518 Analyst, August, 1968, Vol. 93, pp. 518-521 Molecular-emission Spectroscopy in Cool Flames Part In.* The Emission Characteristics of Tin in Diffusion P b e s BY R. M. DAGNALL, K. C. THOMPSON AND T. S. WEST (Chemistry Department, Imperial College, London, S. W. 7 ) Tin can be determined in the range 3 to 3000 p.p.m. by molecular-band emission from the SnH species in a nitrogen - hydrogen diffusion flame. The emission a t 609-5 nm, which is almost line-like, gives a limit of detection of 1.5 p.p.m. of tin. Spectral interference from sodium is eliminated by use of a didymium filter. The presence of oxygen gives rise to a much broader spectrum, caused by tin(I1) oxide formation, with a visual limit of detection of 5 p.p.m. of tin. Atomic emission can be observed only in the presence of alcohols, e g ., isopropyl alcohol, but a high concentration of ground-state tin atoms exists in the diffusion flame. Mechanisms are discussed to explain the production of tin atoms, SnH and the resonance-line emission. TIN is an example of an element that is difficult to excite thermally in conventional pre-mixed flames. Herrmann and Alkemade,l for example, state that only in an oxy-cyanogen flame are the tin lines intense enough to be used analytically. In other flames, weak resonance emission can only be observed in the inner reaction zones. However, the cool diffusion flames of nitrogen and hydrogen used in earlier papers in this series for the determination of sulphur2 and phosphorus3 have been found to promote emission from molecular species not encountered in most other flames.The purpose of this paper is to report the results of a study involving tin in diffusion flames, which is of both analytical and spectroscopic interest. The diffusion flame obtained by burning a mixture of hydrogen and nitrogen in air at atmospheric pressure yields an intense red colour in its cool inner region when a tin(I1) chloride solution is sprayed into it. A thin blue outer mantle may also be observed when there is a high enough concentration of oxygen to form the tin oxide species. SPECTRAL CHARACTERISTICS BAND EMISSION- The spectrum, recorded on a Unicam SP9OOA flame spectrophotometer, consisted of a sharp peak at 6096 nm and a much weaker band with a maximum a t 621 nm (Fig. 1 (b)). Sodium is included to show the line-like nature of the Sn-H emission.The former peak is listed by Herrmann and Alkemadel and corresponds to the SnH species. Pearse and Gaydon4 also list the weaker band at 621 nm, which has only been observed previously in a tin arc operating in a hydrogen atmosphere. The emission at 609.5 nm is not very dependent on hydrogen pressure or on the position of measurement in the flame, although it reaches a maximum about 1 cm above the top of the burner head. The intensity of the emission is directly proportional to the concen- tration of tin(I1) and also to the square of the slit width (which is to be expected for a band- emission signal). Linear calibration graphs may be obtained over the range 3 to 3000 p.p.m. of tin with a detection limit of about 1.5 p.p.m.(signal-to-noise ratio = 1). The effects of extraneous cations and anions were not investigated but, unless a monochromator of fairly high resolution is used, spectral interference may be expected from sodium. However, in this instance a didymium filter interposed between the flame and monochromator slit was found to reduce the sodium emission at 589 nm by 99.5 per cent., while at the same time transmitting about 60 per cent. of the emission caused by SnH at 609.5nm. * For details of earlier parts of this series, see reference list, p. 521. 0 SAC and the authors.DAGNALL, THOMPSON AND WEST 519 Because the flame temperature is low2 (down to 280" C) there is little thermal emission, even from the most easily excited elements. The noise level caused by the flame is also low because the flame exhibits negligible background, even over the OH band region.Some matrix effects caused by aluminium and magnesium have been observed in former studies5 with this flame, and they were overcome then by introducing a small volume of oxygen through an auxiliary jet in the burner stem6 to increase the flame temperature. In this instance this would not be possible, as the presence of a small amount of oxygen replaces the red colour of the SnH species by a blue coloration. The recorded spectrum of the latter showed a continuum with superimposed bands from 360 to 540 nm, with a maximum intensity at 484nm, and corresponded to the tin(I1) oxide species,l cf. Fig. 1 (a). No analytical use can be made of the blue colour with the Unicam SP9OOA spectrophotometer because of the absence of sharp bands, but it is sensitive visually with a limit of detection of about 5 p.p.m.of tin. I I I 1 I 3 (a) (b) x v) C W U c .- - I I I I I Fig. 1. Emission spectrum of Sn-0 and Sn-H species: (a), spectrum of 3 x 1 0 - 2 ~ tin(I1) ( 0 . 5 ~ hydrochloric acid) in the nitrogen - air -hydrogen flame. Slit, 0.03 mm, gain 3,10, bandwidth 1; (b) spectrum of 3 x M tin(I1) (0.5 M hydrochloric acid) in the nitrogen - hydrogen flame. Slit, 0.03 mm, gain 3,lO. band width 1 Sodium Solutions of tin(1V) chloride produced an exactly similar spectrum [i.e., SnH and tin(I1) oxide], but the emission was about 60 per cent. more intense at 609.5 nm than that obtained with tin(I1) chloride. The increased response to tin(1V) is probably caused by the greater volatility of its chloride, which would be more in evidence in a cool flame than in a hot one.As a result of these observations it may be concluded that a substantial number of elements whose salts do not break down at the temperature of this flame (about 280" to 400" C) would cause matrix effects and thus interfere although in many instances it should be possible to produce calibration graphs of lesser gradient if the amount present were approximately known. The emission characteristics of the SnH species were also observed in the air - hydrogen separated flame.2 However, under these conditions, the emission signal was only about half of that in the diffusion flame. Further, concentrated solutions of tin caused severe staining of the quartz separator, and the flame exhibited a higher background than the diffusion flame.The SnCl and SnBr bands4 are not observed in these flames, unless the halogen concen- tration is high, e.g., 0.1 M in hydrochloric or hydrobromic acids. Under theselatter conditions a thin dark blue cone appears when chloride is present and a similar shaped green cone when bromide is present. Each cone extends about 1 cm above the top of the burner head in the coolest portion of the flame. The red SnH bands were present as before, extending much higher in the flame. It would seem that the SnCl and SnBr bands arise from the chemi- luminescent reduction of tin chloride and bromide clotlets by hydrogen. Anionic interference might be more marked, however.520 DAGNALL, THOMPSON AND WEST MOLECULAR-EMISSION [AndJJSt, VOl.93 LINE EMISSION- No thermal emission of tin at the principal resonance lines 284 and 286.3 nm was observed from aqueous solutions of tin(I1) or tin(1V) salts in any diffusion flame. However, a 1000 p.p.m. tin solution, 40 per cent. w/v in isopropyl alcohol, gave not only strong emission signals at 284 and 286.3 nm, but also many other tin lines extending far into the ultraviolet region. The major lines observed were at 286.3, 284, 270.6, 254.7, 243, 235.5, 226.9 and 224.6 nm. These lines have previously been observed by Gilbert from isopropyl alcohol solutions of tin in air - hydrogen flames with a total-consumption burner.' The chemiluminescence emission from the diffusion flame is clearly observable at 284nm even in the presence of 1 per cent.isopropyl alcohol. For emission at the 224.6-nm line, an excitation energy of 5.5 eV is required which is considerably greater than that available in the diffusion flame. Although tin atoms are produced in the diffusion flame from aqueous solutions [from atomic-absorption studies of tin in diffusion flames (R. M. Dagnall, K. C. Thompson and T. S. West, unpublished work)], no emission can be observed as excitation does not occur. It was not possible to determine whether the emission from the isopropyl alcohol solutions emanated from the relatively cool oxygen-free central flame regions or from the outer reaction zone. As a result of these observations we are led to conclude that the production of tin atoms in such diffusion flames must arise as the result of chemical reaction rather than thermal decomposition of tin(I1) oxide or tin(I1) chloride.In the presence of isopropyl alcohol the tin atoms most probably undergo excitation by decomposition products of isopropyl alcohol, e g . , CH. Gibson, Grossman and Cookeg have suggested a three-body reaction to account for this emission from isopropyl alcohol solutions in an air - hydrogen flame, viz.: Sn + CH + OH -+ Sn* + CO + H,. The emission observed in our studies extended throughout the flame and was quite intense, being as high as 4 cm above the burner head. When air was supplied to the diffusion flame through an auxiliary jet in the burner baseJ6 the emission higher up in the flame decreased with increasing air pressure, while the emission just above the burner head increased.The presence of air created an inner (ie,, primary) reaction zone from which the tin emission emanated. Ethanol, methanol and acetone gave emission signals similar in intensity to isopropyl alcohol. Although the emission increased with increasing alcohol concentration, it was not linear, the emission for a 40 per cent. w/v isopropyl alcohol - tin solution being only 50 per cent. higher than a 20 per cent. w/v solution. Glycerol gave only slight emission at 284 nm from a 1000 p.p.m. tin solution, but this could have been caused by increasedviscosity. EXPERIMENTAL APPARATUS- A Unicam SP9OOA atomic-absorption - flame-emission spectrophotometer was used in its emission mode of operation, coupled to a 0 to 10-mV Servoscribe recorder. The normal EM1 9529B photomultiplier was replaced by a more ultraviolet-sensitive EM1 9601B photomultiplier.FLAME CONDITIONS- The nitrogen-hydrogen diffusion flame was obtained with an acetylene jet in the standard burner base, with the 1.8 x 7.5-cm air - acetylene emission head supplied. The nitrogen pressure was set at 15 p s i . , and the hydrogen pressure corresponded to areading of 16cm on the dibutyl phthalate filled manometer gauge. The top of the burner head was set about 06cm below the bottom of the monochromator slit. REAGENTS- to AnalaR specifications were suitable. hydrochloric acid and dilute to 1 litre with distilled water. The reagents used were of analytical-reagent grade ; we found that reagents conforming Tirt(I1) soZution, 2000 p.p.m.-Dissolve 2 g of tin metal in 200 ml of concentratedAugust, 19681 SPECTROSCOPY IN COOL FLAMES.PART I11 521 PREPARATION OF CALIBRATION CURVE (20 TO 200 P.P.M. OF TIN)- Transfer by pipette 1 to 10ml of a 2000 p.p.m. tin(I1) solution into a series of 100-ml calibrated flasks and dilute to volume with distilled water. Nebulise into the nitrogen- hydrogen diffusion flame by using a slit width of 0.15 mm, gain 3,lO and band width 3, and measure the emission at 609.5 nm. No background correction is necessary. The readings should be taken soon after solution preparation because a decrease in response is obtained on standing (e.g., overnight). This is probably caused by the hydrolysis of the tin(I1) chloride. Other concentration ranges can be prepared by suitable dilution and appropriate adjustment of the slit and gain controls.CONCLUSION Although the SnH species, which are shown to be prevalent in diffusion flames of nitrogen and hydrogen, provide a sensitive method of analysis for tin, the mechanism of their production is perhaps of even greater interest. The cool diffusion flame (burning on a Perkin-Elmer triple-slot burner) is known to provide as good an atom reservoir for tin as an air - hydrogen pre-mixed flame for atomic-absorption studies (R. M. Dagnall, K. C. Thompson and T. S. West, unpublished work), The production of tin atoms, and subsequently SnH, presumably occurs via an exothermic reduction process involving hydrogen and tin( 11) chloride. The thermal energy of the diffusion flame is insufficient to dissociate tin(I1) oxide (dissociation energy 5.7 eV), and hence the breakdown must be caused by the highly reducing nature of the flame and the almost complete absence of molecular or atomic oxygen. We have suggested this type of breakdown in a previous paper to explain the dissociation of certain metal phosphate^.^ The air - hydrogen pre-mixed flame gives more than twice the absorbance at the major tin resonance lines of the nitrous oxide - acetylene or air - acetylene flames and this again could be caused by a lower concentration of atomic oxygen.The emission produced from alcoholic tin solutions in the &flusion flame appears to favour a reaction mechanism involving a carbonaceous species such as the CH radical, but not the CN radical as this species is not observed in the flame spectrum.Support is given to this theory because there is insufficient thermal energy available to give excitation at wavelengths as low as 224.6 nm. In addition, resonance-line emission can be observed from aqueous solutions of tin in the nitrogen - acetylene diffusion flame maintained on the same burner system as before. This flame exhibits a light blue region extending 2 to 3 cm above the burner top and is completely luminous above this. The tin emission may be observed in the lower light blue region that shows only CH and C, emission. In general, the emission intensities are only about 10 per cent. of those observed from the inner reaction zones of a slightly fuel-rich air - acetylene flame maintained on the same burner. There is negligible emission above the inner cones of the pre-mixed flame. One of us (K.C.T.) wishes to thank the Science Research Council for the award of a grant. We also thank Unicam Instruments Limited for the loan of the spectrophotometer used in this study. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. NOTE-References 2 and 3 are to Parts I and I1 of this series, respectively. Herrmann, R., and Alkemade, C . T. J., “Chemical Analysis by Flame Photometry,” Interscience Dagnall, R. M., Thompson, K. C., and West, T. S., Analyst, 1967, 92, 506. -,-,- , Ibid., 1968, 93, 72. Pearse, R. W. B., and Gaydon, A. G., “The Identification of Molecular Spectra,” Chapman and Dagnall, R. M., Thompson, K. C., and West, T. S., Talanta, 1967, 14, 1467. Mackison, R., Analyst, 1964, 89, 764. Gilbert, P. T., “Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy,’’ Gibson, J. H., Grossman, W. E. L., and Cooke, W. D., Analyt. Chem., 1963, 35, 269. Publishers Inc., New York and London, 1963. Hall Ltd., London, 1950. March, 1961. Received February 1st 1968