Anal. Chem. 1987, 59, 1483-1485
mental chamber, the material gained weight up to 1.8% or lost up to 0.15% depending upon the conditions employed. No significant weight gain was observed at 55% RH or lower. When a single exposure to 60% RH or more occurred, approximately 80% of the observed gain could be lost by reequilibration at 35% relative humidity. Repeated or cyclic exposures, however, led to a permanent weight gain. A previously untreated sample of SRM 329 was then subjected to three cycles of 40% to 80% to 40% RH at 37 OC with a 1-h exposure at each condition. The sample was then dried for 1 h a t 105 "C as recommended by the certificate and a weight loss of 1.72% was recorded. Of this, 1.66% may be accounted for by the weight gain during the humidification/dehumidification experiment. Thus, the net loss on drying is only 0.06% compared to 0.39% for the untreated SRM. This discrepancy of 0.33% is quite significant and would invalidate the certificate values by the same relative amount. Thus, conditions have been found that for some users could lead to erroneous results. SRM 329 is a zinc sulfide which slowly oxidizes to zinc sulfate. The mechanism of this reaction is not clear but high-humidity conditions are at least one of the probably causes. SRM 329 had been withdrawn from sale as a result of this study and customer feedback. In addition, we have used the apparatus to assist in developing sample processing methods for new SRM's where there was insufficient prior experienceto predict the behavior of the material under dry to humid conditions. A sample of 10 g will yield precision of *0.01% with a balance of 1 mg precision. A more sensitive balance would permit either better precision with the same sample size or the use of smaller
1483
samples. Smaller sample sizes should achieve a more rapid equilibration. This apparatus differs from other approaches primarily because of the dynamic nature of the sample equilibration caused by forced air movement. A smaller chamber should permit better humidity control and perhaps faster sample equilibration. The present apparatus was only meant to demonstrate the feasibility of the experiment with available apparatus. For reference materials, ideal behavior with respect to repeatable moisture content can be demonstrated by the cyclic humidificationfdehumidification process described here. Other processes, such as bacterial degradation or oxidation, occur over a longer time scale and may be influenced by the presence of water. This procedure does not address those problems.
LITERATURE CITED (1) Michaelis, R. E. The SRM Story at NBS; NBS Special Pubiicatlon 408, National Bureau of Standards: Washington, DC, 1975; pp 246-257. (2) Ward, R., Chairman Non Stoichiometric Compounds; 141st Meeting of the Amerlcan Chemical Society; Advances in Chemistry Series No. 39; American Chemical Society: Washington, DC, 1963. (3) Barnes, I. L.; Murphy, T. J., personal communication, National Bureau of Standards, Gaithersburg, MD, M a y 1986.
RECEIVED for review November 18,1986. Accepted February 3, 1987. Certain commercial equipment, instruments, or materials are identified in this paper to specify adequately the experimental procedures. Such identification does not imply recommendation or endorsement by the National Bureau of Standards, nor does it imply that the materials or equipment identified are necessarily the best available for the purpose.
Indirect Cationic Chromatography with Fluorometric Detection Jeffrey H. Sherman and Neil D. Dadelson*
Department of Chemistry, Miami University, Oxford, Ohio 45056 Indirect photometric chromatography (IPC) is a technique by which analyte separation is effected through an ion exchange process, following which, analyte detection is achieved through a photometric process. IPC uses light-absorbingions in the mobile phase, with which photometrically inactive, injected sample ions substitute in the column effluent as sample elution occurs causing a decreased absorbance at the detector and negative peaks to be recorded. A number of different eluents such as Cu(I1) (1, 2 ) , benzyltrimethylammonium ion (3-5), picolinic acid (6, 7), and benzylamine (7) have been used for the separation of either inorganic or organic cations by IPC. The advantage of IPC is that it enables chromatographs equipped with UV detectors to perform ion chromatography. There are two major ways peaks in chromatograms generated by IPC can be obscured. Fmt, some analyte ions might absorb light a t the wavelength of detection. Second, the sample matrix may contain an interferent which could also absorb at this wavelength. However, few alternative detection modes for IPC have been reported (8). Indirect fluorescence chromatography (IFC) using sodium salicylate as the eluent has been reported for the separation and detection of anions (9, 10). Detection in IFC is achieved by simply monitoring the column effluent at the emission wavelength of the counterion. As an analyte ion elutes, the background fluorescence of the mobile phase decreasesand a negative peak is recorded. However, none of the eluents cited previously for IPC would be expected to be good candidates for IFC of cations. 0003-2700/87/0359-1483$01.50/0
The fluorescence of aqueous solutions of cerium(II1) has been known for some time and its use in analytical chemistry has been reported (11-15). Prevously in our laboratory, cerium(II1) was characterized as an eluent for indirect photometric chromatography of cations (16). We now wish to extend the utility of Ce(II1) to include serving as a counterion for indirect fluorescence chromatography. The advantage of IFC over IPC for the determination of Na+, NH4+,and K+ in urine is demonstrated.
EXPERIMENTAL SECTION Reagents. All solutions were prepared with doubly distilled, deionized water and were kept in Pyrex glassware. Cerium mobile phase solutions were prepared from the sulfate salt unless otherwise indicated, while all sample solutions were prepared from the respective chloride salt. These salts were obtained from a variety of sources and were reagent grade or better in quality. Instrumentation. The chromatographic system consisted of an IBM LC/9533 ternary gradient liquid chromatograph and operator station (IBM Instruments, Danbury, CT). An IBM Model 9522 (254 nm) UV detector was employed for IPC and a Kratos Spectroflow 980 programmable fluorescence detector was used for IFC (Kratos Analytical Instruments,Ramsey, NJ). The chromatograms were recorded on a Fisher Recordall Series 5000 recorder (Houston Instruments, Austin, TX). The separations were effected with a 10 cm X 3.2 mm i.d. ION-210 Transition Metals column (Interaction Chemicals, Mountain View, CA) at room temperature. The column was packed with 5 pm sulfonic acid derivatized poly(styrene4ivinylbenzene) resin particles. The capacity of this strong cation exchanger was about 500 pequivjg. 0 1987 American Chemical Society
1484
ANALYTICAL CHEMISTRY, VOL. 59,NO. 10,MAY 15, 1987 MINUT E S INJ
4
MINUTES B
INJ
e
4
T
0 ,002 REL FLUOR
1
Flgure 1. Separation of (1)sodium, 0.16 ppm, (2)ammonium, 0.15 ppm, (3) potassium, 0.21 ppm, (4) rubidlum, 0.71 ppm, and (5)cesium, 1.2 ppm, ions with indirect fluorescence detection: eluent, 0.01 mM Ce(II1); flow rate, 1.0mL/min; sample volume, 20 pL.
Table I. Comparison of Detection Limits? UV Fluorescence sample
ppm
pmol
ppm
pmol
Na+ NH4* K+ Rb'
0.003 0.004 0.01 0.07 0.16
2.6 3.5 5.0
0.004
3.5 5.3 10.0
CS'
14.0 24.0
0.006 0.02 0.1 0.2
I' Figwe 2. Separation of (1)sodlum, (2)ammonium, and (3) potassium ions in two dlfferent urine samples, indirect UV detection: sample dilution, 1 to 1000: eluent, 0.01 mM Ce(II1): flow rate, 1.0 mL/min; sample volume, 20 pL.
vs.
u VC
fluorescenceb
R
20.0
OCalculation based on three times the noise signal. [Ce(III)] = 0.01 rnM. ICe(III)l = 0.1 mM (from ref 16).
0 0 0 1 AU
Because the optimum excitation and emission wavelengths for Ce(II1) were determined to be 247 and 350 nm, respectively, the detector was operated with the excitation wavelength set to 247 nm and was equipped with a 370 nm long pass emission filter.
RESULTS AND DISCUSSION It was determined that a 0.01 mM Ce(II1) mobile phase offered the best separation of monovalent cations in the shortest amount of time. Higher concentrations up to 0.1 mM of cerium resulted in a noisy base line, as well as a high, difficult to offset background. Figure 1 shows a separation of the alkali metals and the ammonium ion, in 7 min, with practically base-line resolution for each component. An attempt was made to include lithium in this separation by decreasing the Ce(II1) concentration; however, lithium still continued to elute in the column void, while the peaks for rubidium and cesium became excessively broad. Table I displays the detection limits for the Ce(II1) eluent system with both fluorescence and ultraviolet indirect detection. The fluorometric detection limits are quite similar to those afforded by UV. Although the smaller concentration of Ce(II1) in the mobile phase for fluorescence should give a lower detection limit than that for IPC (I), the background noise of the fluorescent signal offsets this advantage. Additionally, the low ionic strength of the IFC eluent results in longer retained, broader peaks which become difficult to distinguish from the background noise at low sample concentrations. A lower capacity cation exchange column could possibly be used to alleviate this problem; however, retention of the small alkali metals might be compromised. However, it should be pointed out that these detection limits in Table
i T
30.0
v
-++T+-F
INJ
4
HI NUTES
A
+-+T+F 4
INJ
MINUTES
0
Flgure 3. (A) Separation of sodlum, ammonium, and potassium ions in a third acidified urine sample (pH 2.5 wRh HCIO,), indlrect UV detection: sample dilutlon, l to 1000:eluent, 0.01 mM Ce(CI0,h; pH 3.0 adjusted with HCIO,; flow rate, 1.0 mL/mln: sample volume, 20 pL. (B)Chromatogram of 10 ppm caffeine (pH 2.5 with HCIO,). Other conditions are as in 3A. Explanation of chromatograms is in text.
I are comparable with if not better than those obtained by conductivity methods (17). Because fewer organic compounds fluoresce than absorb UV light, IFC should be useful for samples possessing a complex matrix. An example of this situation is the separation and detection of sodium, ammonium, and potassium ions in urine. Figure 2 shows the separation of two diluted urine samples from two different people by using indirect UV detection at 254 nm. Clearly, there is a significant W-absorbing interference in the sample matrix for both samples although the magnitude of the positive peak does vary. This is not surprising since the composition of urine varies dependent on diet, medication, and other factors. The major component of the UV interference was suspected to be caffeine. To attempt to eliminate this peak, both the sample and the
Anal. Chem. 1987, 59, 1485-1488
1485
ammonium, and potassium ion chromatogram of a second aliquot of the urine sample in Figure 2B this time using indirect fluorescence detection. No interference of the separated cation peaks was noted. Quantitation of the peaks in Figure 4 resulted in the following values: Na+, 980 ppm; NH4+,490 ppm; K+, 850 ppm. While the concentration of sodium ion in this specimen was slightly low, the amounts of ammonium and potassium ions are well within expected values (19). A second urine specimen was found to contain values of 2400 ppm of Na+, 140 ppm of NHd', and 2270 ppm of K'. These values are also within normal accepted limits with the exception of the ammonium ion. Therefore, indirect fluorescence chromatography can be a rapid, viable method for determining the levels, of sodium, ammonium, and potassium ions in urine without prior sample cleanup.
Figure 4. Separation of (1) sodlum, (2) ammonium, and (3) potasslum ion In urine, indirect fluorescence detection: sample dilution, 1 to 1000; eluent, 0.01 mM Ce(II1); flow rate, 1.0 mL/min; sample volume, 20
ACKNOWLEDGMENT Donation of the cation exchange column by James Benson of Interaction Chemicals, Inc., is gratefully appreciated. The authors thank Thomas Trosper of Kratos Analytical for the use of the fluorescence detector.
PL.
Ce(II1) mobile phase were acidified to convert any aromatic amines such as caffeine into cations which would then be retained on the column. Figure 3A shows a chromatogram of a third acidified urine sample analyzed by IPC using a cerium perchlorate mobile phase adjusted to pH 3.0. The perchlorate salt of cerium(II1) was used here, as acidification of the mobile phase sulfate salt with sulfuric acid results in the formation of sulfato-cerium complexes which changes the eluting characteristics of Ce(II1). The negative peak eluting a t 10 min was found to correspond to caffeine as shown by comparison to the chromatogram of a standard solution of caffeine (Figure 3B). A positive peak is still evident (Figure 3B) due to the unprotonated form of caffeine which, having a high UV absorbance, adds to the base-line absorbance of the Ce(II1) mobile phase. Although caffeine was identified as one possible interferent, it is likely other poorly retained compounds such as organic anions identified previously in urine (18)also will contribute substantially to the problem. In any case, quantitation of sodium, ammonium, and potassium was still not possible by IPC. Figure 4 shows a sodium,
(1) (2) (3) (4) (5) (6) (7) (8)
LITERATURE CITED Small, H.; Mlller, T. E. Anal. Chem. 1982, 5 4 , 462-469. Larson, J. R.; Pfeiffer, C. D. Anal. Chem. 1983, 55, 393.
Larson, J. R.; Pfeiffer, C. D. J . Chromtogr. 1983, 259, 519. Iskandaranl, 2.; Miller, T. E., Jr. Anal. Chem. 1985, 57, 1591. McAleese, D. L. Anal. Chem. 1987, 5 9 , 541. Benson, J. R.; Woo, D. J. J . Chromatogr. 1984, 22, 386. Foley, R. C. L.; Haddad, P. R. J . Chromatogr. 1986, 366, 15. Dasgupta. P. Ion Chromatography; Tarter, J. G., Ed.; Marcel Dekker: New York, 1987; Chapter 6, p 250. (9) Mho, S.; Yeung, E. S. Anal. Chem. 1985, 57, 2253. (10) Takeuchi, T.; Yeung, E. S.J . Chromatogr. 1988, 370, 83. (11) Kirkbright, G. F.: West, T. S.;Woodward, C. Anal. Chim. Acta 1966,
36, 293. (12) Katz, S. S.; Pitt, W. W., Jr. Anal. Lett. 1972, 5(3),177-185. (13) Wolkoff, A. W.; Larose, R. H. J . Chromatogr. 1974, 99, 731-743. (14) Katz, S.; Pitt, W. W.. Jr.; Mrochek, J. E.; Dinsmore, S.J . Chromtogr. 1974, 101, 193-197. (15) Lee, S.H.; Field, L. R. Anal. Chem. 1984, 56, 2647. (16) Sherman, J. H.; Danielson, N. D. Anal. Chem. 1987, 59, 490. (17) Jupille, T. Am. Lab. (FairfeM, Conn.) 1988, 17, 114. (18) Miyagl, H.; Miura, J.; Takata, Y.; Kamitake, S.; Ganno, S.; Yamagata, Y. J . Chromatogr. 1982, 239, 733. (19) Modern Urine Chemistry; Ames Company, Division of Miles Laboratories, Inc.: Elkhart, IN, 1982; pp 96-98.
RECEIVED for review November 10,1986. Accepted February 17, 1987.
Low-Volume Fluorescence Detector for High-Performance Liquid Chromatography Alain Berthed,' Kuang Pang Li,2 Tiing Yu, and James D. Winefordner*
Department of Chemistry, University of Florida, Gainesville, Florida 3261 1 The need for a variety of suitable detection techniques has induced many researchers to develop detectors for highperformance