Indirect photometric chromatography - Analytical Chemistry (ACS

---

462

Anal. Chem. 1982, 5 4 , 462-469

ange-yellow color of the solution to a greenish blue one, indicating the partial conversion to the four-valent state, is clearly visible even after only 1 h of equilibrium time. After 24 h equilibration the reduction is almost complete and the distribution coefficients obtained are very similar to those for vanadium(IV). Coefficients for chromium(II1) depend very markedly on the history of the solution. Very indistinct bands representing various chromium ion species appear in column operations at room temperature. This is due to the slow ligand exchange rates of chromium(II1). Although the potential for element separation in the tartrate system is high more attractive alternatives are usually available. A molecule such as tartrate often is troublesome to remove. Only when it is already present in the system or has been added to prevent hydrolysis of elements such as tantalum and niobium should separations in tartrate media be considered as a first choice. The outstanding exception to this philosophy is the group separation of Zr, Sn(IV), etc. from all other elements that was considered earlier in the discussion. The only alternative would mean the use of a

system containing hydrofluoric acid.

LITERATURE CITED Tompkins, E. R.; Mayer, S. W. J . Am. Chem. SOC.1947, 69, 2859. Ryabchikov, D. I.; Eukhtiarov, V. E. Zh. Anal. Khim. 1952, 7, 377. Kreshkov, A. P.; Sayushkina, Tr Mosk Khim .-Tekhnol Inst. im D . I . Mendeleeva 1956, No. 22, 116; Chem. Abstr. 1957, 51, 16197. Marezcnko, 2. Chem. Anal. (Warsaw) 1957, 2 , 255; Chem. Abstr. 1958, 52, 1654t. Alimarin, I . P.; Tsintsevich, E. P. Zavod. Lab. 1956, 22, 1276. Tsintsevich, E. P.; Nazarova, G. E. Zavod. Lab. 1957, 23, 1068. Kimura, K.; Saito, N.; Kaklhana, H.;Ishlmori, T. Nippon Kagsku Zasshi 1953, 74, 305; Chem. Abstr. 1953, 47, 9850. Khorasani, S. S. M. A.; Khundkar. M. H. Anal. Chim. Acta 1959, 21. 406; 1961, 25, 292. Shiokawa, T.; Sato, A. NippOn Klnzoku Gakkalshi, Ser. 8 . 1951. 15, 264; Chem. Abstr. 1953i.47, 2272. Strelow, F. W. E.; Van der Walt, T. N. Anal. Chem. 1975, 47, 2272. Qureshl, M.; Varshney, K. G.; Kaushlk, R. C. Anal. Chem. 1973, 45, 2433. Dadone, A.; Eaffl, F.; Frache, R. Talanta 1976, 23,593. Slmek, M. Collect. Czech. Chem. Commun. 1977, 42, 798. Rouchaud. J. C.; Revel, G. J . Radioanal. Chem. 1973, 16, 221. Strelow, F. W. E. Anal. Chem. 1960, 32, 1185.

.

.

.

RECE~VED for review July 13,1981. Accepted October 26,1981.

Indirect Photometric Chromatography Hamlsh Small" and Theodore E. Mlller, Jr. 1712 Building, M. E. Pruiti Research Center, The Dow Chemical Company, Mldland, Michigan 48640

Indlrect photometric chromatography is a sensitive singiecolumn ion analysis method developed from the concept that photometers may be used to detect transparent ionlc species. The use of light-absorbing eluent ions in an ion-exchange mode enables sample Ions to appear as "troughs" in the base line absorbance as transparent sample ions substitute for the light-absorbing displacing ions. The elution times of these troughs vary with the ion injected and their depths (or areas) are proportional to the amount of sample Injected. Notable advantages of the new technlque are its slngie column simpilclty, its applicability to a wide range of ionic specles, and an inherently greater sensltivity than slngie-column conductometrlc approaches.

Ion determination by liquid chromatography is often frustrated not by separation problems but by detection problems. An example is the problem of determining the many important inorganic ions that are not light-absorbing. Whereas the separation of such transparent ions may be conveniently effected on ion exchange resin columns, their detection and measurement by conventional photometric means are thwarted since they are optically indistinguishable from the transparent eluents commonly prescribed. The technique known as ion chromatography (I, 2) was developed to circumvent the detection problem posed by transparent sample ions. In just 6 years it has become a widely practiced and popular method addressing problems in a great variety of areas (2). Ion chromatography (IC) usually comprises a two-column arrangement followed by a conductance detector where the first column serves to separate the ions of interest while the second column, the suppressor, serves to lower the conductance of the eluent while usually increasing 0003-2700/82/0354-0462$0 1.2510

the conductance of the eluted sample ions. The suppressor column in IC becomes exhausted in the course of normal usage and must be periodically regenerated or replaced-usually regenerated. Whereas this regeneration step has been automated in commercial instruments so that it is relatively unobtrusive or is made continuous as in the recently developed hollow fiber suppressor (3), it would nevertheless be desirable and advantageous for the following reasons to develop single column (suppressorless) methods for the many nonchromophoric ions. (1) Decreased complexity of the instrumentation should yield a concomitant increase in reliability. This is a very important factor in penetrating the process control area with chromatographic methods where the demands for unattended and relatively maintenance-free operation have high priority. (2) Reduced dead volume as a result of eliminating the suppressor will yield faster analysis and somewhat improved resolution. Suppressorless single-column conductometric methods of ion analysis have been described in earlier literature (4-7). The limitations of these approaches have been elaborated by Pohl and Johnson (8)who point to the problem inherent in attempting to determine accurately the often times small changes in eluent conductivity that accompany the replacement of eluent ions by sample ions-both of which are conducting. They argue the sensitivity advantage to be gained by adding a suppressor that effectively "removes" the conductance of the eluent. We would like to report a single column approach that solves the monitoring problem in a different manner while retaining much of the sensitivity of the original ion chromatography method. This new technique is derived from a comprehensive development of the concept that photometers may be used to monitor the many "transparent" ionic species 0 1982 American Chemical Soclety

ANALYTICAL CHEMISTRY, VOL. 54, NO. 3, MARCH 1982

i Cone'"

463

Minutes Ne'

0

12

24

I

1

I

i-

J,

1

I,

I

V O I --P

A

-,;:.

I

VOI

+ B

Flgure 1. Principle of Indirect photometric chromatography.

commonly thought not to be amenable to this type of detection, Although mentioned earlier by Lament and Bourdon (9),lack of sensitivity was expected to limit application of the new idea. The independent discovery and development of the method as reported here demonstrate its abundant versatility, particularly in optimizing sensitivity. From the standpoint of' sensitivity we will also discuss how this novel means of detection overcomes the problems that are intrinsic to suppressorless conductometric monitoring. A feature of this new photometric approach is the use of light-absorbing (usually UV absorbing) eluents, made so by including in the eluent light-absorbing ions of the same charge as the ions to be separated. These light-absorbing ions have a dual role: (1)of selectively displacing the sample ions from the chromatographic column; (2) of revealing the sample ions in the effluent. The appearance of sample ions in the effluent is signaled by "dips" or "troughs" in the base line absorbance of the effluent as the transparent sample ions substitute for the light-absorbing displacing ions. In applying this new approach to the very sensitive detection and determination of ions, it is essential to understand the interplay of such factors as concentration of eluent, concentration of sample, capacity of the ion exchanger, and the optical properties of the eluent. The definition of these critical relationships occupies a large part of this contribution. We have chosen the name indirect photometric chroma tography (IPC) to describe this rapid, sensitive, broad scopk exploitation of ion exchange and photometric monitoring.

PRINCIPLE OF THE METHOD Consider an ion exchange column-for illustrative purposes specifically an anion exchanger--which has been pumped and equilibrated with an electrolyte denoted Na+E- so that the sites in the exchanger are occupied exclusively by eluent ionsi E-. A concentration monitor capable of sensing all ionic species and placed at the outlet of such an operating column would reveal a steady level of Na+ and E-if the feed concentration of the eluent in maintained constant (Figure 1A). If an injection is made of sample electrolyte, denoted Na+S-, then the sample anion, f3-,will generally be retarded by the stationary phase and will exit at a characteristic elution volume determined by such factors as the capacity of the exchanger, the concentration of the solution, and the affinity of the! stationary phase for S- relative to E-. A suitable monitor at the column exit would indicate the concentration of S-to rise and fall in a familiar fashion as it leaves the column (Figure 1B). Conventional ion exchange LC art has been concentrated on devising suitable detectors for directly monitoring the magnitude (height or area) of these sample peaks. Generally ignored has been the fact that accompanying the appearance of S-there must be a concerted and equivalent change in Esince, by the principles of electroneutrality and equivalence of exchange, the total equivalent concentration of anions (Sand E-) must remain fixed since the concentration of sodium coions is fixed. It therefore follows that the concentration of S-in the effluent could be indirectly monitored by continu-

I

a

e

m 0

ID

Figure 2. Separation and indirect photometric detection of several "transparent" sample ions: (a) chloride, (b) nitrite, (c) bromide, (d) nitrate, (e) sulfate.

ously monitoring the level of eluent ion E-. On the basis of this argument it follows how this somewhat latent feature of the ion exchange mode may be usefully tapped in the case of problematical sample ions. Thus, if sample ions are inconvenientlylacking in a particular property, for example, optical absorbance, one may exploit this deficiency in the sample species by deliberately choosing an eluent ion that is light absorbing and monitoring the "troughs" generated in the base line absorbance as transparent sample ions elute. The development of this combination of ion exchange and indirect photometric monitoring is the main concern of this contribution. An example may serve to illustrate how the method works. A column containing an anion exchanger was equilibrated with a dilute M) solution of sodium phthalate until the effluent absorbance was stable as indicated by the UV photometer monitoring the column effluent. When a sample containing chloride, nitrite, bromide, nitrate, and sulfate was injected, the chromatogram of Figure 2 was obtained. By making separate injections of the individual anions in the mixture, we established the identities of the troughs in phthalate absorption. The off-scale positive deflection is due to the ion exchange displacement of phthalate by the injected sample anions as a whole. Since the total equivalent concentration of the sample exceeded that of the eluent, the void disturbance was positive. When the total concentration of the sample is less than that of the eluent, the disturbance is negative. An elution order was established experimentally for a large variety of anions using phthalate as displacing ion. This is recorded in Table I which may be used as a rough guide in predicting the feasibility of certain separations. But it must be used judiciously for, as will be evident later, elution order depends upon a variety of experimental conditions. The feasibility of determining ions by this indirect approach brings a number of questions to mind, notably, how does one choose an eluent species from the enormous number that qualify through being ionic and having appropriate spectral properties? What are appropriate spectral properties? What is the sensitivity of the method and how is it related to the various elution conditions? These questions were addressed in a

464

ANALYTICAL CHEMISTRY, VOL. 54, NO. 3, MARCH 1982

Table I. Elution Volume of Anions’

V,, mL

ion

ion fluoride glycolate iodate maleate malonate nitrate nitrite o-phosphate propionate succinate sulfate sulfite

VE,

mL 3.0 3.4

void 2.0 acetate 3.2 azide 55 3.8 bromate 14 78 bromide 69 62 carbonate 6.5 84 chlorate 86 24.1 chloride 17.1 7.5 monochloracetate 8.0 3.3 dichloracetate 27.5 47 trichloracetate b 114 citrate b 86 cyanate 6.5 a Eluent: M sodium phthalate, M boric acid, pH 9. Column: 4 X 250 miii, SAR-20-0.6. Very large.

I

1

s B

I

Volume

-

A

systematic manner-the

results and recommendations follow.

CONSIDERATIONS IN THE CHOICE OF ELUTING IONS AND CONDITIONS In IPC, sample ions are revealed and quantified by the decrements they produce in eluent concentration. Since the displacing species is usually in much greater abundance than sample species-a feature of elution chromatography-these decrements would ordinarily represent rather small fractional changes in eluent level. Thus, the successful application of IPC is directly related to how precisely we can measure these fractional differences (the signal) in the presence of the random fluctuations (the noise) of the base line response. To this end it is critical to the understanding of IPC to appreciate how signal to noise ratio and, in turn, sensitivity are related to an important variable in the system, namely, the concentration of the eluent. (A) Concentration of the Eluent. Let us consider the case of elution of a sample ion through an anion exchanger operating in the IPC mode. It is assumed that the conditions have been chosen so that the eluted sample is adequately remote from the void disturbance. Typically this would result in a Gaussian-shaped change in sample ion concentration, but for simplicity of treatment we will assume that it emerges as a square wave pulse with maximum concentration, Cs. This pulse of sample will cause a concomitant and identically shaped pulse change in the eluent level as indicated in Figure 3A. The signal to be measured, AS, is the difference between the signals due to the eluent a t base line concentration, CE, and when sample elutes, C E - Cs. This may be expressed as follows: SIGNAL = AS a CsAs (CE - Cs)AE - CEAE = Cs(As - AE) (1)

+

where As and AE denote the absorptivities of the sample and eluent ions, respectively. Equation 1assumes that the signal response is directly proportional to species concentration. Figure 3A depicts an ideal, that is, a noiseless detection situation. Reality is represented in Figure 3B which depicts the noise within which the signal must be detected, that is, the uncertainty in measuring the concentration of eluent CE. At a given base line absorbance, noise represents a fixed (random) fluctuation, represented by N . The signal as a fraction of the base line absorbance is given by the expression

CS(AS - AE) CEAE

(2)

from which it follows that signal Cs(As - AE) noise

NCEAE

(3)

Figure 3.

For transparent ions As is zero so, neglecting signs which are not significant for our purposes signal Cs anoise

NCE

(4)

This simple expression incorporates the conclusion that sensitivity improves, the lower the concentration of eluent employed. There are other important considerations, however, that impose lower limits to the concentration of eluent preferred. Perhaps most significant among them are in the penalties of overlong run times and loss of sensitivity due to band spreading that result from using an eluent that is too dilute. Ideally, the run time should be no longer than t h e time necessary to adequately resolve the troughs. When using LC in an ion exchange mode, there is a further very fundamental consideration to be kept in mind when choosing the concentration of the eluent, that is, the saturation level in the eluting capacity of the eluent. How close one operates to this level, or to what extent one exceeds it, exercises strong control over the sharpness of the eluted peaks and hence resolution. Basically the source of the limitation derives from the inability of eluent to displace sample ion at a higher concentration than that of the eluent. (B) The Relative Affinity of E- and S-. As well as its concentration, the displacing power of the eluent ion with respect to the sample ions is an extremely important factor in the practice of IPC. Different ions vary widely in their displacing power. In an attempt to reduce these eluent options to a reasonable number, we examined a large variety of candidate eluents having a wide range of ion exchange affinities and from these limited our recommendations, somewhat arbitrarily, to just a few. For anion separations, in addition to 0-phthalate, the 1,2-sulfobenzoate, 1,3,5-benzenetricarboxylate (trimesate), and iodide ions are useful displacing species. Trimesats and sulfobenzoate are generally more potent displacing ions than phthalate while iodide is less so. This is illustrated in Figure 4 which is a plot of the elution volume of sulfate ion on a surface agglomerated anion exchanger for the four different eluent species as a function of eluent concentration. Within the concentration range of eluent depicted, the eluting power of the displacing ion is seen to follow the expected trend of polyvalent ions being more potent displacing species than monovalent ions. This order of potency is not always obeyed for, as will be seen shortly, it depends on the charge on the eluent and sample ions and on the concentration of the eluent. (C) The Effect of the Charge of Ex-and SY- on Elution Rate and Elution Order. The elutability of several sample

ANALYTICAL CHEMISTRY, VOL. 54, NO. 3, MARCH 1982

2

void volume of the column. A number of features of the data are noteworthy: (1) the linear dependence of log (VE - V,) on log (concentration of eluent); (2) the differing slopes of the log ( VE - V,) vs. log C plots (There are in fact five different slopes in the 15 plots of Figure 5); (3) the cross-over in certain elution orders. The elution can thus be described in the convenient mathematical form log (VE - V,)

Phthalate

V,-

465

a

-m log [E"-]

where [Er-] is the molarity of the eluent. A relationship of this form, namely log (VE - V,) = constant - y / x log [E"-]

(5)

Sulfobenzoate

may be developed from basic ion exchange theory (IO). This expression, besides accounting for the experimentally observed linearity in the log ( VE - V,) vs. log [eluent] plots, Trimesate also defines the slope as the ratio y / x of the charge of the sample ion to that of the eluent ion. Consequently, it is of k , M interest to examine how the experimentally observed slopes 10 10 3 102M agree with the values predicted by the expression. Accord[Eluent] ingly, the slopes of the plots in Figure 5 were measured and Flgure 4. Elution volumes of sulfate ion using four different eluents. are compared with the values predicted by eq 5. The comResin: 2.8 X 250 mm, SAR-40-0.6. parison is provided in Table I1 and with one exceptiontrimesate eluent, bromide sample ion-the agreement between ions by the four candidate eluents was measured, and the theoretical and observed is good to excellent. results are illustrated in Figure 5. The quantity VE - V, is the corrected elution volume of the ion, that is, the effluent This has an important practical implication. It suggests that from a knowledge of the charges of the sample and eluent volume between sample injection and trough elution less the

\

4

100

B

& ooc

coo-

10

VE-VY

01 104

10'M

103

10'

[Phthalatel

10'M

103

[Trimesatel

100

10

- V"

0.1 10-4

lo2M

103

[Sulfobenroatel

Flgute 5. Elution volumes of sample ions with: (A) phthalate; (8)trimesate; (C) sulfobenzoate; (D) Iodide. Resin: 2.8 X 250 mm, SAR-40-0.6.

466

ANALYTICAL CHEMISTRY, VOL. 54, NO. 3, MARCH 1982

Table 11. Comparison of y / x (Theoretical) with y / x (Observed) S Y-

E X -

I1-

trimesate3trimesatejtrimesate 3 phthalate*-

Br-

so,2-

y/x

y / x obsd

1

1.00 2.00 0.21

2

Br-

113

TPPY -

?

Br-

1/2

0.60 1.9 0.47

1

0.98

Sod2-, S20,z-

s0,z-

213

ions and a single determination of V, at a single concentration of eluent, it is possible to predict V, for that particular species at other eluent concentrations. Furthermore, the dependence of slope on ion charge ratio explains the observed cross-overs in elution order. Knowing the origin of this effect in turn affords us another means of controlling resolution and avoiding the condition of nonresolvability that is represented by a cross-over point.

PHOTOMETRIC FACTORS The precise determination of eluent absorbance (concentration) is a very important part of IPC, so principles that apply to conventional spectrophotometric measurements also apply to IPC. It is known from classical spectrophotometry that the most accurate measurements are obtained when the optical density is 0.43 (11). Actually this value is not critical and the error of measurement varies little within a range of optical density from about 0.2 to 0.8. For this reason it is important to the accuracy of IPC to monitor the eluent under conditions where its absorbance falls within this range. Since the concentration of eluent to be used will generally be dictated by such other considerations as column capecity and eluent ion affinity and since cell path length is fixed, by what means may optical absorbance be adjusted? The molar absorptivity of a given eluent ion generally exhibits so great a dependence upon wavelength, that the desired eluent absorbance is obtained merely via selection of detector wavelength. A photometric detector with multiwavelength monitoring capabilities is therefore a very useful adjunct to IPC although under certain conditions fixed wavelength devices have operated quite effectively. Eluent concentrations as dilute as lo4 M and as concentrated as 1 M have conformed to the “optimum absorbance” requirement of IPC through appropriate choice of detection wavelength. APPLICATIONS Experimental Section. The apparatus used in IPC is conventional, usually consisting of an eluent reservoir, a pump, a sample injection valve, an ion exchange column, a flowthrough photometric detector, and a recorder. A trough integrator is optional. Any of a large variety of LC pumps is suitable-in this work the Laboratory Data Control (LDC) Constametric I, the LDC minipump, and the Altex Model llOA were used. Photometers found to be useful were two fixed-wavelength types, the LDC Model 1203 and the Altex 153, and two multiwavelength instruments, the Perkin-Elmer LC-75 and the Varian UV-50. Note that the photometric detector responds to small fractional reductions in an elevated base line as solute ion zones traverse the optical cell. This feature calls for techniques to suppress the base line level without reducing or otherwise obscuring the small troughs. We have achieved this in two ways. In dual beam instruments with reference cells we have operated with the reference cell containing either air or eluent. Operating with air in the reference cell we found that some of the commercially available photometers did not have enough zero adjustment to null out the high absorbance of some eluents, and we found it necessary, therefore, to develop

additional base line suppression circuitry to handle this problem. Operating with eluent in the reference cell avoids the problem. In a few cases we found careful temperature control of the column to be necessary. In one eluent system-the trimesate at pH 10-we found an exceptional sensitivity to temperature change in the column. For example, it would manifest itself in marked biphasic base line disturbances on even slight warming of a short section of column as might be brought about by briefly gripping the column between finger and thumb. Enclosing the column in a thermostating bath maintained at 37 f 0.1 OC eliminated this problem. As a rule, however, we found such close temperature control to be unnecessary. The eluent in IPC is only slightly changed on passing through the column and may be recycled to the eluent reservoir with negligible penalty. This ability to recycle eluent is desirable in unattended process control applications. Stainless steel columns (4.1 X 250 mm) or glass columns (2.8 X 250 mm) were used to contain the ion exchangers, and sample was introduced via a Rheodyne Model 7010 injector. The column packings are one of the most important features of IPC and a variety of materials are effective. We have made wide use of ion exchangers originally developed for ion chromatography (1,12-14). Especially useful are the surface agglomerated pellicular anion exchangers which are prepared by depositing a monoparticulate layer of submicron anion exchanging spheres onto a much larger (20-50 pm) substrate bead whose surface is anionic. In the original IC work ( I ) , surface sulfonated styrene divinylbenzene copolymer particles were much used as substrate spheres but in the present research, conventional strong acid cation exchangers such as Dowex 50 have been used exclusively for this purpose. The very small particle anion exchangers of uniform particle size distribution were prepared by quaternizing emulsion copolymers of vinylbenzyl chloride and divinylbenzene (14). Surface agglomeration was carried out by adding a quantity of the substrate cation exchange resin to a suspension of the colloidal anion exchanger. The anion exchange capacity of these surface agglomerated separating resins is controlled by the size of the colloidal particles and by the size of the substrate, being greater the smaller the substrate size and decreasing as the colloidal resin size decreases. Surface agglomerated resins are described in the text by a code that indicates the sizes of the substrate and of the colloidal anion exchanger. For example, the designation SAR-20-0.6 denotes a surface agglomerated resin prepared by coating a 20 p m diameter cation exchanger with a 0.6 pm diameter colloidal anion exchanger. Ion exchangers found to be useful in IC for cation analysis are also useful in IPC applications. In this regard, surface sulfonated styrene divinylbenzene copolymer spheres (12) have been used to develop a number of IPC analysis schemes. Besides these low capacity ion exchangers, high specific capacity materials have been successfully applied to IPC. They are especially useful in applications where direct injection of concentrated samples is desirable and column overloading becomes an important consideration. A number of commercial anion and cation exchangers have been used with success. The following applications have been chosen to exemplify the scope, selectivity, and sensitivity of IPC. Details on columns, packings, eluents, etc. are provide in Table 111. Anion Separations. Figure 6 represents the separation of a mixture of five anions. Noteworthy is the trough produced by carbonate ion. This ability to determine anions of high pK acids gives IPC an advantage over ion chromatography which by its nature is very insensitive to such species.

-

ANALYTICAL CHEMISTRY, VOL. 54, NO. 3, MARCH 1982

467

Minutes

Minutes

12

0

O+-r*

24

24

r--JF-l

0

Figure 6. IPC of several anions: (a)carbonate, 1.8 pg; (b) chloride, 1.4 pg: (c) phosphate, 3.8 pg; (d) azide, 5.0 pg; (e) nitrate, 10 pg. Minutes 0

12

6

0

6

12

n

Figure 8. Determination of sulfate in 1% sodium chloride: (A) 100 ppm sulfate: (B) 10 ppm sulfate. Minutes

0 I

A.

2 I

4 I

6 I

8 I

S.

Figure 7. Determination of nitrite In chloride: (A) 58.5 ppm chloride, 4.6 ppm nitrite: (B) 585 ppm chloride, 4.6 ppm nltrlte. Figures 7 and 8 show the determination of nitrite in chloride and of traces of sulfate in a high background of salt. Both illustrate the excellent selectivity and sensitivity of which IPC is capable. As discussed earlier, sensitivity may be improved by decreasing eluent concentration. An example of this is provided in Figure 9. The fairly potent sulfobenzoate ion was chosen as the displacing specieei so that relatively low concentrations would suffice to elute the sample ion, in this case sulfate, with reasonable speed. An appropriately low capacity column was also used. The trough due to sulfate and showing good signal to noise resulted from 11. 100-pLinjection of lo4 M sodium sulfate. The detectability of sulfate under these conditions is approximately 1 ng, attesting to the high sensitivity attainable by the technique. IPC has exceptional capabilities for handling ionic species with high affinity for anion exchangers. Noteworthy is the rapid elution of the polyphosphate species (Figure lo), an ion that is normally very difficult to displace. Cation Separations. A number of cation separation schemes were developed by using the IPC approach. Figure 11illustrates a rapid separation of sodium, ammonium, and potassium. This separation is noteworthy in that it was ob-

Figure 9, Sensltlvlty of IPC. Sulfate peak due to a 0.1 mL injection of lo-' M sodium sulfate. tained on a very small column (2.8 X 20 mm) of high specific capacity cation exchanger. The eluent was copper sulfate (0.01 M), copper being the UV-absorbing displacing ion. A column and eluent such as this have been used to determine small amounts of sodium and potassium in concentrated (20%) calcium chloride solution. Only moderate dilution of the sample is required in view of the high capacity of the resin employed. After the monovalent ions had been eluted, the resin was flushed briefly with a concentrated (1M) solution of copper nitrate in order to displace divalent ions which would otherwise have appeared at a much later time and interfered with subsequent chromatograms. We found it convenient to introduce this purge solution by way of another sample injection valve equipped with a large (0.5 mL) loop. A separation of sodium, potassium, calcium, and magnesium (Figure 12) was achieved by using a split column technique wherein two columns of equal length but containing resins of different specific capacities were connected in series and appropriately switched. Joint Anion and Cation Determination. We have demonstrated that indirect photometric chromatography may be

468

ANALYTICAL CHEMISTRY, VOL. 54, NO. 3, MARCH 1982

Table 111. Details on Chromatographic Conditions in Various Separations Figure 2 column: 4 X 250 mm, SAR-40-0.6 eluent: lou3M sodium phthalate, pH 7-8 flow rate: 2 mL/min; sample size 0.02 mL detector: LDC 1203, 0.032 AUFS M in each of chloride and sample: 2 X nitrite; 5 X 1 O - j M in each of bromide, nitrate, and sulfate Figure 6 column: 4 X 250 mm, SAR-40-0.6 M sodium phthalate, M boric eluent: acid, pH 1 0 flow rate: 5 mL/min: sample vol 0.02 mL detector: Perkin-Elmer LC-75, 0.32 AUFS Figure 7 column: 4 x 250 mm, SAR-40-0.6 M sodium phthalate, M boric eluent: acid, pH 9 flow rate: 2 mL/min; sample vol 0.02 mL detector: LDC 1203, 0.32 AUFS Figure 8 column: 4 X 250 mm, SAR-40-0.6 M sodium phthalate, M boric eluent: acid, pH 9 flow rate: 2.5 inL/min; sample vol 0.1 mL detector: LDC 1203 sensitivity: (A) 0.016 AUFS; (B) 0.004 AUFS Figure 9 eluent: M sodium sulfobenzoate pH 8 flow rate: 1 mL/min detector: Varian UV-50 set at 224 nm sensitivity: 0.005 AUFS Figure 10 column: 2.8 X 250 nm, SAR-40-0.6 eluent: M sodium trimesate, pH 8 flow rate: 1mL/min; sample size 0.02 mL sample: 2 g/L potassium pyrophosphate, 2 g/L sodium tripolyphosphate detector: Varian UV-50 set at 296 nm sensitivity: 0.02 AUFS Figure 11 column: 2.8 X 20 mm, Dowex 50 resin eluent: 0.01 N copper sulfate flow rate: 0.7 mL/min; sample size 0.02 mL sample: 0.01 N in each of sodium, ammonium, and potassium detector: Varian UV-50 set at 252 nm sensitivity: 0.05 AUFS Figure 1 2 first column: 2.8 x 250 mm surface sulfonated styrene-DVB 0.015 mequiv/g second column: 2.8 x 250 mm surface sulfonated styrene-DVB 0.087 mequiv/g M copper sulfate eluent: 1.25 X detector: Perkin-Elmer LC-75 set at 216 nm Figure 1 3 columns: 4.6 X 250 mm Partisil SAX followed by 4.6 X 250 mm Partisil SCX eluent: 5 X M copper nitrate detector: Perkin-Elmer LC-75 set at 241 nm sample: 0.02 mL; 0.2 M NaF, 0.2 M RbCl, 0.1 M MnC1, extended to simultaneous joint anion and cation analysis by combining chromophoric anion and cation mobile phase ions with suitable ion exchange columns. For extension of IPC to combined analysis of anions and cations in a single chromatograph, a special eluent must be chosen. In accordance with the principles already outlined, the eluent for analysis of, say, only anions in a sample has a 2-fold function: to displace anion bands individually from the column and to render them detectable as transparencies in contrast to eluent anion UV absorbance. Joint analysis of anions and cations with a single eluent and UV detection, then, requires mobile phase anion and cation both with UV absorbance and appropriate sample elution power. A necessary further consideration is that mobile phase anion and cation each contribute approximately equally to the absorbance a t base line since we have seen already how sensitivity to eluting sample ions relates to mobile phase absorbance. Taking the various factors into account, we found copper nitrate to be a suitable eluent. As determined by a Cary 15 spectrophotometer at 241 nm, eCu = 37.5 L equiv-l cm-l and

Minutes

2

0

4

PP

Flgure 10. Separation of pyro- (PP) and tripoly- (TPP) phosphates.

NO^ = 77.5 L equiv-' cm-l. I t follows that a 5 X M Cu (NO,), solution would produce a base line absorbance of

ATOT= Acu + A N O=~ (ecu

+ eNOJ

X

normality X pathlength

and so

AToT = 1.15 in a detector with a 1-cm pathlength. This is an acceptable base line absorbance value and thus a 5 x M copper nitrate eluent was selected along with a detection wavelength of 241 nm. The separating columns chosen were commercially available strong anion and cation exchange columns arranged in series. The cation exchanger was 4.6 X 250 mm Partisil 10-SCX from Whatman containing 10-pm microparticulate packing with siloxane-bonded sulfonic acid exchange groups. The anion exchanger was a 4.6 X 250 mm Partisil 10-SAX with 10-pm microparticulate packing and siloxane-bonded quaternary anion exchange sites. The chromatogram obtained from an injection of a synthetic mixture comprising 0.2 M NaF, 0.2 M RbC1, and 0.1 M MnClz is shown in Figure 13. The first three troughs are the deficiencies in the copper absorbance caused by the emergence of the three sample cations, the nitrate absorbance (concentration) remaining constant within this region. The last two troughs are the deficiencies in nitrate absorbance due to sample anions while the copper absorbance remains constant. Calibration. Calibration runs for the three ions, sulfate, nitrate, and phosphate yielded curves that indicate a convenient linear dependence of trough depth on the amount of ion injected. There is an interesting aspect to calibration in the IPC mode in that for many ions the area of the trough is not dependent on the ion injected but only on its amount. This is a natural result of the method of monitoring since each equivalent of sample ion displaces the same amount of monitoring ion from the mobile phase irrespective of the sample ion. To demonstrate this, separate injections of accurate amounts of nitrate, sulfate and phosphate were eluted by sodium phthalate (pH 8) and the areas of the troughs measured. The results are shown in Table IV. The area of trough per equivalent of ion is indeed approximately independent of the ion injected-for these three ions. On the basis of this observation we expect anions to adhere to this rule. Anions of acids with medium to high pKs should give responses determined by their valence at the ambient pH of the eluent. Phosphate, for example, exists predominantly as the HPOd2-species a t pH 8 so that 1mol

ANALYTICAL CHEMISTRY, VOL. 54, NO. 3, MARCH 1982

--

Minutes

Minutes

0

3

469

10 I

0

6

20 I

30 I

40 I

SI

I;(

CI

+t

N ti:

Na'

Flgure 13. Joint determination of anlons and cations by IPC.

,

a+

Flgure 11. IPC of cations. Minutes

5

0

10

7-77

ACKNOWLEDGMENT The authors have appreciated the experimental assistance of D. F. Scheddel during the course of this work.

I K+

Na+

Flgure 12. Separation of mono- and divalent cations by a split column technique.

--

Table IV. Calibration Data for Nitrate, Sulfate, and Phosphate area of trough/ area of trough (mequiv/L) of injection (arbitrary units) ion injected

5 x 10-3 M sodium nitrate 2.5 x 10-3 M sodium sulfate 1.67 x 10-3 M sodium orthophosphate

tometric chromatography as we have called it, shows considerable potential in the area of inorganic and organic ion analysis. In IPC, detection of the eluting sample is accomplished by monitoring a change in property of the column effluent as the eluent ion is displaced by the sample ions. Clearly, sensitivity in such a case is impaired when the eluent and the sample ions possess this property to a comparable extent. Such is the case with suppressorless conductometric monitoring-both sample ion and eluent ion are conducting and the closer the values of their equivalent conductances the poorer the sensitivity for that sample ion. In IPC on the other hand one chooses a monitoring wavelength where the displacing ion is absorbing and the sample ion is not, hence satisfying a condition for obtaining maximum sensitivity. This we claim places IPC in a superior position among suppressorless ion chromatographic methods. Indirect photometric chromatography is a promising new approach to a number of ion analysis problems.

117.5

23.5

111.0

22.2

80.4

24.1

of phosphate injected would be expected to displace 2 equiv of monitor ion. The data of Table IV support this expectation.

CONCLUSIONS This work has demonstrated that photometers may be used as chromatographic detectors for accurate, sensitive determination of transparent ionic species commonly considered photometrically undetectable. The technique, indirect pho-

LITERATURE CITED (1) Small, H.; Stevens, T. S.; Bauman, W. C. Anal. Chem. 1975, 4 7 , 1801-1809. (2) Small, H. In "Applications of Ion Chromatography In Trace Analysis"; Lawrence, J. F., Ed.; Academic Press: New York, 1981. (3) Stevens, T. S.; Davis, J. C.; Small, H. Anal. Chem. 1981, 53, 1488-1492. (4) Gjerde, D. T.; Fritz, J. S. J . Chromatogr. 1979, 176, 199-206. (5) Gjerde, D. T.; Fritz, J. S.; Schmuckler, G. J . Chromatogr. 1979, 186,

509-519. (6) Gjerde, D. T.; Schmuckler, G.; Fritz, J. S. J . Chromatogr. 1080, 787, 35-45. (7) Fritz, J. S.; Gjerde, D. T.; Becker, R. M. Anal. Chem. 1980, 52, 1519-1522. (8) Pohl, C. A.; Johnson, E. L. J . Chromatogr. Scl. 1980, 18, 442-452. (9) Laurent, A.; Bourdon, R. Ann. Pharm. Fr. 1078,36(9-IO),453-460. (10) Ringbom, A. "Complexation In Analytical Chemistry"; Interscience: New York. 1963;pp 198-199. (11) Kolthoff, I. M.; Sandeli, E. B. "Textbook of Quantltative Inorganic Analysis": Macmillan: New York, 1949;pp 862-668. (12) Small, H.; Stevens, T. S. U S . Patent 4 101 460,July 18, 1978. (13) Stevens, T. S.; Small, H. J . Llq. Chromafogr. 1978, 1 (2),123-132. (14) Small, H.; Solc, J. Proceedings of an International Conference on "The Theory and Practice of Ion Exchange"; Streat. M., Ed.; The Soclety of Chemlcal Industry: London, 1976;pp 32-1 to 32-10.

RECEIVED for review September 16,1981. Accepted November 6, 1981. The methods described in this publication are the subject of pending patents which are licensed to Dionex Corporation for commercial use.