example, in the case of Hg, one could employ the Hatch and Ott method (20), generating the sample atoms at room temperature. For other elements, one might employ a
(20) W. R. Hatch and W. L. Ott, ANAL.CHEM., 40,2085 (1968).
region downstream from the atomization cell. Both of these possibilities will be investigated in the near future.
Work supported in part by the Robert A. Welch Foundation and in part by the National Science Foundation.
Arsenic Determination at Sub-Microgram Levels by Arsine Evolution and Flameless Atomic Absorption Spectrophotometric Technique Richard C. Chu, George P. Barron, and Paul A. W. Baumgarner Pesticide Research Laboratory and Graduate Study Center, The Pennsylvania State Unicersity, University Park, Pa. 16802
THEDETERMINATION of arsenic by Atomic Absorption Spectrophotometry (AAS) using the conventional air/acetylene flame presents some difficulties because of strong flame absorption at the far ultraviolet region where the most sensitive resonance lines of arsenic lie. The introduction of an argon/hydrogenentrained air flame considerably reduced the flame absorption in this region of the spectrum ( I ) . However, because of the low temperature of the argon/hydrogen flame compared to the air/acetylene flame, interferences due to molecular absorption and incomplete salt dissociation were inevitable. The use of a nitrogen/hydrogen-entrained air flame and a nitrogenseparated airlacetylene flame have been reported (2, 3) with the advantage of minimizing the interferences. Recently, the chemical conversion of arsenic to arsine and its subsequent introduction into an argonlhydrogen flame has been developed (4-6). This technique eliminated the interference resulting from the matrix effect and improved the detection limits. This paper describes a flameless AAS method for arsenic determination which also involves the chemical conversion of arsenic to arsine. The arsine evolved is swept into an electrically-heated absorption tube by means of an argon carrier gas. Since no flame is employed in this technique, the background absorption is less than that of the argonlhydrogen-entrained air flame method, and a considerable increase of sensitivity is obtained. EXPERIMENTAL
Apparatus. All measurements were made with a Perkin-
Elmer Model 403 Atomic Absorption Spectrophotometer equipped with an arsenic hollow cathode lamp and a digital readout and printer. The apparatus used for the generation (1) H. L. Kahn and J. E. Schallis, A t . Absorption Newslett., 7 , 5 (1968). ( 2 ) A. Ando, M. Suzuki, K. Fuwa, and B. L. Vallee, ANAL.CHEM., 41,1974 (1969). (3) G. F. Kirbright, M. Sargent, and T. S . West, At. Absorption Newslett., 8,34 (1969). (4) W. Holak, ANAL.CHEM., 41,1712 (1969). ( 5 ) E. F. Dalton and A. J. Malonoski, Ar. Absorption Newslett., 10,92 (1971). (6) F. J. Fernandez and D. C. Manning, ibid.,p 36. 1476
ANALYTICAL CHEMISTRY, VOL. 44, NO. 8,JULY 1972
of arsine and the electrically-heated absorption tube is shown in Figure 1. The absorption tube (part No. 7, Figure 1) was constructed from a 2.5-cm i.d. X 15-cm piece of Vycor glass tubing with both ends left open and with a gas inlet inserted in the tubing at approximately the mid-point or 7.5 cm from each end. The absorption tube was placed above the burner head of the instrument and aligned to obtain minimum background absorption. Heat was applied to the absorption tube through asbestos-covered wire (part No. 8, Figure 1) that was coiled around the middle section and connected to a variable transformer (The Superior Electric Co., Bristol, Conn.). The temperature in the tube was measured with a thermocouple (Therm0 Electric Mfg. Co., Dubuque, Iowa). Instrument settings used in the atomic absorption were hollpw-cathode arsenic lamp Furrent, 18 mA; wavelength, 1937 A ; spectral band width, 7 A; and argon flow, 6 SCFH. Reagents. Analytical grade reagents were used. The arsenic standard was a stock solution of 1.O mg of arsenic per ml prepared by dissolving 0.1465 gram of NasHAsO,, 7 H 2 0 in 5 % H2S04-20% HC1 solution in a 100-ml volumetric flask. Also used were: stannous chloride solution (40% SnC12. 2 H 2 0 in concentration HCl) ; potassium iodide solution (15% KI in glass-distilled water); and zinc granules (20 mesh). The diluent for stock solution was 5% H2s04-20% HC1 solution, prepared by adding 50 ml of concentrated HzS04 and 200 ml of concentrated HC1 to 60 ml of glass-distilled water in a liter volumetric flask. Dilute arsenic standards were prepared immediately before use by dilution of the stock solution. All glassware was washed with concentrated HzS04 and rinsed several times with glass-distilled water. Procedures. An amount of arsenic ranging from 0.05 to 1.0 pg was added to a 250-ml beaker containing 25 ml of 5 % Hzs04-20% HCI solution. To this solution, 2 ml of 15 % KI and 1 ml of 20 % SnClz.2 H 2 0were added and mixed well. The solution was heated to 85 "C for 5 minutes and cooled to room temperature to ensure complete reduction of arsenic from the pentavalent to the trivalent state. After the solution was transferred to the reaction flask (part No. 1, Figure l), the flask was connected to the apparatus with the argon flow by-passing it, One gram of 20-mesh zinc granules was added through the side neck of the flask and an adapter with a balloon tied at one end was immediately inserted into the
/
8
Figure 1. Diagram of arsine generation apparatus and the electrically-heated absorption tube The parts are (1) generation flask, round bottom 100 ml; neck (29/42 joint); (2) 3-way stopcocks; (3) side neck (14/20 joint); (4) drying tube as adapter; (5) collection balloon; (6) flowmeter; (7) absorption tube; (8) chrome1 A asbestos-covered wire, size 26; (9) Vycor glass tubing, 8 mm 0.d. and 4.5 mm i.d.; (10) tygon tubing
0.500
0.090
t
0.080
-
-
0.060 0.070
0 W
z : 0.050 a
2 m
a
0.040
-
0.020 -
0.030
0.010
200
300
400
500
600
700
800
TEMPERATURE C '
500
600
700
800
TEMPERATURE *C
Figure 2. Effect of temperature on absorbance of 0.4 pg arsenic side neck (part No. 3, 4, and 5, Figure 1). The reaction was allowed to continue for 10 minutes to ensure complete evolution and collection of arsine. The 3-way stopcocks (part No. 2, Figure 1) were then turned, allowing the argon to flow through the flask and carry the arsine collected in the balloon into the electrically-heated absorption tube. The absorbance was read from the digital readout and recorded on the printer. After the absorbance reached a maximum, the stopcocks were turned to the bypass position and the reaction flask was removed. For samples containing organically bound arsenic, prior oxidation either by dry ashing or wet 'digestion was necessary. Into a 250-ml beaker, 1 ml of concentrated H&O4 and 5 ml of concentrated HNO, were added. The digestion was facilitated by heating at 120-140 "C on a hot-plate. Approximately 5 ml of HN03 were added whenever the mixture became brown or darkened. Heating and adding HNOI continued until the solution no longer darkened. After cooling slightly, 1 ml of 70z perchloric acid was added and the solution was heated until the digest was clear. The digest was allowed to cool at room temperature, then 2 ml of a saturated solution of ammonium oxalate were added and heat was again applied until fumes of SO3 were evolved.
Figure 3. Effect of temperature on background absorbance due to air When the digest cooled, 5 ml of concentrated HC1 acid were added and the volume was increased to 25 ml with glass-distilled water. Analysis was conducted on the acid solution containing KI and SnC12,with the zinc granules added, to obtain values for a reagent blank. All analyses were then blank-corrected. RESULTS AND DISCUSSION
Figure 2 illustrates the effect of the temperature in the absorption tube on the absorbance of arsenic. The increased absorbance observed with the increasing temperature represented the rate at which arsenic was released from arsine. The graph suggests that the greatest sensitivity occurred at higher temperatures. In the present study, the temperature in the absorption tube was maintained at 700 "C for all analyses by a setting of 90 on the control scale of the variable transformer. The resglts of extensive pretesting indicated that variac setting 6s. temperature was linear, and temperature stability and reproducibility were excellent. Figure 3 shows the effect of temperature on the background absorption due to air. As the temperature was increased, the background absorption was elevated geometrically. However, this background could be lowered by ANALYTICAL CHEMISTRY, VOL. 44, NO; 8, JULY 1972
1477
A R G O N (SCFH)
Figure 4. Effect of argon flow on background absorbance
Table I. Reproducibility of Five Replicate Samples of 0.4 pg Arsenic
Sample 1
2 3 4 5
Absorbances 0.319 0.321 0.318 0.322 0.326
Coefficient of variation,
Time, minutes 5
0.36
flowing argon gas through the absorption tube. By varying the argon flow rate from 1 to 6 SCFH, the lowest background was obtained when the argon flow rate was at 6 SCFH, as illustrated in Figure 4. Figure 5 shows the two calibration curves ( A and B ) of arsenic from 0.05 to 1.0 pg. Curve A , obtained by the flameless AAS technique, was linear up to 0.4 pg while calibration curve B, achieved by sweeping arsine into the argon/hydrogen-entrained air flame, was linear up to 1.0 pg. Fernandez and Manning (6) used the arsine evolution and the argonhydrogen flame method and reported that the calibration curve was linear up to 0.6 pg. The present results indicated, however, that the flameless AAS method has more than twice the sensitivity of the argonlhydrogenentrained air flame method. The reproducibility of the present technique, shown in Table I, was determined by measuring the absorbance of 5 replicate samples of 0.4 pg arsenic standards. The coefficient of variation was calculated to be 0.36%. A reproducible and stable operating temperature (700 "C) in the absorption tube and a constant argon flow rate (6 SCFH) contributed significantly to the attainment of good reproducibility. After the desired operating temperature and gas flow rate were once 1478
Table 11. Reaction Time Required for Complete Evolution of 0.4 pg of Arsenic to Arsine
ANALYTICAL CHEMISTRY, VOL. 44, NO. 8, JULY 1972
6 8 10
12
Absorbances 0.227 0.304 0.318 0.321
0.319
established in the system, the introduction of the sample and the subsequent flow of arsine into the absorption tube did not appreciably affect the tube temperature. By using the argon/hydrogen flame method, Fernandez and Manning (6) reported that the coefficient of variation of a 1.0-1.18sample of arsenic was 4.8 %. The time required for the complete recovery of arsenic was measured, and a reaction time of 8 minutes was required as indicated in Table 11. A period of 7.5 minutes for the recovery of arsine has been observed by Madsen (7). However, Fernandez and Manning (6) found that a reaction time of 4-5 minutes was sufficient for the liberation of all arsenic. In the present study, a reaction time of 10 minutes was used in order to compensate for the different rate of hydrogen production due to small variations in the sizes of the zinc granules. Two levels of arsenic, 0.1 and 0.2 pg, were added to vitaminfree casein to determine the per cent of recovery when wet digestion and the present method of analysis were used. Recoveries of 94.7 and 92.6% were obtained from the 0.1-pg and 0.2-pg levels, respectively.
(7) R. E. Madsen, Jr., At. Absorption Newsletr., 10, 57 (1971).
0.2
0
0.4
0.6
As
Figure 5.
I. 0
0.8
(pg)
Calibration curves of arsenic
A . By flameless AAS B. By argonhydrogen-entrained air AAS
Several precautions should be noted when applying the arsine evolution and the electrically-heated absorption tube methods. Since the absorbance of arsenic varies with the temperature, as previously mentioned, a constant temperature should be maintained in the absorption tube. Second, during wet digestion, the arsenic must be maintained in As5+state to avoid losses and this can be accomplished by constant addition of nitric acid to maintain oxidizing conditions (8). However, the nitric acid must be removed in order to maintain arsenic in a trivalent state before the chemical conversion of arsenic to arsine. This can be achieved by addition of a concentrated ammonium oxalate solution and heating the digest to expel the oxides of nitrogen. Third, since arsine is immediately evolved after the addition of zinc granules (5-9, the adapter with the balloon should be inserted immediately to prevent the escape of arsine and the reduction of absorbance. Finally, the balloon should be flushed with argon before each analysis to ensure that no residues of arsine are left in the balloon. Fernandez and Manning (6) reported that a balloon would lose elasticity after 5-10 determinations because of acid fumes. With a curved adapter as used in the present study, a balloon could be used for 15-20 determinations before it needed to be replaced. In conclusion, the flameless AAS method for arsenic determination not only significantly lowered the background __ (8) E. B. Sandell, “Colorimetric Determination of Traces of Metals,” 3rd ed., Interscience, New York, N. Y., 1959, p 278.
absorption compared to that of the argonlhydrogen flame method, but also increased the sensitivity and gave an excellent reproducibility and recovery. RECEIVED for review December 13, 1971. Accepted March 17, 1972. Authorized for publication on Dec. 3, 1971, as paper no. 4101 in the journal series of the Agricultural Experiment Station.
Correction Intrinsic End-Point Errors in Titration with Ion Selective Electrodes. Chelometric Titrations In this paper by Peter W. Carr [ANAL. CHEM.,44, 452 (1972)], an error appears in Equation 12, p 453. This should read -2ba3 -
cy2
- 4bpa
- 4bzp2
=
0
The data presented in the paper are correct and were computed on the basis of the above equation.
ANALYTICAL CHEMISTRY, VOL. 44, NO. 8, JULY 1972
1479