Mechanism of Chemiluminescent Reactions Involving Nitric Oxide-the H +NO Reaction BY M. A. A. CLYNE AND B. A. THRUSH Dept. of Physical Chemistry, Lensfield Road, Cambridge Received 23rd January, 1962 The kinetics of overall combination and light emission (due to excited HNO) in the reaction of H with NO are closely similar to the kinetics of the combination of 0 with NO and its associated " air after-glow " chemiluminescence. Both reactions were studied by measurement of the in- tensities of their respective light emissions in a fast flow system at 1-2 mm Hg total pressure. Pre- liminary work has been extended and shows that in both OfNO and H+NO reactions there is a significant dependence of emission intensity and combination rate upon the nature of the carrier gas used as third body.Negative temperature coefficients for both the O+NO and H+NO emission intensities were observed, and the dissociation energy of the ground state €€NO molecule was thus determined as 48.6 kcal/mole. In a previous publication 1 it has been shown that the kinetics of the nitric-oxide- catalyzed recombination of hydrogen atoms can be represented by the reactions, H+NO + M-+HNO + M H + HNO -+ H2 + NO, where reaction (3) is rate determining. These reactions were studied by following HNO emission, the intensity of which was proportional to hydrogen atom and nitric oxide concentrations, where 10 is independent of total pressure. The kinetics of light emission and of combination in the reaction, are analogous.3 In this work, values of combination rate constants and of absolute emission intensities and their dependences upon temperature and the nature of the inert gas (M) were determined for the O+NO+M and H+NO+M systems. From these data, information regarding the nature of the reactions was obtained.In a forthcoming publication, results relating to the chemiluminescent combination reaction, will be presented. (3) (4) I = ~o[Hl"Ol, O+NO +M+NO2+ My 0 + CO + M-+CO2 + M, EXPERIMENTAL THE REACTION SYSTEM The experiments were conducted in a fast flow system of 28 mm int. diam. Pyrex tubing and similar to those previously used.1 The flow tube had several inlets along its length for admission of nitric oxide from calibrated capillary flowmeters. Hydrogen atoms were produced by a radio-frequency electrodeless electric discharge in the pure dry molecular gas or in mixtures of hydrogen (1-10 %) with purified argon, neon or helium carrier.With both arrangements, hydrogen-atom concentrations from the discharge were +-2 % of total flow. Measurements were made with total flow rates of 50-200pmole/sec and at total 139140 CHEMILUMINESCENT REACTIONS pressures of 0.7 to 3 mm Hg, corresponding to linear flow rates of the order of 1 mlsec. For experiments at low temperatures (220-290°K) the flow tube could be totally immersed in an insulated brass trough containing alcohol and solid carbon dioxide. The atomic hydrogen from the discharge passed through 30 cm of cooled reaction tube before entering the reaction zone. Observations of emission intensity were made through a double window of Pyrex placed normal to a 5 cm length of reaction tube using a 9558-B photomultiplier cell.The photomultiplier was carefully screened from scattered light emitted by the dis- charge. Ten cm downstream from the observation window was an isothermal wire calori- meter 1 for measurement of hydrogen atom concentrations. The decrease in the HNO emission intensity in passing over the isothermal calorimeter was at least 90 % in every experiment, indicating the high efficiency of the detector. THE PHOTOMULTIPLIER CELL The 9558-B photomultiplier cell was operated from a stabilized power supply at potentials between 800 and 1300 V with a constant potential of 200 V between the photocathode and first dynode. Under these conditions the dark current was very much less than the photo- electric currents due to the €€NO emissions observed in this work.The spectral response of the cell was determined from 5000 to 8000A, using Wratten filters 61, 45, 29, 88A and 74+ 29, which transmit different relatively narrow bands in this wavelength region and a tungsten strip filament lamp operating at 2800°K. Using the known distribution of radiant energy from such a source, the relative sensitivity per incident photon of the photo- multiplier was computed from the photocurrents observed with the various filters. The relative quantum efficiency of the cell is shown in fig. 1, which also shows the corresponding data for the 1P28 cell used in the earlier work.1 FIG. 1.-The intensity distribution (full line) of the H+NO and O+NO emissions as a function of wavelength.The relative quantum efficiencies (dotted line) of the photomultiplier cells are also shown. A, H+NO emission bands, this work; B, O+NO emission, this work and data of Fontijn and Schiff ; 4 X, 1P28 cell ; Y, 9558-B cell. DETERMINATION OF THE ABSOLUTE VALUE OF 10 FOR THE HNO EMISSION Before the absolute value of 10 = I/[H]mO] can be computed, it was desirable to in- vestigate the intensity distribution of the most intense bands in the HNO emission spec- trum. The relative intensities of the 6272, 6925, 7625 and 7965A bands were selected for investigation. Measurements of the same HNO emission from the H+NO reaction were made using the photomultiplier fitted with various filter combinations, selected toM. A . A . CLYNE A N D B . A . THRUSH 141 give sharp discrimination between these bands.The intensity distribution of the €€NO bands is shown schematically in fig. 1 and is in agreement with the work of Clement and Ramsay.2 No dependence of distribution on [HI, WO] or [MI was observed, nor was there any difference between systems in which hydrogen and helium respectively were used as carrier gases. There was no detectable radiation below 6000A. Fig. 1 shows that a large fraction of the emission intensity resides in the 7625 A band. The absolute intensity of the air-afterglow emission (NO+ O+ M-+N02+ M+hv) has been determined actinometrically between 4000 A and 6200 A.4 The absolute intensity of the HNO emission can, therefore, be determined by a comparison of the I0 values for the two chemiluminescent processes using the same reaction tube and photomultiplier arrangement.lo-values for the HNO emission due to the 7625 A and 7965 A bands were determined with the photomultiplier and an 88 A infra-red transmission filter when nitric oxide was added to H+H2 mixtures under conditions of negligible HNO emission decay. It was assumed that 20 % of the total €€NO emission resides in bands above 80008, and the 7625A and 7965A bands then contribute 55 % and 5 % of the total HNO emission respectively. On this assumption, the overall HNO emission intensity was determined using the known sensitivity of the photomultiplier at these two wavelengths. The photomulti- plier, now fitted with a 61 filter, was then used to determine the photocurrent due to the air-afterglow spectrum when nitric oxide was added to a stream of oxygen atoms with argon carrier under the same conditions.Oxygen-atom concentrations were measured by NO2 titration.3 Using the obtained value of 10 = I N o , / [ O ] ~ O ] , the transmission character- istics of the 61 filter, the known intensity distribution of the air-afterglow spectrum from 4000 A to 6200 A,4 and the photomultiplier sensitivity over this wavelength range, the total emission intensity of the air-afterglow spectrum in this region was computed in the same units as the value obtained for the HNO emission intensity Io. Using our determined relative values of I0 for A and H2 as third bodies and Fontijn and Schiff's 4 value for I; from 4000 to 6200 A, a value of 10 = (3.2fl-5)x 105 cm3 mole-1 sec-1 at 293°K was ob- tained for argon as carrier, as compared to 16 = 1.0 x 107 cm3 mole-1 sec-1 (4000-6200 A) for the same carrier at 293°K.It was also of interest to determine the relative intensity of the air-afterglow emission in the near infra-red, since it is known that the infra-red emission contributes considerably to the overall air-afterglow emission. Using the known photomultiplier sensitivities at various wavelengths from 5000 8, to 8000 A, the photocurrents from the air-afterglow emis- sion were measured using several filters under different conditions. The filters used were Wratten no. (74+29), 88A, 29 and 61 which when used with the 9558 photomultiplier had peak transmissions at about 7800, 7500, 6300 and 5300 A respectively. These data showed that the intensity of the emission in the near infra-red from 6200A to 80OOA is consider- able, and the approximate form of the variation of intensity with wavelength is shown in fig.1. Thus, the overall intensity of the air-afterglow emission is probably about 4 times greater than that observed in the region 4000-6200 A and we shall adopt a value for the total 16 of 4 x 107 cm3 mole-1 sec-1. RESULTS KINETICS OF HYDROGEN ATOM REMOVAL In this work, as previously, the rate of removal of hydrogen atoms in the presence of nitric oxide was determined by measurement of the rate of decay of the HNO emission produced. However, whereas in the earlier work measurements were made using the weak bands at 6272 and 6453 A, the superior red sensitivity of the 9558-B photomultiplier used in the present work enabled intensity measurements of the principal emission band at 7625 A to be made.The intensity I of this band was found to fit the rate law, with I0 independent of total pressure, as for the intensity of the shorter wavelength bands. I = ~o[Hl"Ol,142 CHEMILUMINESCENT REACTIONS The intensity I of the 7625 A band was found to obey the same rate equations as that of the 6272 and 6453A bands, where [NO] is the concentration of added nitric oxide, t the reaction time, and [HI0 is the concentration of hydrogen atoms at the photomultiplier housing in the absence of nitric oxide. Values of k3 obtained in this way at 293°K were identical to those obtained previously, within the experimental error of the determinations. Measurements of the rate of hydrogen-atom removal were also made at tempera- tures down to 231°K.Experiments to determine the conservation of nitric oxide during the combination reaction, using the two-inlet method described previously,1 were made at 2 mm Hg pressure and at temperatures of 226"K, 237"K, 246"K, 250"K, and 270°K. In each case the steady-state concentration of HNO was found to be less than 0*2[NO] and thus at all temperatures from 226 to 294"K, kq>5k3[M], leading to a lower limit of k4 >3 x 1010 cm3 mole-1 sec-1 at 226°K. It follows that the activation energy of reaction (4) is less than 4 kcal/mole. Since nitric oxide was conserved down to 226"K, measurements of the rate of decay of HNO emission intensity under these conditions enabled values of k3 to be calculated from eqn. (1).The values obtained at 231 and 265°K are given in table 1, together with previous data for higher temperatures. Fig. 2 shows the variations of k3 with temperature from 231 to 704°K. The observed negative temperature co- efficient can be expressed as an Arrhenius activation energy of - 0.6 & 0.2 kcal/mole, or as k3 = 1.5 x 1016(T/273)[-0'9*0*31 c m 6 mole-2 sec-1 over this temperature range. WlIENOI) = 1n(1o[HIo) - 2k3[NOl[M1t9 (1) TABLE 1 THE VARIATION OF I0 AND k3 WITH TEMPERATURE FOR M = H2 I0 (cm3 mole-1 sec-1) T (OK) k3 (cm6 mole-* sec-1) T (OK) 23 1 (2-04 f0.20) x 1016 224 265 (1 -45 f 0.20) 239 294 (1.48 f0.15) 248 340 (1.49 f0-15) 257 433 (1-17 &0*15) 268 704 (0.75 f0-10) 28 1 296 318 333 (7.8 f0.4) x 10s (7.6 f0.4) (6.6 f0-3) (5.6 f0.3) (5.0 -f 0- 3) (4.0 f0-2) (3.6 f0-2) (2.9 50.1) (4.4 f0.2) The absolute valms of 10 given in table 1 and 2 are based on the determination (for M = H2) at 293°K as described in this paper.The standard deviations of l o given in the tables are thus those of determinations of the relative values of 10. DETERMINATION OF THE TEMPERATURE COEFFICIENT OF 10 Measurements of the HNO emission intensity were made at nine temperatures from 224 to 333°K for several concentrations of added nitric oxide, the hydrogen- atom concentrations being simultaneously determined at the isothermal wire calori- meter. The pressure in the flow tube was 0-80 mm Hg, under which conditions the rate of removal of hydrogen atoms was negligible. In each determination the intensity of HNO emission 10 cm downstream from the isothermal calorimeter was also measured.At this second observation point the temperature of the gas was always within 2 deg. of that of the room. At each reaction temperature used the emissionM. A . A . CLYNE AND B . A. THRUSH FIG, 26, 3 FIG. 2.-The logarithmic variation of 10 and k3 with temperature. 10 is expressed in c m 3 mole-1 sec-1 and k3 in cm6 mole-2 sec-1. 143 FIG. 2a.144 CHEMILUMINESCENT REACTIONS intensity at the second observation point was less than 5 % of that at the low tempera- ture observation point, and it was found that I/[NO] measured at the second observa- tion point was accurately proportional to the heat liberated on the calorimeter wire. It follows that the efficiency of the isothermal calorimeter was high and accurately constant in runs performed at different temperatures.It has been found that I is given by I = Io[H][NO], whcre I0 is independent of total pressure. Using the known efficiency of the isothermal calorimeter for hydrogen-atom recombination, values of [HI were determined for each temperature at which I/[NO] had been measured at the first observation point. Fig. 2 shows the variation of I0 = I/[H][NO] with tempera- ture. These data lead to a negative temperature coefficient, expressible as an activa- tion energy of - 1-4+0.3 kcal/mole or in the form I0 = 4 x 105(T/273) [-2'8f0*4] cm3 mole-1 sec-1. VARIATION OF I. AND k3 WITH THE NATURE OF M Values of I0 = I/[H]mO] were determined at 293°K for mixtures of hydrogen atoms with A, Ne and He at a total pressure of 1 mm Hg.Measurements of k3 at pressures of 1-5-3 mmHg and at 293°K were also made for similar mixtures. There was no significant difference between the distributions of HNO emission bands (6272, 6925 and 7625 A) for hydrogen and helium as carrier gases. The data are shown in table 2. The absolute values of I0 given in the table are based on the determination of 10 (for M = H2) at 293"K, as described in the experimental section of this paper. The standard deviations given in tables 1 and 2 are thus those of determinations of the relative values of l o . TABLE 2 THE DEPENDENCE OF k3 AND 10 ON THE NATURE OF M AT 293°K M k3 (cm6 mole-2 sec-1) 10 (cm3 mole-1 sec-1) H2 (1.48 f0.15)~ 1016 (4-3 fo-3) x 105 A (0.87 &0-15) (3.2 f0-3) Ne (0.72 f0-10) (2.0 f0.4) H e (0.66 -+0*10) (2.4 f0.2) DISCUSSION The observation that Io, the proportionality constant for light emission, depends on the nature but not the pressure of the carrier gas which provides the majority of the third bodies in this reaction, indicates that the carrier gas must be involved in the formation and quenching of excited HNO, but in such a way that the pressure de- pendences cancel.This point is also brought out by the similarity of the variations of k3 and I 0 with the nature of the carrier gas. On this basis we can write the follow ing mechanism for the recombination. H + NO + M = HNO + M (non-radiative) H + NO + M = HNO*(lA") + M HNO"(1A") = HNO(1A') + h~ HNO"(1A") + M = HNO(1A') + M H+HNO = Hz+NO In this scheme the reverse reaction of (3b) has been omitted, since light emission occurs predominantly from levels far enough below the dissociation limit that the rate of redissociation will be lower than the rate of quenching.M.A. A. CLYNE AND B . A . THRUSH 145 Applying the stationary-state hypothesis to HNO* we obtain and to HNO,1 Of these rate constants, k5 can clearly be identified with A, the Einstein probability of spontaneous emission which is a constant in these experiments, whilst k6 the rate constant for quenching will depend on the nature of M and should increase slowly with temperature (with T* if k6 depends only on the collision number). Since table 1 and fig. 2 show that 10 has a slightly larger negative temperature coefficient than k3, k3,& and k3b must have very similar negative temperature dependences of the type frequently encountered in third order recombinations.Since the HNO(lA”-+lA’) emission spectrum in the H + NO reaction breaks off at an energy of 48.6 kcal/mole above the ground state, corresponding to the onset of predissociation in the absorption spectrum 5 the absence of an activation energy for this emission shows clearly that this energy corresponds to the dissociation energy of HNO into H (2s) + NO(2II). As the vibrational levels of the 1A” state of HNO continue above this predissocia- tion limit, it is clear that this state cannot correlate with ground state H and NO except via an appreciable potential maximum, and excited HNO must therefore be formed by a different path. As k3 has a small negative activation energy, it is prob- able that HNO is formed in a state from which it can readily cross into the 1A” state.Evidence for the existence of such a state from which crossing to the 1A” could readily occur is provided by the presence of marked perturbations in various vibrational levels of the 1A” including the lowest.5.6 This state is probably responsible for the onset of predissociation in the 1A” state which occurs close to the dissociation limit and must therefore be caused by a state which is not entirely repulsive. Spectroscopically, the only known electronic states of HNO are 1A” and 1A’ states, whereas the combination H(2S)+N0(2ll) can lead to 1A’, 1A”, 3A’ and 3A” states. Molecular-orbital considerations 7 suggest that the three lowest states of bent HNO would be 1A’, 3A” , 1A”.Thismay also be deduced by comparison with the three lowest states of the isolectronic molecule 02, from which the following states of bent HNO would be expected to be derived: 3A” (from the 3C; ground state), lA’+lA’‘ (from Id,) and 1A’ (from 1X:). Repulsion between the 1A’ states explains the observed 1A’ ground state but the 3A” state would be expected to lie below the 1A” state and to correlate with ground state H+NO. It is suggested that the radiative combination H + NO proceeds via this state and it is this state which is responsible for the perturbations and predissociations of the 1A” state, a view which is consistent with the weakness of the observed predissociation. A representation of the potential energy of suggested low-lying states in terms of r(H-NO) is given in fig.3. It is assumed that the 3A’ state derived from ground- state products is entirely repulsive, since it is related to a more highly excited state of 0 2 . The potential maximum in the 1A’ state of HNO is explained in terms of an avoided crossing. The negative temperature coefficient observed for 10 shows that no such potential maximum can exist in the 3A” state, but it is possible that the 1A’ ground state has one. If this were the case k3a would be very much less than k3b and combina- tion would occur almost exclusively in the excited 1A” state. If data were available146 CHEMILUMINESCENT REACTIONS on quenching of fluorescence of the 1A” state of HNO, the ratio kS/k6 could be evaluated and k3b determined. Such data are, however, available for the closely related combination reaction, 0 +NO + M+N02 + M, (3‘) which has a similar overall rate constant (kj = 2 x 1016 cm6 mole-2 sec-1 at 293°K for M = 02).For this reaction the quantity l o (denoted I;) is also independent of the pressure of inert gas 3 and exhibits a small negative temperature dependence.* Fontijn and Schiff have determined 1; to be 1-0 x 107 cm3 mole-1 sec-1 at 293°K with M = 0 2 I \ / \ / \ \$ ,’ internuclear distance Y (H-NO) FIG. 3.Potential energy curves of electronic states of HNO in the H+NO reaction. for wavelengths up to 6200A. Our measurements of the intensity distribution of the air-afterglow spectrum up to 8000A. indicate that the total I; is about 4 x 107 c m 3 mole-1 sec-1 which is about 100 times larger than for H+NO. The potential curves concerned in the NO2 emission 9 are closely similar to those for HNO, and it is reasonable to assume that 0 and NO approach on a relatively shallow stable potential curve and are stabilized by a third body into this state or the corresponding vibrational levels of the excited electronic state responsible for the absorption, fluorescence and afterglow spectra of N02.As in the HNO system, crossings between the states will occur readily due to perturbations and it is to be expected that at a similar pressure and temperature, the relative probabilities of radiation and quenching of an excited NO2 molecule having an energy a little below the dissociation limit would be independent of whether it was formed from Q+NO and stabilized at that energy by a third body or whether it was excited to that energy by absorption of a quantum of light.Since the dissociation limit of NO2 corresponds to a wavelength of approx. 3975 A, fluorescence quenching of Hg4047 and 4358A light is relevant to this discussion and this fluorescence spectrum appears to have a similar distribution to the air-afterglow emission.10 Baxter’s data 11 on the quenching by 0 2 of NO2 fluorescence excited with these wavelengths gives k;/ki = 3.34 x 10-9 mole c m - 3 , and hence k;b = 1.2 x 1016 cm6 mole-;! sec-1. This shows that approximately half the recombination proceeds via an excited electronic state of N02, and it is tempting to infer from this that the dis- tribution between stable electronic states of the NO2 molecules formed in this reactionM. A.A. CLYNE AND B. A. THRUSH 147 is largely statistical. The accuracy of the combined measurements is not sufficient to exclude the possibility of all the molecules being formed in the excited state which leads to radiation. In the H +NO reaction, the value of 10 is a hundred times lower than for O+NO, whilst the overall rate constant k3 of combination for similar third bodies is only reduced by a factor of two. The value of one or more of the constants k3a, k5, k6 for H+NO must differ by at least an order of magnitude from its value in the 0 +NO system. For the excited state of NO2 which we have considered, k; = 2.3 x 104 sec-1,10 and k; = 9 x 1012 cm3 mole-1 sec-1 for M = 0 2 , corresponding to a collision diameter of 0.7 A for quenching.11 The corresponding rate constants for HNO (]A”) are not known, but an indication of the probable magnitude of k5 can be made by comparison with values for the characteristic red absorption bands of monomeric organic nitroso-compounds which lie very close to the HNO transition.Thef-values of these transitions are close to 5 x 10-4 for aliphatic and aromatic nitroso-com- pounds 12 and about 2 x 10-4 for perfluoro-alkyl derivatives,l3 giving values of k5 of 5 x 104 and 2 x 104 sec-1 respectively which are close to that for NO;. Whilst the transition probability for HNO could be much lower than this, alternative explana- tions of the low I0 must also be sought. It has been shown that reaction (3a) involves no energy barrier, and it would therefore be expected from the data on the O+NO reaction that the process, ~ ( 2 s ) + N O ( ~ ~ I ) + M = HNO(~A”) + M, (34 would account for a large fraction of the recombination rate.The lower value of I0 is therefore associated with either slow crossing from the 3A“ state to the 1A” state or with an additional quenching process since it is unlikely that electronic quenching (reaction (6)) is much faster for HNO (1A”) than for NO,*. The presence of several strong perturbations in the 1A” state of HNO make the former explanation unlikely, since the crossing rate would be much higher than the collision rate of 107 to 108 sec-1 in these experiments. The vibrational-energy distribution in the 1A” state would then be similar to that in the corresponding energy levels of the 3A” state from which it is populated.Since the observed emission spectrum is predominantly from the lowest vibrational level of the 1A” state, most of the electronically excited HNO molecules formed must have lost at least 10 kcal of vibrational energy within the radiative (and electronic quenching) life-time of the 1A” state. It follows that a large number of molecules in the 3A” state will have lost more than this amount of vibrational energy, and as a result are below the (O,O,O) level of the 1A” state and cannot therefore cross into the 1A” state to radiate. This effect would not occur in the O+NO reaction, where the minimum of the excited state arising from normal reactants is believed to be above that for the radiating state.9 It is clearly not possible to distinguish between vibrational energy lost in the initial stabilizing process (3c) and that energy lost in subsequent collisions, but available data indicate that vibra- tional energy exchange occurs in most collisions for excited electronic states 14,15 and for levels near the dissociation limit.16 Such processes therefore have about ten times the collisional efficiency of the electronic quenching of NO2 and so vibrational quenching of HNO (3A”) is probably the dominant process in limiting the light emission in the combination reaction H+NO, even if more than one collision is normally needed to remove the critical amount of energy to prevent radiation.To summarize, in the combination reactions H + NO and 0 + NO stabilization by a third body to an electronically excited molecule makes a major contribution to the total reaction rate. In both cases these excited molecules undergo rapid radiationless transition into the state from which emission occurs and which does not148 CHEMILUMINESCENT REACTIONS dissociate directly into unexcited products. In the case of NO; the emission in- tensity is governed by competition between radiation and electronic quenching of this state, but for HNO* vibrational quenching of the non-radiating excited state (3A") to levels below the minimum of the radiating state (1A") is the dominant quenching mechanism. The authors thank Dr. D. A. Ramsay for sending them the manuscript of an unpublished paper and thank the D.S.I.R. for a maintenance grant to one of them (M. A. A. C.). 1 Clyne and Thrush, Trans. Firaday Soc., 1961,57, 1305. 2 Clement and Ramsay, Can. J. Physics, 1961, 39, 205. 3 Kaufman, Proc. Roy. SOC. A, 1958,247, 123. 4 Fontijn and Schiff, Symp. Upper and Lower Atmosphere Chemistry (Stanford Research In- 5 Bancroft, Hollas and Ramsay, J. Chem. Physics, 1962, 40, 322. 6 Dalby, Can. J. Physics, 1958, 36, 1336. 7 Walsh, J. Chem. SOC., 1953,2288. 8 Clyne and Thrush, Proc. Roy. SOC. A , 1962, 269,404. 9 Broida, Schiff and Sugden, Trans. Faraduy SOC., 1961,57,259. 10 Neuberger and Duncan, J. Chem. Physics, 1954,22, 1693. 11 Baxter, J. Amer. Chem. SOC., 1930, 52, 3920. 12 Keussler and Luttke, 2. Elektrochem., 1959, 63, 614. 13 Mason, J. Chem. Sac., 1957, 3904. 14 Rossler, 2. Physik, 1935, 96, 251. 15 Durand, J. Chem. Physics, 1940, 8, 46. 16 Pritchard, J. Physic. Chem., 1961, 65, 504. stitute, 1961).