The present invention relates to coated abrasives and specifically to coated abrasives in which the abrasive particles are held in position by a UV-curable binder.

In the manufacture of coated abrasives, abrasive particles are usually adhered to a backing material by a maker coat and a size coat is placed over the abrasive particles to anchor them in place. Sometimes a supersize coat is applied over the size coat to impart some special property such as anti-loading, antistatic character or to place a grinding aid at the point at which the abrasive particles contact a work piece during use.

Binders most frequently used for the maker and size coats in such structures were and still are phenolic resins though other thermosetting resins have also been used at times. However such binders are slow to cure and require expensive drying and curing equipment to be effective. For this reason in part faster curing binders including those cured using UV radiation have been proposed and to some extent adopted.

As used herein it is understood that the term "UV-cured or UV-curable" embraces resins that can be cured by exposure to actinic light in the visible or ultraviolet part of the spectrum and to electron beam radiation.

Cure of such binder is accelerated by the use of one of a number of classes of photoinitiators which generate free radicals when exposed to UV light. These groups of free-radical generators include organic peroxides, azo compounds, quinones, benzophenones, nitroso compounds, acryl halides, hydrozones, mercapto compounds, pyrylium compounds, triacrylimidazoles, bisimidazoles, chloroalkyltriazines, benzoin ethers, benzil ketals, thioxanthones and acetophenones, including derivatives of such compounds. Among these the most commonly employed are the benzil ketals such as 2,2-dimethoxy-2-phenyl acetophenone (available from Ciba Specialty Chemicals under the trademark Irgacure 651) and acetophenone derivatives such as 2,2-diethoxyacetophenone ("DEAP", which is commercially available from First Chemical Corporation), 2-hydroxy-2-methyl-1-phenyl-propane-1-one ("HMPP", which is commercially available from Ciba Specialty Chemicals under the trademark Darocur 1173), 2-benzyl-2-N,N-dimethylamino-1-(4-morpholinophenyl)-1-butanone, (which is commercially available from Ciba Specialty Chemicals under the trademark Irgacure 369); and 2-methyl-1-(4-(methylthio)phenyl)-2-morpholinopropane-1-one, (available from Ciba Specialty Chemicals under the trademark Irgacure 907).

With the assistance of such photoinitiators such resins cure essentially completely in minutes rather than hours and therefore afford the opportunity for significant cost saving. They do however have a drawback in that, in the presence of solid materials, the cure is often incomplete in areas shielded from the activating light. This can happen as the result of the incorporation of pigments or fillers but it can also happen in the absence of solid materials and merely because the resin layer is particularly thick.

The shielding effect is perhaps acceptable where the resin is applied over abrasive grains such that the greater bulk of the resin is exposed to the UV light during cure. However certain newer products depart from the maker/abrasive particles/size structure by adding the binder and the abrasive particles in the form of a mixture in which the cured binder both adheres the mixture to the substrate backing and acts as a matrix in which the abrasive particles are dispersed. This mixture may be deposited in the form of a uniform layer on the substrate or in the form of a pattern comprising a plurality of composites in repeating patterns, each composite comprising abrasive particles dispersed in the binder, to form the so-called structured or engineered abrasives. It will be appreciated that the shielding effect in such products is quite significantly greater and tends to limit the size of the abrasive particles that can be used and the thickness of the abrasive/binder layer that may be deposited on a substrate.

Incomplete cure is particularly disadvantageous in portions of the structure where the resin contacts the substrate since it leads to poor adhesion to the substrate and poor adhesion leads to poor grinding performance. However this is precisely where the effect is at its most pronounced because it is where the depth of cure and shielding effects are most pronounced.

A new photoinitiator has now been discovered to be surprisingly effective in curing UV-curable resins to greater depths than hitherto considered possible without the assistance of thermal cure initiators. This leads to the possibility that relatively large composites can form part of engineered abrasive products. It also makes possible the elimination of thermal initiators to complete cure of the resin.

The present invention comprises a process for the production of an abrasive tool comprising abrasive particles bonded by a UV-curable resin binder in which the resin binder is present in a formulation which includes bis(2,4,6-trimethylbenzoyl) phenylphosphine oxide as a photoinitiator.

The invention is particularly well adapted to use in the production of coated abrasives but it is also adaptable to the production of other abrasive tools such as thin wheels, and relatively thin segments. Wheels in which a solid wheel-shaped substrate is given a relatively thin abrasive coating around the circumference are also included. The invention however is most readily adaptable to the production of coated abrasives in which a slurry of abrasive particles in a radiation-curable binder is used to provide an abrasive surface on a substrate material. The coated abrasive is preferably one which is laid down with a relief patterned surface, or upon which a patterned surface, (an engineered abrasive), has been imposed such as is described in for example USP 5,014,468; USP 5,152,917; USP 5,833,724 and USP 5,840,088.

The radiation-curable binder can be any one of those that cure by a radiation initiated mechanism. Such resins frequently include polymers and copolymers of monomers with pendant polymerizable acrylate or methacrylate groups. They include acrylated urethanes, epoxy compounds, isocyanates and isocyanurates though these are often copolymerized with monomers such as N-vinyl pyrrolidone that have no (meth)acrylate group. Acrylated polyesters and aminoplasts are also known to be useful. Certain ethylenically unsaturated compounds are also found to be polymerizable by photoinitiated techniques. The most frequently employed binders are based on acrylated epoxies and/or acrylated urethanes and the formulation is chosen to balance rigidity, (primarily reflecting the density of cross-links between polymer chains), and modulus which reflects the lengths of the polymer chains. Achievement of a suitable rigidity can be accomplished by selection of suitable proportions of mono- and/or di- and/or tri-functional components for the binder formulation. Modulus control can be effected for example by selection of oligomeric components and/or by incorporation of a thermoplastic resin into the formulation. All such variations are understood to be embraced by the present invention, provided that radiation-cure of the formulation is accelerated by the use of the bis(2,4,6-trimethylbenzoyl) phenylphosphine oxide photoinitiator.

Polymerization of the resin component of the binder formulation is initiated as a rule by UV radiation to which the bis(2,4,6-trimethylbenzoyl) phenylphosphine oxide used in the present invention is quite susceptible. However the resins can be polymerized under the influence of other radiation such as visible light, electron radiation or other actinic radiation. All such resins are understood to be embraced by the term "radiation-curable".

The initiator that is an essential component of the binder formulations used to make the abrasive tools of the invention is an acylphosphine oxide and this term is understood to embrace compounds having the formula:

• R-CO.-- , wherein R is a hydrogen or a substituted or unsubstituted alkyl, aryl, alkaryl, aralkyl or heterocyclic goup, and any one of X, Y and Z not comprising such an acyl group, is a hydrogen or a substituted or unsubstituted alkyloxy or phenoxy group or a substituted or unsubstituted alkyl, aryl, alkaryl, aralkyl or heterocyclic group.
• BTBPPO (bis(2,4,6-trimethylbenzoyl) phenylphosphine oxide) is available from Ciba. Specialty Chemicals under the trademark Irgacure 819.
• Phosphine oxides are available from BASF as 2,4,6-trimethylbenzoyl-diphenyl phosphine oxide, (as Lucirin TPO) and 2,4,6-trimethylbenzoyl-ethoxyphenyl phosphine oxide, (as Lucirin LR8893).

Thus the bis(2,4,6-trimethylbenzoyl) phenylphosphine oxide initiator can be used alone or also in combination with photoinitiators or even thermal initiators if desired.

Where an abrasive/binder formulation is employed, this can also incorporate other components including but not limited to: fillers such as silica, talc, aluminum trihydrate and the like; and other functional additives such as grinding aids, adhesion promoters, antistatic or anti-loading additives and pigments.

The invention is now described with reference to certain preferred embodiments which are provided to illustrate the invention and the advantages that it affords. They are not however intended to imply any necessary limitation of the scope of the invention.

This Example illustrates the depth of cure of various photoinitiators. A standard slurry of an acrylate-based binder comprised a predetermined amount of aluminum oxide abrasive particles with a grit size of P320 grit. The proportion of abrasive particles in the slurry was 17.39 % by volume and the proportion of potassium tetrafluoroborate particles in the slurry was 27.29% by volume.

The slurry was made up in several samples differing only in the amount of 9R75 Quinn Violet pigment in the slurry. Four Irgacure photoinitiators were evaluated: 819 (bis(2,4,6-trimethylbenzoyl) phenylphosphine oxide); 651 (a benzyl ketal), 369 (an α-amino-acetophenone); and 907 (an α-amino-acetophenone). For each the depth of cure was determined at a number of pigment and photoinitiator concentrations. In each case the mixture was coated on a J-weight polyester woven substrate and passed beneath a UV light source (Fusion UV Systems, Inc., MD) consisting of a 600 watt V-bulb and a 300 watt H-bulb at a speed of 50 feet/minute, (15.2 meters/minute). Depth of cure was determined by the following method. The mix was poured into a foil container (1.5 inch (3.81 cm) in diameter by 0.375 inch (0.95 cm) deep) to a depth of 0.25 inch (0.635 cm). This was passed through UV unit. Any excess uncured resin was removed and the thickness of cured portion was then measured as the depth of cure.

The results are reported on the 3-Dimensional graphs attached as Figure 1 (a, b and c). In each case the plot shows the ratio of cure depth for two photoinitiators. Thus a depth ratio of more than one indicates that one gives a greater cure depth than the photoinitiator against which it is being compared.

From Figure 1(a) which compares the formulation containing the bis(2,4,6-trimethylbenzoyl) phenylphosphine oxide photoinitiator, (819), against one with a conventional benzyl ketal initiator, (651), the acylphosphine oxide photoinitiator gives a uniformly greater cure depth. Figure 1(b) shows that a formulation containing an α-amino-acetophenone photoinitiator, (369), outperforms 651 by almost the same amount as does 819. Figure 1(c) shows that not all α-amino-acetophenone perform equally well since 907 is largely inferior to 651.

To give a more complete picture of the performance of the photoinitiators, the strength of adhesion between the cured coating and the polyester backing was determined. This test is a simple pass/fail test in which the cured material is subjected to an adhesion test by flexing the product over a sharp edge at 90 degree and a value of 1 was accorded to a product that did not separate and 0 was accorded if any separation occurred. Figure 2 (a, b, c, d) records the results in a 3-Dimensional chart for each of the four photoinitiators, 819, 369, 907 and 651 respectively. This shows that for the acylphosphine oxide photoinitiator, (Figure 2 a), failure only occurred at the highest pigment loading and the lowest photoinitiator content. Above 0.2% pigment content the 651 product, (Figure 2 d), failed consistently as did 369, (Figure 2 b), at pigment concentrations of 0.8% or greater except when the photoinitiator concentration was 4% in which case up to 1.6% pigment could be tolerated before failure occurred. Photoinitiator 907, (Figure 2 c), failed under all conditions except when the pigment content was below 0.1% and the photoinitiator concentration was at least 4%. These charts clearly confirm the evaluation from Figure 1 and add the insight regarding adhesion to a substrate which demonstrates convincingly that the 819, bis(2,4,6-trimethylbenzoyl) phenylphosphine oxide photoinitiator gives a much better range of satisfactory adhesion values than the very best α-amino-acetophenone, (369).

In this Example three formulations are used to produce a coated abrasive with a engineered surface. In each case the same acrylate binder was used along with P320 grit alumina abrasive grits in a volume percentage of 17.39% and potassium tetrafluoroborate in a volume percentage of 27.79%.

The backing used was an X weight woven cotton and the engineered abrasive surface was applied using the embossing technique described in USP 5,833,724. The pattern applied was a trihelical design with 10 lines per cm (25 lines per inch).

The performance of three engineered abrasives which differed only in the photoinitiator incorporated into the binder/abrasive formulation was evaluated using the following procedure.

The Examples described above were subjected to grinding tests using a modified 121 Fss Ring Test procedure. In each case a 6.4 cm x 152.4 cm belt was used and the belt was moved at a rate of 1524 smpm. The belt was contacted with a 304 stainless steel ring workpiece, (17.8 cm O.D., 15.2 cm I.D., and 3.1 cm width), at a pressure of 16 psi (110 KN/m²). The contact wheel behind the belt was a 7 inch (17.8 cm) plain face rubber wheel with 60 durometer hardness. The workpiece was moved at a speed of 3 smpm.

Twenty rings were pre-roughened to an initial Ra of 2.032 µm (80 micro inch). The grinding intervals of one minute were followed by measurements of cut amount. With the twenty rings a total of 20 minutes grinding was performed with each belt and the total stock removal were reported.

In each case the initial cut after one minute and the total cut after 20 minutes were measured. The results are given in the Table below. The formulations are identified by the Irgacure photoinitiator used. The coated abrasive made according to the present invention appears in bold characters. The last line on the Table evaluates a conventional, commercial, non-engineered abrasive coated abrasive product. COATED ABRASIVE INITIAL CUT CUMULATIVE CUT Irgacure 819 11.9gm 163.6gm Irgacure 369 11.4gm 150.6gm Irgacure 651 10.4gm 130.3gm R245 10.3gm 68.6gm

As will be appreciated from this Table the coated abrasive according to the invention handily outperformed similar products made using the better performing formulations as evaluated in Example 1 in this very critical "real-world" test.

In this Example the depth of cure and adhesion of formulations containing the same acrylate-based binder and silicon carbide abrasive grits, (grit size 150), in a volume percentage of 17.62% with potassium tetrafluoroborate in a volume percentage of 27.62% were evaluated. Figure 3 compares the depth of cure of these formulations. These formulations differed only in the nature of the photoinitiator used. Each was deposited on an X weight woven cotton backing. Each was evaluated under two conditions: with no surface treatment; and with a surface treatment in which a mixture of silicon carbide abrasive grits (similar to those in the formulation) and a grinding aid, potassium tetrafluoroborate in a 2:1 weight ratio was present.

The adhesion test described in Example 1 was applied to these products. In the Table below "1" indicates a pass and "0" indicates a failure. PHOTOINITIATOR UV SOURCE* NO COATING COATED SURFACE 2% 819 400watt (V-bulb) 1 1 2% 819 600watt (V-bulb) 1 1 4% 819 400watt (V-bulb) 1 1 4% 819 600watt (V-bulb) 1 1 2% 819/1173* 400watt (V-bulb) 1 0 2% 819/1173* 600watt (V-bulb) 1 0 4% 819/1173* 400watt (V-bulb) 1 1 4% 819/1173* 600watt (V-bulb) 1 1 2% 369 400watt (V-bulb) 1 0 2% 369 600watt (V-bulb) 1 1 4% 369 400watt (V-bulb) 1 0 4% 369 600watt (V-bulb) 1 0 2% 369/1173* 400watt (V-bulb) 1 0 2% 369/1173* 600watt (V-bulb) 1 0 4% 369/1173* 400watt (V-bulb) 1 0 4% 369/1173* 600watt (V-bulb) 1 0 2% 651 600watt (D-bulb) 1 0 4% 651 600watt (D-bulb) 1 0 1173* refers to Darocure 1173 (2-hydroxy-2-methyl-1-phenyl propane-1-one, or HMPP) which is a photoinitiator available under that trade name from Ciba Special Chemicals. UV SOURCE* In addition to the radiation source indicated, radiation from a 300watt H-bulb was included in each case. Where a blend is indicated the components were present in the following ratio: 819/1173 (1:3) and 369/1173 (1:3).

In this Example various engineered abrasives are evaluated for their cutting power on 6AL-4V titanium using an evaluation technique in which a 5/8" x 23/8" x 9 ¾" (15.9 mm x 60.3 mm x 247.7 mm) titanium workpiece was ground under 20 psi (138 KN/m²). A plain face rubber contact wheel with a 40 D durometer hardness was used as the contact wheel. The belt speed was 3000 sfpm (914.4 smpm) and the work piece moved reciprocally at 7 sfpm (2.1 smpm).

The formulations were deposited on an X-weight woven cotton backing in one of two patterns: trihelical (TH) with 10 lines per cm (25 lines per inch); and a pyramidal pattern (P) with 10 lines of pyramids per cm (25 lines of pyramids per inch). The patterns were created by embossing the pattern on a surface of the slurry deposited on the substrate. The UV cure in each case was carried out using 300 Watt V bulb and 300 Watt H bulb from Fusion UV Systems, Inc., MD.

The total cut in each case after 15 minutes was measured in each case. The results are set forth in the Table below. PATTERN PHOTOINITIATOR USED TOTAL CUT (gm) 25P 4% 819 21.7 25P 1% 819 + 3% 1173 19.6 25P 1% 819 + 3% 651 18.3 25P 1% 819 + 3% 184 19 25TH 4% 819 29.1 25TH 2% 819 23.0 25TH 1% 819 + 3% 1173 22.6 25TH 1% 819 + 3% 651 21.5 XCF0457* 12.9 *XCF 047 is a commercial non-engineered abrasive made using silicon carbide abrasive grits.

In this Example the depth of cure achieved by three different photoinitiators was compared. Each initiator was added to at the binder used in Example 1 but with no other additives or components being present with the initiator. The amount added was 1 wt% and the binder/initiator blend was applied to a substrate and the coated substrate was subjected to the radaition provided by a 300 W D bulb as the substrate moved under the source at 13.4 meters/minute. In a second evaluation the radiation source was a 600 W D bulb and the rate of passage under the source was also 13.4 meters/minute.

The initiators evaluated were Irgacure1700, (25% DMBAPO WITH 75% HMPP) and Irgacure 4265, (50% TPO with 50% HMPP), and these were compared to Irgacurel 173, (HMPP) alone.

The Results are set out in the following Table: UV SOURCE DEPTH OF CURE 1700 1173 4265 300W D BULB 2.75mm 1.35mm 1.85mm 600W D BULB 3.95mm 1.8mm 2.12mm

Thus it is apparent that the blends of the acylphosphine initiators with other initiators provides a deeper cure than the same total amount of either of the blended components.

From the data provided in the above Examples it is very clear that the acylphosphine oxide photoinitiators can be used alone or in conjunction with other photoinitiators to secure an improved depth of cure and better adhesion to the substrate and, as a consequence, to provide a good total cut that fully meets or exceeds commercial expectations.