On the Photolysis of Acylphosphine Oxides: 1. Laser Flash Photolysis Studies with 2,4,6-Trimethylbenzoyldiphenylphosphine oxide

On the photolysis of acylphosphine oxides: 1. laser flash photolysis studies with 2,4,6trimethylbenzoyldiphenylphosphine oxide T. S u m i y o s h i * and W . S c h n a b e l Hahn-Meitner-lnstitut fi~r Kemforschung Berlin, Bereich Strahlenchemie, Postbox, D- 1000 Berlin 39, Federal Republic of Germany and A. H e n n e and P. L e c h t k e n BASF Aktiengesellschaft, D-6700 Ludwigshafen, Federal Republic of Germany (Received 27 February 1984) 2,4,6-trimethylbenzoyldiphenylphosphine oxide (TMDPO) was irradiated in dilute solutions of benzene, methanol and dichloromethane with 20 ns flashes of 347 nm light. By optical absorption measurements, a transient spectrum with a maximum at about 330 nm detected at the end of the flash was ascribed to the diphenyl phosphonyl radical (Ph)2P=0). This radical is formed by the reaction. Evidence was obtained for fragment radicals being formed, to some extent at least, from triplet states. Both the singlet and the triplet lifetimes were estimated as T< 1 ns. The singlet energy was estimated as 288 kJ mo1-1. In the-absence of radical scavengers (Ph)2.P=0 radicals undergo a self reaction (2k2,,~5x109 I mol -~ s - l ) . Rather high rate constants 7(in I mo1-1 s -1) were found for reactions of (Ph)2.P=0 radicals with oletinic compounds: 6.0×10 (styrene), 6.0×107 (methyl methacrylate), 2.3 × 107 (methacrylate), 1.8 x 107 (acrylonitrile), 5.0 x 106 (t-butyl vinyl ether), 2.0 × 106 (vinyl acetate). The quantum yield of radical formation is ~(-TMDPO)=0.5 to 0.7. These results clearly show why TMDPO is highly appropriate to be applied as an initiator for photocuring of coatings of various compositions. (Keywords: photoinitiators; acylphosphine oxides; diphenylphosphonyl radicals; absorption spectra; reactivity towards olefinic compounds) INTRODUCTION Recently, acylphosphine oxides of the general structure Rl__C__p/R2 II II\R 3 O O have been developed as a new class of u.v. photoinitiators1. Compounds with R t = 2,4,6-trisubstituted benzoyl and R 2 and R3=phenyl, such as 2,4,6-trimethylbenzoyl diphenyl phosphine oxide (TMDPO) * On leavefromHokkaido University,Facultyof Engineering,Sapporo, Japan 0032-3861/85/010141-06503.00 © 1985 Butterworth & Co. (Publishers) Ltd. were shown to be particularly suitable initiators for photocuring compositions based on acrylates of styrenecontaining UP resins. T M D P O , as well as other acylphosphine oxides, exhibits a characteristic absorption in the near u.v. range with a maximum at 380 nm 2'3" as can be seen from Figure 1. This absorption at relatively long wavelengths makes T M D P O particularly suitable for the initiation of photopolymerization in TiO2-pigmented coatings 2 and of thick-walled glass fibre-reinforced polyesters 3b. In toluene or methanol solution, the quantum yield of T M D P O decomposition is 0.6-0.7. Photolysis in methanol/HzO 2 gave strong evidence for ct-scission to occur as is depicted in Scheme 13" . In the presence of hydrogen peroxide the fragment radicals were converted to diphenylphosphinic acid (IV) and 2,4,6-trimethylbenzoic acid (V). The present work is aimed at elucidating kinetics and mechanism of the photolysis of T M D P O in the presence of unsaturated compounds capable of undergoing radical polymerization. It was intended to measure the lifetime of the excited species giving rise to ~.-scission and, in POLYMER, 1 985, Vol 26, J a n u a r y 141 Photolysis of acylphosphine oxides: T. Sumiyoshi et al. 1.0 -5 r\ 2,4,6-trimethylbenzoylphosphonic (TMPDM) were synthesized. Xexc= 320 nm - / " 0'2 t \ 0.1 -3 - 2 .d \\ 01-.--~ 300 250 ~00 300 ---U-----t-----a-- 0 500 600 k(nm) 350 Z,00 /,50 X (nm) Figure 1 Ground state absorption spectrum of TMDPO (1.0x10 -4 mol 1-1 ) in solution of CH2CI2 recorded at 23°C. Inset: Ground state absorption spectrum ( - - ) , excitation ('0'0'0") and emission ( ..... ) spectrum in benzene solution of TMDPO (4.2× 10 4 mol I 1). The excitation spectrum was recorded at 2=440 nm. The ground state absorption spectra in benzene and CH2CI2 above 300 nm were identical 11 IV I _ co. I11 V Scheme 1 Photolysis of TMDPO in methanolic solution containing hydrogen peroxide addition, to evaluate rate constants of the reaction of excited initiator molecules and/or of fragment radicals with various monomers. EXPERIMENTAL Materials 2,4,6-trimethylbenzoyldiphenylphosphine oxide (TMDPO) was recrystallized from ether solution before use. Its synthesis was described previously 1'2. In a similar manner, pivaloyldiphenylphosphine oxide (PDPO) and 142 P O L Y M E R , 1985, Vol 26, J a n u a r y H3C--C --C --P" H3C O \ ci 0.5 0 I (PDPO) acid dimethylester / • /OCH3 C--P - - - ~ - ) 1 )--- II 'IX'OCH3 --~ o o (TMPDM) Benzene (E. Merck, p.a. 99.5%) was washed with concentrated H2SO4, dried over CaCI 2 and sodium, and subsequently distilled over a fractionation column (Fischer, Bonn). Methanol (E. Merck, Uvasol Ira) and dichloromethane (E. Merck, p.a.) were used as received. The following monomers were used: methyl methacrylate (MMA), methyl acrylate (MA), styrene (St), acrylonitrile (AN), vinylacetate (VAc) and t-butyl vinyl ether (BVE). They were freed from stabilizers as has been described before4. Irradiations Dilute solutions of the initiator were irradiated in rectangular quartz cells, which were freed from oxygen by bubbling with purified argon. Irradiation was carried out at 347 nm with the aid of a ruby laser in conjunction with one frequency doubler. The duration of the flash was about 20 ns (halfwidth). Formation and decay of transients were followed by optical absorption and emission measurements. In addition, photocurrent measurements were conducted in order to test whether ionic intermediates were formed during or after the flash. The sensitivity of the set-up was checked with Ar-saturated solutions of anthracene (3 x 10-5 M) in methanol, where ions are readily formed upon irradiation with light of 2i,c = 347 nm 5. As can be seen from Figure 1, the absorption spectrum of T M D P O possesses two absorption bands with maxima at about 380 and 295 nm, which probably correspond to So---~S1 and So--*$2 transitions. Accordingly, irradiation of T M D P O with light of 2~.c= 347 nm produces mainly vibronically excited states of the first excited singlet state. Fluorescence spectra were recorded under continuous irradiation with the aid of a spectrofluorimeter (MPF-4, Perkin-Elmer). Fluorescence lifetimes were determined with a single photon counting apparatus (199M, Edinburgh Instruments). Actinometry In the flash photolysis experiments at 2i,c = 347 nm, the absorbed dose per flash was determined as follows: the optical absorption of naphthalene triplets formed by energy t ransfe r from benzophenone (1.15 x 10 - 3 mol 1- 1) at [ N a p h ] = l . 0 3 x l 0 : l m o l 1 - 1 , was measured at 2=425 nm (eT_V=2.3 X 104 1 mo1-1 cm -1 6, ~b(T)= 1.0). The maximum laser output at 2 = 3 4 7 was 5.7 x 1016 photons per flash corresponding to an absorbed dose per flash, at O.D.=0.1, of Dab,= 3.6 x 10 -5 einstein 1-1 RESULTS Emission measurements with TMDPO The emission spectrum shown in the inset of Figure 1 was recorded upon continuous irradiation of a benzene Photolysis of acylphosphine oxides. L Sumiyoshi et c 03 0.3 0.2 ~. 0.1 0.10 _ /o . /o f ° /o -io 0 °1 ° , , , I 2 3 Oabs(lO -sE/l) d 0.05 - / / o 0 °\ 0 I 300 350 400 )k(nm) F i g u r e 2 Transient absorption spectrum recorded immediately at the end of the laser flash with an Ar-saturated solution of TMDPO in CH2CI 2 (4.2x10 4 mol i-1). Dabs=9.4xl0-6 Einstein 1-1 . Inset: the optical density measured at the end of the flash as a function of the absorbed dose per flash solution of T M D P O at 2,xc=320nm. It possesses a maximum at about 440 nm. A similar emission spectrum was obtained with a CHzCI 2 solution. The inset of Figure 1 also shows the excitation spectrum recorded at 2 = 4 4 0 n m . It is seen that excitation and absorption spectra are identical. Therefore, it is concluded that the emission band corresponds to the ground state absorption band due to the So---*S~ transition (2m~x=380nm). From the intersection of the two spectra at 415 nm the singlet energy (o-o transition) of T M D P O was estimated as ca. 290 kJ m o l Upon flash irradiation at 2~xc= 347 nm a quite similar emission spectrum was observed. The emission formed and decayed during the flash (halfwidth ca. 20 ns). In order to search for a long-lived emission, a photomultiplier system (RCA, type 1P28) operated initially with two-dynodes and, after a delay time of 400 ns, with 9 dynodes, was employed. However, a long-lived emission was not detected. The single photon counting technique also did not reveal a long-lived emitting transient. The kinetic analysis of the decay of the luminescence showed that about 99% of the emission decayed exponentially with a lifetime of about 0.3 ns. These results show that excited singlet states of T M D P O are very short-lived. Absorption measurements with T M D P O Upon irradiation of T M D P O in dilute solution with 20ns flashes of 2i,c=347 nm a transient difference spectrum in the wavelength range between 280 and 380nm was observed. The difference spectrum was built up during the flash. In the absence of scavengers it decayed with 1st halflives between 5 and 20/~s depending on the absorbed dose per flash. A typical difference spectrum (2m~x= 331 nm) recorded with a C H 2 C I 2 solution is shown in Figure 2. Similar transient differential spectra were observed with solutions of T M D P O in benzene (2m~,= 336 nm) and methanol (2m~x= 328 nm). From the al. inset in Figure 2 it is seen that the transient optical density measured at the end of the flash increased linearly with increasing absorbed dose. In Figure 3 the ground state absorption spectra of unirradiated and irradiated TMD P O are compared. Obviously the photolysis of TMD P O leads to products with absorption spectra different from that of T M D P O . The isosbestic point at about 330 nm is identical to the maximum of the transient difference spectrum (see Figure 2). Therefore, this wavelength was considered quite appropriate for kinetic measurements (vide infra). At wavelengths of maximum difference, on the other hand, e.g. at 295 nm and 380 nm, the quantum yield of the conversion of T M D P O could be estimated. The value obtained, ~b(-TMDPO)~0.5-0.6 compares satisfactorily with the value obtained at a much lower absorbed dose rate in solution of toluene and methanol, namely ~b(-TMDPO) ~ 0.6-0.73". With respect to the assignment of the transient absorption formed during the flash, the question arose whether bond cleavage according to Scheme 1 occurred directly from excited singlet states or whether triplet states formed by intersystem crossing were involved. Evidence for the formation of triplets was obtained from experiments with benzene solutions of T M D P O containing naphthalene. At naphthalene concentrations greater than 0.5 mol l- 1 the absorption spectrum of naphthalene triplets (2max=425 nm) was observed. Typical data are shown in Table I. Concurrently with the formation of the absorption at 425 nm the maximum at 336 nm decreased. It appears that naphthalene triplets were formed by energy transfer from T M D P O triplets to naphthalene in competition with fragmentation into radicals according to reaction (2) kQ 3TMDPO* + N a p h ~ I T M D P O + 3Naph (1) 1°f/A I Flash --'-I c5 0.5 d I~10/~s 5 mV 1 Flash f 0 E~ I i 250 i i '%. i I 300 • 350 ~00 , I 450 X (nm} Figure 3 Ground state absorption spectra of TMDPO in Arsaturated CH2Cl 2 solution (A) before and (B) after irradiation with a laser flash of 347nm light (Dabs 5.2x 10 5 Einstein I-1). Optical density at d - 1 . 0 cm vs. wavelength. Inset: oscilloscope traces depicting the change of the O.D. at 2=335nm, U o 104mV and ).-380 nm, Uo-152mV P O L Y M E R , 1985, Vol 26, J a n u a r y 143 Photolysis of acylphosphine oxides: T. Sumiyoshi et al. Table 1 Transient absorption (at h = 425 nm) observed at the end of the flash due to naphthalene triplets. TMDPO (4.2 x 10 -4 tool 1-1) in Ar-sat. benzene solution, Dabs: 1.9 x 10 -5 einstein 1-1 [Naph] (tool 1-1) O.D.(h = 425 nm) 0.5 1.0 2.0 0.006 0.01 0.02 3TMDPO* ~ R- + R' (2) By considering the two reactions, the O.D. of naphthalene triplets observed at the end of the flash should depend on the naphthalene concentration according to equation (3) (O.D.)total (O.D.) kT = 1+ - - ko.[Q ] (3) where (O.D.)total is the optical density that would be observed if all T M D P O triplets were quenched by naphthalene. With the data of Table 1 the lifetime of T M D P O triplets is estimated as 0.3 ns by assuming ko= 1 x 109 1 mo1-1 s -1, i.e. the lifetime is much shorter than the duration of the flash (ca. 20ns). This result implies that the absorption band at 336 nm observed at the end of the flash is to be attributed to fragment radicals. The triplet energy of T M D P O must be greater than that of naphthalene, i.e. ET(TMDPO) > 251 kJ mol- 1. In this connection it has to be considered (a) that naphthalene does not absorb light at 2 = 347 nm, and (b) that singlet quenching can be excluded for energetical reasons (Es(Naph) = 385 kJ mol- 1 > Es(TMDPO) 290 kJ mol- 1). Because it was not possible for practical reasons to employ naphthalene concentrations high enough to suppress the formation of the transient absorption at 336nm completely it cannot be decided whether bond breakage according to Scheme I involves both singlet- and triplet-excited or just triplet excited states of TMDPO. However, it is obvious that triplet states are involved in bond breakage. The assignment of the absorption band at 336 nm observed with T M D P O solutions at the end of the flash to fragment radicals was affirmed by the following findings: The rate of decay of the transient O.D. is significantly increased by 02 and unsaturated compounds capable of reacting with free radicals (vide infra). In the absence of scavengers, the decrease of the transient O.D. follows 2nd order kinetics as can be seen from Figure 4, where the reciprocal halflife is plotted vs. the initial O.D. (measured at the end of the flash). A linear relationship between these two parameters is expected on the basis of equation (4) transient differential absorption spectrum to ionic intermediates can be discarded because experiments with methanolic solutions of T M D P O showed that neither during nor after the flash a photocurrent was formed. Reactions of radicals produced from TMDPO with oxygen and unsaturated compounds As has been mentioned above, the transient absorption decayed very rapidly in the presence of 0 2. By recording the decrease of the absorption in benzenic solution of TMDPO, at 2obs= 335 nm, at two oxygen concentrations (1.9 × 10 -3 and 9.5 x 10 -3 mol 1-1) a bimolecular rate constant kR +o, = 2.5 × 109 1 mol- 1 s- 1 was determined. The reactivity of initiator radicals towards various unsaturated compounds was investigated by measuring the rate of the decrease of the O.D. at 2obs= 335 nm with Ar-saturated benzene solutions containing 4.2 × 10-4 mol 1-1 T M D P O and, in addition, a monomer at a concentration greater than 10-2 tool 1-1. Under this condition pseudo-lst order kinetics prevailed: d[R.] dt = kR.+M[M][R'] = kl [R'] with (6) k1 kl = k, +M[M] and (z,/2)- 1 = ln~ In Figure 5 the reciprocal halflife is plotted vs. the monomer concentration. Bimolecular rate constants obtained from the slopes of the straight lines are compiled in Table 2. For comparison, Table 2 also contains values of rate c o n s t a n t s kR.+U which were determined in former work with 1-phenyl-2-hydroxy-2-methyl-propanone-1 (PHMP) 7 and benzoinmethylether (BME) s. These compounds also undergo c~-scission upon irradiation with u.v.-light: 20~ I0~ I I Y (~1/2)- 1 = 2kR[R'] o (4) I I which is derived from equation (5) I o¢ 1 1 JR'], [R']o + 2kRt (5) where [R.]o(oc(O.D.)o) is the initial concentration of radicals and kR is the bimolecular rate constant for the reaction R- + R-. The possibility of an assignment of the 144 POLYMER, 1 9 8 5 , V o l 26, J a n u a r y 0 , m 2 m ! , /, l 6 I I 8 (O.D.)o-1 Figure 4 Plots of first 2nd order halflives vs. the reciprocal initial optical density, measured with Ar-saturated solutions at the end of the flash at different absorbed doses per flash at ).obs=335nm in benzene at 23°C. [ T M D P O ] = 4 . 2 × 1 0 4 tool 1-1 Photolysis of acylphosphine oxides." T. Sumiyoshi et al. CH 3 - CH 3 --CH 3 - ) • + o OH .C--CH I O 3 (7) OH H H O O O I O I CH3 (8) CH 3 The rate constants compiled in Table 2 are, in principle, composite rate constants comprising the reactions of both fragment radicals with the respective monomer. A comparison of the values reveals that the radicals produced by the photolysis of T M D P O are much more reactive towards monomers than the radicals generated from the other two initiators. If it is taken into account, moreover, that P H M P and BME form benzoyl-type radicals as well as does T M D P O , it becomes obvious that the very high reactivity of radicals stemming from T M D P O must be due to the reactivity of radicals of the type (Pb)2P = O (Ph = phenyl), and that these radicals must be much more reactive than the partner radicals, i.e. 2,4,6-trimethylbenzoyl radicals, that are generated simultaneously. Following this line, it turns out that the transient absorption observed in the flash photolysis of T M D P O is essentially due to the absorption of radicals of the type (Ph)2.P = O. This point will be dealt with in more detail in the next section. Experiments with PDPO and TMPDM These two compounds were studied in order to substantiate the assumption that the transient absorption spectrum observed with T M D P O was essentially due to the absorption of phosphonyl radicals (Ph)z.P = O. As a matter of fact, P D P O yielded almost the same transient absorption spectrum with 2max=325 nm as T M D P O upon irradiation in Ar-saturated methanolic solution. In this case, ~-scission causes the generation of the radical CH3 \ HaC--C C-(besides the phosphonyl radical) / A I o % O' I I I I 0.1 0.2 0.3 0.Z, [M] ( t o o l / I ) Figure 5 Reciprocal p s e u d o - l s t order halflives as a function of the monomer concentration [ M ] . The measurements were carried out in Ar-saturated benzene solutions latl 2°bs= 335 nm at an absorbed dose of 2.2 ×10 5 Einstein [TMDPO]=4.2×10 -4 mol I 1. Curve ( A ) = S t M M A ; curve ( B ) = M A ; curve ( C ) = A N ; curve (D) = BVE Table 2 JI CHa 0 which possesses a much lower absorptivity in the wavelength range between 300 and 350nm than the benzoyl radical, which is produced in the photolysis of TMDPO. Because the same transient absorption spectrum (of comparable intensity) was generated both with PDPO and TMDPO, it is concluded that this spectrum can be attributed to the phosphonyl radical (Ph)2P = O. A quite different situation was encountered with TMPDM. After the rapid decay of the absorption of a shortlived species (z~/2: 30-40 ns) a long-lived transient absorption spectrum with a maximum at a significantly longer wavelength, i.e. at 380 nm, was detected upon irradiation of T M P D M in solution of methanol, dichloromethane and benzene. It decayed with a lifetime much longer than that measured in the cases of T M D P O and PDPO. This finding indicates that, provided dimmethoxy phosphonyl radicals, (H3CO)2p = O, were formed, their absorptivity is much lower than that of (Ph)2P=O radicals. The possibility of assigning this Bimolecular rate constants (in i tool - t s-1 ) of the reaction of initiator radicals w i t h various monomers Monomer TMDPO PHMP* BME** H (( Styrene Methyl methacrylate Methylacrylate Acrylonitrile t-Butylvinyl ether Vinyl acetate * ** (6.0 (6.0 (2.3 (1.8 (5.0 (2.0 + 0.3) + 0.3) + 0.2) -4--0.2) _+0.3) ± 0.2) 10 "/ 10'/ 10 7 10 7 10 6 10 6 l)--c-c\ II I C H o OH O 3 O I CHz 4.7 x 10 5 6.3 x 10 5 1.6 x 10 s 0.9 x 10 5 1.6 x 10 6 0.2 x 10 s 1.5 x 10 s 1 -Pheny I-2-hydro xy-2-methyl-propanone-1 ? Benzoinmethylether 8 POLYMER, 1985, Vol 26, January 145 Photolysis of acylphosphine oxides: T. Sumiyoshi et al. transient absorption spectrum to ketyl radicals of the structure R1 ~__p/R2 I II " R a HO o is discarded because similar results were obtained in different solvents. Ketyl radicals could be formed via hydrogen abstraction from solvent molecules by excited T M P D M molecules. This reaction should be much less probable in benzene than in the other two solvents. The fact that no solvent effect was observed indicates that photoreduction does not occur to a detectable extent 13. With respect to the assignment of the transient optical absorption found with TMDPO, similar conclusions were arrived at by experiments with other acylphosphine oxides having different substituents. Details will be reported in a forthcoming paper. DISCUSSION The important results of this paper concern (a) the assignment of the transient absorption spectrum with its maximum around 330nm, detected during the flash photolysis of T M D P O to diphenyl phosphonyl radicals, (Ph)2.P = O, and (b) the finding that these radicals are of relatively high reactivity towards compounds having olefinic double bonds. The assignment of the transient absorption spectrum is essentially based on the fact that the extinction coefficient of (Ph)z.P=O radicals around 330nm is significantly higher than that of the partner radicals, generated simultaneously with the diphenylphosphonyl radicals. While this fact became obvious by comparing the transient absorption spectra observed with T M D P O and P D P O it can also be substantiated by results obtained with other compounds yielding benzoyl radicals upon photolysis by u.v. light. Such compounds are, e.g., pivalophenone 9, poly(phenyl isopropenyl ketone) 9 and hydroxyalkyl phenones ao. In these cases, transient absorption spectra with maxima around 340, 370 and 410 nm, which were assigned to benzoyl radicals, were observed. However, the extinction coefficients were, generally, very low and amounted at the maxima only to a few hundred 1 mol-1 cm-1. A much higher e value is estimated, for example, from the transient optical density (O.D.)0- 0.09 recorded at 335 nm with a solution of T M D P O in benzene at the end of the flash at Dabs=9.4x 10 -6 einstein/l. From equation (9) (O.D.)0 e335 Dabs~b(R.)d biomolecular rate constants compiled in Table 2 demonstrate that, due to the high reactivity of (Ph)2.P = O radicals, T M D P O is most appropriate among other initiators to initiate the free radical polymerization of various monomers. It has been pointed out by Bentrude 1t that phosphonyl radicals of the general structure R2.P = O are expected to dimerize readily. This becomes feasible since e.s.r, investigations 12 revealed a pyramidal configuration of PhE.P = O and similar radicals with a high degree of spin density on phosphorus: Ph < This expectation was substantiated in this work. The kinetic analysis of the decay of the absorption at 2 = 335 nm resulted in 2k2/e = 3.8 x 105 s- t (in benzene solution at 23°C). With the aid of the extinction coefficient of Ph2.P=O (vide ante) at 335nm one obtains 2k2=3.6x109-7.3x1091mol-ls-1, i.e. k2 refers to an encounter-controlled reaction. ACKNOWLEDGEMENTS The authors are grateful to Dr G. Beck for maintaining the ruby laser and to Dr R. B. Frings for carrying out the fluorescence lifetime measurements. The experiments with vinyl acetate were carried out by Mr W. K. Wong. One of the authors (T.S.) expresses his thanks to the Nishina Memorial Foundation for supporting him financially. Note added in proof The long-lived absorption is due to the triplet of an enol generated by intramolecular hydrogen abstraction. REF ERENCES la b c 2 3a b 4 146 POLYMER, 1985, Vol 26, January ph...~%0 Ph (9) one obtains, with d = 1 cm, e33 s = 9.6 × 103/~b(R'). Because of the uncertainty about the exact value of th(R'), a lower and an upper limit ofe can be given, only: with tk(R') = 1.0: e335=9.6x10 a 1 mo1-1 cm -x (lower limit), and with qb(R-)=0.5: eaas = 1.92 x 104 1 mol -I cm -1 (upper limit). Therefore, it is concluded that the transient absorption spectrum observed in the present work with T M D P O dissolved in various solvents is to be assigned to diphenylphosphonyl radicals. The high reactivity of diphenylphosphonyl radicals towards olefinic compounds is remarkable. Indeed, the > 5 6 7 8 9 l0 11 12 DOS 2830927 (1980) (BASF AG), Leehtken, P., Buethe, I. and Hesse, A. DOS 2909994 (1980) (BASF AG), Lechtken, P., Buethe, I., Jacobi, M. and Trimborn, W. DOS 3023 486 (1980) (Bayer AG), Heine, H.-G., Rotenkranz, H.J. and Rudolph, H. Jacobi, M. and Henne, A. Conference Report, Radcure 83, May 9-11 (1983), Lausanne, Switzerland, Radiat. Curing 10, 16 (1983) Henne, A. and Lechtken, P., 8. Vortragstagung der Fachgruppe Photochemie, Gesellschaft Deutscher Chemiker, Nov. 16-18, 1983, Book of Abstracts, p. 173 Nickolaus, W., Hesse, A. and Scholz, D. Plastverarbeiter 1980, 31, 723 Kuhlmann, R. and Schnabel, W. Polymer 1976, 17, 419; Polymer 1977, 18, 1163; Angew. Makromol. Chem. 1977, 57, 195; Angew. Makromol. Chem. 1978, 70, 145 Beck, G. and Thomas, J. K., J. Chem. Phys. 1972, 57, 3649 Land, E. J. Proc. Roy. Soc. Set. A 1968, 305, 457 Salmassi, A., Eichler, J., Herz, C. P. and Schnabel, W, Polym. Photochem. 1982, 2, 209 Kuhlmann, R. and Schnabel, W. Angew. Makromol. Chem. 1978, 70, 145 Naito, I., Kuhlmann, R. and Schnabel, W. Polymer 1979, 20, 165 Eichler, J., Herz, C. P., Naito, I. and Schnabel, W. J. Photochem. 1980, 12, 225 Bentrude, W. G. 'Phosphorus Radicals', in J. K. Kochi (Ed.),'Free Radicals', John Wiley and Sons, New York, (1973), Vol II, p. 595 Geoffroy, M. and Lucken, E. A. C. Mol. Phys. 1971, 22, 257