J . Chem. Soc., Faradqy Trans. 1, 1982, 78, 3509-3518 Evidence for the Participation of an Isomerization Pathway in Diazirine Photolysis Study of Primary Processes and Energy Partitioning BY JUAN M. PEREZ Department of Physical Chemistry, Facultad de Ciencias, Universidad de Alicante, Apartado 99, Alicante, Spain Receiued 22nd February, 1982 CNDO/S calculations have allowed the band centred at 322.9 nm in the photolysis of diazirine to be considered as an A” c A’ transition. Correlation diagrams constructed using 3H-diazirine as a model and with the conservation of individual electronic angular momenta are used to explain some of the unusual features in the decomposition reactions of diazirines. Calculated energy-partitioning data indicate that there are two photodecomposition pathways with differing efficiencies in converting available energy into internal energy.Evidence shows that diazo-compounds are intermediates in diazirine photolysis and that the efficiency of this pathway accounts for 25”/, of the total reaction. Substituted diazirines are known to cleave, whether thermally or photochemically, into a carbene fragment and a nitrogen molecu1e.l However, the detailed mechanism of this fragmentation remains obscure. It has been shown experimentally that an excited 3H-diazirine molecule can isomerize into diazomethane or break up to give methylene and a nitrogen molecule. Two different reaction paths have been postulated. In the first, an excited 3H-diazirine molecule forms excited diazomethane directly2 which then loses its excitation energy via radiationless processes giving ground-state diazomethane, or breaks up into CH, and N,.In the second, the primary process is again the fragmentation of excited 3H-diazirine again into CH, and N,. In a subsequent step these species recombine and ground-state diazomethane is f ~ r m e d . ~ From ab initio SCF-CI calculations4 these sequencies both seem realistic. Recently, Frey and Penny5 have reinterpreted the photolysis of 3-chloro-3- methyldiazirine. In their opinion, both the fission and isomerization pathways of excited diazirine play a cruci,al role in the reaction mechanism. Moreover, Figuera et aL6 point out that in the same system at least two pathways to the formation of vinyl chloride are required to explain the experimental results, but they do not provide information on the nature of both pathways.Correlation diagrams constructed using 3H-diazirine as a model can be used to explain some of the unusual features displayed by these decomposition reactions. However, increased knowledge of the primary processes occurring in diazirines could be obtained by studying how energy is partitioned among the photolysis products along a series of compounds. This paper supplies theoretical calculations which in conjunction with experimental data available should provide a better insight into the decomposition of diazirine. 35093510 DIAZIRINE PHOTOLYSIS CALCULATIONS MOLECULAR-ORBITAL CALCULATIONS OF EXCITED STATES OF DIAZIRINES The CNDO/S method was used, and a configuration interaction over 30 excited states was carried out.The molecular geometry used was taken from the literat~re.~ The molecules were placed on the YZ plane and the Z axis was chosen as the C, axis. In this way, the molecular orbital perpendicular to the molecular plane become a” and b, in the point groups C, and C,,, respectively. The electronic transitions of low energy are shown in table 1. Table 2 includes the electronic populations of the atoms directly involved in the dissociation (C and N) with 3H-diazirine as a representative case, and table 3 gives the electron configuration of the low lying states of 3H-diazirine. TABLE 1 .-ELECTRONIC TRANSITIONS BETWEEN SINGLET STATES OF ALIPHATIC DIAZIRINES ~~ ~~~~ oscillator band maximuma strength orbital state molecule transitions transitions calc.expt. calc.b expt. 3H-diazirine a, + b, B, +- A , 326 322.9“ 0 - a, + a, A , t A , 231 - a2 t b, B,4- A , 226 320.6 0.04 - 3,3’-dimethyldiazirine a,+ b, B, t A , 327 358.1d 0 - a, 4- a, A , +- A , 212 - B , t A l 169 359.8 0.03 - a, + b, - 0 - 0 a In nm; evaluated from the integrated molar extinction coefficient of the band, see J. G. Calvert and J. N. Pitts, Photochemistry (John Wiley, New York, 1967), p. 172; see ref. 12(a); see ref. 12(b). TABLE 2.-ELECTRONIC POPULATIONS OF THE CARBON AND NITROGEN ATOMIC ORBITALS OF 3H-DIAZIRINE excited state ground atom/orbital state 1 st 2nd 3rd C S 1.011 1.01 1 P, 0.851 0.854 PY 1.130 0.671 pz 0.995 0.994 Na s 1.642 1.623 pz 1.172 0.944 Py 1.235 1.218 pz 0.995 1.488 1.010 0.825 1.129 0.994 1.574 1.159 0.835 1.494 1.010 0.850 1.131 0.918 1.642 1.174 1.235 1.100 a Both nitrogen atoms have the same populations because of symmetry.J. M.PEREZ 351 1 TABLE 3.-ELECTRON CONFIGURATIONS IN c2, SYMMETRY OF THE LOW-LYING STATES OF THE MOLECULES INVOLVED IN THE 3H-DIAZIRINE PHOTODECOMPOSITIONa 3 H-diazirineb methenec nitrogend a 1s electrons have been omitted; obtained from the CNDO/S results described above; S. Yan Chu, A. K. Q. Sin and E. F. Hayes, J. Am. Chem. SOC., 1972,94,2969; M. Orchin and H. H. Jaffe, The Importance of Antibonding Orbitals (Houghton Mifflin Co., Boston, 1967), p. 31. PHOTODECOMPOSITION CORRELATION DIAGRAMS Correlation diagrams were constructed for 3H-diazirine photodecomposition. First we correlated the molecular orbitals of 3H-diazirine with those of the same symmetry in methylene and nitrogen. Then we constructed the electron configurations of the low-lying states of the species involved and the related states of reactants and products whose molecular orbitals were intercorrelated.The ordering and symmetry of the 3H-diazirine molecular orbitals and the electronic composition of its excited states were obtained from CNDO/S calculations. Those of nitrogen and methylene were taken from the literature (see table 3). We assumed, for energetic reasons, that nitrogen is formed in its ground electronic state. It can be shown that this assumption does not exclude any significant correlations. Correlation diagrams were constructed assuming conservation of the C,, symmetry along the reaction path and also for paths of lower symmetry (see fig.1). The basic principle of individual electronic-orbital angular momenta conservation was described.8 PHOTOISOMERIZATION CORRELATION DIAGRAMS A correlation diagram was also constructed for the photoisomerization of 3H- diazirine and diazomethane, assuming conservation of CS(osy plane) symmetry along the reaction path, this plane being the only symmetry element (see fig. 2). The ordering and symmetry of diazomethane molecular and the electronic composition of its excited states were obtained from CNDO/S calculations. Table 4 shows the electronic configuration in C, (oZy plane) symmetry of the low lying states of diazomethane.3512 DIAZIRINE PHOTOLYSIS 7 2 --. x 2 6 'u 0 7 2 --- x P C e, 0 lene (3) FIG. 1 .-Correlation diagrams for 3H-diazirine decomposition constructed under the assumption that nitrogen is formed in its ground electronic state, maintaining: (a) C,, symmetry, (b) the plane out and (c) the plane oZu.cS&z plane) 3H -diazirine, d i azomethane I / 4 1 FIG. 2.-Correlation diagram for the 3H-diazirine isomerization to diazomethane, assuming conservation of the cue plane. TABLE 4.-ELECTRON CONFIGURATIONS IN c,, SYMMETRY OF THE LOW-LYING STATES OF DIAZOMETHANE~ a 1s electrons have been omitted; the molecular geometry was taken from M. Martin, V. Menendez and J. M. Figuera, Rev. Roum. Chim., 1976, 21, 31.J. M. PEREZ 3513 ENERGY-P ARTITIONING CALCULATIONS The study of the partition of energy amongst photolysis products along a series of compounds can be used to elucidate the connection between structure and energy partitioning.In order to calculate the percentage of energy removed by the organic species in photodissociation we must know the total energy available for distribution among the fragments and the part that is carried off by the organic fragment. The total energy available for partitioning will be the sum of the light absorbed by the reactant, its thermal energy and the exothermicity of the reaction at 0 K. The contribution of thermal energy was neglected. The light intensity absorbed was calculated following procedures already de~cribed.~ The exothermicity of the reaction was summarised as the difference between the standard heats of formation of reactants and products. These data were calculated according to the MIND0/3 method with the Rinaldi optimization procedure.Heats of formation, together with the most important characteristics of the respective geometries, are shown in table 5. TABLE 5.-GEOMETRIES AND STANDARD HEATS OF FORMATION IN SUBSTITUTES DIAZIRINES AND RELATED CARBENES AH? bond lengths/nm angles/rad /kJ mol-l C1-N C1-C2 Cl-HI ClNN HlClN C2ClN me th yldiazirine 0.145 0.150 0.110 t-butyldiazirine 0.145 0.157 0.110 3,3’-dimethyldia~irine~ 0.147 0.152 - Cl-C2 C1-HI 1.148 2.039 2.202 186.4 1.151 - 2.056 161.8 1.149 2.007 2.251 248.9 HlClC2 C2ClC2 methylcarbene 0.145 0.1 12 3,3’-dime thylcar benea 0.145 - t- bu t ylcarbene 0.150 0.124 1.915 - 322.8 - 2.190 249.8 1.954 - 360.7 “C,, symmetry in 3,3’-dimethyldiazirine and 3,3’-dimethylcarbene; C, symmetry in the others. The method of determining the distribution of energy is based on the dependence of the rate of unimolecular decomposition of the species formed upon their energy content.The calculation of the energy distribution functions of the excited olefins was carried out by fitting the experimental ratio of the amount decomposed to the amount stabilised (DIS), using the relationship between this quantity and the microscopic unimolecular rate constants, k(E), where ( k i ) is the average rate constant for the ith process. The rate constants k(E) were calculated using R.R.K.M. theory. All the internal vibrations of the molecule were taken to be active degrees of freedom. Activated complexes were chosen under the constraints that they should reproduce available Arrhenius pre-exponential factors, since R.R.K.M.results depend mostly on the A factor and not on the exact model of the activated complex.1o Vibrational assignments of molecules and activated complexes and collision parameters used are given in the Appendix.3514 DIAZIRINE PHOTOLYSIS The weighting function employed in this work is one which corresponds to the two-channel mechanism of olefin formation P(E) = PF(E) + (1 -P) F’(E’) (2) where p is the weighting factor, and F(E) and F ( e ’ ) are assumed to have normal gaussian functions. Calculations were carried out in the following systems: 3-methyldiazirine, 3,3’- dimethyldiazirine and t-butyldiazirine, for which suitable data are available in the 1iterature.ll The values ofweighting factors and gaussian widths employed in the energy distribution function, P(E), are similar for all cases studied.They are determined by a trial-and-error method, which consists of modifying the location of two gaussians, Emp and Ekp, until a reasonable reproduction of the experimental results is obtained. The calculated energy-partitioning data are shown in table 6 . The theoretical curves that best fit the experimental values of the rate constants are plotted in fig. 3. TABLE 6.-ENERGY PARTITIONING DATA (kJ m0l-l) E:,, pathb E& oc efficiency (Eolelin)d ~~~~ methyldiazirine 513 (a) 463 21 0.90 430 3,3’-dimethyldiazirine 523 (a) 417 21 0.80 42 1 (b) 350 50 0 . 6 8 (b) 334 50 0.64 (b) 409 50 0.64 t-but yldiazirine 636 (a) 500 21 0.79 537 a Energy available for partitioning between olefin and nitrogen; path (a), weight factor equal to 0.25; path ‘b) weight factor equal to 0.75; parameters of energy distribution function of excited olefins; olefin average energy relative to path (a) and calculated by eqn (3), see text.6 2 pressure/Torr FIG. 3.-Plot of the rate constants for ethylene (O), propylene (a) and 1,l’-dimethylcyclopropane (*) decomposition produced by photolysis of 3-methyldiazirine, 3,3’-dimethyldiazirine and t-butyldiazirine,” respectively. Experimental data from Frey et al.” Smooth lines, calculated values using distribution function made up of two gaussians with weights 1 :3.J. M. PEREZ 3515 DISCUSSION The low-resoIution electronic spectra of 3-methyldiazirine, 3,3’-dimethyldiazirine and t-butyldiazirine have been recorded by Frey et a1.l1 All spectra exhibit a similar shape.Robertson et ~ 1 . ~ have studied the high-resolution electronic spectra of 3H-diazirine and 3,3’-dimethyldiazirine ; they are very diffuse and rotational structure is not observed. The lowest n*-n+ transition allowed appears to be B, +- A, and is observed to occur at 322.9 nm. The agreement of CNDO/S calculations in band intensity and position with experimental and calculated values is about that expected from the method used. The lowest-energy transition increases the electronic population of the two p z nitrogen orbitals, while that corresponding to the py carbon orbital decreases. In addition to this band, Robertson et al. observed another transition 2.5 nm away from the former. They consider the second band to have arisen from the other non-bonding n- orbital.The above conclusion is based on the iissumption that both transitions originate from non-bonding orbitals and that the separation of these levels is small, owing to a symmetric and antisymmetric combination. However, Vasudevan et ~ 1 . ’ ~ found from SCF-CI calculations that there exists a large n+-n- separation of 288 kJ mol-l. A separation of 410 kJ mol-1 is also obtained from the results of Robin et al.,14 thus confirming that the second band observed in the neighbourhood of the n*-n+ transition has a different origin. Finally, Lombardi et ~ 1 . ’ ~ have pointed out the possibility that this band arises from a triplet-singlet transition, n*-n. The following two transitions calculated according to the CNDO/S method (see table 1) appear far apart from the second band, and it is impossible to attribute such a variation to the method of calculation employed.It seems reasonable to ascribe these transitions to the electronic system that appears below 200 nm. Starting with the electronic populations of the atoms directly involved in the dissociation (C and N) with 3H-diazirine as a representative case (see table 2), the interpretation of primary photochemical processes in terms of the bonding and antibonding character of the molecular orbitals involved is not without its difficulties. To investigate why these compounds photodecompose so effectively we constructed the correlation diagrams shown in fig. 1. These show photodecomposition from the first excited state to be forbidden. If photodecomposition from this state takes place retaining C, symmetry (the y z plane) throughout the reaction path, the process is endothermic and a potential-energy barrier arises between 193 and 338 kJ 11101-l.On the other hand, some diazirine substitutes are shown to fluoresce6*16 during the B, -+ A , transition. Thus it is evident that a fraction of the excited molecules undergo a process by which they decompose. However, Frey5 has reported that in 3-chloro- 3-methyldiazirine photolysis the quantum yield, aD, is close to unity (at 325 nm), and Over a considerable range of pressures CD, did not vary with pressure. To investigate the possibility of another photodecomposition pathway, we constructed the photoisomerization diagram shown in fig. 2. This diagram shows that photoiso- merization from the first excited state of 3H-diazirine is allowed when C, symmetry ( y z plane) is retained throughout the reaction path and that it yields diazomethane in the excited state A”(1).A similar conclusion is propounded by D e ~ a q u e t , ~ who indicated that if the excited system does not undergo intersystem crossing and remains on the ‘n-n* curve no dissociation will occur and the n,n* singlet state of diazomethane will then be populated. Frey et ~ 1 . ~ presented experimental evidence that in 3-chloro-3-methyldiazirine photolysis, dichloroethane results from the reaction of chloromethyldiazomethane3516 DIAZIRINE PHOTOLYSIS with HCI produced in the system. Also, direct formation of the diazocompound has been observed in the thermolysis of 3-butyl-3-pheny1dia~irine.l~ In order to gain information about the distribution between the two fragments of the excess energy due to the overall exothermicity of the process, we studied the energy partitioning between the photolysis products in a series of compounds, as this can be used to determine the connection between structure and energy partitioning. Thus we calculated the energy-distribution functions for those systems for which suitable data are available.The calculated energy-partitioning data are shown in table 6. Two main observations emerge from an inspection of this table. The first concerns the rather different efficiencies of the two pathways in converting available energy into internal energy. Path (a) is of greater efficiency than path (b).The second is connected with the relative importance of both pathways, and shows that path (a) is responsible for almost 25% of the total reaction. In the above discussion we have shown (by arguments based on the symmetry of the electronic states of species involved) that a continuous increase in the NNC angle on going from 3H-diazirine to diazomethane opens an allowed path for decomposition from the first singlet electronic state of diazirine compounds. This state is reached when the diazirine compound is irradiated in the long-wavelength band whose photon energy is close to 380 kJ mol-l. Thus, the average energy of the olefin formed by means of this pathway can be calculated by eqn (3) ( Eolefin) = [hv + AH*(diazirine - carbene)] x F+ AH*(carbene - olefin) (3) where F stands for the energy percentage carried out by the carbene group.For the photodecomposition of diazocompounds from the first singlet electronic state F is found to be ca. 65%.971a In addition, these systems show a very narrow energy distribution within the photolysis fragments and depend only slightly on the charac- teristics of the alkyl chain. If adequate data in table 5 are substituted in eqn (3) the values shown in the last column of table 6 are obtained. The agreement between calculated values and those obtained by application of the R.R.K.M. theory can be considered as satisfactory, bearing in mind the approximate nature of methods used. Thus there is a striking similarity between efficiencies for energy conversion in diazoalkane photodecomposition and those corresponding to path (a).This fact supports the above conclusion that the diazo compound is an intermediate product in diazirine photolysis. In addition, Frey et al. report that in 3-chloro-3-methyldiazirine photolysis the low-pressure results show the photoisomerization pathway to be over 27% of the primary process (this figure becomes 23% at very low pressures). The value found in this work is 25%, which can satisfactorily be compared with the experimental value reported by Frey. Path (b) represents a second means of photodecomposition of excited substituted diazirines. The results shown in table 6 show that there is a fairly wide spread of energies among the excited olefins. These data indicate that the primary process of photodissociation is not as simple as in the case in which the fragments fly apart with little interaction.The existence of intersystem crossing between the non-dissociative singlet state ln - z* and the dissociative triplet 3n - W* has been pr~pounded,~ but the participation of triplet states seems to be ~n1ikely.l~ In any case, this path must involve the intermediate formation of a carbene, followed by rapid intramolecular hydrogen transfer, yielding an olefin with sufficient internal energy to allow further decomposition.J. M. PEREZ 3517 This work was sponsored by Comision Asesora de Investigacion Cientifica y Tecnica, Spain. APPENDIX ETHENE The high-pressure Arrhenius parameters given by Benson and Hangerz0 were used [log(A/s-l) = 13.5 and E, = 354.5 kJ mol-'1.The vibrational frequencies (in cm-l) of the moleculez1 and activated complex are given below. ( S / D ) ratios are calculated using Z = 2.7 x lo7 Torr-l s-l. molecule complex 825 1443.5 780 730 943.2 2989.5 800 2989 949.2 3019.3 840 3019 995 3105.5 900 r.c. 1050 3272.3 620 3272 1342.4 1623.3 620 1623 PROPENE The high-pressure Arrhenius parameters given by Chappell and Shawz2 were used [log (Als-l) = 16.1 and E, = 357.8 kJ mol-'1. The frequencies used both for the molecule and activated complex are given below. ( S / D ) ratios are calculated using 2 = 3 x lo7 Torr-l s-l. complex (C-C rupture) 3090 1378 920 3272 1444 c.r. 3013 1443 912 3184 61 1 943 2954 1419 578 3019 1419 825 2954 1652 963 3002 1623 250 2992 1298 99 1 3002 1050 250 2933 1172 1045 2989 995 200 1474 428 225 1420 145 200 The activated complex frequencies for C-H scission were taken from ref.(23). 1, D DIME THY LCY c LO PRO PA NE The high-pressure Arrhenius parameters given by Flowers and FreyZ4 were used [log ( A / s - l ) = 14.8 and E, = 261 kJ mol-'1. The vibrational frequencies (in cm-l) of the mo1eculeZ5 and activated complex are given below. ( S / D ) ratios were calculated using 2 = 3.2 x lo7 Torr-' s-l. The two internal rotors corresponding to the rotations of methyl groups have been neglected because their reduced inertia moments cancel out in the expression for the rate constant. 1143518 DIAZIRINE PHOTOLYSIS molecule complex 3099 3080 3018 3018 2977 2976 2976 2975 2899 2899 1502 1470 1648 1466 1464 1418 1381 1378 1358 1281 1102 1070 1040 995 987 983 970 964 930 866 844 73 1 675 269 33 1 314 290 - - r.c.301 8 3018 3018 2970 2976 2976 2975 2899 2899 1502 1470 1468 1466 1464 1418 1381 1378 700 700 900 700 1040 990 980 983 970 964 930 866 400 73 1 350 369 33 1 314 290 - - 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 H. Durr, Top. Curr. Chem., 1975, 55, 87. H. M. Frey, Adv. Photochem., 1966, 4, 225. C. B. Moore and G. C. Pimentel, J. Chem. Phys., 1964,41, 3503. Bigot, R. Ponec, A. Sevin and A. Devaquet, J. Am. Chem. SOC., 1978, 100, 6575. H. M. Frey and D. E. Penny, J. Chem. SOC., Faraday Trans. I , 1977, 73, 2010. J. M. Figuera, J. M. Perez and A. Tobar, J. Chem. SOC., Faraday Trans I , 1978, 74, 809. L. Pierce and Sr. V. Dobyns. J. Am. Chem. SOC., 1962, 84, 2651. D. M. Silver, J . Am. Chem. SOC., 1974, 96, 5959. J. M. Figuera, J. M. Perez and A. P. Wolf, J . Chem. SOC., Faraday Trans. I, 1975, 71, 1905. T. B. Alfasi, S. W. Benson and D. M. Golden, J . Am. Chem. SOC., 1973, 95, 4784. H. M. Frey and I. R. D. Stevens, J. Chem. SOC., 1965, 1700; 1963, 3514; 1965, 3101. L. C. Robertson and J. M. Merritt, (a) J. Mol. Spectrosc., 1966, 19, 372; (6) J. Chem. Phys., 1972, 56, 2919. K. Vasudevan and W. E. Kammer, Chem. Phys., 1976, 15, 103. M. B. Robin, H. Basch, N. A. Kuebler, K. B. Wiberg and G. B. Ellison, J. Chem. Phys., 1969, 51, 45. J. R. Lombardi, W. Klemperer, M. B. Robin, H. Basch and N. A. Kuebler, J. Chem. Phys., 1969, 51, 33. P. H. Hepburn, J. M. Hollas and S. N. Thakur, J . Mol. Spectrosc., 1975, 54, 483. B. M. Jennings and M. T. H. Liu, J . Am. Chem. SOC., 1976, 98, 6416. J. M. Figuera, E. Fernandez and J. M. Avila, J. Phys. Chem., 1974, 78, 1348. J. M. Figuera and A. Tobar, J. Photochem., 1979, 10, 473. S. W. Benson and G. R. Hanger, J. Phys. Chem., 1967, 71, 1735. G. Herzberg, Infrared and Raman Spectra of Polyatomic Molecules (Van Nostrand, New York, 1946). G. A. Chappell and H. Shaw, J. Phys. Chem., 1968, 72,4672. J. W. Simons, B. S. Rabinovitch and F. H. Dorer, J. Phys. Chem., 1966, 70, 1076. M. C. Flowers and H. M. Frey, J . Chem. SOC., 1959, 3953. L. M. Sverdlov and E. P. Krainov, Opt. Spektrosk., 1959, 7 , 460. (PAPER 2/326)