322
ANALYTICAL CHEMISTRY, VOL. 50, NO. 2, FEBRUARY 1978
and provide the means to their identification. Both HPLC and 13C-NMR have proved to be valuable quantitative techniques in dealing with aqueous solutions of amygdalin. These techniques avoid the derivatization procedures that are required for the GC/FID and GC-MS methods and provide a distinct advantage for the analysis of injectable preparations of amygdalin. Although thermal epimerization has not been observed when derivatization procedures are used, analysis of aqueous solutions by HPLC and 13C-NMRavoid the risk of induced epimerization. The major utility of the GC/FID and GC-MS techniques for the quantitative determination of amygdalin and its epimers is for those situations where HPLC is not available to the analyst. The correlation between the various techniques firmly establishes each as an independent analytical method for the qualitative and quantitative determination of amygdalin.
ACKNOWLEDGMENT The authors are grateful t o Milton A. Luke and Gregory M. Doose for their helpful suggestions and laboratory assistance. The authors also express their thanks to Roselyn
Erneta of the
U.S.Customs for the
amygdalin samples.
LITERATURE CITED (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15)
E. E. Conn, J . Agric. FoodChem., 17, 519 (1969). "The Merck Index", 9th ed., (1976), p 81. A. Viehoever and H. Mack, Am. J . Pharm., 107, 397 (1935). W. N. Haworth and Wylam, J . Chem. SOC.,123, 3120 (1923). C. S.Hudson, J . Am. Chem. SOC., 46, 483 (1924). E. T. Krebs and E. T. Krebs, Jr., U S . Patent Office, No. 2,985,664 (1961). E. T. Krebs and E. T. Krebs, Jr., Br. patent, 788,855 (1958). W. Benson, FDA Bureau of Drugs, 200 C Street, S.W., Washington, D.C. 20204, private communication. C. Fensebu, Departmentof Pharmacology and ExperimentalTherapeutics, The Johns Hopkins University, Baltimore. Md. 21205, private communications. M. A. Luke, J . Assoc. Off. Anal. Chem., 54, 937 (1971). A. Nahrstedt, J . Chromatogr., 50, 518 (1970). W. 0.McReynolds, J . Chromatogr. Sci., 8, 685 (1970). D. S. Seigler, Phytochemistry, 14, 9 (1975). J. W. Turczan. FDA New York District Laboratory, 850 Third Avenue, Brooklyn, N.Y. 11232, private communications. J. I . Kroschwitz, M. Winokur, H. J. Reich, and J. D. Roberts., J . Am. Chem. SOC.,91, 5927 (1969).
RECEIVED for review July 27, 1977. Accepted November 11, 1977.
Ionic Interference in the Determination of Arsenic in Water by the Silver Diethyldithiocarbamate Method Shingara S. Sandhu" Claflin College, Orangeburg, South Carolina 29 115
Peter Nelson South Carolina State College, Orangeburg, South Carolina 29 1 17
The extent of lonlc Interference In the use of the standard sllver diethyldithiocarbamate (SDDC) method for arsenic determinatlon In water and waste water Is quantltatlvely evaluated. The absorbance peak at 410 nm Is asslgned to the formation of a hydrogen-SDDC rather than a chromium-SDDC complex. The SDDC method provides reliable data for arsenic, even in fairly polluted water and Its recovery up to 0.5 pg (0.01 mg L-') from such waters Is quantitatlve. Antlmony and mercury Interfere wlth arsenlc color development by yleldlng complexes, wlth maxlmum absorbance at 510 and 425 nm, respectively. Recovery of arsenic, released on digesting standard solutlons of natural water as well as cacodyllc acld, (sodium salt) is also quantltatlve.
Arsenic is widely distributed in the human environment and any sample of water, if analyzed by a suitably sensitive method, will be found to contain a t least a small quantity of arsenic ( I ) . The U.S. Public Health Service recommends that arsenic concentration in drinking water should not exceed 0.01 mg L-' and that water with an arsenic concentration greater than 0.05 mg L-' should be rejected for human consumption ( 2 ) . The arsenic concentration of potable water is generally less than 0.005 mg L-' although a concentration as high as 0.1 mg L-' has been reported (3). Arsenic is a suspected carcinogen (3);consequently, there is growing interest in arsenic 0003-2700/78/0350-0322$0 1.OO/O
contamination of the environment. Several methods (4-13) are available for the determination of arsenic but silver diethyldithiocarbamate (SDDC) ( I O ) is the only recommended method for the separation and quantitative determination of arsenic in water samples. This method consists of reducing inorganic arsenic in a water sample by acid zinc reaction to arsine (ASH,) which is scrubbed through lead acetate impregnated glass wool and is absorbed in silver diethyldithiocarbamate dissolved in pyridine. The color developed due to the arsine (ASH,) silver diethyldithiocarbamate reaction is photometrically measured a t 535 nm. Conflicting reports (12,14)have been published on the role of chromium as an interfering ion in the standard method. Though chromium supresses arsine generation ( I O ) , a color enhancement ascribed to chromium interference in the determination of arsenic by the SDDC method was observed. The authors (14) speculated that the chromium-SDDC complex was responsible for the absorbance peak a t 410 nm. Certain other metals-cobalt, copper, mercury, molybdenum, nickel, platinum, and silver-have also been reported to interfere in the generation of arsine (IO) but their limits of interference have not been thoroughly investigated. I t has been indicated that the presence of antimony in a water sample interferes in the development and photometric measurement of the SDDC-arsenic complex color (IO),but a recent report indicates that antimony concentrations up to 0.2 mg L-' do not show any significant interference. This
D 1978 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 50, NO. 2, FEBRUARY 1978
323
I
400
500
603
-'
400
700
Absorbance spectrum of a reagent blank recorded vs. a SDDC solution a s reference
Figure 1.
603
50C WIVELEMGTH.
WAVELENGTH, nm
70C
nm
Absorbance spectrum of 2 pg of arsenic alone and in combination with 1 mg L-' each of Cr(VI), Co(II),Cu(II), Mo(VI), and Ni(I1) recorded vs. a SDDC solution as reference Figure 2.
article presents data to resolve the confusion that surrounds the use of the SDDC method for the determination of arsenic in water and waste water.
'Oi
EXPERIMENTAL Apparatus. The generator and absorber assembly have been previously described (IO) and were purchased from Fisher Scientific (Catalogue No. 1-405). A Beckman Model 24/25, double-beam scanning spectrophotometer with 1-cm cells, equipped with a digital read out system and a strip chart recorder, was used for spectrum study and absorbance measurements. Materials. The stock solutions for arsenic (1111, arsenic(V), chromium, cobalt, copper, mercury, molybdenum, nickel, phosphate, and antimony containing 1 g L-' of ionic concentration were prepared from arsenic trioxide (As203),sodium arsenate (Na2HAs04.7H20), potassium chromate (K2Cr04),cobalt chloride (CoCl2.6H2O),copper nitrate ( C U ( N O ~ ) ~ . ~ H mercury(I1) ~O), chloride (HgC12),ammonium molybdate (NH4)6M07024-4Hz0), nickel nitrate (Ni(N03)J, potassium dihydrogen phosphate (KH2P0,) and antimony trichloride (SbC13)respectively. Intermediate solutions were prepared by diluting the stock solutions l:lO, and working solutions containing requisite concentrations of various ions were obtained by diluting the intermediate solutions. Analytical grade reagents were used. Procedure. The procedure adopted for the present study is similar to the one given for the standard method (10)except that a 50-mL water sample is used which requires the use of 7.5 mL of concentrated hydrochloric acid, 2.0 mL of potassium iodide and 0.5 mL of stannous chloride in hydrochloric acid. The reaction is allowed to proceed for 15 min at room temperature, following the addition of 3.0 g of zinc,after which the generator is transferred to a water bath at about 50 "C for another 15 min. The solutions, from the absorber tubes, are poured directly into a 1-cm cell and scanned for a complete absorbance spectrum ('iOC350 nm), using a SDDC solution and a reagent blank (SDDC solution treated in the absorber tube similar to the experimental procedure but without arsenic) as references. Though the concentration of arsenic(II1) as well as arsenic(V) in the prepared standards and river water is maintained at 2.0 K g or less per 50 mL of sample (0.04 mg L-' or less), the amount of interfering ions is varied up to 350 wg per 50 mL (7.0 mg L-l). An absorbance calibration curve using 0.0, 1.0, 2.0, 4.0, and 5.0 b g of arsenic was prepared.
RESULTS AND DISCUSSION Complete absorbance spectra for a reagent blank, arsenic(III), and arsenic(II1) in combination with chromium(VI), cobalt(II), copper(II), molybdenum(VI), and nickel(II), each a t 1 mg L-l level (total interfering ionic concentration 5.0 mg L-'), vs. an SDDC solution, as reference, are given in Figures
O2t
WAYELEYGTH
"m
Absorbance spectrum of 2 pg of arsenic alone and in combination with 1 mg L-' each of Cr(VI), Co(II),Cu(II),Mo(VI),and Ni(I1) recorded vs. a reagent blank as reference Figure 3.
1 and 2, respectively. Each spectrum shows a n absorbance peak around 410 nm which disappears, or is considerably reduced, (Figure 3) when the same absorbance is recorded vs. a reagent blank. An absorbance peak a t 410 nm similar to Figures 1 or 2 is seen in all observations, no matter which one of the numerous interfering ions is used in the generator, either alone, or in combination with arsenic, if the absorbance of the SDDC complex developed in the absorber tube is read vs. a SDDC solution rather than vs. a reagent blank. Although no mechanism for the transport of chromium from the arsine generator to the absorber, which contains SDDC solution, was suggested, the absorbance peak a t 410 nm was speculated to be due to the formation of a chromium diethyldithiocarbamate complex ( 1 4 ) . The information presented here (Figures 1-3) does not substantiate the previous report ( 1 4 ) , but rather demonstrates that none of the interfering ions is responsible for the absorbance peak a t 410 nm. The absorbance peak at 410 nm is observed even when the generator has nothing in it except the deionized water, hydrochloric acid, and zinc. The addition of stannous chloride alone or in combination with
324
ANALYTICAL CHEMISTRY, VOL. 50, NO. 2, FEBRUARY 1978
Table I. Recovery of Various Forms of Arsenic from Demineralized and River Waters, Concentration mg L-'
The ore tical
Demineralized water Recovery, Determined' 9 i
Std dev
River water Recovery, Determinedb
%
Std dev
As(II1) 0.01
0.0102
102
0.02
0.0196 0.0410
98
0.04
102
5.7 4.3 3.4
...
...
...
0.0214 0.043
107 108
5.6 4.5
5.8 4.9
... ...
... ...
...
... ...
... ...
... ...
109
5.8
MV) 0.01 0.02
0.04
0.0100 0.0208 0.0395
100 104 99
4.0
...
...
... ...
As(II1) + h ( V )(50:50) 0.01 0.02
0.04
0.0095
95
0.0201 0.0405
100 101
' Mean of 35 determinations.
5.6 3.1 3.5
0.044
Mean of 20 determinations.
potassium iodide does not change the nature of this absorbance peak in any way. Accordingly, it is suggested that the absorbance at 410 nm is due to the reaction of hydrogen (generated by the acid zinc interaction) with SDDC reagent. It is further observed that the presence of arsenic in the generator, alone or in combination with the other ions, remarkably increases the peak height a t 410 nm over reagent blank or interfering ions alone (Figures 1 and 2). I t is speculated that during the reaction of arsine (ASH,) with SDDC, a certain amount of additional atomic hydrogen becomes available for simultaneous combination with SDDC, resulting in the formation of the additional hydrogen SDDC complex responsible for increased absorbance a t 410 nm. SDDC + ASH, -+ SDDC-Arsenic complex SDDC-Hydrogen complex
+
The peak a t 410 nm is not precisely reproducible (i3070), and is seen to be preferentially destroyed over the arsenicSDDC complex peak of 535 nm. The arsenic-SDDC complex is developed as described in the procedure and the absorbance spectrum is recorded. The solution is poured back into the absorber tube and concentrated hydrochloric acid is added, a drop a t a time, to the solution in the absorber tube and the absorbance spectrum is taken again. This is repeated until the peak a t 410 nm disappears. Molarity of hydrochloric acid, calculated from the volume of acid added, when plotted vs. the change in absorbance a t 410 nm, shows a negative linear relation, whereas the absorbance at 535 nm, except for a minor dilution effect, remains the same. The instability of the complex, responsible for absorbance at 410 nm, is an additional indication that a hydrogen-SDDC complex rather than a metal-SDDC complex is probably responsible for an absorbance peak a t 410 nm. The single ion absorbance spectra for Cr(VI), Co(II), Cu(II), Mo(VIj, and Ni(I1) are not different from the reagent blank spectrum and do not show any noticeable peak within 700 to 350 nm, if the readings are taken vs. a reagent blank. Therefore, it appears that these ions should not show any positive interference, as suggested previously (14), in the evaluation of arsenic by the standard method. However, if the absorbance spectrum of the complex formed in the absorber tube is recorded, using a SDDC solution instead of a reagent blank as reference, a large absorbance peak at about 410 nm is noticed, which causes the elevation of the baseline (Figures 1 and 2), leading to an inaccurately high arsenic estimate (about 10% j for a given water sample.
The water samples containing 0.30 mg L-' of antimony and 2.0 mg L-' of mercury show absorbance peaks a t 510 nm (average absorbance 0.007) and 425 nm (average absorbance 0.005) respectively, thus these ions are expected to interfere positively in the determination of arsenic by the standard method. Possible positive interference by antimony has been suggested in the past (10)but similar interferences by mercury have not been reported. Ionic mercury reacts with stannous chloride in hydrochloric acid to produce metal mercury (15). 2Hg2++ SnC1, + 2C1-- HgZ2+ + SnCl, Hg," + SnC1, + 2C1-- 2Hg0 + SnCl, There is a distinct possibility that mercury vapors are carried over from the generator to the absorber tube containing SDDC reagent, resulting in the formation of a mercury-SDDC complex, responsible for an absorbance peak a t 425 nm. The amount of arsenic recovered from demineralized water is quantitative up to an arsenic concentration of 0.5 pg or 0.01 mg L-' (Table I). Arsenic recovery from demineralized water in the presence of various interfering single ions is presented in Table 11. This information suggests that the recovery of arsenic is not affected in the presence of chromium(VI), cobalt(II), copper(II), molybdenum(VI),nickel(IIj, nitrate, and phosphate, up to an individual ion concentration of 5.0 mg L-l. But when the single metal ion concentration is increased beyond 5.0 mg L-', a decrease in arsenic recovery from standard solutions is observed and a t the 7.0 mg L-' level, the arsenic recovery decreases by about 10%. Nitrate and phosphate concentrations up to 100 mg L-' do not show any observable effect on the arsenic recovery. It has been reported (12) that antimony concentrations up to 0.2 mg L-' did not produce any observable interference in determining arsenic by the SDDC method. In the present study, antimony concentration of 0.2 mg L-' does not show any significant change in the recovery of arsenic from standard solutions, however, when the antimony concentration is raised to 0.3 mg L-I, the apparent rate of recovery for arsenic increased by about 10%. The location of the arsenic absorbance peak shifts toward 510 nm in the presence of antimony. This shift seems proportional to the amount of antimony present in a solution containing 0.04 mg L-' of arsenic. The mercury concentrations at 1.5 mg L-' and above show a significant positive interference in the recovery of arsenic. Interference by several ions in combination was studied by preparing standard solutions, containing 0.04 mg L-' of arsenic and varying the concentrations of each ion, chromium(VI), cobalt(II),copper(II),molybdenum(VI), nickel(II), phosphate,
ANALYTICAL CHEMISTRY, VOL. 50, NO. 2, FEBRUARY 1 9 7 8
Table 11. Recovery of Arsenic from Demineralized and River Water in the Presence of Interfering Ions, Concentration mg L-l Demineralized Water River Water Std Ions" added dev Recovery, %' Theoretical Recovery, % b individually 0.02 0.04 0.04
99.6 98.9 89.5
5.9 5.7 4.3
0.2
0.04 0.04
99.6 109.3
4.7
0.3
0.04 0.04
107.5
4.5
134.2
4.4
CollectivelyC 5.6, 0.8 each 7.0; 1.0 each Cr(V1) + Co(I1) 3.5 each
0.04 0.04
96.2 89.4
4.7 4.0
0.04
88.2
3.5
3.5 each
0.04
90.2
4.2
5 5 7
325
Std de v
107 108 97
6.6 6.5 8.3
103.3 97.1
7.2 9.3
Sb
Hg 1.5 5.0
M O ( V I ) ~+ CU(II)
4.9
Mean of 12 determinations for each system. Ions used: Co(II), Cr(VI), Cu(II), Mo(VI), NO,-, Ni(I1) and PO,-3. Mean of 6 determinations for each svstem. Results for other combinations. not reDorted here. a
and nitrate, up to 1.0 mg L-'. The results on the recovery of arsenic in the presence of a combination of ions are also given in Table 11. There is no significant interference by these ions up to a combined concentration of 5.6 mg L-' (0.8 mg L-' each ion) but when the combined concentration is increased above this level, there is an observable decrease in the recovery of arsenic from demineralized water. A combined concentration of 7.0 mg L-' (1.0 mg L-' each ion) decreases the arsenic recovery from standard solutions by about 10%. A combined concentration of two ions, chromium(V1) and cobalt(II), each at 3.5 mg L-', decreases the arsenic recovery by about 10%. A molybdenum(V1) and copper(I1) combination, as well as the combination of any other two metal ions in this category a t a total concentration of 7.0 mg L-', shows similar results (Table 11). It appears that under the present set of experimental conditions, the ionic specificity of the elements that interfere in arsine generation, is not as important as their total concentration in inhibiting the recovery of arsenic from the standard solutions. The studies were also carried out using natural water from the Edisto River, Orangeburg, S.C., which was spiked with arsenic as well as with interfering ions. Because only inorganic arsenic is reduced to arsine (IO),a digestion step, as suggested in the literature, using nitric and sulfuric acids (12) was tried for the release of organically bound arsenic. Data on the recovery of arsenic using this digestion method are acceptable for standard solutions but the recovery of arsenic, added to the river water, is inconsistent and reproducibility is very poor. Consequently, the potassium permanganate method (16) with minor modification is adopted for the digestion of water samples. One-tenth milliliter of 5.0% potassium permanganate, followed by 5.0 mL of concentrated sulfuric and 0.3 mL of nitric acids, is added to 50 mL of water sample, containing a t least 1Fg of arsenic. The water sample is placed in a water bath a t 35 "C for an hour and the excess of potassium permanganate is destroyed by adding 1.5% hydroxylamine hydrochloride. Hydroxylamine hydrochloride is added a drop at a time to avoid its excess in the generator. The reliability of the potassium permanganate method was
further investigated using known concentrations (2 fig and 1 pg of arsenic per 50 mL) of dimethylarsenic acid (cacodylic acid, sodium salt). This method not only gives an acceptable recovery and reproducibility for arsenic in natural water as well as in organoarsenicals (88.8% arsenic recovery for cacodylic acid), but is less laborious than the acid digestion method previously used (12). The recovery of arsenic from the river water is greater than the amount added (Tables I and 11). The difference approximately equals the amount of arsenic found in the Edisto River.
LITERATURE CITED H. K. 0. Lee, "Metallic Contamination and Human HeaW, Academic Press, New York, N.Y., 1972. United States Department of Heath, Education and Welfare, Public Heatth Service, "Drinking Water Standards", Public Health Service Publication No. 956, U.S. Government Printing Office Washington, D.C., 1962. Subcommittee on Air and Water Pollution of the Committee on Public Works, United States Senate, "Water Pollution" Part 4. United States Senate, Ninety-first Congress, Second Session, U S Government Printing Office, Washington, D.C., 1970. J. C. Caklwell, R. Lishka, and E. McFarren, J . Am. Water Works Assoc., 6 5 . 73 (1973). A. L. Chaney'and J. M. Harold, Ind. f n g . Chem., Anal. Ed., 12, 691 (1940). I. M. Kolthof, and A. Elias, Ind. Eng. Chem., Anal. Ed., 12, 177 (1940). D. Liederman, F. E. Brown, and 0. I . Milmer, Anal. Chem., 30, 1543 (1958). G. W. Powers, Anal. Chem., 31, 1590 (1959). D. C. Baliinger, R. J. Lishka, and M. E. Gales, J. Am. Water Works A m . , 5 4 , 1424 (1962). American Public Heath Association, American Water Works Association and Water Pollution Control Federation, "Standard Methods for the Examination of Water and Waste Water", 13th ed..American Public HeaW Association, Washington, D.C., 1971. K. C. Tam, Environ. Sci. Techno/.,8 , 73 (1974). J. F. Kopp, Anal. Chem., 45, 1786 (1973). S. S. Sandhu, Ana/yst(London), 101, 856 (1976). G. C. Whitnack and H. H. Martens, Science, 171, 383 (1971). A. B. Ganet. H. H. Slsler, J. Bonk, and R. C. Stouter, "Semimicro Qualitbe Analysis", 3rd ed., Blaisdell Publishing Company, Waltham. Mass., 1957. Perkin-Elmer Corporation, "Analytical Method for Atomic Absorption Spectrophotometer,Mercury Anatysis System:" 303-31 19, Norwalk, Conn.. 1972.
RECEIVED for review July 7, 1977. Accepted November 21, 1977. This work is supported by EPA Grant R8041640-10.
