Photosensitive Composition of Polynitrate Ester, Aromatic Amines and Organic Esters

The present invention relates to non-silver photosensitive compositions. More particularly, it relates to such compositions comprising, as components, certain aromatic amines and nitrate ester type chain polymers, which components co-react on suitable exposure to ultraviolet light, thereby sufficiently altering the physical properties of the composition to either directly yield a visible and useful image or, with lesser required exposure, be simply and easily developed and fixed to useful images, including compositions useful in direct service as photoresists.

The admixture of small amounts of certain aromatic amines with smokeless powder nitrocellulose has been variously suggested for stabilizing the powder against degradation effects ("Cellulose and Cellulose Derivatives", E. Ott, p 643 Interscience Publishers, 1943). The development of color upon subsequent oxidation of the amines is not a factor in such service.

The oxidation of aromatic amines to form dyes and colorants is old and widely practiced art ("The Chemistry of Synthetic Dyes", K. Verkataraman, especially pp 767-779, Academic Press, 1952). Many oxidants, including nitric acid and its decomposition products, have been employed.

It was also previously known that organic polyhalogen compounds could be combined with aryl amines in photosensitive compositions, wherein the organic polyhalogen compounds were activated by visible or ultraviolet light to generate free radicals, which reacted with the aryl amines to form colored products. See, e.g. Kosar, "Light Sensitive Systems", pp 361-380 (1965); U.S. Pat. No. 3,042,515 to Eugene Wainer. In such systems, the polyhalogen compound and amine were generally encompassed within a plastic film former or base of some sort. The polyhalogen compounds, in addition to providing a material which is photolytically activated to react with the aryl amine to form the colored products, may also form coordination compounds with certain amines, which coordination compounds may themselves further photolytically react. The effect of the ability to form coordination compounds is to extend the photolytic sensitivity of the composition to longer wavelengths of light.

There are a wide variety of plastics which have been used as the binder in systems such as Wainer's. Wainer preferred that the base material be a hydrogen donor (i.e., reducing agent), in which case the overall color-forming reaction was postulated by Wainer as one in which the polyhalogen compound is photolytically activated to form free radicals, in the presence of which the plastic base gives up free hydrogen radicals, with the halogen adding to the amine side chain and the hydrogen of alkyl groups produced adding either to the aryl nucleus or to the plastic base. See U.S. Pat. No. 3,042,515, the disclosure of which is hereby incorporated by reference, especially columns 2-3.

Not all of the plastic binders used have been hydrogen donors even though such materials have been preferred. In fact, a large number of other materials, which are not normally so classified, have been used, including commercially available cellulose polymers, such as ethyl cellulose, cellulose acetate and cellulose nitrate.

Thus, while not preferred, there have been photoreactive systems containing certain aromatic amines, a photoactivatable polyhalogen compound, and a celluse nitrate binder. However, so far as we are aware, in all prior systems where a nitrate ester polymer such as nitrocellulose is suggested or used as a component of photooxidation imaging systems, including those in which copresent amines are photooxidized to a color, e.g., as in Wainer, it serves solely as in inert substrate or binder, the photooxidation being accomplished by other, more readily activated ingredients such as the halocompounds of Wainer. Thus nowhere is it indicated that the color or course of the reaction differs when nitrocellulose is used as the binder in such systems.

It has also been known to form solvent soluble esters of cellulose by using organic acids, all or part of which contain unsaturated carbon linkages of a type adapted to photoinduce crosslinking, e.g., cinnamic acid, for service as a conventional photocrosslinkable photoresist polymer coating.

Apart from prior knowledge in the field of photographic or photoimaging compositions, the chemistry of photoinitiated reactions between monomeric compounds having nitrate ester groups with certain diphenylamines was studied in Hayward et al., "Photolysis of Nitrate Esters, Part I. Photonitration of Diphenylamine", Canadian Journal of Chemistry, 40, pp 434-440 (1962). That paper indicates that ultraviolet photon absorption by diphenylamine peaks at 285 nm and extends to 335 nm, but that model nitrate ester compounds with ring structure comparable with cellulose do not absorb above 240 nm. Hence, though impurities or local sites of degradation reactions, etc. may make a nitrocellulose absorb longer wavelengths somewhat, the nitrate ester groups of nitrocellulose are not photoactivated by wavelengths above 240 nm either. Howard et al. state that mixtures of diphenylamine and their monomeric nitrate ester compounds do develop color on prolonged exposure to ultraviolet wavelengths which are not absorbable by the nitrate esters but are absorbable by diphenylamine. Their work concerns possible uses in chemical manufacture and contains no suggestion either of possible photographic or imaging applications or of using nitrate ester polymers such as nitrocellulose.

We have found that when nitrate ester chain polymers, especially nitrocellulose, are molecularly intimately admixed with certain aromatic amines and exposed to ultraviolet light of wavelengths absorbed by the amine, the thus-activated amine molecules chemically react with the nitrate ester groups of the polymer, yielding products which are more light absorbing. Not only does this permit the direct production of usefully visible images, but the chemically altered polymer polynitrate is rendered less thermally mobile and less readily swollen or dissolved by certain solvents, especially those of relatively large molecular size which also are marginally active (as evidenced by loss of their polymer polynitrate swelling or dissolving activity upon slight dilution with non-solvents free of co-solvent tendencies such as, for service with nitrocellulose, heptane). Further, we find that these changes in the properties of the polymer polynitrate are sufficiently pronounced that a much lessened light exposure, sufficient only to produce a faintly visible image, can be sufficient to permit extractive or thermal evaporative removal of color formers from unexposed areas, yet leave the exposed areas relatively unaffected, thus permitting subsequent exhaustive reaction to develop maximum color potential in the exposed areas either by prolonged overall irradiation or simply by heating sufficiently to thermally react the ingredients. Selective solvent removal of all ingredients from only the unexposed areas yields a system with the necessary properties for use in photoresist service, for which purpose color development is unimportant.

An object of the invention is to provide a method of converting spatial variations in the intensity of actinic radiation into physical images in a high molecular weight material.

Another object of the invention is to provide a means of utilizing the above process to produce high resolution images of high and uniform contrast with sharply defined edges from actinic irradiation patterns of relatively low and uneven contrast.

Another object of the invention is to provide a novel photoactivated recording and copying means havng advantageous properties over the prior art.

Another object of the invention is to use the above processes to photoproduce stencil-like voids or openings in an elsewhere physically protective coating for photoresist type applications.

Other objects and advantages of the present invention will be apparent to those skilled in the art from the following description or from practice of the invention herein disclosed.

The present invention takes advantage of the attack on nitrates by certain photoactivated aromatic amines intimately admixed therewith, and the consequential effects such attack has on the physical properties of the composition. A number of factors affect the selection of amines best suited for a given end use.

While not wishing to be bound by theory, standard theories on the effects of structural changes upon the color of dyes (see, e.g., "The Chemistry of Synthetic Dyes", K. Venkataraman, Ch. VIII, pp 323-400, Academic Press, 1952) appear helpful in deciding which amines would be expected to absorb the longer wavelengths of ultraviolet light. These wavelengths are more conveniently used because they are passed by ordinary glass used for lamps, lenses, photographic plates, etc. Most normal dyes are considered to possess an electrically resonant structure such as an aromatic ring (which can be considered a conjugated double bond chain structure of somewhat moderated electroconductive activity), coupled with one or more additional, preferably electrically polar, structures (chromophores) containing double bonds which can act as uninterrupted extensions of the ring "conjugation", plus additional polar groups (auxochromes), such as amine, hydroxyl, methoxyl and even methyl groups. The greater the uninterrupted length of the conjugation chain, the lower the frequencies for which natural resonant photon absorption bands will occur, and the presence of auxochromes will lower and broaden these absorption bands still further. For present purposes, however, as we wish to maximally activate the amine group, chromophore groups are generally undesirable because part of the absorbed photon's energy would resonate through them instead. However, the absorption frequency of aromatic amines can be lowered to the longer U.V. wavelengths by use of the coupled "conjugation" of fused aromatic rings (such as naphthalene), and additional auxochromes, such as methoxyl or additional amino groups.

Practical considerations dictate that the amine should be sufficiently compatible with the type and grade of polymer used so as to not crystallize out. Nor should it be so volatile that, as with the lowest molecular weight aromatic amine, aniline, it is difficult to retain during drying of a solvent-based coating of the composition. Neither should it be so strong in basic character as to rapidly strip off the nitrate ester groupings during storage by reason of basic activity alone. Handbook tables of ionization constants of amines and standard theory of the effects of structure changes on basicity are helpful here. Conversion to the hydroxylamine form lowers basicity, but the storage stability of most hydroxylamines is quite poor. Neither should the amine be so reactive with atmospheric oxygen as to degrade excessively before it can be used. Most ortho, and especially para, primary diamines and hydroxyamines are bad in this respect. Also, the amine is not active in the salt form so even inner salt formation, including with strongly acidic phenolic groups, can greatly depress activity. However, the color, once produced, may, in some instances, be more advantageous. Presence of an activated ring hydrogen ortho or, preferably, para to the amine group can also be helpful in color development. Peridiamines, such as 1, 8 diaminonaphthalene, are especially desirable. Such amine groups are actually closer spaced than if in the ortho position and can envelop a nitrate ester group, clinging tightly despite their (otherwise desirable) low basicity. Color development is strong and resistance to air oxidation is much better than with the ortho positioning which can form a quinone.

Aromatic amines which are useful in this invention may generally be described by the formula: ##STR1## wherein R is an aromatic nucleus such as benzene, naphthalene or anthracene nucleus, which may be either substituted or unsubstituted, e.g., with alkyl, halogen or aryl groups, and X and Y are the same or different and are selected from the group of alkyl, aryl, and alkaryl, which may be either substituted or unsubstituted, and hydrogen, or either X or Y, but not both, may be hydroxyl. It is believed that to be effective the aromatic amine should be relatively basic, and should not have its effective basicity substantially nullified, e.g. by formation of salts or amides.

In general, all else being equal, activity and color development tends to be better for primary amines. The highly polar nitrate group would be expected to attract the complimentary polarity of the amine group to form a rather close pair in the absence of steric hinderance from other substituents. Additionally, tertiary, and, to a lesser extent, secondary amines, unless photodegraded, are more resistant to oxidation, principally yielding only the milder, generally yellow, color development effects of nitration and nitrosolyation. The free base and leuco forms of tertiary amine basic dyes are a particular exception, of course, because, for the free base form, the full dye color develops immediately on salt formation with a liberated acid group or, for some leuco types, additionally only after a simple oxidation.

Nitrocellulose is commercially available in several degrees of nitration and several average chain length ranges. The latter is specified commercially by stating the product's solution viscosity under arbitrary standard measurement conditions. The present invention is believed workable with the whole range of commercially available nitrocelluloses, which are believed to range from about 10.0% to about 13.5% nitrogen and from less than about 0.1 seconds to over 5000 seconds viscosity. It has been demonstrated as workable using commercial nitrocellulose grades ranging from 10.6 to 13.5% nitrogen content and with viscosities from 1/8 to 1500 seconds. Best results have generally been obtained using material analyzing 11.7 to 12.0% nitrogen content, indicating that about one-third of the available hydroxyl groups were not esterified. With regard to chain length, the nitrocellulose preferably has a viscosity of from about 1/4 seconds to 10 seconds, best results with preferred amines being obtained using nitrocellulose having what is commercially designated as 1/2 second viscosity, corresponding to an intrinsic viscosity of about 0.77 deciliters per gram. Generally, too long chain lengths give compositions which are more difficult to remove by solvents or heat, and nitrocellulose having chain lengths which are too short will not be acceptable for use as a binder.

Solvents are conveniently used to dissolve the ingredients of the composition to facilitate the required molecularly intimate admixture and to permit coating the material on the desired substrate. The solvents and solvent mixtures generally recommended for nitrocellulose coatings dissolve most aromatic amines. However, there are special requirements that make some more desirable than others. The solvent chosen should be readily and completely removable under the thermally mild drying conditions required to avoid both loss of the amine by volatilization and thermal reaction between ingredients. Esters work, but we generally prefer medium volatility ketones. Diethylketone diluted with half of its volume each of toluene and n-propyl alcohol is an example of an acceptable solvent mixture.

Coating may be done by any of the well known standard techniques suitable for nitrocellulose lacquers. For the examples to be described, coating was done either using a conventional plate spinner or a wirewound metering rod, results being the same for a given dry coating thickness. Spraying, dripping and roll printing have also been successfully used.

A number of typical aromatic amines of widely varying structure were tested for color development by this invention. Six parts by weight of amine was intimately mixed with 10 parts by weight of nitrocellulose in methyl ethyl ketone solution, made into a 15% total solids solution in methyl ethyl ketone, and coated onto glass and dried. The coatings were exposed through a suitable imaging mask for 20 minutes at close range to a pair of 71/2 watt fluorescent light bulbs having an ultraviolet emitting phoshor to convert the co-generated shortwave ultraviolet to longwave ultraviolet light capable of being transmitted through the lamp's glass envelope (General Electric type F4T5-BL). The amines tested, together with the color obtained, if any, are given in Table I below:

Examples 1 through 11 are primary amines or diamines and, for the most part, yielded deeper colors than the typical simple secondary and tertiary amines, Examples 12 and 13 respectively. Examples 14 and 15, both tertiary amines, indicate that this is not always so, however. The course of the secondary reactions in Example 14 is not known, but Example 15 represents a leuco (reduced) free amine form of crystal violet, a triphenyl methane type of basic dye. Color production is inappreciable unless both acidification and oxidation (of the remaining hydrogen of the methane carbon) occurs, so fair stability against premature color development is possible with proper precautions. Though they are much weaker bases and less stable, hydroxy-substituted amines, i.e., the hydroxylamines, are simply partially peroxidized amines and also work as shown by Examples 16 and 17.

In Example 18, indole, ##STR2## the nitrogen is part of a ring, but it is not an aromatic ring, and so could also be classed as a secondary amine directly attached to an aromatic ring. Response is weak but definite. Example 19, carbazole, ##STR3## could be similarly classified however, and it gave no apparent development of color until reexposed an additional 15 minutes, Example 20, this time using shortwave ultraviolet from a small (4 watt) germicidal lamp (Example 20). While not wishing to be based on theory, there may be several causes for this much weaker than usual response. The (secondary) amine nitrogen of carbazole is only very weakly basic and also is sufficiently hindered structurally to make close approach to the nitrate ester groups difficult. Perhaps for these or other reasons, it is quite poorly compatible with the nitrocellulose. This is evidenced by the cloudy appearance of the coating and visibly obvious development of carbazole crystals. Thus, a cardinal requirement of the invention, namely, a molecularly intimate admixture of the ingredients, is not properly met. Additionally, the nitrogen atom is located midway between two benzene rings. When the adsorbed photon energy resonates electrically between them, it is believed that it would do so almost exclusively via the direct carbon bond between the two rings because the path via the two nitrogen single bonds in series is an electrically insulative interruption in the conjugation chain. Thus, most of the excitation energy would bypass the nitrogen atom, preferring the highly conductive "short circuit" path instead. Hence, though the compound absorbs the U.V. photon energy well, as evidenced by a bright blue-white fluorescence, very little of the energy would be available to excite the amine group, and, in addition, most of the amine group would probably be out of contact with the nitrate ester groups and so cannot attack them anyway. Despite all this, given a sufficiently short wavelength to excite a diagonal mode between the amine group and one ring instead of between the two rings, the system does work, proving the validity of the inventors' claims to generality. By attaching a vinyl group to the carbazole's nitrogen, ##STR4## Example 21, it is converted to a tertiary amine, but with an added double bond. This greatly improved compatibility with the nitrocellulose, the vinyl group perhaps acting as a conductive "wand" to extend the "reach" of the amine's electrical charge. It would also make it easier for energy to resonate into the wand and thence back through the amine nitrogen. Color development rate is greatly increased.

Examples 22, 23, 24 and 25 show that certain compounds which are closely related to, but which are actually not aromatic amines, do not work under otherwise comparable conditions. Example 22, isatin, ##STR5## showed no color either. However, unlike indole, the aliphatic ring carbon attached to the nitrogen is also attached to a carbonyl group. Thus, in addition to the previously mentioned generally inhibitory effects of chromophore-type groupings, the nitrogen can be classed as an amide rather than an amine, and so is more acidic in nature than it is basic. Some amides do give a slight response, however, as shown in Example 23, diphenyl acetamide, probably reflecting impurities or a partial breakdown of the amide structure to yield the amine. Examples 26 and 27 show that, though the presence of chromophore groups on the ring (in one case an aldehyde, and in the other the much more potent nitro group) do reduce the response of the system, they do not completely block it.

As illustrations of an aromatic amine having a sufficiently high ratio of volatility to reaction rate at elevated temperatures to permit simultaneous image fixing and intensification by simple uncovered heating of the previously imaged coating, the following solutions were coated on glass with the aid of a No. 18 wirewound metering rod (wet laydown, approximately 0.0010 inch) and well dried.

Portions of the slides were imaged close to the pair of 71/2 watt U.V.-fluorescing mercury lamps for 15 and 30 minutes. Though the images were markedly darker for the longer exposure times, they otherwise were about equal for the two samples. They were fixed by heating 15 minutes with the uncoated back side against a 130.degree. C hot plate. The exposed areas increased greatly in color density, those with the shorter exposure closely approaching the density of the longer exposure areas.

The Santicizer MHP altered the developed color from the yellow of Example 28 to a greenish yellow in Example 29. Addition of the Santicizer MHP plasticizer also aided diffusion to the surface and, thereby, the evaporation of the last traces of amine in the unexposed areas, leaving them clear with very sharp transitions to full color. In Example 28 this transition was less sharp and the unexposed areas were slightly discolored. Fixing was complete as indicated by absence of detectable differences after reexposing one-half of the sample, unmasked, for an additional 30 minutes. Optical density and contrast was excellent for the blue and ultraviolet lines of a mercury lamp, making it well suited for use as a working master for exposing ordinary photoresist coatings (which are only sensitive to blue and U.V. light). Its visual transparency at longer wavelengths presents an advantage in conventional microfabrication of integrated semiconductor circuits. This see-through characteristic facilitates visual microalignment of the mask pattern with any pattern present on the wafer from prior photofabrication steps.

When coated using a No. 10 rod, this more dilute formulation left only 28% of the dry film weight of Example 28, but yielded similar results after identical imaging and fixing. It showed good blue and U.V. opacity for use as a photoresist working master plate.

Solvent fixing of the exposed image permits much wider choice of the amine used, including ones that are but little volatilized from the unexposed areas on heating. Because less of such amines is lost from the exposed areas during subsequent thermal intensification of the image colors, acceptable optical densities are possible thereby from lighter weight, thinner films.

The following solutions were spin-coated at 1040 RPM on glass plates and well dried before imagewise exposure to ultraviolet light:

The exposed plates, bearing weak but plainly visible images, were "cleared" by solvent-selective preferential extraction of amine from the unexposed areas by gently agitating the plates for 1 minute in a 9:1 weight ratio mixture of toluene:n-butanol. Any slight discolorations of the unexposed areas, whether caused by atmospheric oxidation of the amine or premature thermal reactions during drying and storage of such slides, leach out, along with the amine, from the slightly solvent-swollen unexposed nitrocellulose matrix. However, the exposed areas resist such swelling and are relatively little affected. The slides were quickly rinsed in heptane and dried. A strong overall exposure to ultraviolet light, e.g., 30+ minutes, serves to intensify the image. However, the particular amines specified in these examples were selected for masking service for exposing photoresist coatings on the basis of the ultraviolet absorption effectiveness of the image made using them. Consequently, as the image intensified, it absorbed most of the ultraviolet light, leaving very little available to activate residual unreacted amine and the color development rate was greatly slowed. Hence, thermal image intensification, e.g., 15 minutes on a 130.degree. C hotplate, is more appealing. Slightly greater image intensity was obtained when an overall ultraviolet exposure was used between the clearing and thermal intensification steps.

The "shutter" (exposed) areas of both these examples were sufficiently opaque to ultraviolet and blue light for commercial photomask service in exposing ordinary photoresist coatings. The "window" (unexposed) areas were quite clear, and transition from one to the other appeared very sharp when inspected under the microscope. There was no difficulty in contact printing the finest pattern of the standard resolution test target used, 228 line pairs/mm. These fine clear and opaque lines were reproduced, respectively, as opaque and clear lines of the same relative widths.

The nitrocellulose can be extracted from the unexposed areas along with the amine. The formulations listed in Table 2 were spin-coated on glass plates at 1040 RPM. The plates were hot-dried thoroughly, exposed for the indicated times, and extracted by gently agitating in diisobutyl ketone for one minute, rinsed a few seconds in heptane containing 10% of isopropyl ether, dried, intensified by placing the uncoated side of the glass down against the surface of a 260.degree. C hot plate for 1 minute, and then cooled. The coating was effectively cleared from the unexposed areas in each case, leaving clear window areas.

Examples 34, 35 and 36 show that the ability to selectively dissolve the nitrocellulose from the unexposed areas, leaving the exposed areas relatively unaffected, is not peculiar to one amine. Because of the greated basicity of the amines used in Examples 35 and 36, only mild warming was permissable during drying or the unexposed areas would not extract properly. The intensified image of Example 36 was very dark blue, appearing virtually black when placed in contact with a white background. However, this is true only if intensification is by mild heating or, preferably, by a secondary strong ultraviolet light exposure. On strong heating, the colorant became brown and was much reduced in intensity.

The foregoing description is contained in our copending application, U.S. Ser. No. 560,565, entitled "PHOTOSENSITIVE COMPOSITIONS," filed concurrently, herewith, now abandoned. As good as the compositions described therein are, it has now been found that substantial improvements in fixation can be obtained by addition of certain organic acids. For example, while solvent removal in Example 33 was fairly specific, still some of the amine was also leached from the surface and edges of exposed areas in Example 33. By adding a small amount of picric acid, Example 37, this loss was considerable moderated, and by including o-nitrobenzoic acid in the extractant, Example 38, it was nearly eliminated. This sharpened the edges of exposed areas by virtually eliminating the transition zone of the cleared but undissolved nitrocellulose present with Example 33. Also, the decreased amine loss resulted in markedly better optical density after intensification of the exposed areas despite the use of more solvent in the coating mix to decrease the coat weight. Additionally, these changes decreased 4-fold the minimum exposure time required for full image intensity after processing. The effective weighted optical density of an Example 38 slide, for use as a photoresist processing master with a medium pressure mercury arc light source, measured 3.6 at a coating thickness of less than one micrometer. It permitted a contact printing resolutin approaching 1000 line pairs/mm. Compositions sharing such acid addition with various amines are shown in Table III below:

Examples 39, 40 and 41 show the use of such acids with the amines of Examples 34-36.

The action of 2,6 dinitro benzoic acid is greater than that of o-nitrobenzoic acid, and can be modified, if desired by combined use with 3,5 dinitrobenzoic acid. The choice of acid to be used depends in some measure on the solvent system in which it is to be incorporated. Generally, organic acids having a pK.sub.A of from about 0.1 to 3.5, preferably from 0.5 to 2, will be workable, although the pK.sub.A measured in water is not always a reliable gauge of behavior in other solvent systems. Some highly polar materials, such as acetic acid or propionic acid, may require the addition of materials to counteract the increased solvency imparted by the polarity of the acid. Acids which have been successfully used include citric, citraconic, oxalic, maleic, fumaric, malonic, chlorinated acetic, chlorinated propionic, diesters of phosphoric acid, and o-benzic sulfimide. The latter is actually an imide, but, as is well known, acts as a strong acid.

The acid appears not to simply act as a sensitizer, but rather the opposite. Some of the amine is consumed to form the salt of the acid and in this form it is photochemically inactive. Also, in the use of picric acid, it competitively absorbs some of the ultraviolet light so the extent of the photoinduced color reaction is actually reduced somewhat. Though light absorption of the composition was made strong for blue light, it remained insensitive to visible light. Photoresponse is greatly reduced at high levels of acid addition.

The improved performance obtained is believed to depend on the formation of molecularly bulky salts of the amine which are low in solubility in the ketone extractant used. These salt molecules are thought to block the gel pores, thus hindering both the diffusion out of more amine and the diffusion in of more solvent-swellant. A wide range of acids have been found to serve the purpose, but the ones used in Examples 37-41 represent a fair compromise of conflicting requirements. For incorporation in the coating, the salt must be sufficiently compatible with the coating solvent, the amine and the nitrocellulose to remain molecularly dispersed. For both uses it must be of sufficient ionic strength to remain largely in the salt form. On momentary dissociation, the two components can be separately dissolved and then kept apart by the vastly larger number of solvent atoms. However, too great an ionic strength acid, e.g., anhydrous toluene sulfonic acid, has been found to slow the rate of solution of the unexposed areas virtually to zero.

In sum, the inclusion of the organic acids in the aromatic amine/polymer polynitrate ester compositions or the extractant used therewith in accordance with the present invention results in a drastic reduction of exposure necessary, preferentially inhibits dissolution of the exposed areas, and results in substantially improved contrast, even at low exposure.

That the photoreaction between accessible nitrate ester groups and the present class of amines is general rather than specific to the cellulose polymer backbone is apparent from the cited work of Hayward et al on monomeric nitrate esters. Hence, it is apparent that the invention should similarly serve for direct imaging applications if any otherwise suitable polymer bearing accessible nitrate ester groups were substituted for the nitrocellulose. Further, because of the photoproduced drastic changes in the nature of the polymer, including greatly increased opportunity for formation of hydrogen bond crosslinks between polymer chains, its susceptibility to suitable-chosen border-line solvents would vary markedly between exposed and unexposed areas. Thus, the same post-exposure processing techniques used with the nitrocellulose-containing embodiments of the invention would, appropriately modified, serve similarly. Special advantage accrues in the case of photoresist service because a polymer backbone resistant to acid hydrolysis could be used, for example, the nitrate ester derivative of poly(vinyl alcohol) or polybutadiene modified by adding nitric acid to its double bonds.

The only commercially offered polymeric nitrate esters readily available to the applicants were the various grades of nitrocellulose. Accordingly, to support the applicants' contention of generality with respect to the polymer used, corn starch and poly(vinyl alcohol) were each highly but non-exhaustively nitrated at 25.degree. C for 1 hour using 30 times their weight of a mixed acid composed of 3 parts of 98% strength sulfuric acid to one part of 60% strength nitric acid. To avoid runaway oxidative decomposition, it was necessary to moderate the heat of reaction with poly(vinyl alcohol) by substituting poly(vinylacetate), from which the acetate groups are readily displaced by the much stronger and more plentiful nitrate groups. The reacted mixes were drowned and neutralized by slowly pouring into a large excess of strongly stirred cool aqueous sodium carbonate solution. The insoluble polymeric nitrate ester was recovered, washed, dried, dissolved in diethyl ketone, filtered, assayed for solids, and tested in coatings formulated as follows, being spin coated on glass at 1000 RPM and dried before imaging:

In each case, a colored image was clearly visible after the patterned exposure to ultraviolet light, one which could be intensified by heat after auto-selective extraction of amine from the unexposed areas as described for Examples 31 and 32.

The stated nitration techniques proved far from optimum for both polymers. Extensive chain degradation occured as indicated by very low solution viscosities. Such shortened chains made these nitrate esters somewhat marginally soluble in water. Consequently, the usual inevitable slight denitration such as occurs when an acid mix of a typical nitrocellulose batch is "drowned" into water was, for these supposedly analogous preparations, much more severe, especially for the more soluble shorter chain fragments. The resultant products were, therefore, far from homogeneous as regards both chain length and degree of nitration. Thus, it was not possible for any one solvent or solvent mixture to simultaneously be acceptably marginal in dissolving power for all chains of such heterogeneously constituted polymer mixtures and attempts at complete selective solvent extractive removal of all materials from the unexposed areas alone were, therefore, only partially successful. It will be obvious to those skilled in the art, however, that such difficulties would not be expected had commercially homogeneous nitrate ester polymer preparations of these chemical classes been available and used.

The specific embodiments described herein are meant to be exemplary only, and various modifications will be apparent to those of any skill in the art. The claims below are intended to cover all such modifications as fall within the true spirit and scope of the invention.