J. MATER. CHEM., 1991, 1(6), 1045-1050 Photogenerated Amines and their Use in the Design of a Positive- tone Resist Material based on Electrophilic Aromatic Substitution Stephen Matuszczak,a James F. Cameron,a Jean M. J. Frechet*” and C. Grant Wilsonb a Department of Chemistry, Baker Laboratory, Cornell University, Ithaca, NY 74853-7307, USA IBM Almaden Research Laboratory, San Jose, CA 95720-6099, USA The photogeneration of an active amine within a cationically curable polymer coating can be used to design a novel positive-tone resist material. The resist is based on a copolymer containing 4-hydroxystyrene as well as 4-acetoxymethylstyrene units; when heated in the presence of an acid, this copolymer crosslinks through an electrophilic aromatic substitution process.Therefore, a small amount of 2-nitrobenzyl toluene-p-sulphonate, that decomposes upon heating to produce toluene sulphonic acid, is added to the resist along with a thermally stable but photoactive carbamate that liberates an amine upon irradiation. Exposure of a film of the resist to 254 nm UV radiation results in the formation of a latent image consisting of amine molecules dispersed within the polymer film. The latent image is ‘fixed’ by heating; this liberates acid, which is neutralized where amine has been formed, but causes crosslinking of the polymer by a cationic process in those areas of the film where no amine has been produced. This resist, based on an image-reversal concept applicable to numerous cationically activated resists, can be developed in aqueous base and shows a good sensitivity of ca.19 mJ ern-'. Keywords: Microlithography; Resist; Image reversal ; Photogenerated base The development of highly sensitive resist materials suitable for modern microlithography has proceeded rapidly in recent years with the design of several families of chemically amplified resists.’-3 Perhaps the best known chemically amplified resist to date is based on poly(4-tert-butyloxycarbonyloxystyrene),4 a polymer which is susceptible to acid-catalysed thermolysis of its side-chain tert-butyloxycarbonyl protecting groups in a process that results in a significant change in polarity and solubility. Imaging is achieved through the radiation-induced generation of acid within the polymer film.’-’ This concept has been extended to numerous other functional polymers that are susceptible to acid-catalysed therm~lysis.~.’ More recently, another approach to chemically amplified resists has involved the use of photogenerated acid to crosslink coatings through electrophilic aromatic substitution processes.6-8 This type of reaction is very interesting as it involves a catalytic process that leads to extremely high resist sensitivities.Unfor- tunately, this concept has only been applied successfully to negative-tone resists, since imaging results from crosslinking of the polymer in the exposed areas of the film. In order to extend this mode of imaging to the formation of positive-tone images, we are exploring novel polymer structures that are susceptible to electrophilic aromatic rearrangements.8 The concept of image reversal’ has sometimes been used to change the image tone of a resist material.The best known examples of such processes are the ‘monazoline’, process which is used to produce negative-tone images from the classical novolac-diazonaphthoquinone resists,’ and the gas-phase modification processes that are frequently used in combination with plasma This report describes a simple image-reversal process involving the concept of in situ base photogeneration to effect the local neutralization of an acidic crosslinking catalyst. A preliminary report involving the photogeneration of base to devise a four-component acid- hardening photoresist has appeared.12 Experimental Instrumentations Infrared spectra were recorded using a Nicolet FTIR/44 spectrometer while UV spectra were obtained using a Nicolet UV spectrophotometer. An IBM-Brucker AF 300 spec-trometer operating at 75.4 MHz for 13Cwas used to obtain H and I3C spectra.Gel permeation chromatography was carried out using a Waters 150C gel permeation chromato- graph equipped with a differential refractometer and four PL gel columns of lo6, lo’, lo3 A porosity with tetrahydrofuran as the mobile phase. The molecular-weight data are relative to polystyrene standards. Microanalyses were performed by M.H.W. Laboratories, Phoenix, AZ. Materials Poly(4-acetoxymethylstyrene-co-4-hydroxystyrene), 1, con-taining the two repeating units in a molar ratio of 1 :4 was prepared as described previ~usly~*~ The polymer had M, = 25 000 and M, = 11 000 (GPC with polystyrene standards).The thermally activated acid precursor (TAAP), 2-nitroben- zyl toluene-p-sulphonate, 2, was prepared by the method of Houlihan et a2.13The photochemically activated base precur- sors (PABP) 3-7 were synthesized as reported previou~ly.’~ All materials gave satisfactory elemental analyses and spectra consistent with the proposed structures. 3 4 .No2 J. MATER. CHEM., 1991, VOL. 1 7 Sensitivity Measurements and Imaging Experiments Solutions of 20 wt.% of poly(4-acetoxymethylstyrene-co-4-hydroxystyrene) (1 :4 ratio of repeating units) in 2-methoxy- ethyl ether containing various amounts of 2-nitrobenzyl tolu- ene-p-sulphonate and of the base photoprecursor were prepared.The resulting solutions were filtered through a 0.45 pm Teflon filter and applied to standard silicon, sodium chloride and quartz discs with a Headway Research spin- coater. All films were pre-baked at 80 "C for 1 min. Film thicknesses were measured on a Tencor Sigmascan and were of the order of 0.7 pm. Resist films were irradiated using an Optical Associates exposure system comprising a mercury- xenon lamp with a shutter system, an intensity controller, and an exposure timer. Photon flux was measured using an Optical Associates OAI-3 54 exposure monitor equipped with a 254 nm probe. The lamp output was filtered via a 254 nm narrow bandwidth filter from Oriel Corporation.The light passed through a multidensity resolution filter (Ditrich Optics) in order to produce a range of doses impinging on the resist film. The films were given a post-exposure bake at 120 "C for 1 min to release the acid and effect reaction within the film of copolymer 1. The last step, image development, was achieved by immersion of the films into a stirred solution of aqueous trimethylammonium hydroxide MIF3 12 (Hoechst- Celanese) diluted with water (1 :I) for 130 s, followed by washing with distilled water and air drying. The thickness of the film remaining was measured as a function of the exposure dose received. Film thicknesses were normalized relative to the initial film thickness and plotted against the log of the exposure dose.From these curves, the sensitivity value, Do, and the contrast (slope of the linear portion of the sensitivity curve), y, of each system were determined. Quantitative Infrared Monitoring of the Thermal Generation of Acid A solution of polystyrene in diglyme containing 2-nitrobenzyl toluene-p-sulphonate (5 mol%) was used to cast films of ca. 0.7 pm thickness onto sodium chloride discs. The films were dried at 80°C for 1 min and then subjected to increasing heating periods at 120°C to mimic the post-exposure bake step. By using a Nicolet FTIR/44 spectrometer equipped with software for quantitative analysis, the percentage conversion resulting from each heating cycle was deterrninedI5 by measur- ing the decrease in the asymmetric nitro absorption band at 1530 cm-I relative to the constant intensity C-H defor-mation band at 700 cm-of the polystyrene. Earlier studies have confirmed that the rate of decomposition of 2 upon heating in a variety of polymeric substrates can be measured accurately and in reproducible fashion.' Results and Discussion The overall process to obtain a positive tone image from copolymer 1 that was initially designed to give negative tone images via acid-catalysed crosslinking is shown in Scheme 1.In the first step, the film is irradiated to liberate the base within the polymer matrix. This is followed by application of heat which causes the release of the toluene-p-sulphonic acid from its thermally labile precursor. In the irradiated areas of irradiation photogenerated amine f ifih maskli!JJ1Jlrrrrrlll::I...__.. /\resist -heat c r-7- crosslinked aqueous acid + base positive-tone Scheme 1 the film, the acid is neutralized by the photogenerated base to produce a neutral salt that does not react with copolymer 1.In contrast, acid released in the unirradiated areas of the film catalyses the electrophilic aromatic substitution reaction within 1, which results in crosslinking of the polymer matrix (Scheme 2). Finally, the image is developed using an aqueous base which dissolves the non-crosslinked polymer from exposed areas and affords a non-swollen positive-tone image. Copolymer 1, chosen as the crosslinkable matrix for this study, contains the two repeating units 4-acetoxymethylstyr- ene and 4-hydroxystyrene, in a ratio of 1 :4.This composition was chosen because of its desirable reactivity and solubility properties. It shows a high reactivity, due to the large number of latent electrophilic groups, and it contains enough phenolic units to allow aqueous base development. In addition, this polymer has been studied exten~ively'.~ as a chemically ampli- fied negative resist that is capable of producing high-resolution images well below 0.5 pm in size. The choice of thermally activated acid precursor (TAAP) was governed by the avail- ability of 2. Although a family of interesting thermally labile sulphonium salts has been described recently by Sundell and co-workersl6 the more readily available 2-nitrobenzyl toluene- p-sulphonate was selected.This compound is suitable as a TAAP since it decomposes rapidly'3.'5 at temperatures above its melting point. Scheme 3 outlines the processes involved in the thermal release of acid from 2, as well as in the photochemical release of base from 3 within the resist film. Five different photoactive base precursors were tested, the first three, 3-5, being used to generate cyclohexylamine, and the last two, 6 and 7 being capable of producing 1,6-hexanediamine by exposure to UV light. All five base precursors release carbon dioxide as well as a carbonylated nitrosoarene in addition to the free amine or diamine upon UV irradiation. In order to simplify termin- ology in the case of the diamine precursors, the terms molar OAc crosslinked polymer Scheme 2 J.MATER. CHEM., 1991, VOL. 1 NO2 NO/ \I 2 ,No2 NO 3 Scheme 3 amount or molar ratio will refer to amounts involved in the release of a single equivalent of amino functionality. At first glance, the selection of 2-nitrobenzyl toluene-p- sulphonate as TAAP may appear unusual since this compound is also light sensitive =0.1 1) and its chromophore is similar to that used for one of the base14 photoprecursors (@254=0.13). Exposure to light will therefore result in the release of both acid and base within the exposed areas of the resist, leading to a small reduction in the sensitivity of the overall system if some photons are also consumed in the generation of acid.In practice, this is not a significant problem as the molar ratio of PABP :TAAP (2) is kept relatively high and, therefore, most of the light is in fact absorbed by the base precursor. In order to determine the minimum amount of acid required to crosslink the polymer, different blends containing increasing amounts of 2 as well as thermally stable base precursor such as 3 were tested. Films prepared from these formulations were baked at 80 "C for 1 min, cooled and then baked again at 120 "C for 1 min, thus duplicating the conditions used for the thermal crosslinking reaction. These experiments confirm that films of 1 containing 2 wt.% of 2 become fully crosslinked as a result of this treatment, which mimics that used in the imaging process.Monitoring of the extent of crosslinking was accomplished through film-thickness measurements, no loss of film thickness being observed with the 2% acid precursor formulation upon immersion of the film in aqueous base (undiluted MIF 312) for periods longer than 3 min. Following this study, all resist formulations tested subsequently incorpor- ated 2 wt.% of 2 with respect to copolymer 1. In terms of lithography, an important factor to consider is the UV absorption of the resist films. Values for the absorb- ency per pm of film thickness of the various resist blends used are given in Table 1. The absorbency of the copolymer itself is in the range 0.3-0.4 pm-' (film thickness), a value which is well suited for lithographic applications.However, the relatively high concentrations of PABP required to obtain reasonable sensitivity values lead to a significant increase in the absorption of the films. In addition there is no photo- bleaching of the photoactive compounds upon exposure to 1047 light. Indeed, the products resulting from the 2-nitrobenzyl photorearrangement tend to have a slightly greater absorb- ency than their precursors. To offset this problem of light attenuation, films of lower thicknesses (0.7 pm) were employed. The absorbencies of the various formulations incorporating varying molar ratios PABP :2, with a constant 2 wt.% loading of 2 with respect to the copolymer, are reported in Table 1. Since both the photochemical release of the amine and the thermal liberation of the acid rely on the same type of 2-nitrobenzyl chemistry, it was of interest to attempt to monitor and quantify the overall process by infrared spec- troscopy.The resist films, coated on sodium chloride sub- strates, were treated under the standard imaging conditions of UV exposure at 254 nm, followed by a post-exposure bake at 120 "C, while monitoring the decrease in the nitro absorp- tion band at 1530 cm-'. On increasing UV exposure (254 nm), the asymmetric N-0 stretch of the 2-nitrobenzyl chromo- phore is seen to decrease gradually. As the photoactive base precursor is present in a much higher concentration than the acid precursor 2, this change is primarily the result of the photochemical release of cyclohexylamine.On the other hand, the thermal liberation of acid is not visible by infrared spectroscopy in these imaging formulations since only 2 wt.% of the thermally sensitive 2-nitrobenzyl toluene-p-sulphonate, 2, is present, and the changes that occur during the brief post- exposure bake are too small to be measured. In fact, the thermal decomposition of 2 into toluene-p-sulphonic acid during the baking step can be measured in model studies using 0.7 pm thick films of polystyrene more highly loaded with 2. Monitoring the conversion by quantitative infrared spectroscopy shows that in the typical post-bake period of 1 min at 120 "C only 5% of the available acid precursor 2 is converted into the free acid. As can be seen in Fig.1, this conversion increases with increasing time of post-bake. There- fore, it is not surprising that in the actual resist with only 2% of compound 2, the change in the nitro absorption due to the thermal generation of acid within the resist films cannot be observed. However, if the resist blend is highly loaded with 1.0128 z.I-1.000.-c -7 0.980 2 v 0.960 0 2 0.940.-E 0.920 E 4-11~1111~1111~1111~1111~1111~111 1600 1550 1500 1450 1400 1350 wavenurnber/crn -' Fig. 1 Change in nitro band absorptions upon heating a 0.7 pm thick film of polystyrene doped with 2 at 120°C. (a) Before heating; (b) after 1 min; (c) after 5 min; (d)after 10 min Table 1 UV absorbance per pm of film thickness for resists containing 1, with 2 wt.% of 2 and varying amounts of PABP UV absorbance per pm of film thickness ~ ~~~~~~~ ~~~~ PABP base precursor %4 PABP:2=6: 1 8: 1 12: 1 16: 1 3 0.13 0.74 0.74 0.82 4 0.62 0.76 0.90 1.13 5 0.11 - 0.72 0.82 6' - 0.80 0.95 1.18 7b - - 0.72 0.80 ~~~~~~~~~~~~~ ~ ~~~~~~ From ref.14; for these diamines this molar ratio represents the ratio of latent amino groups to acid groups. 1048 12 wt.% of 2 and 8 molar equivalents of 3, the expected changes in nitro absorption due to both the thermal gener- ation of acid and the photochemical release of amine are readily observable (Fig. 2). Representative positive-tone sensitivity curves obtained for resist formulations containing a cyclohexylamine photoprec- ursor as well as 2 in a ratio of 16: 1 are shown in Fig.3. These curves illustrate that the sensitivity, Do, is strongly dependent upon the nature of the light-sensitive base precur- sor. Table 2 summarizes the resist sensitivity and resist con- trast values that were measured in a series of experiments involving different photoprecursors of amines as well as different molar ratios PABP :2. The resist contrast values are very high, varying from 2.4 to 4.0; although best contrast values are obtained for the simple 2-nitrobenzyl carbamate 3, the differences in contrast between the various amine precur- sors are not significant. Fig.4 illustrates the changes in sensitivity values as a function of the concentration of the photoactive cyclohexylamine precursor.The overall trend is an increase in sensitivity with increasing concentration of the PABP. This effect appears to be most pronounced in the case of 3. The sensitivity of these positive imaging systems is greatly affected by the exact structure of the chromophore used for the cyclohexylamine precursor with compound 5 the most sensitive. The overall order of sensitivity, at each respective concentration, is as follows: 5 >4 >3. The differences in sensitivity between the 2-nitrobenzyloxy carbamate 3 and the related 2,6-dinitrobenzyl carbamate 4 is probably a reflection of their relative photo-efficiencies, 0.13 and 0.62 re~pectively,'~for the photochemical release of c.-t E-0.750 0.7242' I I I I 1 I I I 1 YI I 1 I I I I 1560 1540 1520 1500 1480 wavenumbericm-' Fig.2 Change in the nitro band at 1527 cm-' upon imaging of the photoresist containing 12 wt.% 2 and 8 molar equiv. of 5. (a) Before irradiation; (b) after 20 mJ cm-2; (c) after 40 mJ cmP2; (d) after 60 mJ dm-2; (e) after irradiation as in (d) and heating 1 min at 120 "C 1 J. MATER. CHEM., 1991, VOL. 1 10 100 exposure dose/mJ cm-' Fig. 3 Sensitivity curves for base precursors: a,3; .,4 A,5 cyclohexylamine. In the case of the 2-nitro-a-methylbenzyl- oxycarbonyl-masked cyclohexylamine 5 the quantum efficiency is only 0.1 1, yet the resist sensitivity is highest and this appears to be the most effective PABP for this resist design. This finding suggests that factors other than the quantum efficiency of the PABP are responsible for the ultimate sensitivity (19 mJ cm-2) of the system.This sensitivity value corresponds to an exposure dose (19 mJ cm-2) that is one order of magnitude higher than that required for imaging of the negative-tone resist7 based on the same copolymer 1 and an onium salt. Among other factors, this reflects the lower quantum yield of the overall imaging scheme that requires the photogeneration of an amine in a non-chemically amplified process. An important factor that must be considered is the actual availability of free amine. The 2-nitro-a-methylbenzyl carba- mate 5 is the most effective because the photo by-product that is released upon photolysis is a ketone, 2-nitrosobenzo- phenone (Scheme 4), rather than a nitrosoaldehyde6 as is the case with the other active carbamates 3 and 4.This difference is important since recombination of the photogenerated amine Table 2 Sensitivity and contrast values for different resist compositions base precursor ratioa polymer :2 ratiob PAPB :2 sensitivity D,/mJ cm-2 ~~ contrast, y 3 14.2 8.1 99 4.0 3 12.9 12.0 70 3.2 3 14.3 16.1 54 3.7 4 14.1 8.0 51 3.5 4 14.2 12.0 40 3.2 4 13.9 16.0 35 2.4 5 14.8 5.9 50 2.7 5 13.6 12.0 27 3.1 5 13.6 16.0 19 2.8 6 14.1 8.0 67 3.2 6 14.3 12.2 45 2.7 6 13.5 16.0 36 3.4 7 14.3 12.0 31 2.4 7 14.4 15.8 21 2.8 a Molar ratio of acetoxymethyl groups in polymer 1 to acid precursor 2; molar ratio of amino group precursor to acid precursor 2.Note that for diamines the molecular weight of the precursor is divided by 2 to account for its formation of two amino groups per molecule. J. MATER. CHEM., 1991, VOL. 1 10 15 20 PABP :2 Fig. 4 Changes in sensitivity as a function of loading of base precur-sor. The amount of 2 is constant at 2 wt.% with respect to polymer 1: 0,3; m,4; A,5 NO2 NO '0 5 NO NO / Scheme 4 with this carbonylated by-product to form an imine (Scheme 4) is less likely to occur with the ketonic than with the aldehydic nitroso compound. It must be emphasized that this problem of partial photoproduct recombination is not signifi- cant in most applications of these photoactive carbamates, as our earlier studies have confirmed that, under non-acidic conditions, free amines are indeed produced and little, if any, imine by-products are A problem arises in the specific case of this resist system as acid is also generated in the vicinity of the nitrosocarbonyl and amine photoproducts. Since imine formation is an acid-catalysed process, compe- tition between amine and imine production becomes more significant in the case of compounds 3 or 4 that liberate an aldehydic by-product than is the case with 5.This observation is also supported by earlier work in which low yields of free amino acids were reported for the photochemical deprotection of 2-nitrobenzylcarbamoyl-protected amino acids.I7 There- fore, the 2-nitro-a-methylbenzyl carbamate, despite its rela- tively low quantum efficiency, is the PABP of choice in this application as it affords a significantly greater concentration of free amine with which to neutralize the thermally generated toluene-p-sulphonic acid.Another factor that may play a role in the overall process is the possible photo-dimerization of the nitrosobenzaldehydes by-products of 3 and 4 to azodicarboxylic acids which can then act as efficient internal filters.'* Again, this process, which does not appear to be taking place to a significant extent under the conditions used in this and similar studies,14 is eliminated with 5 since the photo by-product is not an aldehyde. Finally, another potential source of concern is the volatility of the photogenerated cyclohexylamine (b.p.130 "C). Conse- quently, photoprecursors 6 and 7 which liberate the less volatile hexane- 1,6-diamine were also tested. Table 2 also contains the results obtained in resist sensitivity and contrast measurements with these compounds using varying molar ratios of protected amine functionalities and 2. The changes in sensitivity and contrast as a function of the photolabile group and its concentration are similar to those observed for the corresponding cyclohexylamine PABP. Therefore, the volatility of the photogenerated cyclohexylamine is not a significant factor under the conditions used in our imaging experiments. An interesting phenomenon was observed in the amine photoprecursor 7.At higher exposure dose, imaging became increasingly difficult as the film thickness was maintained also in the exposed areas upon development.This problem became acute as the exposure dose was increased further, and full development of the exposed areas to afford a positive image became impossible. The sensitivity curve for this process is shown in Fig. 5. As can be seen in this figure, only a small window exists between 20 and 40 mJ cmP2 for the positive- tone imaging. Doses exceeding 40 mJ cm-2 lead to the forma- tion of an insoluble residue in the exposed areas of the film, as is also the case for exposure doses below 20 mJ cmP2. The exact nature of the reaction leading to this observation is unknown but a possible explanation may lie in the crosslinking of the polymer by the diamine through a yet unconfirmed pathway.No such insolubilization is seen in the case of photogenerated cyclohexylamine. In addition the fact that no crosslinking is observed, even at very high doses, in the case of the 2,6-dinitrobenzyloxycarbonyl-protecteddiamine 6, which is more prone to acid-catalysed imine formation, sug-gests that a high concentration of the free diamine might be required for this crosslinking process to occur. 1 10 100 exposure dose/mJ cm-' Fig. 5 Sensitivity curve with difunctional base precursor 7 1050 J. MATER. CHEM., 1991, VOL. 1 Financial support of this research by the Office of Naval Research and under a gift from IBM Corporation (Materials and Processing Sciences Program) is acknowledged with thanks.8 J. M. J. Frechet, N. Kallman, B. Kryczka, E. Eichler, F. M. Houlihan and C. G. Willson, Polym. Bull., 1988,20,427; H. D. H. Stover, S. Matuszczak, R. Chin, K. Shimizu, C. G. Willson and J. M. J. Frechet, Polym. Muter. Sci. Eng., 1989, 61, 412; J. M. J. Frechet, S. Matuszczak, H. D. H. Stover, B. Reck and C.G. Willson, Polymers in Microlithography, ACS Symp. Ser. 41 2, References 9 C. G. Willson, in Introduction to Microlithography, ed. L. F. 1989, pp. 74-86. C. G. Willson, H. Ito, J. M. J. Frechet, T. G. Tessier and F. M. Houlihan, J. Electrochem. SOC., 1986, 133, 18 1. J. M. J. Frechet, H. Ito and C. G. Willson, Proc. Microcircuit Eng., 1982, 260; C. G. Willson, H. Ito and J.M. J. Frechet, Proc. Microcircuit Eng., 1982, 261. E. Reichmanis, F. M. Houlihan, 0.Nalamasu and T. X. Neenan, Chem. Muter., 1991, 3, 394. J. M. J. Frkchet, E. Eichler, C. G. Willson and H. Ito, Polymer, 1983,24,995; H. Ito, C.G. Willson, J. M. J. Frtchet, M. J. Farrall and E. Eichler, Macromolecules, 1983, 16, 510. J. M. J. Frechet, C. G. Willson, T. Iizawa, T. Nishikubo, K. Igarashi and J. Fahey, Polymers in Microlithography, ACS Symp. Ser. 412, 1989, pp. 100-113; J. M. J. Frechet, E. Eichler, M. Stanciulescu, T. Iizawa, F. Bouchard, F. M. Houlihan and C. G. Willson, in Polymers for High Technology in Electronics and Photonics, ACS Symp. Ser. 346, 1987, pp. 138-148. W. Feely, Proc. SPIE, 1986, 631, 48; R. Dammel, K. F. Dossel, G. Lingnau, J. Theis, H. Huber, H.Oertel and J. Trube, Microelec-tronic Eng., 1989,9,575; A. Bruns, M. Luethje, F. A. Vollenbroek and E. J. Spiertz, Microelectronic Eng., 1987, 6, 467. H. D. H. Stover, S. Matuszczak, C. G. Willson and J. M. J. Fre- 10 11 12 13 14 15 16 17 18 Thompson, C. G. Willson and M. J. Bowden, ACS Advances in Chemistry 2 19, American Chemical Society, Washington D.C., 1983, p. 87 and references therein. J. M. J. Frechet, S. A. MacDonald and C. G. Willson, US. Pat., 4 657 845, 1987. S. A. MacDonald, H. Schlosser, H. Ito, N. J. Clecak and C. G. Willson, Chem. Muter., 1991, 3, 435. W. R. Winkle and K. A. Graziano, J. Photopolym. Sci. Technol., 1990, 3, 4 19. F.M. Houlihan, A. Shugard, R. Gooden and E. Reichmanis, Macromolecules, 1988, 21, 2001. J. F. Cameron and J. M. J. Frkchet, J. Am. Chem. SOC., 1991, 113,4303. J. F. Cameron and J. M. J. Frkchet, Polym. Bull., 1991, 26, 297. P. E. Sundell, PhD. Thesis, Royal Institute of Technology, Stock- holm, Sweden, 1990; H. Jonsson, P. E. Sundell, V. Percec, U. W. Gedde and A. Hult, Polym. Bull., 1991, 25, 649. A. Patchornik, B. Amit and R. B. Woodward, J. Am. Chem. SOC., 1970,92, 6333. V. N. R. Pillai, Synthesis, 1980, 1; W. Reid and M. Wilk, Justus Liebigs Ann. Chem., 1954, 590. chet, Macromolecules, 1991,24, 1741; J. M. J. Frechet, S. Matusz- czak, B. Reck, H.D. H. Stover and C.G. Willson, Macromolecules, 1991, 24, 1746. Paper 1/03059K; Received 20th June, 1991