NITROCELLULOSE LACQUERS Evaluation of Low-Viscosity-Type Nitrocelluloses in Furniture Finishing Compositions' WILLIAM KOCH, H. C. PHILLIPS, AND RLFLS W I S T Hercules Pourder Company. l?'ilrriirrgton 99, D e l .
blend oi equal parts of dibutyl phthalNitrocellulose of RS type, having viscosiITROCELLULOSE lacquers ate and raw castor oil, a t a constant ties of second, second, 30-35 centihold an outstanding position proportion of 0.5 part of mixed poise, and 1%-25 centipoise, were evaluin the furniture finishing field due to ated in experimental lacquer compositions plasticizer t o 1 part of nitrocellulose such properties as speed of air-dry, in direct comparison with three repreby weight. I n formulas.5 to 8, ingood temperature-change resistance, sentative commercial furniture lacquers. clusive (Table 111),dioctyl phthalate gloss, toughness, and print resistance. Nonvolatile content at a spraying viscosity was substituted for the mixed plastiTheir principal disadvantage has been of 80 centipoises, temperature-change cizer used in the other formulas, and low nonvolatile content a t spraying resistance, print resistance, Sward hardAroplaz 905 was replaced by a differviscosity, principally because RS ness, sandability, and water-spot resistent nonoxidizing alkyd (Glyptal 1i2-second nitrocellulose has been ance data were obtained. It was demon2477). Thereason for these substituthe type generally used in their strated that furniture lacquers, equivations will be discussed later. formulation. lent in performance to the best comThe solvent used throughout was h previous article ( 2 ) showed that mercial compositions, with materially formulated as follows: replacement of RS 1/2-second by increased nonvolatile content, can be tower viscosity nitrocelluloses (I/*formulated with nitrocellulose lower in Ingredient 70by Weight second, 30-35 centipoise, and 18-25 viscosity than R S 1 / ~second; the miniButyl acetate 20 centipoise) resulted in a marked inEthyl acetate 10 mum practical level of nitrocellulose visMethyl ethyl ketone 25 crease in non volatile content of cosity appeared to be 30-35 centipoise in Butanol 10 Denatured ethanol lacquers a t spraying viscosity. The 30 the kind of composition studied. Proper Toluene use of these. lower viscosity types Total 100 selection of resin and plasticizer was imoffers a promising means of increasportant in formulating high quality lacing nonvolatile content, provided This is a typical high-solvency solquers based on RS nitrocellulose '/r-second other desirable characteristics are vent blend. It was used to take and 30-35 centipoise. not sacrificed. T h e purpose of advantage of the increase in nonthis article, therefore, is to show volatile content made possible by how the properties of furniture lacquers are affected by viscosity employing a rich nitrocellulose solvent (3). Experimental types below RS 1/2-second nitrocellulose, and how the lower lacquers were adjusted to n spraying viscosity of 80 ccntipoises viscosity types can be employed, through the proper selection of a t 25" C. for application. resin and plasticizer, to increase nonvolatile content without A D e Vilbiss BfBC type spray gun, equipped with a n FX fluid losing other desirable properties. tip, an F fluid needle, and a No. 70 air cap, was employed with a pressure feed cup. The air pressure a t the gun was 50 pounds per FORMULATIOK 4 N D FINISHING square inch, and the pressure on the lacquer in the feed cup ranged between 5 and 10 pounds. Three representative commercial furniture lacquers were obLacquer films were applied to 3 x 6 inch mood panels for cold tained from the trade. The suppliers were requested to furnish check, print, and water spot tests. A single lot of three-ply, l/,formulas representing the best prenrar quality available (latter inch maple plywood was used throughout in order to minimize part of 1940). These compositions (Table I) served as a basis for as far as possible the variations inherent in different lots of wood: establishing a minimum practical standard of quality. ExperiNo filler or sealer coats Tvere employed, the lacquer being applied mental lacquers were evaluated by comparing their performance directly to the bare wood surface. Steel panels, 3 X 6 inch, were directly with that of the commercial compositions. used for the hardness and sanding tests. These panels, of 20The experimental lacquers in Table I1 vere prepared by formugage auto body steel, were sanded with a Model A Skilsaw sander lating RS nitrocellulose of '/2-second, l/r-second, 30-35 centipoise, and 18-25 centipoise viscosity with resin in ratios, respectively, of 1:1, 1:1.5, 1:2, and 1:3 nitrocellulose toresin by weight. The resin consisted of a blend of a typical commercial nonoxidizing alkyd (Aroplaz 905), and a typical commercial maleic-modified ester gum (Lewisol 33). I n the ratios of nitrocellulose to resin, exemplified by the 1:1, 1 :1.5, and 1:2 ratios, the proportion of nonoxidizing alkyd resin t o maleic-modified ester gum in the blend was maintained a t 60:40 by weight. In the 1:3 nitrocellulose t o resin ratio, this proportion was changed to 66.7:33.3. These resin proportions were selected on the basis of performance in preliminary testing. These lacquers were plasticized with a
OF COMMERCIAL FURNITURE IiACQUERB TABLE I. ET-ALCATIOK
Lacquer designation Nonvolatile content as received, Viscosity, (25' C.)as received, centipNoises Estd. solids a t 80-cp. viscosity, % Cold check, cycles Print test, 2 lb./sq. in. for 3 hr. Sward hardness, 70, after: 2 hr. force-dry a t 130' F. 24 hr. air-dry Sanding (double rubs) after: 2 hr. force-dry a t 130' F. 24 hr. air-dry Water spot
Previous articles in this series appeared in August, 1944 (page 756). and in January, 1945 (page 82). 1
19.5 63 21 25 f Slight
20.5 31 25 13 Bad
21.5 51 23.5
4 4 None
5 5 None
INDUSTRIAL AND ENGINEERING CHEMISTRY
TABLE 11. EFFECT OF VISCOSITY TYPEOF NITROCELLULOSE Ratio, nitrocellulose: resin Formula number RS l/rsec. nitrocellulose RS l/r-sec. nitrocellulose RS 30-35 cp. nitrocellulose RS 18-25 cp. nitrocellulose Aroplaz 905 Lewisol 33 Dibutyl phthalate Raw castor oil
1:1.5 6 7
1. . I
.. 0:6 0.4 0.25 .0.25 23 25+
. . . . . . ..
0:6 0.4 0.25 0.25 28
0.6 0.4 0.4 0.25 0.25 0.25 0.25 29.5 31 10 5
Solids a t 80 cp., % 10 Cold check cycles N N " Print test b lb/sq. in. for 3 hr.n Sward hddness %, after: 2 hr. force-dr$ a t 130' F. 38 40 52 24 hr. air-dry 49 50 54 Sanding (double rubs) after: 2 hr. force-dry a t 130' F. 1 3 8 8 8 6 6 24 hr. air-dry Water spot" N " , N 5 N = none, SI = slight, B = bad, VB = very bad.
. . . . . .
. . . . 0:Q 0.6 0.25 0.25 26
. . . .
0:9 0.6 0.25 0.25 30
0.9 0.6 0.6 0.25 0.25 0.25 0.25 31.0 32.5 ' 9s1 0:9
6 5 N 5 . N
equipped with a No. 120 grit belt, washed with xylene, and airdried before lacquering. Film thickness was controlled on steel panels. by measuring directly with a General Electric film thickness gage (4). The thickness of films laid down on wood was controlled within reasonably close limits by measuring the thickness of similar films laid down on steel using a corresponding number of passes of the spray gun. This procedure was shown to be fairly accurate by actual micrometer measurement of representative coated panels, removal of the film with solvent, and measurement of the bare panel. TEST METHODS
Nonvolatile content a t the spraying viscosity selected (80 centipoises) was obtained from the viscosity-concentration curve determined for each lacquer by the method described in a previous article (2). The cold-check test cycle consisted of exposing each lacquered panel for 1 hour at 110' F., changing the temperature suddenly to -6' F., and keeping i t there for 1 hour. Then close examination was made at room temperature for cold checks brought about by the sudden change. This procedure constituted one cycle. Lacquers which withstood twenty-five such cycles were considered to have excellent temperature-change resistance, and the test was not carried beyond twenty-five cycles. All lacquers were tested in triplicate on 3 X 6 inch, three-ply maple plywood panels which were dry-sanded with No. 400 A Tri-M-ite paper before application of lacquer. It is important t o repeat that a single lot of maple plywood was used for all cold-check tests. Preliminary work showed that different lots of the same type of wood can cause considerable variation in performance. Three coats of lacquer totaling approximately 3 mils in thickness were sprayed on each panel within an 8-hour period. These were airdried between coats and overnight after the last coat. The following day each panel was force-dried a t 130' F. for 8 hours. After cooling to room temperature, a minimum conditioning perio,d of 48 hours at 77" F. and SOYo relative humidity followed before exposure t o the cold-check test. Print resistance was determined by applying a pressure of 2 pounds per square inch to a single thickness of gauze laid directly on the surface of the lacquer film. The panels were inspected after 3 hours for printing. These test panels were pre-.. pared by applying three coats (totaling 3-mil thickness) of lacquer in an 8-hour period to three-ply maple plywood panels which had been dry-sanded lightly with 400 A Tri-M-ite paper. Each coat of lacquer was force-dried 30 minutes at 130' F. After additional overnight air-drying a t room temperature, all lacquer films were weesanded (water) with 400 A Tri-M-ite paper and polished with Simoniz Kleener before print resistance was measured.
. . . . . . .. 1 . . . . . . . . 1 ..
1.2 0.8 0.8 0.8 0.8 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 27.5 31.5 32.5 34 7 4
0.25 0.25 31 25+ VB
0.25 0.25 0.25 0.25 0.25 0.25 35 35.5 36.5 10 9 12 VB
5 5 4 4 5 4 3 3 " "
. . . . . . . . . .
. . . .
I:3 14 :1
Hardness was determined with the Sward rocker hardness meter (1)on 1-mil films sprayed on steel panels. Hardness values shown in the tables were determined after 24 hours of air-drying a t room temperature or after 2 hours of force-drying a t 130' F. Sanding data were also obtained on 1-mil films of lacquer on steel panels. The values indicate the number of double rubs which could be made with dry 400 A Tri-M-ite paper before the film began to roll 'up beneath the sandpaper. One double rub consisted of moving a piece of sandpaper under the pressure of the index and middle finger across the test area and returning to the starting point. All sanding tests were made by the same operator to reduce the factor of human error to a minimum. The values reported were determined after 24 hours of air-drying a t room temperature.or after 2 hours of force-drying a t 130' F. Water spotting was tested by confining water under a watch glass on a limited area of the film surface for 24 hours. The water was then removed, and panels were inspected for blush and grain raising. This test was run on a panel prepared exactly like the panel used for the determination of print resistance. COMMERCIAL LACQUERS
The three commercial furniture lacquers were evaluated, and results are given in Table I. As supplied, the nonvolatile content of the commercial samples ranged from 19.5 t o 21.5%, and viscosity ranged from 31 to 63 centipoises. The samples submitted fall within the class known in the trade as 21% solids lacquers. Since none of them were in the class known as 27y0 solids lacquers, it may be presumed that the samples submitted represented a generally higher level of quality in accordance with the request to supply the best prewar lacquers available. Since each of the commercial compositions differed from the others in both viscosity and nonvolatile content, there was no common basis by which they could be compared with one another or with experimental lacquers. Therefore the nonvolatile content of each lacquer was estimated at 80-centipoise viscosity a t 25" C. by the following method: The determined viscosity and nonvolatile content for each lacquer were plotted as a point on semilog paper, on which was already plotted a viscosity-concentration curve for a lacquer containing a ratio of RS 1/1second nitrocellulose t o resin t o plasticizer of 1 t o 1 to 0.3. A line ww then drawn through the single point parallel t o the experimental lacquer curve. The approximate nonvolatile content at 80 centipoises was then indicated a t the point where the newly drawn parallel line intersected the horizontal line corresponding t o 80 centipoises (Figure 1). The justification for this estimation lay in the fact that numerous viscosity-concentration curves had been plotted on semilog graph paper for experimental lacquers, and that these were practically straight parallel lines within the range of viscosity measured, 40 t o 400 centipoises. Furthermore, this relation
LACQUER O F KNOWN COHPOSITIOH
cellu1ose:resin ratios ranging from 1:l t o 1: 4,and with
values obtained for lacquers containing the three lower viscosity types of nitrocellulose were not considered t o be significant brcause even under the most favorable conditions, there were con-
INDUSTRIAL AND ENGINEERING CHEMISTRY
showed an average gain of approximately 35% in nonvolatile content over the average for the commercial lacquers. Coldcheck resistance wm intermediate between lacquers B and C. In other respects formulas 6 to 8 were comparable to commercial lacquers. Although formulas 10 to 12, inclusive (1:2 nitrocellu1ose:resin ratio), and 14 to 16, inclusive (1:3 ratio), showed progressively higher nonvelatile content, they were, in general, inferior t o the commercial lacquers. EFFECT O F RESIN AND PLASTICIZER
The only property affected adversely to any marked extent by the use of viscosity types below RS 1/2-second nitrocellulose was temperature-change resistance. Improved performance of lacquers with respect t o this property was accomplished by proper selection of both resin and plasticizer (Table 111). The direct substitution of Glyptal2477'for Aroplaz 905 and of dioctyl phthalate for a mixture of equal parts of dibutyl phthalate and raw castor oil improved temperature-change resistance in formulas based on 1/4-second and 30-35 centipoise types without changing the other properties of the lacquers to any extent. Excellent cold-check resistance was obtained in the composition containing RS 1/4-second nitrocellulose and good cold-check resistance in the lacquer containing the RS 30-35 centipoise type. The lacquer based on the RS 18-25 centipoise type was not improved by these changes. This indicates that RS 30-35 centi-
poise nitrocellulose probably represents the lowest viscosity material that can be used in lacquers of this general type without sacrificing cold-check resistance. A comparison of data for formulas 6 and 7 (Table 111) with data for the commercial lacquers indicated the following pertinent facts: The average nonvolatile content of 29% was 26% greater than the average (23%) of the three commercial lacquers. The cold-check resistance of formulas 6 and 7 was much better than that of commercial compositions B and C, and experimental lacquer 6 compared favorably with commercial lacquer A in this respect. Print resistance of the experimental lacquers was superior to that of any of the commercial compositions. Both Sward hardness and sanding of theae experimental lacquers was slightly better than the commercial compositions. Water-spot resistance was alike for all lacquers tested. LITERATURE CITED
Gardner, H. A,, "Physical and Chemical Emmination of Paints, Varnishes, Lacquers and Colors", 9th ed., p. 117, Washington, Inst. of Paint and Varnish Research, 1939. Koch, Wm., Phillips, H. C., and Wint, Rufus, IND,ENG.CHEM., 37, 82-6 (1945). Lowell, J. H. (to du Pont Co.), U. S. Patent 2,291,284 (July 28, 1942). Rush, 11. A . , BUZZ. Am. Cemm. SOC.,14, 365-7 (1935). PRESBNTED on the program of the Division of Paint, Varnish, and Plastics CHEMICAL SOCIETY. Chemistry of the 1945 Meeting-in-Print, AMERICAN
Asbestos as Filter Aid in
Sugar Refining TOH LIU China United Sugar Refining Company, Neikiang, China
F T E R Japan seized the Burma Road, the China United Sugar Refining Company, largest modern sugar refinery in free China, was confronted with the problem of carrying on its filtration processes without the necessary filter aid and filter cloth from abroad. An attempt was made to filter sugar sirups through the remaining supply of imported filter cloth without filter aid, but the flow rate was so low as to render the operation practically impossible. Chinese filter cloth, not so thick or so closely woven as the imported varieties, was then tried; the result was a higher rate of flow, but the filtrate was too turbid to yield a satisfactory product. At this point, the author was called in as consultant. From previous experience in filtering starchy materials, asbestos seemed to be the solution of the difficulty. I n the district where this plant is located, asbestos is abundant
and comparatively inexpensive; furthermore, for sugar refining only the cheaper grades are needed, which consist of various short fibers left when long, uniform fibers are prepared for other applications. Repeated tests have shown that asbestos not only incre:ses the flow rate enormously, but also clarifies and partially decolorizes the sirup by adsorption. *During the early part of the investigation, this company's practice of constant-pressure filtration was followed in order that the tests be made under plant conditions; i.e., a pressure of 50 pounds per square inch was applied a t the @art of filtration and maintained constant throughout the run. It wm soon found, however, 'that better results could be obtained by constant-rate filtration with the pressure gradually raised to 50 pounds a t the end of the operation; a still better procedure was to combine
Imported filter aids (such as Filter-Cel, Hyflo Super-Cel, etc.) had not been available in China since the spring of 1942, so that locally produced asbestos was used as a substitute in sugar refining. Asbestos increases the flow rate and reduces the turbidity and color of the sirup when mixed with the sludges to be filteted. Filtration pressure has a profound influence on the flow rate and clarity; 0.3% asbestos on the weight of raw sugar, together'with careful pressure control, can raise the clarity of sirup
85% and the flow rate at least 943%; a larger proportion of asbestos would undoubtedly have a greater effect under similar conditions. Chinese elt'er cloth may be substituted for imported types in filtering sugar sirups if asbestos is used as a filter aid to remedy the defects of the former. The asbestos in the filter cake can be recovered by washing and re-used several times. The sugar in the cake can be recovered by washing and used as a raw material for the manufacture of industrial alcohol.