NATURE. VOL. 218. JUNE 29. 1968

For MBS 35/40 diamond a t least, there seems to be a correlation between the nitrogen impurity content and the crystal habit. The correlation suggests that the presence of nitrogen probably encourages the growth of cube faces.

One of us (J. J. C.) thanksMessrs Industrial Distributors (1946), Johannesburg, for financial support.

B. R. ANGEL Department of Physics, Plymouth College of Technology.

School of Physics, Universit,~ of Warwick, Coventry.

Department of Physics, University of Lovanium, Kinshasa, Congo.

Received May 7, 1968.

' Kaiser, W., and Bond, My. L., Phys. Rev.,ll5,857 (1959). Smith, W. V., Sorokin, P. P., Gelles, I. L., and Lasher, G . J., Phus. Rev.,ll5,

1546 (1959). a Evans, T., and Phaal, C., Proc. Roy. Soc., A, 270, 538 (1962). a Dyer, H. B., Raal, F. A., Du Preez, L., and Loubser, J. H. N., Phil. Mag.,

11. 763 (1965). " Smith, 51. J. A,, and Angel, B. R . , Phil. Mag., 15, 783 (1967).

Trapping of Free Radicals formed by ?-Irradiation of Organic Compounds NITROSO compounds (I) possess scavenger properties for free radicals, and react to form unsymmetric nitroxide radlcals (11) stable enough to be detected by electron spin resonance (ESR) spectroscopy (refs. 1-3, and unpublished lvork of Forshult, Lagercrantz and Torssell).

(ref. 3 and unpublished work of Forshult, Lagercrantz and Torssell). In these experiments 2-methyl-2-nitroso- butanone-3 or t-nitrosobutane was used as the nitroso scavenger (I). As well as the main triplet splitting caused by the 14N nucleus of the nitroxide group, the ESR spectra showed secondary splittings which originated exclusively from magnetic nuclei of the part of the nitroxide radical derived from the radical R.. From these splittings it was in many cases possible to settle the structure of the trapped radicals.

This report describes the trapping (as nitroxide radicals) of the radicals formed by y-irradiation of some organic compounds. Numerous studies435 have been carried out using ESR on the free radicals produced by high-energy irradiation of organic substances in the solid state. I t is well known that these radicals disappear immediately on dissolution of the irradiated samples, in a series of more or less complicated reactions.

The substances were irradiated at room temperature in the polycrystalline state by y-rays from a B°Co source, and were given about 5 Mrad each. The samples were then dissolved in an aqueous solution of the nitroso scavenger t-nitrosobutane (0.01 M), and the ESR spectra recorded at room temperature. For some of the substances, the nitroso scavenger was dissolved in 0.25 N NaOH to a.void too high an acidity of the resulting reaction mixture, which hampered the formation of the nitroxide radicals.

In this way, y-induced radicals were trapped froin sodium acetate, glycine, DL-a-alanine, p-alanine, dipotas- sium malonate, succinic acid and L-glutamic acid hydro- chloride. The coupling constants of the nitroxide radicals and the structure of the trapped radicals, as derived from the secondary splittings, are collected in Table 1, together with the structure of the radicals known to be present in solid state samples after y-irradiation at room tempera- t ~ r e ~ - ~ . The spectra obtained with glycine and DL-cr- alanine are shown in Figs. 1 and 2. No radica.1~ were

R'NO + R.+-R f--N-R obtained with non-irradiated samples.

I The results indicate that radicals trapped by the

0. ( l ) nitroxide method have a structure very similar to that of

I I1 the corresponding species present in the y-irradiated solid samples. We therefore conclude that apart from a pos-

\Ye ha.ve a(ppliec1 t,his technique to the trapping of sible difference in added protons, the trapped radicals are short -li\-ed free ra,dicals produced in a rariet,y of react,ions the primary species of the solid samples. Obviously,

the reaction leading to nitroxide radicals, equation (I) , competes effectively with

I R other reactions in which the radicals take part and then disappear on dissolution of the crvstalline matrix.

Of <he two radical species, CH,COOH (or CH,COO-) and H,N+-GH-COO-, known to be formed in solid glycine on y -irradiation a t room temperature5, only the former could be trapped as a nitroxide

Fig. I. ESR spcctrum of the nitroxide radicals formed on dissolving a y-irradiated solid sample of ylycine in a water solution of t-nitrosobutane. The secondary triplet splitting

(1 : 2 : 1) indicates trapping of the radical CH,COO(H).

radical. A transient species present ini- mediatelv after dissolving the irradi~.t~ed .. . . .. sample Gf glycine could not be identified. No nitroxide radicals have been detected so far with y-irradiated samples of

'r;lhle 1. COUPLING CONSTANTS OF XITROXIDE RADICALS AND STRUCTURB OF TRAPPED R.4TlICALS DERIVED FROM SECONDARY SPLITTIXG4

Nitroxide radical, t-Ru-N-R, formed upon dissolving the irradiated sanlples

Substance 7-irradiated in

solid st,ate

I 0 . Radicals present in solid state sanlples after

in an aqueous solution of t-BuNO y-irradiation at room temperature a:% a:' (arcording to refs.)

(gauss) Trapped radical R.

Na-acetate 16.1 Qlycine 16.1 DL-a-Alanine 16.2 p-Alanine 15.6 Dipotassium malonate 15.5 Succinic acid 16.6 L-Olutamic acid 15.7

hydrocNoride

8.42 (t) 8.42 (t) 5.2 (d) 5.1 (d) 4-36 (d) 4.16 (d) 4.37 (d)

(1 =Do~~blr t , t = triplet, q = rlnartet.

6~ , -COO(H) 6H,-COO(H) ~ H ~ - ~ O O ( H ) , H,N+-(:H-COO- (ref. 5 )

0.4 (q) CH,-cH-C.OO(H) CH,-CH-COO(H) (ref. 6) 0.7 (t) (H~N)cH,~CH-COO(H)

(COOH)-CH-COO(H) (COOH)--6~-COO(H) (ref. 7 ) 0.6 (t) (COOH)CH~-~H-CO$(H) (C'OOH)CII,-CH-COO(H) ( r ~ f . 8) 0.5 (t) (COOH)CH(H2N)CH,-CH-COO(H) (COOH)CH(H,N+)CH,-CH-W(H) (ref. 9)

or (COOH)CH,-CHsCH-COO(H)

NATURE. VOL. 218. JUNE 29. 1968

Fig. 2. ESR spectrum of the nitroxide radicals formed on dissolving a y-irradiated solid sample of DL-a-alanine in a water solution of t-nitroso- butane. The secondary splitting is a doublet further split into a quartet (1 : 3 : 3 : 1) .indicating trapping of the radical

CH,-CH-COO(H).

cysteine, cystine and methionine, which very probably contain radicals with the odd electron localized chiefly on the sulphur atom. These findings seem to conform with the fact that in our experience the most stable nitroxide radicals are those formed by the trapping of radicals in which the odd electron is localized on a carbon atom, and which have structures such as CH,, R-CH,, R , ~ H and R,C (R=alkyl, aryl or COOH).

We thank Dr K. Torssell for supplying t-nitrosobutane, Professor E. von Sydow for providing y-irradiation facilities and Mr G. Gater for helpful assistance. This work was in part supported by grants from the Swedish Natural Science Research Council and Wilhelm och Martina Lundgrens Vetenskapsfond.

CARL LAOERCRANTZ STIG FORSHULT

Department of Medical Physics, University of Goteborg, Goteborg, Sweden.

Received April 11,1968.

' Mackor, A., Wajer, Th. A. J. W., de Boer, Th. J., and van Voorst, J. D. W., Tetrahedron Lett., 2115 (1986); ib~d., 585 (1967).

' Mackor, A. , Wajer, Th. A. J. W., and de Boer, Th. J., Tetrahedron, 24,1623 (1968).

* Lagercrantz, C., and Torssell, K., Acta Chem. S c a d . (in the press). Morton, J. R., Chem. Rev., 64,453 (1964). Morton, J. R., J. Amer. Chem. Soc., 86,2365 (1964). Miyagawa, I., and Itoh, K., J. Chem. Phys., 36,2157 (1962). ' Cole, T., and Heller, C., J. Chem. Phys., 34, 1085 (1961).

Pooley, D., and Whiffen, D. H., Mol. Phys., 4.81 (1961). Lin, W. C., YcDowell, G. A., and Rowlands, J. R., J. Chem. Phys., 35, 757

(1961).

Photodecomposition of Trimethylsilyl Azide in Solid Argon WHEN some group IV azide derivatives have been sub- jected to photodecomposition in argon matrices at cryo- genic temperatures, isocyanides have been produced. In the case of methyl azide H,CNNN, hydrogen isocyanide HN=C was proved to result from a secondary photolysis of methylenimine H,C=NH (ref. 1). From silyl azide, H,SiNNN, iminosilicon HNSi has been detectedz, but the mechanism of formation was uncertain; the vibrational absorptions of other products were not then assigned. From germyl azide H3GeNNN, HNGe has been tentatively identified (unpublished work with 8. Cradock). The force constants of the heavy-atom valence-stretching vibrations of HNSi and HNGe are probably appropriate to what might effectively constitute "double bonds" rather than to a "triple bond" as in the case of HNC. Because x-type bonds of atoms beyond the first row are decreasingly stable because of the reduced overlap of the p-orbitals of the larger atoms, it is interesting to know the results of a situation in which there might be unsaturation

either in a chemical bond between only first row atoms or alternatively, in a bond involving a second row atom. Such a case has been discovered following the photolysis of trimethylsilyl azide isolated in solid argon a t 17" K.

Deposition of less than 30 pmoles of (H,C),SiNNN in argon a t matrix ratios of 195 or >300 during 6 h was followed by spectroscopic scanning from 370 to 4,200 cm-l using a modified 'Unicam' SPlOO spectrophotometer3. The temperature of 17" K was maintained by an Air Products Cryotip refrigerator4, model AG-3L-110, by pumping on hydrogen gas from the second Joule-Thoison expansion nozzle with a mechanical pump (450 1. min-I). After the recording of the vibrational absorption, irradia- tion was effected by focusing the emission from an HPK 125 W (Philips) mercury lamp onto the sample. No t,emperature increase ( < 0.2") was noticed during either photolysis or scanning operations. The repeated scanning of the irradiated sample indicated the extent of de- composition of the (H,C),SiNNN and the formation of products. The figures illustrate the spectra recorded before and after 1.5 h of photolysis. The bands due to hexamethyldisiloxane impurity ( - 2 per cent of the sample before deposition) are unchanged, but bands of HNNN are greatly reduced although remaining extremely narrow. Bands tentatively assigned (vide infra) to a specified product are labelled A in Fig. 1.

The complicated overlapped nature of the spectrum after photolysis almost precludes complete analysis, even if many isotopically substituted trimethylsilyl azides had been available to provide further data. A tentative assign- ment is therefore now being attempted. It is pre- sumed that photolysis leads in the primary process to formation of molecular nitrogen which plays no furt,her part in the chemical processes. Neither molecular nitrogen nor hydrogen could be detected spectrometrically in these experiments. but ethane, analogous to hydrogen from the secondary photolysis of H,C=NH, could have been and was not. The intense absorption at 1,671 cm-l, the most notable spectral feature appearing after photolysis, pro- vides a valuable clue to the nature of the products. The following bands have similar intensity behaviour relative to 1,671 om-l in different experiments, and might be assigned to a species (H,C),SiH-N = CH, as follows : 2,815 and 2,860 cm-1 to the methylene valence-stretching vibrations, 1,489 and 916 cm-I to CH, deformation and wag, 2,087 and 794 cm-I to the SiH stretching and bend- ing vibrations, 613 cm-l to the Si-N stretching modo, and 1,671 cm-I to the CN double bond stretching vibra- tion. The latter frequency is most appropriate to a double bond of first row atoms. The comparatively small frequency and intensity of the 2,087 om-' feature are unusual for a SiH mode, but no overtone nor combination can be found to describe this value. The structure appear- ing in the 2,150 cm-l band after photolysis in Fig. 2 is absent before photolysis only because of complete absorp- tion in this particular experiment. Other characteristic absorptions of the Si(CH,), group would probably be of very similar frequency to those of the Si(CH,), group of the parent substance; other observed absorptions can a t present not be assigned.

Milligan suggested that in his experiments methyl- enimine was formed by the rearrangement of a "hot" methylimino radical H,C-N (ref. 5). The 1,671 cm-I fre- quency is not reasonable for any vibrations of t,he trimethyl- silylimino radical, and would be anomalously large for the SiN stretching vibration of a species (H,C),Si=NCH,, particularly in view of the fact that even in iminosilicon, in which the opportunity for a "triple bond" exists, the NSi vibration frequency is only 1,198 cm-1 (ref. 2 ) . The proposed N-(dimethylsily1)methylenimine product might therefore be envisaged to result from what would effectively be a methyl migration and a hydrogen atom back-migra- tion, all presumably occurring in the interval between the absorption by a (H,C),SiNNN molecule of an energetic