Analyst, January, 1973, Vol. 98, $9. 49-52 49 The Direct Determination of Mercury by Atomic-absorption Spectrophotometry at 184.9 nm by Using a Nitrogen-separated Nitrous Oxide - Acetylene Flame BY G. F. KIRKBRIGHT, T. S. WEST AND P. J. WILSON (Chemistry Department, Imperial College, London, SW7 2A Y) A modified atomic-absorption spectrophotometer with a nitrogen-purged optical path has been used for the direct determination of mercury by atomic- absorption spectrophotometry, by using a nitrogen-separated premixed nitrous oxide - acetylene flame and electrodeless discharge lamp source. The high transparency of the fuel-rich flame below 200 nm permits the use of the principal resonance line of mercury at 184-9 nm for its determination. The sensitivity (for 1 per cent. absorption) enhancement obtained by using this line rather than the 253.7-nm line is fifty-fold; detection limits for mercury(1) and mercury(I1) of 0.02 p.p.m. and 0.05 p.p.ni.are achieved in aqueous solution. To date the direct determination of mercury by flame atomic-absorption spectroscopy has been undertaken by using the mercury 253.7-nm line that corresponds to the spin forbidden transition 6s2 lS,- 6p3PI0. The sensitivities achieved at this line have varied from 2.0 to 10.0 p.p.m.1s2 The principal resonance line of mercury corresponding to the allowed transi- tion 6s2 lS, - 6p1P1 lies at 184.9 nm, and the high background absorption from most flames, and from the atmosphere at this wavelength, prevents the use of this line for the deter- mination of mercury.It has been widely predicted, however, that the sensitivity attainable with atomic-absorption spectrophotometry at this line should be as much as fifty times greater than that a t 253.7 nm. Non-flame devices, which are usually purged with a non- absorbing inert gas, can be used for the determination of mercury at the 184.9-nm line, when a vacuum monochromator is also used. L’vov and Khartsyzov3 were able to operate with the 184.9-nm line in this way by using a graphite furnace atomiser and a vacuum monochromator, and achieved a sensitivity thirty times greater than that achieved at the 253.7-nm line. Recent work in this lab~ratory*-~ has demonstrated the remarkable transparency of the separated nitrous oxide - acetylene flame, and its use to determine the elements sulphur and iodine, whose principal resonance lines occur below 200 nm in the vacuum ultraviolet region of the spectrum.The transparency of this flame at short wavelengths is due to the lack of absorbing oxygen species in the interconal zone of the flame and the separation of the oxidising outer mantle by nitrogen shielding. This flame was used for the determination of sulphur a t 180.7 nm and iodine at 183.0 nm; it was expected that it might be used with a commercial atomic-absorption spectrophotometer, as modified by the use of a nitrogen- purged monochromator so as to permit detection of radiation below 200nm for the deter- mination of mercury at 184.9 nm. A microwave-excited mercury electrodeless discharge lamp was included as the source. APPARATUS- A Perkin-Elmer, Model 290B, atomic-absorption spectrophotometer was used.The apparatus was modified as previously described,6 with provision of a nitrogen-purged optical path between the source and flame, and between the flame and the photomultiplier detector via the grating monochromator. An R166 solar blind photomultiplier (Hamamatsu T.V., Japan) was fitted instead of the IP 28a (RCA Ltd.) instrument previously used; although this tube gave rise to a reduced signal intensity, the signal to background and signal to noise ratios observed at the 184.9-nm line were greatly improved. The mercury electrodeless discharge lamp source was made from silica tubing (i.d. 8 mm, wall thickness 1 mm) so as to form a bulb 240mm in length containing about 5mg of elemental mercury. EXPERIMENTAL @ SAC and the authors.50 KIRKBRIGHT 81 d.: DETERMINATION OF MERCURY BY ATOMIC-ABSORPTION [Ana&St, VOl. 98 RESULTS AND DISCUSSION The spectrophotometer used in this work does not have direct wavelength calibration. The location of the mercury 184.9-nm line was therefore found by calibration with sources that produced the mercury 253.7, cadmium 228.8, iodine 206.1, arsenic 193-7 and iodine 183.0-nm lines: the signal to background ratio at 184.9 nm obtained by using a spectral slit width of 20 nm was about 200: 1 with the mercury electrodeless discharge lamp source used. A4s expected, the signal intensity from the source was attenuated by the flame background absorption. When the optical path was purged and the nitrogen-shielding gas to the flame switched on without a flame burning, 100 per cent.transmission was easily set with a 20-nm slit width; the unshielded nitrous oxide - acetylene flame then produced 45 per cent. trans- mission compared with this value, whereas when separated by nitrogen shielding the trans- mission through the flame rose to 80 per cent. These values must be compared with those obtained through a 5-cm path length air - acetylene flame at 184.9 nm with nitrogen purging of the optical path. Fig. 1 shows the effect of the acetylene flow-rate on the flame background absorption in the nitrous oxide - acetylene flame. The numerical values shown are those of the spectrophotometer fuel control scale. Maximum transmission was obtained at values corresponding to a slightly fuel-rich flame that exhibited a red interconal zone about 40 mm in height; all absorption values were made in a region 3 to 6 mm above the primary reaction zone, As the mercury electrodeless discharge lamp produced very intense radiation at 184-9 nm, high sensitivity was obtained even at an operating power of only 10 W from the microwave generator, although above 26 W the absorption signal did begin to drop slowly, probably because of self-absorption broadening of the 184.9-nm line.c; loot- al 0 L- n r' E E l- .- cn .- C 20 6l I I I I 1 fuel flow-rate (instrument settings) Fig. 1. Effect of variation in acetylene flow-rate on flame background absorption a t 184.9 nm. Fuel flow decreases as instrument settings increase 13.6 13.7 13.8 13.9 14 SENSITIVITY, DETECTION LIMITS AND INTERFERENCES- With the optimum fuel flow-rate, spectral slit width and source operating power condi- tions established, the sensitivity values for 1 per cent.absorption shown in Table I were obtained for mercury(I1) in aqueous solution by using the 253.7 and 184.9-nm lines, and TABLE I SENSITIVITY AND DETECTION LIMITS FOR THE ATOMIC-ABSORPTION SPECTROPHOTOMETRY OF MERCURY Sensitivity (1 per cent. absorption), Linelnm Analyte Flame p.p.m. 253.7 HgINO, Air - C2H2 1.0 HgINO, Separated N20 - C,H, 2.5 Hg11(N0,)2 Air - CsHa 3.0 184.9 HgINO, Air - C,H, 0-05 HgINO, Separated N,O - C,H, 0.05 HgIVO,) 2 Air - C,H, 0.1 0.1 HgII(NO,)z Separated NaO - C,H, Detection limit, p.p.m. 2.0 3.0 5-0 0.1 0.02 0.3 0.05January, 19731 61 6-cm path length air - acetylene and separated nitrous oxide - acetylene flames.Fig. 2 shows the calibration graphs obtained for mercury(I1) and mercury(1) by using the 184-9-nm line in the concentration range between 1 and 50 p.p.m. in the separated nitrous oxide - acetylene flame. For comparison, the calibration graph obtained for mercury(1) in the same flame at 253.7 nm is also shown. The detection limits, defined as the concentration in parts per million of mercury in aqueous solution that produces a signal equal to twice the standard deviation in the background recorded near the limit of detection, are also shown in Table I. As shown in Fig. 2, the introduction of mercury as mercury(1) rather than as mercury(I1) produces an enhancement in absorption similar to that previously 0bserved.l The effect has been attributed to the disproportionation of the Hg,2+ ion to produce free elemental mercury : Hg22+ -+ Hg2+ + Hgo and provides a partial pressure of mercury vapour very low in the flame.No significant chemi- cal or physical interference effect has been observed in the absorbance signal produced by a 10 p.p.m. mercury(I1) solution in the presence of fifty-fold mass excesses of copper, cobalt, aluminium, antimony, chromium, sodium, nickel, magnesium, zinc, tin, iron, chloride, nitrate, phosphate, sulphate and fluoride. SPECTROPHOTOMETRY WITH A NITROUS OXIDE - ACETYLENE FLAME I 0 Mercury concentration, p.p.m. Fig. 2. Calibration graph for mercury: A, mercury as Hg(1) by use of a separated nitrous oxide - acetylene flame a t 184.9 nm; B, mercury as Hg(1I) by use of a separated nitrous oxide - acetylene flame at 184.9 nm; and C, mercury as Hg(1) by use of an air - acetylene flame at 253.7 nm CONCLUSION The results of the studies described in this paper suggest that the separated nitrous oxide - acetylene flame is an efficient atomising medium for the direct determination of mercury at its 184.9-nm resonance line.The sensitivity obtained for mercury(I1) is approxi- mately fifty times greater than that of the best results obtained by using conventional atomic-absorption spectrophotometry, and the instrumentation used with an air - acetylene flame for the determination of mercury at the 253.7-nm line. For relatively high concentra- tions of mercury, however, use of either flame at 253.7 nm may still possibly be the most convenient if sample dilution is to be avoided.Although with extremely intense electrodeless discharge lamp sources other flames, such as those of air - acetylene and nitrogen - hydrogen- entrained air, may provide sensitivity for 1 per cent. absorption at 184.9 nm similar to that obtained in the separated nitrous oxide - acetylene flame, by avoidance of the relatively large sample dilution effect by the flame gases that occurs in the separated nitrous oxide - acetylene52 KIRKBRIGHT, WEST AND WILSON flame, this result is achieved only at the expense of a lower flame transmission. The high flame background absorption then results in severe degradation of the signal to noise ratio obtainable, i.e. , poorer detection limits. The enhancement of the mercury absorbance at 253.7 nm when mercury(1) is nebulised rather than mercury(I1) , due to disproportionation of the Hg,2+ ion,l is also observed when the 184.9-nm line is used.The ratio of the 1 per cent. absorption sensitivity for mercury(1) observed at 184.9 nm to that at 253.7 nm in both nitrous oxide - acetylene and air - acetylene flames (50 : 1 and 20 : 1, respectively) corresponds approximately to the value of 30:l expected from calculation of the ratio of the peak absorption coefficients (K0184.9/K0253.,) by. using the classical equation for KO,’ which involves the different wavelengths, Doppler half-wdths and oscillator strengths of the lines. A similar enhancement in 1 per cent. absorption sensitivity was also observed for mercury(I1) in the air - acetylene flame at 184.9 nm compared with 253.7 nm. With the minimum sample volume requirement of 2 ml with our apparatus, the absolute limits of detection attainable for mercury(1) and mercury(I1) become 40 and 100 ng, respectively; the improvement in sensitivity obtained should enable the pre-concentration of mercury by solvent extraction, which is at present required in many analytical procedures when the final determination is made at 253.7 nm in an air - acetylene flame, to be avoided. We are indebted to Beckman Instruments Inc. for the provision of a research grant to one of us (P.J.W.) in support of this work. REFERENCES 1 . 2. 3. 4. 5. 6. 7 . Hingle, D. N., Kirkbright, G. F., and West, T. S., Analyst, 1967, 92, 759. “Perkin-Elmer Methods Handbook,” Perkin-Elmer Corporation, Nomalk, Connecticut, U.S.A. L’vov, B. V., and Khartsyzov, A. D., Zh. Prikl. Spektrosk., 1969, 11, 413. Kirkbright, G. F., and Ranson, L., Amzlyt. Chem., 1971, 43, 1238. Kirkbright, G. F., and Marshall, M., Ibzd., 1972, 44, 1288. Kirkbright, G. F., West, T. S., and Wilson, P. J., Atomic Absorption Newsletter, 1972, 11 ( 3 ) , 53. Walsh, A., Spectrochim. Actu, 1955, 7 , 108. Received July 17th, 1972 Accepted August 31st, 1972