ANALYTICAL CHEMISTRY, VOL. 51, NO. 14, DECEMBER 1979
Table I. Analysis Results of Boron Powder Using Different Filtering Techniques extraction method diol diol
diol water water water water
filter
boric acid, 5%
Whatman 40a Whatman 40 Whatman 4U Whatman 40 Whatman 40 Whatman 40
0.051 0.050
0.45-pm Metricelb 0.4 5-1m Metricel
water water
0.068 0.084 0.055 0.051
0.057
0.45-pm
Metricel 0.052
acid-base tit rat ion a
0.050 0.099
W. R. Balstron Ltd.
Gelman Instrument Company.
powder. The water then was evaporated a t room temperature and the samples were extracted with diol solution, after which the original boron powder and the samples spiked with boric acid were analyzed. The data that were obtained are given in Table 11. These data have been corrected for the original boric acid content of the boron powder. The actual boric acid content of the spiked samples was in close agreement with the theoretical content. The extraction efficiency also was evaluated by determining the boric acid content of boron which had been recovered from a previous extraction and boric acid determination. No boric acid was found in this boron, indicating the boric acid had been quantitatively extracted during the first determination. T h e determination of boric acid using curcumin is unaffected by the presence of other compounds, except for fluoride and nitrate ions. A procedure for removing the nitrate interference is reported by Hayes and Metcalfe ( 2 ) . No satisfactory method has been found to prevent the interference caused by fluoride.
Table 11. Extraction Efficiency of 2- Eth yl-l,3-Hexanediol
LITERATURE CITED
boric acid, w g
sample 1
added 250
2 3
250 2 50
recovered recovery, % 256 102.1 246 9s.4 258
average
2403
103.2 101.3
water. Careful filtration of the samples, therefore, is necessary before analysis. T h e extraction efficiency of 2-ethyl-1,3-hexanediol was evaluated by adding 1mL of 500 ppm aqueous boric acid and 1 drop of 10% NaOH to accurately weighed samples of boron
Spicer, G. S.; Strickland, J. D. H. J . Chem. SOC. 1952, 4644. Hayes, M. R.; Metcalfe, J. Analyst (London) 1962, 87, 956. Kowalenko, C. G.; Lavkulich. L. M. Can. J . Soil Sci 1976, 56, 537. Uppstrom, L. R. Anal. Chlm. Acta 1968, 43, 475. Ostling, G. Anal. Chim. Acta 1975, 78, 507. Agazzi, E. J. Anal. Chem. 1987, 39, 233. Peterson, H. P.; Zoronski, D. W. Anal. Chem. 1972, 4 4 , 1291. Mair, J. W.; Day, H. G. Anal. Chem. 1972, 4 4 , 2015. Feldman, C. Anal. Chem. 1961, 33, 1916.
RECEIVED for review May 25, 1979. Accepted August 20,1979. This work is a result of work performed by the Bendix Corporation a t the Kansas City Division which is operated for the U S . Department of Energy under Contract Number DEAC04-76-DP00613.
Evaluation of Solute Vaporization Interference Effects in a Direct Current Plasma G. W. Johnson, H. E. Taylor,' and R.
K. Skogerboe"
Department of Chemistry, Colorado State University, Fort Collins, Colorado 80523
Quantitative measurements by atomic emission spectrometry depend on the production of free and excited atom populations. Processes which cause shifts in these populations may be broadly labeled as physiochemical interference effects. Solute vaporization effects are included among these. These originate from the conversion of the analyte population(s) to compounds which exhibit some degree of stability in the excitation medium. Classical examples include the suppression of calcium atom populations when phosphorus and/or aluminum are present (2-8). Several previous investigations (4-12) dealing with dc plasma excitation sources have indicated that both phosphorus and aluminum cause supression of calcium excitation. These reports have also shown that the extent of the interferences observed is dependent on operational conditions used. A previous publication (13) from this laboratory has demonstrated that a three-electrode dc plasma can be used with a single set of operating conditions for the simultaneous determination of 18 elements in natural and effluent waters a t requisite concentration levels. Although it was shown in that report that reliable analyses were obtained for reference water 'Present address, U S . Geological Survey, 5293 Ward Road, Arvada, Colo. 80002. 0003-2700/79/0351-2403$01.00/0
samples having reasonably diverse compositions, data that specifically addressed the possibility of solute vaporization interference effects were not presented. T h e present report summarizes the results of experiments used to define: the significance of the interferences of phosphorus on barium, calcium, and strontium emission; the effects of excess barium, calcium, and strontium on phosphorus emission; the interference effect of calcium on aluminum emission; and the extent of the aluminum effect on calcium emission.
EXPERIMENTAL Apparatus. The Spectraspan I11 (Spectrametrics, Inc.) dc plasma spectrometer was used. The unit consists of a dc power supply, a three-electrode plasma torch ( I ' $ ) , a gas regulation and sample nebulization system, a direct-reading echelle spectrometer, and a microprocessor-based control and data acquisition system (13). The operating conditions used in the present study have been described (13). Reagents. All test solutions were prepared from 99.999% pure (or better) metals, oxides, or carbonates. Master solutions were prepared by dissolving the solids in doubly redistilled nitric acid and diluting to volume with distilled-deionizedwater to maintain a final HN03 concentration of 0.1 M. Appropriate dilutions of the master solutions were made with the same water maintaining the same " 0 3 concentration. IC 1979 American Chemical Society
2404
ANALYTICAL CHEMISTRY, VOL. 51, NO. 14, DECEMBER 1979
Table I. Effects of Phosphate on Ba, Ca, and Sr Analyses percent deviation from t r u e value at a P level (mgiL) of true concn, analyte mg/L 0 3.12 6.25 12.5 25 50 100 Ca 1 2 4 8
16 Sr 0.5 1.0
+ 6.0 + 2.0 --1.3 -
3.9 1.6
+ 8.8
4.0
+6.0 +4.0 1- 4.6
8.0
-
2.0
1.9
+9.0 -L 2 . 5 + 1.3 8.4 - 2.8 ~~
+5.2
5.0 + 3.0 - 3.4
i
A
1.0
+ 5.0
+4.3 -~ 7.7 - 5.9
~
~~
-
7
3.6
+
A
4.6
5.5 --4.O 0.0
0.8
-1.8
1.7
-~ 4.6
1.0 0.0 + 0.6 ~- 4.2
10.0
-
I-
5.0
~
+ 3.0
-t
+ 8.5 + 5.0 + 5.5
3.4 3.4
+ 1.4
5.0
-
+0.6
+ 1.6 1.6.0 -c
t
4.1
0.10 0.20 0.40
0.80
0.0
+ 3.8 ? 1.0 16.2 13.0
+ 3.0 ~
~
3.9
+ 4.6
Test Solutions. To evaluate the interferences of P on Ca, Ba, and Sr emission and the collective effects of Ca, Ba, and Sr on P emission, solutions containing P at levels of 0, 3.12,6.25, 12.5, 25, 50, 100, and 200 mg/L were prepared with: (1) Ca present a t 0, 1, 2, 4, 8, and 16 mg/L, (2) Ba present at 0, 0.05, 0.01, 0.2, 0.4, and 0.8 mg/L, and (3) Sr present at 0,0.5, 1,2,4,and 8 mg/L. Each of these solutions contained Cs at 1000 mg/L as the ionization buffer and In at 20 mg/L as the internal standard. In each instance except for the blank, the concentration ratios of Ca/Sr/Ba were maintained at 2O:lOl as an approximation of the concentration distribution of these elements in nature. These 48 solutions consequently covered the concentration range of zero to excess P relative to the alkaline earth elements and zero to excess alkaline earths relative to the P. To evaluate the effect of A1 on Ca and vice versa, two sets of solutions were prepared with Cs and In present at the above levels and in which: (1)the A1 was constant at 0.5 mg/L while the Ca was varied from 0 to 0.5. . . 500 mg/L in steps of 2X and (2) the Ca was held constant a t 2 mg/L while A1 was varied from 0 to 0.5 . . . 128 mg/L in steps of 2x. These combinations covered ranges in which each of the elements increased to concentration excess over the other. Procedures. To determine the effects of P on the alkaline earths, the instrument was calibrated with standard solutions containing only the alkaline earths, Cs and In. To determine the effects of the alkaline earths on P, the unit was calibrated with standard solutions containing only P, Cs, and In. The instrument was calibrated for A1 in the absence of Ca and Ca in the absence of AI to determine the respective effects of these two elements on each other. After calibration, each of the above test solution sets was analyzed via four replicate measurements. The concentrations obtained via these analyses were subsequently compared with the true concentrations to determine the extents of the respective interferences.
RESULTS AND DISCUSSION A summary of the results obtained for P effects on emission for t h e alkaline earths is presented in Table I. These data indicate that the majority of the measured concentrations agree with the true values within the experimental error of %5% and that systematic trends in the deviations d o not occur. Thus, the presence of phosphate in excess does not cause the typical emission suppression effect and does not detract from the analytical accuracy with which Ba, Ca, and Sr can be determined. T h e effects of the alkaline earth elements on phosphorus emission are summarized in Table 11. Again, the measured concentrations agreed with the true values within experimental error and systematic trends were absent. These results also verify that the classical alkaline earth-phosphate solute va-
-
~
- 1.0 7.2
t 6.5
4.0
+ 3.0
+ 10.5
+ 1.5
0.8
2.2 --7.2
2.7
-
3.7
-~
5.1
~
+7.3
+ 5.5 + 4.6 L
4.4 + 4.6
10.6
4.9
+ 5.6
0.0 0.0 + 2.0
5.5
--
1.5 -0.3 6.9 ~- 15.0
t 4.0
~~
3.8 11.5 3.0
6.7
+ 2.0
+ 2.5 + 2.0
-t 6.7
~-
-L
T
2.0
t
--
Ba 0.05
5.5
-~ 4.4
- 2.5
0.5 6.5
*
200
A
8.5
5.4 2.5
+ 4.0
t 1.8
4-4.1
.~
2.8
-
+ 0.3 --
3.1
2.0 7.7 +14.5 -
t
5.5
t4.6 L
-
2.0 4.6 6.9
Table 11. Effects of Alkaline Earths o n Phosphorus Analysis percent deviation from true value at CaiSriBa concentrations (mg/L) of P concn, 1:0.5: 16:8: mg/L 0.05 2:l:O.l 4:2:0.2 8:4:0.4 0.8 3.12 +3.8 +8.6 --5.3 -3.2 --2.4 -8.4 t4.3 -0.6 --4.9 6.25 -6.6 12.5 t6.0 -2.8 +3.6 -1.2 -2.8 25 15.7 +2.5 --1.9 +0.5 -4.0 0.5 ~ 0 . 4 13.8 50 +2.7 -6.0 ~
A i Concentration Img/ll
+2
4
- 14
i 8
tI 0
10
20
30
40
A'/Ca Yolar Conc*ntration
Figure 1. Effect of AI on
50 ' Ratio
Ca emission
porization interference effects are not significant for the present excitation medium. When A1 (0.5 mg/L) was determined in the presence of 0, 125,250, and 500 mg/L of Ca; the measured values deviated from the true value by 0, -1.0, -4.0, and -2.5% for the respective Ca concentrations. These measurements were made a t the A1 308.2-nm wavelength to avoid the calcium stray light effect previously reported (13)for the AI 396.1-nm wavelength. Again, the data indicate the lack of a n interference effect. In spite of the fact that excess Ca failed to interfere with AI analysis, the presence of excess A1 did cause suppression of Ca emission a t 393.3 nm. The results summarized in Figure 1 demonstrate that the suppression effect attains a constant value of -10% at Al/Ca molar concentration ratios in excess of approximately 20. In natural or effluent water samples and
ANALYTICAL CHEMISTRY, VOL. 51, NO. 14, DECEMBER 1979
in most biological materials, the Al/Ca concentration ratio is typically far less than unity. In view of the fact that the ratio must exceed approximately 2-3 (Figure 1) before the suppression effect becomes significant, the effect may be regarded as unimportant for the above types of samples. These evaluations generally indicate that solute vaporization interference effects are generally of minimal significance for the three-electrode dc plasma system used. This observation indicates that the three-electrode design change has been instrumental in eliminating interferences effects found to be significant for the two-electrode system ( I I ) .
2405
T. C. Rains, ref. 1, Chapter 12. D. J. David in "Flame Emission and Atomic Absorption spectrometry", J. A. Dean and T. C. Rains, Eds., Vol. 111, Marcel Dekker, New York, 1975, Chapter 2. E. M. Bulewicz and P. J. Padley, Spectrochim. Acta, Part B , 28, 125 (1973). Roland Herrmann and C. T. J. Alkemade, "Chemical Analysis by Flame Photometry", Interscience, New York, 1963. M. Marinkovic and B. Dimitrijevic, Spectrocheim. Acta, Part 6 ,23, 257 (1968). W. E. Rippetoe, E. R. Johnson, and T. J. Vickers, A m i . Chem., 47, 436 (1975). R . K. Skogerboe and I. T. Urasa, Appl. Spectrosc. 32, 527 (1978). Hugo L. Felkel and Harry L. Pardue, Anal. Chem., 50, 602 (1978). G. W . Johnson, H. E. Taylor, and R. K. Skogerboe, Spectrochim. Acta, Part B , 34, 197 (1979). (14) Joseph Reednick, A m . Lab., 11(3), 53-62 (1979).
LITERATURE CITED C. T. J. Alkamade in "Flame Emission and Atomic Absorption Spectrometry", J. A. Dean and T. C. Rains, Eds., Vol. I, Marcel Dekker, New York, 1969, Chapter 4. R. N. Kniseley, ref. 1. Chapter 6. A. N. Hambly and C. S. Rann, ref. 1, Chapter 8. Ivan Rubeska, ref. 1, Chapter 11.
RECEIVED for review June 1, 1959. Accepted August 24, 1979. This investigation was partially supported by the Water Resources Division of the U S . Geological Survey.
Loss of Polychlorinated Biphenyl Homologues during Chromium Trioxide Extraction of Fish Tissue Michael J. Szelewski, David R. Hill, Stuart J. Spiegel," and Edwin C. Tifft, Jr. O'Brien & Gere Engineers, Inc., 1304 Buckley Road, Syracuse, New York
Polychlorinated biphenyls (PCBs) are synthetic organic compounds produced by the chlorination of biphenyls. There are 209 possible chlorobiphenyls containing from one to ten atoms of chlorine ( I ) . These compounds are homologues, various mixtures of which are registered in the United States by the Monsanto Chemical Company under the trade name, Aroclor. T h e compounds are characterized by relative nonflammability, and useful heat exchange and dielectric (insulating) properties, the primary purpose for their development in 1929. They have been found to be toxic to a variet,y of organisms, although the greatest danger may be presented by the phenomenon of bioaccumulation, or bioconcentration, in the environment as the containment is traced through the food chain (1). Chemical characteristics of the individual compounds and Aroclors are dependent upon the degree of chlorination. Identification of the P C B mixture is by numerical nomenclature, for example, Aroclor 1221, Aroclor 1242, etc., with the number indicating the structure and composition of the compound. The first two digits represent the type of molecule - 12 = chlorinated biphenyl, 54 = chlorinated terphenyl. The last two digits give the average percentage, by weight, of chlorine. The exception to this nomenclature is Aroclor 1016 which contains 41% chlorine by weight, b u t in which the penta-, hexa-, and heptachlorobiphenyl content has been significantly reduced from Aroclor 1242. T h e analysis of environmental samples for polychlorinated biphenyls can be complicated by the presence of organochlorine pesticide residues, which are characterized by gas chromatographic (GC) retention times identical with many PCB homologues. One of the established methodologies for the identification and quantitation of PCBs in animal tissue in the presence of organochlorine pesticides, such as DDE, includes a preparative step involving chromic acid digestion (1-3). The purpose of this step is to oxidize these sources of interference. 0003-2700/79/0351-2405$01 OO/O
13221
In this work, the chromic acid treatment was being employed in the analysis of freshwater and marine fish for PCBs when an alteration of Aroclor chromatographic patterns was observed. These alterations included changes in expected peak areas and the disappearance of several P C B homologues. A confirmatory investigation was performed under controlled conditions to discover the extent of the effect.
EXPERIMENTAL A Tracor model 550 gas chromatograph equipped with a model 700 Hall electrolytic conductivity detector was used in this study (Tracor, Austin, Tex.). The GC was interfaced to a HewlettPackard (Palo Alto, Calif.) model 33808 integrator. The chromatographic column was 183 cm long, 6.4-mm 0.d. by 4-mm i.d., glass, and packed with 3% OV-1 on 80-100 mesh Chromosorb W-HP (Applied Science Laboratories, State College, Pa.). The column was conditioned at 275 "C, 50 mL/min N2 until satisfactory resolution and response were obtained. Operating temperatures for the GC were: furnace, 850 "C; column, 190 "C; inlet, 225 "C; outlet, 250 "C, and auxiliary, 260 "C. The carrier flow was 50 mL/min N2 at 3.5 kg/cm2; the reaction flow, 50 mL/min H2 at 0.7 kg/cm2. All solvents were spectrograde (Mallinckrodt, St. Louis, Mo.) without further purification, while other reagents were reagent grade. The chromic acid solution was prepared by dissolving 135 g of chromium trioxide in 90 mL of water and adding 750 mL of glacial acetic acid ( 3 ) . Twenty-five milliliters of Cr203solution was added to each of fifteen 5-mL hexane solutions containing prepared Aroclor standards, as follows: eight replicates of Aroclors 1016 and 1254, respectively, and six replicates of both Aroclor 1221 arid a mixture containing all three Aroclors in equivalent concentrations. The solutions were heated at 90-100 "C for 45 min with constant stirring and the addition of Cr203 solutions as necessary to maintain an excess quantity of oxidant in the solution. After heating, each solution was added to 100 mL of water, respectively, and extracted four times with petroleum ether, with removal of the aqueous phase after each extraction. The organic 1979 American Chemical Society
