THE EXPLOSIOX OF ACETYLENE AND NITROGEN. PART IV. 77 XIIL-The Explosion of Acetylene and 2l’itrogen. Part I V . Spectra of Explosions of Gases con-taining Hydrogen Carbon Nitrogen and Oxygen. B)7 WILLISM EDWARD GARXER alld SIDNEY FF7ALTER SAUNDERS. THE origin of the Swan and cyaiiogeii band spectra has frequently been discussed and although the former is usually ascribed to free carbon (Watts Phil. Mag. 1914 [vi] 28 117) and the latter to carbon and nitrogen opinion is not unanimous on these questions. Before 1914 i t was believed that the cyanogen bands were due to carhon niid nitrogen b u t Grotriaii and Runge (Yhysikal. Z. 1914, 15 545) reopened the discussion on the origin of this spectrum, and concluded as a result of their experiments with the electric arc that the so-called cyanogen spectrum was due to nitrogen alone.This view seems to have been widely accepted although neither Barratt (Proc. Roy. Soc. 1920 [A] 98 40) who experi-mented with flames nor Hernsalech (PhiZ. X a q . 1920 [vi] 39, 241) nor Kolst and Oosterhuis ( K . Acad. Amsterdam Proc. Sci., P!E1 23 1 727) who suppressed the cyanogen bands by immersin 78 GARNER AND SAUNDERS : their nitrogen and argon discharge tubes in liquid oxygen and so freezing out the cyanogen have confirmed it. The results obtained in this paper also indicate that the presence of both carbon aiid nitrogen is essential for the production of the " cyanogen '' spectrum .* It has been found that the nature of the spectrum emitted during explosions of acetylene in oxygen and nitrogen is very largely dependent on the ratio of gram-atoms of carbon to oxygen in the mixture.With low percentages of oxygen where large quantities of carbon are deposited during the explosion a continuous spectrum only is obtained. The carbon being an almost perfect black body acts as a screen cutting off the radiation emitted from the molecules activated during the explosion. As the ratio of oxygen to acetylene approaches unity the Swan CH and cyanogen bands, and metallic absorption and emission lines appear on the back-ground formed by the continuous spectrum which simultaneously diminishes in intensity. These band spectra persist into the region where oxygen)carbon but gradually fade away as the oxygen percentage is increased until finally with a still larger proportion of oxygen only the line emission and absorption spectrum of the metallic impurities are present.The Swan and cyanogen spectra appear and disappear together as the oxygen percentage increases. Their disappearance occurs approximately at the concentrations found previously (J. 1924, 125 1634) for the disappearance of free carbon and hydrocyanic acid from the h a 1 products of the explosion uiz. when the ratio of oxygen to carbon becomes unity. Thus the presence of free carbon in the products of the explosion is essential not only for the production of hydrocyanic acid but also for both the Swan and the cyanogen spectra. The Swan and cyanogen spectra persist into the region where oxygen is in excess (up to C 0 = 0.956 for mixtures of C2H2 and 0, and C 0 = 0.968 for mixtures of 0, N, and C,H,) but this does not invalidate the above conclusion for the spectrum of the explosion is a record of the radiation emitted throughout the whole of the explosive process and hence includes the radiation emitted before equilibrium is reached.On the molecular collision view of the propagation of the explosion wave undecomposed acetylene molecules immediately in front of the wave will be bombarded by swiftly moving molecules of carbon, hydrogen oxygen carbon monoxide etc. from the explosion wave itself. An oxygen molecule in collision with an acetylene mole-* Since this paper was written Freundlich and Hochheim (2. Physik 1924, 26 102) in furnace experiments have shown that elementary carbon is essential for the production of this spectrum THE EXPLOSION OF ACETYLENE AND SITROCEX.PART I V . 79 cule would be likely to give carbon monoxide but no free carf;on, C,H -j- 0 + 2CO + H, but any other molecule may bring about the change C,H + 2C + H,. Thus evenin the presence of' excess of oxygcn free carbon may be momentarily formed and it' the averags life of the carbon particles (C possibly) beiore rtxiioval as carbon nionoxide is long enough the Swan spectrum may be emitted. This would explain our results (Table I). The results of this investigation thus support the vie17 that carbon is necessary for the production of the cyanogen spcctrunl. That nitrogen also is necessary is evident from the abseiice of the cyanogen spcctrurn in the spectra of explosions of acetylene and oxygen alone.E X P E R I M E w T A L . examination has been made of the spectra from the follovVring series of explosive mixtures ( a ) pure acetylene ( b ) acct) Icnc~ md nitrogen (c) acetylene and oxygen (a?) acetylene o x ) - p i antl nitrogen ( e ) electrolytic gas (f) carbon monoxide and oxyge~i, (9) coal gas and oxygen. Appumtus.-A 2-litre phosphor-bronze bomb similnr to that described by Wheeler (J. 191 8,113 855) was fitted with a c.j~!indr.ic~al quartz window 6 mm. in diameter and 8 mm. thick placed oppo.-il e the collimator tube of a Hilger Constant Deviation Spectronictc.r.* Ignition of the gaseous mixture was usual1~- accomplishctl by fusion of a thin iron wire placed across the terminals in the ceritre of the bomb. In some experiments the position of firing of the mixture was changed and in others the iron wire was replaced by ~)latitiuin.Preparation of the Gases.-The acetxlene from a cylinder con taining the gas compressed in acetone was washed with water, dried over phosphorus pentoxide and stored over merciiry. It was completely soluble in a solution of animoniacal cuprous cliloritlc. Oxygen and nitrogen were prepared as described elsewhere (Loc. cit.). Carbon monoxide obtained by the &ion of sulphuric acid on sodium formate was washed with sodium hydroxide antl dried over phosphorus pentoxide. Electrolytic gas was prepared froin baryta. The gases were introduced into an encuated bomb, which had been previoiisly washed out six times with oxygen. This repeated washing was necessary in order to ensure the absence of nitrogen in the first series of experiments.The composition of the gaseous mixture was determined by pressure measureiuents. Xeasurement of Spectra.-Each photograph of the explosion spectrum was braclieted by photographs of a neon-helium lamp, the wave-lengths of the lines in this spectrum being knoxn an(l * A few measiirernents were made with a quartz spcctrometer SO GARNER AND SAUNDERS : the lines were measured by means of a travelling microscope. The extreme errors of measurement were & l k for the band spectra and 0.2 8. for the metallic lines. Results. (a) Acetylene and Oxygen.-With pure acetylene and also acetylene and oxygen mixtures up to 41% of oxygen a perfectly continuous spectrum was obtained (Table I). With I00 yo acetyl-ene the spectrum was rather more brilliant in the red than is the case with the mixture containing 41% of oxygen.The effect of the addition of oxygen on the position of maximum light intensity is to shift this maximum slightly towards the blue end of the spectrum, this being in agreement with Wien’s displacement law. TABLE I. Oxygen and Acetylene. P = 1 atm. % 0,. 76 C,H,. 76 C,H,/“/O 0,. Spectrum. - 100.0 * 35.22 64.78 41.10 58.90 Continuous with no line or 1.433 spectra. Swan and CH bands with metallic lines. 49-33 60.67 49.58 50.42 49.69 50.3 1 1.013 band faint 50.30 49.70 0-983 Swan and CH bands with metallic 51.12 48.88 0.956 } lines. 53.14 46-86 0-852 59.86 44.50 40.14 ::~:ftl afetallic lines only. 55.50 62-91 37.09 0-5SOtj * 3-2 atm.Platinum wire ignition. As the ratio (column 3) approaches unity the Swan ‘‘ the three,” and the CH bands appear on the continuous spectrum. The bands 61911,5635,5165,4737,4382,4371,4365 and 4314 A. were observed. These bands first appear when the value ~OC,H,/~OO, is between 1.433 and 1.035; the exact value was not found. It was anticipated that the Swan spectrum would disappear when ~0C,H2/~002<1 since a t the temperature of explosion, ca. 3000” the equilibrium constant for the reaction C + H20 5= CO + H is Kp = PmP,,/P, = 105 and hence practicaLy no free carbon can be present if the gaseous mixture attains equilibrium. Since the Swan spectrum is considered as being due to free carbon it was expected that this together with “the three ” and CH bands would disappear when the ratio 74,C,H2/~00 (1.It was found however that these bands persisted even when the acetylene-oxygen mixture had the ratio 70C2H,/7002 = 0.956, although they had all disappeared when the ratio became 0.882 THE EXPLOSION OF ACETYLENE AND KITROGEN. PART IV. 81 The cyanogen spectrum was not observed in this series of experi-ments. (b) Acetylene Xitrogen a.nd Oxyyen.-When nitrogen was added until the mixture had the composition 82% C,H, 187; N, and the niixture was fired under a pressure of about 3 atmospheres several units per cent. of hydrocyaiiic acid were ohtaineci and yet the spectrum was still continuous. Since the amount of free carbon could not be reduced sufficiently by the addition of nitrogen alone to enable emission and absorption lines to be obtained oxygen was added to the niixture until the ratio ~OC!,H,/~OO was between 1.1 and 1.This diminished the carbon sufficiently for the Swan, " the three," and the CH bands to be observed and in addition, the cyanogen bands 4216 A. and 3883 A. These all persisted when the ratio :bC',H,/Yb@ = 0.968 (Table 11). Further addition 0 ' 1 0 o-. -47.98 40.i2 48.G5 49.28 39-55 48.73 45.97 39.96 42.92 4l*T,!) 43.1 3 45.29 51.60 5ti-04 0; C,H,. s2.2 49.56 41-81 49.22 49.99 40.06 4G.14 45-25 39.1 1 41.49 40.28 39.45 37-3G 39.79 34.73 :b C21-I,/~~> 0,. Fpcctrum. Continuous spec t rii ni. -. Swan. CH cy-anogen bands and 1.033 faint metallic lines. 1.012 0.988 0.983 'I Swm CH cyanogen baiids and ~ ~ ~ ~ * j nietollic lines.0.968 0.915 Metallic lines only. 0.620 * S o band spectra. of oxygeii causes the disappearance of these spectra,. If Grotrian and Runge's contention be correct that nitrogen will give rise to the " cyanogen " spectrum if no appreciable amount of oxygen be present then the above-mentioned disappearance of this spectrum, when %C,K,/'300 = 0.915 is not easy to explain for in such a mixture at the temperature of the explosion there can be but little free oxygen." The disappearance of this spectrum is understand-able however if i t be due t o both carbon and nitrogen. The nature of the spectrum is the same whatever the position of the iron wire used for igniting the gases; placing the iron wire close to the side of the bomb or directly in front of the quartz n-indow made no difference.* Approsiinntely 0.001 yo of oxygen a t 3000" abs 82 MCPENZIE AND STRATHERN REACTIONS OF The only lines present in the spectra of the explosions of hydrogen and oxygen and carbon monoxide and oxygen were those due to the metallic impurities present e.g. sodium calcium iron etc. As would be expected on account of the lower temperature of these fiames the metallic lines were not so numerous as in the acetylene explosions. Summary. The spectra emitted during the explosion of mixtures of acetylene, The results indicate that nitrogen and oxygen have been studied. the cyanogen spectrum is due to both carbon and nitrogen. The authors wish to express their indebtedness to the Depart-ment of Scientific and Industrial Research for a maintenance grant to one of them (S.W. S.) and for an equipment grant towards the cost of apparatus. J THE SIR WILLIAM RAMSAY INORGANIC AND PHYSICAL CHEMICAL LABORATORIES, UNIVERSITY COLLEGE LONDON. [Received October Sbh 1924. THE EXPLOSIOX OF ACETYLENE AND NITROGEN. PART IV. 77 XIIL-The Explosion of Acetylene and 2l’itrogen. Part I V . Spectra of Explosions of Gases con-taining Hydrogen Carbon Nitrogen and Oxygen. B)7 WILLISM EDWARD GARXER alld SIDNEY FF7ALTER SAUNDERS. THE origin of the Swan and cyaiiogeii band spectra has frequently been discussed and although the former is usually ascribed to free carbon (Watts Phil. Mag. 1914 [vi] 28 117) and the latter to carbon and nitrogen opinion is not unanimous on these questions.Before 1914 i t was believed that the cyanogen bands were due to carhon niid nitrogen b u t Grotriaii and Runge (Yhysikal. Z. 1914, 15 545) reopened the discussion on the origin of this spectrum, and concluded as a result of their experiments with the electric arc that the so-called cyanogen spectrum was due to nitrogen alone. This view seems to have been widely accepted although neither Barratt (Proc. Roy. Soc. 1920 [A] 98 40) who experi-mented with flames nor Hernsalech (PhiZ. X a q . 1920 [vi] 39, 241) nor Kolst and Oosterhuis ( K . Acad. Amsterdam Proc. Sci., P!E1 23 1 727) who suppressed the cyanogen bands by immersin 78 GARNER AND SAUNDERS : their nitrogen and argon discharge tubes in liquid oxygen and so freezing out the cyanogen have confirmed it.The results obtained in this paper also indicate that the presence of both carbon aiid nitrogen is essential for the production of the " cyanogen '' spectrum .* It has been found that the nature of the spectrum emitted during explosions of acetylene in oxygen and nitrogen is very largely dependent on the ratio of gram-atoms of carbon to oxygen in the mixture. With low percentages of oxygen where large quantities of carbon are deposited during the explosion a continuous spectrum only is obtained. The carbon being an almost perfect black body acts as a screen cutting off the radiation emitted from the molecules activated during the explosion. As the ratio of oxygen to acetylene approaches unity the Swan CH and cyanogen bands, and metallic absorption and emission lines appear on the back-ground formed by the continuous spectrum which simultaneously diminishes in intensity.These band spectra persist into the region where oxygen)carbon but gradually fade away as the oxygen percentage is increased until finally with a still larger proportion of oxygen only the line emission and absorption spectrum of the metallic impurities are present. The Swan and cyanogen spectra appear and disappear together as the oxygen percentage increases. Their disappearance occurs approximately at the concentrations found previously (J. 1924, 125 1634) for the disappearance of free carbon and hydrocyanic acid from the h a 1 products of the explosion uiz. when the ratio of oxygen to carbon becomes unity.Thus the presence of free carbon in the products of the explosion is essential not only for the production of hydrocyanic acid but also for both the Swan and the cyanogen spectra. The Swan and cyanogen spectra persist into the region where oxygen is in excess (up to C 0 = 0.956 for mixtures of C2H2 and 0, and C 0 = 0.968 for mixtures of 0, N, and C,H,) but this does not invalidate the above conclusion for the spectrum of the explosion is a record of the radiation emitted throughout the whole of the explosive process and hence includes the radiation emitted before equilibrium is reached. On the molecular collision view of the propagation of the explosion wave undecomposed acetylene molecules immediately in front of the wave will be bombarded by swiftly moving molecules of carbon, hydrogen oxygen carbon monoxide etc.from the explosion wave itself. An oxygen molecule in collision with an acetylene mole-* Since this paper was written Freundlich and Hochheim (2. Physik 1924, 26 102) in furnace experiments have shown that elementary carbon is essential for the production of this spectrum THE EXPLOSION OF ACETYLENE AND SITROCEX. PART I V . 79 cule would be likely to give carbon monoxide but no free carf;on, C,H -j- 0 + 2CO + H, but any other molecule may bring about the change C,H + 2C + H,. Thus evenin the presence of' excess of oxygcn free carbon may be momentarily formed and it' the averags life of the carbon particles (C possibly) beiore rtxiioval as carbon nionoxide is long enough the Swan spectrum may be emitted.This would explain our results (Table I). The results of this investigation thus support the vie17 that carbon is necessary for the production of the cyanogen spcctrunl. That nitrogen also is necessary is evident from the abseiice of the cyanogen spcctrurn in the spectra of explosions of acetylene and oxygen alone. E X P E R I M E w T A L . examination has been made of the spectra from the follovVring series of explosive mixtures ( a ) pure acetylene ( b ) acct) Icnc~ md nitrogen (c) acetylene and oxygen (a?) acetylene o x ) - p i antl nitrogen ( e ) electrolytic gas (f) carbon monoxide and oxyge~i, (9) coal gas and oxygen. Appumtus.-A 2-litre phosphor-bronze bomb similnr to that described by Wheeler (J. 191 8,113 855) was fitted with a c.j~!indr.ic~al quartz window 6 mm.in diameter and 8 mm. thick placed oppo.-il e the collimator tube of a Hilger Constant Deviation Spectronictc.r.* Ignition of the gaseous mixture was usual1~- accomplishctl by fusion of a thin iron wire placed across the terminals in the ceritre of the bomb. In some experiments the position of firing of the mixture was changed and in others the iron wire was replaced by ~)latitiuin. Preparation of the Gases.-The acetxlene from a cylinder con taining the gas compressed in acetone was washed with water, dried over phosphorus pentoxide and stored over merciiry. It was completely soluble in a solution of animoniacal cuprous cliloritlc. Oxygen and nitrogen were prepared as described elsewhere (Loc. cit.). Carbon monoxide obtained by the &ion of sulphuric acid on sodium formate was washed with sodium hydroxide antl dried over phosphorus pentoxide.Electrolytic gas was prepared froin baryta. The gases were introduced into an encuated bomb, which had been previoiisly washed out six times with oxygen. This repeated washing was necessary in order to ensure the absence of nitrogen in the first series of experiments. The composition of the gaseous mixture was determined by pressure measureiuents. Xeasurement of Spectra.-Each photograph of the explosion spectrum was braclieted by photographs of a neon-helium lamp, the wave-lengths of the lines in this spectrum being knoxn an(l * A few measiirernents were made with a quartz spcctrometer SO GARNER AND SAUNDERS : the lines were measured by means of a travelling microscope.The extreme errors of measurement were & l k for the band spectra and 0.2 8. for the metallic lines. Results. (a) Acetylene and Oxygen.-With pure acetylene and also acetylene and oxygen mixtures up to 41% of oxygen a perfectly continuous spectrum was obtained (Table I). With I00 yo acetyl-ene the spectrum was rather more brilliant in the red than is the case with the mixture containing 41% of oxygen. The effect of the addition of oxygen on the position of maximum light intensity is to shift this maximum slightly towards the blue end of the spectrum, this being in agreement with Wien’s displacement law. TABLE I. Oxygen and Acetylene. P = 1 atm. % 0,. 76 C,H,. 76 C,H,/“/O 0,. Spectrum. - 100.0 * 35.22 64.78 41.10 58.90 Continuous with no line or 1.433 spectra.Swan and CH bands with metallic lines. 49-33 60.67 49.58 50.42 49.69 50.3 1 1.013 band faint 50.30 49.70 0-983 Swan and CH bands with metallic 51.12 48.88 0.956 } lines. 53.14 46-86 0-852 59.86 44.50 40.14 ::~:ftl afetallic lines only. 55.50 62-91 37.09 0-5SOtj * 3-2 atm. Platinum wire ignition. As the ratio (column 3) approaches unity the Swan ‘‘ the three,” and the CH bands appear on the continuous spectrum. The bands 61911,5635,5165,4737,4382,4371,4365 and 4314 A. were observed. These bands first appear when the value ~OC,H,/~OO, is between 1.433 and 1.035; the exact value was not found. It was anticipated that the Swan spectrum would disappear when ~0C,H2/~002<1 since a t the temperature of explosion, ca.3000” the equilibrium constant for the reaction C + H20 5= CO + H is Kp = PmP,,/P, = 105 and hence practicaLy no free carbon can be present if the gaseous mixture attains equilibrium. Since the Swan spectrum is considered as being due to free carbon it was expected that this together with “the three ” and CH bands would disappear when the ratio 74,C,H2/~00 (1. It was found however that these bands persisted even when the acetylene-oxygen mixture had the ratio 70C2H,/7002 = 0.956, although they had all disappeared when the ratio became 0.882 THE EXPLOSION OF ACETYLENE AND KITROGEN. PART IV. 81 The cyanogen spectrum was not observed in this series of experi-ments. (b) Acetylene Xitrogen a.nd Oxyyen.-When nitrogen was added until the mixture had the composition 82% C,H, 187; N, and the niixture was fired under a pressure of about 3 atmospheres several units per cent.of hydrocyaiiic acid were ohtaineci and yet the spectrum was still continuous. Since the amount of free carbon could not be reduced sufficiently by the addition of nitrogen alone to enable emission and absorption lines to be obtained oxygen was added to the niixture until the ratio ~OC!,H,/~OO was between 1.1 and 1. This diminished the carbon sufficiently for the Swan, " the three," and the CH bands to be observed and in addition, the cyanogen bands 4216 A. and 3883 A. These all persisted when the ratio :bC',H,/Yb@ = 0.968 (Table 11). Further addition 0 ' 1 0 o-. -47.98 40.i2 48.G5 49.28 39-55 48.73 45.97 39.96 42.92 4l*T,!) 43.1 3 45.29 51.60 5ti-04 0; C,H,.s2.2 49.56 41-81 49.22 49.99 40.06 4G.14 45-25 39.1 1 41.49 40.28 39.45 37-3G 39.79 34.73 :b C21-I,/~~> 0,. Fpcctrum. Continuous spec t rii ni. -. Swan. CH cy-anogen bands and 1.033 faint metallic lines. 1.012 0.988 0.983 'I Swm CH cyanogen baiids and ~ ~ ~ ~ * j nietollic lines. 0.968 0.915 Metallic lines only. 0.620 * S o band spectra. of oxygeii causes the disappearance of these spectra,. If Grotrian and Runge's contention be correct that nitrogen will give rise to the " cyanogen " spectrum if no appreciable amount of oxygen be present then the above-mentioned disappearance of this spectrum, when %C,K,/'300 = 0.915 is not easy to explain for in such a mixture at the temperature of the explosion there can be but little free oxygen." The disappearance of this spectrum is understand-able however if i t be due t o both carbon and nitrogen.The nature of the spectrum is the same whatever the position of the iron wire used for igniting the gases; placing the iron wire close to the side of the bomb or directly in front of the quartz n-indow made no difference. * Approsiinntely 0.001 yo of oxygen a t 3000" abs 82 MCPENZIE AND STRATHERN REACTIONS OF The only lines present in the spectra of the explosions of hydrogen and oxygen and carbon monoxide and oxygen were those due to the metallic impurities present e.g. sodium calcium iron etc. As would be expected on account of the lower temperature of these fiames the metallic lines were not so numerous as in the acetylene explosions. Summary. The spectra emitted during the explosion of mixtures of acetylene, The results indicate that nitrogen and oxygen have been studied. the cyanogen spectrum is due to both carbon and nitrogen. The authors wish to express their indebtedness to the Depart-ment of Scientific and Industrial Research for a maintenance grant to one of them (S. W. S.) and for an equipment grant towards the cost of apparatus. J THE SIR WILLIAM RAMSAY INORGANIC AND PHYSICAL CHEMICAL LABORATORIES, UNIVERSITY COLLEGE LONDON. [Received October Sbh 1924.