(unsaturated) coordination (see, however, [18, 23, 95, 96]).
Poeppel et al. [125] attempted to react siliceous MCM-41 by SSIE with Cu(II)
salts. However, investigation with ESEM and comparison with MCM-41
exchanged in aqueous salt solutions proved that Cu2+ ions introduced by SSIE
did not show any interaction with the MCM-41 framework. The authors con-
cluded that in the samples modified by SSIE the Cu cations occupy sites differ-
ent from those in the conventionally prepared Cu-MCM-41 materials.
Catalysts for selective catalytic reduction (SCR) with propene were prepared
by Liese and Grünert [126] through solid-state reactions of CuCl, Cu(NO3)2 or
Cu(CH3COO)2 with Na-ZSM-5 or Na-ZSM-5/H-ZSM-5 mixtures and character-
ized by XPS and X-ray-induced AES (surface) and IR with pyridine as a probe
(bulk). Introduction of copper was most intense for CuCl, but intrazeolite cop-
per was also stabilized through SSIE in the chlorine-free systems. For compari-
son, Cu-ZSM-5 conventionally exchanged in aqueous solutions was used. It
turned out that the reaction rates over both types of catalysts, when normalized
to the same degree of exchange, were in the same order of magnitude. Even
though incorporation of copper via SSIE predominantly occurred into H-ZSM-
5, when it was present in a mixture with Na-ZSM-5, the Brønsted acidity proved
to be irrelevant in the SCR under study.
108 H.G. Karge · H.K. Beyer

Price et al. [127] prepared Cu-MFI catalysts for NO decomposition by high-

temperature treatment of a mixture of CuO and H-MFI in the absence of oxygen.
The process included most probably auto-reduction, and the copper cations were
introduced as Cu+. The upper limit was close to 1 Cu/1 framework Al. The results
of ESR, TPR and thermal desorption of 1-propylamine were consistent with this
assumption. The materials produced were very active in the decomposition of
NO into N2 and O2 and behaved similarly as so-called overexchanged Cu2+-MFI.
Copper cations were also introduced into zeolites by solid-state reaction of a
complex oxide such as CuCrO4 [93]. The chromate was mixed with H-ZSM-5 and
heated in air to 823–1073 K. ESR spectroscopy yielded two superimposed spec-
tra from Cu2+ and Cr5+ cations (cf. Sect. 5.3.5). These cations were randomly
located in the cation positions of the ZSM-5 structure and showed negligible
dipole-dipole interaction. In another experiment, a consecutive solid-state reac-
tion was carried out first with CrO3 and then with CuO which resulted first in the
appearance of the Cr5+ signals and then, after reaction of CuO, in a decrease of
their intensities. Simultaneously, the signals of Cu2+ cations developed. This
indicated that a fraction of the initially incorporated Cr5+ cations were replaced
by Cu2+ cations, and the latter were more strongly held than Cr5+. Reduction by
hydrogen caused both the Cr5+ and Cu2+ signals to disappear.
Copper cations were incorporated by solid-state ion exchange into the galli-
um analog of ZSM-5, i.e., into the gallosilicate H-[Ga]ZSM-5 [128, 129]. The ESR
results reported by Kucherov et al. [128, 129], however, were essentially the same
as those obtained with the aluminosilicate, i.e., H-[Al]-ZSM-5. For example,
after SSIE with a mixture of CuO/H-[Ga]-ZSM-5 at 823 K, the same hyperfine
splitting was observed. Also, two types of isolated Cu2+ cations were identified,
one in a square-planar and the other in a five-coordinated environment. Inter-
action with gases, e.g., O2 , showed that the copper ions were accessible.

5.3.2.2
Introduction of Silver

There is, to our best knowledge, only one report on solid-state ion exchange of a
silver compound with zeolites. Ag+ was incorporated into ZSM-5 by solid-state
reaction of AgCl with H-ZSM-5 [130]. Figure 37 shows the result of a tempera-
ture-programmed heat-treatment of an AgCl/H-ZSM-5 mixture.
The evolving gases, H2O (M = 18) and HCl (M = 36), were monitored by a
mass spectrometer. A very pronounced HCl peak occurred at 890 K, but no de-
hydroxylation peak was observed. This demonstrated that a solid-state ion
exchange, viz., Ag+ for H+, had taken place. It is worth mentioning, however, that
this solid-state reaction proceeded with a salt that is insoluble in water. This, in
turn, demonstrated that the presence of H2O, which might have been physically
adsorbed on the zeolite material, is not a prerequisite for SSIE to occur (cf.
Sect. 7.1). Solid-state ion exchange between AgCl and H-ZSM-5 was also proven
by IR spectroscopy. Figure 38 demonstrates the significant decrease in the
absorbance of the OH band at 3610 cm–1 which occurred when the parent zeo-
lite, H-ZSM-5, was reacted with AgCl at 675 K. The difference in the absorbances
corresponded to a degree of exchange of about 70%.
Solid-State Ion Exchange in Microprobes and Mesoporous Materials 109

Fig. 37. Solid-state ion exchange in the system AgCl/H-ZSM-5 monitored by mass spectro-
metric analysis of the gases evolved from an AgCl/H-ZSM-5 mixture (nAg+/nAl = 1.0) as a func-
tion of the temperature; heating rate: 10 K min–1 (after [130], with permission)

Fig. 38. Solid-state ion exchange in the system AgCl/H-ZSM-5 shown by the decrease in the
intensity of the OH bands of the parent zeolite (after [130], with permission)
110 H.G. Karge · H.K. Beyer

5.3.2.3
Introduction of Gold

Na-Y zeolites containing gold species were prepared by SSIE by subjecting a

physical mixture of Na-Y and AuCl3 to a thermal treatment under vacuum. The
product was investigated by XAFS spectroscopy. The XANES spectra provided
evidence for the presence of Au0, AuCl and Au2Cl6 (cf. [131] and Vol. 4, Chap. 5,
this Series, see [69] therein).

5.3.3
SSIE of Zinc, Cadmium and Mercury Compounds with Zeolites

5.3.3.1
Introduction of Zinc

Introduction of zinc by solid-state ion exchange into ZSM-5 was investigated by

Salzer [132], Salzer et al. [133] and Roessner et al. [134] using diffuse reflectance
and transmission IR spectroscopy. A mixture of ZnO and H-ZSM-5 was subject-
ed to heat-treatment at 823 K well below the sublimation point of ZnO (2223 K).
It was shown that (i) the solid-state ion exchange was completed after 1 h, (ii)
ZnO seemed to be the migrating species, and (iii) all types of OH groups present
in the zeolite reacted simultaneously rather than consecutively [133]. Zinc-mod-
ified ZSM-5 materials were also produced by Hagen and Roessner [135], who
thermally treated mechanical mixtures of ZnO and H-ZSM-5 (with and without
template) at about 720 K under evolution of H2O, characterized the products
containing 2 wt.% Zn by IR using pyridine as a probe, TPD of ammonia to evi-
dence the absence of Brønsted acid sites and X-ray absorption spectroscopy
measuring the Zn K edge in transmission of excited synchrotron radiation. The
latter technique (XANES analysis) allowed the authors to distinguish between
Zn2+ in ZnO and Zn2+ in cationic positions in the zeolite structure.After 5 h reac-
tion time about 0.06 mmol Zn per gram zeolite was located in cationic sites.
Zn-Y zeolites were prepared via SSIE by Seidel et al. [136] and Boddenberg
and Seidel [137] and subjected to an investigation by quantitative 129Xe NMR
spectroscopy. Completely Zn-exchanged Y-zeolite could be obtained by solid-
state reaction between ZnCl2 and NH4-Y at 693 K in high vacuum. Zn-Y zeolites
obtained via solid-state ion exchange exhibited a more homogeneous cation dis-
tribution than samples conventionally prepared from aqueous zinc salt solu-
tions. Analysis of CO and Xe adsorption data in combination with 129Xe NMR
results, using a multisite adsorption model [138], yielded the population of sites
II and III by Zn2+ in the supercages. It was shown that up to an exchange degree,
d, of 50% only sites II were populated, whereas at higher Zn-contents up to
d = 100%, sites III were increasingly occupied (cf. Fig. 39). ‘Overexchange’
(d>100%) led to significant salt imbibition (cf. Sects. 5.1.5 and 5.2.3). An excess
of ZnCl2 in the mixture with the parent zeolite NH4-Y was occluded in the
supercages and, thus, decreased the accessibility of the charge-compensating Zn
cations for CO and Xe. A maximum concentration of zinc cations was reached
2+/n = 0.5 [137].
for a ratio of nZn Al
Solid-State Ion Exchange in Microporous and Mesoporous Materials 111

Fig. 39. Concentrations of Zn2+ cations at supercage positions in samples prepared by solid-
state ion exchange in ZnCl2/NH4-Y mixtures by heating the mixtures at a rate of 20 K min–1 to
393 K and maintaining there for 5 h and subsequently heating to 693 K for 24 h in high vacu-
um (final pressure: p £ 10–3 Pa); a = 2 · nZn2+/nAl · 100, at SIII sites: ●; at SII sites: ▲; at SIII +SII
sites: ■ (after [138], with permission)

EXAFS techniques were employed by Förster and Hatje [105] for the deter-
mination of the cation coordination in Zn-Y obtained via SSIE; in fact, in a sim-
ilar way and with similar conclusions as described for Cu-Y (cf. Sect. 5.3.2.1) and
Ni-Y (cf. Sect. 5.3.4.3). The authors assumed that upon SSIE Zn2+–O–Zn2+ com-
plexes were likely to form. The distribution of the Zn2+ cations over the various
sites of Y-type zeolites modified through solid-state reaction of ZnCl2 with H-Y
and NH4-Y was also investigated by far-infrared spectroscopy [81, 104].
In a study by Onyestyák et al. [139], it was also reported that zinc cations were
incorporated via SSIE into zeolite Y through reaction of ZnCl2 and NH4-Y.
In a recent contribution, Beyer et al. [140] carried out solid-state exchange
between metallic zinc (zinc dust) and hydrogen forms of faujasite-type zeolite
or mordenite. Ground mixtures of Zn0 (zinc dust) and the respective zeolites
were dehydrated at 523 K and then heated to 873 K in vacuum or, preferentially,
in a flow of nitrogen. XRD patterns and SEM micrographs taken from the mix-
tures before and after heat-treatment showed that the reflections of metallic zinc
and the zinc particles had disappeared upon heating of the mixture. Moreover,
comparison with the XRD patterns of Zn-Y samples obtained by conventional
ion exchange revealed changes in the reflections of the zeolite lattice which were
obviously due to an incorporation of Zn2+ cations into the zeolite structures. The
originally greyish color of the mixture changed to white. Quantitative evidence
for a solid-state exchange according to Eq. (11) was provided by measurements
of the released hydrogen in temperature-programmed evolution of hydrogen
(TPEH):
Zn0 + 2 H+Z– Æ Zn2+Z 2– + H2≠ (11)
An analogous reaction had been reported earlier by Jacobs et al. [141] and by
Sárkány and Sachtler [142] for the oxidation of small aggregates of Ag0 and Cu0
112 H.G. Karge · H.K. Beyer

in the interior of zeolite structures, respectively. Interestingly, and at some vari-

ance with the findings of Salzer [132, 133] and Roessner et al. [134], Beyer et al.
[140] reported that they did not observe SSIE between ZnO and acid zeolites
when they conducted pertinent control experiments.
The hydrogen evolved according to Eq. (11) was measured during heat-treat-
ment of the Zn0/NH4 ,Na-Y mixtures by a heat conductivity detector; alterna-
tively, it may be determined by mass spectrometry. IR spectroscopy with and
without pyridine as a probe demonstrated that, on heating the Zn0/NH4 ,Na-Y
mixture, the intensities of the OH bands originating from acidic OH groups and
those of the corresponding pyridinium ion bands decreased, while simultane-
ously bands due to pyridine attached to Zn2+ cations (at 1445 cm–1) developed.
Similar results were obtained with fine powders of cadmium, indium and galli-
um. Beyer et al. pointed out the potential of the described reactions for develop-
ing a new method to determine qualitatively and quantitatively acidic OH
groups in zeolites or related materials. Such a technique would use the amounts
of hydrogen released on temperature-programmed heating of the mixtures of
fine metal powders and acid solids as a measure of the density of Brønsted acid
sites. Moreover, it should provide a tool to distinguish Brønsted acid sites with
respect to the acidity strength according to the characteristic temperature peaks
seen during TPEH (cf. also [136] and Sect. 6.3).
Roessner et al. [134] found that Zn,H-ZSM-5 or Zn-ZSM-5 obtained via SSIE
possessed catalytic activities in n-hexane isomerization similar to those exhib-
ited by Zn,H-ZSM-5 catalysts that had been prepared by conventional methods,
i.e., either through ion exchange in aqueous solutions of Zn(NO3)2 or by the
incipient-wetness technique. The latter method seemed to be in between
exchange in aqueous solution and a solid-state reaction (cf. Sect. 6.1). Also, Zn-
ZSM-5 produced via SSIE proved to be almost as equally active in ethane arom-
atization as conventionally modified Zn-ZSM-5 prepared by exchange in aque-
ous Zn(NO3)2 solution [135]. Finally, as was shown by Rojasova et al. [143],
incorporation of zinc into NH4-Y by solid-state reaction with, e.g., ZnO yielded
catalysts active in n-hexane aromatization. Under equal conditions, ZnO alone
did not catalyze this reaction.
Microwave irradiation during solid-state reactions of ZnCl2 with Y-type zeo-
lites (H-Y or Na-Y) was employed in a series of studies by Yin et al. [144–146]
aimed at producing and testing Zn-containing Y zeolite catalysts for special
organic reactions. In the case of Na-Y, XRD and IR showed that ZnCl2 was com-
pletely dispersed onto the surface of the zeolite. Part of the dispersed salt (up to
25% ZnCl2 loading) underwent SSIE with the Na+ cations of Na-Y, the rest of
them coordinated with oxygen atoms of the Y-type framework. The catalysts
were used for the Diels-Alder reaction between myrcene and acrolein. At load-
ings above 25% ZnCl2 and up to 37.5%, when finely dispersed ZnCl2 particles
were left after SSIE, a significantly improved performance of the catalysts was
observed, suggesting that the activity and selectivity of those particles were even
higher than that of Zn-Y generated by SSIE. In the case of the system ZnCl2/
H-Y, the products of SSIE were used to catalyze the a-pinene conversion. The
selectivity for carvenol increased with increasing ZnCl2 loading, i.e., with de-
creasing Brønsted and increasing Lewis acidity.
Solid-State Ion Exchange in Microporous and Mesoporous Materials 113

Fig. 40. IR spectra of pure NH4-Y and after solid-state reaction of mixtures of cadmium com-
pounds with NH4-Y at 793 K and 10–2 Pa; a prior to and b, c after adsorption of H2S under
6.6 kPa at 298 K for 10 min and subsequent 10 min evacuation at 298 K (after [139], with per-
mission)

5.3.3.2
Introduction of Cadmium
In similar experiments as with zinc, Onyestyák et al. [139] reacted NH4-X, NH4-
Y and NH4-MOR with various cadmium compounds. The reactions were moni-
tored by IR spectroscopy. Again, very high degrees of exchange were achieved
with the chlorides of cadmium. In the case of Cd compounds, the degree of
exchange was shown to decrease in the sequence Cd(NO3)2 >CdCl2 >CdO>
CdS>CdSO4 (cf. Fig. 40).
The Cd-zeolites obtained in this manner were compared with conventionally
prepared ones with respect to dissociative adsorption of H2S, hydrosulfurization
of olefins and decomposition of ethylthiol (EtSH) or ethylthioether (Et2S). It
turned out that Cd-Y samples formed by SSIE were more active in hydrosulfur-
ization, dissociative adsorption of H2S and decomposition of EtSH and Et2S than
conventionally prepared Cd-zeolite catalysts.
5.3.3.3
Introduction of Mercury
As an example of solid-state ion exchange with mercury compounds, the result
of a reaction between Hg2Cl2 and H-ZSM-5 is reproduced in Fig. 41 [130]. The
114 H.G. Karge · H.K. Beyer

Fig. 41. Solid-state ion exchange in the system Hg2Cl2/H-ZSM-5 shown by the decrease in the
intensity of the OH bands of the parent zeolite (after [130], with permission)

mixture was prepared with a ratio of nHg /nAl = 1. From the decrease in the
absorbance of the OH band at 3610 cm–1 a degree of exchange of 70% was
derived. No attempt was made to enhance the degree of exchange by, e.g.,
increasing the nHg/nAl ratio, the reaction temperature (Treact) and/or the duration
of the solid-state reaction (treact). However, the most interesting result is that,
similar to the reaction in the system AgCl/H-ZSM-5 (vide supra), the salt com-
ponent was insoluble in water, but nevertheless the solid-state reaction took
place. This supports the earlier statement that the presence of water does not
play a decisive role in SSIE (cf. Sect. 7.1).

5.3.4
SSIE of Iron, Cobalt, Nickel and Manganese Compounds with Zeolites
5.3.4.1
Introduction of Iron
As early as in the work by Clearfield et al. [18] in 1973, solid-state reactions of
Fe2+, Co2+, Ni2+ and Mn2+ chlorides with ammonium (hydrogen) forms of zeo-
lites A, X and Y were studied to demonstrate the phenomenon of SSIE. Those
authors monitored the reactions through titration of HCl evolved. More recent-
ly, interest in zeolites containing these transition metals, especially cobalt and
Solid-State Ion Exchange in Microporous and Mesoporous Materials 115

iron, has been stimulated by possible applications of the modified zeolites in

catalysis (vide infra). Therefore, the possibility of preparation of such catalysts
via solid-state ion exchange was explored in more detail.
In 1987, solid-state interactions of Fe cations with ZSM-5 were studied by
Wichterlová et al. [147]. These authors employed mixtures of Fe2O3 and H-ZSM-5
or Na-ZSM-5, which were heat-treated in a stream of oxygen. The results ob-
tained with the zeolites prepared in this way were compared with those from
conventionally exchanged Fe-containing ZSM-5 samples. Measurements of Xe
adsorption on the parent zeolites and on the materials obtained via solid-state
interaction yielded the same adsorption capacities indicating that no structural
degradation or plugging of the zeolite channels had occurred during the proce-
dure of post-synthesis modification.
No indication of solid-state ion exchange was observed when Wichterlová et
al. [147] used the sodium form of ZSM-5: calcination of a Fe2O3/Na-ZSM-5 mix-
ture did not give rise to an ESR signal indicating isolated Fe3+ ions. In contrast,
in the ESR spectra of samples produced via calcination of mixtures of Fe2O3 with
the hydrogen form of ZSM-5, a signal appeared at g = 4.27 which was ascribed
to Fe3+ ions isolated in the zeolite matrix and in tetrahedral coordination. In fact,
the intensity of the ESR signal was significantly lower than in the case of con-
ventionally exchanged Fe-containing ZSM-5 samples. However, those Fe3+ ions
were not believed to be part of the framework. Rather, from the ESR results and
the observed decrease of the concentration of acid OH groups as evidenced by
temperature-programmed desorption of ammonia, the authors concluded that
solid-state interaction in the system Fe2O3/H-ZSM-5 led to an incorporation of
trivalent Fe3+ ions in extra-framework cation sites, but in tetrahedral coordina-
tion. This is at variance with the report of Kucherov and Slinkin [148] (vide
infra). The Fe3+ cations in Fe,H-ZSM-5, as described by Wichterlová et al. [147]
after solid-state interaction in the system Fe2O3/H-ZSM-5, were easily reduced
in vacuum. Part of the tetrahedrally (Td) coordinated extra-framework Fe3+
cations introduced via SSIE retained this ligand field symmetry (Td) even after
adsorption of H2O. The catalytic activity in toluene disproportionation and
methanol conversion was lower than that of the acidic parent zeolite, H-ZSM-5,
but comparable to that of Fe-containing ZSM-5 catalysts that had been conven-
tionally prepared from H-ZSM-5 suspended in aqueous solutions of FeCl3 .
Incorporation of trivalent Fe3+ into H-ZSM-5 and Fe2+ into NH4-Y was exten-
sively studied by Kucherov and Slinkin [148] and Lazar et al. [149, 150], respec-
tively. The former authors investigated the possibility of Fe3+ introduction into
H-ZSM-5 by calcination of mixtures of FeCl3 , FeO or Fe3O4 and the zeolite. The
heat-treatment of the mixtures was carried out in air or vacuum at temperatures
up to 1073 K for 4 h. ESR spectroscopy was employed to check whether or not
SSIE had occurred. Isolated Fe3+ cations were not detected after calcination of Fe
oxides with H-ZSM-5. The authors suggested that this was due to the high lattice
energy of these compounds and/or the low mobility of the ions in the oxide lat-
tices even at 1073 K. However, Kucherov and Slinkin [148] provided evidence for
a migration of Fe3+ cations into the zeolite matrix when they reacted the chlo-
ride, FeCl3, with H-ZSM-5 at 793 K. It was shown that the Fe3+ cations introduced
in this fashion were isolated (not clustered) and resided in a strong crystal field
116 H.G. Karge · H.K. Beyer

of low symmetry. Even at 1073 K, XRD did not indicate any amorphization of the
zeolite structure upon calcination of the mixture FeCl3/H-ZSM-5. Since the ESR
signals were strongly affected by admission of gases and vapors such as O2 , NH3 ,
H2O, pyridine and p-xylene (vide infra), it was obvious that the Fe3+ cations
incorporated via SSIE were accessible to adsorbates.
Fe,H-ZSM-5 materials obtained via SSIE with FeCl3 were compared with the
ferrisilicate analog of H-ZSM-5, i.e., H-[Fe]-ZSM-5 (nSi/nFe ª 50), where Fe3+ was
incorporated into the MFI framework by isomorphous substitution during syn-
thesis and, therefore, tetrahedrally coordinated [148, 151]. In fact, both Fe3+-con-
taining materials, viz., Fe,H-ZSM-5 with Fe residing on extra-framework cation
positions and H-[Fe]-ZSM-5 with Fe in tetrahedral positions in the framework,
gave rise to similar ESR spectra, even though they showed some differences in
the g-values. The main ESR signal of Fe,H-ZSM-5 appeared at g1 = 4.27, which
is close to g1 = 4.25 as observed for H-[Fe]-ZSM-5 by Kucherov et al. [148, 151],
as well as by Wichterlová et al. [147] for Fe,H-ZSM-5. More severely oxidized
Fe,H-ZSM-5 samples exhibited additional ESR lines at g2 = 5.65 and g3 = 6.25; the
corresponding g-values for the ferrisilicate analog H-[Fe]-ZSM-5 were g2 = 5.2
and g3 = 7.9. However, the Fe,H-ZSM-5 materials behaved completely differently
from H-[Fe]-ZSM-5: (i) Upon admission of the above-mentioned adsorbates,
the ESR spectrum of the former zeolite (prepared via SSIE) dramatically
changed (vide infra), whereas that of H-[Fe]-ZSM-5 remained practically the
same. (ii) Fe3+ on cation sites of Fe,H-ZSM-5 could be replaced by solid-state
reaction with CuO, i.e., the spectrum of Cu2+ in ZSM-5 developed (cf.
Sect. 5.3.2.1). This did not happen when a mixture CuO/H-[Fe]-ZSM-5 was sub-
jected to heat-treatment. (iii) In contrast to H-[Fe]-ZSM-5, the sample prepared
via SSIE, i.e., Fe,H-ZSM-5, exhibited an anomalous temperature effect, i.e., upon
cooling to 77 K, the intensity of the main signal at g1 = 4.27 was very much
enhanced (cf. Fig. 42), which is typical of Fe3+ on extra-framework cation sites.
These criteria (i)–(iii) enable us to distinguish between Fe3+ in extra-frame-
work and Fe3+ in (tetrahedrally coordinated) framework sites. This would even
hold if Fe3+ occurred in tetrahedrally coordinated extra-framework positions.
However, in the opinion of Kucherov and Slinkin [148], the occurrence of the
anomalous temperature effect strongly suggests that Fe3+ introduced via SSIE
was in fact not located in tetrahedral but rather in distorted octahedral environ-
ment, in contrast to the conclusion by Wichterlová et al. [147] (vide supra).
The interaction of Fe3+ in Fe,H-ZSM-5 with NH3 and pyridine led to a com-
plete disappearance of the low-field lines at g2 = 5.65 and g3 = 6.25, and interac-
tion with H2O to their considerable decrease. In any event, the intensity at g1 =
4.27 was markedly enhanced. This was especially pronounced with NH3 and
pyridine indicating an increase of the crystal field symmetry upon adsorption
of these powerful ligands. Interaction with O2 resulted in a considerable but
reversible broadening of the Fe3+ ESR lines caused by dipole-dipole interaction
of Fe3+ with O2 . With NH3, the samples of Fe,H-ZSM-5 were reduced at higher
temperatures (823 K) as indicated by the disappearance of the signals of Fe3+
and formation of Fe0 clusters. Reoxidation did not fully restore the original spec-
trum. Interaction with p-xylene yielded an ESR spectrum characteristic of p-
xylene cation radicals.
Solid-State Ion Exchange in Microporous and Mesoporous Materials 117

Fig. 42. Changes in the Fe3+ ESR signal upon calcination of the mixture FeCl3/H-ZSM-5
(2.3 wt.% FeCl3) a at 293 K; b after heat treatment in vacuum at 593–793 K for 2 h; c sample
(b) measured at 77 K (after [148], with permission)

From the spin density, Kucherov and Slinkin [148] derived an ion concentra-
tion of 1 Fe3+ per 30–80 Al. They assumed that it is difficult to compensate by
one Fe3+ three rather separated negative charges of the framework and hypo-
thesized that not a naked Fe3+ but one complex cation such as FeCl+2 neutralizes
one negative charge. Upon oxidative calcination a gradual substitution of the
anionic ligand gives rise to a transformation FeCl+2 Æ FeO+ accompanied by the
appearance of signals with g2 = 5.65 and g3 = 6.25. As the signal at g1 = 4.27 was
assigned to isolated Fe3+ cations in tetrahedral or orthorhombic coordination
stabilized in a crystal field of low symmetry, the development of signals with
g > 4.3 (cf. the above g2 and g3 signals) was assumed to be due to further lower-
ing of the symmetry of the environment of the ions.
Kucherov and Slinkin also showed that in the solid-state reaction between
CuO and Fe,H-ZSM-5 (vide supra) at least 99% of the Fe3+ ions were replaced by
Cu2+ cations. The ESR signals with g2 = 5.65 and g3 = 6.25 were entirely elimi-
nated, and only a trace of the signal with g1 = 4.27 was left. This shows that essen-
tially all of the Fe was in extra-framework positions (vide supra) and no or only
negligible insertion of Fe3+ into the framework had occurred. Thus, SSIE of FeCl3
and H-ZSM-5 did not result in any isomorphous substitution.
The conventional ion exchange in aqueous solutions of easily oxidizable
cations such as, e.g., bivalent Fe2+, might require the exclusion of oxygen during
the whole exchange procedure. This frequently leads to experimental compli-
cations that possibly can be avoided by solid-state ion exchange (cf. Cu+,
Sect. 5.3.2.1). Therefore, the incorporation of Fe2+ into zeolites via SSIE was
118 H.G. Karge · H.K. Beyer

investigated in more detail. Similar to earlier work on conventionally exchanged,

Fe-containing zeolites (cf., e.g., [152, 153]), in the pertinent studies by Lázár et
al. [149, 150] on SSIE of Fe2+ into zeolites, Mössbauer spectroscopy was applied
as a powerful diagnostic tool. However, XRD and TPE were also employed. In
these studies, an NH4-Y zeolite prepared from Na-Y by almost complete (99.9%)
exchange in aqueous NH4Cl solutions was used as a starting material. It was
ground with FeCl2 ◊ 4H2O (nFe/nAl, tetr. = 0.5) followed by heat-treatment at tem-
peratures between 420 and 720 K. Most of the sample preparation steps were
undertaken without exclusion of ambient atmosphere; subsequent Mössbauer
and TPE experiments, however, were carried out in vacuum.
Schematic XRD patterns of the parent NH4-Y zeolite and its mixture with
FeCl2 ◊ 4H2O as prepared by grinding at ambient temperature and after heat-
treatment are presented in Fig. 43. As can be seen in this figure, the incorpora-
tion of iron (cf. patterns a and b) led to a change of the absolute intensities of the
reflections of the lattice, partly caused by changes of the respective structure fac-
tors [54] and partly by the higher absorption factor of Fe. However, the intensi-
ties re-increased upon heat-treatment, and no broad peak at 2Q = 20–30° was
observed which would have been indicative of amorphous silica. Both these
findings proved that the zeolite lattice did not measurably deteriorate upon

Fig. 43. Schematic representation of XRD patterns of a the parent NH4-Y zeolite, a = 24.785 Å;
b the mixture FeCl2 ◊ 4H2O/NH4-Y ground in air at ambient temperature, a = 24.742 Å; c the
material (b) heated in air up to 720 K (heating rate: 10 K min–1), a = 24.542 Å; *FeCl2 ◊ 4 H2O;
**NH4Cl
Solid-State Ion Exchange in Microporous and Mesoporous Materials 119

SSIE. The most striking features in the XRD patterns were, however, the appear-
ance of the (110) reflections of crystalline NH4Cl at Q = 32.68° and weakening of
the reflections of FeCl2 ◊ 4H2O in pattern b and their complete disappearance
upon heat-treatment at 720 K (cf. pattern c). These features confirmed that sol-
id-state ion exchange had indeed occurred.
The system FeCl2/NH4-Y turned out to be a very interesting example of the
application of Mössbauer spectroscopy in the field of solid-state ion exchange in
zeolites [149, 150]. The spectra (cf. Fig. 44) were decomposed (vide infra; cf., as
an example, Fig. 45), and the oxidation states and coordination of incorporated
iron were deduced on the basis of assignments reported earlier [154]. Results are
presented in Table 10. When the above mixture was investigated as prepared, the
Mössbauer spectrum provided evidence that 57% of the iron was oxidized to the
trivalent state (cf. Table 10, RI = 57). From the Mössbauer parameters it was con-
cluded that the Fe(III) species were almost perfectly octahedrally coordinated.
The remaining non-oxidized iron occurred as three different species, viz.,
(i) partially dehydrated Fe(II) chloride (21%); (ii) also partially dehydrated,

Table 10. Mössbauer parameters of iron species present in FeCl2/NH4-Y mixtures after grind-
ing and subsequent heat-treatment in high vacuum. IS, isomer shift; QS, quadrupole splitting;
RI, relative intensity

Species Parameter As-prepared After treatment in high vacuum at

420 K 520 K 620 K 720 K

Fe(III)oct IS 0.37 0.35 0.31 0.30 0.33

QS 0.63 0.72 0.75 0.72 0.58
RI 57 44 38 5 2
Fe(III)trig IS – – 0.25 0.23 0.22
QS – – 1.53 1.71 1.73
RI – – 11 11 2
Fe(II)tetr IS 0.69 0.72 – – –
QS 0.35 0.36 – – –
RI 13 11 – – –
Fe(II)trig IS – – 1.08 0.88 0.92
QS – – 0.78 0.62 0.68
RI – – 21 21 30
Fe(II)oct-1 IS 0.83 0.79 – – –
QS 1.95 2.00 – – –
RI 9 22 – – –
Fe(II)oct-2 IS – – – 0.95 0.99
QS – – – 2.20 2.13
RI – – – 37 45
Fe(II)oct-3 IS – – 1.06 1.23 1.27
QS – – 2.60 2.19 2.16
RI – – 9 26 20
FeCl2 ◊ xH2O IS 1.13 1.04 1.06 – –
QS 1.83 1.79 1.90 – –
RI 21 23 21 – –
120 H.G. Karge · H.K. Beyer

Fig. 44. Mössbauer spectra of a FeCl2 · 4H2O/NH4-Y mixture. a ground in air at ambient tem-
perature; material (a) after heat treatment in vacuum at b 420 K, c 520 K, d 620 K, and e 720 K
(after [149], with permission)

tetrahedrally coordinated Fe(II) ions (13%) residing in the small and large
cavities with probably one H2O molecule in their coordination shell (Fe(II)tetr);
and (iii) an octahedrally coordinated Fe(II) species (Fe(II)oct-1 , 9%), where the
octahedral environment included H2O and OH ligands.
When, however, the FeCl2/NH4-Y mixture was heated in vacuum to increas-
ingly higher temperatures (420, 520, 620, 720 K) and the Mössbauer measure-
ment conducted after cooling to 300 K, the set of spectra shown in Fig. 44 was
obtained.
For the spectrum obtained after heat-treatment at 720 K, Fig. 45 illustrates the
decomposition into individual signals.
From the data of the whole set of spectra accumulated in Table 10, one real-
izes that, after heat-treatment at 720 K, only about 4% of Fe(III) species were left
and an almost pure Fe(II)-Y zeolite was obtained exhibiting Fe(II) in trigonal
(30%) or octahedral (65%) coordination. The octahedrally coordinated iron
was either Fe(II)oct-2 , being due to interaction with oxidic extra-framework Al-
containing species (released from the framework during auto-reduction, vide
infra) or Fe(II)oct-3 , being due to coordination with framework oxygen atoms in
the hexagonal prisms of the faujasite structure.
Solid-State Ion Exchange in Microporous and Mesoporous Materials 121

Fig. 45. Mössbauer signals of individual iron species giving the best fit to the spectrum in
Fig. 44e of ground FeCl2 · 4H2O/NH4-Y heat-treated at 720 K. (●) Fe(III)oct-1 ; (■) Fe(III)trig ; (ƒ)
Fe(II)trig ; (▼) Fe(II)oct-2 ; (ƒ) Fe(II)oct-3 ; QS quadrupole splitting; IS isomer shift (after [149],
with permission)

TPE investigations of the system FeCl2 ◊ 4H2O/NH4-Y, where the evolved H2O,
NH3 and HCl were monitored by mass spectrometry, revealed some special fea-
tures: (i) besides the usual H2O peak around 400 K originating from the release
of adsorbed water, a second one at 520 K was detected (cf. Fig. 46).
This second peak was ascribed to H2O molecules stemming from hydroxyl
groups formed intermittently through hydrolysis in which Fe cations and crys-
tal water were involved; (ii) at temperatures above 520 K, the evolution of ammo-
nia declined faster than that of HCl (cf. Fig. 46). The resulting small but signi-
ficant and reproducible delay in HCl evolution is a peculiarity of the system
FeCl2 ◊ 4H2O/NH4-Y in SSIE chemistry and was not observed with, e.g., CaCl2 ◊
2H2O/NH4-Y. This peculiar behavior can be explained in view of the results
obtained by Mössbauer spectroscopy (vide supra):
1. During grinding and/or subsequent heating hydrolysis occurred under for-
mation of hydroxy iron cations according to Eq. (12).
Fe2+ + H2O Æ Fe(OH)+ + H+ (12)
2. Mössbauer spectroscopy (vide supra) provided evidence for oxidation of Fe2+
to Fe3+; the latter may also be present as hydroxyl cations, e.g., Fe(OH)2+ or
Fe(OH)+2 .
122 H.G. Karge · H.K. Beyer

Fig. 46. Curves of temperature-programmed evolution of HCl (m/e = 36, –––), NH3 (m/e = 17,
– – –) and H2O (m/e = 18, – · – · –) evolved A from a ground mixture of FeCl2 ◊ 4 H2O/NH4-Y
(nFe2+/nAl = 0.5) and B from crystalline NH4Cl (after [149], with permission)

3. These as well as Fe(OH)+ may react with HCl formed by decomposition of

NH4Cl, resulting in the observed evolution of H2O and NH3 at 520 K:
Fe(OH) +n–1 + (n–1) HCl Æ FeCl +n–1 + (n–1)H2O (13)
where n is the valence state of the iron (i.e., n = 2 or 3).
4. Finally, FeCl+n-1 undergoes at somewhat higher temperatures a solid-state
reaction with Brønsted acid sites according to:
FeCl+n–1 + (n–1)H+ Æ Fen+ + (n–1)HCl (14)
In summary, a fraction of the HCl which had formed by decomposition of the
SSIE product, viz., NH4Cl, is consumed by reaction (13) and only at higher tem-
peratures released according to Eq. (14). This explains the shift of the HCl evo-
lution curve to higher temperatures compared with the evolution curve of NH3
(cf. Fig. 46). It is striking that the decomposition of ammonium chloride
observed in TPE experiments around 520 K and the drastic change in the oxi-
dation state of iron at 520–620 K as revealed by Mössbauer spectroscopy (vide
supra, Table 10) occurred in almost the same temperature range.
The interest in iron exchange into zeolites, particularly into ZSM-5 (MFI-
type) zeolites, was stimulated by the possible application of the products in
DENOX processes. Thus, incorporation of iron by solid-state ion exchange was
Solid-State Ion Exchange in Microporous and Mesoporous Materials 123

discussed in a series of recent papers in relation to adsorption and catalytic con-

version of nitrogen oxides.
Fe,H-ZSM-5 samples with different Fe-loadings were prepared by Bell and
co-workers via solid-state ion exchange and characterized by IR spectroscopy of
NO, TPD of NO and NH3 and TPR with H2 [155]. At low loadings, Fe cations
exchanged with protons of the acid Brønsted centers on a one-to-one basis. NO
adsorption and TPR experiments suggested that the primary form of incorpo-
rated iron was Fe3+, i.e., Fe3+(OH)–2 . At loadings above nFe2+/nAl ≈0.56, small FeOx
particles were formed (vide infra, cf. [156, 157]) and at increasingly higher load-
ings the concentration of Fe2+ cations was found to increase because of auto-
reduction. Upon NO adsorption at room temperature, three different types of
Fe2+ sites were observed. Furthermore, the interaction of NO, NO2 and O2 as well
as the reduction of NO by C3H8 on the modified zeolites was investigated.
A high activity of Fe-MFI catalysts produced by SSIE in NO decomposition,
even though lower than that of conventionally exchanged materials, was re-
ported by Varga et al. [118]. Turek and colleagues [156–160] reported on NOx
decomposition and/or reaction under various conditions over Fe-MFI or Fe-
FER (ferrierite) produced via SSIE. The activity for NO oxidation was inhibited
by the presence of SO2 or H2O, but could be completely restored by removal of
the poisons. Fe-FER was more active but also more sensitive to poisoning than
Fe-MFI [158]. Kögel et al. [160] have found that Fe-ZSM-5 prepared through sol-
id-state ion exchange in the presence of air exhibits activities in simultaneous
catalytic reduction of NO and N2O comparable to those of Fe-ZSM-5 catalysts
obtained under anaerobic conditions. Upon reduction of NO by iso-butane over
Fe-MFI catalysts obtained by SSIE, HCN was detected as a substantial product
[159]. The catalysts produced by SSIE with FeSO4 · 7H2O to various Fe-contents
[156, 157] exhibited a high activity in decomposition of N2O in the absence of
reducing agents, but were also active in reduction of NOx with hydrocarbons,
especially propane. Above a ratio of nFe2+/nAl ≈ 0.5, however, iron was no longer
incorporated as catalytically active Fe2+species, but hematite was formed from
the Fe2+ excess (vide supra, [155]), which was inactive in both NO and N2O
reduction.
Incorporation of iron into zeolites by SSIE was also studied by Varga et al.
[114, 118, 121] and Dandl [161] in relation to catalytic decomposition and
reduction of NO and N2O. Varga et al. [114] reported the preparation of Fe-,
Co- and Ni-containing ZSM-5 zeolites by SSIE for NO adsorption and decom-
position (see Sects. 5.3.4.2 and 5.3.4.3). Preparation and characterization were
identical to those reported by the same authors for Cu-ZSM-5 (SSIE) in
Sect. 5.3.2.1. Also, the results were similar, except for some differences in NO
adsorption.
The properties of Fe-[Ga]ZSM-5 prepared by SSIE of FeCl3 and the gallium
analog of H-ZSM-5, i.e., H-[Ga]ZSM-5, were essentially the same as those of
the corresponding Fe-ZSM-5 (i.e., Fe-[Al]ZSM-5) with Al instead of Ga in the
framework [128, 129]. The incorporated Fe Cl2+ cations or, after calcination in air,
FeO+ cations were replaced by Cu2+ when a mixture of CuO and Fe-[Ga]ZSM-5
was heated. However, by Kucherov et al. [128, 129], the Brønsted acid sites of H-
[Ga]ZSM-5 provided less stable traps for the introduced cations.
124 H.G. Karge · H.K. Beyer

Crocker et al. [28] reported on the incorporation of Fe3+ and Al3+ into mont-
morillonite (MM) via solid-state reaction with Fe(NO3)3 · 9H2O and Al(NO3)3 ·
9H2O, respectively. Oven-dried Na-MM was co-ground with the metal salt at
room temperature for 30 min in air, subsequently dried in vacuum (1 Pa; 24 h)
and then analyzed by XRD. Ion exchange was indicated by the appearance of a
reflection at 2Q = 29.4°, characteristic of NaNO3. Furthermore, reflections typi-
cal of the starting nitrate [Fe(NO3)3 ◊ 9H2O] were no longer observed in the XRD
pattern of the ground mixture. From their results, the authors concluded that
SSIE had occurred according to Eq. (15), yielding a high degree of exchange:
3 Na-(MM) + Fe(NO3)3 Æ Fe-(MM)3 + 3 NaNO3 (15)
After careful drying, the materials prepared in this way exhibited significant
amounts of Brønsted and Lewis acid sites as evidenced by IR spectroscopy and
pyridine adsorption. With Al(NO3)3 ◊ 9H2O similar results were obtained. The
authors pointed out that the acidity, in particular the high Lewis acidity, may
render the Fe-MM (and Al-MM) materials produced via SSIE valuable catalysts
for Lewis acid catalyzed Friedel-Crafts acylation and alkylation reactions.
Preparation of such catalysts by a solid-state reaction should be preferred over
conventional ion exchange since, at least on a large scale, ion exchange with clays
in aqueous solutions is hampered by the tendency of aqueous clay suspensions
to form intractable gels.

5.3.4.2
Introduction of Cobalt

Sachtler and co-workers [162] studied the redox chemistry of cobalt ions intro-
duced into MFI via SSIE with nCo/nAl = 0.4–1.0, employing IR, ESR and UV-Vis
diffuse reflectance spectroscopy. The coordination of Co2+ was found to be tetra-
hedral; the ions were in their high-spin state, detectable at 60 K.
In a recent contribution, Enhbold et al. [109, 163] investigated the incorpora-
tion of cobalt into clinoptilolite (CLIN) via solid-state reaction of CoCl2 or
Co(NO3)2 with hydrogen and sodium forms of this zeolite obtained through
repeated conventional exchange of natural clinoptilolite (from Tzaaga, Mongo-
lia) with 1 M aqueous solutions of NH4Cl and NaCl, respectively. The degree of
subsequent solid-state ion exchange was determined by chemical analysis of the
starting materials and, after thorough washing, by back-exchange of Co2+ with
NH+4 (cf. Sect. 4.1). The degree of exchange was studied as a function of reaction
temperature (Treact), reaction time (treact) and the type of cation (Na+, H+) of the
parent clinoptilolite. Moreover, the process of SSIE upon heat-treatment of mix-
tures of the cobalt salts and Na-CLIN or NH4(H)-CLIN was evidenced by for-
mation of NaCl or NaNO3 crystallites and HCl as determined by XRD and chem-
ical analysis, respectively. In fact, introduction of cobalt cations was achieved,
but only into those channels of the heulandite-like structure which are formed
by 10-membered rings; Co2+ could not enter the channels built by 8-membered
oxygen rings. It was shown that SSIE rendered preparation of Co-CLIN possible
with cobalt contents similar to that of conventionally exchanged materials. SSIE
with clinoptilolite occurred not only with the hydrogen but also with the sodi-
Solid-State Ion Exchange in Microporous and Mesoporous Materials 125

Fig. 47. Degree of solid-state ion exchange of cobalt into clinoptilolite as a function of the ratio
nCo2+/nOH (nCo2+/nNa+). 1 H-form of clinoptilolite; 2 Na-form of clinoptilolite (after [163], with
permission)

um form, even though a higher degree of exchange was reached with H-CLIN
(cf. Fig. 47).
Surprisingly, the amount of Co2+ incorporated into the clinoptilolite structure
was not very dependent on the cobalt content in the mixture (cf. Fig. 47). A very
fast replacement of the Na+ and H+ cations by Co2+ was observed in the initial
stage of SSIE. The subsequent steady state degree of SSIE was temperature-
dependent (cf. Fig. 48).
Both XRD and IR spectroscopy of the lattice vibrations confirmed that SSIE
did not cause a loss of crystallinity. Only slight differences between the IR spec-
tra of Co-CLIN samples prepared via solid-state ion exchange on the one hand
and conventional ion exchange on the other were found. No formation of cobalt
hydroxide or oxide was detected. At 373 K, SSIE with Co(NO3)2 resulted in an
exchange degree twice as high as with CoCl2 in accordance with the lower lattice
energy of the former salt (cf. Sects. 5.1.3 and 5.4.3). Adsorption isotherms mea-
sured with the parent Na-CLIN and Co-CLIN (prepared through SSIE) as adsor-
bents suggested that the micropore volume was lower in the case of Co-CLIN
due to partial blockage of the 10-membered ring channels. No formation of
mesopores was observed.
Introduction of cobalt and nickel into zeolites by solid-state ion exchange was
investigated by Jentys and colleagues and compared with preparations via con-
ventional exchange in aqueous solution, impregnation and direct synthesis
([164–170], vide infra). For their studies, the authors employed TPD of NH3
(TPDA), IR spectroscopy, X-ray absorption spectroscopy (XANES and EXAFS)
and XRD. TPDA revealed that 100% of the strong Brønsted acid sites were elim-
inated at an nCo/nAl ratio of ca. 1.0 and 0.5 in the case of SSIE of CoCl2 in mix-
tures with H-ZSM-5 and NH4-Y, respectively. Therefore, the authors concluded
126 H.G. Karge · H.K. Beyer

Fig. 48. Solid-state ion exchange of cobalt into clinoptilolite as a function of reaction time and
temperature. A Amount of Co2+ incorporated per gram; B degree of exchange (after [163],
with permission)

that in the case of ZSM-5 the negative charge of the framework originating from
the Al content was balanced not by bivalent Co2+ cations but by monovalent
CoCl+ complexes. This was ascribed to the fact that, due to the high nSi/nAl ratio
in their H-ZSM-5 samples (nSi/nAl = 26), the distance between two AlO4/2 – tetra-
2+
hedra was too large as to be neutralized by one Co cation (cf., e.g., Sects. 5.2.2
and 5.3.5). No attempt was reported [164, 165] to confirm the existence of
charge-balancing CoCl+ species by chemical analysis for chlorine of the Co-
ZSM-5 samples prepared via solid-state ion exchange. However, the interpreta-
tion given by Jentys et al. was supported by EXAFS results. These indicated in
Co-ZSM-5 obtained through SSIE a coordination number of N ª 1 for CoCl and
a distance Co–Cl slightly shorter than in CoCl2 where N = 4 was observed. In
contrast, in H-Y (nSi/nAl = 2.5), the spatial separation of AlO4/2 – tetrahedra
2+
allowed neutralization of two negative charges by only one Co (cf. Sect. 5.2.2,
SSIE of La3+ into H-Y vs. H-ZSM-5). This holds true even for lower concen-
trations of CoCl2 in the mixtures, corresponding to 0.25 £ nCo/nAl £ 0.5; for nCo/
nAl ª 0.25, some Co cations replaced only one proton. This may be the preferred
SSIE ratio at very low cobalt salt contents in the mixture.
The IR spectra (with and without application of pyridine as a probe mole-
cule) of H-ZSM-5 and Co-ZSM-5 prepared via SSIE confirmed the suggestion
that in the case of nCo/nAl ª 1.0 the acid Brønsted OH groups were completely
consumed. Two types of newly formed acid sites, most likely Lewis acid sites,
were indicated, viz., CoCl+ species (vide supra) and possibly ‘true Lewis sites’,
i.e., Al-containing extra-framework species (cf. [171, 172]).
Microcalorimetry and XPS measurements of NH3 adsorption were used by
Auroux et al. [106] to characterize Co-ZSM-5 obtained through solid-state reac-
tion in a similar way as reported for introduction of copper (vide supra) and
Solid-State Ion Exchange in Microporous and Mesoporous Materials 127

nickel (vide infra). After ion exchange, a strong increase in Lewis acidity was
observed. The results were compared with those observed with Co-ZSM-5 pro-
duced through conventional exchange or impregnation.
Jentys et al. [168] also compared the properties of Co-ZSM-5 prepared by dif-
ferent methods such as conventional exchange in aqueous solutions of cobalt
salts, impregnation and solid-state ion exchange. It is worth mentioning that
with samples produced by the impregnation method metal oxide clusters were
observed, mainly located on the external surface of the zeolite crystallites [166].
In contrast, SSIE resulted in a highly dispersed distribution of cobalt cations
inside the channels. Over reduced bifunctional Co,H-ZSM-5 catalysts prepared
via SSIE, high overall activity in hydroconversion of n-heptane and highest
selectivity to isomerization to iso-C7 were observed by Lugstein et al. [168].
Main products were, besides 2-methylhexane, iso-butane and propane. An in-
creased metal loading gave rise to pronounced activity in hydrogenolysis upon
hydroconversion of n-heptane [167]. Furthermore, it was found that Co-con-
taining zeolites produced via SSIE between H-Y or H-ZSM-5 and cobalt salts
were active catalysts in thiophene hydrodesulfurization [166].
Another interesting catalytic application of Co-containing zeolites was
reported by Li and Armor [173]. These authors used dealuminated H-ZSM-5, H-
BETA and H-Y, modified by solid-state reaction with solid cobalt salts or con-
ventional exchange in cobalt salt solution, as catalysts for ammoxidation of
ethane to acetonitrile. Even though the catalysts prepared by SSIE had, in gener-
al, a lower nCo/nAl ratio, they produced under equal conditions more acetonitrile
and also showed higher selectivity for incorporation of NH3 into acetonitrile.
Introduction of cobalt into NH4-Y via SSIE has also been reported by Varga et
al. [114] and Onyestyák et al. [139]. In the first publication [114], it was reported
that Co-ZSM-5 (SSIE) could bind oxygen only loosely, so that oxygen could be
easily removed and re-adsorbed in contrast to Cu-ZSM-5. In the opinion of the
authors, this behavior renders Co-ZSM-5 produced via SSIE a better catalyst for
selective reduction of NO by methane in the presence of oxygen than Cu-ZSM-
5. In fact, Co-ZSM-5 prepared via SSIE was also employed as catalyst for catalyt-
ic NO decomposition and/or reduction by, e.g., propylene (cf. [121, 174]) and
other hydrocarbons such as iso-C4H10 in O2-rich or O2-free streams, both with
dry and wet feeds. Wang et al. [175] compared the properties of Co-containing
ZSM-5 (or ferrierite) catalysts that had been prepared by various methods, viz.,
conventional ion exchange in aqueous solution of Co-salts, impregnation, solid-
state ion exchange and sublimation (cf. Sect. 6.3.1).

5.3.4.3
Introduction of Nickel

Solid-state ion exchange between H-ZSM-5 (nSi/nAl = 13.6, sample I and 22.5,
sample II) and nickel compounds such as NiCl2 , NiSO4 , Ni(CH3COO)2 and NiO
was studied by Wichterlová et al. [176]. In their investigations, these authors
employed IR spectroscopy and TPDA to monitor changes in the concentration
of Brønsted acid sites, mass spectrometry and back-titration to determine the
gases evolved upon SSIE, e.g., HCl and decomposition products of acetate
128 H.G. Karge · H.K. Beyer

anions.With the system NiCl2/H-ZSM-5 it was observed that almost 100% of the
acid OH groups were eliminated through solid-state reaction in a mixture with
nNi /nOH = 0.5 at 770 K in a flow of oxygen. In a mixture with nNi /nOH = 0.33, 64
and 66% of the initial OH groups were consumed at 670 and 770 K, respectively.
These IR results were confirmed by back-titration of evolved HCl. Wichterlová
et al. [176] concluded that SSIE of both H-ZSM-5 samples with NiCl2 resulted in
neutralization of two Brønsted acid OH groups by one Ni2+ cation. This finding
is different from the interpretation of SSIE of bivalent cobalt with H-ZSM-5
(vide supra), where a ratio of nCo/nAl ª 1.0 was required for complete exchange
and the charge balance was assumed to be achieved by replacement of the pro-
ton of one Brønsted acid OH group by one CoCl+ species. The reason for this dif-
ferent behavior of the systems CoCl2/H-ZSM-5 and NiCl2/H-ZSM-5 in solid-
state reaction has not yet been clarified. Wichterlová et al. [176] demonstrated
further that via back-exchange of Ni-ZSM-5 (produced by SSIE) with NH4NO3
solution and subsequent deammoniation the original density of Brønsted acid
OH groups was almost completely restored. Similarly, a full regeneration of the
original Brønsted acid OH groups was accomplished when the Ni2+ cations
introduced via SSIE were subsequently reduced by hydrogen. The fact that the
original OH groups were regained in both cases showed that neither SSIE nor
subsequent reduction affected the integrity of the zeolite structure or caused a
measureable dealumination of the framework. Solid-state reaction of NiSO4
with H-ZSM-5 (nNi /nOH = 0.5) resulted in a consumption of only 50% of the acid
OH groups, whereas SSIE did not proceed at all in the systems Ni(CH3COO)2 /
H-ZSM-5 or NiO/H-ZSM-5. Nickel acetate decomposed at temperatures above
520 K into CO, CO2 , H2 , CH3COOH, CH3COCH3 and NiO. NiO did not react, most
likely because of its rather high lattice energy.
Introduction of Ni2+ into Y-type zeolite by SSIE was proven by XAS in exper-
iments similar to that carried out with Cu+ and Zn2+ (vide supra, [105]). In the
case of Ni compounds it was observed that solid-state ion exchange started
already during compression of the powdered components. Analogously to the
cases of Cu+ and Zn2+, the authors tentatively assumed that upon SSIE Ni–O–Ni
complexes formed.
Partial structures of fully dehydrated Ni-containing Y-zeolite, which was pre-
pared via SSIE, and of its D2O sorption complex, were determined by pulsed-
neutron diffraction in a study by Haniffa and Seff [177]. In Ni(30)-Y, with the
unit cell composition of Ni30Na7Cl12Al55Si137O384 , the Ni2+ cations occupy crys-
tallographically different positions, viz., I, I¢ and II¢ sites with population num-
bers 4, 18 and 8, respectively. Six of the eight sodalite cages of the unit cell con-
tain (Ni-Cl-Ni-Cl-Ni)4+ clusters. The site population by Ni2+ in the D2O complex
is different: 4 at site I, 11 at site I¢, 4 at site II¢ and 11 at site III¢.
Solid-state reaction was also reported for the system NiCl2/H-SAPO-34,
where Ni(I) was introduced into extra-framework sites, in contrast to as-syn-
thesized samples of NiAPSO-34, where Ni(I) is incorporated into the framework.
This difference was evidenced by ESR and ESEM [178].
The study by Wichterlová et al. [176 ] on SSIE of nickel compounds with H-
ZSM-5 was essentially motivated by problems of the preparation of Ni-contain-
ing zeolite catalysts for isomerization of C8 aromatics. The authors reported that
Solid-State Ion Exchange in Microporous and Mesoporous Materials 129

Ni-ZSM-5 was easily obtained with an exchange degree of almost 100% by SSIE,
whereas conventional exchange yielded a material with only a maximum of 60%
exchange degree. Moreover, Ni-ZSM-5 catalysts, where the protons have been
completely exchanged by Ni2+ via SSIE, exhibited after reduction in a hydrogen
flow the same catalytic activity as the parent H-ZSM-5 in isomerization and
dealkylation of C8 aromatics in a mixture of o-xylene and ethylbenzene. This
was ascribed to full restoration of the original acid OH groups upon reduction
(vide supra). Furthermore, it was found that the reduced Ni-ZSM-5 samples pre-
pared by SSIE were similarly active in hydrogenation of ethylene as reduced
Ni-ZSM-5 materials obtained through conventional exchange of H-ZSM-5 with
aqueous solutions of Ni(CH3OOH)2 .
Incorporation of nickel into ZSM-5 was also studied by Jentys et al. [164, 165,
167], who reacted NiCl2 ◊ 6H2O with H-ZSM-5 and obtained a ratio of nNi 2+/n3+ = 1
Al
2+
in a similar way as reported for the introduction of Co (vide supra). The Ni-
ZSM-5 prepared this way was more easily reduced with H2 (at 573 K) than Co-
ZSM-5 (at 773 K) but exhibited a much lower hydrogenolysis activity. Jentys et al.
[169] carried out a comparison between the results of nickel introduction via dif-
ferent methods into zeolites similar to that reported for modification by cobalt
(vide supra). The samples were prepared by conventional exchange, impregna-
tion and solid-state ion exchange. The comparison was extended to other zeolite
structures, viz., H-MOR and H-BETA. A 100% exchange could be achieved only
in the case of SSIE. Again, SSIE led to products with highly dispersed metal
cations and, after reduction, to small metallic clusters of nickel inside the zeolite
structure. The reduced materials showed in the hydroconversion of n-heptane an
enhanced selectivity for isomerization. In n-nonane hydroconversion, the differ-
ently prepared materials exhibited similar activities and selectivities with Ni,H-
MFI as the most active and Ni,H-BETA the most selective catalyst.
Bock [44] and Bock et al. [179] introduced Ni2+ cations into zeolite-like
SAPO-42, a narrow pore silicoaluminophosphate with zeolite A (LTA) structure
(cf. [180]), i.e., 8-membered ring pore openings (0.41¥0.41 nm), and the follow-
ing ratios of the tetrahedrally coordinated framework T-atoms: nSi /(nAl + np) =
1.0, nAl /nP = 2.3 [180]. They reacted powdered mixtures of SAPO-42 and NiCl2 or
NiO. Solid-state ion exchange was monitored by TPE of HCl and H2O and char-
acterized by the starting temperature, Tstart , i.e., the temperature of initial
gas evolution. Evolved HCl was trapped in an excess of 0.1 M aqueous NaOH
solution and determined via back-titration. In the case of NiCl2 , a ratio nCl-,out /
nCl-, in = 0.97 of the out-going vs. in-going Cl– was measured. This means that
Ni2+ cations were not incorporated as monovalent complex cations, (NiCl)+, in
analogy to incorporation of (CoCl)+ into ZSM-5 (vide supra), but just as they
were, i.e., as bare Ni2+. This was possible because of the higher Al content com-
pared with ZSM-5. Surprisingly, the starting temperature was lower in the case
of SAPO-42/NiO (Tstart = 842 K) than with SAPO-42/NiCl2 (Tstart = 930 K), even
though the lattice energies of NiO and NiCl2 are 4010 and 2772 kJ/mol, respec-
tively. This was explained by the fact that the oxide possesses the lower ion pair
size (0.30 ¥ 0.38 nm) compared to the chloride (0.38 ¥ 0.84 nm). The higher
mobility resulting from the smaller size of the migrating species probably over-
compensated the effect of the higher lattice energy.
130 H.G. Karge · H.K. Beyer

Similarly, introduction of nickel into SAPOs was investigated by Kevan and

his associates [181–185]. These authors prepared Ni,H-SAPO-n (n = 5, 8, 11, 34)
via SSIE between NiCl2 and H-SAPO-n. The materials were characterized by
ESR. Ni was incorporated as Ni2+; however, after dehydration (thermal reduction
in vacuum at 473–873 K), Ni+ was stabilized in the cation positions of the SAPO
structures. Azuma et al. prepared Ni,H-SAPO-5 by SSIE between NiCl2 and H-
SAPO-5 as well as by conventional exchange using H-SAPO-5 and an aqueous
solution of NiCl2 ◊ 6H2O [181]. The Ni-containing materials were investigated in
detail by ESR and ESEM (electron spin echo modulation spectroscopy).With the
sample prepared via CE no ESR signals were observed. However, in the case of
samples produced by SSIE, after thermal reduction, various Ni(I) species could
be detected. Their existence and location were dependent on the temperature. In
another study, Hartmann et al. compared Ni incorporated into SAPO-5 either by
adding Ni(OCOCH3)2 ◊ 4H2O to the synthesis mixtures for the SAPO resulting
in NiAPSO-5 or by SSIE of NiCl2 and H-SAPO-5 at 873 K yielding Ni,H-SAPO-5
[182]. The nickel content in the Ni,H-SAPO-5 prepared in this manner,
Ni0.01H0.01(Si0.03 Al 0.50 P0.47)O2 , appeared to be the upper limit for SSIE with respect
to the silicon content of the sample. Even though the g-values of Ni+ in the syn-
thesized NiAPSO-5 and the Ni,H-SAPO-5 were almost identical, the different
incorporations in framework and extra-framework positions, respectively,
could be clearly evidenced by 31P nuclear modulations, differences in the ESR
spectra after adsorption of ammonia, CH3OH or C2D4 , contrasting coordination
properties derived from the ESEM data and the fact that Ni,H-SAPO-5 decom-
posed water at room temperature to generate O2– (cf. [182]). Similar results were
obtained when NiAPSO-11 and Ni,H-APSO-11 were prepared in an analogous
manner and investigated by ESR and ESEM [183, 184]. The Ni-containing SAPOs
and APSOs were employed as catalysts for ethylene dimerization to n-butenes.
The catalytic activity was shown to depend on Ni(I) species incorporated. Com-
pared with NiAPSO-n, the catalysts obtained via SSIE exhibited a lower selectiv-
ity with respect to n-butenes [185].
Ni(II) cations were also introduced by SSIE into [Si]MCM-41 and
[Al,Si]MCM-41. After thermal or hydrogen reduction, Ni(I) species were stabi-
lized in the products. The results were compared with those obtained with con-
ventionally exchanged MCM-41 samples [186].
Similarly to the zeolite modifications with Fe2+ and Co2+, Ni-zeolites pro-
duced via SSIE were also tested as catalysts for NOx decomposition and/or
reduction. Varga et al. [114, 121] investigated by IR the adsorption behavior and
transformation of NO on the resulting Ni-ZSM-5 and showed that this material
is indeed a good catalyst for NO decomposition. In their study on selective cat-
alytic reduction of NO by Raney-nickel-supported Cu-ZSM-5, Ma et al. [187]
observed that the NO conversion decreased only at low temperatures but
remained constant or even improved at high temperatures. The authors ascribed
this effect to a migration of Ni cations into the zeolite matrix via SSIE that led to
a stabilization of the lattice.
Solid-State Ion Exchange in Microporous and Mesoporous Materials 131

5.3.4.4
Introduction of Manganese

A few reports have been published dealing with solid-state ion exchange of
manganese cations into hydrogen forms of zeolites. Wichterlová et al. [147] and
Beran et al. [188] preferentially employed ESR spectroscopy to monitor intro-
duction of Mn2+ from MnCl2 , MnSO4 , Mn3O4 or Mn(CH3COO)2 into H-ZSM-5.
In addition, other techniques such as XPS, IR spectroscopy, TPD of ammonia
(TPDA), TPE, back-titration of evolved HCl and test reactions have been used.
IR spectroscopy for monitoring the changes of the absorbance in the OH
stretching region upon SSIE was also employed in the work by Onyestyák et al.
[139] mentioned above on incorporation of various bivalent cations (Ca2+, Cd2+,
Zn2+, Co2+ and Mn2+) into faujasite-type zeolites via solid-state reaction in order
to prepare adsorbents for H2S (vide supra).
Application of XPS to mixtures of Mn(NO3)2 or MnSO4 with H-ZSM-5 after
solid-state reaction at 700–920 K in a stream of oxygen or hydrogen revealed
that the same concentration of Mn2+ occurred at the external surface of the zeo-
lite crystallites as that determined by chemical analysis for the bulk. Therefore,
the authors [147] concluded that the incorporated cations were homogeneously
distributed: neither surface enrichment nor depletion had taken place.
An instructive set of ESR spectra is displayed in Fig. 49. It is seen that the
starting mixture of solid MnSO4 and H-ZSM-5 did not exhibit any hyperfine
splitting of the signal at g = 2.0 originating from Mn2+ in crystalline MnSO4

Fig. 49. X band ESR spectra of Mn2+: a a physical mixture of MnSO4 and H-ZSM-5; b sample
(a) heat-treated at 770 K, c sample (a) heat-treated at 870 K; and d Mn,H-ZSM-5 prepared by
conventional ion exchange and calcined at 770 K (after [147], with permission)
132 H.G. Karge · H.K. Beyer

(spectrum a). Only progressive heating up to 770 and 870 K (spectra b and c,
respectively) and subsequent rehydration at ambient temperature caused the
hyperfine splitting to appear. This indicated disaggregation of the MnSO4 crys-
tallites and migration of Mn2+ into the zeolite structure. The six-line spectrum
with a hyperfine splitting constant of A = 9.8 mT is considered to be character-
istic of isolated Mn2+ cations in Oh coordination. This was substantiated by com-
parison with the ESR spectrum of a conventionally prepared and equally treat-
ed M,H-ZSM-5 sample (cf. spectrum d). The low-intensity signal at g = 4.27 was
ascribed to Mn2+ in distorted tetrahedral (Td) coordination. This assignment
was based on a comparison with ESR spectra of borate glasses [189].
Essentially the same ESR features were observed by Beran et al. [188] when
studying the system MnCl2/H-ZSM-5. In contrast, no ESR signals ascribable to
isolated Mn2+ cations on exchange sites of ZSM-5 were detected when the sodi-
um form of the zeolite, i.e., Na-ZSM-5, was heated in a mixture with MnSO4 and
subsequently hydrated [147].
Solid-state reaction of MnCl2 , MnSO4 , Mn3O4 or Mn(CH3COO)2 with H-ZSM-
5 was also monitored by IR spectroscopic measurements of the consumption of
acidic OH groups, i.e., the decrease in the intensity of the respective IR band at
3610 cm–1 (cf. Fig. 50 and Table 11).
In no case, however, was a 100% degree of exchange reached, as measured via
the fraction of protons replaced by Mn2+, i.e., as dOH , or through the incorporat-
ed fraction of available manganese cations, i.e., as dMn (cf. Table 11). From Fig. 50
and Table 11 it can be seen that an increase in the reaction temperature consid-
erably enhanced the consumption of OH groups, i.e., dOH . In contrast, an
increase in the amount of applied MnCl2 from a sub-stoichiometric ratio
(0.33 mmol Mn2+ vs. 0.91 mmol OH groups per gram) to a stoichiometric ratio
(0.45 mmol Mn2+ vs. 0.91 mmol OH groups per gram) did not bring about a mea-

Fig. 50. Number of bridging OH groups consumed as a function of the reaction time in solid-
state reaction in the system MnCl2/H-ZSM-5 (nMn2+/nOH = 0.33) upon heating in vacuum at
570, 670 and 770 K (after [188], with permission)
Solid-State Ion Exchange in Microporous and Mesoporous Materials 133

Table 11. Chemical composition of the original mixtures of H-ZSM-5 with MnCl2 , MnSO4 ,
Mn3O4 , and their characteristics after heat-treatment in high vacuum (SSIE)

Composition of the mixture a,b Heat-treatment c OH consumed (%)

Salt /oxide Mn (mmol g–1) Mn2+/OH b T (K) In vacuumd In N2 flow e

MnCl2 0.33 0.35 570 0.21 –

0.33 0.35 670 0.38 0.41
0.33 0.35 770 0.56 0.57
0.45 1.00 770 0.57 –
MnSO4 0.47 0.50 770 0.16 0.18
Mn3O4 0.60 0.66 770 0.46 –
a Mn content per gram of dry zeolite.
b Number of Mn cations related to the original number of the zeolite bridging OH groups
(0.91 mmol per gram of dry zeolite).
c Under vacuum at 10–4 Pa.
d Determined from the intensity of the IR band at 3610 cm–1 after heat-treatment in high
vacuum (10–4 Pa).
e Determined from TPDA, high-temperature peak approximately at 700 K.

surable increase in the consumption of OH groups (0.56 to 0.57 mmol OH

groups per gram, cf. Table 11).
Figure 50 shows clearly that most of the Mn2+ cations were introduced during
the initial stage of the reaction (i.e., within the first hour). Subsequently, the
reaction proceeded rather slowly (compare also the system CoCl2/H-CLIN, vide
supra). This kinetic behavior in solid-state ion exchange was confirmed by TPE
measurements of HCl evolution [188].
As in previous cases, the anion bound to the in-going cation also played a sig-
nificant role in SSIE with manganese compounds. Similar to the observations with
solid-state incorporation of alkaline, alkaline earth, copper and iron cations, it was
found that SSIE proceeded most easily with manganese chloride and nitrate but to
a significantly lesser extent with sulfate and oxide. It is worth mentioning, how-
ever, that at variance with the case of nickel (vide supra), an exchange did take
place with Mn oxide.As a consequence, reaction with acetate also led to some solid-
state ion exchange because Mn oxide resulting from decomposion of Mn acetate
was able to react with the Brønsted acid OH groups (cf. Table 11).
Consumption of acid OH groups upon introduction of Mn2+ cations led to a
decrease in Brønsted and a concomitant increase in Lewis acidity, as indicated
by IR using pyridine as a probe. The decrease in Brønsted acidity was indicated
by a decrease in the intensity of the pyridinium ion band at 1540 cm–1 observed
upon pyridine adsorption on Mn-ZSM-5 samples that had been prepared by
SSIE. A corresponding increase in the Lewis acidity effected an enhancement of
the absorbance around 1450–1454 cm–1, which is indicative of pyridine coordi-
nated to electron pair acceptors or Lewis sites such as Mn2+. This was compared
with similar IR measurements on the parent H-ZSM-5. Figure 51 shows a linear
correlation between the density of Brønsted and Lewis acid sites of the parent
H-ZSM-5 as well as of Mn,H-ZSM-5 samples measured by this technique.
134 H.G. Karge · H.K. Beyer

Fig. 51. Plot of the maximum absorbance log(T0/T) (T, T0 transmittances at the frequency of
the band minimum and the correlated base line) of the PyH+ band against log(T0/T) of the PyL
band measured for 1 H-ZSM-5; 2 Mn3O4/H-ZSM-5; 3 MnSO4/H-ZSM-5; 4 MnCl2/H-ZSM-5;
and 5 conventionally exchanged Mn,H-ZSM-5; materials heated in vacuum at 770 K for 6, 6,
60, 2, and 6 h, respectively (after [188], with permission)

In fact, the result for SSIE of Mn3O4/H-ZSM-5 deviated from this correlation
in that the measured density of Lewis acid sites was lower than was expected
according to the measured decrease in the density of Brønsted acid sites. The
value of the density of Brønsted acid sites after solid-state reaction of Mn3O4/H-
ZSM-5 was, however, in agreement with IR spectroscopic and TPDA determina-
tion of the consumption of OH groups. Therefore, the authors ascribed the devi-
ation to an underestimation of the Py Æ Mn2+ value, and tentatively assumed
that this was due to a partial blockage of incorporated Mn2+ by unreacted Mn3O4
species. It is important to note in the context of these investigations that the
authors established that SSIE did not generate ‘true’ Lewis sites (cf. [171, 172,
188]) by dealumination of the framework. They showed via re-exchange of the
introduced Mn2+ in NH4NO3 solution and subsequent deammoniation that the
original density of OH groups could be fully restored. Completeness of re-
exchange was proven by ESR, which indicated that only traces of Mn2+ were left
in the re-exchanged zeolite samples.

5.3.5
SSIE of Vanadium, Niobium, Antimony, Chromium, Molybdenum
and Tungsten Compounds with Zeolites

5.3.5.1
Introductory Remarks

Introduction of transition metal cations of high oxidation states via conven-

tional ion exchange is usually rather difficult if not impossible, since the
required cations are frequently not available in simple metal salts. If, however,
Solid-State Ion Exchange in Microporous and Mesoporous Materials 135

such salts are available, the respective cations may exist only in strongly acidic
solutions, i.e., the metal cations have to compete with protons for ion exchange
and the zeolite lattice is often damaged by acid attack. Therefore, solid-state ion
exchange may be a promising route to prepare zeolites containing polyvalent
and homogeneously distributed cations with changeable oxidation states. This
is particularly true in view of possible applications of such zeolites as valuable
redox catalysts. Kucherov and Slinkin [23, 92, 190, 191] were the first to study the
incorporation of cations with (possibly) high oxidation states, such as V5+ or V4+,
Cr6+ or Cr5+ and Mo6+ or Mo5+, into zeolites via solid-state exchange. Again, they
preferentially employed ESR spectroscopy (vide supra). Huang et al. [192–194]
as well as Marchal et al. [195–197] extended the investigations on SSIE with
vanadium compounds, by also monitoring solid-solid reactions between tung-
sten and zeolites. Studies on SSIE with chromium cations were conducted by
several authors [93, 100, 198, 199] using various experimental techniques. In-
corporation of Mo-containing cationic species into zeolites attracted much
attention. As various oxidation states of molybdenum may occur in zeolites, the
materials resulting from introduction of Mo were expected to exhibit interesting
redox properties. Incorporation of Mo-containing species into particular micro-
porous structures could generate valuable shape-selective redox catalysts.
Therefore, after the early work by Dai and Lunsford [200], a number of reports
on SSIE of zeolites with molybdenum compounds appeared [44, 92, 179, 190, 197,
201]. Finally, systematic research into SSIE using cations with a high oxidation
state was extended to solid-state reactions with niobium compounds [202].
Interestingly, in most examples of solid-state reactions between compounds of
tetra-, penta- or hexavalent cations and zeolites, a reduction of the cations
occurred (vide supra; e.g., V5+ Æ V4+, Cr6+ Æ Cr5+ and Mo6+ Æ Mo5+). Most
likely, this is due to an auto-reduction (cf. [203, 204]), i.e., the reduction is
accompanied by an oxidation of framework oxygen anions to molecular oxygen,
although this was not discussed by the respective authors.

5.3.5.2
Introduction of Vanadium

In order to introduce vanadium into hydrogen forms of zeolites (H,Na-MOR, H-

MOR with nSi/nAl = 5 and H,Na-ZSM-5, H-ZSM-5 with nSi/nAl = 35), mixtures of
V2O5 and the zeolites were subjected to heat-treatment at 1073 K in air [92, 190,
191]. Electron spin resonance spectroscopy (ESR) yielded a spectrum (as shown
for the example of H-ZSM-5 in Fig. 52) exhibiting a well-resolved hyperfine (HF)
signal of vanadyl cations with g = 1.93, g^ = 2.02,A = 19.8 mT, and A^ = 8.3 mT.
These parameters are typical of isolated V4+ cations in an almost square-planar
coordination.
Identical signals were seen after heating in air, hydrogen or vacuum. When
the sample was cooled to 77 K, no change was observed. In particular, vanadyl
cations proved to be very stable against oxidative or reductive treatment. How-
ever, introduction of oxygen resulted in a considerable line broadening (cf.
[190]). This shows that the incorporated vanadium-containing species are acces-
sible to gases.
136 H.G. Karge · H.K. Beyer

Fig. 52. ESR spectra of isolated vanadyl cations (51VIV) introduced into H-ZSM-5 (nSi/nAl ª 35)
by solid-state reaction between V2O5 and H-ZSM-5 at 1025 K (after [191], with permission)

In the case of H-MOR with an nSi/nAl ratio of 5, which is considerably lower

than in H-ZSM-5 and, thus, gives rise to a much smaller distance between the
cation sites, the ESR signals were broadened due to dipole-dipole interaction of
the V4+-containing species.
In a situation where the isolated paramagnetic ions on extra-framework sites
of a zeolite structure are able to electronically interact with adjacent Al3+ ions of
the framework, one may expect a super-hyperfine splitting (SHFS) of the signal.
This was, indeed, observed with CrO +2 , i.e., Cr(V), in H-ZSM-5 (vide infra and cf.
[92, 95]). In view of the analogy of the electronic configuration (3d1), occurrence
of a SHF-splitting was also expected for ESR spectra of vanadyl species such as
VO(OH)+, i.e., V(IV), in the MFI structure. In fact, Kucherov and Slinkin [191]
succeeded in detecting SHF-splitting in the ESR spectra of V,H-ZSM-5 obtained
via SSIE in a V2O5/H-ZSM-5 mixture with nV/nAl = 0.3 after calcination at 1023 K
in air and subsequent evacuation. A satisfactory resolution was achieved when
the spectra were run at 473 K; this produced a signal with a SHF-splitting con-
stant of 0.7 mT (see Fig. 52, spectrum c). Figure 53 shows the proposed possible
arrangement of the vanadyl species and its unpaired electron in the dxy orbital
adjacent to Al in the framework.
As mentioned above, admission of dry oxygen to V,H-ZSM-5 prepared
through SSIE led to a line broadening because of dipole-dipole interaction of
VO(OH)+ and the paramagnetic O2 molecules. Adsorption of ammonia or pyri-
dine significantly changed the ESR parameters and made the SHF-splitting
completely disappear. Kucherov and Slinkin [190] ascribed this to an effect of
Solid-State Ion Exchange in Microporous and Mesoporous Materials 137

Fig. 53. Schematic representation of the close proximity of a vanadyl cation to framework Al
in a zeolite structure giving rise to the super-hyperfine splitting of the ESR signal shown in
Fig. 52 (after [191], with permission)

the additionally coordinated strong ligands on the density distribution of the

unpaired electron.
An attempt was made to co-introduce Cu into the above-characterized
VO(OH),H-ZSM-5 via SSIE by calcining a mixture of Cu compounds and the V-
containing zeolite [93]. This, however, led to a decrease in the VO(OH)+ concen-
tration as was indicated by a decrease in the intensity of the signals related to
vanadyl species (cf. Fig. 54) and to the appearance of the ESR signal of isolated Cu+
in ZSM-5 (vide supra, Sect. 5.3.2.1). Obviously, Cu2+ is more strongly held in the
zeolite structure than VO(OH)+ and is able to easily replace the vanadyl cation.
In a similar way as that used by Kucherov and Slinkin [92, 191], Huang et al.
[192–194] prepared vanadium-containing ZSM-5 and mordenite via SSIE
between V2O5 and the zeolites mentioned above. The products were character-
ized by XRD, XPS, and determination of their surface areas. At variance with the
interpretation given by Kucherov and Slinkin, Huang et al. [192] and Shan et al.
[193] claimed from their ESR results that the vanadyl species existed in two dif-
ferent coordination structures both in ZSM-5 and mordenite, viz., in a com-
pressed hexacoordinated octahedral and a pentacoordinated pyramidal sym-
metry. Furthermore, the authors studied the acidity of the zeolites after SSIE
with V2O5 by FTIR of adsorbed pyridine and titration with butylamine in apro-
tic solution. In addition, the capacities for adsorption as well as the activities and
selectivities for oxidation of toluene to benzaldehyde were investigated. The
authors showed that the differences in adsorptive and catalytic behavior were
related to the differences in acidity of the materials produced by SSIE [193].
In similar experiments, Ignatovych et al. [205] modified H-ZSM-5 and char-
acterized the oxidation state, coordination and dispersion of vanadium in the
products of the solid-state reaction by UV-Vis diffuse reflectance spectroscopy
as well as photoluminescence spectroscopy. It was shown that the reaction con-
ditions, i.e., calcination of, e.g., V2O5/NH4-ZSM-5 in (i) a flow of O2 at 973 K or
(ii) N2 at 773 K, affected the properties of the products: in the case of (i) isolated
VIV and tetrahedral VV were predominant, whereas in case (ii) mainly polyvana-
dates and clusters with VIV and VV valences were detected.
Finally, it is worth mentioning that, in a zeolite in which Al had been isomor-
phously substituted by Ga, i.e., in H-[Ga]ZSM-5, vanadyl cations could not be
138 H.G. Karge · H.K. Beyer

Fig. 54. ESR spectra of a V2O5/H-ZSM-5 calcined in air at 1073 K for 1 h; b after calcination of
sample (a) in a mixture with CuO at 823 K for 1 h; c after calcination of sample (b) at 1023 K
for 1 h; and d after reduction of sample (c) with H2 at 400 K for 2 h (after [93], with permis-
sion)

stabilized [128, 129]. Thermal treatment of VO(NO3)2 with H-[Ga]ZSM-5 gave

rise to only very weak ESR signals of isolated VIV species.
Introduction of vanadium into AlPO4-5 was attempted by Whittington and
Anderson [206] by grinding a mixture in air at room temperature and subse-
quent heating at 823 K for 2 h. The products were investigated by XRD and ESR.
The results depended on the vanadium loading. For molar ratios of nV/(nV + nAl
+ nP) = 0.025 the AlPO4-5 structure remained intact, whereas, at higher ratios
(0.039 and 0.089), collapse of the lattice and formation of new phases were
observed. The ESR spectrum revealed that part of the V was incorporated as
isolated vanadyl (= V = O) species and part of it as species with interacting VIV,
most probably as some form of oxide. These findings essentially agreed with
Solid-State Ion Exchange in Microporous and Mesoporous Materials 139

those obtained by Hong et al. [207] for vanadium-loaded AlPO4-5 prepared by

impregnation of this aluminophosphate with aqueous solutions of ammonium
vanadate and subsequent calcination at 803 K.
The possibility of incorporating vanadium into the sodium form of Y-type
zeolite, i.e., Na-Y, by solid-state reaction at 690–750 K in air was investigated by
Marchal et al. [195, 196] and Thoret et al. [197]. XRD, TEM, 29Si MAS NMR, 129Xe
NMR, and ESR techniques were used to study the solid-state reaction between
V2O5 and Na-Y (nSi/nAl ª 2.5) and to characterize the products, in particular
their crystallinity and phase composition. The authors found that only at low
vanadium contents in the V2O5/Na-Y mixtures (i.e., nV/(nSi + nAl) £ 0.05 or nV/nAl
£ 0.175) did introduction of V-containing species (V2O5) proceed without mod-
ification of the zeolite lattice and/or formation of amorphous phases.After reac-
tion in such a system with nV/nAl = 0.175, ESR showed a signal, the hyperfine
splitting of which was hardly visible because of the strong interactions between
neighboring tetravalent vanadium ions. A sample with a lower concentration of
vanadium was obtained by heating a mixture with a ratio of nV/nAl £ 0.018. This
sample produced a spectrum exhibiting a much more resolved HF structure
because of the decreased dipole-dipole interactions. The hyperfine splitting was
further increased when the observation temperature was lowered from 293 to
77 K (Fig. 55).

Fig. 55. ESR spectrum obtained at 77 K after solid-state reaction in the mixture V2O5 /Na-Y at
690–750 K in air; nV/(nAl + nSi) = 0.005; numbers assigned to maxima and minima indicate the
magnetic induction in 10–4 T (after [196], with permission)
140 H.G. Karge · H.K. Beyer

Then, the spectrum looked qualitatively like that reported by Kucherov and
Slinkin for the product of a solid-state reaction between V2O5 and H-ZSM-5 (cf.
Fig. 52), confirming the axial symmetry of the V4+ ions. However, the character-
istic parameters were assumed to be different because of the different environ-
ments of the V-containing species introduced either into the fausasite-type (Na-
Y) or MFI (H-ZSM-5) structure. It was claimed that at a somewhat higher load-
ing, corresponding to nV/(nSi + nAl) = 0.2, XRD patterns showed the coexistence
of three phases, viz., (i) NaV5VVIVO15, formed by a reaction of weakly acid V2O5
and Na+ of Na-Y, (ii) V2O5 , and (iii) slightly amorphitized Na-Y. The authors sug-
gested that at even higher vanadium loadings, i.e., nV/(nSi + nAl) = 0.6 and T =
870 K, the lattice is completely destroyed and some of the sodium cations of Na-
Y are involved in the formation of vanadium bronzes such as Na5V11VVIVO32 [195,
197]. Analysis of the XRD pattern of the (hydrated) materials after solid-state
reaction showed an increase in the cubic unit cell parameter from 2.4639 to
2.4673 ± 0.0015 nm. This effect could originate from an isomorphous substitu-
tion of Si4+ by V4+ in the framework, a possibility that cannot be excluded, even
though the authors did not obtain further experimental evidence for such a
replacement.
Narayanan and Sultana [208] as well as Narayanan and Deshpande [209]
admixed V2O5 to H,Na-X, H-Y, H-ZSM-5 and H-MOR and subsequently con-
ducted solid-state reaction. Compared to the parent hydrogen zeolites, consid-
erably enhanced activity of the catalysts prepared via SSIE was observed in
vapor-phase alkylation of aniline.
Introduction of vanadium into microporous materials is also discussed in
Sect. 6.3.1.

5.3.5.3
Introduction of Niobium and Antimony

Niobium-containing catalysts have attracted considerable interest because of

their activity in reactions such as dehydrogenation of alcohols, photooxidation
of propene, oxidative decomposition of methyl-tert-butyl ether and selective
catalytic reduction of NO by NH3. Zeolites with niobium as an active component
are difficult to prepare since niobium salts are rather sensitive to moisture.
Therefore, Ziolek et al. [202] explored the possibility of introducing Nb into
hydrogen forms of zeolites via solid-state reactions between Nb2O5 and NH4 ,
Na-Y, NH4-Y (nSi/nAl = 2.56) and dealuminated NH4-Y (NH4-Y(D), nSi/nAl = 4.25).
Mixtures of Nb2O5 and zeolites with ratios of nNb/(nSi + nAl) = 0.03, 0.06 and 0.3
were heated at 975 K in air or in vacuum. For these investigations various tech-
niques were employed, viz., high-temperature in situ XRD, IR, TPE/MS and
TPD/MS. TPE/MS and TPD were used to monitor ammonia and water evolved
during SSIE and to determine the acidity of the products, respectively.
The authors found that SSIE indeed occurred with zeolite samples containing
acid protons. Evidence was derived from the decrease in the intensity of the
Nb2O5 reflections in the XRD pattern of the heat-treated Nb2O5 /zeolite mixtures
and the evolution of water. Also, XRD proved that the crystallinity of the zeolite
samples did not deteriorate upon SSIE. TPD of ammonia and IR spectroscopy
Solid-State Ion Exchange in Microporous and Mesoporous Materials 141

showed a partial elimination of Brønsted acid sites and a concomitant formation

of Lewis acid centers. The activity of the niobium-containing samples in Brøn-
sted acid-catalyzed dehydration of isopropanol was distinctly lower than that of
the parent zeolites, but introduction of Nb generated redox properties of the
zeolite samples that were evidenced by the appearance of acetone among the
products of isopropanol conversion. No incorporation of niobium was detected
when Nb2O5 was mixed with the sodium form of Y-type zeolite and heat-treated
at 975 K. However, in contrast to the findings of Marchal et al. [195, 196] for the
case of the system V2O5/Na-Y at high vanadium loadings, no loss of crystallini-
ty was observed with Nb2O5/Na-Y.
Thoret et al. also studied the interaction of antimony oxide, Sb2O3, with Na-Y
or La,Na-Y [210]. However, similar to their findings with the systems V2O5/Na-Y
and MoO3/Na-Y, no real solid-state ion exchange was observed with Sb2O3/Na-
Y. Rather, the oxide is inserted as such. In the case of Sb2O3/La,Na-Y, where the
zeolite possesses residual Brønsted acid sites, the oxide can react in a certain
range of Sb2O3 content and temperatures to produce Sb2O42+ cations.

5.3.5.4
Introduction of Chromium

There are several reports on solid-state ion exchange of chromium into zeolites
[92, 93, 190, 198, 199]. Thus, Kucherov and Slinkin reacted chromium oxides
(CrO3 , CuCrO4 , Cr2O3) with hydrogen forms of mordenite (nSi/nAl = 5) and ZSM-5
(nSi/nAl = 35 or nSi/nAl = 140) at 773–1093 K in air or vacuum. The proton-con-
taining samples were H,Na-ZSM-5(40), H-ZSM-5(95), Na,H-MOR(50) and H-
ZSM-5 (95), where the figures in brackets designate the degree of replacement of
Na+ by protons, and the gallosilicate and ferrosilicate analogs of H-ZSM-5, i.e.,
H-[Ga]ZSM-5 and H-[Fe]ZSM-5.
Solid-state reaction with Cr(VI) oxide was indicated by the appearance of
ESR signals characteristic of isolated Cr(V)-containing ions [92].As an example,
the ESR spectrum after SSIE of the system CrO3/H-ZSM-5 (nSi/nAl = 35) is shown
in Fig. 56. The signal exhibited a typical hyperfine splitting (HFS). However, in
the case of CrO3/H-ZSM-5 (nSi/nAl = 140), a particularly well-resolved splitting
occurred with 15 components and a splitting constant of ca. 0.7 mT. This type of
splitting was ascribed to the presence of 27Al located in close proximity to the
Cr(V) cation (cf. also [211, 212]), and is analogous to the effect of super-hyper-
fine splitting (SHFS) observed in the system V2O5/H-ZSM-5 (vide supra).
The ESR spectrum obtained after SSIE in the system CrO3 /H-ZSM-5 (cf.
Fig. 56) was almost identical to that observed with an H-ZSM-5 sample loaded
with 1.8 wt.% Cr via impregnation and calcined in air for 3 h at 773 K (cf. [213,
214]). This latter spectrum was characterized by g = 1.96 and 15 SHFS compo-
nents and a splitting constant of 0.7 mT and ascribed to octahedrally coordinat-
ed Cr(V) interacting through its unpaired electron with the nuclear spin (I =
5/2) of the adjacent 27Al in the second coordination sphere of the Cr-containing
ion. However, at higher loadings (2.5 wt.%) and calcination in air at 773 K for
1 h, both octahedrally and tetrahedrally coordinated Cr(V) cations were
observed with g^1 = 1.98, g 1 = 1.95 and g^2 = 1.91, g 2 = 1.98, respectively. It is quite
142 H.G. Karge · H.K. Beyer

Fig. 56. ESR spectra obtained at 77 K a after solid-state reaction in the mixture CrO3/H-ZSM-5
at 823 K in air followed by evacuation at 293 K; b after calcination of the mixture at 1023 K
followed by evacuation at 293 K; and c after admission of air (after [92], with permission)

probable that such coordinations occur with Cr,H-ZSM-5 samples obtained

through SSIE as well, in view of the similarity of the ESR spectra of Cr,H-ZSM-5
obtained via solid-state reaction (vide supra) and impregnation.
Calcination of CrO3/H-ZSM-5 under more severe conditions (e.g., at 1073 K
instead of 823 K) resulted in an increase of the signal intensity (cf. spectrum a
and b in Fig. 56). In addition, the intensities were affected by the number of
available Brønsted acid sites. From a comparison of the number of spins per
gram and the initial number of Brønsted acid sites as given in the report by
Kucherov and Slinkin [92], one can deduce that about 3 and 15% of the protons
were replaced by CrO+2 in the hydrogen form of ZSM-5 and mordenite samples,
respectively. The spins measured per gram, i.e., the absolute number of intro-
duced cations, increased with the number of Brønsted acid centers per gram,
and the upper limit of the cation concentration was, indeed, determined by the
concentration of these acid sites [92, 190]. However, in no case did the number
of isolated Cr(V), i.e., CrO+2 cations located inside the zeolite channels, exeed
20% of the Al cations in the framework corresponding to a maximum of 20% of
Solid-State Ion Exchange in Microporous and Mesoporous Materials 143

the available Brønsted acid centers. Excess chromium present after completion
of SSIE was assumed to be located as compact Cr2O3 aggregates on the external
surface of the zeolite crystallites [21, 128, 213, 214].
When, after solid-state reaction yielding the spectrum a or b in Fig. 56, the
sample was contacted with air at 293 K, a considerable but reversible line broad-
ening occurred (spectrum c in Fig. 56). Upon admission of water, the ESR signal
completely disappeared as in similar cases (vide supra). These effects were taken
as proof for the accessibility of the cations incorporated via SSIE. It is worth
mentioning that heat-treatment of a mixture of CrO3 and H-MOR did not result
in the appearance of an ESR signal with SHFS similar to that of Fig. 56b, whereas
such a signal being indicative of SSIE was observed with the system CrO3/H,Na-
MOR. This was most likely due to the lower density of Brønsted acid sites and,
consequently, Cr(V) species in H,Na-MOR. The higher dilution of chromium-
containing species in H,Na-MOR in turn decreased the signal broadening due to
dipole-dipole interaction and, thus, improved the resolution.
Similar to their interpretation of the incorporation of V(IV) or Mo(V) cations
into zeolites by solid-state reaction, Kucherov and Slinkin suggested that in high-
silica zeolites the respective cation sites are too distant to allow a balance of the
negative framework charges by bare Cr5+ cations. Rather, they proposed incorpo-
ration of Cr(V) in the form of complex cations such as CrO2+ under participation
of framework ligands, in analogy to VO(OH)+ (vide supra) or MoCl4+ (vide infra).
Co-introduction of Cr(V) and Cu(II) into H-ZSM-5 was possible by SSIE
using either CuCrO4 in a one-step solid-state reaction or a consecutive reac-
tion with CrO3 and CuO [93, 198]. In the former case, the ESR signals of
Cr(V) (cf. Fig. 56) and Cu(II) (cf. Fig. 28) were superimposed, appeared simul-
taneously and exhibited well-resolved hyperfine splitting. The Cr- and Cu-con-
taining species were randomly distributed over the same types of sites with
negligible dipole-dipole interaction and exhibited the same coordinations as
observed with the individual cations after SSIE in CrO3/H-ZSM-5 and CuO/H-
ZSM-5. From the intensities of the signals obtained upon solid-state reaction
of CuCrO4 and H-ZSM-5, a ratio of nCu(II)/nCr(V) = 2–3 was estimated. The pref-
erence for incorporation of Cu(II) was even more pronounced when Cr(V) and
Cu(II) were successively introduced. When Cr,H-ZSM-5, prepared via SSIE
of CrO3 and H-ZSM-5 at 1023 K, was subsequently mixed with CuO and
calcined in air, the Cr signal dropped considerably and concomitantly the signal
of Cu(II) developed. This replacement resulted in a ratio of nCu(II)/nCr(V)
= 20–30. When the sequence of the two SSIE steps was reversed, only weak
Cr(V) signals were detected. Upon reduction of the samples prepared by co-
introduction of Cr(V) and Cu(II), the ESR signals of both cations were rapidly
eliminated.
There have been no successful attempts to react chromium oxides with sodi-
um forms of ZSM-5 or MOR to incorporate chromium-containing cations. No
solid-state ion exchange was indicated by ESR signals of stabilized isolated
Cr(V) species similar to those observed upon the solid-state reaction of CrO3
with the hydrogen forms of ZSM-5 or mordenite (vide supra).
Kucherov et al. [128, 129] carried out a comparative study of solid-state reac-
tion between CrO3 and the gallosilicate analog of H-ZSM-5, i.e., H-[Ga]ZSM-5,
144 H.G. Karge · H.K. Beyer

Fig. 57. ESR spectra obtained at 293 K of the mixture CrO3/H-[Ga]ZSM-5 a after calcination
at 693 K in air for 1 h followed by brief evacuation (5 min) at 293 K; b after calcination at 973 K
in air for 2 h followed by brief evacuation (5 min) at 293 K; and c after admission of O2 at 293 K
(after [129], with permission)

where via synthesis Al of H-ZSM-5 (i.e., H-[Al]ZSM-5) was isomorphously sub-

stituted by Ga. Samples with nSi/nGa = 35 were preferentially used. After calcina-
tion of a mixture of CrO3 and H-[Ga]ZSM-5 at 693 K, ESR spectra with
two superimposed signals (g^ = 1.99, g = 1.94, A^ = 2.8 mT, A = 2.65 mT) and at
973 K with one signal (g = 1.97, A ª 2.4 mT) were observed (Fig. 57).
These spectra indicated formation of isolated Cr(V) cations (CrO+2 ) through
SSIE on cation sites of the [Ga]ZSM-5 structure. In analogy to the cases of V(IV)
Solid-State Ion Exchange in Microporous and Mesoporous Materials 145

and Cr(V) introduced via SSIE into H-[Al]ZSM-5 (vide supra), super-hyperfine
splitting with a splitting constant of 2.4–2.8 mT in Cr,H-[Ga]ZSM-5 was
ascribed to an interaction of the unpaired electron of Cr(V) with the nuclear
spin of closely adjacent Ga3+ (I = 3/2) ions of the framework located in the sec-
ond coordination sphere of Cr(V). Indeed, the lower nuclear spin number (I =
3/2) of Ga compared to that of Al (I = 5/2) caused, according to the relationship
Nrel = (2IGa + 1)/(2IAl + 1), a reduction in the number of components, of the split
signal by a factor of 2/3, i.e., from 15 in the case of Cr,H-[Al]ZSM-5 to 10 in the
case of Cr,[Ga]ZSM-5. The concomitant increase in the splitting constant from
0.7 to 2.4–2.8 mT is due to the greater radius of Ga3+ (0.062 nm) compared to
that of Al3+ (0.05 nm). Admission of O2 to Cr,H-[Ga]ZSM-5 resulted in a consid-
erable but reversible broadening of the ESR signal. Adsorption of NH3 on Cr,H-
[Ga]ZSM-5 caused an anisotropic spectrum to develop identical to that found
after adsorption of NH3 on Cr,H-[Al]ZSM-5 (vide supra). The effects upon
adsorption of O2 and NH3 showed that the Cr(V) species incorporated by SSIE
into H-[Ga]ZSM-5 were coordinatively unsaturated and accessible to gaseous
molecules. The most significant difference between the Cr-containing Al- and
Ga-forms is, however, the lower thermal stability of Cr,H-[Ga]ZSM-5 which
exhibited a rather rapid decrease in the concentration of isolated Cr(V) on
cation positions already at 1073 K.
The structure of H-[Fe]ZSM-5 was even less thermally stable than that of H-
[Ga]ZSM-5 and possessed even weaker acid sites [148, 151]. This zeolite was,
therefore, not able to stabilize cations such as Cr5+ or Cu2+ (cf. Sect. 5.3.2.1).
The behavior of Cr(III) oxide on heat-treatment in mixtures with hydrogen
forms of zeolites was different from that of Cr(VI) oxide [190]. At 353 K, an
uncalcined mixture of 1.5 wt.% Cr2O3 produced only a broad ESR signal which
was identical to that of bulk Cr2O3 . Also, thermal treatment of Cr2O3/H-ZSM-5
or Cr2O3/H-MOR in vacuum did not give rise to the development of ESR signals
of isolated chromium cations. However, when such mixtures were calcined in air
at 1093 K, signals of isolated Cr(V) species appeared similar to those shown in
Fig. 56. Obviously, Cr2O3 did not react in the solid state, most likely because of
the high lattice energy of this compound (Tmelt = 2613 K). Calcination in air,
however, produced oxide species of Cr with higher valency that were sufficient-
ly mobile to penetrate into the channels of H-ZSM-5 or H-MOR and react there
with the acid Brønsted sites.
Weckhuysen and Schoonheydt [100, 199] conducted systematic studies of
the preparation of Cr-containing zeolites. These authors employed diffuse
reflectance UV spectroscopy (DRS) and ESR spectroscopy. For a comparative
investigation, they loaded various zeolites (X, Y, [Ga]Y and MOR-type zeolites)
not only by conventional ion exchange or impregnation-incipient wetness tech-
niques but also via solid-state ion exchange with CrCl3 ◊ 6H2O. After calcination
of differently prepared samples at 823 K, similar DRS spectra were obtained (cf.
Fig. 58).
These spectra showed two pronounced bands at 28,000 and 38,000 to
39,000 cm–1, which are typical of chromate-like species ascribed to charge trans-
fer processes (O Æ Cr6+), and were assigned to 1t1 Æ 2e and 6t2 Æ 2e transitions,
respectively. Spectrum c of Fig. 58 was obtained from a Cr,H-Y sample prepared
146 H.G. Karge · H.K. Beyer

Fig. 58. Diffuse reflectance spectra (DRS) of chromium-containing zeolites. a Cr-X obtained
by conventional ion exchange (CE); b Cr-Y obtained by impregnation; and c Cr-Y obtained by
solid-state ion exchange in the mixture CrCl3 ◊ 6 H2O and NH4-Y (H-Y); the samples were cal-
cined at 823 K (after [100], with permission)

via solid-state ion exchange. It showed an additional weak band around

10,000 cm–1 possibly due to d-d transitions of Cr(III). Detailed analysis of the
ESR spectra and decomposition of the DRS bands revealed that Cr was present
in various oxidation states, viz., Cr(III), Cr(V) and Cr(VI). However, the ESR
parameters found for Cr,H-Y obtained via SSIE (g^ = 1.99, g = 1.93, Cr5+ in
supercages, site II; g^ = 1.98, g = 1.94, Cr5+ in b-cages, site I¢, cf. [199]) were sim-
ilar to those reported by Kucherov and Slinkin for Cr,H-ZSM-5 (cf. [92, 190], vide
supra) and by Hemidy et al. for Cr-Y [211, 212]. Similar to the interpretation giv-
en by Kucherov et al. [129], Weckhuysen and Schoonheydt assumed that Cr(V)
species occurred inside the zeolite structure as complex cations such as CrO2+
coordinated to two framework oxygen atoms. Irrespective of the method of
preparation, chromate-type species were formed upon calcination, i.e., this
holds also for the zeolite exchanged through solid-state reaction. The authors
assumed the following two-step process as the most probable pathway of chro-
mate formation [cf. Eqs. (16) and (17)]:
CrCl3 + 3 H+Z– Æ Cr3+Z–3 + 3 HCl ≠ (16)
Cr3+Z–3 + 3/
4 O2 + 1/
2 H2O Æ (O2Cr=)2+Z–2 + H+Z– (17)

Scheme 2. Coordination of a complex CrO2+ to two framework oxygen atoms

Solid-State Ion Exchange in Microporous and Mesoporous Materials 147

The reaction (Scheme 2) results in the formation of two extra-framework oxy-

gen atoms and anchoring of the chromate-like species to two structural oxygen
ions (each bound to one framework Al) and regeneration of one Brønsted acid
site, H+Z– [100].
An interesting result of the studies reported by these authors was that Cr,H-Y
obtained through SSIE showed the lowest reducibility compared to samples pre-
pared via conventional ion exchange or the impregnation-incipient wetness
technique. This behavior was attributed to residual acidity of Cr-containing
Y-zeolite produced by solid-state reaction between CrCl3 ◊ 6H2O and NH4-Y.
Chromium was also introduced by SSIE into SAPO-11 by Kevan and co-work-
ers [215], and the products were investigated by ESR, ESEM and UV-Vis spec-
troscopy. The Cr-containing SAPO-11 resulting from SSIE showed Cr(V) in
square-pyramidal coordination that could be reduced by H2 to Cr(III).

5.3.5.5
Introduction of Molybdenum

Introduction of molybdenum into hydrogen forms of zeolites by SSIE was inves-

tigated by Dai and Lunsford [200], Kucherov and Slinkin [92, 190], Bock [44] and
Bock et al. [179]. Mo-containing compounds, such as MoCl5 , (MoOCl4), MoCl3
and MoO3 , were mixed with H-ZSM-5, H-MOR, H-ZSM-35, H-EU-1, H-ZSM-48
and H-L and subsequently subjected to heat-treatment. Solid-state ion exchange
with Mo compounds was monitored, and the respective SSIE products were
characterized by ESR, XPS, IR, XRD, CA, TPE/MS and titration of evolved HCl.
As mentioned above, the earliest experiments of solid-state ion exchange
between MoOCl4 and H-Y were those reported by Dai and Lunsford [200]. These
studies were occasioned by the interest in Mo-containing solid catalysts because
of their possible applications as redox catalysts in, e.g., epoxidation of cyclohex-
ene. MoCl5 (MoOCl4) was reacted with NH4-Y (deammoniated during SSIE to H-
Y), dealuminated, ultrastabilized NH4-Y(D) (deammoniated during SSIE to H-
Y(D)) and Co,H-Y (obtained by impregnation of NH4-Y with Co(NO3)2 and
deammoniated during SSIE to Co,H-Y). Since the preparation of the MoCl5/zeo-
lite mixtures was carried out in the presence of ambient air and moisture, the
authors assumed that most of the Mo compound was, prior to SSIE, converted
to MoOCl4 . From ESR and XRD experiments they concluded that Mo species
indeed entered the interior of the zeolite structure upon solid-solid reaction.
Differing from the results reported by Kucherov and Slinkin [92, 190] (vide
infra), XPS data showed that Mo cations were mainly present as Mo(VI) species,
whereas the results of ESR spectroscopy suggested that only about 3.5% of
the incorporated molybdenum was in the form of Mo(V). Because the ESR sig-
nal was affected by adsorption of O2 and N2O molecules, which cannot enter
the small cavities, the authors concluded that a fraction of the Mo-containing
species were located in the supercages. This was supported by their finding that
twice as many MoOCl4 molecules reacted as the number of OH groups that were
available on the external surface of the zeolite crystallites, so that at least part of
these entities must have migrated into the interior of the structure. Moreover,
from their XRD data, they drew the conclusion that these charge-compensating
148 H.G. Karge · H.K. Beyer

Mo species resided on sites SII. Back-titration of HCl evolved during solid-state

ion exchange with MoCl5 (or, in fact, MoOCl4) revealed that four Cl– anions per
Mo species had reacted. This would be compatible with either the involvement
of four hydroxyl groups (silanol groups and/or Brønsted acid groups) or of four
water molecules, since ambient moisture was not excluded, according to
Eqs. (18) and (19).
MoVIOCl4 + 4 H+Z– Æ [O=MoVI ∫Z–3]+Z– + 4 HCl≠ (18)
MoVIO Cl4 + 4 H2O Æ MoVIO(OH)4 + 4 HCl≠ (19)
where Z– is a negatively charged fragment of the zeolite framework.
After heat-treatment at 873 K of mixtures of MoCl5 (MoOCl4) and H-Y or
Co,H-Y a new band at 900 cm–1 appeared in the mid-infrared spectra. This band
was assigned to a Mo = O bond vibration. It was missing when mixtures of
MoCl5 (MoOCl4) and H-Y(D) were heat-treated. Therefore, Dai and Lunsford
proposed two model structures for Mo-containing cations in the solid-solid
exchanged zeolites, viz., structure I for Mo,H-Y and Mo,Co,H-Y and model II for
Co,H-Y(D):

XRD, IR of lattice vibrations and O2 adsorption experiments provided evi-

dence for a partial loss of crystallinity which occurred upon SSIE with MOCl4 ,
increasing in the sequence Mo,H-Y(D)<Mo,Co,H-Y ª Mo,H-Y. Possible reasons
for the degradation in crystal integrity were seen in a partial dehydroxylation of
the above hydrogen forms of Y-zeolite followed by rehydration in ambient air as
well as in the action of acidity produced in the exchange reaction with MoOCl4 .
At variance with the experiments by Dai and Lunsford [200], Kucherov and
Slinkin [92, 190] used an experimental procedure which enabled them to
exclude ambient air and water vapor. This might explain most of the differences
in their results, in particular with respect to the ESR signals and oxidation state
of the incorporated Mo cations. After heating at 423 K an evacuated ampoule
containing MoCl5/H-ZSM-5, Kucherov and Slinkin observed a color change of
the sample to red-violet and the development of the strong ESR signal shown in
Fig. 59, indicating uptake of Mo-containing species.
The above spectrum is typical of isolated Mo(V) cations. It exhibits a splitting
constant of 0.8 mT that is ascribed to the presence of 95Mo and 97Mo (spin quan-
tum number I = 5/2). A maximum of 4 ¥ 1020 spins per gram was reached, which
corresponded to about 100% replacement of protons from the acid Brønsted
sites of the H-ZSM-5 sample. Increasing the temperature of solid-state reaction
to 623 K caused sublimation of excess MoCl5 and a color change to pale yellow.
However, the intensity of the ESR signal shown in Fig. 59 did not decrease. This
finding suggested that the Mo-containing cations were relatively strongly held in
the zeolite structure. At elevated temperatures the cations were stable only in
Solid-State Ion Exchange in Microporous and Mesoporous Materials 149

Fig. 59. ESR spectra (obtained at 293 K) after solid-state reaction in the mixture MoCl5 /
H-ZSM-5 a on calcination in vacuum at 423 K and b upon admission of air to sample (a) (after
[23, 190], with permission)

vacuum. Oxidation at 723–823 K made the Mo(V) ESR signals irreversibly dis-
appear. A spectrum similar to that shown in Fig. 59 was obtained when H-MOR
was used instead of H-ZSM-5. In analogy to the interpretation given for the
incorporation of polyvalent cations such as vanadium or chromium cations
(vide supra), in the case of SSIE with Mo compounds it was also assumed that
not bare Mo5+ but complex cations such as MoCl4+ were the charge-compensat-
ing species inside the zeolite structures (cf. [92, 190]).
In a systematic study on incorporation by SSIE of molybdenum into hydro-
gen forms of narrow- or medium-pore zeolites, Bock [44] and Bock et al. [179]
investigated the effects of the nSi/nAl ratio, size of the zeolite crystallites, diame-
ter of the pores and dimensionalities of the pore systems of zeolites on their sol-
id-state reaction with MoCl3 . As parameters characterizing the reaction, the
temperatures of onset (Tstart) and completion (Tend) of the process were deter-
mined by temperature-programmed evolution of the gases (HCl, H2O) released
via SSIE. Thus, rather than steady-state experiments,dynamic experiments with
a heating rate of 5 K/min in the range 773–1073 K were carried out and, conse-
quently, the kinetics of SSIE likely affected the parameters mentioned above.
Bock [44] and Bock et al. [179] obtained the following results:
1. The acid strength of the H-ZSM-5 samples used (characterized by TPD of
ammonia) decreased somewhat when the nSi/nAl ratio increased from 15 to
110. In the same sequence Tstart decreased from an average value of 725 K to
697 K. This was surprising, since higher Brønsted acid strength was expected
to facilitate SSIE. However, this converse effect was ascribed to the decrease in
extra-framework Al-containing species with increasing nSi/nAl ratio as mea-
sured by 27Al MAS NMR. Extra-framework species were visualized as barriers
for the diffusion of the in-going entities. This would significantly affect the
kinetics of SSIE.
2. Tstart and Tend decreased with decreasing crystallite size (cf. Table 12). Again,
this effect was attributed to the influence of diffusion. Tstart and Tend were
150 H.G. Karge · H.K. Beyer

affected by both the heating rate and rate of uptake of the in-going Mo-con-
taining species. Thus, at a constant heating rate, one would expect an effect of
the diffusion pathways. Since larger crystallites present longer diffusion path-
ways, higher temperatures were reached before measurable amounts of HCl
were evolved indicating the onset of the reaction (Tstart) and, similarly, the
evolution of HCl ceased only at higher temperatures indicating the comple-
tion of solid-state reaction (Tend).
3. A significant influence of the pore architecture of the zeolites on SSIE existed
with MoCl3 . The effect of pore width and/or dimensionality is illustrated by
the data compiled in Tables 12 and 13. They were obtained with the following
zeolites: ZSM-35 (structure of ferrierite, FER, with intersecting pores or
channels built from 8- and 10-membered rings), EU-1 (one-dimensional pore
system with channels of 10-membered rings and large side-pockets), ZSM-48
(one-dimensional pore system of channels with 10-membered rings), and L
(one-dimensional pore system with channels of somewhat narrowed 12-
membered rings; cf. [180]). As expected, solid-state ion exchange was facili-
tated, i.e., Tstart and Tend decreased when the pore size increased (Table 13).An
exception seemed to be the case of cylindrical crystallites of the L-type. How-
ever, this is most likely due to the increased diffusion pathway along the axes
of the cylinders compared to the disk-shaped L-crystallites. A striking result
of these investigations were the considerably increased values of Tstart and
Tend observed with the one-dimensional channel systems (Table 13) com-
pared to those of the three-dimensional channel system of H-ZSM-5 (cf.
Table 12). Obviously, the uptake of the in-going Mo-containing species pro-
ceeds more rapidly into pore systems of higher dimensionality.

Table 12. Effect of crystallite size on (diameter, d) on the temperature of on-set and completion
of solid-state ion exchange between MoCl3 and H-ZSM-5

dcrystallite nSi/nAl Tstart , SSIE Tend , SSIE nCl–, out/nCl–, in

(µm) (K) (K)

10 110 550 1065 0.65

2 110 499 1021 0.66
Prevailing 1 (up to 3) 30 479 1006 0.66

Table 13. Effect of the pore size on the temperature of on-set of solid-state ion exchange
between MoCl3 and H-ZSM-5

Zeolite nSi/nAl Pore size Tstart , SSIE nCl–, out/nCl–, in

(nm ¥ nm) (K)

H-ZSM-35 8 0.42 ¥ 0.54 668 0.53

H-EU 15 0.41 ¥ 0.57 631 0.33
H-ZSM-48 40 0.53 ¥ 0.56 601 0.65
Zeolite L, disk-shaped 3 0.71 ¥ 0.71 559 0.53
Zeolite L, cylindrical 3 0.71 ¥ 0.71 653 0.54
Solid-State Ion Exchange in Microporous and Mesoporous Materials 151

Dai and Lunsford [200] as well as Kucherov and Slinkin [92, 190] hypothesized
that cations from molybdenum chlorides were not exchanged in solid state as bare
Mo6+ or Mo5+ cations but as complex cations, e.g., as [MoOCl2]+ and [MoCl4]+.
One of the most important results of the work by Bock [44] and Bock et al. [179]
is that this hypothesis was elegantly confirmed for the systems MoCl3/H-zeolites
by measuring the numbers of in-going (nCl-, in) and out-going (nCl-, out) Cl– anions.
In these experiments the disturbing effect of ambient water vapor (vide supra)
was excluded.According to Tables 12 and 13, reaction of MoCl3 (or its dimer) with
H-ZSM-5 resulted in an evolution of HCl corresponding to two-thirds of the intro-
duced Cl– anions. This result supports the proposition that not Mo3+ but
[Mo2Cl2]4+ or [MoCl]2+cations were incorporated. The chemistry is somewhat
more complicated in the cases of H-ZSM-35, H-L and H-EU-1. Here, perhaps only
one-half or one-third of MoCl3 (or Mo2Cl6) reacted to form [Mo2Cl3]3+ and
[Mo2Cl4]2+, respectively, which reside on cation sites of the zeolite structure. The
assumption that, upon solid-state reaction of molybdenum chlorides with hydro-
gen forms of zeolites, Cl-containing complex cations were generated inside the
zeolite pore systems was further supported by EDX analysis after solid-state reac-
tion between MoCl3 and H-ZSM-5. This analysis clearly proved the presence of
chlorine in the materials obtained via SSIE between MoCl3 and H-ZSM-5.
Many experiments aimed at the introduction of molybdenum via SSIE of a
mixture of molybdenum oxide (MoO3) and H-ZSM-5 were unsuccessful, even
though MoO3 is volatile and has a relatively low lattice energy (Tm = 1068 K).
Kucherov and Slinkin [92, 190] ascribed the failure of this experiment to the fact
that MoO3 easily forms large polymeric aggregates that would be unable to enter
the zeolite pores. Reduction of precalcined MoO3/H-ZSM-5 mixtures by H2 at
573–673 K produced ESR signals similar to that shown in Fig. 59 but of very low
intensity [92]. Yuan et al. [201] and Wang et al. [216] reported that the reaction
between MoO3 and H-Y was significantly facilitated by the presence of water
vapor. Such a modification of solid-state ion exchange, however, will be dealt
with in Sect. 6.3.2.
In contrast to the reports by Kucherov and Slinkin [92, 190], Harris et al. [217]
claimed to have successfully introduced Mo-containing cationic species via SSIE
between MoO3 and H-ZSM-5. At low loadings (£8 wt.%), MoO3 appeared to be
highly dispersed on both the internal and external surfaces of the zeolite crys-
tallites, where the oxide species reacted with bridging OH groups (acid Brønsted
centers) and terminating silanol groups, respectively. This was deduced from the
reduction in the band intensities around 3601 cm–1 (by 70%) and 3740 cm–1.
Moreover, in the mid-infrared region bands appeared around 915 and 950 cm–1.
The former was seen as indicative of Mo = O stretching vibrations of incorpo-
rated cationic Mo-containing species such as:

The band at 950 cm–1 was assigned to T–O–T vibrations. XPS results showed
an essentially symmetric doublet originating from Mo(3d5/2) and Mo(3d3/2)
152 H.G. Karge · H.K. Beyer

electrons with binding energies typical of Mo(VI) species. However, ESR spectra
revealed that, in addition, a small number of Mo(V) species had formed upon
SSIE between MoO3 and H-ZSM-5.
In their study on the genesis of methane-activating sites in Mo-exchanged H-
ZSM-5, Borry et al. [218] and Kim et al. [219] prepared the zeolite catalysts by
reacting dry solid mixtures of MoO3 and H-ZSM-5 (nSi/nAl = 14.3) powders at
973 K in air. This post-synthesis modification was preferred by the authors over
the conventional technique of impregnation with ammonium hexamolybdate
because it eliminated the evolution of large amounts of N2 , NH3 , and H2O dur-
ing subsequent calcination in air. Moreover, it allowed accurate measurements of
the nature and rate of MoO3 exchange. Techniques employed to confirm SSIE in
the system MoO3/H-ZSM-5 were XRD, CA by AAS, N2 physisorption at 77 K, and
TPE of H2O and D2-OH exchange for the determination of remaining acid Brøn-
sted OH groups. Mixtures of MoO3 and H-ZSM-5 with Mo contents below 4 wt.%
exhibited spreading of Mo oxide to produce MoOx species on the external sur-
faces of the zeolite crystallites. These species gradually migrated into the H-
ZSM-5 channels and formed anchored [MoOx]n complexes. More specifically,
from the result of TPE of H2O and D2-OH it was derived that each Mo6+ replaced
1.2 (±0.1) protons in H-ZSM-5 as long as nMo/nAl £ 0.37. 27Al MAS NMR, X-ray
absorption and Raman spectroscopy (as previously reported [218, 220]) con-
firmed that, during heat-treatment of the MoO3/H-ZSM-5 mixture, isolated dite-
trahedral [Mo2O7]2– dimers formed, which contained framework oxygens asso-
ciated with two Al sites and were located at exchange centers according to
Eq. (20):

(20)

Thus, Mo,H-ZSM-5 samples prepared via SSIE proved to be catalyst precur-

sors which, after activation (reduction, carburization), activated CH4 molecules
and produced hydrocarbons such as C2H4 , C2H6 , C3H6 , C3H8 , benzene, toluene,
and naphthalene. At higher Mo contents, however, (inactive) aluminum molyb-
dates, Al2(MoO4)3 , formed by extracting framework Al that led to a loss of crys-
tallinity.
No real solid-state ion exchange was observed by Thoret et al. [197], who used
mixtures of molybdenum oxide, MoO3, and sodium forms of zeolites. Similar to
what was found by Marchal et al. [196] with the system V2O5/Na-Y, upon reac-
tion of MoO3 with Na-Y at low temperature (790 K) and low loading corre-
sponding to nMo/(nSi + nAl) = 0.075, molybdenum trioxide seemed to be incor-
porated as such into the porous zeolite structure. This was indicated by an
increase in the lattice parameter ao from 2.4649 to 2.4673 nm. For higher load-
Solid-State Ion Exchange in Microporous and Mesoporous Materials 153

ings, i.e., nMo/(nSi + nAl) = 0.125, and higher temperatures (above 750 K), Thoret
et al. observed increasing amorphization and finally total destruction of the
Na-Y lattice accompanied by the appearance of several new crystalline phases.
Harris et al. (vide supra, [217]) also did not observe solid-state ion exchange in
the system MoO3/Na-ZSM-5. In particular, no bands at 915 and 950 cm–1 devel-
oped upon heat-treatment of the MoO3/Na-ZSM-5 mixture, i.e., no zeolite-an-
chored MoOx species exhibiting Mo=O stretching vibrations formed.

5.3.5.6
Introduction of Tungsten

As Thoret et al. [197] observed, even tungsten trioxide, WO3 , was unable to
migrate into the Na-Y structure. Thus, the initial solid phases in a WO3/Na-Y
mixture remained unaffected by heating and coexisted up to 850 K. This was
ascribed to the relatively high lattice energy of WO3 (Tm = 1746 K), which was
supported by the behavior of oxides having properties similar to those of WO3 .
Thus, it was observed that ThO2, UO2, Nb2O5, and Ta2O3 did not enter the pore
system of Na-Y whatever the temperature and the composition of the mixture.
With respect to Nb2O5/Na-Y, however, this system should be compared with
Nb2O5/NH4-Y (vide supra, cf. Sect. 5.3.5.2). It seems possible that in the case of
acid Y-zeolite a penetration of WO3 into the structure would occur. In any event,
upon heating a mixture of WO3 and Na-Y beyond 850 K and at a loading corre-
sponding to nW/(nSi + nAl) ≥ 0.05, the zeolite lattice was distroyed.

5.4
SSIE of Noble Metal Compounds with Zeolites

5.4.1
Introductory Remarks

A particularly important post-synthesis modification of zeolites is that of ion

exchange with noble metal compounds. Zeolites modified this way are, after
reduction, frequently employed as catalysts for hydrogenation/dehydrogenation
of, e.g., unsaturated hydrocarbons. Introduction of noble metals into zeolite
structures can be achieved not only by conventional methods, e.g., ion exchange
in aqueous solution or by the incipient-wetness technique (cf. Sect. 6.1), but also
by solid-state ion exchange, usually between noble metal halides or oxides and
hydrogen forms of zeolites. In fact, SSIE is in several cases the only way to carry
out post-synthesis modification of zeolites by incorporation of noble metal
cations. Residual Brønsted acidity, which is formed on reduction of the noble
metal cations incorporated through SSIE, may be eliminated by subsequent SSIE
with alkaline metal compounds (cf. Sect. 5.1). However, for some categories of
reactions such as hydroisomerization or hydrocracking of hydrocarbons, both
hydrogenation/dehydrogenation and acidic functions are required. The respec-
tive bifunctional catalysts may be produced via SSIE as well. Thus, Zhang [221]
and Karge et al. [222–224] were the first to prepare noble-metal-containing zeo-
154 H.G. Karge · H.K. Beyer

lite catalysts via solid-state ion exchange in their study aimed at formulating
bifunctional catalysts (vide infra).

5.4.2
Preparation of Noble-Metal-Containing Large-Pore Zeolite Catalysts by SSIE

Palladium- and platinum-containing zeolites were prepared by Zhang [221] who

reacted successfully PdCl2, Pd(NO3)2 , PdO, PtCl2 , PtCl4 and PtO2 with NH4-Y or
H-ZSM-5 in experiments that provided the groundwork for the production of
bifunctional catalysts by SSIE (cf. Sect. 5.4.4). Hatje et al. [40, 225] investigated
solid-state ion exchange of platinum oxide, chloride and bromide with NH4-Y.
The process was monitored by TPR, XRD and, particularly interestingly, by dis-
persive extended X-ray absorption fine structure (DEXAFS) measurements and
time-resolved XAS. It was found that in the presence of oxygen reduction was
inhibited but, starting at 555 K, the Pt-Cl shell at 0.175 nm was replaced by a
Pt-O shell at somewhat lower distances (0.150 nm). The reflections of PtCl2 dis-
appeared (e.g., at 2Q = 12.85°). Concomitantly, the (111) and (220) reflections
increased, mainly due to the occupation of cation sites by Pt2+. In the absence of
air, PtO and PtCl2 were reduced at approximately 573 K to Pt0 under formation
of highly dispersed Pt clusters. In the case of reacting PtBr2 it was shown that
some bromide remained in the zeolite matrix. Thus, from their combined exper-
iments, the authors obtained evidence that the salt or oxide components migrat-
ed into the zeolite pores before reduction occurred provided the exchange was
conducted in atmospheric air. In contrast, when the reaction was carried out in
a flow of helium, the Pt cations were reduced by NH3 to Pt0, forming clusters
which could be characterized by EXAFS.
Mkombe et al. [226] investigated the dispersion of Pt0 on K-L prepared by var-
ious techniques of loading and calcination. They reported that at a level of
1.5 wt.% Pt, different loading techniques, viz., ion exchange in aqueous solution,
incipient-wetness impregnation and solid-state ion exchange, resulted in simi-
lar Pt dispersions. The dispersion was well correlated to the n-hexane conver-
sion in the aromatization reaction.
Solid-state ion exchange of rhodium chloride with the hydrogen form of
highly dealuminated Y-zeolite (DAY, nSi/nAl ª 300) was studied by Schlegel et al.
[227]. These authors used FTIR to monitor solid-state reaction through the
decrease in the intensities of the OH bands at 3631 cm–1 (HF band) and
3568 cm–1 (LF band). In vacuum and oxygen, exchange degrees of only about
25% were obtained as indicated by the weakening of both the HF and LF bands.
Because of the low Al content of DAY, introduction of complex Rh cations (vide
supra) was proposed according to the following chemistry:
[∫Si–(OH)–Al∫] + RhCl Æ [∫Si–(O)–Al∫]– RhCl+ + HCl ≠
3 2 (21)
[∫Si–(OH)–Al∫] + RhCl3 + 1/2O2
Æ [∫Si–(O)–Al∫]– RhO+ + HCl ≠ + Cl2 ≠ (22)

After adsorption of CO, IR bands at 2118 and 2053 cm–1 appeared which were
assigned to well-defined Rh(CO)+2 complexes. Simultaneously, the exchange
Solid-State Ion Exchange in Microporous and Mesoporous Materials 155

degree increased. The effect of CO on DAY materials modified by SSIE with

RhCl3 as described above prompted Schlegel et al. [227] to conduct the solid-
state reaction itself in the presence of CO. This will be dealt with in Sect. 6.3.2 on
exchange processes related to solid-state ion exchange.
Wasowicz and Kevan [228] showed that SAPO-11 has only a low exchange
capacity for Rh(III) ions in aqueous solution, whereas via SSIE significantly
higher loadings can be achieved, which enabled the authors to detect and study
by ESR the stabilized paramagnetic Rh species in SAPO-11. Rh(II) and Rh(I)
species were observed which were analogous to those found earlier in X zeolite.

5.4.3
Preparation of Noble-Metal-Containing Narrow-Pore Zeolites by SSIE

With respect to small-pore zeolites, especially those in which the pore mouths
are formed by eight-membered oxygen rings (8-MR openings), conventional ion
exchange in aqueous solutions of noble metal salts usually fails or provides only
low degrees of ion exchange. This is ascribed to geometric constraints: Fre-
quently, the (solvated) cations or complexes such as Pt(NH3)2+ 4 are not able to
penetrate the narrow-pore openings. In some cases these difficulties in prepar-
ing the desired metal-containing 8-MR zeolites may be circumvented by adding
the respective metal compound to the synthesis gel [229, 230].Application of this
method is, however, limited because of possible effects of the metal cations on
the crystallization process. Solid-state ion exchange offers an alternative route
and is frequently the only way for a post-synthesis modification of zeolites by
incorporation of noble metals. Systematic studies of the introduction of noble
metals (Pt, Pd, Rh) into narrow-pore (8-MR) zeolites have been conducted by
Bock [44] and Weitkamp et al. [45, 231]. Zeolites and zeolite-like materials used
in the above studies and some of their relevant properties are listed in Table 14.
PtCl2 , PdCl2 , PdO and RhCl3 were employed for SSIE. Ground mixtures of the
salts and zeolites with a composition such as to achieve a metal loading of
1 wt.% in the case of complete solid-state reaction were heated in a flow of heli-

Table 14. Properties of narrow-pore zeolites and zeolite-like materials used in solid-state ion
exchange with noble metal salts

Zeolite nSi/nAl Pore structure Pore size

(nm¥nm)
Ring type a Dimensionality b

H-RHO 3.0 8-MRc 3 0.36 ¥ 0.36

H-ZK5 2.2 8-MR 3 0.39 ¥ 0.39
H-ZSM-58 30; 50; 100 8-MR 1 0.36 ¥ 0.44
H-ALPHA 2.5 8-MR 3 0.41 ¥ 0.41
nAl/nP; nSi/(nAl + nP)
H-SAPO-42 2.3; 1.0 8-MR 3 0.41 ¥ 0.41
a Type of oxygen rings forming the channels of the structure.
b Dimensionality of the channel system.
c Eight-membered oxygen rings of the framework.
156 H.G. Karge · H.K. Beyer

um or nitrogen at 553–623 K. Before conducting a reaction over the exchanged

materials, they were dried at 393 K and reduced at 573 K in a current of hydro-
gen at atmospheric pressure with a typical flow rate of 15 cm3 min–1. The SSIE
process and the materials obtained were characterized by measurements of HCl
evolved, the temperatures of the onset of the solid-state reaction and, as a test
reaction, hydrogenation of a mixture of hex-1-ene and 2,2,4-trimethylpent-1-
ene. Prior to the hydrogenation experiments, any residual Brønsted acidity of
the catalyst was eliminated by poisoning through chemisorption of NH3 . Thus,
any effect of a competing acid-catalyzed double-bond isomerization was avoid-
ed. Results are presented in Table 15 and in Fig. 60.
The effects of (i) the nSi/nAl ratio of the zeolite, (ii) the lattice energies of the
metal compounds, (iii) the ion-pair sizes of the metal compounds, and (iv) the
pore width of the zeolites were considered. From Table 15 it is evident that SSIE
did occur in all of the investigated systems. This was indicated by evolution of HCl
except for ALPHA/PtCl2 and H-SAPO-42, where, above 705 and 773 K, respective-
ly decomposition of PtCl2 under evolution of Cl2 was observed. Furthermore, the
ratio of nCl–,in/nCl–,out was in almost all cases close to 1, showing that indeed bare
Pt2+, Pd2+ or Rh3+ cations were incorporated. For the systems PtCl2/H-ZSM-58 and
PdCl2/H-ZSM-58, Tstart increased with increasing nSi/nAl ratio (= 35, 50, 100) in the
sequence 598, 614, 636 K and 668, 682, 704 K, respectively. Bock [44] and Weitkamp
et al. [45] tentatively ascribed this behavior to a decrease in the driving force for
solid-state reaction with increasing nSi/nAl ratio: A lower Al content corresponded
to a lower number of extraframe-work cation sites and, therefore, to a lower met-

Table 15. Introduction of noble metals into narrow pore zeolites via solid-state ion exchange
Zeolite nSi /nAl Salt EL of salt Tstart nCl–, out /nCl–, in
(kJ mol–1) (K)

H-RHO 3.0 PtCl2 2819 639 1.01

PdCl2 2778 780 0.97
H-ZK-5 2.2 PtCl2 2819 677 1.0
PdCl2 2778 749 1.0
RhCl3 5644 891 1.0
H-ALPHA 2.5 PtCl2 2819 –a –a
PdCl2 2778 738 1.01
RhCl3 5644 864 0.97
H-ZSM-58 30 PtCl2 2819 589 0.98
50 PtCl2 2819 614 0.98
100 PtCl2 2819 636 0.98
30 PdCl2 2778 668 1.06
50 PdCl2 2778 682 1.06
100 PdCl2 2778 704 1.06
nAl/nP; nSi /(nAl + nP)
H-SAPO-42 2.3; 1.0 PtCl2 2819 –b –b
PdCl2 2778 776 0.6
RhCl3 5644 867 0.9
a Decomposition of PtCl2 at ca. 700 K.
b Decomposition of PtCl2 at ca. 773 K.
Solid-State Ion Exchange in Microporous and Mesoporous Materials 157

Fig. 60. Time-on-stream behavior of palladium-containing microporous catalysts in the com-

petitive hydrogenation of an equimolar mixture of hex-1-ene and 2,4,4,-trimethylpent-1-ene
at 343 K and W/Falkenes = 10 g h/mol (after [45], with permission)

al concentration on the surface and a smaller concentration gradient for the in-
going noble metal cations, respectively. In general, one would expect that Tstart
increases with increasing lattice energies of the metal compounds, i.e., in the
sequence PdCl2 <PtCl2 <RhCl3 (cf. Table 15). However, exceptions were PdO/H-
SAPO-42 and systems involving PtCl2 PdO possesses a significant lower ion-pair
size, which facilitates diffusion of the in-going Pd-containing species (most likely
PdO) through the pore structure. In the case of PtCl2/zeolite systems the above
deviation is possibly due to uncertainties in the determination of Tstart caused by
(partial) decomposition of the salt. Tstart for RhCl3/ZK-5 is higher than for
RhCl3/ALPHA and RhCl3/SAPO-42, because ZK-5 has narrower pores than the
other two microporous materials. In most cases, highly selective hydrogenation of
the slim n-hex-1-ene vs. 2,2,4-trimethylpent-1-ene occurred (see, e.g., Fig. 60).
This was in contrast to hydrogenation over other Pt catalysts, where the metal was
supported by amorphous carriers and both alkenes were hydrogenated at almost
the same rate. In the case of zeolitic supports, obviously only the slim linear n-hex-
1-ene had access to the noble metal clusters inside the microporous structures,
whereas the bulky tribranched 2,2,4-trimethylpent-1-ene could not enter the nar-
row 8-MR pores.(cf.[232,233]).The tribranched alkene could,therefore,be hydro-
genated only on the outer surface of the zeolite crystallites (cf. hydrogenation over
Pd/H-ZSM-5, vide infra). A significantly lower selectivity was found over
Pt/ALPHA, viz., about Y = 50% n-hexane vs.Y = 15% 2,2,4-trimethylpentane after
30 min on stream [44]. This might be due to differences in the incorporation of Pt
because of the decomposition of PtCl2 , as mentioned in Sect. 5.4.4. In summary,
these results led to the conclusion that all of the noble-metal-containing 8-MR
zeolite catalysts were active in alkene hydrogenation and, moreover, the majority
of the noble metal clusters were located in intracrystalline voids, i.e., in the interi-
or of the zeolite structures. Solid-state ion exchange of noble metal compounds
158 H.G. Karge · H.K. Beyer

with hydrogen forms of narrow-pore zeolites and zeolite-like materials such as H-

SAPO-42 proved to be a valuable route for the preparation of shape-selective
hydrogenation catalysts.

5.4.4
Preparation of Bifunctional Zeolite Catalysts by SSIE
Zhang [221] and Karge et al. [222–224] prepared via SSIE bifunctional zeolite
catalysts which possessed both a hydrogenation/dehydrogenation and an acid
function, i.e., noble metal aggregates and Brønsted acid OH groups, respective-
ly. The following mixtures were used for SSIE: PdCl2/H-ZSM-5; CaCl2, PdCl2/H-
ZSM-5; LaCl3, PdCl2/H-ZSM-5; PdO/H-ZSM-5; Pd(NO3)2/H-ZSM-5; PtCl2/NH4-
Y and PtCl4/NH4-Y. Solid-state reactions were carried out by heating (10 K/min)
the finely dispersed mixtures in high vacuum (10–5 Pa) to the desired reaction
temperature of 625 K for 2 h. The reactions were monitored by in situ IR and TPE
of evolved HCl. Solid-state ion exchange was successful in all systems indicated
above. For several mixtures the data in Table 16 show the decrease in intensity
(relative adsorbance, Arel) of the OH stretching band around 3605 cm–1 which is
typical of Brønsted acid OH groups in H-ZSM-5. XRD patterns as well as re-
exchange experiments with aqueous NH4Cl solutions confirmed that no loss of
crystallinity occurred upon SSIE (see Table 16).
TPE of HCl evolved and subsequent titration of trapped HCl during solid-
state reaction between PdCl2 and H-ZSM-5 is illustrated by Fig. 61. The experi-
ment shown in Fig. 61 demonstrates that no complete exchange of Pd2+ for the
protons of the Brønsted sites was achieved; rather an exchange degree of about
64% was reached. This was ascribed to the difficulty of establishing a full balance
of the negative framework charges by bivalent cations in ZSM-5, if the distances
between these charges are relatively long because of a low Al content such as in
the sample used (nSi/nAl = 33.7). The Pd,H-ZSM-5 materials obtained via SSIE
had a Pd content of 1 wt.% and were subsequently reduced in a flow of H2 (575 K,
3 h, 60 ml H2/min). After reduction, the catalysts were tested for hydrogenation
of ethylbenzene and dehydrogenation of ethylcyclohexane. It turned out that,
with respect to conversion of ethylbenzene, even in the presence of H2, acid-cat-
alyzed dealkylation and disproportionation of ethylbenzene predominated.
Main products were light paraffins, benzene and diethylbenzenes. Only minor
fractions of ethylcyclohexane and dimethylcyclohexanes were detected.
These results seemed to indicate that the acid and hydrogenation functions
were not properly balanced and larger palladium particles had formed at the
Table 16. Relative absorbance, Arel , of the OH band at 3610 cm–1 and ion exchange capacity for
NH4+ before and after solid-state ion exchange

Sample Parent zeolite Mixture of H-ZSM-5 with

H-ZSM-5
PdCl2 CaCl2 PdO Pd(NO3)2

Relative absorbance, Arel , (%) 100 42 37 55 40

NH4+ ion exchange capacity (mmol g–1) 0.37 0.36 0.36 0.39 0.34
Solid-State Ion Exchange in Microporous and Mesoporous Materials 159

Fig. 61. Titration of HCl evolved upon temperature-programmed heating of the mixture
PdCl2/H-ZSM-5 (after [223], with permission)

external surface of the zeolite crystallites. This was confirmed by TEM images.
Earlier work had shown that SSIE of alkaline earth and lanthanum chlorides
with hydrogen forms of zeolites was able to reduce the acidity of the zeolite (cf.
Sects. 5.1 and 5.3). In fact, solid-state ion exchange of H-ZSM-5 with increasing
amounts of, e.g., CaCl2 , reduced the density of Brønsted acid OH groups as indi-
cated by the decrease of the intensity of the OH band at 3605 cm–1, A(OH), and
the decrease in the conversion in ethylbenzene disproportionation, X(EB). How-
–
ever, temperature-programmed desorption of NH3 (Tmax , Ed and microcalori-
metry of NH3 adsorption (DHad) revealed that the acid strength remained essen-
tially unaffected (cf. Table 17).
Thus, to reduce the density of Brønsted acid sites in the bifunctional zeolite cat-
alysts,introduction of both Ca2+ and Pd2+ was carried out,and this was done either
simultaneously (method A) or successively (method B). Best results were obtained
by method B.As an example, Fig. 62 shows the conversion of ethylbenzene and the
yields of ethylcyclohexane, xylenes, dimethylcyclohexanes, alkanes, benzene and
diethylbenzenes over a Pd,Ca,H-ZSM-5 catalyst prepared by a two-step solid-state
ion exchange. In a first step, CaCl2 was incorporated via SSIE, followed by a second
step, viz., SSIE of PdCl2 into Ca,H-ZSM-5 obtained in the first step.
Dealkylation (or hydrogenolysis) of ethylbenzene was largely and dispropor-
tionation almost completely suppressed. Similarly, dehydrogenation of cyclo-
hexane over Pd,Ca,H-ZSM-5 prepared via successive SSIE proceeded with high
selectivity and very low catalyst deactivation (cf. Fig. 63).
160 H.G. Karge · H.K. Beyer

Table 17. Acidic properties of H-ZSM-5 and Ca,H-ZSM-5 samples obtained by solid-state ion
exchange.A(OH), maximum absorbance of the band of acidic OH groups at 3610 cm–1; X(EB),
conversion of ethylbenzene; Tmax , peak temperature obtained by TPD of NH3 from Brønsted
–
acid sites (cf. [29]); Ed , most frequent energy of activation for desorption of NH3 from Brøn-
sted acid sites (cf. [29, 222, 223]); DHad , differential heat of adsorption of NH3 (cf. [222, 223])
–
Sample nCa/nAl A(OH) X(EB) Tmax Ed DHad
no. (arb. units) (%) (K) (kJ · mol–1) (kJ · mol–1

1 0.0 0.30 4.0 585 99 146

2 0.2 0.20 2.2 610 102 146
3 0.3 0.14 2.0 605 101 138
4 0.4 0.11 1.9 600 102 138

Fig. 62. Hydrogenation and hydroisomerization of ethylbenzene over a Pd,Ca,H-ZSM-5 cata-

lyst prepared by successive solid-state ion exchange of (first) CaCl2 and (second) PdCl2 with
H-ZSM-5 followed by reduction in H2 (after [223], with permission)

It was assumed that the introduction of Ca2+ preceding the solid-state reaction
with PdCl2 not only affected the acidity but also facilitated the generation of a
more homogeneous distribution of small palladium particles inside the zeolite
matrix on subsequent reduction by H2 . By a test reaction, i.e., competitive hydro-
genation of branched 2,4,4-trimethylpent-1-ene vs. slim n-oct-1-ene, it was
shown that, indeed, essentially all of the platinum particles were located in the
interior of the zeolite crystallites (cf. Sect. 5.4.3 and [45, 221, 232, 233]). Electron
micrographs obtained by TEM indicated well-dispersed relatively small metallic
palladium aggregates. Figure 64 presents evaluations of such micrographs.
It is evident from Fig. 64 that, in contrast to case A of simultaneous SSIE, the
two-step solid-state reaction (case B) and subsequent reduction resulted in a
shift to smaller particles and narrower particle size distributions with increas-
Solid-State Ion Exchange in Microporous and Mesoporous Materials 161

Fig. 63. Dehydrogenation and isomerization of ethylcyclohexane over a Pd,Ca,H-ZSM-5 cata-

lyst prepared by successive solid-state ion exchange of (first) CaCl2 and (second) PdCl2 with
H-ZSM-5 followed by reduction in H2 (after [223], with permission)

Fig. 64. Size distribution of Pd0 particles from electron microscopy micrographs of Pd,Ca,
H-ZSM-5 catalysts obtained by A simultaneous and B successive introduction of Ca2+ and Pd2+
via solid-state ion exchange with CaCl2 and PdCl2 at 675 K (2 h, HV) followed by reduction in
H2 (after [223], with permission)
162 H.G. Karge · H.K. Beyer

ing Ca2+ preloading. Concomitantly, the catalytic performance of the materials

produced via two-step SSIE improved in that they exhibited higher selectivity
and stability. A tentative explanation of this finding was given as follows: Simul-
taneous SSIE of CaCl2 and PdCl2 with H-ZSM-5 probably generated ‘islands’ of
either adjacent Ca2+ or Pd2+ populations, whereas in the case of successive
exchange in the first step a relatively homogeneous population of the cation sites
by Ca2+ and, correspondingly, a homogeneous distribution of the remaining
protons was reached, which reacted in the second step with PdCl2 . Upon reduc-
tion of Pd2+ to Pd0, the neighboring Ca2+ cations could act as anchors for the
rather mobile Pd0 atoms and, thus, favor the formation of small palladium
aggregates. Sachtler and co-workers [234, 235] were the first to advance the idea
of polyvalent cations acting as anchors for small aggregates of noble metals in
zeolite matrices. Results similar to those reported for SSIE of PdCl2 with H-ZSM-5
were obtained when the other mixtures discussed above were studied. Further-
more, preceding introduction of LaCl3 in a two-step SSIE with palladium com-
pounds and H-ZSM-5 had effects similar to those described for CaCl2 .

6
Modified SSIE and Related Processes
6.1
Introductory Remarks
There are several processes which seem to be more or less closely related to sol-
id-state ion exchange in zeolites. An early reported example is that of the so-
called contact-induced ion exchange (cf. Sect. 6.2). In 1986, Kokotailo et al. [19]
and Fyfe et al. [20] showed that mere physical contact of crystallites of two sam-
ples of the same zeolite structure but loaded with different cations led finally to
one sample of crystallites exhibiting a homogeneous distribution of both types
of cations. However, later, Karge and Koy [236] demonstrated that the phenom-
enon of contact-induced ion exchange mentioned above does not occur in the
total absence of water. Thus, contact-induced ion exchange is, in fact, mediated
by water filling the zeolitic micropores. In essence, the mechanism is the same as
that of conventional ion exchange. In the frequently employed incipient-wetness
technique of impregnation (cf. [237–241]), two mechanisms are consecutively
in operation. In the first stage, i.e., at lower temperatures when the zeolite pow-
der is impregnated with an aqueous solution of a compound of the in-going
cation, a fraction of the ions are exchanged as in the conventional process except
that, due to omission of the washing step, the out-going cations and the anions
of the impregnating salt remain in the product of this procedure. However, when
the paste produced via impregnation is subsequently heated to complete the
exchange, the mixture of residual salt and zeolite powder is dried and real solid-
state ion exchange can then take place. However, if no volatile compounds are
formed in this second step, an equilibrium with an exchange degree lower than
100% will be reached (cf. Scheme 1b, c in Sect. 2). In Sect. 7 it will be shown that
real solid-state ion exchange does not require any presence of water but will pro-
ceed in an absolutely water-free system as well.
Solid-State Ion Exchange in Microporous and Mesoporous Materials 163

In a sense, even the introduction of cations mediated through the vapor phase
of a metal or metal compound is related to solid-state ion exchange (cf. Sect. 6.3).
In fact, in the former case, the in-going entities are adsorbed from the vapor
phase onto the external surface of the zeolite crystallite. In the case of solid-state
ion exchange, i.e., on thermal treatment of a physical mixture of (water-free) salt
or oxide of the in-going cation and the (water-free) zeolite powder, small entities
(ions or, more likely, molecules; see Sect. 7.2) will be firstly separated from the
salt or oxide lattice and then cover the external surface of the zeolite crystallites.
Indeed, here also it cannot be excluded that the transport from the solid salt or
oxide to the zeolite surface may proceed through the gas phase, even though it is
generally assumed that the separation of these small entities is facilitated by the
intimate contact of both solids, and their transport may occur via surface diffu-
sion. In any event, once the external surfaces of the zeolite crystallites are cov-
ered with the in-going entities, the situation is the same in both cases, i.e., the
species must migrate into the (water-free) pores of the zeolite and react there.
An interesting modification of solid-state ion exchange as described in
Sects. 2 and 5 is to be seen in those cases where the exchange requires in addi-
tion to the two solids (zeolite and salt or oxide) the presence of certain vapors or
gases such as H2O, O2 , H2 , CO, NH3 , or volatile hydrocarbons (cf. Sect. 6.3). Very
important examples are encountered in cases of reductive solid-state ion
exchange (RSSIE).

6.2
Contact-Induced Ion Exchange

Employing an interesting experiment, Kokotailo et al. [19] and Fyfe et al. [20]
were able to show that even at room temperature cation exchange occurred
between, e.g., Li-A and Na-A simply upon intimate contact beween the crystal-
lites of the dry powders of both different cationic forms of zeolite A. This phe-
nomenon was demonstrated by 29Si MAS NMR and XRD measurements. As
expected, the mixture of Li-A and Na-A exhibited initially two well-separated
29Si MAS NMR signals, viz., at –85.1 and –88.9 ppm (referenced to TMS). This

reflected the different local environments of Si in Li-A and Na-A. Similarly, the
XRD patterns showed separated reflections (cf. Fig. 65).
After the length of time required for equilibration under ambient conditions,
only one sharp 29Si MAS NMR line was observed, and the corresponding split-
ting of the XRD reflections disappeared. These observations confirmed that
after equilibration only one single phase existed where Li+ and Na+ cations were
homogeneously distributed over all zeolite A crystallites. Similar findings were
reported for the pairs Li-A/NH4-A, Li-X/Na-X, Li-X/NH4X, and Li-A/Na-MOR
and by Huang et al. [33] for the systems Li-A/Na-Y, Li-A/Na-X, and Li-A/Ca-X.
Koy and Karge [236], however, were able to prove that this type of contact-
induced ion exchange requires the presence of residual adsorbed water in the
pores of the zeolite crystallites. These authors used completely dried Li-A and
Na-A samples and carried out all experimental steps (mixing, filling of the cap-
illaries for XRD measurements, sealing of the capillaries, etc.) in an efficiently
working glove box (PH2O <10–7 Pa). The samples then exhibited the expected
164 H.G. Karge · H.K. Beyer

Fig. 65. Contact-induced ion exchange between crystallites of Li-A and Na-A zeolites; 29Si
MAS NMR and XRD patterns of Li-A/Na-A mixtures immediately after mixing (initial) and
after equilibrating at ambient conditions (equilibrium) (after [20], with permission)

Fig. 66. Check of contact-induced ion exchange between crystallites of Li-A and Na-A: A XRD
patterns obtained under exclusion of even traces of water and B after admission of ambient
moisture (cf. [236])

splitting of the reflections in their XRD patterns. However, with the samples pre-
pared in this way, even after very extended observation times, no collapse of the
two separated sets of reflections was found (Fig. 66A). Only when the capillaries
were opened and moisture from the ambient air had access to the samples did
the splitting disappear, showing that the presence of water was required to make
the Li+ cations of Li-A crystallites migrate into Na-A crystallites and vice versa
until an equilibrium was established (Fig. 66B).
Solid-State Ion Exchange in Microporous and Mesoporous Materials 165

Contact-induced ion exchange was also studied by Fraissard et al. [242].

These authors used the system Rb,Na-X/Na-Y and monitored the exchange by
129Xe NMR. Interestingly, the exchange was not observed upon mixing at 300 K,

but only after heat-treating the Rb,Na-X/Na-Y mixture under vacuum to 673 K.

6.3
Gas-Phase-Mediated Processes Related to SSIE

6.3.1
Introduction of Cations into Zeolites Through a Vapor Phase Containing
the In-Going Species
In Sect. 6.1 it was indicated that incorporation of cations via an interaction of a
vapor phase (containing the in-going species) and the solid phase (the zeolite)
is, in a certain sense, related to SSIE in that in both cases the in-going species or
their precursors must be first adsorbed onto the external surface of the zeolite
crystallites in order to subsequently migrate into the zeolitic pore system.
Adsorption may occur in the case of volatile metals, oxides or halides. Examples
are the introduction of Zn2+, Ga3+ and Fe3+ into hydrogen forms of zeolites via
the reaction of vapors of ZnCl2 , GaCl3 and FeCl3 , respectively, with zeolites
[243–245], the incorporation of V-, Ti- and Cr-containing species from VOCl3 ,
TiCl4 and CrO2Cl2 gas phases, respectively [206, 246], and the reaction of vapors
of metallic zinc or cadmium with, e.g., H-Y or H-ZSM-5 [136, 247–249].
For instance, Guisnet et al. [244] prepared Zn-doped ZSM-5 by contacting the
degassed zeolite H-ZSM-5 with a flow of ZnCl2 in N2 at 558 K (sublimation tem-
perature of ZnCl2). Incorporation of vanadium into silicalite-1 and H-ZSM-5
was achieved by contacting silicalite and H-ZSM-5 at 593–793 K with a nitrogen
stream saturated with VOCl3 vapor at 293 K (pVOCl3 ª 2.1 kPa). IR spectroscopy
showed that in silicalite silanol groups were efficiently removed under forma-
tion of (∫ Si)3 ∫ VO species, which were relatively resistant to hydrolysis. In addi-
tion, in H-ZSM-5 the protons of the acid Brønsted sites associated with the
framework aluminum were eliminated. No degradation of the crystallinity was
indicated by XRD. According to the ESR spectra of the modified and hydrated
MFIs, small amounts of paramagnetic vanadyl species were present. However,
most of the vanadium remained as V5+. The materials were active in catalytic
gas-phase oxidation of toluene with O2 . The behavior of TiCl4 and CrO2Cl2 was
similar, but CrO2Cl2 was less efficient in removal of the silanol groups and Brøn-
sted acid sites [206, 246].
As reported by Boddenberg and co-workers, vapor of Zn0 did not undergo
any reaction with Na-Y (at 693 K and pZn ª 0.2 hPa), while contact with H-Y led
to a degree of exchange of more than 90% of the initial acidic protons [136].
129Xe investigation, volumetric determination of the adsorption isotherms of CO

and Xe on the resulting Zn-Y samples, and application of the multi-site adsorp-
tion model [137] yielded the distribution of the Zn2+ cations on the cation sites
in the supercages. Under anhydrous conditions, an unusually high population of
the SIII and SII sites was found [247], distinctly higher than in samples prepared
through conventional or solid-state ion exchange (cf. [136]). Contact with water
166 H.G. Karge · H.K. Beyer

vapor reduced the population of the supercage sites. In this context, Seidel and
Boddenberg discussed the possibility of developing the reaction between zinc
vapor and hydrogen forms of zeolites to a titration procedure for zeolitic pro-
tons under anhydrous conditions [247]. Similar results to those obtained with
zinc vapor were reported for the reaction of cadmium in the system Cd0-
vapor/H-Y [248].With respect to the interaction of zinc vapor with H-Y and sub-
sequent oxidation of Zn0 to Zn2+ by the protons of the H-Y zeolite, Seidel et al.
reported that this technique of zinc incoporation led to a more homogeneous
distribution of the Zn2+ cations within the macroscopic sample than conven-
tional ion exchange in aqueous solutions of Zn(NO3)2 [136]. In a more recent
contribution, Rittner et al. [248] derived from a 129Xe NMR investigation of
Zn-containing Y zeolite, which was prepared via reaction between zinc vapor
and H-Y, that in different regions of the sample zinc cations occurred in at least
two different formal oxidation states, viz., as Zn2+ cations and Znx2+ clusters with
x ≥ 2. The most probable cluster was assumed to be Zn2+ 2 . According to
Beyer et al. [250] it is possible to determine the density of the reacting Brønsted
acid OH groups by measuring the hydrogen evolved.

6.3.2
Effect of Additional Molecules in the Vapor Phase on SSIE at Elevated Temperatures

In Sect. 5.3.5.5 it was reported that attempts to introduce molybdenum into

hydrogen forms of zeolites via SSIE with MoO3 were usually unsuccessful (cf.
[92, 190]). The failure of the respective experiments was ascribed to the propen-
sity of MoO3 to form polymeric aggregates which are too bulky to penetrate into
zeolitic pores. However, Yuan et al. [201] and Wang et al. [216] reported that an
exchange reaction between MoO3 and H,Na-Y did occur in the presence of water
vapor. This effect was most likely due to a reaction at higher temperatures
(723 K) of water vapor with the molybdenum oxide that prevented MoO3 from
polymerization and/or led to a disaggregation of initially formed bulky poly-
meric species. Thus, it seemed possible that smaller Mo-containing entities
formed which were able to migrate into the zeolite pores and react there with the
protons of, e.g., H,Na-Y. The fact that incorporation of Mo into the zeolite had
occurred was evidenced by XRD, FTIR, ESR and back-exchange of the intro-
duced Mo-containing cations.
In the presence of carbon monoxide, CO, metals such as Mo, Re or Ru fre-
quently form carbonyls, which subsequently may be encaged in zeolites and,
after decomposition, transformed into small metal aggregates. However, in sev-
eral cases, the encapsulated carbonyls are transferred into cationic species and
anchored as such in the zeolite matrix. For example, Co2(CO)8 was oxidized by
the protons of H-Y to Co2+ and held at sites in the supercages, as suggested by IR
investigations [251]. Formation of the cations may be due to a reaction as
described by Eq. (23).

4Co2(CO)8 + 4H+Z– Æ 2Co2+Z –2 + 12CO + 2H2 + 2[Co(CO)4] + [Co4(CO)12]

(23)
Solid-State Ion Exchange in Microporous and Mesoporous Materials 167

Enhanced solid-state ion exchange in the presence of CO was also observed by

Schlegel et al. [227] who attempted to introduce rhodium into dealuminated
Y-type zeolite via reaction with RhCl3 (vide supra, Sect. 5.4.2). When SSIE
between RhCl3 and dealuminated H-Y was conducted in vacuum or oxygen, the
degree of exchange was limited to 25%, even though the reaction temperature
was as high as 675–875 K. When, however, SSIE was carried out in the presence
of a CO-containing gas phase, the ion exchange occurred at much lower tem-
peratures (375–475 K), and more than 50% of the protons were exchanged for
cationic rhodium species. The authors assumed that on the outer surface of the
zeolite crystallites during calcination in the presence of residual H2O and/or
oxygen rhodium oxychloride or rhodium oxide had formed that was subse-
quently reduced by CO to yield dicarbonyl species:
RhCl3 + H2O Æ RhOCl + 2 HCl ≠ (24)
RhOCl + 3 CO Æ Rh(CO)+2 Cl– + CO2 ≠ (25)
These dicarbonyl species were able to migrate into the zeolitic pores and react
there with acidic OH groups, as evidenced by FTIR. Thus, stable Rh(CO)+2 were
finally formed on cation positions:
Rh(CO)+2 Cl + H+Z– Æ Rh(CO)+2 Z– + HCl ≠ (26)
A new route for introduction of polyvalent cations from the respective oxides
into hydrogen forms of several zeolites, viz., H-ZSM-5 (nSi/nAl = 15), H-BETA
(nSi /nAl = 8), H-Ferrierite (nSi/nAl = 4), H-USY (nSi/nAl = 5), was found and
described by Kucherov et al. [252]. For instance, samples were prepared by
grinding precalcined MoO3 with the zeolites mentioned above. Then the mix-
tures were calcined in an air flow at 773 K for 2 h, cooled to 293 K, heated in an
air stream saturated at 293–296 K with tetrachloromethane, CCl4 , (pCCl4 =
10–13 kPa) up to 323–403 K for 2 h and finally purged in pure dry air for 2–3 h.
The samples produced in this manner exhibited well-resolved ESR signals typi-
cal of isolated Mo5+-containing ionic species, as evidenced by comparison with
standard samples prepared via impregnation with aqueous ammonium molyb-
date solutions. In contrast, no such signals were observed previously when the
experiments were conducted without the assistance of a gas phase containing
CCl4 (vide supra, Sect. 5.3.5.5). It was proposed that the interaction between
MoO3 and CCl4 resulted in the formation of rather mobile reactive species such
as molybdenum oxychlorides. In view of the large distances between neighbor-
ing framework Al atoms, however, it was assumed that the incorporated isolated
Mo species were not bare Mo5+ cations but rather complex entities such as
(MoOCl2)+ or MoCl +4 (vide supra). CCl4-assisted incorporation of Mo5+ was
much more efficient with H-ZSM-5 and H-BETA than with H-ferrierite and
H-USY. In the case of ferrierite, this was explained by the low effective pore
diameter (0.4 nm vs. 0.6 nm in ZSM-5 and BETA). For USY (0.8 nm) it was
assumed that the low degree of exchange was probably due to a significant
aggregation of the Mo5+ species in the large cavities of the structure of the
Y-type zeolite.
168 H.G. Karge · H.K. Beyer

6.3.3
Oxidative and Reductive SSIE

6.3.3.1
Oxidative SSIE of Ag0, Cu0 and Pd0 in the Presence of O2 or Cl2

In earlier studies, Beyer and colleagues [253, 254] observed that tiny silver and
copper particles (Ag n0 , Cu n0 ) formed on the external surfaces of zeolite crystal-
lites upon reduction by H2 of Ag-Y and Cu-Y, respectively. Concomitantly, acid
zeolitic OH groups were restored. Upon calcination in oxygen, Ag0 remigrated
into the interior of the zeolite structure and reacted there with the protons of
these OH groups under formation of Ag+ on cation sites and water, whereas
external copper aggregates were irreversibly oxidized to CuO. In contrast,
Kucherov et al. [255] claimed that in the case of reduced Cu,H-ZSM-5 bulky Cu0
aggregates, which had formed upon reduction on the external surface of the zeo-
lite crystallites, could also be reoxidized to cupric species slowly remigrating
into the zeolite channels.
Similarly, Feeley and Sachtler [256] showed that cation exchange of solid pal-
ladium with H-Y was mediated by an oxidative gas phase. These authors
assumed that PdCl2 formed from Pd0 and chlorine and that PdCl2 subsequently
diffused into the pore system, reacted there and replaced the protons under for-
mation of Pd2+ on cation positions and HCl.

6.3.3.2
Reductive SSIE of Ga2O3

A very interesting and promising modification of solid-state ion exchange is

reductive solid-state ion exchange (RSSIE) which proceeds in a reductive atmos-
phere that is in contact with both solids, usually an oxide of the in-going cation
and the hydrogen form of a zeolite. Price et al. [257, 258] and Kanazirev et al.
[259–261] were the first to report on RSSIE in the systems Ga2O3 /H-ZSM-5,
In2O3/H-ZSM-5 and CuO/H-ZSM-5 in the presence of hydrogen. These exchange
processes were assumed to occur according to Eq. (27) or (28):
M2O3 + 2 H2 + 2 H+Z– Æ 2 M+Z– + 3H2O (27)
MO + 1/2 H2 + H+Z– Æ M+Z– + H2O (28)
where Z– again represents a monovalent negatively charged zeolite fragment.
RSSIE in the above systems was investigated by TPR/TG (temperature-pro-
grammed reduction monitored by a thermobalance), XRD, TEM/EDAX, XPS
and IR. However, in the case of CuO/H-ZSM-5, it turned out that employment of
a reducing agent was not necessary for incorporation of Cu cations, since auto-
reduction occurred.
Interest in the preparation of gallium-containing and indium-containing zeo-
lites by reductive solid-state ion exchange was stimulated by the fact that these
materials might also be promising catalysts for aromatization of alkanes and
redox reactions, respectively.
Solid-State Ion Exchange in Microporous and Mesoporous Materials 169

Gallium-containing catalysts were obtained by ball-milling b-Ga2O3 with H-

ZSM-5 [257]. Investigations using XRD allowed the exclusion of any phase
transformations of the thermodynamically most stable modification, b-Ga2O3
into a-, g-, d- or e-Ga2O3 , and any ion exchange induced by ball-milling. Howev-
er, it was shown that an intimate mixing of the fine powders of the oxide and
zeolite was necessary for subsequent reduction to Ga+Z–: no reduction was
observed when Ga2O3 and H-ZSM-5 were placed on the microbalance pan with-
out previous mixing. Also, no reduction was observed in the system Ga2O3/Na-
ZSM-5, which proved that zeolite acidity is likewise required for RSSIE. Under
appropriate conditions, Price and co-workers [262] achieved by RSSIE in the
system Ga2O3 /H-ZSM-5 a degree of exchange close to 100%, similar to the
results obtained for CuO/H-ZSM-5 (vide supra) and In2O3 /H-ZSM-5 (vide
infra). TPR experiments further showed that with Ga2O3 the stoichiometry of
the process is most likely described by Eqs. (29) and (30):
Ga2O3 + 2 H2 Æ Ga2O + 2H2O (29)
Ga2O + 2 H+Z– Æ 2 Ga+Z– + H2O (30)

Only a small fraction of Ga2O3 conversion may occur according to Eq. (31) upon
calcination under high vacuum:
Ga2O3 + 6 H+Z– Æ 2 Ga3+Z–3 + 3 H2O (31)

Subsequent reduction with H2 would restore two-thirds of the protons replaced

according to Eq. (31) (cf. [258]):
Ga3+Z3– + H2 Æ Ga+Z– + 2 H+Z– (32)

Step 1 according to Eq. (29) is consistent with the well-established gallium

chemistry in that the suboxide, Ga2O, is a product of reduction of Ga2O3 with
H2 . Further reduction of Ga2O to elemental Ga0 could not be completely ex-
luded, but appeared unlikely, since even with a large excess of Ga2O3 it was
impossible to exceed an upper limit of reduction. Such an upper limit of re-
duction, which was significantly lower than what had corresponded to the
total amount of admixed Ga2O3 , could not be rationalized if the reduc-
tion could proceed from Ga+ to Ga0. Thus, it is obvious that the zeolite was a
stoichiometric reactant, and the produced Ga+ was stabilized in the zeolite
matrix. Further, it was shown by IR spectroscopy, simply using the KBr pellet
technique, that after RSSIE of Ga2O3/H-ZSM-5 the typical IR bands at 648 and
456 cm–1 of Ga2O3 were no longer detectable if Ga2O3 had been admixed in
sub-stoichiometric amounts.When Ga2O3 was admixed in excess, its reduction
proceeded only to an upper limit corresponding to the Al content of the frame-
work or the respective density of Brønsted acid OH groups [259].
An IR investigation of the OH stretching region confirmed through the
decrease of the intensity of the 3610 cm–1 band that the acid OH groups were
largely consumed upon RSSIE. This was in contrast to the results of earlier
experiments [263] that failed to provide such IR evidence for an incorporation
170 H.G. Karge · H.K. Beyer

of Ga into H-ZSM-5 through the decrease of the OH band intensity. However, in

those experiments, an attempt was made to introduce Ga under non-reductive
conditions, i.e., by conventional ion exchange or impregnation.
TEM/EDAX investigations demonstrated the occurrence of RSSIE in Ga2O3/
H-ZSM-5 in that the Ga2O3 particles of the mixture had essentially disappeared
and Ga transfer into the zeolite particles had taken place [257]. Neither Ga-con-
taining phases on the external surfaces of the zeolite crystallites nor Ga enrich-
ment in their surface layers were detected.
Finally, XPS of the reduced Ga2O3/H-ZSM-5 mixture before and after Ar+
bombardment showed a remarkable increase in the Ga(3d) signal caused by a
large increase in the Ga-containing surface. This was due to the spreading of Ga
oxide over the internal zeolite surface and thus clearly indicated the incorpora-
tion of Ga into the bulk of the zeolite [259].
In fact, Price et al. [258] also reported on successful introduction of gallium
into extra-framework cation sites of the Ga analog of H-ZSM-5, i.e., into H-
[Ga]ZSM-5. However, in no case was evidence found for incorporation of Ga into
the framework via reaction of Ga2O3 with internal silanol-type hydroxyls of
ZSM-5, as suggested by Endoh et al. [264]. This should, inter alia, result in an
expansion of the unit cell of the zeolite, which was not observed by XRD. Rather,
those internal silanol-type groups that were especially abundant in the case of
H-ZSM-5 were assumed to also generate Ga+ cations and to be responsible for
the observed Ga2O3 reduction beyond the Al content corresponding to bridging
acid OH groups. Similar findings were reported for the system In2O3/H-BETA
(vide infra). Introduction of Ga into SAPOs is discussed in the Sect. 6.3.3.3 in the
context of incorporation of In by RSSIE.
Preparation of Ga-containing catalysts through RSSIE of b-Ga2O3 in physical
mixtures with hydrogen forms of zeolites was proposed as a particularly inter-
esting technique, because other popular methods such as impregnation with
Ga(NO3)3 solutions or conventional ion exchange may not have followed identi-
cal phase transformations and reduction pathways, especially when different
conditions for calcination and reduction were employed. Thereby, catalysts with
quite different properties (that have proved difficult to reproduce) may have
been obtained.
Ga-containing zeolites, i.e., Ga-ZSM-5 materials prepared from ground mix-
tures of Ga2O3 and H-ZSM-5, proved to be active and very selective catalysts for
the conversion of propane [257, 260, 262, 265] and n-pentane [258] to aromatics
under atmospheric pressure and at 750–780 K. In such catalytic experiments it
was observed that calcined mixtures of Ga2O3/H-ZSM-5 even without preceding
reduction by H2 exhibited, after some time on stream, increasing activity. This
suggested that the initial reduction of Ga2O3 was achieved by the hydrocarbons
of the feed or H2 generated from the hydrocarbons. However, admission of H2
greatly enhanced the activity.
In addition, desorption and decomposition of propylamines on Ga-ZSM-5
were investigated [260, 261, 266]. The features of these processes were com-
pletely different from those observed upon interaction of propylamines with
H-ZSM-5. Also, Ga-ZSM-5 prepared via interaction of Ga2O3 and H-ZSM-5 or
H-MOR in a physical mixture was employed as a catalyst for NO reduction by
Solid-State Ion Exchange in Microporous and Mesoporous Materials 171

CH4 [267, 268]. According to Kikuchi et al. [267], the activity of such catalysts is
lower than that of conventionally exchanged Ga-ZSM-5.

6.3.3.3
Reductive SSIE of In2O3

It has been shown by Kikuchi et al. [267, 268] that ground mixtures of Ga2O3 and
H-ZSM-5 as well as of In2O3 and H-zeolites (H-mordenite, H-ZSM-5, H-BETA,
H-Y) were, after thermal treatment, active and selective catalysts for the reduc-
tion of NO2 (NOx) with hydrocarbons in the presence of oxygen. They assumed
incorporated GaO+ to be the catalytically active species. A reduction step lead-
ing to Ga+Z– (or In+Z–, vide infra) was not explicitly carried out. However, most
likely, the materials were in fact reduced by the hydrocarbon involved in the cat-
alytic reaction. The degree of exchange of InO+ for protons was claimed to be
correlated to the strength of the Brønsted acid sites of the zeolites, as character-
ized by TPD of ammonia, thus determining the observed sequence in catalytic
activity for reduction of NO2 by CH4, viz., H-MOR>H-ZSM-5>H-BETA ª H-
Y>SiO2-Al2O3. In contrast to Ga-containing ZSM-5 (vide supra), the activities of
In-ZSM-5 prepared via RSSIE on the one hand and by conventional exchange on
the other were found to be similar [267–269]. Furthermore, such In-containing
zeolites were proven to catalyze the conversion of methanol to hydrocarbons
[270]. These observations caused an increased interest in the details of RSSIE in
In2O3/H-zeolite systems. Zatorski [270] has suggested that a direct reaction
between In2O3 and the hydrogen forms of zeolites discussed above occur that
lead to an incorporation of bare In3+ cations into the zeolite according to
Eq. (33):
In2O3 + 6 H+Z– Æ 2 In3+Z3– + 3 H2O (33)
However, this was not confirmed by subsequent studies of Kanazirev and
associates [260–262] and the work of Beyer’s group [271–275] (vide infra).
Kanazirev et al. [258, 259] investigated the incorporation of indium through
thermal treatment of mixtures of In2O3 and hydrogen forms of ZSM-5, zeolite Y,
mordenite and offretite under reductive conditions, in complete analogy to their
studies on the introduction of gallium into zeolites via reductive solid-state ion
exchange (RSSIE). Techniques used for these investigations were TEM/EDAX,
TPR/TG and IR. Indium was also introduced via RSSIE into large crystallites of
H-ZSM-5 when finely powdered In2O3 was loosely mixed with the zeolite crys-
tallites and subjected to reduction in H2 at temperatures as low as 623 K [276].
Similar to what was found with Ga2O3/H-ZSM-5, it was shown by TEM cou-
pled with EDAX that upon heat-treatment of In2O3/H-ZSM-5 under H2 the In2O3
particles disappeared and indium was tranferred into the zeolite.
Differential TPR profiles (obtained by subtraction of the DTG curves of the
indium-free zeolite) exhibited three peaks, viz., the LTP at low, MTP at medium
and HTP at high temperature, corresponding to dehydration, reduction/incor-
poration of In+ into the zeolite [cf. Eq. (27)] and reduction of (excess) indium
oxide to elemental In0, respectively. LTP and HTP always occurred when In2O3
was present in excess and also when samples of pure In2O3 or In2O3-loaded pro-
172 H.G. Karge · H.K. Beyer

Fig. 67. DTG/TPR curves (dehydration, reduction to In+ stabilized in zeolites, reduction to In0)
of In2O3 , In2O3/Al2O3 , and mixtures of In2O3 with H- or Na-forms of zeolites (offretite,Y, mor-
denite). Designation of samples: A sample 1, 25 In/H-OFF, i.e., 0.025 g In2O3 per gram dry zeo-
lite; sample 2, 35 In/H-Y, i.e., 0.035 g In2O3 per gram dry zeolite; sample 3, 25 In/H-MOR, i.e.,
0.025 g In2O3 per gram dry zeolite. B sample 1, In2O3 in Ar; sample 2, In2O3 in H2 ; sample 3, 25
In/Na-Y, i.e., 0.025 g In2O3 per gram dry zeolite; sample 4, 25 In/Al2O3 , i.e., 0.025 g In2O3 per
gram dry alumina. LTP low-tempreature peak; MTP medium-temperature peak; HTP high-
temperature peak (for details see text; after [260], with permission)
Solid-State Ion Exchange in Microporous and Mesoporous Materials 173

ton-free materials such as Al2O3 or Na-ZSM-5 were employed (cf. Fig. 67). The
middle peak (MTP) appeared only if a mixture of In2O3 and the acid hydrogen
form of a zeolite was subjected to RSSIE.
From the TPR experiments it was concluded that the upper limit of In incor-
poration was determined by the proton content. The ratio nIn, incorp /nH was
always found to be close to 1 despite the fact that various zeolite matrices and
In2O3 loadings were used. This ratio was never close to 0.33, which is the value
that would be expected if three protons were replaced by one In3+ as suggested
by Zatorski for SSIE in a non-reductive atmosphere [270].
Profiles of TPD of ammonia from H-ZSM-5 exhibited two peaks at about 500
and 670 K corresponding to weak and strong Brønsted acid sites [260]. Upon
RSSIE of In2O3/H-ZSM-5 at 9 wt.% of In2O3 loading, the high-temperature peak
at 670 K disappeared completely. At increasing In2O3 contents, the low-tempera-
ture peak at 500 K vanished as well. This decrease in the total concentration of
Brønsted acid sites was confirmed by IR spectroscopy through the decrease in
the intensity of the 3610 cm–1 band. When pyridine was adsorbed after the
reductive solid-state ion exchange, a corresponding decrease in the band at
1540 cm–1 was measured. This band originated from pyridinium ions and was,
therefore, related to the density of the acid OH groups. Concomitantly, a band at
1446 cm–1 developed indicating pyridine bonded to Lewis acid sites. In the case
of zeolites containing appreciable amounts of (internal) silanol-type hydroxyls,
e.g., H-ZSM-5 or H-BETA, these hydroxyl groups were assumed to also react to
some extent.
The increase in the MTP in the TPR profile mentioned above corresponded
well with the decrease in the total density of Brønsted acid sites. The consump-
tion of Brønsted acid sites was explained by the stoichiometry of the solid-state
reaction in analogy to the case of Ga incorporation [cf. Eqs. (29) and (30)]:

In2O3 + 2 H2 Æ In2O + 2 H2O (34)

In2O + 2 H+Z– Æ 2 In+Z– + H2O (35)

As a consequence of the elimination of acid OH groups upon RSSIE in In2O3/

H-ZSM-5 or In2O3/H-Y mixtures, the activity of these materials in acid-cat-
alyzed reactions significantly decreased (vide infra).
In an analogous way to Ga2O3 and In2O3, experiments were conducted to
prepare Cu-containing catalysts. However, while for preparation of Ga- and In-
containing catalysts the exposure of the oxide/H-zeolite mixture to a reductive
environment was an indispensible requirement, during thermal treatment, the
system CuO/H-ZSM-5 underwent so-called auto-reduction in that a certain
fraction of the framework oxygen anions were oxidized to molecular oxygen
under reduction of Cu2+ to Cu+ (vide supra, cf. [203, 204, 261]). Therefore, no
admission of a reductant to the system was necessary. In this context, it is worth
mentioning that, in contrast to other cations associated with extra-framework
oxygen (such as Cu(OH)+, [Cu-O-Cu]2+, FeO+), InO+ residing on cation sites in
zeolites did not give rise to auto-reduction, i.e., no molecular oxygen was
released.
174 H.G. Karge · H.K. Beyer

The studies on incorporation of indium into zeolites via reductive solid-state

ion exchange were considerably amended by the work of Beyer and colleagues
[271–275, 277, 278]. To confirm and elucidate RSSIE, these authors investigated
mixtures of In2O3 and hydrogen forms of zeolites such as H,Na-Y, H-ZSM-5, H-
MOR and H-BETA. For their measurements they employed IR spectroscopy with
and without pyridine as a probe as well as TPR and TPO, which was monitored
either thermogravimetrically by a microbalance or with a thermal conductivity
detector determining the hydrogen and oxygen consumed. Furthermore, acid-
catalyzed isomerization of m-xylene was used as a test reaction to characterize
the density of Brønsted acid OH groups. First, stoichiometric (nIn/nNH4 = 1; sam-
ple I) and non-stoichiometric (nNH4/nIn = 2; sample II; nNH4/nIn = 3; sample III)
mixtures of In2O3 and H,Na-Y were investigated. In agreement with the findings
of Kanazirev and associates [260, 261], the requirement of a reductive atmo-
sphere as well as that of the presence of Brønsted acid OH groups for introduc-
tion of In into H,Na-Y in the temperature range up to about 770 K was estab-
lished by TPR/DTG and IR.At higher temperatures, rapidly progressing thermal
dehydroxylation prevented any investigation of the indium incorporation in a
non-reductive atmosphere or vacuum. In this context it is worth mentioning
that, very recently (cf. [278]), it has been shown that ∫Al-O(H)-Si∫ hydroxyls
are more resistant to thermal dehydroxylation in mixtures of In2O3 with H-ZSM-
5 than in pure H-ZSM-5. Thus, incorporation of indium into this zeolite was
found to proceed at 810 K in vacuum. However, strong evidence was provided
that the incorporation is preceded by an auto-reduction of In2O3 (vide infra).
The DTG peak observed on reduction of pure In2O3 or In2O3/Na-Y with H2 at
730 K was, in the presence of acid OH groups, i.e., in the case of In2O3/H,Na-Y,
shifted to a lower temperature of about 630 K. This suggested a modification of
the reduction process according to Eqs. (34) and (35) (vide infra). The decrease
in the intensities of the OH bands in the IR spectrum of In2O3/H,Na-ZSM-5
upon RSSIE can be recognized from Fig. 68. This figure shows quite clearly that
both types of OH groups were affected. They were completely eliminated in the
case of the stoichiometric sample I.
Correspondingly, with the samples II and III, which possessed a 50 and 66%
excess of NH+4 (or H+), only 50 and 33%, respectively, of the total amount of the
initial OH groups were consumed during RSSIE. In fact, the OH groups indicat-
ed by the HF band at 3640 cm–1 seemed to be somewhat more reactive. After
pyridine adsorption, the non-reduced sample I exhibited a strong pyridinium
ion band at 1544 cm–1, whereas this band did not appear after RSSIE because of
the complete consumption of acid OH groups during reductive solid-state ion
exchange. Similar to the observation by Kanazirev et al. in the system In2O3/H-
ZSM-5 ([260], vide supra), a band at 1445–1455 cm–1 was seen instead, which is
indicative of pyridine coordinatively bonded to cationic In-containing species.
This band was better resolved in other experiments, in particular with H-BETA,
and will be discussed below. Concomitantly with the band at 1445–1455 cm–1, a
new band was observed at 1441 cm–1. It originated from pyridine attached to Na+
and, thus, demonstrated that residual sodium cations had been expelled from
their original sites inside the sodalite cages by the In-containing cationic species
which were formed by RSSIE.
Solid-State Ion Exchange in Microporous and Mesoporous Materials 175

Fig. 68. IR spectra of the OH stretching frequency range of a stoichiometric mixture of In2O3
and NH4-Y (nIn/nNH+4 = 1; sample I) a after calcination at 670 K for 1 h; b sample (a) reduced
with hydrogen at 760 K for 1 h; c sample (b) contacted with water vapor (2.7 kPa) at 420 K for
30 min; d sample (a) reoxidized with oxygen at 620 K for 1 h and subsequently degassed at
760 K for 1 h; e sample (d) contacted with water vapor (2.7 kPA) at 470 K for 30 min; and f sam-
ple (d) contacted with water vapor (2.7 kPa) at 620 K for 30 min (after [271], with permission)

Reduction was shown by TPR, IR and MS to be also possible to some extent

with NH3 and with CO [272]:
In2O3 + 2 CO Æ In2O + 2 CO2 ≠ (36)
In2O + 2 H+Z– Æ 2 In+Z– + H2O ≠ (37)
Beyer et al. [271–273] assumed a reduction mechanism for the system
In2O3/H,Na-Y similar to that proposed by Kanazirev et al. [cf. Eqs. (34) and (35)],
even though no evidence for the intermediate formation of In2O could be
obtained; this suboxide is probably unstable and non-existent in the crystalline
state [279]. The univalent indium cation, In+, was proven to be absolutely resis-
tant to further reduction by hydrogen at least up to 870 K, whereas In2O3 on non-
acidic zeolites (e.g., Na-Y) turned out to be reducible in one step to elemental In0
at temperatures somewhat higher than required for reductive solid-state ion
exchange in the system In2O3/NH4,Na-Y. Since In+ cations introduced by RSSIE
into zeolites were not further reducible, it was impossible to regenerate the orig-
inal OH groups according to Eq. (38):
M+Z– + 1/2 H2 Æ M0 + H+Z– (38)
in contrast to the case of Cu+, where
Eq. (38) applied (cf. [280, 281]).
However, the samples produced via RSSIE (InI-Y from sample I; InI,H-Y from
samples II and III) could be reoxidized in a flow of O2/Ar. Reoxidation com-
176 H.G. Karge · H.K. Beyer

menced already at room temperature and was completed at 350 K. Since the
original OH groups were not restored, it was suggested that the reoxidation took
place according to Eq. (39):
2 In+Z– + O2 Æ 2 (InO)+Z– (39)
The validity of Eqs. (34), (35) and (39) in describing the reduction and reoxida-
tion process in the system In2O3/H,Na-Y was unequivocally confirmed by mea-
surements of H2 and O2 consumption during the RSSIE and oxidation steps,
respectively. During RSSIE of sample I, that started at 370 K and was completed
at 670 K, two H2 molecules per In2O3 molecule were consumed. The consump-
tion of oxygen in the reoxidation step corresponded exactly to one O2 molecule
per In2O3 molecule or two In+ in agreement with Eqs. (34), (35) and (39). Sub-
sequent reduction of InO+Z– in H2 proceeded at 370–670 K. Thus, in the case of
the stoichiometric sample I, a fully reversible redox cycle was confirmed by mea-
surements of oxygen and hydrogen consumption during TPO and TPR, respec-
tively:
+O
2
2In+Z– ¨––––––––Æ 2InO+Z– (40)
+2H2 ,–2H2O

In the case of samples II and III, where an excess of Brønsted acid OH groups
was present, after reoxidation a secondary dehydroxylation was found by IR
measurements, which most likely resulted in the formation of [In–O–In]4+ com-
plexes [cf. Eq. (41)]. However, upon reduction of these reoxidized and dehy-
droxylated samples, IR evidenced a complete restoration of the excess Brønsted
acid OH groups [271]. Thus, in this case, the reversible redox cycle is described
by the following scheme:
+O2 –H2O
2In+Z– + 2H+Z– –––––––Æ 2InO+Z– + 2H+Z– –––––––Æ [In – O – In]4+Z–4
1442443 1442443 (41)
+ 2H2 , –H2O
≠0000000042 |

In view of the above redox cycles, In-containing zeolite catalysts prepared by

RSSIE are obviously promising candidates for redox reactions, because of their
fully reversible changes in the oxidation state of the indium species incorporat-
ed and stabilized in the zeolite matrices.
RSSIE was also carried out in the system In2O3/NH4(H)-BETA [274]. As in the
case of zeolite Y, incorporation of indium was achieved by thermal treatment of
mixtures of template-free NH4-BETA and In2O3 in a reductive environment.
Reduction proceeded by reaction with H2 or CO and, to a minor extent, by NH3
evolved during deammoniation. RSSIE in the system In2O3/H-BETA was moni-
tored by IR through consumption of OH groups. The acid Brønsted OH groups
(band at 3610 cm–1) preferentially reacted.
Consumption of acid OH groups was also proven by the decrease in the activ-
ity of the heat-treated In2O3/H-BETA samples in the acid-catalyzed isomeriza-
tion of m-xylene in comparison with equally treated pure zeolite H-BETA.
Concomitantly with the pronounced decrease in the bands associated with
bridging and silanol-like hydroxyls, upon RSSIE a strong band at 1446 cm–1
Solid-State Ion Exchange in Microporous and Mesoporous Materials 177

developed after adsorption of pyridine which was shifted to 1453 cm–1 by treat-
ment in O2 at 670 K.
The opposite effect occurred when the reoxidized sample was subjected to
another reduction. Applying Eqs. (34), (35) and (39), this enabled Beyer et al.
[274, 275] to assign the band at 1446 cm–1 to pyridine coordinatively bonded
to In+ (Py Æ In+) and the band at ca. 1455 cm–1 to pyridine attached to InO+ (Py
Æ InO+). The band at 1446 cm–1 was easily removed by degassing at 380–470 K,
i.e., the interactions (Py Æ In+) were similar to (Py Æ Na+). In contrast, the band
at 1455 cm–1 disappeared only on evacuation at temperatures higher than 570 K.
This suggested that extra-framework InO+ species exhibited higher Lewis acid
strength than In+ and behaved more like ‘true’ Lewis sites, i.e., extra-framework
AlO+ (cf. [171, 172]).
RSSIE in the system In2O3/BETA exhibited some pecularities. When as-syn-
thesized samples of BETA were employed, the zeolite still contained cations
(TEA+) derived from the template (tetraethylammonium hydroxide, TEAOH) as
used for the synthesis. Thus, RSSIE proceeded without an additional reductant,
because the decomposition products of TEA+ (ethene, alkylamines, hydrogen)
acted as reducing agents. Thus, In-containing BETA zeolites can be prepared
simply by heating a mixture of In2O3 and as-synthesized, template-containing
BETA. Furthermore, not only the acidic Brønsted OH groups associated with
framework aluminum (band at 3610 cm–1) but also part of the less acidic inter-
nal silanol-like hydroxyls occurring in large amounts in template-free H-BETA
were involved in the RSSIE process. This was indicated by the decrease in the
intensity of the band at 3730–3738 cm–1 that are typical of the latter kind of
hydroxyl groups [275].
Simultaneously, the band at 1453 cm–1 assigned to Py Æ InO+ interactions
(vide infra) developed to some degree during the RSSIE step, i.e., without admis-
sion of an oxidative reactant such as O2 . This was not observed when instead
of H-BETA other zeolites that were devoid of internal silanol groups (H-MOR,
H-Y) were subjected to RSSIE. Therefore, Beyer et al. [275] assumed that these
hydroxyls reacted with the strong reductant In+ similarly to the case of silica-
lite-1 [278] according to Eq. (42):
In+ + 2 [∫SiOH] Æ H2 + InO+ + [∫Si–O–Si∫] (42)
The incorporation of indium into NH4-MOR, H-ZSM-5 and NH4-Y by RSSIE as
well as the redox behavior of the cationic In species introduced in this way were
exclusively studied by IR (cf. [277]). Again, the progress of RSSIE was monitored
by the change in the intensity of bands typical of pyridine interacting with acid
hydroxyl groups and with incorporated indium cations acting as Lewis acid
sites. In their study, the authors stressed the outstanding suitability of the n8a
ring vibration mode of adsorbed pyridine in the range between 1590 and
1630 cm–1 for the detection of indium cations of different oxidation states and
their discrimination from other cations frequently occurring in zeolites, such as
Na+ and AlO+. The bands assigned to the n19b ring vibration mode of Py Æ Na+
and Py Æ In+ at 1442 and 1446 cm–1, respectively, and of Py Æ InO+ and Py Æ
AlO+ at 1554 and 1556 cm–1, respectively, generally strongly overlap. At variance,
the resolution of the spectra in the range of the n8a ring vibration mode is much
178 H.G. Karge · H.K. Beyer

Fig. 69. IR spectra of pyridine retained after adsorption at 470 K and subsequent degassing at
370 K on In2O3/NH4Na-Y which was A thermally pretreated in high vacuum at 720 K for 1 h,
B subjected to RSSIE at 720 K in H2 and C oxidized with O2 at 570 K subsequent to RSSIE, and
after successive degassing at 420 K (B1), 470 K (B2, C1), 570 K (B3, C2), 670 K (B4, C3) and 770 K
(C4) and after pyridine adsorption at 470 K and after immediately subsequent degassing at
670 K (C5) (after [277])

better due to larger differences in wavenumbers: 1591, 1600, 1612 and 1623 cm–1
for pyridine attached to Na+, In+, InO+ and AlO+, respectively.
Bands at 1442 and 1591 cm–1 appearing in the spectra of pyridine adsorbed
on In2O3/H,Na-Y were low in intensity prior to RSSIE due to the preferential
location of the residual Na+ cations in the sodalite cages at sites inaccessible to
pyridine. They became intense, however, after RSSIE (Fig. 69). At the same time,
when Na+ was indicated, bands typical of In+ incorporated according to Eqs. (34)
and (35) developed at 1446 and 1600 cm–1 (Fig. 69). To explain these effects it was
suggested that Na+ cations were expelled by a fraction of the in-going In+ cations
from sites in the sodalite cages into the large cavities. Upon treatment with O2 at
570 K, the bands characteristic of Py Æ In+ disappeared and intense bands at
1454 and 1612 cm–1 were seen which again proved the formation of InO+ species
according to Eq. (39). The broad bands in the range 1440–1480 cm–1, which only
appeared at a higher degassing temperature (670 K) in the spectra of pyridine
interacting with InO+ (cf. C3 and C5 in Fig. 69), were ascribed to strongly
adsorbed compounds formed by oxidation of adsorbed pyridine by InO+.
In one report by Beyer et al. [277] the same conclusions were consistingly
drawn with respect to the systems In2O3/H-ZSM-5 and In2O3/NH4-MOR from IR
spectra obtained after RSSIE, subsequent oxidation and another reduction by
H2. However, in the case of In2O3/H-ZSM-5, the bands at 1454 and 1612 cm–1
developed already immediately after the RSSIE process, even though with minor
intensities and besides the bands typical of In+. This phenomenon, already
Solid-State Ion Exchange in Microporous and Mesoporous Materials 179

Fig. 70. IR spectra of pyridine retained after adsorption at 470 K and subsequent degassing at
370 K on In2O3/NH4-MOR which was A thermally pretreated at 720 K in high vacuum for 1 h,
B subjected to RSSIE in H2, C oxidized with O2 subsequent to RSSIE, and D reduced with H2
after preceding oxidation (steps B-D were performed at 720 K for 0.5 h) (after [277])

observed upon RSSIE in BETA (vide supra), was also attributed to the reaction
of In+ with silanol groups [cf. Eq. (42)] known to be generally present in ZSM-5.
An apparently unusual chemical behavior was exhibited by In+ cations intro-
duced into NH4-MOR (Fig. 70) inasmuch as no bands characteristic of Py Æ
InO+ were observed after oxidation. Nevertheless, oxidation to the trivalent state
must have occurred, since the spectrum after subsequent reduction was identi-
cal with that obtained after RSSIE, i.e., In+ cations were restored (Fig. 70).
To explain these surprising results it was suggested that InO+ cations may
occupy hidden sites in the mordenite structure that are not accessible for pyri-
dine. Alternatively, indium ions may be bound in mordenite to five framework
oxygen atoms. In this case, the coordination of trivalent indium cations would
be fully saturated due to the additional oxygen atom belonging to InO+.
Neinska et al. [282–284] succeeded in introducing gallium and indium into
SAPO materials (SAPO-5, SAPO-34 and SAPO-37) via reductive solid-state ion
exchange in a flow of H2 . RSSIE was evidenced by TPR/TGA and the acidity
properties of the modified SAPOs characterized by TPD of ammonia and
TGA/TPD of propylamine [284]. The exchange reactions proceeded according to
Eqs. (29), (30) and (34), (35), respectively, where now Z– has to represent a nega-
tively charged fragment of the SAPO framework. It was shown that, in agree-
ment with Eqs. (30) and (35), an upper limit for the incorporation of Ga and In
existed, determined by the specific number of exchangeable protons. This num-
ber decreased in the sequence H-SAPO-34>H-SAPO-5>H-SAPO-37, as mea-
sured by TPR/TG and TPD of ammonia. XRD and TEM/EDAX proved that the
180 H.G. Karge · H.K. Beyer

crystal structures of SAPO-5 and SAPO-34 were not markedly damaged during
RSSIE. Ga and In were, indeed, introduced into these silicoaluminophosphates.
In contrast, SAPO-37 completely lost its crystallinity during RSSIE and/or upon
subsequent rehydration, even though TPR clearly indicated that ion exchange
had occurred.
In contrast, a procedure similar to the template-induced introduction of indi-
um into BETA zeolite [275] described above seemed to be more promising.
Neinska et al. [282–284] carried out such experiments with mixtures of In2O3
and as-synthesized, template-containing SAPO-5, SAPO-34 and SAPO-37. Again,
the organic templates (SAPO-5: triethylamine; SAPO-34: tetraethylammonium
hydroxide; SAPO-37: tetrapropylammonium hydroxide) or their decomposition
products acted as reducing agents. The presence of In2O3 facilitated the decom-
position of the templates in a flow of pure inert gas at 873 K. In the case of
In2O3/as-synthesized SAPO-5 and In2O3/as-synthesized SAPO-37, the DTG fea-
tures at ca. 700 K, which corresponded in In2O3-free SAPOs to the decomposi-
tion of the most strongly bound template species, were significantly shifted to
lower temperatures. Neinska et al. observed a marked decrease in the acidity and
activity in m-xylene isomerization of the materials prepared in this way com-
pared with those of H-SAPO-n (n = 5, 34, 37). This clearly indicated that ion
exchange of In+ had indeed occurred. The procedure of template-induced RSSIE
might be particularly helpful for post-synthesis modification of materials such
as SAPO-37, the template-free hydrogen form of which suffers from low stabili-
ty against hydration.
Incorporation of indium into MCM-41 by RSSIE was also studied using XRD,
FTIR, TPR and TPD techniques [285]. The process occurred as easily as with
zeolites. Some typical differences were associated with the particular acid prop-
erties of the mesoporous material and were discussed in terms of the peculiar
structure and composition of the MCM-41 framework.
In view of the general reaction described by Eqs. (29) and (30) attempts were
also made to introduce Fe, Cr, La or Eu cations into zeolites by thermal treatment
of mixtures of the hydrogen forms of zeolites with Fe2O3, Cr2O3 , La2O3 or Eu2O3
in a reductive atmosphere [273]. Indeed, in high vacuum at 760–790 K, solid-
state reactions were observed, but the respective M+ cations could not be stabi-
lized in the zeolite matrix. Rather, the solid-state reaction was accompanied by a
collapse of the zeolite structure. Thus, to the best present knowledge, reductive
solid-state ion exchange seems to be restricted to reactions with Ga2O3 and
In2O3 .
Incorporation of Ga or In via RSSIE into hydrogen forms of zeolites signifi-
cantly affected their catalytic properties. As already mentioned, Ga-ZSM-5 sam-
ples obtained by reductive solid-state ion exchange were active and selective cat-
alysts for aromatization of propane and n-pentane [257, 260, 265]. However, a
drop in activity was observed with respect to conversion of n-pentane after
loading H-ZSM-5 with In via reductive solid-state ion exchange. Moreover, in
sharp contrast to the Ga-ZSM-5 discussed above (cf. [259]), reduced In-ZSM-5
prepared via RSSIE yielded almost no aromatics upon reaction of propane or n-
pentane. Surprisingly, a catalyst prepared by simultaneous introduction of Ga
(4 wt.%) and In (3 wt.%) through RSSIE into ZSM-5 proved to be superior in
Solid-State Ion Exchange in Microporous and Mesoporous Materials 181

aromatization of n-pentane to any Ga-ZSM-5 sample with similar Ga loadings.

It seems that the combination of both cations (Ga+ and In+) in one and the same
zeolite offers a possibility for regulating the catalytic properties, viz., hydrogen
transfer ability (by Ga+) and modification of the acidity (by In+).
It was also demonstrated by Kanazirev and Price [261] that Ga- and In-con-
taining ZSM-5 samples prepared via thermal treatment of mixtures of H-ZSM-
5 with Ga2O3 and In2O3 , respectively, resulted in active catalysts for reactions of
amines if heat-treatment was carried out in a reductive atmosphere. When the
Ga-, In- or Cu-ZSM-5 materials obtained through reductive or auto-reductive
solid-state ion exchange were employed for catalytic amine conversion, the reac-
tion pathways were completely changed compared to those over H-ZSM-5.
While over H-ZSM-5 at low coverages unimolecular decomposition of adsorbed
alkylamines to alkenes and ammonia occurred, dehydrogenation, transalkyla-
tion and condensation reactions were the prevailing processes over Ga-, In- or
Cu- containing ZSM-5 catalysts produced by RSSIE. This change in the pathways
was ascribed to the elimination of Brønsted acidity and formation of Lewis acid-
ity upon RSSIE [261].
Similarly, the catalytic behavior of hydrogen forms of zeolites in methanol
conversion was completely changed upon loading with In2O3 via RSSIE. Instead
of acid-catalyzed selective dehydration to olefins, dehydrogenation became pre-
dominant resulting in products such as H2 , CO, CO2 and CH4 [260].

7
The Role of Water in and Mechanisms of SSIE

7.1
Role of Water in SSIE

In most experiments on solid-state ion exchange with zeolites as discussed in

Sect. 5, the reactants (salts, oxides, zeolites) were indeed ‘dry’, but this does not
necessarily mean that water was totally excluded. Usually, these experiments
were carried out under ambient conditions, so that water could be adsorbed into
the zeolitic pore system. However, in particularly designed experiments, which
will be discussed below, it has been shown that the presence of adsorbed water,
even though not detrimental to SSIE, was in no case a prerequisite for solid-state
ion exchange to occur. Seemingly, an exception was the reaction of MoO3 with
hydrogen forms of zeolites; here, however, the reaction with water vapor was
required to produce small and mobile Mo-containing species, which were able
to easily enter the micropores (cf. Sect. 6.3.2). In fact, there are quite a number of
examples that suggest that SSIE proceeds well without any assistance from
water. Thus, in Sects. 5.3.2.2 and 5.3.3.3 we saw that salts insoluble in water such
as AgCl and HgCl2 reacted (under high vacuum and at elevated temperatures)
with H-ZSM-5. Similarly, Kucherov and Slinkin [21, 23, 92] reported on solid-
state ion exchange with oxides such as CuO and V2O5 , which are also insoluble
in water. Many experiments for SSIE were carried out between hydrogen forms
of zeolites and chloride in dynamic high vacuum (p£10–5 Pa). Examples were
182 H.G. Karge · H.K. Beyer

discussed, for instance, in Sects. 5.1, 5.2 and 5.3 with respect to the exchange with
MICl (MI = Li, Na, K, Rb, Cs), MIICl2 ◊ xH2O (MII = Mg, Ca, Cu, Fe) and MIII ◊ xH2O
(MIII = La, Fe). In these cases one would expect that, at the elevated temperatures
of solid-state reaction (T ≥ 670 K), water is removed from the reactants. How-
ever, it cannot be completely excluded that traces of residual water persist in the
salt/zeolite mixtures or, in the case of salts with crystal water, partial hydrolysis
of the chlorides occurs.
In this context it is, therefore, worth mentioning that a few investigations have
been reported in which special measures were taken to carefully exclude water
from the very initial steps of the SSIE experiment and then throughout the
whole subsequent procedure: The reactants were separately evacuated and heat-
ed at about 670 K in ampoules until water vapor was no longer detected by MS
and/or IR. The ampoules were then sealed and transferred into an efficiently
working glove box (pH2O £ 10–7 Pa).All the subsequent steps of the sample prepa-
ration were then carried out in this glove box, i.e., breaking the ampoules; mix-
ing the salts and zeolite powders; filling the mixtures into capillaries for XRD
runs and sealing them; pressing wafers for IR measurements; transferring the
wafers into sample holders and these into an ultra-high vacuum-tight IR cell (cf.
[286]). One example was the solid-state reaction of a mixture of NaCl and H-
MOR (cf. [130]). After heating the IR cell, which was connected to an ultra-high

Fig. 71. IR spectra of the OH stretching frequency range of a the parent zeolite, H-L (NH4-L
heated at 500 °C in high vacuum), and a mixture of LaCl3 (water-free) and H-L after evacua-
tion (10–5 – 10–6 Pa) at b 300; c 400; and d 525 °C (after [287], with permission)
Solid-State Ion Exchange in Microporous and Mesoporous Materials 183

Fig. 72. IR spectra of the OH stretching frequency range of a the parent zeolite, H-MOR (NH4-
MOR heated at 500 °C in high vacuum), and a mixture of LaCl3 (water-free) and H-MOR after
evacuation (10–5–10–6 Pa) at b 300; c 400; and d 525 °C (after [287], with permission)

vacuum device, IR spectroscopy revealed that the OH groups of the H-MOR

component had completely disappeared. This was ascribed to a 100% exchange
of the protons of H-MOR for Na+ from NaCl. Another set of experiments was
reported for SSIE with lanthanum chloride under complete exclusion of H2O
(cf. [287]). Here, water-free LaCl3 was prepared from LaCl3 ◊ 7H2O by heating the
salt in a flow of HCl to avoid any hydrolysis and the formation of lanthanum oxy-
chlorides. Subsequently, this water-free LaCl3 was sealed under high vacuum in
an ampoule, transferred to a glove box and further treated as indicated above.
Mixtures of water-free LaCl3 with H-L, H-MOR, ultrastabilized (dealuminated)
H-Y, i.e., H-S-Y, or ferrierite (FER) were employed. The IR measurements pro-
vided clear evidence that the intensities of the bands indicating the reactive OH
groups had significantly decreased, i.e., SSIE has occurred in the case of H-L,
H-MOR and H-SY under these conditions of absolute exclusion of water vapor
(cf. Figs. 71 and 72).
The acidic OH groups of H-L (nSi/nAl = 3.18) completely disappeared and
were markedly decreased in the case of H-MOR (nSi/nAl = 6.78). The fact that
with H-MOR not all of the OH groups were replaced by La3+ is easily understood
if one takes into account the high nSi/nAl ratio: similar to H-ZSM-5 (cf. Sect.
5.2.4), the distance between the acid OH groups is larger than in regular fauja-
site-type H-Y (nSi/nAl = 2.5) or H-L and it is, therefore, more difficult to com-
184 H.G. Karge · H.K. Beyer

Fig. 73. IR spectra of the OH stretching frequency range of a the parent zeolite, stabilized
H-Y (H-S-Y), and a mixture of LaCl3 (water-free) and stabilized H-Y (H-S-Y) after evacuation
(10–5 – 10–6 Pa) at b 400; c 500; and d 600 °C (after [287], with permission)

pensate the negative charge of three of those sites by one La3+ cation. For the
same reason also with ultrastabilized faujasite-type zeolite (H-S-Y, nSi/nAl = 8.9)
only a partial solid-state ion exchange was possible. However, this zeolite con-
tained a large amount of silanol groups that also reacted (cf. Fig. 73).
In contrast, water-free LaCl3 did not exchange with H-ferrierite at all, i.e., no
decrease in the OH bands was observed indicating the failure of SSIE. The latter
result will be discussed in Sect. 7.2.
By a comparative experiment similar to that reported in Sect. 5.2.5, the incor-
poration of La3+ cations from water-free LaCl3 into H-L, H-MOR and H-S-Y was
qualitatively confirmed by the changes in the intensities of the correspon-
ding framework reflections in the XRD patterns. These were obtained in situ
under high vacuum in a heatable XRD chamber [287]. No changes in the XRD
reflections of the framework of ferrierite were observed upon calcination of a
LaCl3/H-FER mixture which confirms the IR results reported above.

7.2
Possible Mechanisms of SSIE

In solid-state ion exchange, crystallites of salts or oxides are the sources of the
in-going cations and must be brought into intimate contact with the zeolite crys-
Solid-State Ion Exchange in Microporous and Mesoporous Materials 185

tallites into which the cations should be exchanged. In Sect. 6.1 it was already
mentioned that the transport of the involved species from the salt or oxide crys-
tallites to the zeolite surface may occur through the vapor phase or via surface
diffusion. In most cases the latter possibility appears more likely, since many of
the salts and oxides employed in SSIE have, even at the temperatures of solid-
state ion exchange of about 530–730 K, a rather low vapor pressure.
Another important question arises with respect to the nature of the species
separated from the salt or oxide crystallites and subsequently diffusing into
the (adsorbate-free) channels and cavities of the zeolite crystallites, viz., as to
whether these species are molecules or ions. Usually, the effort required to sep-
arate a molecule from the kink of the surface of a salt or oxide crystal (i.e., from
a so-called “Halbkristall-Lage”, i.e., the position of the “half-crystal”; cf. [288])
is lower than that necessary to remove a cation and an anion in sequence. This
was first computed by Stranski [288] for the case of a sodium chloride crystal in
contact with its diluted vapor. A related question is, whether the species migrate
from the external surface of the zeolite crystallites to the interior of the structure
as molecules or whether cations and ions travel separately? With respect to this
question, mainly two mechanistic models seem to be conceivable: (A) Cations
and anions of the salt or oxide migrate simultaneously (most likely as mole-
cules) reacting in the interior of the zeolite, in that the original cations are
replaced by the in-going ones, combine with anions of the salt or oxide and leave
the structure together with them. (B) Cations stemming from the salt or oxide
migrate into the pores via hopping from site to site to replace cations of the zeo-
lite which have to move in the opposite direction and combine (possibly outside
the zeolite structure) with anions of the salt or oxide. This counter-diffusion of
the in-going and out-going cations must proceed in such a way that no excessive
electrical gradients occur, i.e., the charge balance must be sustained. Both main
models are schematically represented in Fig. 74.

Fig. 74. Schematic representation of two possible models of the mechanism of solid-state ion
exchange in microporous materials. Mechanism A, (top): NaCl molecule diffuses. Mechanism
B, (bottom): Na+ and H+ counter-diffuse (see text)
186 H.G. Karge · H.K. Beyer

Indeed, modifications of both models may be visualized. Models A and B are

difficult to discriminate experimentally. However, there were some indications
that at least in several cases model A applied. When, for instance, a salt with a
bulky anion such as Cs3 [PW12O40] was reacted with H-ZSM-5, solid-state ion
exchange was suppressed to a large extent. This was ascribed to the fact that the
large [PW12O40]3– anion prevented the salt molecule from entering the zeolite
channels [43]. (In fact, solid-state reaction did occur to a minor extent because
of partial decomposition of Cs3 [PW12O40].) Similarly, the failure of ion exchange
between LiX and NaX in the absence of adsorbed water (vide supra, Sect. 6.2)
could be interpreted as a sign that the cations were unable to separate from the
anionic framework of the zeolite (visualized as a bulky polymeric anion) and
thus could not migrate. Solid-state ion exchange with a highly covalent com-
pound such as CuCl suggested that most likely intact CuCl molecules migrated
into the zeolite structure (cf. Sects. 5.3.2.1 and 8). Finally, we have seen that
water-free LaCl3 could be easily introduced into water-free zeolites such as large-
pore H-L, H-MOR and H-S-Y, but an analogous experiment with H-ferrierite
was unsuccessful (Sect. 7.1). In fact, La3+ cations, having a diameter of 0.208 nm
(after Goldschmidt) or 0.230 nm (after Pauling), should be small enough to enter
the structure through its 10-membered and even through its 8-membered rings
with diameters of 0.42¥0.54 nm and 0.35¥0.48 nm [180], respectively. Thus, the
failure of SSIE in the system LaCl3/H-FER supports the assumption that the in-
going species are salt molecules rather than separated ions, in that the LaCl3
molecule is obviously too bulky to enter the ferrierite pores [287].

8
Kinetics of SSIE
In general, only qualitative observations have been reported with respect to the
kinetics of solid-state ion exchange. Thus, it was frequently recognized that SSIE
was initially fast and then its rate levelled off. Increasing the amount of the salt
or oxide in the mixtures with the zeolite powders usually resulted in an increase
in the rate of exchange, possibly due to an enhanced concentration gradient of
the in-going cation. Similarly, the exchange was reported to accelerate when the
temperature of the solid-state ion exchange was raised.
However, systematic investigations of the kinetics of SSIE carried out to date
are rather scarce. In principle, kinetics of SSIE could be determined through in
situ measurements of, e.g., the time-resolved changes in the intensities of XRD
reflections in the pattern of the salt (or oxide)/zeolite mixtures upon heat-treat-
ment (cf. Sects. 5.2.5 and 5.3.2) or of typical IR bands. The IR method may use
the signals of lattice vibrations (cf. [60–63]), bands in the OH or NH stretching
region (in the case of H- or NH4-forms of zeolites), or IR bands characteristic of
interactions between the in-going and/or out-going cations and probe mole-
cules. With respect to the latter method, however, a tacit assumption is made,
viz., that the presence of the probe molecules does not affect the kinetics of sol-
id-state ion exchange.
IR spectroscopy using pyridine as a probe was employed in the investigation
of SSIE of CuCl with Na-Y and Na-MOR (cf. [289, 290]): The respective experi-
Solid-State Ion Exchange in Microporous and Mesoporous Materials 187

Fig.75. IR spectra of pyridine adsorbed on a CuCl/Na-Y mixture at 533 K as a function of time

(for details see text; after [289], with permission)

ments were conducted in a cell, where in the upper part the CuCl/Na-zeolite
wafer could be dehydrated at 390–425 K, while the lower compartment with IR-
transmittant CaF2 windows was brought to the reaction temperature, Treact (cf.
[286, 291]). At zero time, the sample was moved from the upper part of the cell
into the lower compartment preheated to the reaction temperature, and simul-
taneously the probe (pyridine) was admitted. Figure 75 displays a set of selected
spectra for the system CuCl/Na-Y run during the solid-state reaction.
Initially, only the Na+ ions were indicated by bands at 1592 and 1442 cm–1
originating from Py Æ Na+ complexes. However, when the temperature in the
lower compartment was above ca. 450 K, at first shoulders and, after a period of
time, bands at 1604 and 1451 cm–1 developed. These were indicative of Cu+ (Py
Æ Cu+) populating a fraction of the cation sites in the Y-zeolite structure where
they had replaced the Na+ cations. As a consequence, the intensities of the IR
bands typical of Py Æ Na+ concomitantly decreased until a steady state was
established. At temperatures above 670 K, however, the changes in the band
intensities were reversed, i.e., the bands due to Py Æ Cu+ were weakened and
those indicative of Py Æ Na+ increased. This was ascribed to a shift of the
exchange equilibrium [cf. Eq. (43)], i.e., to a remigration of Na+ to and removal
of Cu+ from the cation sites (cf. Scheme 1b; Sects. 5.1.10, 5.2.5 and 5.2.6 with
respect to the systems BeCl2/Na-Y and LaCl3/Na-Y):
CuCl + Na-Y ¤ NaCl + Cu-Y (43)
188 H.G. Karge · H.K. Beyer

Fig. 76. Example of a correction of the reaction isotherms (at Treact = 493 K) for the solid-state
reaction of CuCl/Na-Y as monitored by the IR spectra of adsorbed pyridine; the correction
accounts for the temperature dependence of the pyridine adsorption (for details, see text; after
[289], with permission)

Even though the IR bands overlapped, it was possible to determine the proper
integrated absorbances after appropriate decomposition of the spectra. This was
achieved via fitting the spectra by mixed Gaussian-Lorentzian functions [290].
To obtain curves describing the exchange kinetics, the integrated absorbances
had to be plotted as a function of the reaction time. However, one had to be
aware of the fact that, for a given exchange temperature, the absorbances, At , of
the bands not only depended on the amount of Cu+ introduced and Na+ replaced
at a given time, t, but also on the temperature-dependent adsorption equilibri-
um. Thus, in order to compare the rates of uptake of Cu+ and replacement of Na+
for different temperatures, the influence of the adsorption equilibrium of pyri-
dine on the band intensities had to be taken into account. Since the effect of
minor changes of the pyridine pressure during an SSIE experiment (from, e.g.,
500 to 400 Pa) turned out to be negligible, the necessary correction could be
achieved with the help of an experimentally determined adsorption isobar of
pyridine. As adsorbents, CuCl/Na-Y and CuCl/Na-MOR wafers were used which
had been previously heat-treated at 533 K until the steady state of exchange was
reached. With the isobars obtained in this way, it was possible to relate all the
absorbance data measured at lower temperature to the adsorption equilibrium
at 533 K; an example of such a correction is illustrated in Fig. 76. The curves with
Solid-State Ion Exchange in Microporous and Mesoporous Materials 189

Fig. 77. Normalized and corrected integrated absorbances from pyridine adsorption on
CuCl/Na-Y during solid-state ion exchange as a function of reaction time (for details, see text;
after [289], with permission)

the open symbols are plots of ion exchange kinetics (absorbances vs. reaction
time) providing the corrected data that would have been obtained after
exchange at 493 K but under the adsorption equilibrium of pyridine at 533 K.
From Fig. 76 the correction necessary because of the temperature dependence
of the adsorption equilibrium of the probe is obvious. For the correlation tem-
perature of 533 K, the correction would be zero, i.e., the respective curves would
coincide. The above correction procedure enabled a comparison of all measure-
ments at Treact £ 533 K. Results for the system CuCl/Na-Y are shown in Fig. 77.
The data were normalized to equal sample thickness (5 mg cm–2).
The effect of the reaction temperature on the rates of the Cu+ introduction
and Na+ replacement alone, i.e., after removal of the temperature effect on the
adsorption equilibrium, can be readily recognized. As expected, the rates
increase with increasing temperature, whereas the final steady state of SSIE was
independent of the temperature. For the temperature of 453 K this was con-
firmed by extending the reaction time to 25 h. Similar results to those obtained
for CuCl/Na-Y were obtained for the system CuCl/Na-MOR [289, 290] and for
zeolites containing K+, Rb+ or Cs+ as (out-going) cations [292]. It was tentative-
ly assumed that the exchange kinetics were diffusion-controlled.
The reasonable assumption was made that the absorbances of the Py Æ Cu+
bands at a given time, t, and at steady state, i.e., At and At Æ • , are proportional to
190 H.G. Karge · H.K. Beyer

Fig. 78. Description of the kinetics of solid-state ion exchange in the system CuCl/Na-Y
through a diffusion model; the symbols represent experimental data derived from the mea-
sured integrated absorbances of the probe (pyridine), the broken lines represent results of the
fitting to the diffusion model (for details, see text; after [289], with permission)

Fig. 79. Arrhenius plot of the diffusion coefficients evaluated from the description of the
kinetics of solid-state ion exchange in the systems CuCl/Na-Y and CuCl/Na-M through a dif-
fusion model (for details, see text; after [289], with permission)
Solid-State Ion Exchange in Microporous and Mesoporous Materials 191

the amounts of incorporated Cu+, i.e., Mt and Mt = • . Then, an attempt was made
to describe the kinetics curves (cf., e.g., Fig. 77) by solutions of Fick’s second law
(cf. [290]). Appropriate solutions were provided by Crank [293] for sphere-like
(Na-Y) and membrane-like (Na-MOR) adsorbent particles. Indeed, it turned out
that such a description is possible (cf. Fig. 78) which, however, does not neces-
sarily mean that SSIE is in fact a process controlled by Fickian diffusion.
Figure 79 shows an Arrhenius plot of the thus-determined values of ln
(Dt0 /R2) vs. 1/T, where D, t0 , R, T represent the diffusion coefficient, selected time
after beginning of the exchange process, particle radius and the exchange tem-
perature, respectively. From the slopes of the straight lines activation energies of
EA ~ 70 kJ mol–1 were derived.
Even under the assumption that the process is properly described by the dif-
fusion model described above, the magnitude of EA unfortunately did not pro-
vide unambiguous support for the above proposal that molecules (CuCl) rather
than cations (Cu+) are the diffusing species, since the activation energies for
cation diffusion were found to be of about the same magnitude [294, 295].

9
Conluding Remarks
As we have seen, a great variety of zeolites and related materials can be modified
via solid-state reactions with a similarly broad variety of compounds, i.e., salts
or oxides of the desired in-going cations. Solid-state modifications occur most
easily when halides (sometimes nitrates) of the cations and hydrogen forms of
the materials to be modified are employed. Often a 100% degree of exchange can
be achieved in one step. In several cases, however, sodium forms and complex
cations may also be used. Also, an extended arsenal of techniques is now avail-
able for monitoring and quantitative analysis of solid-state ion exchange. Thus,
solid-state reactions of microporous (and mesoporous) materials have become,
during the past decade, a well-established method for their post-synthesis mod-
ification. Furthermore, related methods such as oxidative or reductive incorpo-
ration of cations into microporous solids through solid-state reactions have
been developed. Similar modifications of the procedure of cation introduction
into zeolites as well as the extension to other systems are likely to come. How-
ever, a number of open questions remain to be answered, concerning a deeper
understanding of solid-state modifications of zeolites and related materials;
pertinent problems are, for instance, the thermodynamics, kinetics and mecha-
nisms.
192 H.G. Karge · H.K. Beyer

References
1. Breck DW (1974) Zeolite molecular sieves – structure, chemistry, and use. John Wiley
and Sons, New York, especially p 5
2. Dyer A (1988) An introduction to zeolite molecular sieves. John Wiley and Sons, New
York, especially p 13
3. Meier WM (1986) In: Murakami Y, Iijima A, Ward JW (eds) New developments in zeolite
science and technology. Proc 7th Int Zeolite Conference. Kodansha, Tokyo, Elsevier,
Amsterdam, p 13
4. Flanigen EM (1986) In: Murakami Y, Iijima A, Ward JW (eds) New developments in zeo-
lite science and technology. Proc 7th Int Zeolite Conference. Kodansha, Tokyo, Elsevier,
Amsterdam, p 103
5. Szostak R (1998) In: Karge HG, Weitkamp J (eds) Molecular sieves – science and tech-
nology, vol 1. Synthesis. Springer, Berlin Heidelberg New York, p 157
6. Sulikowski B (1996) Heterogeneous Chem Rev 3:203
7. Ziołek M, Sobczak I, Decyk P, Nowak I (1999) In: Kiricsi I, Pál-Borbély G, Nagy JB, Karge
HG (eds) Porous materials in environmentally friendly processes. Proc 1st Int FEZA Con-
ference, Eger, Hungary, Sept 1–4, Elsevier, Amsterdam, p 633. Stud Surf Sci Catal 125:633
8. Uytterhoven JB, Christner LC, Hall WK (1965) J Phys Chem 69:2117
9. Guth J-L, Kessler H (1999) In: Weitkamp J, Puppe L (eds) Catalysis and zeolites – funda-
mentals and applications. Springer, Berlin Heidelberg New York, p 1
10. Robson H (ed) (1998) Verified synthesis of zeolitic materials Elsevier, Amsterdam, 1998.
Microporous Mesoporous Mater 22:495
11. Townsend RP (1991) In:Van Bekkum H, Flanigen EM, Jansen JC (eds) Introduction to zeo-
lite science and practice. Elsevier, Amsterdam, chap 10, p 359. Stud Surf Sci Catal 58:359
12. Townsend RP (2001) In: Karge HG,Weitkamp J (eds) Molecular sieves – science and tech-
nology, vol 3. Modification. Springer, Berlin Heidelberg New York, p 1
13. Borgstedt EvR, Sherry HS, Slobogin JP (1997) In: Chon H, Ihm S-K, Uh YS (eds) Progress
in zeolite and microporous materials. Proc 11th Int Zeolite Conference, Seoul, Korea,Aug
12–17, 1996. Elsevier, Amsterdam, p 1659. Stud Surf Sci 105:1659
14. Adams CJ, Araya A, Carr SW, Chapple AP, Franklin KR, Graham P, Minihan AR, Osinga
TJ, Stuart JA (1997) In: Chon H, Ihm S-K, Uh YS (eds) Progress in zeolite and micro-
porous materials. Proc 11th Int Zeolite Conference, Seoul, Korea, Aug 12–17, 1996. Else-
vier, Amsterdam, p 1667. Stud Surf Sci 105:1667
15. Zeolite as catalysts, sorbents and detergent builders – applications and innovations
(1989) Karge HG, Weitkamp J (eds) Proc Int Symposium, Würzburg, Germany, Sept 4–8,
1988. Elsevier, Amsterdam, chap III. Stud Surf Sci Catal 46:645
16. Rabo JA, Poutsma ML, Skeels GW (1973) In: Hightower JW (ed) Proc 5th Int Congress on
Catalysis, Miami Beach, FL, USA, Aug 20–26, 1972, North-Holland Publishing Co., New
York, p 1353
17. Rabo JA, Kasai PH (1975) Prog Solid State Chem 9:1
18. Clearfield A, Saldarriaga CH, Buckley RC (1973) In: Uytterhoven JB (ed) Proc 3rd Int
Conference on Molecular Sieves – Recent Research Reports, Zürich, Switzerland, Sep-
tember 3–7, 1973. University of Leuwen Press, Leuwen, Belgium, p 241
19. Kokotailo GT, Lawton SL, Sawruk S (1977) In: Katzer J (ed) Proc 4th Int Conference on
Molecular Sieves, Chicago, IL, USA, April 18–22, 1977, Am Chem Soc, Washington DC,
p 439. ACS Symp Series 40
20. Fyfe CA, Kokotailo GT, Graham JD, Browning C, Gobbi GC, Hyland M, Kennedy GJ,
DeSchutter CT (1986) J Am Chem Soc 108:522
21. Kucherov AV, Slinkin AA (1986) Zeolites 6:175
22. Beyer HK, Karge HG, Borbély G (1988) Zeolites 8:79
23. Kucherov AV, Slinkin AA (1994) J Mol Catal 90:323
24. Karge HG, Beyer HK (1991) In: Jacobs PA, Jaeger NI, Kubelkova L, Wichterlova B (eds)
Zeolite Chemistry and Catalysis. Proc Int Symposium, Prague, Czechoslovakia, Sept
8–13, 1991, Elsevier, Amsterdam, p 43. Stud Surf Sci Catal 69:43
Solid-State Ion Exchange in Microporous and Mesoporous Materials 193

25. Karge HG (1997) In: Chon H, Ihm S-K, Uh YS (eds) Progress in Zeolite and Microporous
Materials. Proc 11th Int Zeolite Conference, Seoul, Korea, August 12–17, 1996. Elsevier,
Amsterdam, 1997, p 1901. Stud Surf Sci Catalysis 105C: 1901
26. Kosanovic C, Bronic J, Subotic B, Smit I, Stubicar M, Tonejc A, Yamamoto T (1993) Zeo-
lites 13:261
27. Kosanovic C, Cizmek A, Subotic B, Smit I, Stubicar M, Tonejc A (1995) Zeolites 15:51
28. Crocker M, Herold RHM, Emeis CA, Krijger M (1992) Catal Lett 15:339
29. Karge HG, Dondur V (1990) J Phys Chem 94:765
30. Karge HG (1968) Z Phys Chem Neue Folge 76:133
31. Karge HG, Hunger M, Beyer HK (1999) In: Weitkamp J, Puppe L (eds) Catalysis and zeo-
lites – fundamentals and applications. Springer, Berlin Heidelberg New York, chap 4, p 198
32. Ward JW (1968) J Colloid Interface Sci 28:269
33. Huang Y, Paroli RM, Delgado AH, Richardson TA (1998) Spectrochim Acta A 54:1347
34. Karge HG, Lange J-P, Gutsze A, Łaniecki M (1988) J Catal 114:144
35. Karge HG, Pál-Borbély G, Beyer, HK (1994) Zeolites 14:512
36. Weitkamp J, Ernst S, Hunger M, Röser T, Huber S, Schubert UA, Thomasson P,
Knözinger H (1998) In: Hightower JW, Delgass WN, Iglesia E, Bell AT (eds) Proc 11th
Int Congress on Catalysis, Baltimore, MA, USA, Elsevier, Amsterdam, p 731. Stud Surf
Sci Catal 101: 731
37. Ernst H, Freude D, Mildner T, Wolf I (1995) In: Beyer HK, Karge HG, Kiricsi I, Nagy JB
(eds) Catalysis by Microporous Materials. Proc ZEOCAT ’95, Szombathely, Hungary, July
9–13, 1995, Elsevier, Amsterdam, p 413. Stud Surf Sci Catalysis 94:413
38. Hunger M, Seiler M, Horvath T (1999) Catal Lett 57:199
39. Hunger M, Horvath T (1997) J Catal 167:187
40. Hatje U, Ressler T, Petersen S, Förster H (1994) In: Proc European Symposium, Frontiers
in Science and Technology with Synchrotron Radiation. J de Physique IV, supplément au
J de Physique III, p 141
41. Karge HG, Beyer HK, Borbély G (1988) Catal Today 3:41
42. Schöllner R, Nötzel P, Herden H, Körner G (1978) In: Fejes P (ed) Application of Zeolites
in Heterogeneous Catalysis and Related Fields. Proc Symposium on Zeolites, Szeged,
Hungary, Sept 11–14, 1978. Acta Universitas Szegediensis, Acta Pysica et Chemica, Nova
Series, 24:293
43. Karge HG (1994) In: Hattori T, Yashima T (eds) Zeolites and Microporous Crystals, Proc
Int Symposium, Nagoya, Japan, Aug 22–25, 1993, Kodansha, Tokyo, Elsevier, Amsterdam,
p 135. Stud Surf Sci Catal 83:135
44. Bock T (1995) Dissertation (PhD thesis), University of Stuttgart
45. Weitkamp J, Ernst S, Bock T, Kiss A, Kleinschmit P (1995) In: Beyer HK, Karge HG, Kiric-
si I, Nagy JB (eds) Catalysis by Microporous Materials. Proc ZEOCAT ’95, Szombathely,
Hungary, July 9–13, 1995, Elsevier, Amsterdam, p 278. Stud Surf Sci Catalysis 94
46. Beran S, Wichterlova B, Karge HG (1990) J Chem Soc Faraday Trans 86:3033
47. Jia Ch, Massiani P, Beanier P, Barthomeuf D (1993) Appl Catal A 106:L 185
48. Hari Prasad Rao PR, Massiani P, Barthomeuf D (1994) In: Weitkamp J, Karge HG, Pfeifer
H, Hölderich W (eds) Zeolites and Related Microporous Materials: State of the Art 1994.
Proc 10th Int Zeolite Conference, Garmisch-Partenkirchen, Germany, July 17–22, 1994,
Elsevier, Amsterdam, p 1449. Stud Surf Sci Catal 84:1449
49. Mavrodinova VP (1988) Microporous Mesoporous Mater 24:1
50. Xu J, Yan A, Xu Q (1998) Wuji Huaxue Xuebao 14:422
51. Mavrodinova VP (1988) Microporous Mesoporous Mater 24:9
52. Ogawa M, Handa T, Kuroda K, Kato C (1990) Chem Lett 71
53. Rabo JA (1976) Salt occlusion in zeolite crystals. In: Rabo JA (ed) Zeolite chemistry and
catalysis. Am Chem Soc, Washington DC, USA, chap 5, p 332, ACS Monograph 171
54. Karge HG, Hunger M, Beyer HK (1999) In: Weitkamp J, Puppe L (eds) Catalysis and zeo-
lites – fundamentals and applications. Springer, Berlin Heidelberg New York, p 296
55. Engelhardt G, Michel D (1987) High-resolution solid-state NMR of silicates and zeolites.
John Wiley and Sons, New York, p 212
194 H.G. Karge · H.K. Beyer

56. Jiang M, Tatsumi T (1998) J Phys Chem B 102:10879

57. Sauer J (1994) In: Weitkamp J, Karge HG, Pfeifer H, Hölderich W (eds) Zeolites and Relat-
ed Microporous Materials: State of the Art 1994. Proc 10th Int Zeolite Conference,
Garmisch-Partenkirchen, Germany, July 17–22, 1994, Elsevier, Amsterdam, p 2039. Stud
Surf Sci Catal 84:2039
58. Karge HG, Pál-Borbély G, Popovic E, unpublished results
59. Jiang M, Ye S, Shan S, Fu Y (1997) Wuli Huaxue Xuebao 10:377
60. Esemann E, Förster H (1994) J Chem Soc Chem Commun 1319
61. Esemann E, Förster H (1995) Z Phys Chem 189:263
62. Esemann E, Förster H, Geidel E, Krause K (1996) Microporous Mesoporous Mater 6:321
63. Geidel E (1997) Habilitation Thesis (Habilitationsschrift) University of Hamburg
64. Hirschler A (1963) J Catal 2:428
65. Plank CJ (1965) In: Sachtler WM, Schuit GCA, Zwietering P (eds), Proc 3rd Int Congress
on Catalysis, Amsterdam, The Netherlands, July 20–25, 1964, North-Holland Publishing
Comp, Amsterdam, p 727
66. Li Y, Xie W, Shen Y (1997) Shiyou Huagong 26:87
67. Yang C, Xu Q-H (1997) Huaxue Xuebao 55:562
68. Yang C, Xu Q-H (1996) J Chem Chinese Univ 17:1336
69. Yang C, Xu Q-H (1996) Gaodeng Xuexiao Huaxue Xuebao 17:1336
70. Akolekar DB, Bhargava S (1997) J Mol Catal A Chem 122:81
71. Karge HG (1977) In: Katzer J (ed) Proc 4th Int Conference on Molecular Sieves, Chicago,
Ill, USA, April 18–22, 1977, Am Chem Soc, Washington DC, p 584. ACS Symp Series 40
72. Canizares P, Duran A, Dorado F, Carmona M (2000) Appl Clay Sci 16:273
73. Borbély G, Beyer HK, Radics L, Sandor P, Karge HG (1989) Zeolites 9:428
74. Beyer HK, Pál-Borbély G, Karge HG (1993) Microporous Mater 1:67
75. Hunger M, Schenk U, Burger B, Weitkamp J (1999) Angew Chem Int Ed Engl 36:2504
76. Karge HG, Ladebeck J, Sarbak Z, Hatada K (1982) Zeolites 2:94
77. Blauwhoff PMM, Gosselink JW, Kieffer EP, Sie ST, Stork WHJ (1999) In: Weitkamp J,
Puppe L (eds) Catalysis and zeolites – fundamentals and applications. Springer, Berlin
Heidelberg New York, chap 7, p 437
78. Karge HG, Borbély G, Beyer HK, Onyestyak G (1988) In: Philips MJ, Ternan M (eds) Proc
9th Int Congress on Catalysis, Calgary, Alberta, Canada, June 26 – July 1, 1988, Chemical
Institute of Canada, Ottawa, p 396
79. Karge HG, Beyer HK (1991) In: C1-Chemie-Angewandte Heterogene Katalyse-C4-
Chemie, DGMK-Berichte-Tagungsbericht 9101, DGMK – Fachbereichstagung, Leipzig,
FRG, Februar 20–22, 1991, ISBN 3-928164-07-4, p 191
80. Jia Ch, Beaunier P, Massiani P (1998) Microporous Mesoporous Mater 24:69
81. Esemann H, Förster H (1995) Z Phys Chem (München) 189:263
82. Difallah M, Ginoux JL (1998) J Soc Alger Chim 8:21
83. Pires EM, Wallau M, Schuchardt U (1998) J Mol Catal A Chem 136:69
84. Van Kooten WEJ, Liang B, Krijsnen HC, Oudshoorn OL, Calis HPA, van den Bleck CM
(1999) Appl Catal B 21:203
85. Nassar EJ, Serra OA, Souza-Aguiar EF (1998) Quim Nova 21:121
86. Welsh B, Lambert SL (1988) In: Flank WH, Whyte TE Jr (eds) Perspectives in molecular
sieve science. American Chemical Society, Washington DC, p 33
87. Challoner R, Harris RK (1991) Zeolites 11:265
88. Hunger M, Engelhardt G, Weitkamp J (1994) In: Weitkamp J, Karge HG, Pfeifer H,
Hölderich W (eds) Zeolites and Related Microporous Materials: State of the Art 1994.
Proc 10th Int Zeolite Conference, Garmisch-Partenkirchen, Germany, July 17–22, 1994,
Elsevier, Amsterdam, p 725. Stud Surf Sci Catal 84A:725
89. Streitwieser A Jr, Reif L (1960) J Am Chem Soc 82:5003
90. Weiss U, Weihe M, Hunger M, Karge HG, Weitkamp J (1997) In: Chon H, Ihm SK,
Uh YS (eds) Progress in Zeolite and Microporous Materials. Proc 11th Int Conf Zeoli-
tes, Seoul, Korea, August 12–17, 1996, Elsevier, Amsterdam, p 973. Stud Surf Sci Catal
105: 973
Solid-State Ion Exchange in Microporous and Mesoporous Materials 195

91. Karge HG, Mavrodinova V, Zheng Z, Beyer H (1991) Appl Catal 75:343
92. Kucherov AV, Slinkin AA (1987) Zeolites 7:38
93. Kucherov AV, Slinkin AA (1987) Zeolites 7:43
94. Krüerke U, Jung P (1968) Z Phys Chem 58:53
95. Kucherov AV, Slinkin AA, Kondrat’ev DA, Bondarenko TN, Rubinstein AM, Minachev KM
(1985) Zeolites 5:320
96. Slinkin AA, Kucherov AV, Chuvylkin ND, Korsunov VA, Kliachko AL, Nikishenko SB
(1989) J Chem Soc Faraday Trans 85:3233
97. Jirka I, Wichterlova B, Maryska M (1991) In: Jacobs PA, Jaeger NI, Kubelkova L,
Wichterlova B (eds) Zeolite Chemistry and Catalysis. Proc Int Symposium, Prague,
Czechoslovakia, September 8–13, 1991, Elsevier, Amsterdam, p 269. Stud Surf Sci Catal
69:269
98. Karge HG, Wichterlova B, Beyer HK (1992) J Chem Soc Faraday Trans 88:1345
99. Haniffa RM, Seff K (1998) Microporous Mesoporous Mater 25:137
100. Weckhuysen BM, Spooren HJ, Schoonheydt RA (1994) Zeolites 14:450
101. Hartmann M, Boddenberg B (1994) In: Weitkamp J, Karge HG, Pfeifer H, Hölderich W
(eds) Zeolites and Related Microporous Materials: State of the Art 1994. Proc 10th
Int Zeolite Conference, Garmisch-Partenkirchen, Germany, July 17–22, 1994, Elsevier,
Amsterdam, p 509. Stud Surf Sci Catal 84:509
102. Hartmann M, Boddenberg B (1994) Microporous Mater 2:127
103. Borovkov VY, Jiang M, Fu Y (1999) J Phys Chem B 103:5010
104. Esemann H, Förster H (1999) J Mol Structure 483:7
105. Förster H, Hatje U (1997) Solid State Ionics 101/103:425
106. Auroux A, Gervasini A, Guimon C (1999) J Phys Chem B 103:7195
107. Shi Q, Li C (2000) Cuihua Xuebao 21:113
108. Li C, Li D, Shi Q, Zhu Q (1999) Fenzi Cuihua 13:115
109. Astrelin IM, Enhbold T, Sychev M (1999) In: Treacy MMJ (ed) Proc 12th Int Zeolite Conf,
Baltimore, USA, July 12–17, 1998, vol 3, Materials Research Society, Warrendale, Pa,
p 2129
110. Kanazirev VI, Price GL (1995) J Mol Catal A Chem 96:145
111. Kanazirev VI, Price GL (1994) J Catal 148:164
112. King ST (1997) Catal Today 33:173
113. King ST (1996) J Catal 161:530
114. Varga J, Fudala Á, Halász J, Schöbel G, Kiricsi I (1995) In: Beyer HK, Karge HG, Kiricsi I,
Nagy JB (eds) Catalysis by Microporous Materials. Proc ZEOCAT ’95, Szombathely,
Hungary, July 9–13, 1995, Elsevier, Amsterdam, p 665. Stud Surf Sci Catal 94:665
115. Varga J, Halasz J, Kiricsi I (1998) In: Van der Hoek KW (ed) 1st Int Nitrogen Conf, Else-
vier, Oxford, UK, p 691
116. Varga J, Halász J, Kiricsi I (1998) Environment Pollution 102 (Suppl 1): 691
117. Varga J, Nagy JB, Halász J, Kiricsi I (1997) J Mol Struct 410:149
118. Varga J, Halász I, Horvath D, Mehn D, Nagy JB, Schobel G, Kiricsi I (1998) In: Kruse N,
Frennet A, Bastin J-M (eds) Proc 4th Int Symp Catalysis Automotive Pollution Control,
Brussels, Belgium, April 9–11, 1997, Elsevier, Amsterdam, p 367. Stud Surf Sci Catal
116:367
119. Schay Z, Knözinger H, Guczi L, Pál-Borbély G (1998) Appl Catal B 18:263
120. Halász I, Pál-Borbély G, Beyer HK (1997) React Kinet Catal Lett 61:27
121. Varga J, Fudala A, Halász J, Kiricsi I (1996) J Therm Anal 47:391
122. Halász I, Brenner A (1998) Catal Lett 51:23
123. Halász J, Varga J, Schobel G, Kiricsi I, Hernadi K, Hannus I, Varga K, Fejes P (1995) In:
Frennet A, Bastin J-M (eds) 3rd Int Symp Catalysis Automotive Pollution Control, Brus-
sels, Belgium, April 20–22, 1994, Elsevier, Amsterdam, p 675
124. Setzer C, Demuth D, Schüth F (1997) Chem Ing Tech 69:79
125. Poeppel A, Newhouse M, Kevan L (1995) J Phys Chem 99:10019
126. Liese T, Grünert W (1997) J Catal 172:34
127. Price GL, Kanazirev V, Church DF (1995) J Phys Chem 99:864
196 H.G. Karge · H.K. Beyer

128. Kucherov AV, Slinkin AA, Beyer HK, Borbély G (1989) Kinet Katal 30:429 (Engl Trans
30:367)
129. Kucherov AV, Slinkin AA, Beyer HK, Borbély G (1989) J Chem Soc Faraday Trans I85:2737
130. Karge HG, Mavrodinova V, Zheng Z, Beyer HK (1990) In: Barthomeuf D, Derouane EG,
Hölderich W (eds) Guidelines for mastering the properties of molecular sieves – rela-
tionship between the physicochemical properties of zeolitic systems and their low
dimensionality, Plenum Press, New York, p 157. NATO ASI Series B Physics 221: 157
131. Salama TM, Shido T, Ohnishi R, Ichikawa M (1996) J Phys Chem 100:3688
132. Salzer R (1992) In: Proc SPIE – Int Opt Eng 1992. 8th Int Fourier Transform Spectroscopy,
1991, p 50, ISSN 0277-786X
133. Salzer R, Finster U, Roessner F, Steinberg K-H, Klaeboe P (1992) Analyst 117:351
134. Roessner F, Hagen A, Mroczek U, Karge HG, Steinberg K-H (1993) In: Guczi L, Solymosi
F, Tétényi (eds) Proc Int Congress on Catalysis, Budapest, Hungary, July 19–24, 1992,
Elsevier, Amsterdam, p 1707
135. Hagen A, Roessner F (1994) In: Hattori T, Yashima T (eds) Zeolites and Microporous
Crystals.Proc Int Symp, Nagoya, Japan, August 22–25, 1993, Elsevier, Amsterdam, p 313.
Stud Surf Sci Catal 83:313
136. Seidel A, Rittner F, Boddenberg B (1996) J Chem Soc Faraday Trans 92:493
137. Boddenberg B, Seidel A (1994) J Chem Soc Faraday Trans 90:1345
138. Seidel A, Kampf G, Schmidt A, Boddenberg B (1998) Catal Lett 51:213
139. Onyestyák G, Kalló D, Papp J Jr (1991) In: Jacobs PA, Jaeger NI, Kubelkova L, Wichterlova
B (eds) Zeolite Chemistry and Catalysis. Proc Int Symposium, Prague, Czechoslovakia,
September 8–13, 1991, Elsevier, Amsterdam, p 287. Stud Surf Sci Catal 69: 287
140. Beyer HK, Pál-Borbély G, Keindl M (1999) Microporous Mesoporous Mater 31:333
141. Jacobs PA, Uytterhoeven JB, Beyer HK (1977) J Chem Soc Faraday Trans I 73:1755
142. Sárkány J, Sachtler WMH (1994) Zeolites 14:7
143. Rojasova E, Smieskova A, Hudec P, Zidek Z (1999) Collect Czech Chem Commun 64:168
144. Yin D, Yin D, Li Q, Fu Z (2000) Cuihua Xuebo 21:113
145. Yin D, Yin D (1998) Microporous Mesoporous Mater 24:123
146. Yin D, Yin D, Fu Z, Li Q (1999) Cuihua Xuebao 20:419
147. Wichterlova B, Beran S, Bednárová S, Nedomová K, Dudíková, Jíru P (1988) In: Grobet PJ,
Mortier WJ, Vansant EF, Schulz-Ekloff G (eds) Innovation in Zeolite Materials Science.
Proc Int Symposium, Nieuwpoort, Belgium, September 13–17, 1987, Elsevier, Amster-
dam, p 199. Stud Surf Sci Catal 37: 199
148. Kucherov AV, Slinkin AA (1988) Zeolites 8:110
149. Lázár K, Pál-Borbély G, Beyer HK, Karge HG (1994) J Chem Soc Faraday Trans 90:1329
150. Lázár K, Pál-Borbély G, Beyer HK, Karge HG (1995) In: Poncelet G, Martens J, Delmon B,
Jacobs PA, Grange P (eds) Preparation of catalysts VI. Scientific bases for the preparation
of heterogeneous catalysts. Elsevier, Amsterdam, p 51. Stud Surf Sci Catal 91:51
151. Kucherov AV, Slinkin AA (1987) Kinet Katal 28:1199
152. Morice JA, Rees LVC (1968) Trans Faraday Soc 64:1388
153. Garten RL, Delgass WN, Boudart MJ (1970) J Catal 18:90
154. Lázár K (1991) Struct Chem 2:245
155. Lobree LJ, Hwang I-C, Reimer JA, Bell AT (1999) J Catal 186:242
156. Rauscher M, Kesore K, Monnig R, Schwieger W, Tissler A, Turek T (1999) Appl Catal
184:249
157. Kögel M, Monnig R, Schwieger W, Tissler A, Turek T (1999) J Catal 182:470
158. Giles R, Cant NW, Kögel M, Turek T, Trimm DL (2000) Appl Catal B 25:L75
159. Liu IO, Nant NW, Kögel M, Turek T (1999) Catal Lett 63:241
160. Kögel M, Sandoval VH, Schwieger W, Tissler A, Turek T (1998) Catal Lett 51:23
161. Dandl H (1997) PhD thesis, University of Erlangen-Nürnberg
162. El-Malki El-M, Werst D, Doan PE, Sachtler WMH (2000) J Phys Chem B 104:5924
163. Enhbold T, Sychev M, Astrelin IM, Rozwadowski M, Golembiewski R (1988) In: Rozwa-
dowski M (ed) Proc 3rd Polish-German Zeolite Colloquium, Torun, Poland, April 3–5,
1997, Nicholas Copernicus University Press, Torun, p 137
Solid-State Ion Exchange in Microporous and Mesoporous Materials 197

164. Jentys A, Lugstein A, Vinek H (1997) J Chem Soc Faraday Trans 93:4091
165. Jentys A, Lugstein A, Vinek H (1997) Zeolites 18:391
166. Koranyi TI, Pham NH, Jentys A,Vinek H (1997) In: Froment GF, Delmon B, Grange P (eds)
Hydrotreatment and Hydrocracking of Oil Fractions, Proc 1st Int Symp/6th European
Workshop, Oostende, Belgium, February 17–19, 1997, Elsevier, Amsterdam, p 509. Stud
Surf Sci Catal 106:509
167. Jentys A, Lugstein A, Vinek H (1997) Zeolites 18:391
168. Lugstein A, Jentys A, Vinek H (1998) Appl Catal A General 166:29
169. Kinger G, Lugstein A, Swagera R, Ebel M, Jentys A, Vinek H (2000) Microporous Meso-
porous Mater 39: 307
170. Jentys A, Lugstein A, Dusouqui OE, Vinek H, Englisch M, Lercher JA (1996) In: Absi-
Halabi M, Beshara J, Qabazard H, Stanislaus A (eds) Catalysts in Petroleum Refining
and Petrochemical Industries 1995, Proc 2nd Int Conf Catalysts in Refining and Petro-
chemical Industries, Kuweit, April 22–26, 1995, Elsevier, Amsterdam, 1996, p 525. Stud
Surf Sci Catal 100:525
171. Kühl GH (1973) In: Uytterhoven JB (ed) Proc 3rd Int Conference on Molecular Sieves –
Recent Research Reports, Zürich, Switzerland, September 3–7, 1973. University of
Leuwen Press, Leuwen, Belgium, p 227
172. Kühl GH (1977) J Phys Chem Solids 38:1259
173. Li Y, Armor JN (1999) Appl Catal A 188:211
174. Park Y-K, Goryashenko SS, Kim DS, Park S-E (1999) In: Treacy MMJ (ed) Proc 12th Int
Zeolite Conf, Baltimore, USA, July 12–17, 1998, Materials Research Society, Warrendale,
Pa, 1999, vol 2, p 1157
175. Wang X, Chen H-Y, Sachtler WMH (2000) Appl Catal B 26:L227
176. Wichterlová B, Beran S, Kubelková L, Nováková J, Smiesková A, Sebík R (1989) In: Karge
HG, Weitkamp J (eds) Zeolite as catalysts, sorbents and detergent builders – applications
and innovations, Proc Int Symposium, Würzburg, Germany September 4–8, 1988. Else-
vier, Amsterdam, p 347. Stud Surf Sci Catal 46: 347
177. Haniffa RM, Seff K (1998) J Phys Chem 102:2688
178. Djieugoue M-A, Prakash AM, Kevan L (1999) J Phys Chem B 103:804
179. Bock T, Weitkamp J, Karge HG (2002) Microporous Mesoporous Mater, submitted
180. Meier WM, Olson DH, Baerlocher C (1996) Atlas of zeolite structure types, 4th edn. Else-
vier, Amsterdam
181. Azuma N, Hartmann M, Kevan L (1995) J Phys Chem 99:6670
182. Hartmann M, Azuma N, Kevan L (1995) J Phys Chem 99:10988
183. Azuma N, Lee ChW, Zamadics M, Kevan L (1994) In: Weitkamp J, Karge HG, Pfeifer H,
Hölderich W (eds) Zeolites and Related Microporous Materials: State of the Art 1994.
Proc 10th Int Zeolite Conference, Garmisch-Partenkirchen, Germany, July 17–22, 1994,
Elsevier, Amsterdam, p 805. Stud Surf Sci Catal 84:805
184. Hartmann M, Azuma N, Kevan L (1995) In: Bonneviot L, Kaliaguine S (eds) Zeolites: A
Refined Tool for Designing Catalytic Sites. Proc Int Symp, Québec, Canada, October
15–20, 1995, Elsevier, Amsterdam, 1995, p 335. Stud Surf Sci Catal 97:335
185. Hartman M, Kevan L (1997) In: Chon H, Ihm SK, Uh YS (eds) Progress in Zeolite and
Microporous Materials. Proc. 11th Int Conf Zeolites, Seoul, Korea, August 12–17, 1996,
Elsevier, Amsterdam, p 717. Stud Surf Sci Catal 105:717
186. Hartmann M, Pöppl A, Kevan L (1995) J Phys Chem 99:17494
187. Ma A, Muhler M, Grünert W (2000) Appl Catal B 27:37
188. Beran B, Wichterlová B, Karge HG (1990) J Chem Soc Faraday Trans 86:3033
189. Purans J, Kliava J, Millere I (1979) Phys Stat Solidus 56:K25
190. Kucherov AV, Slinkin AA (1986) Kinet Katal 27: 678 (Engl Trans 27:585)
191. Kucherov AV, Slinkin AA (1987) Zeolites 7:583
192. Huang M, Shan S, Yuan C, Li Y, Wang Q (1990) Zeolites 10:772
193. Shan S, Shen S, Chen K, Huang M (1992) Acta Physico-Chimica Sinica (Wuli Haxue
Xuebao) 8:338
194. Shen S, Shan S, Chen K, Huang M (1993) Cuihua Xuebao (Chin J Catal) 14:455
198 H.G. Karge · H.K. Beyer

195. Marchal C, Thoret J, Gruia M, Dorémieux-Morin C, Fraissard J (1990) In: Symp on

Advances in FCC Technology II. Am Chem Soc, Catalysis Division (ed) Washington DC,
USA, August 26–31, 1990, Am Chem Soc, Washington DC, p 762
196. Marchal C, Thoret J, Gruia MG, Dorémieux-Morin C, Fraissard J (1991) In: Ocelli ML (ed)
Fluid catalytic cracking II. Concepts in catalyst design. ACS Symp Series 452, Am Chem
Soc, Washington DC, p 212
197. Thoret J, Marchal C, Dorémieux-Morin, Man PP, Gruia M, Fraissard J (1993) Zeolites
13:269
198. Kucherov AV, Slinkin AA (1986) Kinet Katal 27:909 (Engl Trans 27:789)
199. Weckhuysen BM, Schoonheydt RA (1994) In: Weitkamp J, Karge HG, Pfeifer H, Hölderich
W (eds) Zeolites and Related Microporous Materials: State of the Art 1994. Proc 10th
Int Zeolite Conference, Garmisch-Partenkirchen, Germany, July 17–22, 1994, Elsevier,
Amsterdam, p 965. Stud Surf Sci Catal 84:965
200. Dai P-SE, Lunsford JH (1980) J Catal 64:173
201. Yuan C, Yao J, Chen K, Huang M (1992) Shiyou Huagong (Petrochemical Technology)
21:724
202. Ziołek M, Novak I, Karge HG (1995) In: Beyer HK, Karge HG, Kiricsi I, Nagy JB (eds)
Catalysis by Microporous Materials. Proc ZEOCAT ’95, Szombathely, Hungary, July 9–13,
1995, Elsevier, Amsterdam, p 270. Stud Surf Sci Catal 94: 270
203. Jacobs PA, De Wilde W, Schoonheydt RA, Uytterhoeven JB, Beyer HK (1976) J Chem Soc
Faraday Trans I 72:1221
204. Jacobs PA, Uytterhoeven JB, Beyer HK (1979) J Chem Soc Faraday Trans I 75:56
205. Ignatovych M, Gomenyuk A, Ogenko V, Chuiko O (1999) In: Kiricsi I, Pál-Borbély G, Nagy
JB, Karge HG (eds) Porous Materials in Environmentally Friendly Processes. Proc. 1st Int
FEZA Conference, Eger, Hungary, September 1–4, Elsevier, Amsterdam, p 261. Stud Surf
Sci Catal 125:261
206. Whittington BI, Anderson JR (1993) J Phys Chem 97:1032
207. Hong SB, Hwang BW, Yeom Y, Kim SJ, Uh YS (1991) In: Inui T, Namba S, Tatsumi T (eds)
Chemistry of Microporous Crystals. Proc Int Symp Chemistry of Microporous Crystals,
Tokio, Japan, June 26–29, 1990, p 179. Stud Surf Sci Catal 60:179
208. Narayanan S, Sultana A (1998) Appl Catal A General 167:103
209. Narayanan S, Deshpande K (2000) Appl Catal A General 199:1
210. Thoret J, Man PP, Fraissard J (1997) Zeolites 18:152
211. Hemidy J-F, Delavennant F, Cornet D (1973) J Chemie Physique 70:1716
212. Hemidy J-F, Cornet D (1974) J Chemie Physique 71:739
213. Slinkin AA, Kucherov AV, Goryashenko SS, Aleshin EG, Slovetskaya KI (1989) Kinet Katal
30:184 (Engl Trans 30:158)
214. Slinkin AA, Kucherov AV, Gorjachenko SS, Aleshin EG, Slovetskaja KI (1990) Zeolites
10:111
215. Zhu Z, Wasowicz T, Kevan L (1997) J Phys Chem B 101:10763
216. Wang Q, Sayers D, Huang M, Yuan C, Wei S (1992) Preprint ISSN: 0569-3772. Am Chem
Soc, Div Fuel Chem 37:283
217. Harris M, Dwyer J, Garforth AA, McAteer CH, Ball WJ (1989) In: Karge HG, Weitkamp J
(eds) Zeolite as Catalysts, Sorbents and Detergent Builders – Applications and Innova-
tions. Proc Int Symposium, Würzburg, Germany September 4–8, 1988, Elsevier, Amster-
dam, p 271. Stud Surf Sci Catal 46:271
218. Borry RW, Kim Y-H, Huffsmith A, Reimer JA, Iglesia E (1999) J Phys Chem B 103:5787
219. Kim H-Y, Borry RW III, Iglesia E (2000) Microporous Mesoporous Mater 35–36:495
220. Li W, Borry RW, Meitzner GD, Iglesia E (2000) J Catal 191: 373
221. Zhang Y (1992) Dissertation (PhD thesis), Technical University of Berlin
222. Karge HG, Zhang Y, Beyer HK (1992) Catal Lett 12:147
223. Karge HG, Zhang Y, Beyer HK (1992) In: Guczi K, Solymosi F, Tétényi (eds) New Fron-
tiers in Catalysis. Proc 10th Int Congress on Catalysis, Budapest, Hungary, July 19–24,
Elsevier, Amsterdam; Akadémiai Kiadó, Budapest, 1993, p 257. Stud Surf Sci Catal 75:
257
Solid-State Ion Exchange in Microporous and Mesoporous Materials 199

224. König A, Karge HG, Richter T (1996) German Patent DE 196 37 032 A1 6:25. Assignee:
Volkswagen AG
225. Hatje U, Hagelstein M, Ressler T, Förster H (1995) Physica B 208 & 209:646
226. Mkombe CM, Dry ME, O’Connor CT (1997) Zeolites 19:175
227. Schlegel L, Miessner H, Gutschick D (1994) Catal Lett 23:215
228. Wasowicz T, Kevan L (1998) In: Rudowict CZ, Yu PKN, Hiraoka H (eds) Proc 1st Asia Pac
EPR/ESR Symp Singapore, 1997, Springer, Singapore, p 279
229. Weisz P, Frilette VJ (1960) J Phys Chem 64:382
230. Weisz P, Frilette VJ, Maatmann RW, Mower EB (1992) J Catal 1:307
231. Weitkamp J, Ernst S, Bock T, Kromminga T, Kiss A, Kleinschmit P (1996) US Patent
5,529,964
232. Dessau RM (1982) J Catal 77:304
233. Dessau RM (1984) J Catal 89:520
234. Tzou MS, Jiang HJ, Sachtler WMH (1986) Appl Catal 20:231
235. Tzou MS, Teo BK, Sachtler WMH (1986) Langmuir 2:773
236. Koy J, Karge HG, unpublished work
237. Ostgard DJ, Kustov L, Poeppelmeier KR, Sachtler WMH (1992) J Catal 133:342
238. Poeppelmeier KR, Towbridge TD, Kao J-L (1986) US Patent 4,568,556
239. Che M, Clause O, Marcilly Ch (1997) In: Ertl G, Knözinger H,Weitkamp J (eds) Handbook
of heterogeneous catalysis, VCH, Weinheim, Germany, p 191
240. Kucherov AV, Slinkin AA, Goryashenko SS, Slovetskaja KI (1989) J Catal 118:459
241. Kucherov AV, Kucherova TN, Slinkin AA (1991) Catal Lett 10:289
242. Fraissard J, Gedeon A, Chen Q, Ito T (1991) In: Jacobs PA, Jaeger NI, Kubelkova L, Wichter-
lova B (eds) Zeolite Chemistry and Catalysis. Proc Int Symposium, Prague, Czechoslovakia,
September 8–13, 1991, Elsevier, Amsterdam, p 461. Stud Surf Sci Catal 69:461
243. El-Malki El-M, van Santen RA, Sachtler WMH (1999) J Phys Chem B 103:4611
244. Guisnet M, Gnep NS,Vasques H, Ribeiro FR (1991) In: Jacobs PA, Jaeger NI, Kubelkova L,
Wichterlova B (eds) Zeolite Chemistry and Catalysis. Proc Int Symposium, Prague,
Czechoslovakia, September 8–13, 1991, Elsevier, Amsterdam, p 321. Stud Surf Sci Catal
69:321
245. Liang J, Tang W, Ying M-L, Zhao S-Q, Xu B-Q, Li H-Y (1991) In: Jacobs PA, Jaeger NI,
Kubelkova L, Wichterlova B (eds) Zeolite Chemistry and Catalysis. Proc Int Symposium,
Prague, Czechoslovakia, September 8–13, 1991, Elsevier,Amsterdam, p 207. Stud Surf Sci
Catal 69:207
246. Whittington BI, Anderson JR (1991) J Phys Chem 95:3306
247. Seidel A, Boddenberg B (1996) Chem Phys Lett 249:117
248. Rittner F, Seidel A, Boddenberg B (1997) Microporous Mesoporous Mater 24:127
249. Sprang TH, Seidel A, Wark M, Rittner F, Boddenberg B (1997) J Mater Chem 7:1429
250. Beyer HK, Pál-Borbély G, Keindl M (1999) Microporous Mesoporous Mater 31:333
251. Ozin GA, Gil C (1989) Chem Rev 89:1749
252. Kucherov AV, Kucherova TN, Slinkin AA (1998) Microporous Mesoporous Mater 26:1
253. Beyer HK, Jacobs PA (1977) In: Katzer JR (ed) Molecular Sieves – II. Proc 4th Int Zeolite
Conf, Chicago, USA, April 18–22, 1977, Am Chem Soc, Washington DC, p 493
254. Beyer HK, Jacobs PA, Uytterhoeven JB (1976) J Chem Soc Faraday Trans I 72:674
255. Kucherov AV, Slovetskaya KI, Goryaschenko SS, Aleshin EG, Slinkin AA (1996) Micro-
porous Mater 7:27
256. Feeley OC, Sachtler WHM (1991) Appl Catal 75:93
257. Price LG, Kanazirev V (1990) J Catal 126:267
258. Price GL, Kanazirev VI, Dooley KM (1995) Zeolites 15:725
259. Kanazirev V, Price GL, Dooley KM (1991) In: Jacobs PA, Jaeger NJ, Kubelkova L,
Wichterlova B (eds) Zeolite Chemistry and Catalysis. Proc Int Symp, Prague, Czechoslo-
vakia. Sept 8–13, 1991, Elsevier, Amsterdam, p 277. Stud Surf Sci Catal 69:277
260. Kanazirev V, Neinska Y, Tsoncheva T, Kosova L (1993) In: von Ballmoos R, Higgins JB,
Treacy MMJ (eds) Proc 9th Int Zeolite Conf, Montreal, Canada, July 5–10, 1992, vol I,
Butterworth-Heinemann, Boston, p 461
200 H.G. Karge · H.K. Beyer

261. Kanazirev V, Price GL (1994) In: Weitkamp J, Karge HG, Pfeifer H, Hölderich W (eds) Zeo-
lites and Related Microporous Materials: State of the Art 1994. Proc 10th Int Zeolite Con-
ference, Garmisch-Partenkirchen, Germany, July 17–22, 1994, Elsevier, Amsterdam,
p 1935. Stud Surf Sci Catal 84:1935
262. Price GL, Kanazirev V, Dooley KM, Hart VI (1998) J Catal 173:17
263. Kazansky VB, Kustov LM, Khodakov AYu (1989) In: Jacobs PA, van Santen RA (eds) Zeo-
lites: Facts, Figures, Future. Proc 8th Int Zeolite Conf, Amsterdam, The Netherlands, July
10–14, 1989, Elsevier, Amsterdam, p 1173. Stud Surf Sci Catal 49:1173
264. Endoh A, Nishimiya K, Takaishi T (1989) In: Karge HG, Weitkamp J (eds) Zeolite as Cat-
alysts, Sorbents and Detergent Builders – Applications and Innovations (1989) Proc Int
Symposium,Würzburg, Germany September 4–8, 1988. Elsevier,Amsterdam, p 779. Stud
Surf Sci Catal 46:779
265. Halász J, Konya Z, Fudala A, Kiricsi I (1996) Catal Today 31:293
266. Kanazirev VI, Price GL, Dooley KM (1994) J Catal 148:164
267. Kikuchi E, Ogura M, Terasaki I, Goto Y (1996) J Catal 161:465
268. Ogura M, Aratani N, Kikuchi E (1997) In: Chon H, Ihm S-K, Uh YS (eds) Progress
in Zeolite and Microporous Materials. Proc 11th Int Zeolite Conference, Seoul,
Korea, August 12–17, 1996. Elsevier, Amsterdam, p 1593. Stud Surf Sci Catalysis 105B:
1593
269. Ogura M, Ohsaki T, Kikuchi E (1998) Microporous Mesoporous Mater 21:533
270. Zatorski LW (1987) Bull Pol Acad Sci 35:325
271. Beyer HK, Mihályi RM, Minchev Ch, Neinska Y, Kanazirev V (1996) Microporous Meso-
porous Mater 7:333
272. Mihályi RM, Pál-Borbély G, Beyer HK, Minchev Ch, Neinska Y, Karge HG (1997) React
Kinet Catal Lett 60:195
273. Mihályi RM, Beyer HK (1998) In: Rozwadowski M (ed) Proc 3rd Polish-German Zeolite
Colloquium, Torun, Poland, April 3–5, 1997, Nicholas Copernicus University Press,
Torun, p 309
274. Mihályi RM, Beyer HK, Mavrodinova V, Minchev Ch, Neinska Y (1998) Microporous
Mesoporous Mater 24:143
275. Neinska Y, Mihályi RM, Mavrodinova V, Minchev Ch, Beyer HK (1999) Phys Chem Chem
Phys (PCCP) 1:5761
276. Kanazirev V, Valtchev V, Tarassov MP (1994) J Chem Soc Chem Commun 1043
277. Mihályi RM, Beyer HK, Neinska Y, Mavrodinova V, Minchev Ch (1999) React Kinet Catal
Lett 68:355
278. Mihály RM, Beyer HK, Pál-Borbély G (2001) submitted to J Chem Soc Chem Commun
279. Broch NC, Christensen AN (1966) Acta Chem Scand 20:1996
280. Herman RG, Lunsford JH, Beyer H, Jacobs PA, Uytterhoeven JB (1975) J Phys Chem
79:2388
281. Beyer H, Jacobs PA, Uytterhoeven JB, Vandamme JL (1976) In: Proc 6th Int Congress
Catalysis, London, UK, July 12–16, 1976, Chemical Society, London, p 1 (A18)
282. Neinska Y, Minchev Ch, Dimitrova R, Micheva N, Minkov V, Kanazirev V (1994) In:
Weitkamp J, Karge HG, Pfeifer H, Hölderich W (eds) Zeolites and Related Microporous
Materials: State of the Art 1994. Proc 10th Int Zeolite Conference, Garmisch-
Partenkirchen, Germany, July 17–22, 1994, Elsevier, Amsterdam, p 989. Stud Surf Sci
Catal 84:989
283. Neinska Y, Minchev Ch, Kozova L, Kanazirev V (1995) In: Beyer HK, Karge HG, Kiricsi I,
Nagy JB (eds) Catalysis by Microporous Materials. Proc ZEOCAT ’95, Szombathely,
Hungary, July 9–13, 1995, Elsevier, Amsterdam, p 262. Stud Surf Sci Catalysis 94:262
284. Neinska YG, Mavrodinova VP, Kanazirev VI, Minchev CI (1998) Bulg Chem Commun
30:357
285. Neinska Y, Mavrodinova V, Minchev C, Mihály RM (1999) In: Kiricsi I, Pál-Borbély G,
Nagy JB, Karge HG (eds) Porous Materials in Environmentally Friendly Processes. Proc.
1st Int FEZA Conference, Eger, Hungary, September 1–4. Elsevier, Amsterdam, p 37. Stud
Surf Sci Catal 125:37
Solid-State Ion Exchange in Microporous and Mesoporous Materials 201

286. Karge HG, Hunger M, Beyer HK (1999) In: Weitkamp J, Puppe L (eds) Catalysis and zeo-
lites – fundamentals and applications. Springer, Berlin Heidelberg New York, chap 4,
p 210–211
287. Sulikowski B, Find J, Karge HG, Herein D (1997) Zeolites 19:395
288. Stranski IN (1928) Z Phys Chem A 136:259
289. Jiang M, Karge HK (1995) J Chem Soc Faraday Trans 91:1845
290. Jiang M, Koy J, Karge HK (1996) In: Occelli ML, Kessler H (eds) Proc. 3rd Int Symp Syn-
thesis of Zeolites, Expanded Layer Compounds and other Crystalline Microporous or
Mesoporous Solids. ACS Meeting, Anaheim, California, USA, April 2–7, 1995. Synthesis
of microporous materials: zeolites, clays and nanostructures. Marcel Dekker Inc., New
York, p 335
291. Karge HG, Niessen W (1991) Catal Today 8:451
292. Jiang M, Karge HG, publication in preparation
293. Crank J (1975) In: The mathematics of diffusion, 2nd edn. Clarendon Press, Oxford, p 96
294. Schoonheydt R, Uytterhoeven JB (1969) Clay Minerals 8:71
295. Simon U, Flesch U, Maunz W, Müller R, Plog C (1998) Microporous Mesoporous Mater
21:111