Dalton

Transactions
PAPER

Tuning the target composition of amine-grafted

Cite this: Dalton Trans., 2015, 44,
CPO-27-Mg for capture of CO2 under
18970 post-combustion and air filtering conditions: a
combined experimental and computational study†
M. C. Bernini,*a A. A. García Blanco,b J. Villarroel-Rocha,b D. Fairen-Jimenez,c
K. Sapag,b A. J. Ramirez-Pastorb and G. E. Nardaa

A computational and experimental screening of hypothetical and real compounds exhibiting different
degrees of ethylenediamine grafted to the CPO-27-Mg or Mg-DOBDC skeleton is performed in order to
determine the target composition that optimizes the CO2 adsorption properties under flue gas and air
filtering conditions. On the basis of the [Mg2(dobdc)] formula, eighteen hypothetical models involving
15–100% of functionalization of the coordinatively unsaturated sites (CUS) were considered by means of
Grand Canonical Monte Carlo simulations to evaluate the CO2 adsorption at 298 K. In addition, post-syn-
thesis modification was applied to CPO-27-Mg leading to three kinds of samples exhibiting 15, 50, and
60% of CUS functionalization with ethylenediamine, named CPO-27-Mg-a, CPO-27-Mg-b and CPO-27-
Mg-c. Compounds were characterized using elemental analysis, TGA, FTIR spectroscopy, PXRD and DSC.
Received 13th August 2015, Finally, bare and functionalized CPO-27-Mg materials were evaluated using gas adsorption and micro-
Accepted 25th September 2015
calorimetry in the 0.001–1 bar range, which is pertinent for the mentioned applications. Valuable infor-
DOI: 10.1039/c5dt03137k mation related to design criteria for synthesis of tuned CO2 adsorbents is derived through this
www.rsc.org/dalton computational and experimental investigation.

1. Introduction step in these technologies is the selective separation of CO2

from the gas mixtures that integrate the different gas flows in
Facing the global warming concern, technologies involving each chemical process.
CO2 capture can be performed in pre-combustion, post- The elimination of CO2 from low pressure flue gases as
combustion or in oxy-combustion routes. Since the operating con- emitted, for example, by fossil fuel burning power plants can
ditions vary greatly with each method, the required features of be accomplished very effectively by means of gas scrubbing.3
materials in each process are also different.1 An interesting The partial pressure of CO2 in these streams lies between
alternative involves the selective removal of CO2 from gas mix- 50–200 mbar.1,4,5 Aqueous amine solutions (e.g. monoethanol-
tures by the carbon capture and sequestration (CCS) processes, amine) are currently being used as absorbents for the removal
which includes the compression of pure CO2 to a supercritical of CO2 from industrial commodities, a technique which has
fluid, transportation to an injection site, and finally its desti- been employed in the natural gas industry for more than 60
nation to an underwater or underground storage.2 A critical years.6 However, this methodology has several disadvantages
such as pipe corrosion, high energy consumption associated
with the regeneration of the absorbent solution, degradation
a
Área de Química General e Inorgánica, Facultad de Química, Bioquímica y in the presence of oxygen and SOx, etc. As such, the use of
Farmacia, Universidad Nacional de San Luis. Instituto de solid adsorbents, which are easy to regenerate and exhibit a
InvestigacionesenTecnologíaQuímica (INTEQUI-CONICET), 5700 San Luis,
high affinity for selective CO2 adsorption under flue gas con-
Argentina. E-mail: [email protected]
b
Departamento de Física. Universidad Nacional de San Luis. Instituto de
ditions, has been proposed as a good alternative for CO2
FísicaAplicada (INFAP-CONICET), 5700 San Luis, Argentina capture in thermal plants.7 Additionally, carbon dioxide scrub-
c
Department of Chemical Engineering and Biotechnology, University of Cambridge, bers are critical to life support systems in confined spaces with
Pembroke St., Cambridge CB2 3RA, UK limited air exchange, such as spacecrafts, submarines, and re-
† Electronic supplementary information (ESI) available: Details of the post-
breathing suits.8 With these applications in mind, research
synthesis procedure, elemental and thermal analysis characterization, XRPD
patterns, adsorption data, force fields and molecular simulations. See DOI: into the direct capture of this gas from more diluted stream
10.1039/c5dt03137k gases is also underway, being studied for its implementation

18970 | Dalton Trans., 2015, 44, 18970–18982 This journal is © The Royal Society of Chemistry 2015
Dalton Transactions Paper

in air filter technologies, with 5 mbar serving as a reference a N2 gas flow) in comparison with the bare material.25 The
pressure value.9 Zeolites,10 activated carbons,11 silicas,12 and Mg2(dobpdc) compound, adopting an isoreticular expanded
new classes of porous materials13 have been investigated for CPO-27 structure (1),26 was also modified by the functionali-
their CO2 adsorption behaviour. Among them, MOFs with zation of 80% of its CUS employing ethylenediamine (1-en)26
high porosity, good thermal stability and containing strong and N,N′-dimethylethylenediamine (mmen-Mg2(dobpdc)),9
Lewis acid-basic sites are emerging as one of the most promis- exhibiting promising features for CO2 adsorption from flue gas
ing candidates and have demonstrated excellent qualities for mixtures and from air under ambient conditions. Moreover,
separation of multicomponent systems, such as CO2–N2 and these functionalized Mg2(dobpdc) compounds showed higher
CO2–CH4.14 stability than the non-functionalized analogues when exposed
Experimental and theoretical investigations have shown to atmospheric conditions.9
that CPO-27-M,15 where M is a divalent metal, is a particularly Taking this into account, we decided to perform a screening
promising material.16 This MOF (also known as M-MOF-7417 of a high number of real and hypothetical materials with
or M-DOBDC15f ) consists of one dimensional channels com- different degrees of ethylenediamine functionalization. We
pletely lined with coordinatively unsaturated sites (CUS), and carried out CO2 adsorption experiments on three fully charac-
it can be synthesized with Mg, Mn, Fe, Co, Ni, Cu, and Zn as terized samples exhibiting 15–60% of CUS functionalization.
the metal center. The enthalpy of adsorption (Δḣads) of CO2 in Besides, eighteen hypothetical models were considered for
these materials at low loadings has been reported in the range complementary Monte Carlo simulations, covering 15–100%
of 37–73 kJ mol−1 (ref. 15a, f, 18 and 19) and such particularly of CUS functionalization. This procedure aims to determine
large enthalpy values have been associated with the interaction the targeted composition that optimizes selective CO2 capture
of CO2 with the CUS in the structure. For post-combustion for flue gas and air filtration technologies, through a systema-
capture, the desired range of Δḣads lies between the physisorp- tic variation of ethylenediamine incorporation in the CPO-27-
tion and chemisorption energies, which would be optimal for Mg skeleton. The influence of –NH2 groups/free CUS-Mg2+
the regenerability of a material having sufficient CO2 speci- ratio, the pore geometry, the pore size distribution (PSD) and
ficity to be economically viable.20 A reliable determination of the remaining void volume fraction in the CO2 adsorption
this parameter is then desirable in order to characterize the performance and CO2/N2 selectivity are analysed. Moreover,
adsorbate–adsorbent interaction and to critically evaluate the once the best composition of amine-grafting was identified,
applicability of these kinds of materials. In general, this para- recycling experiments and determination of differential
meter is derived from isotherm data obtained at two or, enthalpy of adsorption from direct microcalorimetry measure-
less frequently, three temperatures, applying the Clausius– ments of both functionalized and non-functionalized versions
Clapeyron equation. Only a few studies have involved in the of CPO-27-Mg were carried out.
direct measurement of Δḣads of CO2 in MOFs by adsorption
microcalorimetry, and even lesser examples in the case of
amine-functionalized MOFs.21 The CPO-27-M family is also 2. Experimental section
important because expanded versions of these MOFs have
2.1 Synthesis of CPO-27-Mg
been synthesized using longer DOBDC linker analogues, con-
ducive to materials with extremely high porosity.22 Addition- This compound was obtained following a reported proce-
ally, the CUS are suitable for post-synthetic modification, dure,15f in which 1.83 mmol of Mg(NO3)2·6H2O and 0.53 mmol
allowing the incorporation of further polar functional of 2,5-dihydroxyterephthalic acid were dissolved in a mixture
groups that may increase the chemical affinity with specific of solvents composed of 44 mL of dimethylformamide (DMF),
adsorbates, which is desirable for capture of CO2 from diluted 3 mL of ethanol and 3 mL of distilled water. The resultant
( post-combustion) and very diluted (enriched environments) solution was heated at 398 K for 21 h in a Teflon-lined diges-
stream gases. In particular, the grafting of diamine molecules tion Parr bomb (120 mL of internal volume). After that, the
on CUS has demonstrated improvement in CO2 adsorption reactor was immediately cooled to room temperature and the
performance.23 yellow microcrystalline powder obtained was filtered and
In recent papers, one degree of functionalization of CUS washed with DMF and methanol. The solid was left in contact
was studied experimentally by grafting diamine molecules in with methanol (50 mL) and it was replaced four times over
the CPO-27-M members24,25 and expanded analogues9,26 for three days. Finally, the solvent was extracted with a Pasteur
CO2 adsorption applications. Thus, 25% of CUS present in pipette and the solid was dried at room temperature.
CPO-27-Ni were functionalized by anchoring a cyclic diamine
( piperazine), showing good CO2/N2 adsorption selectivity, but 2.2 Post-synthetic modification of CPO-27-Mg with
without improving the adsorption capacity at low pressure, ethylendiamine
when compared to the non-functionalized material. Consider- Samples of CPO-27-Mg were thermally treated under dynamic
ing the analogue CPO-27-Mg,15g about 11.11% of the CUS vacuum conditions in order to remove the solvent (co-
were functionalized by using ethylenediamine, showing an ordinated and non-coordinated). The resulting partially acti-
increased adsorption capacity under ultradilute gas stream vated samples were put in contact with different solutions of
conditions ( pCO2 = 0.0004 bar, equivalent to 400 ppm of CO2 in ethylenediamine in toluene, with the aim of obtaining

This journal is © The Royal Society of Chemistry 2015 Dalton Trans., 2015, 44, 18970–18982 | 18971
Paper Dalton Transactions

materials with different amine-grafted contents. The suspen- criteria proposed by Rouquerol et al.28 were taken into
sions were stirred for 65 h at room temperature, after which account. CO2 (99.995% purity) adsorption isotherms were
the solid was separated by centrifugation and washed with measured using a manometric adsorption apparatus (Micro-
toluene and hexane. The samples obtained by this procedure meritics, ASAP 2050) at 298 and 323 K.
were dried at 333 K for 2 h. The details of the functionalization In order to evaluate the reusability of the functionalized
procedures of each sample (named as CPO-27-Mg-a, CPO-27- and non-functionalized CPO-27-Mg samples, two consecutive
Mg-b and CPO-27-Mg-c) are displayed in the ESI.† CO2 adsorption isotherms at 298 K were obtained, and to
The content of amine grafted was studied by ATG-DSC, evaluate the stability of the porosity of both materials under
FTIR and elemental analysis; see Fig. S1–S3 and Table S1, the recycling conditions, an equivalent experiment was per-
ESI.† Powder X-ray diffraction allowed the confirmation of the formed by adsorbing N2 at 77 K in two consecutive cycles. The
structure and stability of all products (see Fig. S4, ESI†). experimental details are described in section S7, ESI.† All of
All samples, non-functionalized and functionalized, were these experiments were carried out using a manometric
employed in adsorption experiments and characterization adsorption apparatus (Micromeritics, ASAP 2050).
within 24 h of their preparation. During this time all samples
were kept under ambient conditions. 2.8 CO2 adsorption microcalorimetry

2.3 Powder X-ray diffraction (PXRD) CO2 adsorption microcalorimetric experiments were per-
formed at room temperature (∼300 K) using ∼0.07 g of
PXRD patterns were obtained with a Rigaku D-MAX-IIIC dif-
sample. The evolved heat was measured using a Tian-Calvet
fractometer using Cu Kα radiation (λ = 1.5418 Å) and quartz as
microcalorimeter (model CA-100, ITI), which was connected to
an external calibration standard. The best counting statistics
an adsorption manometric setup to measure simultaneously
were achieved using a scanning step of 0.03° between 4° and
the adsorption isotherm and the differential enthalpy of
50° Bragg angles with an exposure time of 5 s per step.
adsorption. A pressure transducer of 1000 Torr that provides
2.4 Fourier-transform infrared (FTIR) spectroscopy high accuracy and high resolution in the 3 × 10−5 to 1.4 bar
range was employed.
FTIR spectra were recorded with a Nicolet Protégé 460 spectro-
The total heat (Q) produced in each adsorption step was
meter in the 4000–225 cm−1 range with 64 scans and a spectral
calculated as:
resolution of 4 cm−1 using the KBr pellet technique.
ð t1
2.5 Thermal analysis Q¼K EV dt; ð1Þ
t0
Thermogravimetric Analysis (TGA) and Differential Scanning
Calorimetry (DSC) were performed between room temperature where EV is the voltage signal (in V) registered between times t0
and 873 K, employing a Shimadzu TGA-51 and a DSC-60 and t1 and K is the microcalorimeter constant (in W V−1),
apparatus, respectively, under air flowing at 50 mL min−1, and determined by electrical calibrations. The so-called discontinu-
a heating rate of 10 K min−1. ous procedure was used, where the differential enthalpy of
adsorption (Δḣads T, n) is given by:28
2.6 Elemental analysis    
dQrev dp
Elemental analyses for C, H and N were conducted in an Δḣads T; n ¼ þ VC  ; ð2Þ
dnσ T dnσ T
Exeter CE 440 instrument. The samples were analyzed both
as-prepared and after activation processes ( performed for where dQrev is the heat reversibly exchanged with the sur-
recording adsorption isotherms) without additional thermal roundings at temperature T, dnσ is the adsorbed amount due
treatment. The proposed stoichiometries for each sample are to the pressure increase dp and VC is the dead volume of the
shown in Table S1, ESI.† adsorption cell immersed in the microcalorimeter. According
to the setting of the instrument, and since the magnitude of
2.7 Gas adsorption the second term of eqn (2) is negligible with respect to the first
Before the measurement of the adsorption isotherms, the term, the differential enthalpy of adsorption (Δḣads T, n) was
CPO-27-Mg samples were degassed at 523 K (or 503 K in the calculated by using:
case of ethylenediamine grafted samples) for 6 h in a dynamic  
dQrev
vacuum (20 µTorr). The most convenient outgassing tempera- Δḣads T; n ¼ ð3Þ
dnσ T
ture for the functionalized samples was determined by con-
sidering the TG-DSC curves of CPO-27-Mg-c (see Fig. S1 and It is important to remark that the enthalpy of CO2 adsorp-
S2, ESI†). tion is frequently expressed as the isosteric heat of adsorption
Measurements of N2 (99.999%) at 77 K were carried out (Qst) as a function of the quantity of CO2 adsorbed.1 The Qst
using a manometric adsorption equipment (ASAP 2000, Micro- value is a parameter that describes the average enthalpy
meritics). The Brunauer, Emmett and Teller (BET) method27 of adsorption for an adsorbing gas molecule at a specific
applied to N2 adsorption data was used to estimate the specific surface coverage and is usually evaluated using two or more
surface area (SBET) of the samples, where the consistency CO2 adsorption isotherms collected at similar temperatures

18972 | Dalton Trans., 2015, 44, 18970–18982 This journal is © The Royal Society of Chemistry 2015
Dalton Transactions Paper

(usually within 10 K of each other). However, even though the vapour–liquid coexistence curves. Lorentz–Berthelot mixing
term Qst is now discouraged and should be replaced with the rules were used for all cross terms, and LJ interactions beyond
isosteric enthalpy of adsorption,28 it is still frequently found in 12.8 Å were neglected. Coulomb interactions were calculated
the related literature. using partial charges on the atoms, calculated by the QEq
method as described above. The Ewald sum method was used
to compute the electrostatic interactions. Up to 4 × 104 Monte
3. Computational details Carlo equilibration cycles plus 4 × 104 production cycles were
performed to calculate the ensemble averages. In one cycle,
3.1 Models of CPO-27-Mg grafted with ethylenediamine
N moves were performed, where N is the number of molecules
Three initial conformations for the ethylenediamine molecule in the system (which fluctuates in GCMC). Monte Carlo moves
adopting three different N–C–C–N dihedral angles (ω) were used with equal probability were translation, rotation, inser-
considered, i.e. (i) ω1 = 60°, (ii) ω2 = −60° and (iii) ω3 = 180°. tion, deletion and random reinsertion of an existing molecule
For each one, six different models were obtained by grafting at a new position. The pore volume was obtained using a
these ethylenediamine molecules into 1 to 6 inorganic chains Widom particle insertion method, by probing the structure
(SBUs, secondary building units) of the framework’s hexagonal with a helium molecule at room temperature, and recording a
channels. Each model is denoted as ωi-J, where i = 1–3, large number of random points not overlapping the van der
depending on the ethylenediamine dihedral angle, and J = Waals volume of the framework.36 The pore size distributions
1–6, depending on the degree of functionalization. The unit were calculated using the method of Gelb and Gubbins,37
cells of each model were then subjected to energy minimiz- where the largest sphere that can fit in a random point within
ation and geometry optimization based on molecular mech- a structure, without overlapping the van der Waals surface of
anics, which include van der Waals and electrostatic the framework, is recorded for a large number of random
interactions, allowing for modification of the unit cell size and points.
the atomic coordinates of the new structures. These calcu-
lations were performed with the Forcite module of Materials
Studio, using an algorithm which is a cascade of the steepest
descent, an adjusted basis Newton–Raphson set, and quasi-
Newton methods.29 The bonded and the short range (van der
4. Results and discussion
Waals) interactions between atoms were modeled using the 4.1 Post-synthesis modification and characterization
Universal Force Field (UFF).30 A cut-off distance of 18.5 Å was Three samples exhibiting different ethylenediamine contents
used for the van der Waals interactions during the geometry after post-synthesis modification were obtained and fully
optimization. The long-range electrostatic interactions, arising characterized. Elemental analysis of %C, %H and %N were
from the presence of partial atomic charges, were modelled conducted for all functionalized samples before and after the
using a Coulombic term. The Ewald sum method was used to adsorption experiments (see Table S1, ESI†). Because the as-
compute electrostatic interactions. Partial atomic charges were synthesized samples and the samples used in the adsorption
derived using the charge equilibration method (QEq) as studies were kept under standard atmospheric conditions,
implemented in Materials Studio.31 and since the elemental analysis was conducted in air, CO2
was present in some samples. This was also observed in the
3.2 Monte Carlo simulations expanded CPO-27-Mg analogue compound functionalized
The CO2 adsorption at 298 K was studied using grand cano- with ethylenediamine.26 According to the elemental analysis,
nical Monte Carlo (GCMC) simulations, performed with the the formulae of the post-synthetic functionalized samples are
multi-purpose code RASPA.32 Atomistic models obtained as [Mg2(DHT)(H2O)1.7(en)0.3]·5H2O (as-synthesized CPO-27-Mg-a),
described above were used as adsorbent models. Framework [Mg2(DHT)(H2O)(en)]·0.2(en)·3H2O (as-synthesized CPO-27-
atoms were kept fixed at their crystallographic positions in all Mg-b) and [Mg2(DHT)(H2O)0.8(en)1.2]·0.2(en)·3H2O (as-
simulations. CO2–CO2 and CO2–framework interactions were synthesized CPO-27-Mg-c). Thus, we were able to prepare
calculated using a Lennard-Jones (LJ) + Coulomb potential. LJ samples exhibiting CUS functionalization degrees of 15, 50
parameters for the C, H and O atoms of the CPO-27 skeleton and 60%, respectively.
were taken from the Dreiding33 force field and the ones corres- Thermogravimetric analysis was also conducted for all
ponding to the Mg atom were taken from the UFF;32 those the samples, so that the proposed formulae from elemental
parameters corresponding to C, H and N atoms belonging to analysis could be confirmed (see Fig. S1, ESI†). Differential
the ethylenediamine portion were taken from the TraPPE force scanning calorimetry (DSC) was performed in order to evaluate
field34 developed for amine molecules (see Table S2, section S8 the energy variation as a function of thermal treatment (see
in ESI†). Carbon dioxide was modeled as a linear triatomic Fig. S2, ESI†). A detailed description of the TGA-DSC curves is
molecule with fixed bond lengths and bond angles. The atoms displayed in section S3 of ESI.†
were modeled by charged Lennard-Jones (LJ) centers using the The incorporation of ethylenediamine into the CPO-27-Mg
TraPPE force field developed by Potoff and Siepmann (see frameworks was also studied qualitatively by FTIR spectro-
Table S3†).35 This force field has been fit to reproduce the scopy. As can be seen in Fig. 1 and S3,† several new bands

This journal is © The Royal Society of Chemistry 2015 Dalton Trans., 2015, 44, 18970–18982 | 18973
Paper Dalton Transactions

which exhibited higher amine contents (CPO-27-Mg-b and -c).

This could be associated with the elongation of the Mg–O
bonds in the MgO5 polyhedra consequential from the acti-
vation followed by the grafting of the ethylenediamine process.
Furthermore, two weak bands at ∼303 and 290 cm−1 can be
assigned to Mg–N stretching vibration, and as can be seen in
Fig. 6 and S3,† they are more notorious in the spectrum of
CPO-27-Mg-c, in keeping with the highest amount of amine
present in this compound.

4.2 CO2 adsorption isotherms

The most convenient outgassing temperature was determined
by analysing the TG-DSC curves, looking for the optimal temp-
erature at which the non-coordinated ethylenediamine and the
eventual solvent molecules remaining at the framework chan-
nels can be evacuated, without eliminating the coordinated
ethylendiamine ones. As an example, Fig. S5† shows the effect
of different outgassing temperatures in the CO2 adsorption
performance of the CPO-27-Mg-c sample. When it is outgassed
at 393 K in a dynamic vacuum, the CO2 adsorption capacity is
low, showing that the pores remain mostly blocked, but at an
outgassing temperature of 503 K, the CO2 adsorption capacity
is strongly increased. Thus, 503 K was selected as the out-
gassing temperature for all functionalized samples, while
523 K was chosen for the non-functionalized samples.
In order to properly correlate the adsorption performance
with the ethylenediamine composition, the elemental analysis
of all samples was studied at the end of the adsorption experi-
ments. Thus, the amine content measured corresponds to the
one kept after the outgassing procedure performed before
adsorption. The corresponding formulae for the recovered
samples are [Mg2(DHT)(H2O)1.7(en)0.3]·3H2O (CPO-27-Mg-a),
[Mg2(DHT)(H2O)1.2(en)0.8]·0.4CO2·4H2O (CPO-27-Mg-b) and
[Mg2(DHT)(H2O)(en)]·CO2·2H2O (CPO-27-Mg-c), which rep-
resent 15%, 40% and 50% of functionalization remaining after
the outgassing treatment. The difference between the amine
contents, of the as-synthesized and post-isotherm samples, is
due to the outgassing treatment, which includes heating
Fig. 1 FTIR spectra of as-synthesized CPO-27-Mg, CPO-27-Mg-a, -b under dynamic vacuum conditions.
and -c and ethylenediamine in selected spectral zones. The bands The CO2 adsorption isotherms measured at 298 K for the
marked with * are assigned to ethylenediamine molecules. four samples under study are displayed in Fig. 2. The amounts
adsorbed at key pressure values are presented in Table 1 along
with the corresponding values obtained for other functiona-
lized CPO-27 compounds9,24,26 for the sake of comparison.
coincident with those of ethylenediamine appear in the IR At ∼1 bar the adsorption capacity of the CPO-27-Mg is
spectra of the functionalized samples. about 7.1 mmol g−1, which is slightly lower than the highest
Those bands located at ∼3350, 3290 and 3175 cm−1 are suit- reported value (8.1 mmol g−1) for this MOF,38 being the differ-
able for the assignment to N–H stretching vibrations, while ence probably related to the storage and manipulation con-
those at ∼2930 and 2865 cm−1 are assigned to the stretching of ditions applied (vacuum in ref. 38 or atmospheric conditions
aliphatic C–H bonds. Moreover, two bands located at ∼1000 in this work). The adsorption capacities for the functionalized
and 965 cm−1 in the spectra of those samples exhibiting samples at ∼1 bar decrease to 5.22, 4.97 and 5.42 mmol g−1,
higher amine contents can be associated with the same for CPO-27-Mg-a, -b, and -c, respectively. The decrease of about
vibration of aliphatic C–C bonds. Interestingly, the bands 25–30% could be associated with the reduction of the free void
located at 407, 374 and 331 cm−1 in the spectrum of the fraction in the ethylenediamine-containing samples. A similar
CPO-27-Mg compound are shifted to 399, 367 and 316 cm−1 in behaviour was observed for the CPO-27-Ni and piperazine-
the corresponding spectra of the functionalized analogues CPO-27-Ni compounds24 with the corresponding adsorption

18974 | Dalton Trans., 2015, 44, 18970–18982 This journal is © The Royal Society of Chemistry 2015
Dalton Transactions Paper

As a consequence of the incorporation of polar functional

groups into the pore, the adsorbent–adsorbate interactions
were modified, affecting the adsorption performance, particu-
larly at low pressures (below 0.2–0.3 bar). For this reason, it is
interesting to analyse carefully the impact of the different
functionalization degrees on the adsorption isotherms focus-
ing on two key pressure zones: 5 mbar (important for air filter-
ing technologies), and 50–200 mbar (for capture under post-
combustion conditions).
4.2.1 CO2 capture in the 5–50 mbar pressure range. In this
pressure range, different adsorption behaviour is observed for
the functionalized samples relative to the non-functionalized
material. As shown in Fig. 2, the isotherm of the sample con-
taining 15% functionalized CUS (CPO-27-Mg-a) is below the
corresponding one to the non-functionalized material over
this pressure range, showing that this degree of amine content
does not improve CO2 affinity compared to the CPO-27-Mg
compound. A similar situation is also observed for the CPO-27-
Ni and the piperazine-grafted compound, since the adsorption
capacity of the functionalized adsorbent is lower than that of
the non-functionalized adsorbent.24 The opposite is true for
those samples with 40% (CPO-27-Mg-b) and 50% (CPO-27-Mg-
c) CUS functionalization, which exhibit a higher CO2 adsorp-
tion up to 50 mbar and 65 mbar, respectively, relative to the
performance of the CPO-27-Mg. Interestingly, the best per-
formance in this pressure range is exhibited by the sample
containing 50% of the Mg2+ CUS grafted with ethylenediamine
molecules (CPO-27-Mg-c). At 5 mbar, a reference pressure
value for CO2-enriched environments,8 this material adsorbs
Fig. 2 Experimental CO2 adsorption isotherms obtained for CPO-27-Mg 2 mmol g−1, which is 3.3 times higher than the amount
and functionalized CPO-27-Mg samples at 298 K. Dashed lines indicate adsorbed by the non-functionalized material. This value is
the corresponding pressure ranges for the proposed applications.
close to the one reported for mmen-Mg2(dobpdc)9 (2.6 mmol
g−1), and significantly smaller than the reported value for
1-en26 (3.3 mmol g−1). Samples with more than 50% of amine
capacities at this pressure being ∼5.5 mmol g−1 and 3.2 mmol functionalization remaining after the outgassing process could
g−1, respectively. It is interesting to note that even when the not be obtained, since the maximum achieved functionali-
functionalization degree of Ni-CUS was only 25%, the influ- zation degree was 60%, which might be a limit value related
ence of employing the bulkier diamine molecule piperazine is with the pore volume availability. Additionally, according to
evident in the more marked uptake decrease, being about 42% A. Das et al.20 primary amines like ethylenediamine exhibit a
less than the corresponding value for the non-functionalized high tendency to form hydrogen bond interactions among
compound. them, reducing the possibility to effectively coordinate the

Table 1 Experimental CO2 adsorption loadings obtained in this study and related compounds

% of functionalized Uptake at 5 mbar Uptake at 50 mbar Uptake at 150 mbar Uptake at 1 bar
Sample CUS (mmol g−1) (mmol g−1) (mmol g−1) (mmol g−1) Ref.

CPO-27-Mg 0 0.6 2.55 3.9 7.1 This work

CPO-27-Mg-a 15 0.5 2.20 3.12 5.22 This work
CPO-27-Mg-b 40 1.6 2.48 3.25 4.92 This work
CPO-27-Mg-c 50 2 2.85 3.6 5.42 This work
CPO-27-Mg 0 — 5 6.1 8.1 38
CPO-27-Ni 0 — 1.5 3.3 5.4 24
pip-CPO-27-Ni 25 — 0.8 1.25 3.2 24
1 0 — — — — 26
1-en 80 3.3 3.45 3.62 4.57 26
Mg2(dobpdc) 0 3.8 4.85 6.42 9
mmen-Mg2(dobpdc) 80 2.6 2.9 3.13 3.86 9

This journal is © The Royal Society of Chemistry 2015 Dalton Trans., 2015, 44, 18970–18982 | 18975
Paper Dalton Transactions

CUS in the MOF framework, precluding the availability of For the in depth characterization of the sample whose com-
higher degrees of post-synthesis modification in microporous position seems to be optimal, and since the main component
systems. However, considering our results, it would be of inter- of air and flue gases is N2, the adsorption isotherms of this
est to perform a screening of a wide range of functionalization gas were also measured at 323 K for both, CPO-27-Mg and
degree to some of the IRMOF-7422 members, since eight out of CPO-27-Mg-c. As shown in Fig. 3, both materials exhibit
nine of those materials exhibit pore diameters in the range of almost coincident N2 adsorption isotherms, the loading below
mesopores (>20 Å). 65 mbar being lower than 0.22 mmol g−1, while at 5 mbar is
4.2.2 CO2 capture in the 50–150 mbar pressure range. At 0.0165 mmol g−1. On the contrary, the CO2 adsorption behav-
150 mbar, which is a reference value for the partial pressure of iour is markedly sensitive to the functionalization, being the
CO2 under flue gas conditions, CPO-27-c adsorbs 3.6 mmol loadings at this pressure 0.24 and 1 mmol g−1 for CPO-27-Mg
g−1, which is the same amount reported for 1-en26 and slightly and CPO-27-Mg-c, respectively. This fact reflects the improve-
more than the amount adsorbed by mmen-Mg2(dobpdc)9 (see ment in the CO2 selectivity that would be achieved with the
Table 1). Considering the different performances found for the composition of CPO-27-Mg-c.
functionalized samples with respect to the bare material, and
taking into account the operative conditions of the flue gas in 4.3 Recycling experiments and differential enthalpy of
industrial facilities, CO2 adsorption isotherms were also adsorption of CO2
measured at 323 K for CPO-27-Mg and CPO-27-Mg-c (see In order to complete the characterization of CO2 adsorption
Fig. 3). At this temperature, the presence of diamine grafted for the optimal composition (50% of CUS functionalization),
molecules improves the CO2 adsorption below 65 mbar of and to properly evaluate the reusability of this material in com-
pressure compared to the non-functionalized material, while parison with that of the non-functionalized material, two con-
at 150 mbar the adsorption capacity of CPO-27-Mg is 20% secutive CO2 adsorption isotherms were performed at 298 K,
higher than that of CPO-27-Mg-c. including an outgassing step between both experiments (see
These results suggest that the latter material could be more Experimental details, section 2.8). As seen in Fig. 4, the
efficient in those applications that require a high chemical adsorption capacity of the CPO-27-Mg-c suffers an important
affinity to selectively adsorb CO2 in the presence of other com- decrease in the second cycle, while the performance of the
ponents (e.g., in air filter technologies). non-functionalized material remains almost unaltered.
Moreover, considering the significant reduction of the
specific surface area and the micropore volume by the
functionalization treatment, i.e. SBET = 1280 m2 g−1and Vo-DR
= 0.49 cm3 g−1 for the bare material to SBET = 400 m2 g−1and
Vo-DR = 0.16 cm3 g−1 for the CPO-27-Mg-c sample (see Fig. S6
in ESI†), it is remarkable to observe the impact of introducing
polar functional groups leading to an optimum composition
and a porous architecture that increase notably the CO2
adsorption at low pressure.

Fig. 3 CO2 (filled symbols) and N2 (open symbols) adsorption iso- Fig. 4 CO2 recycling experiments performed at 298 K for CPO-27-Mg
therms obtained at 323 K for the CPO-27-Mg and CPO-27-Mg-c (up) and CPO-27-Mg-c (down); 1 and 2 denote the first and second
samples. cycles, respectively.

18976 | Dalton Trans., 2015, 44, 18970–18982 This journal is © The Royal Society of Chemistry 2015
Dalton Transactions Paper

Fig. 6 Differential enthalpy of adsorption of CO2 vs. loading deter-

mined by microcalorimetry experiments for CPO-27-Mg and CPO-27-
Mg-c at ∼300 K.

It follows that a reliable determination of the differential

enthalpy of adsorption in both materials is of great impor-
tance, since this value is directly related to the chemical
affinity between the adsorbent and the adsorbate and may
justify the selectivity as well as energy requirements for releas-
Fig. 5 N2 adsorption recycling experiments performed at 77 K for ing CO2 molecules during the regeneration process. Using an
CPO-27-Mg (up) and CPO-27-Mg-c (down); 1 and 2 denote the first and adsorbent that binds CO2 too strongly would increase the
second cycles, respectively. regeneration cost due to the energy required to break the
framework–CO2 interactions. Meanwhile, if the enthalpy of
adsorption is too low, the material could be more readily
This result could be due to: (i) a possible alteration of the regenerated, but the purity of the captured CO2 would be
porous architecture ( partial degradation) by the whole re- lowered due to the decreased adsorption selectivity.1
cycling process; (ii) the permanence of some CO2 molecules The differential enthalpy of adsorption of CO2 as a function
“strongly” adsorbed on the CPO-27-Mg-c MOF surface that of loading was determined as a direct measurement by micro-
would block the strongest adsorption sites and fill partially the calorimetry experiments and the corresponding profiles are
porosity, diminishing the available pore volume and conse- shown in Fig. 6, while the associated adsorption isotherms are
quently, giving rise to a lower adsorption capacity along the displayed in Fig. S7.†
whole pressure range. As can be seen, the differential enthalpy of adsorption at
In order to investigate the hypothesis stated in (i) an ana- ∼zero loading is highly incremented in the case of the functio-
logue recycling experiment employing the functionalized and nalized CPO-27-Mg-c material, being the initial value of 132 kJ
bare materials was performed by means of N2 adsorption at mol−1 associated with a chemisorption process. In the case of
77 K. If the porous architecture of any material is affected by the bare CPO-27-Mg material, this magnitude reaches the
the recycling process, a diminution in the second N2 adsorp- value of 53 kJ mol−1, which is close to the maximum values
tion isotherm should be appreciated. However, Fig. 5 shows found for this MOF by applying the Clausius–Clapeyron deter-
that only for the case of the non-functionalized compound a mination method.1
small diminution during the second cycle is observed, but According to our measurements, the adsorption–desorption
both isotherms of the CPO-27-Mg-c are almost coincident, process involved in the case of the functionalized compound
thus indicating that the porosity is not affected by the re- is not completely reversible, since different magnitudes of
cycling procedure. These results suggest that CO2 molecules energy (as heat) were measured during the adsorption and
have a strong interaction with the framework surface of desorption steps, being higher in the case of the adsorption
CPO-27-Mg-c, since in spite of the outgassing treatment step. This fact suggests that part of the CO2 adsorbed mole-
between the isotherms (at 503 K during 6 hours under cules is strongly chemisorbed and, thus, they remain within
dynamic vacuum conditions), the pores of the functionalized the framework. Regarding the magnitude of the measured
sample would be partially filled with CO2 at the beginning of enthalpy and the observed non-reversibility of the process, the
the second cycle. formation of a stable carbamate species could be occurring.

This journal is © The Royal Society of Chemistry 2015 Dalton Trans., 2015, 44, 18970–18982 | 18977
Paper Dalton Transactions

This fact justifies the decrease in the second cycle of the CO2 Table 2 Functionalization degrees, void fractions, and potential energy
adsorption isotherm and suggests that in spite of the selecti- values for the ωi-J models obtained after energy minimization
vity being highly improved with the targeted composition,
Model % Functionalization Void fraction Energy (kJ mol−1)
stronger energetic requirements may be needed to remove all
the CO2 adsorbed molecules, which is not desirable for tech- ω1-1 16.6 0.640 −29 844.36
nological applications. In contrast, under the applied con- ω1-2 33.3 0.566 −30 002.448
ω1-3 50 0.448 −30 051.21
ditions (adsorption of pure CO2 gas) the non-functionalized ω1-4 66.6 0.356 −29 992.284
material showed a reversible behaviour with a measured differ- ω1-5 83.3 0.274 −30 061.458
ential enthalpy of adsorption at ∼zero loading being about ω1-6 100 Non-porous −8754.606
ω2-1 16.6 0.640 −30 744.546
53 kJ mol−1. This was determined not only by getting the same ω2-2 33.3 0.544 −30 890.244
energy magnitudes on the adsorption and desorption pro- ω2-3 50 0.443 −30 888.48
cesses, but also by keeping the same adsorption capacity ω2-4 66.6 0.350 −30 524.508
ω2-5 83.3 0.272 −30 193.422
during the second cycle (see Fig. 4). However, as it was already ω2-6 100 Non-porous −8881.404
shown,39 the adsorption capacity of the non-functionalized ω3-1 16.6 0.642 −29 410.626
CPO-27-Mg material is strongly diminished by the presence of ω3-2 33.3 0.543 −29 176.644
ω3-3 50 0.442 −28 606.284
water in the post-combustion gas mixture. ω3-4 66.6 0.327 −28 486.248
ω3-5 83.3 0.199 −28 262.346
4.4 Molecular simulations ω3-6 100 Non-porous −7854.126

In order to deeply understand the results obtained for the

functionalized materials and, since no samples exhibiting
more than 60% of CUS functionalization were obtained experi-
mentally, we decided to perform a complementary molecular extended or trans conformations (ω3-J). The ω2-J structures are
simulation study. To this aim, three sets of six hypothetical in turn the most stable ones for all the functionalization
models of ethylenediamine-grafted CPO-27-Mg structures were degrees (see Table 2).
obtained by functionalizing different proportions of CUS in It is interesting to mention the marked diminution in the
the original MOF. This involved the functionalization of 16.6, magnitude of the energy associated with the ωi-6 models,
33.3, 50, 66.6, 83.3 and 100% of CUS (see Fig. 7). The void frac- which reflects the strong steric impediments resulting from
tion varies from ∼0.64 to ∼0.13 (non-porous) for the 16.6% to the saturation of 100% of CUS with the ethylenediamine mole-
100% of functionalization in each of the three sets of models cules. From a conformational point of view, it seems that the
(see Table 2). When studying the geometry of the grafted ethyl- geometry adopted by ethylenediamine molecules confined
enediamine molecules, it can be observed that those models inside the pores and grafted to the CUS is governed by the
coming from the initial gauche conformations (ω1-J and ω2-J) “gauche effect”.40 The corresponding values of the dihedral
are energetically more stable than those obtained by the initial angles for the ethylenediamine grafted molecules after the

Fig. 7 Models of CPO-27-Mg exhibiting different degrees of ethylenediamine grafted molecules, ω1-J (top); ω2-J (middle) and ω3-J (bottom)
models. The initial conformation of the diamine molecules is also shown (left) for each set of models.

18978 | Dalton Trans., 2015, 44, 18970–18982 This journal is © The Royal Society of Chemistry 2015
Dalton Transactions Paper

energy minimization procedure are displayed in Table S4, different adsorption behaviours in porous materials,41 which
ESI.† is expected for the ωi-J systems, with i = 1–3 and J = 2–5.
The simulated PSDs of the three sets of models (Fig. S8†) Simulations of CO2 adsorption isotherms at 298 K were per-
are quite similar for all structures; however as can be seen in formed for the eighteen hypothetical structures and also for
Fig. S9 and S10,† the shape of the pores is equivalent only for the non-functionalized CPO-27-Mg (see Fig. 8). The simulated
the ωi-1 models, while the porous architectures show clear isotherm obtained for the original material is similar to that
differences between ωi-2 and ωi-6, being more notable for the reported by A. O. Yazaydin et al.15e Simulations in MOFs con-
ωi-3 structures. Differences in pore geometry can induce taining open metal sites15e,42 give rise to underestimation of

Fig. 8 Simulated CO2 adsorption isotherms for: ω1-J (top); ω2-J (middle) and ω3-J (bottom) models.

This journal is © The Royal Society of Chemistry 2015 Dalton Trans., 2015, 44, 18970–18982 | 18979
Paper Dalton Transactions

adsorption at low pressure when compared to the experimental evident. Moreover, the simulated capacity for the ω2-3 model is
isotherms, which accounts for deficiencies in the employed higher than that simulated for the non-functionalized CPO-27-
generic force fields that do not properly represent the strong Mg MOF, both at ∼5 mbar and 150 mbar, the key pressure
interaction between the adsorbate molecules and CUS. This values for CO2 capture in spacecraft cabins and flue-gas,
situation was indeed observed when comparing our simulated respectively. This fact suggests that the 50% functionalization
and experimental isotherms. However, as the same force field of CUS in the CPO-27 structure could represent an optimized
is employed for all systems under study, it is appropriate to composition and porous architecture, which may improve the
compare the simulated adsorption performance among all performance of CO2 adsorption for flue gas mixtures and for
models, to investigate the existence of trends as a function of air purification applications. Here, we highlight that, to the
the different degrees of functionalization. best of our knowledge, there are no studies of CO2 adsorption
Since the CO2 molecule has a quadrupole moment in either in any CPO-27-M or in the expanded analogous versions
which the carbon atom is polarized positively and the oxygen that exhibit this particular degree of functionalization.
atoms are polarized negatively, the adsorbate–adsorbent inter- On the basis of the PSD (Fig. S8†), pore shape images
actions can include both atoms of the CO2 molecule and (Fig. S9 and S10†), and simulated isotherms (Fig. 8), the CO2
different parts of the MOF skeleton (i.e., OCO:→Mg or adsorption seems to be highly favoured by the formation of a
H2N:→CO2). Thus, the first one involves the non-functiona- sub-structure of regular cylindrical pores of ∼5 Å in diameter,
lized part of the framework while the second one is dependent in which there is a 1/1 ratio of free Mg-CUS/available –NH2
on the incorporation and proper localization (the real avail- groups, that is achieved for the 50% of CUS functionalization.
ability) of the –NH2 groups of the ethylenediamine grafted Linking both, the experimental and computational results,
molecules. Considering this, and taking into account the and considering the available data on the related CPO-27-M or
observed tendencies in the experiments, there are two possibi- M-MOF-74 family and on the diamine-functionalized versions,
lities. On the one hand, it could be better to functionalize the design criteria to obtain optimized CO2 adsorbents arise. On
majority of the available CUS in the structure to improve the the one hand, it seems a reasonable possibility that applying
chemical adsorbent–adsorbate attraction? On the other hand, the target composition of 50% of CUS functionalization to a
is there a particular composition of free CUS/functionalized CPO-27-M member based on a lower attractive metal ion could
CUS, having a singular pore sub-structure that can effectively give rise to reduce the differential enthalpy of adsorption and
optimize the adsorption properties? In this scenario it is thus, the energy regeneration requirements. For example,
important to highlight that the increase in the functionali- CPO-27-Co was found having the lowest enthalpy of adsorption
zation degree is accompanied by a diminution in the void pore (37 kJ mol−1 (ref. 15f )) and also was found as the most stable
volume and, as such, both factors play a key role in the adsorp- one under the humidity conditions of the flue gas mixtures.39
tion process. On the other hand, if the optimal functionalization degree is
As it is seen in Fig. 8 the incorporation of the amine groups applied to an extended CPO-27-Co compound based on a
in the pore surface affects the adsorbent–adsorbate inter- longer dicarboxylate linker (i.e. a mesoporous member of the
actions and modifies the adsorption performance at low IRMOF-74 series), the increase in framework channel size and
pressure (below 0.2–0.3 bar). The comparison of the isotherms void volume could lead to an incremented adsorption capacity
of the ωi-J systems in this pressure range shows that the along with a lower adsorbent–adsorbate interaction, helping to
adsorption capacity exhibits a complex dependence on the further reduce the differential enthalpy of adsorption and
chemical composition, pore shape and available pore volume. obtaining a good equilibrium between selectivity and regener-
The most remarkable fact is the evidence that the adsorp- ation costs.
tion is not regulated by free void volume in all systems under
study, since different behaviors are observed for systems with
almost equal available pore volume (see Table 2). Thus, for the 5. Conclusions
ω1-J and ω2-J models, there is a tendency to increase the
adsorption up to those containing 33.3% (ω1-2) or 50% (ω2-3) By combining experiments and simulations, a screening of
functionalization, with a consequent diminution when the real and hypothetical materials exhibiting different degrees of
degree of functionalization is further elevated. On the contrary, CUS functionalization in the CPO-27-Mg structure was per-
the adsorption in the ω3-J systems seems to be modulated by formed to identify the optimum features of CO2 adsorbent
the void fraction since the adsorption capacity decreases along materials for air filtering and flue-gas applications. According
with the diminution of the available pore volume when the to experimental and simulated CO2 adsorption isotherms,
functionalization degree increases. This finding suggests that recycling studies and microcalorimetry measurements, the
the shape of the pore is of great importance to the resulting functionalization degree involving 50% of CUS could be a
adsorption performance. tuned composition for CO2 adsorption in flue gas and air filter
An interesting fact is observed if we focus on the ω2-J technologies. However, considering the associated high regene-
systems (Fig. 8, middle), which were the most stable models ration energy cost, we suggest that an optimized adsorbent
for all degrees of functionalization. In the low pressure zone, a having the expanded CPO-27 structure based on a weaker
tendency in the adsorption capacity as ω2-1 < ω2-2 < ω2-3 is Lewis acid metal ion would allow the modulation of adsor-

18980 | Dalton Trans., 2015, 44, 18970–18982 This journal is © The Royal Society of Chemistry 2015
Dalton Transactions Paper

bate–adsorbent interactions as a strategy to overcome the and C. Matranga, Angew. Chem., Int. Ed., 2011, 50, 10888–
energy requirements for the regenerability step. 10892.
14 (a) J. R. Li, R. J. Kuppler and H. C. Zhou, Chem. Soc. Rev.,
2009, 38, 1477–1504; (b) J. Zhang, H. Wu, T. J. Emge and
Acknowledgements J. Li, Chem. Commun., 2010, 46, 9152–9154.
15 (a) D. Britt, D. Tranchemontagne and O. M. Yaghi, Proc.
The research leading to these results has received funding Natl. Acad. Sci. U. S. A., 2008, 105, 11623–11627;
from ANPCyT (Argentina) PICT-2012-1994 and PICT 2013-1678, (b) P. D. C. Dietzel, V. Besikiotis and R. Blom, J. Mater.
Universidad Nacional de San Luis (PROICO 2-1612 and Chem., 2009, 19, 7362–7370; (c) L. Valenzano, B. Civalleri,
PROICO 322000), CONICET (Argentina) PIP 112-201101- S. Chavan, G. T. Palomino, C. Arean and S. Bordiga, J. Phys.
00615. M. C. Bernini, G. E. Narda, A. J. Ramirez-Pastor and Chem. C, 2010, 114, 11185–11191; (d) H. Wu,
K. Sapag are members of CIC-CONICET. J. M. Simmons, G. Srinivas, W. Zhou and T. Yildirim,
J. Phys. Chem. Lett., 2010, 1, 1946–1951; (e) A. O. Yazaydin,
R. Q. Snurr, T. H. Park, K. Koh, J. Liu, M. D. LeVan,
Notes and references A. I. Benin, P. Jakubczak, M. Lanuza, D. B. Galloway,
J. J. Low and R. R. Willis, J. Am. Chem. Soc., 2009, 131,
1 K. Sumida, D. L. Rogow, J. A. Mason, T. M. McDonald, 18198–18199; (f ) S. R. Caskey, A. G. Wong-Foy and
E. D. Bloch, Z. R. Herm, T. Hyun Bae and J. R. Long, Chem. A. J. Matzger, J. Am. Chem. Soc., 2008, 130, 10870–10871;
Rev., 2012, 112, 724–781. (g) E. D. Bloch, L. J. Murray, W. L. Queen, S. Chavan,
2 DOE/NETL Carbon Dioxide Capture and Storage RD&D S. N. Maximoff, J. P. Bigi, R. Krishna, V. K. Peterson,
Roadmap; National Energy Technology Laboratory, 2010. F. Grandjean, G. J. Long, B. Smit, S. Bordiga, C. M. Brown
3 (a) H. Q. Yang, Z. H. Xu, M. H. Fan, R. Gupta, and J. R. Long, J. Am. Chem. Soc., 2011, 133, 14814–14822;
R. B. Slimane, A. E. Bland and I. Wright, J. Environ. Sci., (h) P. D. C. Dietzel, R. Blom and H. Fjellvag, Eur. J. Inorg.
2008, 20, 14–27; (b) R. Carapellucci and A. Milazzo, Proc. Chem., 2008, 23, 3624–3632.
Inst. Mech. Eng., Part A., 2003, 217, 505–517. 16 Y.-S. Bae and R. Q. Snurr, Angew. Chem., Int. Ed., 2011, 50,
4 A. U. Czaja, N. Trukhan and U. Müller, Chem. Soc. Rev., 11586–11596.
2009, 38, 1284–1293. 17 N. L. Rosi, J. Kim, M. Eddaoudi, B. L. Chen, M. OKeeffe
5 (a) D. M. D’Alessandro, B. Smit and J. R. Long, Angew. and O. M. Yaghi, J. Am. Chem. Soc., 2005, 127, 1504–
Chem., Int. Ed., 2010, 49, 6058–6082; (b) Q. Wang, J. Luo, 1518.
Z. Zhong and A. Borgna, Energy Environ. Sci., 2011, 4, 18 Z. B. Bao, L. A. Yu, Q. L. Ren, X. Y. Lu and S. G. Deng,
42–55. J. Colloid Interface Sci., 2011, 353, 549–556.
6 G. T. Rochelle, Science, 2009, 325, 1652–1654. 19 J. A. Mason, K. Sumida, Z. R. Herm, R. Krishna and
7 R. S. Haszeldine, Science, 2009, 325, 1647–1652. J. R. Long, Energy Environ. Sci., 2011, 4, 3030–3040.
8 (a) H. W. Carhart and J. K. Thompson, ACS Symposium 20 A. Das and D. M. D’Alessandro, CrystEngComm, 2015, 17,
Series, 17, American Chemical Society, Washington, DC, 706–718.
1975, p. 1; (b) P. O. Wieland, Controlling Carbon Dioxide. 21 S. V. Vaesen, Q. Guillerm, A. D. Yang, B. Wiersum,
Living Together in Space: The Design and Operation of the Life B. Marszalek, A. Gil, M. Vimont, T. Daturi, P. L. Devic,
Support Systems on the International Space Station; NASA/ P. Llewellyn, C. Serre, G. Maurin and G. De Weireld, Chem.
TM-98-206956/VOL1, NASA Marshall Space Flight Center: Commun., 2013, 49, 10082–10084.
Marshall Space Flight Center, AL, 1998, ch. 3.3.1, p. 121. 22 H. Deng, S. Grunder, K. E. Cordova, C. Valente,
9 T. M. McDonald, W. R. Lee, J. A. Mason, B. M. Wiers, H. Furukawa, M. Hmadeh, F. Gándara, A. C. Whalley,
C. S. Hong and J. R. Long, J. Am. Chem. Soc., 2012, 134, Z. Liu, S. Asahina, H. Kazumori, M. O’Keeffe, O. Terasaki,
7056–7065. J. F. Stoddart and O. M. Yaghi, Science, 2012, 336, 1018–
10 Y. Lee, D. Liu, D. Seoung, Z. Liu, C. C. Kao and T. Vogt, 1023.
J. Am. Chem. Soc., 2011, 133, 1674–1677. 23 (a) T. M. McDonald, D. M. D’Alessandro, R. Krishna and
11 G. P. Hao, W. C. Li, D. Qian, G. H. Wang, W. P. Zhang, J. R. Long, Chem. Sci., 2011, 2, 2022–2028; (b) Q. Yan,
T. Zhang, A. Q. Wang, F. Schüth, H. J. Bongard and Y. Lin, C. Kong and L. Chen, Chem. Commun., 2013, 49,
A. H. Lu, J. Am. Chem. Soc., 2011, 133, 11378–11388. 6873–6875.
12 A. Sayari and Y. Belmabkhout, J. Am. Chem. Soc., 2010, 132, 24 A. Das, P. D. Southon, M. Zhao, C. J. Kepert, A. T. Harris
6312–6314. and D. M. D’Alessandro, Dalton Trans., 2012, 41, 11739–
13 (a) B. Wang, A. P. Côté, H. Furukawa, M. O’Keeffe and 11744.
O. M. Yaghi, Nature, 2008, 453, 207–211; (b) C. Sarah, 25 S. Choi, T. Watanabe, T.-H. Bae, D. Sholl and C. W. Jones,
F. M. D. Joeri, V. B. Gino, R. Tom, G. Jorge and K. Freek, J. Phys. Chem. Lett., 2012, 3, 1136–1141.
J. Am. Chem. Soc., 2009, 131, 6326–6327; (c) K. L. Kauffman, 26 W. R. Lee, S. Y. Hwang, D. W. Ryu, K. S. Lim, S. S. Han,
J. T. Cuip, A. J. Allen, L. Espinal, W. Wong-Ng, T. D. Brown, D. Moon, J. Choide and C. S. Hong, Energy Environ. Sci.,
A. Goodman, M. P. Bernardo, R. J. Pancoast, D. Chirdon 2014, 7, 744.

This journal is © The Royal Society of Chemistry 2015 Dalton Trans., 2015, 44, 18970–18982 | 18981
Paper Dalton Transactions

27 S. Brunauer, P. H. Emmett and E. Teller, J. Am. Chem. Soc., (b) R. Krishna and J. M. van Baten, Phys. Chem. Chem.
1938, 60, 309–319. Phys., 2011, 13, 10593–10616; (c) T. Duren, F. Millange,
28 F. Rouquerol, J. Rouquerol, K. S. W. Sing, P. Llewellyn and G. Ferey, K. S. Walton and R. Q. Snurr, J. Phys. Chem. C,
G. Maurin, Adsorption by powders and porous solids: 2007, 111, 15350–15356; (d) A. L. Myers, Thermodynamics
Principles, methodology and applications, Academic Press, of Adsorption, in Chemical Thermodynamics for Industry, ed.
San Diego, CA, U.S.A, 2014. T. M. Letcher, Royal Society of Chemistry, Cambridge, UK,
29 (a) D. Fairen-Jimenez, Y. J. Colón, O. K. Farha, Y.-S. Bae, 2004, ch. 21, ISBN 0-85404-591-0.
J. T. Hupp and R. Q. Snurr, Chem. Commun., 2012, 48, 37 L. D. Gelb and K. E. Gubbins, Langmuir, 1998, 15, 305–
10496–10498; (b) N. L. Strutt, D. Fairen-Jimenez, J. Iehl, 308.
M. B. Lalonde, R. Q. Snurr, O. K. Farha, J. T. Hupp and 38 D. Yu, A. O. Yazaydin, J. R. Lane, P. D. C. Dietzel and
J. F. Stoddart, J. Am. Chem. Soc., 2012, 134(42), 17436– R. Q. Snurr, Chem. Sci., 2013, 4, 3544–3556.
17439; (c) W. Bury, D. Fairen-Jimenez, M. B. Lalonde, 39 A. C. Kizzie, A. G. Wong-Foy and A. J. Matzger, Langmuir,
R. Q. Snurr, O. K. Farha and J. T. Hupp, Chem. Mater., 2011, 27, 6368–6373.
2013, 25(5), 739–744. 40 (a) E. Juaristi, Introduction to Stereochemistry and Confor-
30 A. K. Rappe, C. J. Casewit, K. S. Colwell, W. A. Goddard and mational Analysis, Wiley & Sons, New York, 1991, ch. 18;
W. M. Skiff, J. Am. Chem. Soc., 1992, 114, 10024–10035. (b) E. Juaristi and G. Cuevas, The Anomeric Effect, CRC
31 Materials Studio v6.1, Accelrys Software Inc., San Diego, CA Press, Boca Raton, FL, 1995, ch. 4; (c) Y. Sasanuma and
92121, USA. K. Sugita, Polym. J., 2006, 38(9), 983–988; (d) C. Aleman and
32 D. Dubbeldam, S. Calero, D. E. Ellis and R. Q. Snurr, Mol. J. Puiggalí, J. Org. Chem., 1997, 62, 3076–3080;
Simul., 2015, 1–21. (e) D. J. Price, J. D. Roberts and W. L. Jorgensen, J. Am.
33 S. L. Mayo, B. D. Olafson and W. A. Goddard III, J. Phys. Chem. Soc., 1998, 120, 9672–9679; (f ) M. C. Bernini,
Chem., 1990, 94, 8897–8909. E. V. Brusau, J. C. Garro, G. Narda and E. L. Varetti, J. Mol.
34 C. D. Wick, J. M. Stubbs, N. Rai and J. I. Siepmann, J. Phys. Struct., 2008, 888, 113–123.
Chem. B, 2005, 109, 18974–18982. 41 A. I. Skoulidas and D. S. Sholl, J. Phys. Chem. A, 2003, 107,
35 (a) J. J. Potoff and J. I. Siepmann, AlChE J., 2001, 47, 1676– 10132–10141.
1682; (b) M. G. Martin and J. I. Siepmann, J. Phys. Chem. B, 42 J. Getzschmann, I. Senkovska, D. Wallacher, M. Tovar,
1998, 102, 2569. D. Fairen-Jimenez, T. Duren, J. M. van Baten, R. Krishna
36 (a) M. Arif Nadeem, A. W. Thornton, M. R. Hill and and S. Kaskel, Microporous Mesoporous Mater., 2010,
J. A. Stride, Dalton Trans., 2011, 40, 3398–3401; 136(1–3), 50–58.

18982 | Dalton Trans., 2015, 44, 18970–18982 This journal is © The Royal Society of Chemistry 2015