Anal. Chem. 1988, 60,1565-1569 lev
eV
E
I 5 t
4 2
0
I
2
3 4 Nominal Oxidation
Figure 3. Correlation plots of
Av
5
6
Number
vs oxidation number.
structures in a multicomponent system is difficult, unless additional information is available to provide representative calibrating spectra. Registry No. Mn, 7439-96-5; MnO, 1344-43-0;MnOz, 131313-9; KMn04,7722-64-7;MnC12,7773-01-5;MnA1204,12068-52-9; [Mn(CO),Cl],, 18535-43-8.
LITERATURE CITED (1) Sham, T. K. J. Chem. Phys. 1986. 84, 7054. (2) Sugiura, C.; Kkamura, M.; Muramatsu, S. J. Chem. Phys. 1986, 84, 4824. (3) Sankar, 0.; Sarode, P. R.; Srinivasan, A,; Rao, C. N. R.; Vasudevan, S.; Thomas, J. M. Roc. Indbn Aced. Sci., Chem. Scl. 1984, 93, 321.
1565
(4) Mansour, A. N.; Cook, J. W., Jr.; Sayers, D. E. J. Phys. Chem. 1984, 88, 2330. (5) Grunes, L. A. Phys. Rev. 8 : Condens. Matter 1983, 27, 2111. (6) Mulier, J. E. I n EXAFS and Near Edge Structure 111;Hcdgson, K. O., Hedman, B., Penner-Hahn, J. E., Eds.; Springer-Veriag: Berlin, 1984; P 7. (7) Proietti, M. G., et ai. I n EXAFS and Near E@ Sfructwe 111; Hodgson, K. O., Hedman, B., Penner-Hahn, J. E., Eds.; Springer-Veriag: Berlin, 1984; p 26. (8) Brown, N. M. D.; McMonagie, J. 6.; Greaves, G. N. J. Chem. Soc., Faraday Trans. 11984, 80, 589. (9) Bauer, S. H.; Chiu, N-S.; Johnson, M. F. L. J. fhys. Chem. 1986, 90, 4868. (10) Leapman, R. D.; Grunes, L. A.; Fejes, P. L.; Siicox, J. I n EXAFS Spectroscopy; Teo, B. K., Joy, D. C., Eds.; Plenum: New York, 1981; p 217. (11) Cramer, S. P., et ai. J. Am. Chem. SOC. 1976, 98, 1287. (12) Cramer, S. P., et. ai. J . Am. Chem. SOC. 1983, 705, 799. (13) Belli, M.; Scafati, A.; Bianconi, A.; Mabiiio, S.; Pailadino, L.; Reaie, A,; Burattini, E. Solid Sfate Commun. 1980, 35, 355. (14) Garcia, J.; Biaconi, A.; Marceiii, A,; Davoii, I.;Bartolome, J. Nuovo Cimento SOC.Ital. Fis. D 1888, 70, 493. (15) Cartier, c.; Verdaguer, M.; Menage, S.; Girerd, J. J.; Tuchagues, J. P.; Mabad, B. J. Phys. (les Uls, Fr.) 1988, 47(Sup. to No. 12), C8-823. (18) Van Nordstrandt, R. A. I n Non-CrystallineSolids; FrOchette, V. D. Ed.; Wiley: New York, 1960; p 168. (17) Bauer, S. H.; Chiu, N-S.; Johnson, M. F. L. J. Mol. Struct. 1984, 725, 33.
RECEIVED for review November 19, 1987. Accepted March 5, 1988. We thank the donors of the Petroleum Research Fund, administered by the American Chemical Society, for support of this research. The extended X-ray absorption fine structure (EXAFS) spectra were recorded at the Cornell High Energy Synchrotron Source, supported by NSF Grant No. DMR-780/267. Our investigation was also supported, in part, by a grant from DOE, Grant No. DE-FG22-86PC90520.
Spectrophotometric Determination of the Second Acid Dissociation Constant of Oxine Bound to Controlled Pore Glass Anne Kristine Kolstad, Patrick Y. T. Chow, and Frederick F. Cantwell*
Department of Chemistry, University of Alberta, Edmonton, Alberta, Canada T6G 2G2
The Ionizable surfacegroup model Is appiled to a sorbent conslstlng of 5-(phenylazo)-8-hydroxyqulnollne covalently bound to controlled pore glass (CPG-OXINE). Capaclty measurements combined wkh informanon from potentlometric titration curves permit calculation of surface charge denslty and swface potenHal as a function of pH. Most of the surface charge comes from residual hydroxyl groups on the glass surface. Spectrophotometrlc measurements on CPG-OXINE give the degree of lonlzation of the phenolic hydroxyl group as a function of pH. The observed of 8.62 f 0.07 for this group Is Independent of surface potentlal and Is close to that reported In homogeneous aqueous solution.
za,'
Over the last 15 years there have been numerous literature reports on the preparation of sorbents by covalent binding of the ligand 8-hydroxyquinoline (oxine) to insoluble porous supports such as organic polymers (1-5),silica gel (6-12),and controlled pore glass (CPG) (7-10,13,14).These sorbents are used to remove trace concentrations of heavy metal ions from electrolyte solutions (15),to preconcentrate trace metal ions prior to their quantitative determination (4,6,8, 14, 0003-2700/88/0360-1565$01.50/0
16-18) and as column packings for batch and chromatographic separations (1,3,12,13, 19,20). In the sorbents for which silica gel or CPG is the support it is not orine itself that is bound to the surface but rather one or another 5-phenylazo derivative of oxine (6-14). In order to explain the observed dependence of metal ion sorption on solution pH and ionic strength, it is necessary to know the acid-base properties of the bound 5-(phenylazo)-8-hydroxyquinoline (5-P-OXINE, Figure 1). In homogeneous aqueous solution 5-(phenylazo)-8-hydroxyquinoliniumcation has pKa, = 3.2 (for cation to neutral species) and pK, = 8.6 (for neutral to anionic species) (21). Substituents at the ortho, meta, and para positions of the phenylazo group do not change these values by more than f0.7 unit unless they are highly electron-withdrawing or -donating (21).Thus, by analogy with homogeneous solution behavior, covalently bound 5-P-OXINE is expected to have in the range 2.5-3.9 and in the range 7.9-9.3. Several attempts have been made to measure and for 5-P-OXINE bound to silica gel and CPG using potentiometric (pH) titration of aqueous suspensions (7,13,19,22). Titration curves are typically ill-defined and yield values showing considerable variability (7,22). This is due to the
sal
sa2 sa, sa, sa
0 1988 American Chemical Society
1566
ANALYTICAL CHEMISTRY, VOL. 60, NO. 15, AUGUST 1, 1988 I _
HOX
0
Glass
+
em-+m
(5)
The bar indicates a species in the vicinity of the charge surface. The species HOX is the neutral species, shown in Figure 1. The acidity constant is given by (30, 32)
Figure 1. Structure of the derivative of 5-(phenylazo)-8-hydroxyquinoline bound to controlled pore glass in CPGOXINE.
presence, on the surface, of other acid-base groups such as hydroxyl groups in residual SiOH and BOH sites ( 7 , 2 3 4 5 ) . Uncertainty about the value of has prompted an investigation of the acid-base behavior of a popular commercially available material in which 5-P-OXINE is bound to controlled pore glass (CPG-OXINE) and the development of a spectrophotometric technique to measure
Substituting for aH,xin eq 6 from eq 4 gives
sa,
za,.
THEORY The degree of ionization of a weak acid depends on the pH of the solution. When the surface lining the pores in CPGOXINE is ionized, the pH immediately adjacent to the surface is different than the pH in bulk solution surrounding the particle of CPG-OXINE. It is pH that determines the degree of ionization of the 5-P-OXINE bound to the surface. In a previous study (24) pH for small-pore diameter (32 A) CPG was related to pH via the polyelectrolyte gel model (26-29). In the present case, however, the pores in CPGOXINE are 500 A. In 0.1 M electrolyte the “thickness” ( 1 / ~ ) of the electrical double layer near the charged surface lining the pores is -10 A (30,31)so that solution in the middle of the pores is uninfluenced by the surface potential. In this case the ionization of the bound 5-P-OXINE can be understood in terms of an ionizable surface-group model (32-34). In this model the relationship between surface charge density (uo, C/cm2) and electrical potential of the charge surface (\ko, V) is given by the Gouy-Chapman equation which, a t 25 OC, has the form sinh
(-)
2 F\ko
2RT
= (8.53 x 104)c-%,
(1)
where Z is the charge on the ionized surface group, R is the ideal gas constant, F is the Faraday constant, T i s the absolute temperature, and c is the ionic strength of the bulk solution (30,32,35).The value of uocan be calculated from the surface concentration of ionized surface groups (Q,mmol/g) as follows UO
= 2 FQ/1000Asp
(2)
where AsP is the specific surface area (7.0 X lo5 cm2/g for CPG-OXINE (36)). The potential \k, at any distance x cm away from the charge surface can be calculated (31). The simplified form of the relationship that is valid at relatively low values of \k0 is \k, = \ko exp((-3.29
x ~o’)c~/~x)
(3)
When uo is constant, changing c alters \ko via eq 1 which, in turn, alters 9,via eq 3. The relationship between the activity of hydrogen ion x cm away from the charge surface ( a H , x ) and in the bulk solution ( a H ) is given by the Boltzmann equation (30, 32, 34):
(4) Since Kal is undoubtedly 6 comes from deprotonation of surface hydroxyl groups than from deprotonation of the phenolic hydroxyl groups of bound 5-P-OXINE. The charge surface, at which uo and Q0 are defined, is identified with the hydroxyl-containing glass surface of the CPG.
EXPERIMENTAL SECTION Chemicals and Reagents. All solutions were prepared with doubly distilled, deionized water. All chemicals were reagent grade. The pH 8.0 buffer used to elute C1- from CPG-OXINE was 0.05 M NH3 and 0.9 M NH4N03. The pH 6.40-9.80 buffers used in the spectrophotometric study of CPG-OXINE were composed of various combinations of NaH2P04,NH3, and/or NH&l and all had c = 0.10 M ionic strength. Apparatus. Measurements of pH were made either with separate glass and calomel electrodes using an Accumet Model 525 pH meter (Fisher Scientific Co.) or with a combination glass/reference electrode using an Accumet Model 520 pH/ion meter (Fisher). EMF measurements in the null-point potentiometry experiment (37) were made between two Ag/AgCl electrodes in an “H-type” cell with 2 M KN03 salt bridge using a digital voltmeter. Copper concentrations were measured at 324.8 nm by using a Model 4000 atomic absorption spectrophotometer (PerkinElmer). A calibration c w e was prepared from electrolytic copper. Spectrophotometric measurements on CPG-OXINE were made at 520 nm in a modified variable path length cell with 1.5-in. diameter round windows (Beckman Instruments). Inlet and outlet holes were threaded to accept standard Cheminert end fittings (Applied Science),and a porous Teflon filter was installed in the outlet hole to permit slurry packing and flushing with buffer. The actual cell geometry was established by cutting a 4 mm wide rectangular slot through the 0.13 mm thick Teflon spacer disk placed between the cell windows. The ends of the slot terminated at the inlet and outlet holes so that liquid was pumped through the length of the packed bed. The bed of particles was packed by pumping a slurry from a stirred reservoir. A Milton Roy Minipump (Applied Science Co.) was used for slurry packing and flushing the bed with buffer. CPG-OXINE. Material from each of two commercial batches (batch 071982-81, called batch A; and batch 850605085, called batch B; Pierce Chemical Co.) with particle diameter 125-177 pm was washed with 1% “OB, rinsed with water, and air-dried prior to use. Microanalyses for C and N were performed on batch A by using standard methods by the Microanalytical Laboratory of the
ANALYTICAL CHEMISTRY, VOL. 60, NO. 15, AUGUST 1, 1986
University of Alberta Chemistry Department. Capacity determination by Cu2+uptake (7, 11, 19) was performed on batch A by equilibrating CPG-OXINE overnight with an acetate/acetic acid buffered solution (pH 5.0) containing an excess of Cu2+. The CPG-OXINE was fitered out and rinsed with water until no Cu2+was detected in the rinsings. The bound Cu2+was then eluted with a mixture containing 1M HC1 and 1 M HN03 and determined in the eluate by atomic absorption spectrophotometry. Determination of C1- content of acid-washed CPG-OXINE was performed on both batches A and B by first equilibrating with 1M HCl overnight and then drying under vacuum, without water washing, to remove all unreacted HCl. Elution of Cl- was achieved with pH 8 NH3/NH4N03buffer, checking for completeness of elution with AgNOB. The Cl- content of the eluate was determined via null-point potentiometry. Potentiometric Titrations. Duplicate 1-g samples of HClwashed and vacuum dried CPG-OXINE from batch B were suspended in 100 mL of water containing 0.10 M NaC104. These were titrated under Nz atmosphere with 0.1 M NaOH by using a 2-ML micrometer buret (Roger Gilmont, Great Neck, NY). Spectrophotometry. Both CPG-OXINE from batch B and CPG (CPG/5OO, upderivatized, batch 841101081, Pierce) were ground and sieved under water. The 38-45 wm fraction of each was slurry packed into the 0.13 mm path length flow cell, and the appropriate buffer (pH 6.40, 7.49, 7.76, 7.98, 8.20, 8.50, 8.61, 9.80,all c = 0.10) was pumped through the packed bed until well after the pH of the effluent was equal to that of the influent. The visible spectrum was recorded with a Model 8451A diode-array spectrophotometer (Hewlett-Packard Corp.). Changing to the next buffer solution was accomplished without disturbing the packing structure of the bed.
RESULTS AND DISCUSSION Capacity. The capacity of batch A was determined by the established literature method involving sorption of Cu2+(7, 11,19). In this experiment copper forms a 1:l complex with bound 5-P-OXINE (5,19,22). The measured capacity was (3.6 f 0.05) X mmol of 5-P-OXINE/g. Measurement of bound 5-P-OXINE capacity cannot be based upon microdetermination of nitrogen or carbon because the synthetic scheme for CPG-OXINE is known to leave a lot of N- and C-containing byproduds as impurities covalently bound to the glass surface (10,ll). For example, from the measured nitrogen content of 0.16 mmol of N/g and from the measured carbon content of 1.18 mmol of C/g, the calculated content of bound 5-P-OXINE has an erroneously high value of 0.040 and 0.062 mmol of 5-P-OXINE/g, respectively. Chloride Content. In order to calculate uo, as discussed below, it is necessary to know the concentration of cationic protonated nitrogens present on the acid-washed CPG-OXINE at the start of the titration. Since cationic groups arising from protonation of basic nitrogens have associated with them an equivalent amount of C1-, the former may be determined by measuring the latter. In this experiment vacuum drying, rather than water washing, was used to remove excess HC1 from acid-washed CPG-OXINE in order to avoid loss of protons from NH+groups via hydrolysis. The Cl- content was found to be 0.011 f 0.001 mmol of Cl-/g for batch A and 0.008 f 0.003 for batch B. This is the quantity needed to calculate 00.
It is interesting to calculate the fraction of protonatable (basic) nitrogens in CPG-OXINE that are associated with bound 5-P-OXINE. It has been shown that in strong acid solution the compound 5(phenylazo)-8hydroxyquinolinemay add up to two protons, the f i t on the quinoline nitrogen and a second on one of the azo nitrogens (21). Thus, on the basis of the measured capacity of 0.0036 mmolfg, the bound 5-POXINE could account for between 0.0036 and 0.0072 mmol of basic N/g, which represents between 33% and 65% of the total basic nitrogen content of 0.011 mmol/g in CPG-OXINE. In addition, the fact that the total basic nitrogen content (0.011
1567
1
h
3.0 0.0
0.2
0.1
mmol NaOH/g Potentiometric titration curve of acid-washed and vacuwndried CPGOXINE (batch B) at ionic strength 0.10 M. Figure 2.
mmol/g) is far less than the total nitrogen content (0.16 mmol/g) indicates that the majority of the nitrogens in the impurity groups on CPG-OXINE are not basic. This is consistept with the facts that an aromatic amide is the major synthetic byproduct (10-12,14) and that aromatic amides are extremely weak bases (38). Potentiometric Titrations. In Figure 2 is presented one of the titration curves for 1.00 g of acid washed CPG-OXINE that was obtained in 0.10 M NaC104 solution. As far as bound 5-P-OXINE is concerned, protons are removed from cationic protonated azo and quinoline nitrogens in the early stages of the titration and from the neutral phenolic group later in the titration. However, the situation is complicated by the fact that protons are being removed from residual surface hydroxyl groups throughout the titration. This precludes the use of titration curves to obtain the value of Nevertheless, the titration curves can be used to measure the total negative surface charge on CPG-OXINE at any pH above that at which deprotonation of cationic groups is complete. Deprotonation is complete at pH >6 since is in the range 2.5-3.9 (7,21). The concentration of negative surface groups, 8, is equal to the mmol of NaOH/g taken up by CPG-OXINE in reaching a given pH, calculated by the method of Berube and de Bruyn (39),minus mmol of OH-/g required to neutralize the protonated cationic sites. The mmol of OH-/g required to neutralize the protonated cationic sites at pH >6 is equal to the measured mmol of Cl-/g (Le., 0.008 f 0.003 for batch B). Calculated values of Q are presented in Table I. Spectrophotometry. Absorbance measurements made through a thin layer of CPG-OXINE particles were subject to light scattering. However, it was found, by using the same size particles of CPG without bound 5-P-OXINE, that the absorbance due to scattering alone was independent of solution pH (see footnote c, Table I). Thus, it was possible to use the change in absorbance as a function of solution pH to measure pK,,' for bound 5-P-OXINE. At a fixed wavelength the absorbance, A , at any pH is given by
zap.
sa,
-
A = [HC)XIbe,m+ [OX-]bto~+ Ascatbr (8) By measuring the absorbances Aaad at pH > 3 ,; and combining eq 8 with eq 7, the following expression is obtained exp( -
_1 -aH
%)
Ka2
A - Amid Aalkali-A
(9)
in which it is assumed that the phenolic group on bound 5-P-OXINE is located at a distance x cm from the charge surface. Aacidwas measured at pH 6.40, and Adui was measured at pH 9.80. A plot of l / a vs ~ ((A - Aacid)/(Adm- A)) should have a slope of exp(-F*,/RT)/Ka2 at any pH. Such a plot for one of the duplicate runs is shown in Figure 3. Presented in Table I are values of Q, uo, and \k0 The quantity -Poincreases with pH. If the distance x remains constant, independent of pH, then -9,should also increase with pH (eq 3) and the plot in Figure 3 should be curved (i.e., changing slope). However, the plot is linear with slope (3.7 f 0.1) X lo* and zero intercept (1 f 5) x lo6. From the reciprocal slope of the line, is found to be 8.57 (run 1). For run 2 (not shown) the slope was (4.6 f 0.1) X lo8, the was 8.67. intercept was (-5 h 7) X lo6 and Sa,' Evidently, the value of 9,remains constant even though q0changes with pH. The probable explanation for this is that x , the distance between the charge surface and the phenol/ phenolate group on bound 5-P-OXINE, increases as both uo
z,,'
Vernon, F.; Nyo, K. M. J. Inorg. Nucl. Chem. 1978, 40. 887-891. Vernon, F.; Eccles, H. Anal. Chlm. Acta 1973, 63, 403-414. Parrish, J. R.; Stevenson, R. Anal. Chlm. Acta 1974, 7 0 , 189-198. Willle, S.N.; Sturgeon, R. E.;Berman, S. S. Anal. Chlm. Acta 1983, 749, 59-66. (5) Utkelov, B. A.; Yergozhin, Ye. Ye.; Mukhitdinova, B. A.; Rafikov, S. R. Pdym. Scl. USSR (Engl. Trans/.) 1978. 2 0 , 532-540. (8) Sturgeon, R. E.; Berman, S. S.;Wlllie, S. N.; Desaulniers, J. A. H. Anal. Chem. 1981, 53, 2337-2340. (7) Marshall, M. A.; Mottola, H. A. Anal. Chem. 1983, 5 5 , 2089-2093. (8) Ryan, D. K.; Weber, J. H. TaLnta 1985, 3 2 , 859-863. (9) Marshall, M. A.; Mottola, H. A. Anal. Chim. Acta 1984, 158, 369-373. (10) Jezorek, J. R.; Faltynski, K. H.; Blackburn, L. G.; Henderson, P. J.; Medlna, H. D. Talanta 1985, 3 2 , 763-770. (1 1) Fulcher, C.; Crowell, M. A.; Bayliss, R.; Holland, K. B.; Jezorek, J. R. Anal. Chim. Acta 1981, 729, 29-47. (12) Hill, J. M. J. Chromatogr. 1973, 7 6 , 455-458. (13) Kuo, M.-S.; Motolla, H. A. Anal. Chlm. Acta 1980, 720, 255-266. (14) Sugawara, K. F.; Weetall, H. H.; Schucker, G. D. Anal. Chem. 1974, 46, 489-492. (15) Moorhead, E. D.; Davis, P. H. Anal. Chem. 1974, 4 6 , 1879-1880. (16) Malamas, F.; Bengtsson, M.; Johansson, G. Anal. Chlm. Acta 1984, 760, 1-10. (17) Marshall, M. A.; Mottola, H. A. Anal. Chem. 1985, 5 7 , 729-733. (18) Guedes da Mota. M. M.; Romer, F. 0.;Griepink, B. Fresenius 2.Anal. Chem. 1977. 267, 19-22. (19) Jezorek, J. R.; Freiser, H. Anal. Chem. 1979, 5 7 , 366-373. (20) Shahwan, G. J.; Jezorek, J. R. J. Chromatogr. 1983, 256, 39-48. (21) Khater, M. M.; Issa, Y. M.; Shoukry, A. F. J. Prakt. Chem. 1980, 322, 470-474. (22) Jezorek, J. R.; Fulcher, C.; Crowell, M. A.; Bayliss, R.; Greenwood, B.; Lyon, J. Anal. Chlm. Acta 1981, 737, 223-231. (23) Cukman, D.; JednaEek-BibEan, J.; Veksli, 2.;Hailer, W. J. Collold Interface Scl. 1987, 775, 357-361. (24) Altug, I.; Hair, M. L. J. Phys. Chem. 1967, 7 7 , 4260-4263. (25) Dawidowicz, A.; Janusz, W.; Szczypa, J. and Waksmundzki, A. J. Colloid Interface Sei. 1987, 175, 555-558. (26) Helfferich. F. Ion Exchange; McGraw-Hill: New York, 1962; Chapter (1) (2) (3) (4)
5. (27) Marinsky, J. A. I n Ion Exchange: A Series of Advances; Marinsky, J. A., Ed.; Marcel Dekker: New York, 1966; Vol. 1, Chapter 9. (28) Marinsky, J. A.; Wolf, A.; Bunzl, K. Talanta 1980, 2 7 , 461-468. (29) Marinsky, J. A. J. Phys. Chem. 1985, 6 9 , $294-5302.
Anal. Chem. 1988, 60,1569-1573 (30) Grahame. D. C. Chem. Rev. 1947, 4 7 , 441-501. (31) Shaw, D. J. Introduction to Colloid and Swface Chemistry, 3rd ed.; Butterworths: London, 1980; Chapter 7. (32) Healy, T. W.; White, L. R. A&. Colbid Interface Sci. 1978, 9 , 303-345. (33) Harding, I. H.; Heaiy. T. W. J . Colbid Interface Sci. 1985, 107, 382-397. (34) Takamura, K.; Chow, R. S. Co/b& Surf. 1985, 15, 35-48. (35) Cantwell, F. F.; Puon, S. H. Anal. Chem. 1979, 51, 623-632. (36) Product Bulletin, Pierce Chemical Co.. Rockford, IL, 1986. (37) Blaedel, W. J.; Lewis, W. E.;Thomas, J. W. Anal. Chem. 1952, 24, 509-5 12.
1569
(38) Albert, A.; Serjeant, E. P. The Determination of Ionization Constants, 3rd ed.;Chapman and Hall: New York, 1984; Chapter 9. (39) Mrubi, Y. G.; de Bruyn, P. L. J . Colbid Interface Sci. lQ68, 2 7 , 305-3 18.
RECEIVED for review September 21, 1987. Accepted March 19,1988. This work was supported by the Natural Sciences and Engineering Research (h~ncil of Canada and by the University of Alberta.
Calcium Sorption by Immobilized Oxine and Its Use in Determining Free Calcium Ion Concentration in Aqueous Solution Patrick Y. T.Chow and Frederick F. Cantwell*
Department of Chemistry, University of Alberta, Edmonton, Alberta, Canada T6G 2G2
Complexation of Ca2+ by 5-( phenylazo)-8-hydroxyqulnollne bound to controlled pore glass (CPGOXINE) Is descrlbed In terms of a modlflcatlon of the Ionizable surface-group model to Include "Me blndlng". Electrical propertles of the surface are obtained by applying the Gouy-Chapman theory to data obtained from capacity measurements and potentiometric tltratlons of CPG-OXINE. The marked dependence of Ca2+ sorption on both pH and ionic strength Is quantltatlvely explained. I n addltlon, CPG-OXINE Is used as a sorbent In a column equlllbratlon method to measure [Ca2+] In solutions containing labile calcium complexes with EDTA and the dlpeptlde carnoslne.
Cation-exchange techniques have been used to measure the concentration of free (hydrated) metal ion (e.g., hP+)in the presence of its complexes (1-11). Measurements are made by the column equilibration technique in order to avoid Perturbing the mea-figmd equilibria in the sample solution (7-21). In principle, the cation exchanger should selectively sorb only M"+ but not sorb any of ita complexes. While this condition is met when only negatively charged complexes of Mn+are present, it is not true in the presence of positively charged and neutral complexes (11). In searching for sorbents that are more selective for the species Mn+than are sulfonated cation exchangers, we have investigated a commercially available sorbent in which a derivative of 5- (phenylazo)-8-hydroxyquinoline(5-P-OXINE) is immobilized on controlled pore glass (CPG-OXINE). In a previous study the ionizable surface group model was used to describe the acid-base behavior of CPG-OXINE (12). In the present study an extension of this model is used to explain the pH and ionic strength dependence of calcium ion sorption by CPG-OXINE, and, in addition, this sorbent is shown to be selective for Ca2+in the presence of both its anionic and cationic labile complexes.
THEORY Site-Binding Model. The ionizable surface-group model (12-15) can be extended to describe "site-binding" of counterions (13,16-18) such as the formation of a metal-ligand 0003-2700/88/0360-1569$01.50/0
ox-
complex between the immobilized species and Ca2+ion. Previous evidence suggests that metal ions form 1:1complexes with bound (19-21)
ox--
Ca2+
+ Ox-+-x+
(1)
where the bar indicates a species in the vicinity of the charge surface (i.e., the silica surface of CPG). The formation constant for the complex is
ICaOX+l The Boltzmann equation relates activity of Ca2+at a distance x cm away from (but close to) the charge surface to its activity in bulk solution (ac,) (13, 18)
(3) where F is the Faraday R is the ideal gas T is absolute temperature, and \k, is the electrical potential at x . The value of \E, is calculated from the surface charge density by combining the Gouy-Chapman equation (eq 1 in ref 12), wKch gives the surface pokntial Q,,, with the following equation
9,= q,, exp(-3.29 x i07ci/2x)
(4)
where c is ionic strength of the bulk solution. In the column equilibration experiment CPG-OXINE is first equilibrated with bulk solution containing Ca2+with an activity aca and is then washed briefly with water to remove bulk solution from the pores and between the particles. The total number of moles of calcium remaining per gram of CPG-OXINE (Nc,) is given by the expression
Nca= [CaOX+],V,
+ iiCa
(5)
where V , is the volume per gram of CPG-OXINE in a thin layer of solution located at a distance x cm from the charge surface and [CaOX+], is the concentration of complexed calcium in this thin layer of solution. The quantity iic, is the moles of dissolved calcium, not complexed by bound 5-POXINE, in the pores of CPG-OXINE that are not removed 0 1988 American Chemical Society
