ANALYST, JUNE 1991, VOL. 116 595 Automated Determination of Sulphide by Gas-phase Molecular Absorption Spectrometry Toyin A. Arowolo and Malcolm S. Cresser* Department of Plant and Soil Science, University of Aberdeen, Meston Building, Aberdeen AB9 2UE, UK An automated method for the determination of sulphide in solution that involves the interfacing of an automatic sampler, a proportioning pump and a gas-liquid separator to an atomic absorption spectrometer is described. Sulphide ions react with 3 mol dm-3 hydrochloric acid and the released hydrogen sulphide is swept into a gas-liquid separator by an air stream. The absorbance was measured at 200nm using a deuterium hollow cathode lamp. The method is relatively free from interference with a detection limit for sulphide of 0.06 pg mi-’ and relative standard deviations of 1.4-3.3% for repeated analyses. The calibration graph is linear up to 100 pg ml-1 of sulphide.Twenty samples can be analysed in 1 h. The method has been applied to the determination of sulphate-sulphur in plants. Keywords: Gas-phase molecular absorption spectrometry; automated sulphide determination; sulphate- sulphur in plants The determination of anions and cations in solution by conversion of the determinant into a volatile molecular species followed by their molecular absorbance measurements in the gas phase has been thoroughly investigated in the past two decades. The gaseous product is carried by a stream of air or nitrogen to a flow-through absorption cell which is positioned in the light path of the spectrometer in the space normally occupied by the flame of the atomic absorption spectrometer. A narrow band of radiation corresponding to an absorption maximum of the evolved compound is passed through the cell and the absorbance signal of the compound is measured.The technique, which is known as gas-phase molecular absorption spectrometry (GPMAS) , was de~elopedl-~ in this laboratory several years ago during a search for a rapid and reliable method for the determination of ammonium-nitrogen in digests of soil and plant samples using the Kjeldahl method. The GPMAS technique has been applied to the determination of several anions and cations in a variety of biological, environmental and food samples.Gl6 Determination of sulphide is important because of its extreme toxicity as hydrogen sulphide and its objectionable odour.There is also considerable interest in its measurement because one of the steps in the accurate determination of total sulphur in soils and plants involves the conversion of the various sulphur compounds in the sample into sulphide by reduction. The analytical chemistry of sulphide has been included in several reviews of the general analytical chemistry of sulphur compounds. 17-21 A variety of analytical techniques have been applied to the determination of sulphide in environmental samples. The most widely used methods can be grouped into three categories: titrimetric,2(”-*2 electrochemical,23-25 and spectro- scopic. The last includes ultraviolet-visible spectrophoto- metric methods2628 and molecular emission methods.29.30 However, most of these methods are manual and involve considerable manipulation. They lack speed, simplicity and precision for routine analysis of a large daily throughput of samples.Of the various colorimetric methods recommended for the determination of sulphate-sulphur in soils and plants, the Methylene Blue procedure developed by Johnson and Nishita‘h is the most sensitive and accurate. It has been widely used for many years and thoroughly investigated.27 Leggett et ~ 1 . 2 8 developed a flow injection method for sulphide determination by using the Methylene Blue method. Although Vijan and Wood31332 have developed an automated * ‘To whom correspondence should be addressed. version of Cresser’s GPMAS method for ammonia, auto- mated GPMAS procedures have apparently not been de- veloped for sulphide determination.This paper describes the automation of the gas-phase molecular absorption spectrometric method for sulphide determination. It is a rapid and specific method which requires minimal sample treatment. The system has been automated by the introduction of an automatic sampler, a proportioning pump and a gas-liquid separator. The proposed method was applied to the determination of sulphate-sulphur in mixed herbage (grass/clover) . Experimental Apparatus A Shandon Southern A3600 atomic absorption spectrometer was used with a deuterium hollow cathode lamp and an Auto-graph S chart recorder. The spectrometer was modified for a non-flame cold vapour analysis by removing the burner head and replacing it with a 13 cm long, quartz-windowed flow-through absorption cell supported by a holder.The carrier gas (air) was introduced into the system (as shown in Fig. 1) via a plastic T-piece similar in dimensions to the gas-liquid separator. The distance between the point at which the carrier gas entered the system and the outlet of the tube carrying the reacting solutions/hydrogen sulphide was about 7cm (see Fig. 2). The control of the lower flow-rates was achieved by passing the compressed air through a flow meter which was connected to a Brooks Flow Controller No. 8943 (Brooks Instrument, Hatfield, PA, USA). All measure- ments were made at 200 nm. Reagents All reagents were of analytical-reagent grade and de-aerated, de-ionized water was used throughout. Sulphide stock solution, 500 pg ml-1.Prepared by dissolving 1.875 g of sodium sulphide (Na2S.9H20) in 500 ml of 25% sulphide anti-oxidant buffer (SAOB) solution. Working standards were freshly prepared each day in 25% SAOB by the least number of dilution steps possible. Interferent solutions, 1000 pg ml-1. Solutions of a range of cations and anions were prepared from analytical-reagent grade salts. Stability of sulphide solutions Sulphide solutions are unstable as they are very readily oxidized by the ambient air. The SAOB23 is a reagent usually596 'I ANALYST, JUNE 1991, VOL. 116 Hollow cathode lamo ml min-, Recorder Monochromator + 0.6 Air 1 1 - PMT 1.0 Sampe 2 0.23 HCI 3 - 1.6OWash 4 - 1.60 5 - - J - - I flow controller - Waste ( t )Waste - . Fig. 1 spectrometer system Schematic diagram of the autoanalyser-atomic absorption TO absorption 1 /pipette tip 5 mm i.d.i , , f ' ' From manifold- T! I4-35 mm From Brooks- flow controller - 14-35 m m 4 I To waste via pump Fig. 2 Gas-liquid separator and associated connections added to samples containing sulphide in order to raise the sample pH, free the sulphide bound to hydrogen, fix the total ionic strength and retard the oxidation of sulphide. In the original procedure from Orion Research,23724 the SAOB consisted of sodium hydroxide, sodium salicylate and ascorbic acid in de-aerated, de-ionized water.33 Baumann25 used an alkaline ethylenediaminetetraacetic acid (EDTA)-ascorbate solution for sample treatment in the direct determination of sulphide at concentrations >30 pg 1-1 with an ion-selective Table 1 Operating conditions Spectrometer- Wavelength setting 200 nm Mode Absorbance Lamp current 10 mA Slit-width 0.50 mm Bandpass 3.0 nm Recorder- Sensitivity Chart speed 2.5-100 mV full scale 5 mm min- Proportioning pump and automatic sampler- Sampling cycle 45 s Wash cycle 45 s Air flow-rate 0.6 ml min-1 3 mol dm-3 HCl flow-rate 0.23 ml min-l Sample or wash uptake rate 1.0 ml min-1 Carrier gas (air) flow-rate 13.3 ml min-l Mixing coil 29 turns 200 250 300 Wavelengthlnm Fig.3 Absorption spectrum of hydrogen sulphide electrode (to increase sensitivity) while Tanaka et al.34 included EDTA in the absorption mixture used to trap atmospheric hydrogen sulphide (to prevent oxidation of the sulphide ions). When sodium salicylate was replaced by EDTA, a noticeable improvement in the stability of the sulphide solution was observed.Consequently, the SAOB used in this study consisted of 2 mol dm-3 sodium hydroxide, 0.2 mol dm-3 ascorbic acid and 0.2 mol dm-3 EDTA solutions. The ascorbic acid serves to prevent the oxidative loss of sulphide by its own oxidation to dehydroascorbic acid35 while the EDTA masks the trace metal ions that are catalysing the oxidation.34 Instrumental Operation The operating parameters for the various parts of the instrument and manifold are shown in Table 1. The manifold tube sizes were selected by trial and error until the optimum sensitivity was achieved. A slit-width of 0.5mm was used throughout. Cressers-7 has already observed that the slit- width is not crucial for sulphide determination owing to the simplicity of the hydrogen sulphide spectrum (see Fig.3). This observation was also confirmed in the present investigation. The instrument was allowed to warm up for 5 min with the sampling needle in the 'wash' position. The manifold tubes were lowered into the appropriate reagent bottles and the solutions were allowed to flow. With the carrier gas (air) flow rate set at 13.3 ml min-1, air was allowed to fill the absorption cell. After the entire cell system had equilibrated, the zero on the instrument and recorder was adjusted.ANALYST, JUNE 1991, VOL. 116 597 Procedure Samples were supplied to the manifold, shown in Fig. 1, by a Technicon Sampler 11 with a 40 sample capacity. The reacting solutions (sample and hydrochloric acid) were carried by a Technicon proportioning pump and were mixed in the mixing coil just before entering the gas-liquid separator.The gas-liquid separator was made, after several preliminary experiments, from a glass T-piece (see Fig. 2) connected to the absorption cell by a pipette tip and a narrow bore tube (i.d. 0.82mm). The evolved hydrogen sulphide was swept by the carrier gas into the flow-through absorption cell which was aligned in the optical path of the spectrometer. Once the sampler tray was loaded with blanks, standards and samples, the sampler was switched on and the analysis commenced. The absorbance values of the standards were measured to prepare a calibration graph. Results and Discussion Optimization of Experimental Conditions Effect of carrier gas flow rate Supplementary air was introduced to sweep the evolved hydrogen sulphide into the absorption cell.The effect of the carrier gas flow rate on the absorbance signal was evaluated by making a series of analyses of a 20 pg ml-1 sulphide solution while varying the flow rate from 10 to 30 ml min-1. The results, shown in Fig. 4, indicate that the sulphide absorption intensity varied with the carrier gas flow rate, i.e., decreasing gradually with increasing flow rate. As expected, the highest absorbance signals were observed at low flow rates. This is because the dilution of the evolved hydrogen sulphide by the carrier gas was less. Flow rates lower than 10 ml min-l gave rise to very broad peaks while higher flow rates (>30 ml min-1) were not suitable because of the consequent dilution of the evolved hydrogen sulphide.Also, high flow rates might force liquid into the cell. Most of the data in this study were collected at a flow rate of 13.3 ml min-I. Effect of temperature The effect of temperature on the rate of evolution of hydrogen sulphide was investigated by lowering the mixing coil into hot water at different temperatures (between 40 and 65 "C) but no significant improvement was observed. Therefore, this study was carried out at room temperature. Choice of wavelength The effect of the choice of wavelength on the slope of the calibration graph for 0-20 pg ml-1 of sulphide solution was investigated and the results are shown in Fig. 5. A linear regression treatment of the data obtained yielded the follow- ing relationship between absorbance, A, and sulphide concen- 0.13 1 0.1 1 o.'2 I 4? 0.09 s 2 0.08 O'I0 I tration, cs2-, for the different wavelengths (A) investigated.The regression coefficient ( r ) was calculated for n = 6. h = 185 nm A = 0.0031~~2- + 0.0010 r = 0.9995 h = 190 nm A = 0.0050~~~- + 0.0009 r = 0.9998 h = 200 nm A = 0.0057~~2- + 0.0009 r = 0.9996 h = 210 nm A = 0.0025~~2- + 0.0005 r = 0.9996 h = 215 nm A = 0.0017~~2- + 0.0008 r = 0.9989 h = 220nm A = 0.0011~~2- + 0.0010 r = 0.9962 The absorption spectrum of hydrogen sulphide3 is shown in Fig. 3 and for optimum sensitivity, a wavelength of 200 nm was employed. Precision Ten replicate analyses of different standard solutions of sulphide were made under the optimum conditions to test the reproducibility of the technique.The results are shown in Table 2. Detection Limit Replicate analyses of a 0.2 pg ml-1 standard sulphide solution gave a standard deviation of 0.0003 pg ml-1. Defining the detection limit as the concentration of sulphide which yields a signal twice the standard deviation for a signal close to the blank, the detection limit of the proposed method was 0.06 pg ml-1 of sulphide. The detection limit could be pushed to lower values by using a timing cam with a longer sampling time at the expense of a reduced number of samples that can be analysed per hour. Dynamic Range and Sensitivity The relationship between the absorbance of the evolved hydrogen sulphide and the concentration of the sulphide was linear up to 100 pg ml-1 of sulphide. A curvature towards the concentration axis was observed at higher concentrations.The 0.14 C,..:"' I 0.04 0.02 0 5 10 15 20 25 Sulphide concentration/pg ml-1 Fig. 5 A, 185; B, 190; C, 200; D, 210; E, 215; and F, 220 nm Effect of choice of wavelength on sulphide calibration graphs: Table 2 Precision of the proposed method at various sulphide concentrations Absorbance Relative Concentration standard of sulphide/ Standard deviation pg ml- 1 Range Mean* deviation (YO) 20 0.1 10-0.115 0.113 0.0017 1.5 12 0.064-0.066 0.065 0.0009 1.4 4 0.020-0.022 0.021 0.0007 3.3 2 0.0 124.013 0.012 O.OOO4 3.3 * Based on ten determinations.598 ANALYST. JUNE 1991, VOL. 116 slope of the calibration graph (the linear portion) was 0.0055 ml pg-1, which represents the sensitivity of the method. The equation of the calibration graph obtained by the method of least squares was A = 0 .0 0 5 5 ~ ~ 2 - + 0.0081 with r = 0.9986. Interferences The effect of various anions and cations on the absorbance of the evolved hydrogen sulphide was studied. A range of solutions was prepared containing 20 pg ml-1 of sulphide and 500 pg ml-1 of the possible interferent. Efforts were made during the preparation of these solutions to avoid premature hydrogen sulphide evolution by the addition of an appropriate volume of SAOB solution to the concomitant element solutions (pH >7). The solution containing the sulphide sample plus the potential interferent ion and another solution containing only the interferent ion (500 pg ml-1) were analysed by the proposed method. The responses were compared with those obtained from an uncontaminated sulphide solution.The anions Cl-, Br-, I-, P043-, SO42-, NO3- and C032- did not cause any interference. Solutions of 100 pg ml-1 of sulphite and 500 pg ml-1 of nitrite interfered with the determination. This is due to the evolution of sulphur dioxide by the sulphite and a mixture of gaseous oxides of nitrogen by the nitrite. When solutions of these anions were analysed alone, noticeable peaks were obtained (100 pg ml-1 of sulphite solution and 500 pg ml-1 of nitrite solution gave absorbance values of 0.083 and 0.259, respectively) thereby confirming that the gases evolved absorb at 200 nm. The cations Na+, K+, AP+, Sr2+, Ca2+, Mn2+, Mg2+ and Zn2+ had no effect on the determination of sulphide. The ions Ni2+ and Cd2+ showed a marginal effect while Fe3+ and Crvl caused a substantial depression of the sulphide signal (see Table 3).A concentration of 100 pg ml-l of Co2+ caused almost complete signal depression while a similar concentra- tion of Cu2+ totally depressed the signal. The effect of various concentrations of Cu2+ (0-70 pg ml-1) on the absorbance signal of 20 pg ml-1 of sulphide is shown in Fig. 6. Other workers6336 have also observed the deleterious effect of Cu2+ on the determination of sulphide. The addition of between 2 and 20 ml of various concentrations of EDTA solution (0.1-0.3 mol dm-3) to the solutions of sulphide and interferent did not prevent the interference. The use of a stronger acid (6 mol dm-3 HC1) did not solve this problem either. This might be due to the stability of the resultant metal sulphide that is formed.Table 3 Effect of other ions on the evolution of hydrogen sulphide and its absorbance at 200 nm; sulphide concentration of test solution, 20 pg ml-I Ion added None c1- Br- I- ~ 0 ~ 3 - sop NO3 - CO32- S032- NOz- NOz- K+ Con- centration/ pg ml-1 500 500 500 500 500 500 500 100 100 500 - 500 Relative absorb- ance* 100 98 103 102 98 102 98 97 173 99 239 100 Ion added Sr2+ Ca2+ Na+ Mg2+ Mn2+ Zn2+ Cd2+ Ni2 + Crvl Fe3 + co2+ cu2+ AP+ Con- centration/ Relative pg ml-1 absorbance* 500 94 500 99 500 100 500 94 500 98 500 97 500 81 500 85 500 99 500 55 500 16 100 4 100 0 * The ratio of the absorbance for the test solution to that for the solution containing the concomitant. Application In order to test the proposed method, sulphate-sulphur in mixed herbage (grassklover) was determined by the pro- cedure described above using a modification of the direct digestion procedure of Johnson and Nishita26 to convert the sulphur into hydrogen sulphide.This method involves the reduction of sulphate to hydrogen sulphide by a reducing mixture containing hydriodic acid, formic acid and red phosphorus. The hydrogen sulphide evolved is collected in a 100 ml glass-stoppered calibrated flask containing zinc acetate and sodium acetate and made up to the mark with SAOB (to prevent aerial oxidation of the sample as explained earlier). The hydrogen sulphide trapped by the zinc acetate and sodium acetate is then determined. A set of standard sulphur solutions (0-100 pg ml-1) (as sodium sulphate) was first analysed by the procedure of Johnson and Nishita26 followed by direct digestion of the mixed herbage.The determination of sulphate-sulphur in the samples was then carried out by interpolation from the calibration graph. The results of the analysis are shown in Table 4. The values obtained are in good agreement with those obtained by the Methylene Blue method of Johnson and Nishita. Conclusion This paper describes the automation of the GPMAS method for the determination of sulphide. It is a rapid and specific spectrometric method which requires minimal sample treat- ment. No reagent other than 3 mol dm-3 hydrochloric acid is required, thereby eliminating all concerns of reagent preser- vation and timed colour development. The method is simple, fast and direct.The manifold is simple and any atomic absorption instrument can be employed. The procedure retains much of the inherent sensitivity of the manual GPMAS technique whilst allowing the efficient processing of a large 0.12 1 I 0.10 8 0.08 - L (D ' 0.06 2 2 0.04 0.02 0 20 40 60 80 20 pg ml-1 S2- + xpg mi-' Cu2+ Fig. 6 Effect of various concentrations of Cu2+ on the absorbance of 20 pg ml-1 of sulphide ~~ ~ Table 4 Sulphate-sulphur in mixed herbage Concentration found*/mg kg-l Automated Sample No. GPMAS 1 765.5 2 774.5 3 202.2 4 266.2 5 176.1 6 763.4 * Mean of duplicate determinations. Methylene Blue method 789.4 764.8 198.7 274.8 176.2 751.5ANALYST, JUNE 1991. VOL. 116 599 number of samples. The precision and accuracy are as good as those of the conventional methods.The most attractive feature of the method is its manipulation-free unattended operation. Twenty samples can be analysed in 1 h. Interfer- ences in the proposed method are few, and it offers a good alternative for the determination of sulphide in environmental samples. The method may also be applied to the determina- tion of total sulphur in soil extracts by using the modified Methylene Blue procedure of Tabatabai and Bremne1-3’ to convert the various forms of soil sulphur into hydrogen sulphide. The latter is then collected and analysed as described above for sulphate-sulphur. Where the concentration of sulphide is much lower than the detection limit of the proposed method, a preconcentration method should be developed. The results reported here, for example, suggest that sulphide in large samples could be trapped in a much smaller volume of zinc solution prior to determination.Automated GPMAS methods for other anions are under development in this laboratory. The authors thank Tony Edwards and Denise Donald of Macaulay Land Use Research Institute, Aberdeen, for supplying the plant samples and the Methylene Blue results shown in Table 4. T . A. A. thanks the Commonwealth Scholarship Commission and the University of Agriculture, Abeokuta, Nigeria, for financial support and leave of absence, respectively. References 10 11 12 13 Cresser, M. S . , Anal. Chim. Acta. 1976, 85, 253. Cresser, M. S., Lab. Pract., 1977, 26, 19. Cresser, M. S . . and Isaacson, P. J . , Talanta, 1976, 23, 885. Cresser, M.S . . Analyst, 1977, 102, 99. Cresser, M. S., Proc. Anal. Div. Chem. SOC.. 1978, 15, 68. Cresser, M. S . . Lab. Pract.. 1978, 27, 639. Cresser, M. S . . Eur. Spectrosc. News. 1978. 19. 36. Syty, A., Anal. Chem., 1973, 45, 1744. Winkler, H. E., and Syty. A., Environ. Sci. Technol., 1976,lO. 913. Syty, A., Anal. Chem., 1979, 51, 911. Syty, A., and Simmon, R. A.,Anal. Chim. Acta, 1980.120,163. Takahashi, M.. Tanabe, K . , Saito, A.. Matsumoto. K., Haraguchi, H., and Fuwa, K., Can. J . Spectrosc., 1980,25,25. Macpherson, H. B.. At. Spectrosc., 1983,4, 150. 14 15 16 17 18 Kupchella, L., and Syty. A.. J . Assoc. Off. Anal. Chem., 1984, 67, 188. Firestone, M. A., Dietz, M. L., and Syty, A., Anal. Lett., 1985, 18, 985. Anwar. J., and Marr, I. L., J. Chem. SOC. Pak., 1986, 8, 67.Szekeres. L., Talanta, 1974, 21, 2. Blasius, E., Horn, G., Knochel, A . , Munch, J., and Wagner, H., in Inorganic Sulphur Chemistry, ed. Nickless, G., Elsevier, New York, 1968, p. 211. Haff, L. V., in The Analytical Chemistry of Sulphur and its Compounds, Part I , ed. Karchmer. J. H., Wiley-Interscience, New York, 1970, p. 285. Heinrich, B. J., Grimes, M. D., and Puckett, J . E., in Treatise on Analytical Chemistry, Part 2, eds. Kolthoff, 1. M.. and Elving, P. J . , Wiley, New York, 1961, vol. 7. p. 49. Williams, W. J . , Handbook of Anion Determination, Butter- worth. London. 1st edn., 1979. p. 568. Greenberg, A . E., Connors, J . J . , and Jenkins, D., Standard Methods for the Examination of Water and Wastewater, Public Health Association. Washington, DC, 15th edn., 1980. pp. 4424.50. 23 Determination of Total Sulfide Content in Water, Applications Bulletin No. 12, Orion Kesearch, Cambridge, MA, 1969. 24 Frant, M. S . , and Ross, J . W., Tappi, 1970, 53, 1753. 25 Baumann, E. W., Anal. Chem., 1974,46, 1345. 26 Johnson, C. M., and Nishita, H.. Anal. Chem., 1952. 24, 736. 27 Zutshi, P. K., and Mahadevan, T. N., Talanta, 1970, 17, 1014. 28 Leggett, D. J . , Chen, N. J.. and Mahadevappa, D. S., Anal. Chim. Acta, 1981, 128, 163. 29 Grekas, N., and Calokerinos, A. C., Anal. Chim. Acta, 1985, 173. 311. 30 Henden, E., Pourreza, N., andTownshend, A., Prog. Anal. At. Spectrosc., 1979, 2, 355. 31 Vijan, P. N., and Wood, G. R., Anal. Chem., 1981, 53, 1447. 32 Vijan, P. N., and Wood, G. R., Anal. Lett., 1982, 15, 699. 33 Donaldson. E. L., and McMullan, D. C., Anal. Lett., Part A , 1978, 11, 39. 34 Tanaka, S., Shimizu, Y . , and Hashimoto, Y., Bcinseki Kagaku, 1978, 27. 599. 35 Glaister. M. G., Moody, G. J., Nash, T., andThomas. J . D. R., Anal. Chim. Acta, 1984, 165, 281. 36 Johnson. C. M., and Arkley. T. H., Anal. Chern., 19.54, 26, 1525. 37 Tabatabai, M. A., and Bremner, J . M., Soil Sci. SOC. Am. Proc., 1970,34, 62. Paper 0/04556J Received October loth, I990 Accepted February 4th, I991 19 20 21 22