View Article Online / Journal Homepage / Table of Contents for this issue

390 Analyst, June, 1973, Vol. 98, p p . 390-411

Pyridylazonaphthols (PANS) and

Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

Pyridylazophenols (PAPS) as Analytical Reagents

Part 11." Spectrophotometric and Solvent-extraction Studies of
Complex Formations

BY D. BETTERIDGE AND D. JOHN?

(Chemistry Department, University College of Swansea, Swansea, Glamorgan, S A 2 8PP)

The reactions between 2-(2-pyridylazo)-l-naphthol(o-oc-PAN), 1-(2-

pyridylazo)-2-naphthol (o-P-PAN) and 2-(2-pyridy1azo)phenol (0-PAP) with
manganese(II), zinc(I1) and lanthanum(II1) and 4-(2-pyridylazo)phenol
(#-PAP) with cobalt(II), nickel(I1) and copper(I1) have been studied. The
spectrophotonietric procedure based on linear extrapolation, as used by
Sommer, has been critically evaluated. The results from the spectrophoto-
metric method have been used to predict the optimum conditions for solvent
extraction. It is shown that this procedure is a valuable approach for systems
in which hydroxy-complexes are common.

SEVERAL studies have shown that o-P-PAN is very useful for the extraction of many cations.1-3
However, the extraction systems are often more complex than some of the early work suggests,
and a t present few detailed solvent-extraction studies have been made. Preliminary studies
showed that 2-(2-pyridylazo)-l-naphthol(0-a-PAN) is often more advantageous than
1-(2-pyridylazo)-2-naphthol(o-P-PAN) and that 2-(%pyridylazo)phenol (0-PAP) and 4-(2-
pyridy1azo)phenol (@-PAP)also form extractable complexes. The conventional method of
determining equilibrium constants from solvent-extraction experiments is time consuming.
It is possible, in principle, to determine the complex formation constants by other means,
to determine the partition coefficient of the reagent and complex experimentally and to use
the values obtained to predict the extraction graphs. One practical difficulty is that the con-
stants that are obtained from solvent-extraction data relate to the aqueous phase saturated
with the organic solvent and as, almost by definition, the complex is only sparingly soluble
in water, the equilibrium constants must be determined in some other medium, e.g., 1 1 +
dioxan - water, so that they cannot be used directly for predicting extraction equilibria.
Spectrophotometric methods can partially overcome this difficulty because solutions can be
used that are so dilute that the complex can be maintained with such a small proportion of
organic constituent that the values of the equilibrium constants are close enough to those
obtained by solvent-extraction procedures to be interchanged. However, most spectrophoto-
metric methods are based on Job's method, which can result in misleading or erroneous results,
or both,4 and the study of a system over a wide range of conditions with these methods
is extremely tedious.
Recently, Sommer and co-workers5-11 have demonstrated that spectrophotometric
procedures based upon linear extrapolation, when used with care, can be used most advan-
tageously to study the complex equilibria of systems based on compounds analogous to the
compounds discussed below. A practical advantage of their approach is that it is based
upon the analysis of pH - absorbance curves so that no superfluous information is gathered
and basic data are interpreted fully.
In this paper, we assess this procedure by analysing several systems, for some of which
results are already available for comparison, and show how it can be used for predicting
extraction curves with an accuracy that is acceptable to the analyst.
EXPERIMENTAL
REAGENTS-
Salts and solvents of analytical-reagent grade purity or better were used throughout.
* For Part I of this series, see p. 377. Parts I11 and I V will appear in the July issue.
t Present address: BP Chemicals (U.K.) Ltd., Llandarcy, Swansea.
@ SAC and the authors.
 View Article Online

BETTERIDGE AND JOHN 391

1-(2-Pyridylaz0)-2-naphthol(0-p-PAN).
2- (2-P yridy1azo)- 1-naphthol (o-a-PA N ).
2- (2-Pyridylazo)phenol(0-PA P).
Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

4-(2-Pyridylazo)phenol (p-PAP).
The above four reagent solutions were prepared and purified as described in Part I.
Solutions in absolute or aqueous ethanol were prepared for spectrophotometric studies and
in carbon tetrachloride or chloroform for solvent extraction. The concentrations are given
in the procedures. The solutions were taken to be standard and the validity of this assumption
was checked occasionally by spectrophotometric titration of a standardised metal-ion solution.
Metal-ion solutions-Solutions of manganese(II), zinc(II), nickel(II), cobalt(I1) and lan-
thanum(11)were prepared and standardised titrimetrically with EDTA by standard procedures.
Bufer solutions-Standard chloride, phthalate, phosphate, borate and hydroxide buffer
solutions to cover the pH range from 0 to 12 were used.
Sodium perchlorate-Solid sodium perchlorate was used in order to maintain an ionic
strength of 0.10 & 0.01.
Radioisotopes-Manganese44 and zinc-65 were obtained from the Radiochemical Centre,
Amersham. They were diluted and mixed with carrier manganese(I1) and zinc(I1) so that
solutions of known concentration and radioactivity were obtained.
APPARATUS-
pH meter-Radiometer, Model M4C.
Spectro$hotometers-Unicam SP500, SP600 and SP800 and Cary, Model 16, instruments
were used as appropriate. Absorbance values used for the calculation of equilibrium constants
were obtained on fixed-wavelength instruments. One-centimetre cells were used.
Radioactivity-A 1-inch well sodium chloride (thallium-activated) crystal connected to
a photomultiplier and a 1DL 1700 scaler was used to measure the radioactivity. Samples
of constant volume were taken so as to ensure constant geometry and sufficient counts were
taken so as to ensure a statistical error of not more than 1 per cent. on all except the very
low count-rates, when an error of 10 per cent. was accepted.
Computer-An IBM 1600 computer was used for the calculation of constants based on the
spectrophotometric data and the statistical analyses of them. The standard program for
linear regression in two variables 1620/6.0.27from the 1620 General Program Library was used.
PROCEDURES-
Spectrophotometric determination of acid-dissociation constants-A 10-ml aliquot of a
solution of o-a-PAN (1 x M) in absolute ethanol or o-PAP (2.5 x M) or $-PAP
(1.25 x 1 0 - 4 ~ )in 12.5 per cent. aqueous ethanol was placed in a 25-ml calibrated flask
and sufficient solid sodium perchlorate to maintain an ionic strength of 0.10 & 0.01 was added.
The solution was then made up to the mark with buffer solution. The solution was mixed well,
the spectrum recorded and the absorbance and pH were measured. A pH - absorbance curve
was plotted for each of the reagents a t suitable wavelengths and the pK, values were calculated
from the inflection point as determined from a graph of the differential AAIApH against pH.
The pK, values were also calculated from the same pH - absorbance curve by the procedure
described below.
Spectrophotometric determination of formation constants-pH - absorbance curves were
obtained with both the metal ion and the reagent in excess. The general procedure was to
place 10 ml of ethanolic solution, 10 ml of metal-ion solution and 5 ml of buffer in a 25-ml
calibrated flask that contained sufficient solid sodium perchlorate to maintain the ionic
-+
strength of the final solution a t 0-10 0.01. These solutions were mixed thoroughly and
the colour was allowed to develop to a maximum (a few minutes). The absorbance remained
constant for the period of the measurement and showed little change after 24 hours. The
spectra were recorded, the absorbances were determined a t suitable wavelengths and the
pH values were measured. The measurements were carried out a t room temperature. The
method of continuous variations12 and the mole-ratio method13 were also applied to some
systems.
Solvent extraction and determination of acid-dissociation constants-The reagent solutions
were made up in carbon tetrachloride (for o-a-PAN, o-p-PAN and o-PAP) or chloroform
(for $-PAP) to concentrations of 1.0 x 1.0 x 2.0 x and 5.0 x 1 0 - 4 ~ ~
 View Article Online

392 BETTERIDGE AND JOHN : PYRIDYLAZONAPHTHOLS AND [Analyst, Vol. 98

respectively (@-PAPis insufficiently soluble in carbon tetrachloride for the distribution ratio
to be measured accurately). A 5-ml aliquot of reagent solution was placed in a vial and
Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

60 ml of buffered aqueous phase of ionic strength 0.10 & 0.01 were added. The solutions
were shaken overnight in a box shaker that was maintained thermostatically a t 24 &- 1 "C. The
phases were allowed to separate and the absorbance of the organic phase was measured with
the Unicam SP600 instrument and that of the aqueous phase, after adjustment of the pH to
zero with concentrated hydrochloric acid, with the Cary 16 instrument. Calibration graphs
were used to convert the absorbance values into concentration values.
The distribution ratio of the reagent, D,was calculated from
D = (CT- c W ) v W / c W v O
for D > 1, or
D = CoVw/CwV,
for D < 1, where CT is the total concentration of the reagent, C, and C, are the concen-
trations of the reagent in the organic and aqueous phase, respectively, and V , and Vw are
the volumes of the organic and aqueous phase, respectively.
Log D was plotted against pH and the partition coefficient of the reagent, K D R , and
acid-dissociation constants, Kal and K,,, were calculated from1*
+ +
D = KDR{ [HI/KaI 1 KaJ [HI1
Solvent extraction of metal ions-Reagent solutions were 5.0 x M in carbon tetra-
chloride. Stock metal-ion solutions of 5 x M were diluted to a suitable concentration
when the extraction was followed spectrophotometrically. When the extraction was followed
radiochemically, a solution that was as dilute as the specific activity of the isotope would
allow was used and a 0.2 M solution of sodium perchlorate was prepared. Equal volumes of
organic and aqueous phase were used, the latter consisting of 2 nil of metal-ion solution,
2 ml of buffer solution and 4 ml of sodium perchlorate solution. The solutions were placed
in a vial and shaken overnight, the layers were separated and the radioactivity of each phase
was measured. The total activity in the two phases was computed and if it was less than
80 per cent. of the total activity in the vial, the results were discarded.

RESULTS AND DISCUSSION

The following symbols are used-
HR, H,R+, R- the neutral, protonated and ionised forms of the reagent, respectively
CCXI the total analytical concentration of species x
the molar absorptivity of species x
the total absorbance
IMlo/lMlw,the distribution ratio
[HR],/ [HR], or [MRn]o/[MRnIw, the partition coefficient of the reagent
and extractable complex, respectively
[HR] [H+]/[H2R+], the first acid-dissociation constant
[R-] [H+]/[HR], the second acid-dissociation constant
[MRI/ CMI [RI
[MR,]/[M] [R-I2, the over-all formation constant of MR2
WR2I [H+l2/[M1WRI2
WR2OHl / WR2I PHI
a, [H,,R]~%~H,-,R]
x=o
P [MI/CMT0J
f the fractional extent of completion of the reaction as written.
Charges are omitted unless possible ambiguity would arise from their absence,
REAGENT EQUILIBRIA IN THE ABSENCE OF METAL IONS
SPECTROPHOTOMETRIC STUDIES-
Principles of graphical analysis of absorbnnce versus pH curves-This method was re-
examined by Sommer5 in 1964 and has since been applied to the study of a number of
reagents.6~7 The principles involved are summarised below.
 View Article Online

June, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I1 393

The method involves a detailed algebraic study of the equilibria of the reagent, and in
the present study these are taken to be:
Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

+ R- + H+
Kaz
HR (Equilibrium 1)
Ka,
H,R+ + H R + H+ (Equilibrium 2)
On the initial assumption that these equilibria represent the true reagent equilibria,
a series of linear algebraic transformations can be derived. Experimental values of absorbance
and pH are introduced into these transformations, which then provide values of the acid-
dissociation constants, Kal and Ka,, and also molar absorptivity values of individual reagent
species. An observed lack of linea.rity in these transformations indicates that the assumed
equilibria are incorrect.
In Equilibrium 1, if it is assumed that only the neutral (HR) and anionic (R-) forms
of the reagent are present, it follows that the total reagent concentration, CR, is given by
+
C R = [HRI

The total absorbance, A , is given by

= [RI {e+
LR1
l} .. .. .. ' (1)

A = EHR [HRI + ER [R-I

.. .. .. ..
Substitute for [R] from equation (1) and re-arrange to give

Multiply throughout by l/eR and re-arrange to give

- (A - €HRCR) .. ..
A = -€R+
CR Ka,ERA .. (1)
A graph of CR/A zleysus [H]A is usually a straight line of intercept l/en, as A EHRCR.
Values of the acid-dissociation constant, K a , , can be obtained by calculation by using corre-
>
sponding values of absorbance and pH.
Alternatively, it can be argued that
r
.. .. .. ..
and
A = [HR] [EHR + 6 3 1 .. .. .. ..
[HI
which give

This equation, on multiplication by and subsequent re-arrangement, gives

l/gHR

CR - 1 Ka, ( A - CRER). 1 .. ..
A -EG + EHR A[H] "

Equation (IA) should also yield a straight line when CR/A is plotted against l/A[H].
The intercept of this line represents the reciprocal of the molar absorptivity, EHR. Values
of Ka, can be calculated from corresponding values of absorbance and pH.
Alternatively, a value of eR, which can be found from other experiments, can be substi-
tuted and &/A plotted against ( A - CR~R)/A[H].A similar procedure can be adopted to
solve graphically all of the subsequent transformations, where the linearity is not obvious.
 View Article Online

394 BETTERIDGE AND JOHN : PYRIDYLAZONAPHTHOLS AND [Analyst, vol. 98

These equations are therefore derived simply from consideration of mass balance and
equilibrium constants, with the one important algebraic manipulation being effected by the
Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

multiplet ~ / E H Ror ~ / E R .
In equilibrium 2, if it is assumed that [R-] is negligible compared with [H,R] and [HR],
then
and
CR = LHR1 + LHzR1
+
A = EHR[HR] EH,R[H&]
By using these equations and proceeding in the same manner as before, it can be
shown that

and

For equation (11),the graph of CE/A veysus l/A[H] should be a straight line. Similarly,
the graph of CR / A versm [H]/A for equation (IIA) should be linear. Values of the first acid-
dissociation constant, &, and of the molar absorptivities, EHR and c H 2 ~can , be calculated
from these equations.
Absorbance curves and results of calculations-The absorbance veysus pH curves are shown
in Fig. 1. The wavelength of maximum absorption for the various species and isosbestic
points are given in Table I. The results obtained for each transformation are summarised
in Table 11. The transformations are represented as straight-line equations of the form
y = mx + c, where y = &/A and x = 1/A [HI or [H]/A multiplied by a power factor to
keep x and y mainly within the range 0.1 to 10.
The values obtained are in good agreement with those obtained from the inflection point
of the pH - absorbance curve. It is slightly advantageous that the molar absorptivity is
calculated simultaneously and that a check on the nature of the equilibrium is provided.
If two protons were to be released simultaneously, for example, the observed transformations
would be curved or extremely ill defined. The least-squares analysis permits the precision
of the result to be calculated. It was found that because the method is based on an extrapola-
tion, it is very sensitive to the choice of points used for the calculation. Inevitably, a t one
end of the part of the pH - absorbance curve that is being used for analysis, the basic assump-
tion that one species is of negligible concentration compared with the others becomes less valid.
A point from this part of the curve will therefore be “bad” and because it will be at the
end of the transformation it can exert a disproportionate effect on the least-squares analysis.
Several checks are possible: (i) the point may be so “bad” that it falls outside the 95 per cent.
confidence limits and can be rejected; (ii) the point can be discarded and the analysis carried
out again, and if the result is then the same the point can be accepted and if it is markedly
different it can be rejected; and (iii) a different transformation can be used and the results
compared. Rejection by the second of these procedures is not entirely satisfactory and care
was taken to reject not more than one result and to confirm, by examination of the experi-
mental pH - absorbance curves, that this result was a marginal value. Checks of (i) and (iii)
were always carried out. A practical disadvantage of the method, therefore, is that a larger
number of “good” experimental points must be obtained than is necessary for the simpler
spectrophotometric procedures.
SOLVENT-EXTRACTION STUDIES-
The experimental curves of log D versus pH were very similar to that already published
for o-a-PAN.14 All of the curves indicated that a protonated species of reagent was formed
under acidic conditions and an anionic species under alkaline conditions. The pK values
obtained are given in Table 111. The logarithm of the partition coefficients for o-a-PAN,
o-p-PAN, o-PAP and +-PAP are 5.20, 5.00, 3.75 and 2.53, respectively. It was found that
although the partition coefficients for 0-a-PAN and o-P-PAN were an order of magnitude
greater than those reported earlier, the acid-dissociation constants were in reasonable agree-
ment. This agreement is surprising, as the values of the partition coefficients are used in
the calculation of the acid-dissociation constants. With such high partition coefficients, the
 View Article Online

June, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I1 395

Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

1-2

1.0

08

0.6

0-4

0.2

2.0

1
0.4 - 1
1
1

1
1 1 1 1 1 1 1 1 1 1
0

(C)

Ka
0.9 - HzR' + 1
HR + H" HR i
R-+ H+
pKaq= 2.70
0.8 -

0.7 -

PH
Fig. 1. Absorbance versus p H graphs: (u)5.0 x M $-PAP in 6 per cent. aqueous ethanol;
(b) 1.0 x lo-* M o-PAP in 5 per cent. aqueous ethanol; and (c) 4.0 x M o-a-PAN in 40 per
cent. aqueous ethanol
 Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

I
TABLE
MAXIMA,
ABSORBANCE MOLAR ABSORPTIVITIES AND ISOSBESTIC POINTS
H,R+ HR R- Isosbestic point
A
r \
Reagent wnm E )czlnm E &bm E H,R 4 HR+ + H+ HR + R- + H+
P-PAP 402 2-38 x 104 358 1.89 x 104 442 2-10 x 1 0 4 373 382
O-PAP 352 1.62 x 1 0 4 328 1-56 x lo4 480 9-16 x 103 337 410
o-a-PAN 365 1.63 x 104 486 1.62 x 1 0 4 520 2.08 x lo4 476 -

*Z
U
TABLEI1 c-c
0
MOLARABSORPTIVITIES AND ACID-DISSOCIATION CONSTANTS OF REAGENTS FROM LEAST-SQUARES FIT OF TRANSFORMATIONS (I), (IA),
(11) AND (IIA) E
Correlation Standard error of Molar Acid-dissociation
Transformation Reagent y = w + c coefficient absorptivity constant
ER PKa,
I p-PAP y = 0.101x + 4.756 0.992 0.0046 0.183 2.10 x 104 7-66 f 0.02
0-PAP y = 0 . 0 2 3 ~ 1.003
+ 0.990 0*0010 0.057 9-16 x 103 8.77 f 0-02
o-a-PAN ~ 4.808
J.J = 0 * 0 8 1+ 0.988 0.0045 0.048 2.08 x 10' 10.17 & 0.02
E HR PKa,
IA $-PAP y = 0 . 2 6 4 ~ 5.056
+ 0.979 0.0192 0.198 1.98 x 104 7.78 f 0.02
0-PAP y = 0 . 0 1 8 ~ 6.638
-+ 0.975 0.0014 0.091 1-51 x 104 8.92 f 0.04
- - - I - -
o-E-PAN
EHIR PKa,
I1 P-PAP y = 0 . 1 8 3 ~ 4.203
+ 0.980 0.0142 0.057 2.38 x 1 0 4 3.00 f 0.03 *Z
O-PAP y = 0 . 0 3 0 ~$- 6.189 0.992 0-0015 0.054 1.62 x 10' 2.72 f 0.06 U
o-a-PAN y = 0 . 0 5 5 ~ 6.154
+ 0-993 0.0020 0.018 1.63 x lo4 2-73 f 0.03
€HR Ph'a,
IIA P-PAP y = 0.044~ + 5.561 0.978 0.0033 0.150 1-80 x 104 2.94 f 0.07 n
0-PAP y = 0.041%+ 6.621 0.989 0~0020 0.035 1.51 x lo1 2.64 f 0.05 b
o-a-PAN y = 0 . 0 9 9 ~+ 6.182 0.990 0.0049 0-023 1.62 x 1 0 4 2.60 f 0.05
x
No calculations were carried out for o-a-PAN by using transformation (IIX) because of the proximity of the Amax. values of the reagent species HR $
and R-.
View Article Online
 View Article Online

June, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I1 397

presence of a small amount (e.g., 0.1 per cent.) of a coloured impurity that is less extractable
than the reagent can profoundly affect the experimentally determined value of the partition
Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

coefficient, but would scarcely affect the log D versus pH curve when the value of the dis-
tribution ratio is less than 100. This effect would result in the experimentally determined
pKa, value being greater than and the pKa, value being less than the true values. As there
is an internal consistency in both sets of studies, which should prevent miscalculation, we
can offer no explanation for the discrepancy unless there is some slight difference between
the samples of reagent-grade carbon tetrachloride used in the two studies. The values
obtained in this study were used throughout this work.

TABLE I11
ACID-DISSOCIATION
CONSTANTS AT AN IONIC STRENGTH OF 0.10
Reagent Method* Medium PK, PKa, Reference
O- p-PAN SP. 20% aqueous dioxan 1.9 12.2 15
Pot. 50% aqueous dioxan <2 12.3 16
S.E. - 2-9 11.5 17
S.E. Chloroform - water 2.9 11.2 18
o-E-PAN Pot. 50% aqueous dioxan 2-54 10.74 14
SP. Water 3.0 9.1 14
S.E. Water - carbon tetrachloride 3.1 9.5 14
S.E. Water - carbon tetrachloride 2.90 9.63 This work
SP- 50% aqueous methanol 2.29 10.00 19
SP. 40% aqueous ethanol 2.67 10.17 This work, graphical
SP. 40% aqueous ethanol 2.70 10.23 This work, conventional
o-PAP SP. 50% aqueous methanol 1.85 9.42 20
SP* 5 yo aqueous ethanol 2.68 8.84 This work, graphical
SP. 5y0 aqueous ethanol 2.68 8.79 This work, conventional
S.E. Water - carbon tetrachloride 2.68 8.68 This work
#-PAP SP. 50% aqueous methanol 2-47 8-20 20
SP. 5% aqueous ethanol 2-97 7-72 This work, graphical
SP. 5% aqueous ethanol 2.86 7.76 This work, conventional
S.E. Water - carbon tetrachloride 3-58 7-55 This work
* Sp. = spectrophotometric; Pot. = potentiomctric; S.E. = solvcnt extraction.

COMPLEX FORMATION EQUILIBRIA

SPECTROPHOTONETRIC
STUDIES-
The graphical method that is used for the determination of acid-dissociation constants
can be applied in the study of the formation 01 complexes.
The method consists in first postulating the chelation reaction and the species formed
over a particular pH range, then algebraically deriving transformations that would necessarily
be followed if the postulates were correct. Typically, these transformations are straight lines,
and false assumptions are rcadily detected by the presence of curvature or random scatter
when experimental values are substituted into the transformations. All parts of the simple
absorbance - PI-I curve can be subjected to such analysis, so that the various conditions of
chelation can be deduced over a wide pH range. Normally, absorbance-pH curves are
obtained in the presence of excess of reagent and excess of metal ions, so that the absorbance
can be expressed generally as
A =f(PH)cma,cx
A =.WM)~H,C~
and
A = ~ ( C Rp> ~cM
,
Sommer and co-workers have summarised a large number of basic equations for a variety
of chelation reactions in the presence of various concentrations of reagent and metal ions,
and have since applied the method successfully to different systems. These systems include
the reactions of 4-(2-thiazolylazo)resorcinol(TAR) and 4-(2-pyridylazo)resorcinol(PAR) with
thallium and uranyl ions.8-f0 The reactions of copper(II), lead(II), cadmiuni(II), zinc(I1)
and bismuth(II1) with TAR and of the lanthanides with PAR have also been reported.6911
 View Article Online

398 BETTERIDGE AND JOHN : PYRIDYLAZONAPHTHOLS AND [.4nd,St, VOl. 98

We have applied the method to the reactions of zinc(II), manganese(II), cobalt(II),
nickel(II),lanthanum(III), copper(I1) and titanium(1V) with o-a-PAN, o-b-PAN, o-PAP and
Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

p - P A P and compared the results when possible with those obtained by different methods.
Most of the findings are given below but some will be dealt with in subsequent papers.
GRAPHICALANALYSIS OF ABSORBANCE CURVES-
The general approach can be illustrated by the derivation of a few basic equations.
One common reaction is chelate formation when the reagent is predominantly in the neutral
form. The reaction can be represented as
*KYR1
M + 2 H R + MR2+2H .. .. .. * * (3)
This reaction can be carried out with either metal ions or the reagent in excess and the un-
reacted reagent may or may not contribute to the absorbance a t the wavelength of maximum
absorbance of the chelate. The effects of these variations are considered below.
I n the presence of excess of metal ions--In this instance, [MI = Cy,the total metal-ion
concentration, and if it is assumed that E H R is negligible a t the wavelength being used, then
A = EMR, [MRJ .. .. .. * - (4)
Furthermore, if it is assumed that the concentration of the intermediate chelate species,
MR, is negligible, then
C R = [HR] + 2[MR,] . . .. .. - (5)
Re-arrange equation (5) and substitute for [MR,] :
[HR] = CREMR, - ZA
EMR,
Divide equation (4) by equation (6) and substitute for [MR,]/[HR] by introducing the
reaction constant, so that

or

Multiply through by ~ / A E ~ and

I R , re-arrange :

When the reagent makes a contribution to the absorbance, the expression for the total
absorbance requires an additional term :
A = EMR,[M&] EHR[HR] 6 - .. . . . . (4a)
The steps outlined above then give
CR - -
-- 1 . .
( A - C R E M K ) [HI2 . . (IIIA)
A - EMR, + " K M R CMEMR, A [HR] ' '

In the presence of excess of reagent-Under these conditions, [HR] = C,, the total reagent
concentration, and if the reagent absorption is negligible, then
A
[MR,] =- .. .. .. .. ' * (7)
EMR,
and if [MR] is negligible compared with [MI and [MR,], then

so that
C M = [RI] [MR,] + .. .. .. -- (8)

Divide equation (7) by equation (9), substitute for [MR,]/[M] and re-arrange, to give
 View Article Online

June, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I1 399

Multiply through by ~ / A E M Rand
, re-arrange, so that
1 ~ 1 - . .. .. ..
A-EMR, *KMR,c~REMR,[HI2 ..
CId - - + (IV)
Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

When the reagent absorbance is not negligible, an extra term can be introduced. In
practice, we found that either the reagent absorbance was negligible or it represented such
a large contribution that internal compensations on the spectrophotometer had to be made.
Another important type of reaction is the hydrolysis reaction
KOH
MR2 + OH- + MR,OH- .. .. .. . . (10)
If it is assumed that the hydroxy-complex makes no contribution to the absorption, then in
the presence of both excess of reagent and excess of metal ions the total absorbance, A , is
given by
A = E M R ~[MR2I
For excess of metal ions, the total reagent concentration, CR,is given by
C R = 2[MR,] + 2[MR,OH]

For excess of reagent, the total metal-ion concentration, CM,is given by

C M = [MI +
In the presence of excess of reagent:
PR2I + [MR,OHI

or
.. .. -- (V)
The derivation of equations in the presence of excess of metal ions is academic because.
in most systems, the metal hydroxide would be formed preferentially and be precipitated:
The transformations for these and other systems are given in Table IV; some are in
logarithmic form because this form is more convenient for use in calculations. All of the
transformations can be expressed as a linear function, although some contain two variables,
[HI and [HR], in one term. I n these instances, as for example transformation (111),EMR, was
calculated from another transformation, (IV), or by simpler conventional means, and this
was used to calculate [HR] by means of equation (6). The calculation procedure is very
similar to that used for the acid-dissociation constants and the same checks were applied.
However, even greater care is necessary because although a pH - absorbance curve reflects the
course of several reactions, which allows these reactions to be detected and studied, it is
easier to introduce “bad” points into the calculations. A further check was therefore always
carried out. The values of the calculated constants were used to derive a pH - absorbance
curve and this curve was then compared with the experimental curve.
Zinc(l1) chelates of o-a-PAN, o-P-PAN and o-PAP-pH - absorbance curves obtained
for the chelates of each reagent in the presence of excess of metal ions and of reagent are
shown in Fig. 2. The absorption maxima of the zinc(I1) chelate with o-a-PAN, o-P-PAN
and o-PAP occurred a t 548 and 590, 514 and 550, and 530nm, respectively, the chelates
of the naphthol derivatives each having two maxima, and the pH - absorbance curves were
obtained a t 590, 550 and 530 nm, respectively. These curves indicate that each chelate has
a stoicheiometry of metal to ligand of 1 :2. At the upper limits of pH, the absorbance values
became inconsistent and no smooth curve could be drawn. Hence, over the pH range 6 to 8
in which the chelate is being formed, the basic reaction would appear to be
Zn + 2HR + ZnR, + 2H
Transformations (111)and (IV) were therefore used for the reactions with o-a-PAN and o-PAP
and transformation (IIIA)was used for o-P-PAN, as the reagent contributed to the absorbance.
The results are given in Table V, which shows that there is good agreement between the molar
absorptivities and stability constants determined in the presence of both excess of reagent and
excess of metal ions. The agreement between these values, together with the linearity of
transformation (IIIA) and hence of transformation (VIA), confirm that the chelation reaction
 Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

IV
TABLE
BASICCOMPLEX FORMATION REACTIONS AND LINEA.R TRANSFORMATIONS W

Principal reaction Condition No. 24

M
M Excess of metal ions 111 w
Y
+ ZHR+MR,+ZH
U
0
IIIA M

Excess of reagent IV

M + 2R+MR, Excess of metal ions VI

VIA
Excess of reagent VII

MR, + OH+MR,OH Excess of reagent V

M + HRsMR + H Excess of metal ions VIII

Excess of reagent IX

M + R+MR Excess of metal ions X

Excess of reagent XI

n
b
View Article Online

W
M,
 View Article Online

June, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I1 401

suggested is correct and that the assumptions made in the derivation of the equations are
valid. Each chelate investigated therefore has a stoicheiometry of metal to ligand of 1:2,
Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

and is involved in a chelation reaction of the type

PH
Fig. 2. Absorbance zwsw p H graphs for zinc(I1) chelates:
( a ) in the presence of excess of metal ions-
cM/M CR/M X/nm
0-a-PAN .. .. 2.0 x 1 0 - 4 4.0 x 10-6 590
0-P-PAN .. .. 2.0 x 1 0 - 4 5.0 x 10-5 550
O-PAP *. .. 2.0 x 10-4 6.0 x 10-6 530
(b) in the presence of excess of reagent-
cbf/M CR/M X/nm
0-M-PAN .. .. 2.0 x 10-6 2.0 x 10-4 590
o-P-PAN .. .. 2.0 x 10-6 2.0 x 1 0 - 4 550
0-PAP .. .. 2.0 x 10-6 2-0 x 10-4 630

The values of the molar absorptivities and stability constants determined in this study
were found to be consistent with the results of other
Mangnnese(l1) chelates of o-a-PAN , o-P-PAN and o-PAY-The absorbance maxima of
the nianganese(I1) chelates of o-a-PAN, o-P-PAN and o-PAP occurred a t 646 and 686, 616

1 .o

1
(a1

12
5 6 7 8 9 1 0 1 1 1 2
PH
Fig. 3. Absorbance versus pH graphs for manganese(I1) chelates:
( a ) in the presence of excess of metal ions-
Cm/hl CR/M h/nm
0-LX-PAN .. .. 2.0 x 10-4 4.0 x 10-6 686
o-P-PAN .. .. 2.0 x 10-4 4.0 x 10-6 546
0-PAP .. .. .. 2.0 x 10-4 4.0 x 10-6 630
(b) in the presence of excess of reagent-
car/M Cn/M hlnm
0-LX-PAN .. .. 2.0 x 2.0 x 10-4 686
o - p - ~
0-PAP ..
~ ..~
.. ..
.. 2.0 x
2.0 x
10-6
10-6
10-5
2.0 x 10-4
2.0 x 10-4
546
530
 Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

TABLEV
RESULTSFOR THE CHELATES OF ZINC(II)
Standard error of
Correlation
Condition Reagent y=mt:+c coefficient m-eat
slope, m EMIR, x log KMR,
Excess of metal ions o-a-PAN y = 0.087~ + 4.515 1.000 0*0003 0-015 4-43 19.13 f 0.07
O- p-PAN y = 0 . 4 5 3 ~+ 4.208 0.999 0.0055 0*108 4.75 21.63 f.0.08
0-PAP y = 0-804X+ 8.349 1.000 0.0052 0.096 2-40 15-52 f 0.09
Excess of reagent o-a-PAN y = 0 . 0 6 6 ~+ 2.212 1.000 0.0002 0.016 4.52 19.12 f 0.07 ?-
O- 8-PAN y = 0.037% + 2.222 1.000 0-0002 0.015 4-50 21.45 f 0.08 z
O-PAP y = 0 . 0 7 6 ~+ 4.165 1.000 0.0003 0.025 2.40 15-48 f 0.09 U
pKa, = 10.20 (o-a-PAN); 12-20 (O-F-PAN); 8-80 (0-PAP). 0
X
Z

VI
TABLE
RESULTSFOR CHELATES OF MANGANESE(II)
Standard error of
Correlation r
Condition Reagent y=m%+c coefficient slope, m estimate EYR, x log KMR,

Excess of metal ions o-a-PAN y = 0.874~ + 4.823 1.000 0.0006 0-029 4-15 13-54 f 0.06
0-8-PAN y = 0.619~ + 4.945 1.000 0.0006 0.011 4.04 15-77 f 0.06
0-PAP y = 0.504~ + 9.201 1.000 0.0012 0.059 2-17 10.52 f 0.02
Excess of reagent o-a-PAN y = 0.086~ + 2.443 1.000 0-0002 0-015 4.09 13.27 f 0.05
0-8-PAN y = 0.045~ + 2.455 1.000 0-0001 0.009 4.07 15-69 f 0.08 n
0-PAP y = 0.265% + 4.808 0.999 0-0032 0-132 2-08 10-45 f 0.05
pKa, = 10.20 (0-a-PAN) ; 12.20 (0-8-PAN) ; 8-80 (o-PAP).
?
p
View Article Online
 View Article Online

June, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I1 403

and 546, and 530 nm, respectively, the chelates of the naphthol derivative each having two
maxima, and the pH - absorbance curves (Fig. 3) were obtained a t 586, 546 and 530 nm,
Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

respectively. Over the pH range 6 to 9 the curves are similar in appearance to those found
for the zinc(I1) chelates and the same analysis was applied. The results are shown in Table VI.
At higher pH values, the absorbance of the neutral chelate was found to decrease with
increase in pH. The decreased absorbance was usually accompanied by the formation of a
precipitate, which was presumed to be a hydroxy-form of the chelate. With o-a-PAN and
o-PAP, it was not found possible to correlate absorbance with pH, and it was observed that
the absorbance tended to zero on standing the solution. The manganese(I1) chelate of
o-P-PAN, however, gave reasonably smooth absorbance curves in the presence of both
excess of reagent and excess of metal ions. The values tended to vary on standing the
solution, but it proved possible to use them in transformation (V) (Fig. 4). From these
values, K M R , ( O H ) was found to be The linearity suggests that the hydrolysis reaction
that was proposed is reasonable, but because of the variation noted above, we have some
reservations about these conclusions.

0.2- ( a )
0-
-0.2 -
-0.4 -

, l (b)
1 2 . o 0.2 ;l.l

0
-0.2
-0.4
-0.6
9.7 9.8 9-9 10.0 10.1 10.2
PH
Fig. 4. Transformation (V) for the hydrolysis reaction of
Mn(o-P-PAN),: (a) excess of reagent; and (b) excess of metal.
Conditions as in Fig. 3

Lalzthanum(III) chelates of o-a-PAN, o-P-PAN and o-PAP-The chelation reactions of

lanthanum(111) with o-a-PAN, o-P-PAN and 0-PAP were investigated and absorbance curves
in the presence of both excess of reagent and excess of metal ions were obtained.
Absorbance measurements were made a t 550 and 514 nm, these wavelengths correspond-
ing to the absorbance maxima of the o-a-PAN and o-PAP chelates, respectively. The
absorbance curves are given in Fig. 5 and show that the extent of chelation is greatly affected
by hydrolysis of the lanthanum species, particularly when the pH is above 7.5. The occurrence
of a sharp hydrolysis effect above this pH may be expected because of the equilibrium
position of lanthanum hydroxide, indicated by21
Log [La3+] = 23-02 - 3pH
When the reaction of o-p-PAN with lanthanum(II1) was investigated] only a slight
colour change occurred. As the formation of the o-p-PAN chelate may be expected to occur
at lower acidities than those for the a-PAN chelate, this lack of reaction may be due to prefer-
ential hydrolysis of the lanthanum(II1).
Unlike the absorbance curves for the manganese(11) and zinc(I1) chelates, those shown
in Fig. 5 give no direct indication of stoicheiometry. However, as the absorbance values in
the presence of both excess of reagent and excess of metal ions are similar, the formation
of 1 : 1 chelates can be anticipated. The neutral reagent species, HR, predominates over
the pH range of chelation and so the suggested chelation reaction is
KMR
La3++HR + LaR2++H+
The validity of this postulate was tested by using transformation (IX) (Fig. 6).
 View Article Online

404 BETTERIDGE AND JOHN : PYRIDYLAZONAPHTHOLS AND [Analyst,Vol. 98

Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

Fig. 6 . Comparison of experimental and theoretical absorbance

curves for LaIIl(o-a-PAN) [upper graphs: ( a ) cxccss of metal; and
(b) excess of reagent] and LaIII(o-PAY) [lower graphs: (c) excess of
metal; and (d) excess of reagent]. Solid lines, experirncntal; and brokcn
lines, theoretical. Concentrations : (u)CR 5.0 x M , CX 5.0 x M;
(b) CR 4.0 x 1 0 - 4 M , CM 4.0 x 10-5 M : ( c ) CR 5.0 x 1 0 - 6 M , CM 5.0 x 10-4 M ;
( d ) CR 5.0 x 1 0 - 4 M, CM 5.0 x 10-5 M

The formation of such an LaR2+species is in agreement with the observation of Sommer

and Novotna,ll who have reported the reactions of lanthanum with PAR. These workers
also found that a 1: 1 chelate was formed and that severe hydrolysis interfered in the accurate
determination of absorbances a t higher pH.

I l l I l l I l l
0 10 20 30 40 50 60 70 80 90 100
[ H I 2 x lo--*
Fig. 6. Graphs of transformation (IX) in the presence of
excess of lanthanum(II1): (a) o-a-PAN; and (b) o-PAP. CE
6.0 x M and CM 5-0x M
 View Article Online

June, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I1 405

Agreement between the constants listed in Table VII is not quite as good as that observed
for the manganese(I1) and zinc(I1) chelates, probably because of the ease of hydrolysis of
Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

lanthanum species. In order to determine precisely the hydrolysis effect, and to verify
further the proposed chelation reactions, theoretical absorbance curves were determined by
using the computed complex formation constants and known values for the hydrolysis of
lanthanum(II1).
TABLEVII
MOLAR ABSORPTIVITIES AND STABILITY CONSTANTS OF LANTHANUM(III)
CHELATE
Reagent Transformation EMR Transformation log KMR

o-E-PAN IIIA 2.50 x 104 VIA 7.16 f 0.02

IX 2-47 x 104 XI 7-08 f 0.04
0-PAP IIIA 1.43 x 104 VIA 6-76 f 0-04
IX 1.56 x 104 XI 6-58 f 0.04
pKa, = 8.80 (0-PAP) ; 10.20 (o-a-PAN).

(i) I n the pyesence of excess of Zanthanum(III)-In this instance, the absorbance is limited
by the concentration of the reagent.
The absorbance, A , is given by
~ f)
A = E M R C H R ~(1~ - - .. .. - . . (11)
where cVR = molar absorptivity of MR; CHRTot= total reagent concentration; andf = fraction
of reaction completed.
f=- CR -
- CR
‘HKTot R
‘ ‘MI?
1
- .. .. .. .. ..
1 CMR/CR + (12)
where CR = unreacted reagent concentration.
The proposed chelation reaction is
* KMR
M+HR + MR+H+ .. .. .. .. (13)
The conditional reaction constant, *KhR, is given by

In the presence of excess of lanthanum(III), CM= CYTot,the total metal-ion concen-

tration, and

where the side-reaction coefficients, a1 and p, are [HR]/C, and [M]/C,, respectively.
In the present study, a1 = 1 and p, if the major hydroxy-complex is La(OH),, is given by
TLaI
[La] + CLa(OH),I
or

The equilibrium constant involving hydrolysis of La3+ is given by21

+
La3+ 3H,O + La(OH), 3H+ +
Log [La3+] = 23.02 - 3pH
Values of obtained from equation (15)were used in equation (14)to calculate Cn values.
Absorbances were then obtained by using equations (11) and (12).
 View Article Online

406 BETTERIDGE AND JOHN: PYRIDYLAZONAPHTHOLS AND [Analyst, VOl. 98

(ii) I n the $resence of excess of reagent-Absorbance values are now limited by the metal-
ion concentration and are given by
Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

1
f =1 + cMR/cM
.. .. .. .. .. (17)

*K&R= - c ~ ~ [ H-l *KMR,~~

CMCR
In the presence of excess of reagent, CR = CHRTot,the total reagent concentration, and

As before, a = 1 and is given by equation (15). Values of CMobtained from equation

(18) were then used to calculate absorbance from equations (17) and (16).
The molar absorptivity, reaction, stability and dissociation constants used for the above
calculations are summarised in Table VII.
Theoretical absorbance curves were drawn and compared with the experimental curves
(Fig. 5). Agreement between the theoretical and experimental curves is good, which confirms
that the chelation reaction suggested is correct and that the decreased absorbance at high
pH is due to hydrolysis of lanthanum species.
Reaction of 4-(2-pyridylazo)phenolwith cobalt(II), nickeZ(II) and copper(II)-Spot tests
show that $-PAP is selective in its reactions with metal ions, and will react principally with
metals of Groups VIII and IB. Cobalt(II), copper(I1) and nickel(I1) react strongly, and the
chelation reactions of these ions have been investigated.
Absorbance curves were obtained in the presence of both excess of reagent and excess
of metal ions.
Absorbance measurements were made a t 548,520 and 517 nm, these wavelengths corre-
sponding to the absorbance maxima of the copper, nickel and cobalt chelates, respectively.
The absorbance curves are shown in Fig. 7. These curves indicate that maximum chelation
occurs a t a pH greater than the pK,, value of the reagent. Also, very little chelation occurs
at pH values below the pK,, value, which suggests that the neutral form of the reagent has
little tendency for chelation and that chelation occurs via the anionic form. This represents
an unusual type of chelation because the protons that are lost do not originate in the chelating
groups themselves.
Preliminary investigations of the nature of these chelates by continuous variation,l2
mole-ratio13 and s l o p e - r a t i ~methods
~ ~ ~ ~ ~showed that each chelate has a stoicheiometry of
metal to ligand of 1 :2. As the anionic form of the reagent is involved in chelation, then the
suggested chelation reaction is
+
M 2R- + MR,
Therefore, the application of transformations (VI) and (VII) was tried. The consistency of
the calculated constants and linearity shown by the correlation coefficient in Table VIII
indicated that this is the predominant reaction over the pH range 7 to 8.
The decrease in absorbance observed a t higher pH indicates the formation of hydroxy-
chelate species :
KOH
MR2+OH- + MR,OH-
This hydrolysis reaction is identical with that of the manganese(I1) and zinc(I1) chelates
with o-p-PAN. The hydrolysis constant, KMR(OH), can therefore be calculated from identical
transformations. The linearity of these transformations is shown in Fig. 8, and the hydrolysis
constants were found to be 104.58 * andOao2 * Oeo2 for the nickel(I1) and cobalt(I1)
hydroxy-species, respectively.
SOLVENT-EXTRACTION STUDIES-
The theory of the solvent extraction of chelates is fairly well established.24-27 If the
distribution ratio, D,is defined as the concentration of the metal in the organic phase divided
 Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

1.2
- 1.0
s
8 0-8
0.6
.ff
8 0-4
o
a 0.2

6 7 8 9 1 0 6 7 8 9 1 0 1

PH
Fig. 7. Absorbance zlersus pH graphs for chelates of p-PAP in 5 per cent. aqueous methanol: (a) with
copper(I1); (b) with nickel(I1); and ( c ) with cobalt(I1). A, excess of reagent: CM 2.5 x M , CR 1.0 X
W 4 M ; B, excess of metal: C x 1.0 x M, CR 2.5 X M ; c, excess of reagent: CM 2.5 X M, CR
1.0 x M ; D, excess of metal: CM 1.0 x M , CB 2.5 x M ; E, excess of reagent: CM 5.0 X M,
CR 1.0 X lo-* M ; F, excess of metal: CM 1.0 X M, CR 5-0 X M

VIII
TABLE
MOLAR ABSORPTIVITIES AND STABILITY CONSTANTS FOR COBALT(11) AND NICKEL(11) CHELATES
Standard error of
Correlation
Condition Metal ion y=mx+c coefficient slope,
e s rnt i m a t e EMR~ x lo4 log KMR,
Excess of metal ions Ni(1I) y = 0 . 0 3 1 ~ 3.820
+ 0.994 0.001 1 0.088 5-23 8.95 f 0-08
WII) y = 0 . 1 9 8 ~ 6-843
+ 0.998 0.0044 0.205 2-92 7.88 f 0.07
Excess of reagent Ni(I1) y = 0 . 0 1 5 ~$- 2.031 0-981 0.0009 0.1 46 4.92 8-69 f 0.10
Co(I1) y = 0 . 2 0 8 ~ 3-887
+ 0.999 0-0031 0.266 2.57 7-82 f 0.06
pKa, = 7.77.
View Article Online
 View Article Online

408 BETTERIDGE AND JOHN : PYRIDYLAZONAPHTHOLS AND [Analyst, VOl. 98

by the concentration of the metal in the aqueous phase, it follows, for the extraction reaction
MC+ + 2HRo + MR2(,, 2H$ +
Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

that
D = [MR210/([MR2]W cM) +
= KDX/(l +
cM/[MR21W)
= KDX/(l +
[Ml/p[MR21)
= KDX/{l +
l/(pKMR,a&D)[HRlz)) .. . . (19)
In equation (19) and subsequent equations K,, and KDRare the partition coefficients of
the chelate and reagent, respectively.
%(D)= [R-l/c€IR
= KalKa2/{ [HI2 + + +
Kal[HI (1 K ~ ~ v o / v w )Ka,Ka, 1
where CHRis the total concentration of reagent present initially in the organic phase of
volume V o and the other terms have been defined previously. It is assumed that the metal
is present in the organic phase only as MR,. In the absence of side-reactions that involve
the metal ion (/3 = 1) and incomplete chelate formation (D < KDx)and on the assumption
that the reagent is present in the organic phase mainly as HR, this expression reduces to
= K1\IR~KDXK~~[HRl~/I(~RIHla .. .. . . (20)
For practical purposes, the percentage extracted, E , is often used, and is related to the
distribution ratio by

The pH value at E = 50 per cent. is designated as pH&,and the difference between

the pH, values of two extractable chelates is a measure of their separability. If it is assumed
that the partition coefficients for chelates of the same reagent are identical and there are
no competing side-reactions, extraction curves ( E veisus pH) for the same reagent and a
series of metal ions of the same charge will have identical shapes. The relative degree of
extraction a t a given pH will be governed by the relative values of the formation constant
ISMn,. Hence, if the values of the constants in equation (20) are known and a value of [HRIo
is defined, curves of E v e w m pH can readily be calculated.

PO H
Fig. 8. Graphs of hydrolysis equations for cobalt(I1) and
nickel(I1) chelates of p-PAP: A, cobalt(I1); and B, nickel(I1)

The values of all the terms in equation (20) except KDRand KDxcan be determined by
the spectrophotometric methods described above. K D R must be determined independently.
Therefore, in order to calculate the extraction curve, one must either assume a value of
KDx or determine it independently. Two approaches seem reasonable: (i) assume that
KDx = K,,, which was applied to o-PAP systems; (ii) assume that KDxis the same for
similar systems, which was applied to o-or-PAN and to o-p-PAN systems, the value of lo4 being
taken from previous work.14-18 The calculated curves were compared with those obtained
 View Article Online

June, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I1 409

experimentally (Fig. 9). As would be expected from the above assumptions and the additional
assumption that KMRI remains constant, not all of the experimental and theoretical curves
Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

coincide. Nevertheless, they are close enough for the analyst to use the method as a guide
in the choice of his experimental conditions.

Fig. 9. Comparison of experimental and predicted extraction

curves for the chelates of zinc and manganese: A, zinc; and B, man-
ganese. Solid lines, experimental ; and broken lines, calculated.
Chelates: 0, o-a-PAN; 0 , o-S-PAN; and x , o-PAP

A more valuable guide is provided when there are competing reactions, /3 < 1. The
simplified equation (20) must be modified by including /3 in the numerator. The competing
reaction might arise from the addition of a masking agent, L, which forms a series of com-
plexes with the metal ion. If the formation constants are known, it is a straightforward
+ n
matter to calculate 1/p, which is given by 1 X k , k , . . . kn[LIn. The spectrophotometric
0
studies should reveal the presence of hydroxy-species, which can be taken into account in a
similar way:
+
l/P = 1 KMR,(OH) +
[OH] KMR,(oH), [OH],
The extraction curve for the manganese(I1) - o-P-PAN system, in which the presence
of hydroxy-complexes had been established spectrophotometrically, was calculated with the
aid of the formation constants determined spectrophotometrically. The curve is compared
with the experimentally obtained curve in Fig. 10. The agreement is satisfactory and clearly

PH
Fig. 10. Comparison of experimental and
predicted extraction curves for Mn(o-,%PAN),. Solid
line, experimental ; and broken line, predicted
 View Article Online

410 BETTERIDGE AND JOHN: PYRIDYLAZONAPHTHOLS AND [Analyst, vOl98

demonstrates the need to select the extraction conditions with care. For more exact com-
parison, the calculated and experimental pH values are given in Table IX.
Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

--
TABLE
IX
COMPARISON O F EXPERIMENTAL SOLVENT-EXTRACTION CONSTANTS WITH
CONSTANTS PREDICTED BY SPECTROPHOTOMETRY
Experimental Predicted
Reagent Metalion LogKf pH,,, Log Kf PHI/,
o-a-PAN Mn(I1) 13.30 7.30 13.40 7-22
Zn(I1) 19-60 4.10 19.13 4.36
O- B-PAN Mn(I1) 16-13 7.86 16.80 7-96
Zn(I1) 22-17 4.86 21.60 6.10
O-PAP Mn(I1) 10.42 7.60 10.60 7-66
Zn(I1) 16.68 6.00 16.60 6-07
pKa, = 9.63 (o-x-PAN); 11.62 (O-p-PAN); 8-88 (O-PAP).
KDx= lo4, assumed for each chelate.
It is more usual to deduce the values of the constants in equations (19) and (20) from
the experimental curve, and this procedure was followed for the manganese(I1) - o-P-PAN
system so as to determine the values of KMR,(OH) and KMR,(OH),. The value of K D X K M R , was
calculated from the points on the rising part of the curve of log D veysus pH and a value of 18
was calculated as a function of pH from equation (19) by using points from the descending part
of the curve. fl was then examined as a function of [OH] and it was deduced that MnR2(OH),2-
could not be detected, that MnR2(OH)was present and that log KMnR,(OH) had a value of 7.64.
The agreement between this value and that obtained spectrophotometrically is good, and
is a little surprising because of the low number of acceptable experimental points in the
solvent-extracti on systern.
CONCLUSIONS
The determination of reliable stability constants is a tiresome and time-consuming
procedure, which can often produce results that are of limited applicability. The most
accurate results are generally agreed to be those obtained by potentiometric titrations, and
these results now have the advantage of the availability of well tested computer programs28
that can be used in order to transmute the experimental points into results. However,
inevitably, relatively large concentrations of reactants are used, which results in the use of
non-aqueous solvents and the increased likelihood of the formation of polynuclear species.
Methods based on solvent extraction and spectrophotometry have direct appeal to the
analytical chemist as he may be able to set up his analytical method while measuring the
stability constants. This is true of solvent extraction, although it may be difficult to obtain
reliable constants without a great deal of experimentation and it is also true of the spectro-
photometric method described above. The pH - absorbance curves upon which the method
is based are essential in the development of the analytical procedure. These curves can also
yield much information about the effect of varying the conditions, e.g., reagent concentration,
pH and masking agent. In this respect, as well as in terms of accuracy, the method involving
these curves is far superior to Job’s method and the related procedures that are commonly
used. It has also been shown that the results can be used directly to predict solvent-extraction
curves. The weakness of the method is that it depends upon extrapolation, and it requires
careful experimental work and alteration of conditions so as to ensure that reliable results
are obtained. We have concluded that it is just as time consuming as the other reliable
methods, but we feel that because of the amount of information of direct analytical interest
that is obtained, it is a very desirable method.
Of the four reagents studied, o-a-PAN is clearly the most suitable as it forms highly
coloured stable complexes a t lower pH values than o-fl-PAN, o-PAP or @-PAP. o-fl-PAN
and o-PAP form complexes a t about the same pH values but the molar absorptivities of
PAN complexes are greater than those of P A P complexes.
Specific analytical applications will be described in subsequent papers.
 View Article Online

June, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I1 411

We are grateful to the T. and E. Williams Scholarship Fund for a maintenance grant
to one of us (D.J.).
Published on 01 January 1973. Downloaded by Karolinska Institutet University Library on 19/06/2014 13:59:01.

REFERENCES
1. Anderson, R. G., and Nickless, G., Analyst, 1967, 92, 207.
2. Shibata, S., Analytica Chim. Acta, 1960, 23, 367.
3. -Ibid., 1961, 25, 348.
4. Rosotti, F. J. C., and Rosotti, H., “The Determination of Stability Constants,” McGraw-Hill,
New York, 1961, pp. 47-51.
5. Sommer, L., Folia Fac. Sci. Natn. Univ. Purkynianae Bmo, 1964, 5, Part 1, 1.
6. Hnilickova, M., and Sommer, L., Talanta, 1966, 13, 667.
7. Chromy, V., and Sommer, L., Ibid., 1967, 14, 393.
8. Sommer, L., and Hnilickova, M., Ibid., 1969, 16, 83.
9. Sommer, L., and Ivanov, V. M., Ibid., 1967, 14, 171.
10. Sommer, L., Ivanov, V. M., and Novotna, H., Ibid., 1967, 14, 329.
11. Sommer, L., and Novotna, H., Ibid,, 1967, 14, 457.
12. Vosburgh, W. C., and Cooper, G. R., J . Amer. Chem. SOL, 1941, 63, 437.
13. Yoe, J. H., and Jones, A. L., Ind. E n g g Chem., Analyt. Edn, 1964, 16, 11.
14. Betteridge, D., Todd, P. K., Fernando, Q., and Freiser, H., Analyt. Chem., 1963, 35, 729.
15. Pease, €5. F., and Williams, M. B., Ibid., 1959, 31, 1044.
16. Corsini, A., Mai-Ling Yih, I., Fernando, Q., and Freiser, H., Ibid., 1962, 34, 1090.
17. Nakagawa, G., and Wada, H., J. Chem. SOC. Jafian, 1963, 84, 639; Chem. Abstr., 1964, 61, 1242h.
18. Betteridge, D., Fernando, Q., and Freiser, H., Analyt. Chem., 1963, 35, 294.
19. Anderson, R. G., and Nickless, G., Analyst, 1968, 93, 13.
20. - -, Analytica Chim. Acta, 1967, 39, 469.
9

21. Pourbaix, M., “Atlas of Electrochemical Equilibria in Aqueous Solutions,” Pergamon Press,
Oxford, 1966.
22. Harvey, A. E., and Manning, D. L., J . Amer. Chem. SOC.,1950, 72, 4488.
23. - -, Ibid., 1952, 74, 4744.
t

24. Morrison, G. H., and Freiser, H. F., “Solvent Extraction in Analytical Chemistry,” John Wiley,
New York, 1962.
28. Starf, J., “The Solvent Extraction of Metal Chelates,” Pergamon Press, Oxford, 1964.
26. Laitinen, H. A. , “Chemical Analysis,” McGraw-Hill, New York, 1960.
27. Marcus, Y., and Kertes, A. S., “Ion-exchange and Solvent Extraction of Metal Complexes,”
Wiley-Interscience, New York, 1969.
28. Childs, C. W., Hallman, P. S., and Perrin, D. D., Talanta, 1969, 16, 1119.
Received April 2nd, 1970
Accepted November 22nd, 1972