Document 7wzDEDprgjQ8R58zo6Jd7oBR

of Buckman Laborstoriaa. Inc. 1953 PREPRINT f o r e s t p r o d u c t s r e s e a r c h s o c ie t y 252 Preliminary copy of a report to b prose&tod at tbe SevooUi Annual KatSosal to 1m bid art MompRbr Tm< IS* tt 17# 11$). Kot to bo sopublialiod oxcopt upon roiooM by FPRS Ex*c. Bocy.* Box M10* Uxdv. Bto.r Madison S. Wis. Tbo Boetoty la not soopaudblo tor bo rbwi oxproaaod. The Absorption of Water, Swelling and Solubility of Free Films of Paint p Q F. L. BROWNE Chemist, Forest Products Laboratory, Forest Service, U. S. Department of Agriculture1 1 1957 Tests substantiating that free films of house paints soaked in distilled water absorb surprisingly large amounts of water are de scribed. Amount is often more than 1 Vs times the weight of linseed oil that the film contains. When such films are young, swelling of many paints may cause expansion of surface area as much as 50 percent. T IS POPULARLY SUPPOSED that I coatings of oil paints and of dry ing oils are inert or nearly inert to water. Because oil and water are prac tically insoluble in each other, it is easy to assume that the incompatibility continues after the oil has become oxi dized, polymerized, and cross-linked into a gel structure. Nevertheless, more than 30 years ago English (23)* and German (14, 29) workers showed that films of linseed oil after drying absorb a third or even half of their weight when immersed in water for a few days, and Dunlap (12, 13) and later Browne (4) at the Forest Prod ucts Laboratory reported that coating wood with paint, varnish, or oil merely retards without preventing changes in moisture content of the wood. Since then there have been many publications on the absorption of water (2, 2, 3, B, 25, 26, 27, 39> 40) and water vapor (16, 17, 31) by coatings of oils and paints and on the leaching of water-soluble ingredi ents from coatings (11, 14, 20, 21). Many of the studies were concerned 1 Miinuined it Madison. Wis.. in cooperation wjth the University of Wisconsin. a Italic figures in parentheses refer to Litera ture Cited at end of toe piper. ABOUT THE AUTHOR: BROWNE, DR. FREDERICK L. B. Chcm. degree in 1917 at Cornell University and the Ph. D. degree in colloid chemistiy in 1921 at the University of Wisconsin. He first joined the staff of the Forest Products Laboratory in 1918-19 to develop casein glues for the wood airplanes of World War I. In 1921-- 22 he was a National Research Fellow at Wisconsin. He is one of the "iron men" of the Chemical Abstracts staff and has edited Section 2, General and Physical Chemistry, since 1929. Besides the Ameri can Chemical Society, he is a fellow of the American Association fox the advancement of Science and the Wisconsin Academy of Sciences. Arts, and Letters, and is a mem ber of Sigma Xi, of the American Society for Testing Materials, and FPRS. with the evaluation of rust-protective paints and the mechanism of protec tion against corrosion for metal (18, 19. 20, 24, 28, 32 to 38, inch). Most observations were made with coatings supported on glass or metal; few on free, unsupported films (30, 31). In all of the publications cited, swelling was measured in terms of the absorp tion of water by the films; no direct measurements of changes in dimen sions are reported. The loss of substance by leaching and the swelling and shrinking from gain and loss of water are surely sig nificant properties of paint coatings about which more should be known before the behavior of paints on ex terior woodwork and the chemistry of coating deterioration can be explored to full advantage. Such developments as checking, cracking, curling, making, and blistering clearly reveal the opera tion of sizeable stresses on the coat ings. It is commonly assumed that the stresses originate outside the coatings, particularly in movement or develop ment of fluid pressure within the wood surface that supports the paint. The discoveries that water swells many of the paints in common use to a much greater extent than it does wood and that painted surfaces on houses are continuously wet for long periods of time (15) strongly suggest that the destructive stresses may arise chiefly within the coatings themselves. In order to limit the present studies to the behavior of the coating mate rials themselves without complications from the effects of the substrate, the tests were made on free unsupported films of clear drying oils and oil paints. All of the films swelled meas urably when immersed in water, many of them very greatly. The swelling observed was the increase in area. calculated from precise measurements of increase in length and width of the films. (Convenient means were not found for measuring increase in thick ness with sufficient precision to calcu late volumetric swelling reliably, though k is hoped that some deter minations of change in volume can be made in future work.) Films at tached to a substrate, of course, can not swell freely in length and width; length and width must remain fixed by the requirements of the substrate if separation is to be avoided, and any necessary adjustment of volume must be made by change in film thickness. In a substantially rigid medium such adjustment of volume by change in one dimension only necessarily de velops internal stresses. By reason of such stresses it may be that films attached to substrates do not absorb so much water as the free films used in these studies, although the data reported in the literature indicate that any such difference is small. Methods of Obtaining Free Films Suitable substrates were coated with oil or paint by brushing or by drawing Hew the length end width of o merited portion of o free film of point held fiat between, gloss plates wet measured before end after swelling In water fey use of a traveling microscope. down with a Bird doctor blade. The coatings were allowed to stand 10 days to dry and harden. Some were then stripped from the substrafe for immediate testing, whereas others were exposed in an accelerated weath ering machine or to natural weather ing out of doors for desired periods of time, after which the coatings were stripped from the substrate for testing. Three kinds of substrate permitted the films to be stripped off at the desired time. 1. Flat tinplate (in pieces 3% by 5'/8 inches) coated on one face held coatings well during natural or arti ficial weathering, after which free films could be stripped off readily by amalgamating the tin with mercury. Enough coating was scraped off near the edges of a specimen to expose tin, and the specimen was floated on mer cury with the coated face down for about half an hour. With care, it was then possible to loosen the coating at one of its edges and slowly peel it from the metal without breaking the freed film, provided that the coating had not been weathered too long to permit moderate bending without breaking. Much trouble was experi enced, however, in stripping unweath ered coatings only 10 days old with out seriously tearing the films. There might be less difficulty if the speci mens with young coatings were chilled to 40 F. for stripping. In these studies the amalgamation process was used chiefly for specimens subjected to natural weathering. 2. Flat pieces of glass, metal, or paper coated with a thin layer of sili cone parting agent DC-7 (DowComing Co.) applied with a finger can be painted readily, will hold coat ings for a limited period of natural or artificial weathering, and yet allow the films to be stripped off when needed for testing. Some of the films for these studies were prepared on siiiconed paper. Shortcomings of the method were a graininess in the layer of silicone that was transmitted to the free films and a tendency for some paints to crack and to begin to peel from embrittlement at an earlier stage of natural or artificial weathering than they did on tinplate or on gummed paper. It is possible that both shortcomings might be over come by coating the paper mechani cally with a thinner and more nearly uniform layer of silicone than was possible with manual application. 3. Commercial gummed (dextrine mucilage) paper, previously saturated with linseed oil, proved the most con venient substrate for most of the present tests. Prior saturation of the paper with linseed oil was necessary because the layer of dextrine muci lage is readily permeable to the oil in paints; if the paper is allowed to draw oil through the dextrine the composi tion of the paint coating is altered undesirably. Gummed-paper specimens held most paints satisfactorily for 1,200 kilowatt-hours (about 360 hours of ultraviolet light with water spray for 11/2 minutes in each 18 minutes), by which time all paints were chalk ing, most of them very freely, and were greatly embrittled, some to the point of cracking. To strip off the tree films for testing, the specimens were placed in a room at 80 F. and 97 percent relative humidity for about 30 minutes to soften the oiled paper, plasticize embrittled coatings, and melt the dextrine mucilage. The films were then peeled from the oiled paper and placed face down on a glass plate while any mucilage clinging to the backs was quickly sponged off with moist absorbent cotton. Alternatively, the specimens were placed on a pad of damp blotting paper, paper-side down, and held flat under the weight of glass plates for half an hour be fore stripping off the films. A few very soft coatings, particularly those of raw linseed oil containing paint drier, required a special technique in which the specimens were immersed in water chilled to 40 F. for 13 to 30 minutes. The films were then stripped off by loosening the film at a comer of the specimen sufficiently to insert a finger tip, by placing the specimen coated side down in the cold water, and' by holding the film firmly against the bottom of the glass water tray while the paper was carefully peeled off. Soft films are very much firmer at 40 F. than they are at 70 or 80 F. But too long a soaking period may blister coatings of paints that swell greatly, and the blisters may remain in the films after they have been stripped from the substrate. Other Experimental Procedures Tests were made with 32 coating materials. Four were clear (unpigmented) materials, namely, raw lin seed oil, bodied linseed oil of viscosity Z3, bodied linseed oil of viscosity Z6, and a commercial alkyd-resin paint vehicle said to contain 77 percent dry ing oil and 16 percent phthalic an hydride by weight These liquids were used also for the vehicles of 23 paints of varying pigmentation ground at the Forest Products Laboratory. Of the 23 laboratory paints, 18 were practicable house paints and 5 were single-pigment paints of scientific in terest only. All of the linseed oils and oil paints contained lead-manganese naphthenate driers; the alkyd resin ana alkyd-resin paints contained leadcobalt naphthenate driers. The remain ing 5 coating materials were tradebrand paints purchased in Madison in July 1952, of which 4 were products of large manufacturers with national distribution and 1 was a so-called "breather type" or "blister-resistant" paint made Iby a smaller manufacturer. All of the commercial paints had the formula printed on die label. In this report the composition of paints is expressed by volume in the short no tation first described by Browne (3) and recently described in this Journal (6) and elsewhere (7). fAiHhora of applying coalings of controlM fnhkttmz by y of Bird doctor bM< and tucHan plate. 2 GLD37969 For most of the experiments re ported, paints were applied with a Bird film applicator or doctor blade drawn manually over a substrate held firmly on a Bird suction plate. With each paint, specimens were prepared in pairs, one with a draw-down with a blade that left a coating nominally 3 mils thick and the other with a 6-mil blade. Actual thicknesses, how ever, varied appreciably from the nominal. When made on gummed paper 8 by 10i/2 inches in size, such coatings were 3'/2 inches wide and IOI/2 inches long and were then trimmed to specimens 3Y2 inches by 8 inches by cutting off the uncoated parts of paper and the first and last parts of tne draw-down, which might be of unrepresentative thickness. When made on tinplate the coatings covered nearly all of the surface of 3% by 5yB inch specimens, which were not trimmed after coating. Although the 3-mil and 6-mil draw downs made much thicker coatings than are usually spread at one time in practical painting, most of the paints dried without wrinkling and none wrinkled too seriously to be accepta ble for testing. The clear liquids, how ever, wrinkled too badly for use. Clear liquids, therefore, were applied by brushing in a sufficient number of successive thin coats to build coatings thick enough for the purpose. Some of the paints for preliminary tests were also applied by brushing. In brush application, coating materials were always spread by stroking both longitudinally and crosswise for even application, but care was taken to see that the brushing ended with longi tudinal strokes over all portions of the areas coated. For exposure to weathering, speci mens of coatings on tinplate and on siliconed paper were fastened with weather-resistant pressure 1 tape to sheets of aluminum-painted] plywood that were attached to outdoor test Table 1--The behavior of fly* .Undf _gf flljuhMl_tMk*d eontlnuouilr In dletnied water at 69.5* E for p+rlo&a of JlatL.uR^JS Mlt Thi film* war* 1.5 by 2.5 lnohaa In alsa, of tha thiokneee indioated below, arwl vara 9 day* old whan they ware etr.'.pped from gummed-paper eubetratee and eubmerged under Aiatilled water. Negative value* for availing after radrylng naan that the fila ehrank to 1M than tha area before lameraion. ---------------n TT Number 11 of day* !t lmaeraed:t Water in water: rabeorbed :: Thinner film* it it. < Swelling tSolubilityt > 1 (Ineraaaa in area) tfLeaohed It i--_------~---- t out by tt tin water 1 Redried t water) 11 t (mg) t (d<pj t () ft Thieker film* Water I Swelling abeorbed: (Inoreaee in area) tin water 1 Redried (A) I (*Q) I Idftt) tSolubility :(L*aoh*d 1 out by !tNUiIgr^|i hwi>t \ iarotPt Ixilllgreaei ntmutraael Tereent t IPeroent iaiiujrou 11 *tr Thlcknc* 3.4 all*. Initial waigni 1 it 23 it 19 12 it 27 20 ti *5 2 it ? it 2 -0.5 -1.0 4 -1.4 5 12 :f:i 9 containing lead-Banganeee drier Bllilgrm nThiokruxi balfr iiar initial weight 371.4 milligram* :: 4.7 it 4.7 it 46 9.7 It 48 12.5 li 65 22.8 it 4 t -1.8 % t 41.4 t - .9 I -1.1 14 t 41.1 &! Ill 16.1 33.8 It 18 l., 38.9 AThulaokvnaenewe 2a. .8 mile, -aaa. initial vei^n^^.^SIxitgbaiS^^ThlSkn^ei^H mile, initial weight 1034 milllgrame 3 6 * tt 27 37 t 10 t 17 I .5 1 2.1 X 6.8 t 7.8 It it 22 t S 5 7 t - .6 t I - .2 I 13.0 13.9 12 tt 50 t 19 1 3.3 l 9.9 tt 61 t 10 t 4.3 1 18.5 n20 tt 57 t 22 1 4.3 t 12.2 It 32 tt 64 t 25 t 5.7 t 12.3 It t 11 X 12 1 4 .8 1 . 1 4 .8 l 22.7 23.4 49 tt 74 t 24 t 6.0 i 14.1 tt 90 t 12 I 4 .9 t 21.8 Taint /nr 0.30 (anatae* titanium dioxide Paint) ^ZXi Ec LIm 3 tt 59 t 20 t 41.6 t 6.2 it 72 t u 6 tt 87 t 35 1 44.2 I 7.8 It 114 t 17 12 tt 92 t 39 1 46.0 1 11.6 tt 152 t 22 20 tt 96 t 43 I 43.1 t 14.9 It 158 t 21 32 tt 95 l 40 I 4 .4 I 21.2 It 177 t 21 49 XX 79 t 28 t -2.9 1 30.6 tt 158 t 20 t - .7 t t 4 .5 I t 41.4 t t 42.0 t i 42.8 t t 45.0 t 7.1 9.4 14.5 19.9 m J ***** ***** 3 it 59 * 70 1 25 6 i i, # 12 ltf. 20 lt*T n.. 49 II.. Ihiekneee .X^SLgLllg. 3 6 12 , 20 2f tt 140 tt 143 tt 152 tt 165 St 180 tt 169 t s 1 I t 1 2! 42 46 46 45 t 11 t 13 t ll t 13 t 12 t 11 : *llt- Initial welaht 8ll.2mir*m* t 10ml tt 284 400 416 362 342 365 t 4? t 68 t l oX t 58 t 59 l 16 l 26 t 25 ut t t 23 t 12.7 1 17.7 t 20.3 t 26.0 1 29.9 I 35.0 p/wv 0.30 (titanlum-leed-slno paint) S 8.2 t 9.4 t 12.2 I 13.6 t 17.0 s 19.3 tt 239 tt 317 IS 432 tt 425 it 430 tt 4?B 3 t 28 s t 21 I 41 1 40 t 40 t8 t 13 t 18 s 16 t 16 s 16 t 8.6 t 12.0 s 17.1 s 21.6 $ 24.1 s 29.6 GLD37970 Effech of cooking in water ond then redrying on free flint of two different kinds of point. The three Aims in the lop row were of Ktonlvm-xinc point (TZsr, p/nv 0.30), tho three on the bottom row were of pure white lead point (l, p/nv 0.30). The two flint in tho left-hand Column were dry and show the markings in drawing ink between which length ond width were measured accurately with a traveling microscope. The two Aims In the center column had just been token from water after soaking for three days ond were still thoroughly wet. The two flint in the right-hand column bad been soaked In water and hod then been allowed to dry out again. The rumpled Aims were flattened out for measurement by sandwiching them between sheets of gloss. racks facing south and sloped back at 45 from the vertical. For exposure to artificial weather ing, one set of specimens consisting of five commercial and five laboratorymade paints spread on gummed paper and on siliconed paper were exposed continuously to ultraviolet light from a flaming carbon arc covered by a globe of Corex D glass (Atlas Weatherometer) in complete dryness, that is, without use of water spray at any time. Samples were taken for test after 400, 800, and 1,600 kilowatt-hours' exposure. The results with this run, in which water-soluble products were allowed to accumulate in the films until the immersion tests, suggested that light without moisture fails to effect the visible changes in paint that are reasonably representa tive of natural weathering. It was therefore decided to use the water spray in the weathering machine after equipping the spray with special noz zles that produced a fine mist rather than jets of water impinging force fully on the paint coating. The mist deposited water gently on the sur faces in sufficient quantity to wet the surfaces thoroughly and then to drain down over the specimens and drip copiously from the bottom edges. The specimens were sprayed in that way about three times every hour. Loss of soluble ingredients by such leaching was measured by weighing specimens of coated tinplate before exposure and after periods of 400, 800, and 1,200 kilowatt-hours of exposure. Specimens of the coatings on gummed paper exposed simultaneously were then used for stripping films for measurements of water absorption and swelling. There was, however, further loss of soluble ingredients from the free films during the immersion tests. The usual procedure with specimens on gummed paper was as follows: Ten days after the coatings had been spread the specimens were trimmed to 3!/2 inches by 8 inches, as already described. Then a 2-inch length was cut off for testing before weathering. The 6-inch specimens remaining were mounted in the weathering machine and exposed for 400 kilowatt-hours (about 120 hours), after which an other 2-inch length was cut S for testing. A third 2-inch length was cut off after 800 kilowatt-hours of weath ering, and the remainder was available for testing after 1,200 kilowatt-hours, provided the coating was able to with stand that much weathering without cracking too seriously to be stripped and tested. Free films, after being stripped from the substrate, were cut into test specimens either 2 inches by 2 inches in size (for the first run) or \t/2 inches by 2l/j inches in size. It was sometimes necessary to depart from the preferred dimensions when very soft young coatings or unusually brit tle weathered coatings were tom or cracked in stripping or became cracked in weathering. As a rule, the lV^-inch dimension was the direction of strok ing with brush or doctor blade dur ing application and is therefore called the length of the specimen in this report. If the direction of stroking lay otherwise than parallel to the 1Vi-inch dimension, the direction was marked on the specimen with black drawing ink. Suitable reference marks were then ruled on the specimens in draw ing ink to record two places, approxi mately 1. inch apart along die length and to points approximately 1 inch apart along the width, between which length and width could be measured exactly before and after immersion in water. The specimens were then placed in a desiccator over calcium chloride for at least 24 hours to dry them thoroughly. The dried specimens were weighed with a duinomatic analytical balance. 'They were then sandwiched between iy2- by 3-inch microscope slides to hold them fiat while the length and width between reference marks were measured precisely with a traveling microscope. The lightest specimens weighed more than 0.12 gram, and the heaviest no more than 1.2 grams; weights were recorded to 0.1 milli gram. Length and width were meas ured to 0.001 inch. Specimens were then submerged in distilled water held at 69.5 F. in a thermostat. As a rule, one 3-mil and one 6-mil film of the same kind of paint were immersed in 300 ml. of water. Paints of different kinds were immersed in separate dishes because the water comes to a pH characteristic of the .kind of paint and the absorption and swelling vary with change in pH. Submerged films were weighted down beneath pads of coarsely woven spun-glass doth to keep some of them from coding too seriously. Except when otherwise stated, films were immersed for 72 hours, after which they were removed, sandwiched between microscope slides while still thoroughly wet, and the length and width were then remeasured with the traveling microscope. The wet films were men pressed lightly between pieces of dry blotting paper or of paper toweling to remove free water from the surfaces and reweighed as rapidly as. possible, always well within 1 minute after Wotting, with the - chainomatic balance. The- films were then allowed to dry in the laboratory for 2 or 3 hours and were then placed GLD37<5?1 in a desiccator over calcium chloride for at least 30 hours. The films were then weighed for the third time and, when the data were desired, the length and width were measured for a third time. There was, of course, an experi mental uncertainty in the second weighings while the swollen films were still wet owing to the fact that water was being lost by evaporation during progress of the weighings. A tedious procedure leading to a differ ent end-point has been proposed (21) for dealing with this difficulty, but the method used, which is similar to those of other careful workers (77) was considered preferable because the uncertainty was found to be a negli gibly small fraction of the differences among paints. The difference between the first and the last weighing of a film gave the weight of water-soluble ingredients extracted during the period of immer sion in water, represented hereafter by S. The difference between the second and the third weighing gave the weight of water absorbed by the film and still retained after loss of the water-soluble ingredients, represented hereafter by A. The swelling in length, a L, and in width, a B, was obtained from the differences between the first and the second measurements of those dimensions. The increase, in area of the marked portion of the film is represented by AQ. On re drying, most films shrank only part of 'the way back to their initial dimensions. None of the films ever reached a long period of constancy in weight and dimensions. The general course of events was a steady gain in absorp tion of water and swelling until maxima were reached, after which absorption and swelling declined. The 6-mil films of the two paints that contained zinc oxide attained their maxima relatively early, perhaps at 12 days, although the thinner film of titanium-lead-zinc paint continued to absorb and swell until the thirty- second day. The other paints and the clear linseed oil did not reach their maxima before the twentieth to thirtysecond day. Some of them gave the highest readings on die forty-ninth and last day of test, but it may be presumed that on still longer immer sion they, too, would begin to decline. On the other hand, the loss of watersoluble ingredients by films of all kinds and thicknesses continued steadily upward throughout the 49 days of test except for the pure white lead paint, for which the solubility was nearly constant from the twentieth to the forty-ninth day. When redried after immersion, the two paints con taining zinc oxide retained a large rt of the great increase in area they d experienced during immersion. Paints and oil that did not swell so much in water, on redrying, shrank more nearly, though seldom quite, to their original dimension; occasionally a zinc-free paint and frequently a clear oil film shrank back to less than its initial length and width (indicated in table 1 by minus signs). Loss of water-soluble ingredients readily ac counts for such shrinkage. The course of the absorption, swelling, and leaching in water sug gests no one period of immersion that is more significant than any other. For the rest of this study a soaking period of 72 hours was selected be cause it has been used by some previ ous workers, is short enough for rea sonable convenience in the conduct of the teste, and is not so long as to per mit unduly serious chemical changes in the films while they are being tested. 2. The temperature of the water in wbidi paint films are immersed markedly affects their absorption, swelling, and solubility, as is shown Table 2.-~Effect of temperature of soaking water on absorption, dwelling, and solu-~* bllity of films Hie films were 2 by 2 inches in size, of the thick ness indicated below, and were nearly 3 months old (stored indoors in a dry, dark place) when they were stripped from the substrates and tested by immersion in water for 72 hours. Preliminary Tests 1.Table 1 shows that free films of linseed oil and oil paints when continuously submerged in distilled water for 49 daw undergo changes all the time and fail to reach a condi tion of equilibrium. The five coating materials were selected to represent widely different behavior. For each material two large specimens on gummed paper were prepared, one. applied with the 3-mil and the other with the 6-mil doctor blade except for the clear linseed oil, which was applied by brushing. After 8 days for hardening, six small specimens were cut from each large specimen, marked, and dehydrated over calcium chloride. On the ninth day they were weighed and measured, and each set of six matched specimens was submerged in its own dish of distilled water held at 69.5 F. At each of the prescribed periods of immersion recorded in table 1, one specimen was drawn from each matched set to be weighed and meas ured while wet, redried over calcium chloride, and weighed and measured again. : Temperature j of soaking : water : Water absorbed (A) i: i Swelling Solubility (Increase i (Leached out i in area) t by water) t (AQ> : (s) *F : Milligrams { Percent s Milligrams Paint L, p/nv 0.30 (white lead paint)? thickness 6. mils, initial weight 1.207 milligrams 33 to 36 i 9 t 1.3 i 2.2 72 to 75 i 20 : 2.2 i 6.8 88 to 98 s 34 : 2,6 1 12.6 Paint TZoo, p/nv 0.30 (titanium--zinc paint); thickness 7,,7 mils. initial weighi 878 milligrams 33 to 36 103 72 to 75 i 369 88 to 98 * 519 s7 s 34 t 78 i 6.5 % 18.1 t 48.5 5 GL 037972 Table 3.--Effect of pH (degree of acidity or alkalinity)' of soaking water on ebaorptlon, swelling, and solubility of films The films were 2 by 2 inches in size, of the thickness indicated below, and were nearly 3 months old (stored indoors in a dry, dark place) when they were stripped from the substrates and tested by immersion in buffer solutions (citric acid and sodium phosphate) of the pH and concentration indicated} temperature 72* to 75* f. :: n it t Soaking solution it Mater it Swelling itSolubU- i Buffer ---------------------------- :t absorbed u (AQ) si ity i salts pH t Concentre- it (A) 11----------11(leached sabsorbed t tion of : t--------------------- :iln 19 tin 72 n out by iby film :buffer salts:tin 19 tin 72 tihours thours itsolution)t : tihours thours ti t it (S) i ------; -------- -----| t ---- I ------I |----I------II------ |---- t Moles per iiMUli-iMUli-ti Per- i Per- t: Milli- t Mllli- : liter ngrama igrams it pent t cent it grams t grams Paint L. p/nv 0,30 (white lead paint)i thickness> 6,0 mils, weight 1,067 milligrams 6.0 : 7.0 i 7.0 i 8.0 s 0.030 .033 .066 .036 :: n 1 29 SI 1.9 s 2.9 it 1.2 s 3.1 it 16 1 33 SI 1.6 t 3.3 it 8.5 t 3.7 ii u 1 43 SI 1.7 s 2.4 it 9.8 t 3.2 n 6 1 41 SI 1.9 s 3.6 it 13.4 t 3.2 Paint T2y>. p/nv 0,30 (titanium-ainc paint)} thickness14.7 mils, weight 5^4 milligrams 6.0 : 7.0 i 7.0 : 8.0 : .030 .033 .066 .036 :: 141 i: 90 : t 218 :: 184 i 242 t (1) i (T) (1) s: 22 SI 22 it 28 ts 27 s 33.2 :: 40 t U) SI 71 s a> ss 0 (i) SI 114 s 3.7 s* s 39.4 s 16.2 -In solutions of pH 7,0 or pH 8.0, films of TZ paint after 1'9 hours acquired such soft and tacky surfaces that they vers no longer manageahle for weighing and measuring while still wet. Table 4.--Loss in weight and shrinkage of free films of five kinds when allowed to stand 53 days in the laboratory in preparation for tests reported in'table 5 The films were 1.5 by 2.5 Inches in sise, of the thickness indicated, and were 7 days old when stripped from gunned paper substrates, marked, weighed, and measured, after which they stood 53 days before making final weighings and measurements. (Intermediate observations were made but are not tabulated.) Kind of film tilt iThick- ;Initial:loss ini Shrinkage in-- : ness t weight :weight :-------- ------ :of film:of films :Length t Width tlrea : Mils iHilli- iHilli- iPercentiPercentiPer: i grams tgrams t s scent Linseed oil and drier s 3.4 t 192 1 4*9 S 1.9 : 2.2 s 44 I 6.1 s 347 1 7.5 s 1.6 s 1.4 l 3.0 Paint L, p/nv 0.30 1 2.9 t 482 1 3.6 s .9 s .9 t 1.8 g 6a S 1,013 1 5.2 I 1.2 * 1.4 t 2,6 Paint p/nv 0.30 s 20. s 312 t 5.7 ' i 2.0 1 1.8 s 3.8 s 4i8 1 696 1 '9.5 % 2.7 1 2.5 s 5.3 Paint Z, p/nv 0,30 s 5 Paint TLj q Zj o , P/nv 3C s 1 2.2 s 5.5 s 2.9,' * 6*6 % 326 1 1.9 % .4 8 .7 s 14 800 * 3.7 % 14 8 1.1 * s 2.2 349 . 2.9 % .8 1 1.2 1 2.0 782 t .2 % .6 8 1.1 t 1.7 6 in table 2. In water at 88 to 98 F. (near body temperature) the changes in films of titanium-zinc paint were 5 to 10 times as great as they were in water at 33 to 36 F. (near freezing). For pure white lead paint, which is a more resistant paint to water, the effect of temperature was somewhat less marked, Ixat the ab sorption, swelling, and solubility at the highest temperature were still 2 to 6 times those at the lowest tem perature. It was therefore necessary m all subsequent tests to hold the water at constant temperature during immersion tests by keeping the dishes in a closely controlled water thermostat 3. The degree of acidity or alka linity, measured customarily in terms of pH, and the presence of soluble salts in the water in which paint films are immersed, greatly modify the ab sorption of water, swelling, and solu bility of paints that contain zinc oxide, although the effects are less on paints that contain much white lead. Table 3 records tests made with films of pure white lead paint and of a titaniumzinc paint immersed in solutions buf fered with mixtures of citric add and sodium phosphate to hold the pH at 6.0, 7.0, and 8.0 with approximately 0.03 moles of solute per liter, and at pH 7.0 with double that concentra tion of solute.. With white lead paint the absorption of water was slightly greater at pH 7.0 than at pH 6.0 or 8.0, but there was little or no effect on the swelling; solubility was very low at pH 7.0 and increased with pH, but it was still moderate at pH 8.0. A small quantity of buffer salts remained in the films after they were removed from die solutions and dried, the amount of which was determined by soaking the films in distilled water for 24 hours and then drying and weighing them again. The films of TZ paint were seri ously affected at pH 7.0 and 8.0 in a manner suggesting break-down of die gel structure. The films soon became softened, the surfaces became tacky enough to stick together where they came in contact, and there was obvious dislodgment of both pigment and vehicle from the surfaces. It was found possible to measure absorp tion and swelling only for the first 19 hours of immersion at pH 7.0 or 8.0, Swelling apparently did not vary gready with pH, but the absorption of water varied gready but unsystemati cally. Solubility increased greatly with pH, except for an anomaly in the more concentrated buffer solution of pH 7.0, where adequate differentia tion between leaching of film sub stance and retention of buffer salts in the film proved experimentally diffi- GLD37973 cult. The quantity of buffer salts re tained in the film after removal and drying was much greater at pH 7.0 and pH 8.0 than at pH 6.0. In another experiment a film of the TZ paint of initial weight of 483 mil ligrams was immersed in distilled wa ter for 23 hours, when the absorption was found to be 117.3 milligrams (uncorrected for loss by leaching), and the swelling was 20.7 percent. While still wet, the film was then immersed for 24 hours in water kept saturated with sodium chloride. The film lost all of its absorbed water together with a loss of 12.9 milli grams by leaching while in the satu rated salt solution, but it shrank back only part way, to a net gain over the original area of 4.3 percent. Only a few exploratory tests have been made so far on the composition of the soaking water after films of paints have been soaked in distilled water for 3 days. It has been tenta tively observed, however, that the soaking water after contact with pure white lead paint comes to pH 7.2, approximately, and contains only enough soluble lead compounds to be slightly discolored by hydrogen sulfide without precipitating an observable amount of lead sulfide. The soaking water from pure zinc oxide paint comes to pH 5.7, approximately, and contains enough soluble zinc com pounds to yield a copious precipita tion of zinc sulfide with hydrogen sulfide. The soaking water from pure titanium dioxide paint comes to a pH lower than 5.2 and is unaffected by hydrogen sulfide. The pronounced effect of pH on the behavior of many paints made it ad visable to use only distilled water for subsequent soaking tests and to avoid immersing films of more than one kind of paint in any one dish of dis tilled water, 4. To observe the rate at which films of coating materials change while kept under ordinary room con ditions, die five materials reported in tables 4 and 5 were applied to gummed paper, each in nominal 3mjl and in nominal 6-mil thickness. Films were stripped, cut to size, and marked after 7 days of diying, which was about the earliest age at which all of them were hardened enough to fable 5.--tff'fseoet of awing durlnr storage indoor# on the absorption of water, ewell&nit. reeorenr after ra&rklng. and loTubTUty o? Tllae of flvelttge-of eoatlnk MUtinr-"***------------- *-----*" Che fllne were 1.5 bp 2.5 Inahee In stce, of the thiokness lndieated, war# atripped from waned paper .7 aape after application, and war# than atorad for tba additional tlnae aeoordad in tba f:'lrat eoluan bafora tasting bp lauaralen in dlatillad watar at 69.5* T. tor 72 boor*, negative valuta for ewelllng after redrping aean that the fila shrank to laea than the area before insertion. ~rr Ties In atorege after : ----- Thinner fllaa tt -II- Thlekair fllaa stripping before teat bp inaaralon Water t Swelling 1Solubility!! absorbed t (Inereaee in area) I (Leaehod is s------------------------------ ! out bv' IS tin water l Redried S water! It t <*Q) I ? ) l IS) It Water t Swelling 1Solubility absorbedi (Znereaee in area) t (Leaehed i out by (A) I19Wi w1! nWaitrfr) S' !- Saga nwilliffraag! Peroent ijiruat iMWtiBMB.il < Proent Willigraae Thiokneae 3.0 nile. initial welgri9ZniuraM 1,>" 1**a~TH&lCkssiiTl ,u,, ltvLtltl 0 12252 3533 ItnS1IsSt1I,e0 28 20 17 si t5 1t 53 X 10 s -1.5 l 7.2 1t --1.*31 tt 5.8 6.7 I - .9 t * io.e ts iitt it 322752 228 30 %tt9t9 3 5 4 33 5 9 1 1 1t ------11112......994844 s 1t ss1 353 ,mirra 10.0 10.6 8.3 10.5 1112..92 Thlekneee 2.9 nils. Initial velaM. AiffeSi? <*11 WJiS'i. V alle. Initial weight 1023 ailligraae 0 5 12 22 33 53 SI IS It tt tt tt 27 20 19 3 15 t t t t s 9 6 7 6 i t t t *.42 0. 9 9 1 6.7 i6*.11 t - .2 t 6.3 t t - .7 -1.1 9 1 ?.? 6.4 it <tt* tt it u 33 3t 5 2257 24 1 t t* i k 3 3 3 9m 9-- 9 * 1 mo S-- t- 1 M 12.5 12.3 Paint Thzo. p/av 0.30 (anataae titanlua dioxide paint) Thiokneae 2.1 alia, lnlt-f--i-l---w--e--i-ght 3W--- --l-l-ll-g--r--a-ba * traftwaolnr' nlllUra.*. 0 tt 52 . 9 19 t 1.3 9 7.0 7* 9 10 5 tt 47 t 18 1 .8 9 2.6 67 I 10 12 22 33 53 tt tt tt tt 44 39 35 32 9 16 9 15 t 13 t 10 9 t 1.1 .7 9 9 56-.?* 0 9 7. K - .6 9 8 66 67 1 64 it 103 99 t t I t 10 - .6 - .8 - .3 - .2 - .5 0 10.0 11.1 10.3 12.6 12.1 17.1 Paint 2. p/nv 0.30 (1- oxide paint) Thlokneea 2.2 alls, initial weight 3x6 ml:lligraae Thleknseaj.S alia, initial weight T95 aliljgraas tl 68 t 40 t 3 t 5 tt 56 t 3f t 4 t 12 22 33 t: 62 9 3$ t 7 tt 61 S 36 1 7 tt 61 t 34 t 7 t 9 t 53 tt 61 t 31 t 6 I III 58 II251 230 230 40 217 37 25 14.7 2k 17 ii:| 3 12.4 16 ll'.l Paint TL20Zgp.- p/nv 0.30 .. ltanlw-lead-elne paint) Thlekneee 2.9 alia. initial waiint j!.i aiiiig-aai mekneei 1.6 MU, initial weight 791 allllgraaa 0 u 167 1! it 137 it 125 22 is 47 42 3258 st I 10.5 t 7.0 10 I 4.8 B 1 4.4 11 is St 79 26 24 5.4 I 4.9 lft- IS 1 18 t if t 11 \ 6 4 5 17.3 . 12.6 10.9 10.1 9.7 10.0 7 GLD37974 be processed without damage. Six matched specimens for each material and thickness were prepared, weighed and measured, and then allowed to rest on a table top near a closed window of the laboratory with northern exposure. From each set of matched specimens one was taken immediately and one after 5, 12, 22, 33, and 53 days, respectively, for test by immersion in distilled water for 3 days. All films except the 6-mil films of paint TLJOZJ0 lost in weight and shrank progressively throughout the period of storage before immersion in water. The 6-mil films of paint TLJOZJO gained slightly in weight for 33 days and swelled in area for 5 days before they began to decrease in weight and dimensions. All of the changes, however, were small when compared with the changes most paints undergo when immersed in water. Table 4 records the changes at the end of the 53-day storage period. Difference, among the coating matc- Tabl* 6. The rline were 1.5 by 2.5 inches in aise, of the thleknosa Indicated, and were prepared on gunned paper. After 10 daye for hardening, eaaplea were atrlpped for the iaaaraion teat before weathering and the rest were exposed to ultraviolet light and perlodlo water epray In a weathering aaohlne. further eaaplea ware atrlpped for laaerelon teate after 400, 600, and 100 kilowatt hours of oxpoeur* respectively. The weight of linseod oil in a 1.5- by 2.5-ineh flla 3.0 alia thlek, V0, le 171 eilllgraas; in a paint of p/nr 0.30 it le 120 allligraaa, p/nv 0.36 it le 107 allllgraae, p/nv 0.40 It le 102 allllgraae. Coating aatarlal t it TT sPlgaentti Thinner tlias it Thicker fllae i TOlUMBi i---- >tt- i in t tlnlttaUAbeorptlon of water, ttXnltialtAbsorption of water, A/Vo, t paint t tthlok-t A/Vo. after expoaura it thlok-iafter exposure for t p/nv it nose ifor lndloatod nuaber ofii note i lndloated raaber t it i kilowatt-hours i* i of kilowatt-hour* t It tot 400 t 800 l 120011 l 01 400 t 800 t 1200 t t...... t-----t--------1----------1--------i---I----!------------- t- *! Hiii > feroent of the oil In 11 Wile t feroent of the oil In flla i file, by weight n by weight Srcup 1 - flnplCTtnfd coating Material* Raw linseed oil Bodied linseed, Z3 Bodied llneeed, Z6 Alkyd reelu It 4.8 t tt 6.3 t it 6.4 t tt 4,1 I 5 7 11 8> 6 62 t 42 tl 6.5 > 21 I 11 22 it 15.5 4 t if! i33 * 16.0 l U10 i 13 >> 7.6 i 4 12 6 6 5 12 8 8 7 Oroup 2 - Linseed oil paints free frea sine oxide Hagneslua silleate Calolua oarbonate L L TL40 TL40 *420 *420 Average for greup S 0.30 tt .30 tt .30 it .36 tt .30 tt .40 it .07 tt .30 it tt 3.4 2.8 3.0 2.4 3.1 2.6 3.1 2.1 125 15 t. ...tt l! 25 26 uu13 29 t t 20 47 25 I 40 tt 74 13 X) t. tt 56 22 10 tt 26 23 tt 2Ttt 1 tt t..** tt 7.6 63 t 6.5 t 126 t 6.4 t 12 I 4.7 12 t 6.5 1 44 t 5.6 < 34 4.2 t 20 > 4.7 ST g: 17 10 U i 13 20 > 29 13 7. 20 I 13 32 < 12 4' 't 4 ` 23 l! 27 4 10 tt Oroup 3 - AUtyd-reeln paints Ware) Ware) Ware) t .30 tl 3.4 t 44 t 18 t 12 1 9 tt t .36 tt 2.9 t 57 t 32 1 13 t 10 tt t ..40 tl 2.7 * 79 1 29 t 14 1 13 tt 4.6 t 42 1 23 1 19 t 3.7 t 73 32 t 27 t 4.3 t 79 1 39 t 26 I 17 12 13 Oroup 4 - Llneeed ell Paints containing sin* oxide and of p/nv 0.30 Zt LZ37 t *450*30 t *430*25 *420*20 t t *410*10 t **20 Average for group 4 t t .10 tt .30 tt .30 St 30 tt .10 tt .10 tt .30 it tt 2.6 t 147 t 114 t 67 l 3.2 1 100 t 3.2 t 130 t f68t 1. t 43 t 3.2 t 130 t 81 1 ?4 t 3.3 127 t 61 t 41 t 3.3 t 154 t 57 t 51 l 3.3 & t t I34 t tt t I tt t s 67 It 28 tt 62 It 37 It 48 tt 71 tt J? tt 5.9 t 6.6 t t 6.8 t 6.7 t 6.3 t 6.8 t t 96 1 ti l ?8 1 86 1 96 t tt! 40 t 3* 40 t t 39 t 3? t S t t t 25 t 21 t 20 t 22 I 22 t 24 s tt t t Oroup 5 - linseed oil paint* containing tine oxide and of p/wv 0.30 or OtM *430*25 *420*20 5-10*10 nao **20 ...ttnit 2.6 141 85 70 I, ti 2.7 148 46 it 1 8:It 2.7 142 62 45 tl & tt 3.0 178 3.0 t 173 t 82 81 I 70 t 6..1. tt 67, t..4..ll *s5.6 t 5.6 t 6.1 t 82 I 5.9 t 106 t 5.8 t 1X3' * 47 1 47 t 3577. I t 57, t 26 26 ni 44 t Average for group 5 t I50 t 78 t 61 I 51 It I TT t Jfl JJt 20 23 15 19 16 16 tt 21 24 11 4* 3l *0. 1 - & f: iSfliS Wo. 5 - *i3jfr*J Oroup 6 - Ooaaerolal palate 1 a; 9i if.35 it 2.7 .31 It 3.2 74 t 62 t tt 12 tt till.30 tt 2.8 1 118 t TO t 82 t .38 tt 2.8 t 145 1 88 t 8! n\6.11 6.21 5.71 100 6.0 > 155 41 t 28 21 81 t li .31 it 1.8 t 73 39 t 46 tt 3.7 33 26 It__ ___ t _____t ____ I tt _____ t_ ____ t 8 28 22 21 27 GLD37975 rials were relatively small. In another preliminary test it was found that films of paint L and of paint TZ20 were still losing' weight very slowly after 8 months in storage. Both of these films increased slightly in weight for about 50 hours after they were applied and then began to fall off steadily, but- the loss after 8 months was only 27 milligrams for paint L and 17 milligrams for paint TZJ0. Table 5 records the results of im mersion tests made with the films after different periods of storage up to 53 days after stripping from the sub strate (60 days after applying the coatings). Absorption of water and swelling diminished steadily for at least the first 3 weeks. After that time films of dear linseed oil and tire thicker film of titanium dioxide paint began to increase again in absorption and swelling, whereas the other films continued to dedine. The solubility of the films and the dimensional changes on redrying after immersion did not change very much for any of the films, although they diminished slightly with age for the films of zinc oxide and of titanium-lead-zinc paint. On the whole, differences among the paints at any age were much larger than the differences Table 7.--Swelling (increase In surface area) of free films of different coating materials during 7fz-hour immersion m distilled water at F.. both before and after exposure of "the Tllisa'to artificial weaCHerlnir----- ----------------------------- The measurements of swelling were made on the same films and at the same times as the measurements of absorption of water reoorded In table 6. By way of comparison, flat-grain surfaces of southern yellow pine boards shrink 6.2 percent (practically all aoross the grain) In drying from the green condition to 6 percent moisture oontent. Most other softwoods shrink less than that. Coating material sPigments Swelling (increase in area)4Q, after exposure for lndloated svolume : number of kilowatt-hours s in :------- : paint s Thinner films Thloker films : p/nv s------- ::--------------- 400 : 800 1.200 st 0 s 400 s 800 1200 s----------- :------------: -------------- : ------- --s------- ---- ----- s--------------- sPercentsPercentSPercentspercentt sPeroentSPercentSPercentt Percent Raw linseed oil S Bodied linseed,Z3: Bodied linseed,Z6t Aikyd resin s Group 1 - TJnplgaented coating materials ?2 n 6 5 :a s3 :3 4 1l3 3 t3 10 * s 8 10 s3 12 SS 3 SS 2 ts 4 s : t t 3 3 2 3 4 4 3 3 2 s s t 6 3 3 3 Linseed oil paints free from tine oxide Magnesium silicate 0.30 : 27 $ 4 s: 16 t 3 t 8 t 10 Calcium carbonate L .30 S 26 .30 t 8 sn S2 : '4 3 1 s X 4 1 SS 5h XI 5 s s 9 2 s 2 2 s t 2 2 L TL40 TL40 T420 Tl20 Average for group 2 .36 s .30 : .40 s .09 : .30 : 8 27 18 13 18 Iff s3 s4 s6 :8 s s "5 s2 3 s4 :? 3 * "3 31 t*_s 3. . s s s 4 15 8 13 10 Iff s s t l 5 2 4 6 7 2 s s 2 2 3 6 1 t t t 2 2 4 4 1 Group 3 - Alkyd-resln paints T(are) T( are) T(are) .30 s 9 t 6 .36 : 12 t 6 .40 : 12 s 4 3s3 3s3 2: 2 S 10 1 5 S 4 s 10 s 4 s 4 : 9s 4 t 4 Group 4 - Linseed oil paints containing sine oxide and of p/nr 0.30 Z t .30 : 70 S 43 S 21 t 16 s: 33 13 8 6 LZ3? TL50Z30 Is .30 : 42 S 29 t........... . s......... , si 27 : .30 s 49 : 28 S 20 t 16 st 30 11 11 7 6 TL30225 .30 s 49 J 32 1 18 t 19 t 27 12 7 9 TL20Z20 t .30 : 40 : 19 s 13 13 SS 26 9 6 4 TLiozio S .30 : 47 t 1? t 16 12 ts 29 9 6 5 TZ20 s .30 S Average for group 4 : ! I 29 I 25 S 19 ! 17 SS 40 s: 3ff Tft 10 Group 5 - Linseed oil paints containing tine oxide and of p/nv 0.30 or 0.40 TL30Z25 TL20Z20 TLioZyo TZ20 TZ20 Average for group 5 Ho. 1 - TL7 Z23 S Ho. 2 - (TL23Zl8)l3k Ho. 3 - (TZ19)126 Ho. 4 - (TZ16>122! Ho. 5 - Ti39(re) 1 .36 .36 .36 .36 .40 .35 * 31 * .30 s .38 : .31 46 S 27 s 13 s. 45 s 25 s 15 1 14 . t s 38 s 18 t 12 I 11 SS 56 : 31 t 23 3 21 s: 8 t *4 3 3 ; t 3 33 Oroup 6 - Commercial paints 38 s 24 X 17 *4 14 33 35 s 18 X 11 t 10 33 26 t 20 t 17 t 21 st 55 * 38 t 29 t 27 33 3s 2t1 I1s 18 22 19 36 $ 24 19 17 49 1 12 s 12 ;8 t 18 s 20 1 tit JS 10 t8 s8 * 26 1 7s6 7t7 5s4 15 t' 15 14 Iff s ft s 18 s 9 f *5 s7 22 s 19 1 1 0.3 GLD37976 Table 8.--Swelling efficiency of water for the free film reported in tables b and 7 The swelling efficiency is the swelling recorded in table 7 divided fey the absorption of water per unit weight of oil in the film recorded in table 6, that is, AW j A/H0. >s Coating material or group!: Average swelling efficiency for the group after of materials as dassi- 11 exposure for indicated number of kllowatt-hoira fi*d in table 6 and 7 It* It Thinner film* 1 1 Thicker film* It" 1 t 0 1 400 800 l 1,200 it 0 400 i 800 i 1,200 SI Percent It Percent Group 1 - Unpigmented II materials It so 61 Group 4 - Zinc-containingi I paints of p/nv 0*30 It 39 1 39 Group 5 - Zinc-containingit 1 paint of p/nv 0*36 or II t 0.40 It 31 I 33 Group 6 - Commercial It line-containing painta II 28 30 Group 2 - Zinc-free SI paints other than uhitett lead IS 34 I 30 1 t1 27 i 29 n 72 1M 35 i 35 it 37 t II t i: 27 t 29 it 27 1 It 27 i 27 it 2$ I II t It 23 t 28 i 38 tI 48 i 46 t 43 11 31 i 31 I 35 1I II 27 i 29 30 1t 25 t 25 t 27 tt lt 29 i 31 29 Group 3 - Alkyd-resin paints II It 19 * 22 White lead paints, L p/nvu 0.30 and 0.36 II 43 I 9 Commercial paint Ho. 5 : ("breather type") II 3 4 1 21 t t 6i t 2i It 26 ti 16 II 5 it 38 II 2 I1 14 i 17 22 1 18 i 13 15 1 3* 4i 1 brought about by aging of any one kind of paint. For this subsequent studies, films were allowed to harden for 10 days after application before they were stripped from the substrate and were then submitted to immer sion tests within 3 days. Tests of 32 Oils and Paints Before and After Artificial Weathering Large specimens of the 32 coating materials on substrates of gummed paper were prepared, each material in nominal 3-mil and 6-mil thicknesses. After 10 days for hardening, a small sample was cut from each large speci men to be stripped and tested immedi ately by immersion in water. The re maining portions of the large speci mens were placed at once in the arti ficial weathering machine, operated with periodic spraying with water. Each large specimen was further sam pled for stripping and testing after 400, after 800, and after 1,200 kilowatt-hours of exposure (approxi mately 5, 10, and 15 days of con tinuous operation). A second com plete set of specimens spread on tin plate instead of gummed paper was exposed in the weathering machine at the same time in order to deter mine the weight of soluble material leached from each specimen by the water spray during the course of the artificial weathering. Absorption of Water: The absorp tion of distilled water on 72-hour soaking of the free films before and after each stage of artificial weather ing is reported in table 6. The weight of water absorbed by each free film is expressed in percentage of the weight of linseed oil in the film, that is, 100A/Wo, where We is the weight of oil in the film. That is considered a more significant basis for comparing the different paints and film thick nesses than the absorption in milli grams of water. A, because absorp tion and swelling are believed to re side primarily in the oil in the film. It would be very misleading, of course, to express the absorption in percentage of the total weight of film in view of the great variation in specific gravity of different pigments. Thus the weight of a free film 1.5 inches long, 2.5 inches wide, and 3 mils thick is approximately 171 milligrams for unpigmented linseed oil, 270 milligrams fot calcium car bonate paint of p/nv 0.30, and 499 milligrams for white lead paint of p/nv 0. 30. 10 All of the free films tested absorbed distilled water in significant quantity. For all of the paints in which there were pigments, the newly prepared and unweathered films absorbed much more water (ban the artificially weath ered films did. The first 400 kilowatthours of weathering usually effected greater reduction, in swelling than the next 400 kilowatt-hours did, but the last 400 kilowatt-hours produced so little further change as to suggest that a minimum had been reached. Some films, in fact, absorbed more water after 1,200 than after 800 kilowatt-hours of exposure. Films of dear oils without pigments, with the possible exception of the bodied lin seed oil of viscosity Z6, passed through a minimum of absorption after 400 kilowatt-hours and usually absorbed more water after 1,200 kilowatt-hours than they did before exposure to weathering. The bodied linseed oils initially were much less absorptive than the raw linseed oil, but after weathering the differences became small. The alkyd resin was only slightly less ab sorptive than the bodied linseed oil of viscosity Z6. Paint made with magnesium sili cate in raw linseed oil absorbed more water than the unpigmented oil did, especially so before the films were weathered. On the other hand, the calcium carbonate paint, although high in absorption at the outset, became less absorptive than unpigmented oil after weathering. Pure titanium dioxide paint, Tno, was initially slightly more absorptive and, after weathering, slightly less absorptive than raw lin seed oil. Pure White lead paint, L, at the outset had a lower absorption than any other paint and a lower absorp tion than raw linseed oil. Weathering produced relatively little change in the absorption of pure white lead paint, so that, in the end, it still re mained less absorptive or, in the thicker film, only slightly more ab sorptive than raw linseed oil. Zinc oxide, on the other hand, whether the sole pigment in a paint or present to the extent of as little as 10 percent of the total pigment by volume, very greatly stimulated ab sorption of water by free films, par ticularly when the films were unweath ered. Thus pure zinc oxide paint, Z, and titanium-zinc paint, TZ,,,, ab sorbed about 5 times as much water initially as raw linseed oil did, and remained 2 to 4 times as absorptive after weathering. The presence of white lead along with zinc oxide usu ally reduced the absorptiveness some what, but the stimulating effect 'of zinc oxide was dominant over the GLD37977 moderating effect of white lead even when there was 1.7 times as much white lead as zinc oxide in the paint. Paints of pigment volume 0.36 or 0.40 usually absorbed a little more water in proportion to their oil con tent, although slightly less in propor tion to their volume than the corre sponding paints of pigment volume 0.30, despite the fact that much of the oil in the former was bodied oil, whereas all the oil in the latter was the more absorptive raw oil. Of the five commercial paints tested, the four that contained zinc oxide proved high in absorption, comparably so with the laboratorymade paints of similar composition. The fifth commercial paint, advertised as resistant to moisture blistering, was decidedly less absorptive than the others but was more absorptive than the paints made in the laboratory with no lead or zinc pigments and alkyd-resin vehicle. Swelling Ini Waters Table 7 re cords the swelling, AQ (increase in area of surface), of the free films paused by the absorption of water reported in table 6. In general, the swelling runs approximately parallel to the absorption of water, as would be expected. The relations already dis cussed! for absorption therefore apply also to swelling, except that commer cial paint No. S exhibited dispropor tionately low swelling for the amount of water it absorbed. Possible reasons for the peculiar behavior of this paint are discussed further on. The swelling of the free paint films may well be compared with the swell ing of wood on which sudi paints might be used. Flat-grain boards of southern yellow pine change 6.6 per- fable 9.--Solubility of fraa fllas of different posting aarn?MTT-*'n-r rjrsw-vw-i* llntia oil In the flli during 72~hour lmirilon In aoaeuronente qf abeorptlon n percentage of the weight Coating Material _ i j........ tPlgaentst :volume tt-- : In tt Solubility In water, B/Vo, after exposure for l.n...d..l.o...a...t.e..d..number :f kilowatt-houra thinner fllae It Thicker filaa Pp/nr tt 0 """40o""T"ioo"s 1200 It ~ 0 t" WO 1 800 t 1200 .j------------tt-----------1------------ 1------------ 1--------- --It---- --t------t----t- ! t tPeroarttPeroenttPeroenttPeroentiIParoanttParoanttParoantt Paroant group 1 - Onplgnented posting aatarlala Daw Unseal oil t Bodied llnaaad. 23 : Bodied llnaaad, Z6 s Alkyd-resln t 0 0 0 0 It. 7.0 t. 9.6 t 10.9 tt 3.9 t 2.5 t 2.0 t tt 4.0 t 2.6 t 4.7 s 7.0 tt 1.4 l 1.2 t 2.0 t tt 3.4 t 2.3 t 3.9 t ?.? It 1.0 t .9 t 1.7 t tt 7.0 t 2.9 t 3.4 t 4.4 It 3.2 t .8 t 2.1 t group 2 - Linseed all palnta free froa tine oxide 2.6 2.3 Magnesium allieate t Calolua carbonate t Lt Lt 0.30 tt 7.5 5.8 .30 :: 14.5 ll.l .30 tt 4.7 t 5.8 .36 tt 6.8 t 4.9 5.9 6.6 10.7 ............ it 7.1 it 9.0 tt 12.0 tt 3.6 9.7 1:1 2:5 1:1 5.5 !:1 4.3 5.6 2.7 2.9 4.7 XL40 * .30 it 5.4 t 6.6 3.3 ............tt 4.4 2.8 2.2 2.0 1LU0 *420 * .40 it 6.5 t 9.8 * .09 it 7.4 l 5.5 2:1 2:1:8 ............tt ............tt 3.T 5.0 4.8 3.0 *420 * .30 it tf i Average for group 2: tt 2$............tt t* lO 3.5 t 5'j00 376 t 3TB group 3 - Alkrd-reeln paint* t(are) T(ara) T(ara) t .30 ti 8.3 t .36 tt 5.6 t .W tt 4.9 4.8 5.3 5.3 5.6 6.0 6.2 3.9 it 4.2 6.8 tt 3.2 7.0 tt 3.2 3.5 t 4.5 t 3.0 t 4.3 t 3.3 t 4.5 t 3.2 2.9 .7 group 4 - Llnaeed oil palnta containing alwo oxide and of P/nv 0.30 2 4237 1J50Z30 W.30Z25 TL20Z20 IL10Z10 *Z20 Average for group .30 tt 11.2 t .30 tt 6.3 t .30 it 5.8 t 3.9 16.0 tt 3.5 2.7 2.1 1:2......it 6.4 3.0 i t 3*^ 1.5 3.2 l 2:1:: s y7.5 5.3 tt 2;? 2.7 3.2 '.6 4.0 3.2 t 2^5 t 3*7 t 2,5 SaO ayag-i-j lAnmmmA oil palnta containing Uno oxide and of P/nv 0.36 or 0^40 TL30Z25 TL20Z20 TLlOZlO T220 TZ20 Average for group 5 .36 .36 tt .36 tt .36 tt .40 tt ss 5.1 t 7.3 4.5 t 7.5 5.4 t 6.6 t 14.9 * 7*2 t o77 t 7.8 t 10.0 t 8.7 t 10.9 > 10.5 t 976 t. ......... II t 8.0 tt t 4.5 tt t 13.2 tt t....... tt t 876 tt 2.2 2.1 2.7 1.5 2.1 2.2 group 6 Ooaaerolal palnta Mo. 1 IL7 Z23 t .35 it 2^.5 t 9.1 7.8 9.6 tt 5.1 3.5 3.3 4.2 Mo. 2 11 nMo. 3 Mo. 4 J2rz3rl8,134 *219>126 * <2216)122 .31 6.6 s 7.9 .30 tt 12.1 t 13.8 .38 si 23.6 s 19.0 4.1 7.9 25.6 4.3 tt 12.3 tt 31.8 tt 12.4 2.1 5.2 7.2 1.7 4.4 6.4 12.8 Mo. 5 .31 * 9.2 t 11.4 9.3 16.1 tt. 6.0 6.4 6.8 GL03797B tent in width (across the grain) and practically not at all in length (along the grain) between 6 percent moisture content and the green or thoroughly water-soaked condition. The moisture content of exterior lumber is seldom as low as 6 percent. The change in area is therefore 6.6 percent,of the dry dimensions or 6.2 percent of the green dimensions. Most other woods used on houses swell less than south ern yellow pine. Thus free films of most paints when newly prepared swell more than wood does on chang ing from the air-dry to the water- saturated condition. After weathering for some time, zinc-free paints usu ally swell less than wood does. Paints containing zinc oxide swell several times as much as wood does when the paints are new, and swell as much or more than wood does even after they have become well weath ered. Of course, the coatings of paint must conform to the wood in length and width as long as the coatings re main in place, and to do so the coat ings must change sufficiently in thick ness to meet the required volumetric changes, a process that necessarily in volves severe internal stresses in a rigid body. To check that conclusion, dry 1/24inch yellow-poplar veneer was painted in one case with two coats of paint L, p/nv 0.30, making a coating 3.6 mils thick when dry, and in another case with two coats of paint TZ^, p/nv 0.30, making a coating 4.8 mils thick when dry. Suitable specimens were marked for measurement of length, width, and coating thickness (by ob servation of sharply cut edges in a microscope with micrometer eyepiece) and were then soaked in water for 2 days and remeasured while still wet. In surface area both paint coatings swelled 10 percent, chiefly across the grain of the wood, which is the ex pected dimensional change for thin veneer of yellow-poplar. In thick ness the coating of paint L shrank 0.2 mil, but that of paint TZa0 swelled 2.7 mils. The result was qualitatively in agreement with the data of table ,7, which give the swelling in area as 8 percent for paint L and 57 percent for paint TZao in free films about 3 mils thick after soaking 3 days. At the end of the soaking period the coating of paint TZ20 was beginning to form tiny blisters, but that of paint L was still entirely sound. Among the paints reported in table 7, paints L, TL<0, and commercial paint No. 5 are known from past tests on laboratory blistering boxes and from experience on houses to be highly resistant to failures by blister ing and peeling under conditions of ' so-called excessive moisture. Work re ported by a large paint-industry lab oratory indicates that the leadless and zincless alkyd-resin paints, T(arc), are similarly resistant to such condi tions. All of these paints are charac terized by a low degree of swelling in water. On the other hand, all of the zinc-containing paints, which have a high degree of swelling, are known from past experience to be sensitive to moisture blistering and peeling. It may well be, then, that the degree of swelling in water is a useful measure of the ease with which a paint yields to blistering and peeling of this kind. During the immersion tests the free films of those paints that exhibited a low degree of swelling remained fiat and free from troublesome curling. Even after redrying they could be kept flat provided the absorbed water was allowed to evaporate equally rapidly from both faces of the film. In contrast, free films of paints that swelled greatly always had a marked tendency to curl, particularly near the edges of the films, and required very careful han dling to straighten them but between glass microscope slides sufficiently to measure the dimensions. Such curling became still more marked during redrying. These differences in behavior correspond to the characteristic per formance of the paints during the later stages of normal weathering. Once cracking sets in, paints that swell greatly also curl markedly at the cracked edges and tend toward scaling in large pieces, whereas paints of low degree of swelling, if they crack at all, usually curl very little and tend toward crumbling or flaking in tiny particles. Extreme curling during immersion in water was exhibited by heterogene ous films made up of consecutive coats of paints of widely different swelling tendencies. Thus free films were prepared by applying in one case two coats of paint , p/nv 0.30, fol lowed by two coats of paint TZ,0, p/nv 0.30, and in the other case two coats of the TZj0 paint, followed by two of paint L. When the free films were soaked in distilled water for a few days, both curled into tight rolls much like rolls of wrapping paper. The two films curled in opposite di rections, so that in both cases the highly swellable TZ paint was on the outside of the curled roll and in both cases the direction of last brushing of the paints during application was par allel to the axis of tne roll. After the films were soaked for a week, the twb layers of paint in each film began to separate and the TZ layer became wrinkled until eventually it was an easy matter to pull the two layers com pletely apart. 12 Some Paints Possess Grain: When free- films swell while immersed in water, the increase in width, a B, usu ally differs from the increase in length, AL. (The length is always the direc tion in which the brush or doctor bktyie was last moved' during applica tion.) The ratio a B/a L is not exactly reproducible quantitatively among dif ferent films of the same kind or even in different sections cut from the same large film; slight variations in the man ner of application or in the way the film is immersed in water apparently exert measurable effects. But such vagaries are small in comparison with certain large differences among paints according as they do or do not contain pigments of strongly anisotropic shape. Acicular (rod-shaped) pigments evi dently become oriented with their long dimensions parallel to the last direc tion in which the liquid paint flows before it hardens, and they retain their orientation in the hardened film even after prolonged weathering. Such ori entation of acicular pigments imparts grain to the paint films that is roughly analogous to the grain in woocT It has long been observed in paint ex posure tests at the Forest Products Laboratory and elsewhere that many paints, including enamels that form coatings entirely free of brush marks, crack on weathering with the predomi nant direction of the cracks roughly parallel to the direction in which the paint was last brushed. Since the last brushing of paint is normally parallel to the grain of the wood, it has usu ally been assumed that the grain of the wood controlled the direction of paint cracking by reason of the greater swelling of wood across than with the grain. The discoveries that many paints swell more than wood does, swell more across than with the last direction of flow, and therefore possess a grain of their own, confirms this Laboratory's earlier observations that it is the grain of the paint rather than the grain of the wood on which the direction of paint cracking commonly depends. The observations of grain in free films may lbe briefly summarized by the following list of approximate ratios of transverse to longitudinal swelling, AB/a L, which are averages for the number of closely agreeing paints or coating materials indicated: Kind of costing material a b /a l Magnesium silicate In Homed ofl............ 2.10 five commend*! paint* (alt contained magnesium silicate)............................ . 1.87 FaintT(are) with all extender magnesium ~ silicate.----------------- -............. -........ - 1.78 Berea TLZ nutate with all extender mag- aeetao* eillcate.........-........................ . 1.60 Calcium carbonate (No. 1) la Hweeeii oU... 1.40 Paint TCare) with half the extender mat* neelum silicate led halt etldtua car bonate (No. 8)..--.--............. 1.45 Three T2* paint* with half the extender magnesium silicate and carbonate (No. 2).......... !h7a.lf...c..a..lc..i.u..m..... 1.1T GLD37979 Kind of coating materia] AB/AL Paint* L, Z. Ttt*. and LZ* with no extender______ ____________________ 1,12 Paint T(are) with all extender calcium carbonate (No. 8) .96 Pour unpigmented vehicle* (Unaecd oil* and alkyd reain)................... -......................... M Magnesium silicate (fibrous talc) is a markedly acicular pigment; it dearly imparted grain to all the paints in which it was used, including all five of the commercial paints. Calcium car bonate No. 1 was a pigment of very high oil absorption and an appreciable degree of acicularity that imparted sig nificant grain to paint; but calcium car bonates No. 2, low in oil absorption, and No. 3, very low in oil absorp tion, were granular and imparted little or no grain to paint. All ofthe opaque pigments used, white lead, zinc oxide, and titanium dioxide, were effectively granular enough to produce little or no grain in paint, though some of them are available in somewhat acicu lar grades. Swelling Efficiency of Water: In order to study the relation between degree of swelling and the quantity of water absorbed per unit weight of oil in the films, a factor that may be called the swelling efficiency was com puted for all of the films listed in tables 6 and 7. The swelling efficiency is defined as the swelling in surface fable 10.--l.oaa In weight by leaohlng with water apray during txpoaura of aoatlnge on tinplate " to artificial weathering, expreaeed ae percentage or the weight orllnaaeBT In the fCSii - ' .... Coating aterlal --------------57 TT Plgnent *: thinner fllaa It Thloker film* voluae 1------------------tt----------- in : :Thlok->Loee by leeohlng per t1 Thiele- tLoaa by leaehlng per unit paint :: neee tunlt weight of oil tvclght of oil after p/nv it tafter expoeure for texpoeure for lndloated ti lIndicated nuaber of tnuaber of kllowatt-houre :: I kllowatt-houre t si t~ 4oo~t ~8oo~t~;zoo t WO I 000 I }200 :i Mila tPescenttPeroenttPeroentt Mile tPeroentt Percenti Percent Oroup 1 - Pnplgaented ooatlng material# Raw llneeed oil t0 Bodied llneeed. Z3 t 0 Bodied llneeed. Z6 t 0 Alkyd-reeln 10 It 6.1 t 15.9 t 22.2 t 28.1 tt 12.3 t 10.3 t 14.7 It 5.8 t 9.9 t 12.7 t 16.3 tt 12.7 t 5.7 1 7.6 tt 6.0 t 7.6 t 10.0 t 12l7 t 13.4 t 5.1 t 6.2 tt 6.3 l 4.2 t 6.2 t 6.4 tt 11.8 t 2.9 t 3.9 $ 1. N I Llneeed oil palnte free fro* sine oxide Magneslun ellloate Celolun carbonate t t 0.30 .30 tt IS 4.8 I 13.7 t 21.8 3.1 t 18.3 t 31.2 t 28.4 1 39.8 tt tt 7.7 t 6.2 t 9.8 9.8 t 14.2 t 15.7 L L t l .30 .36 tt It 4.2 t 13.0 X 21.1 4.2 1 10.2 1 14.4 2 27.0 t 18.8 it tt 7.2 t 5.0 t 2*2 6.8 t 11.9 I 12.5 TL40 t TL40 t T420 t *420 t Average for group 2 t $.09 -5S it*.30 tt tt It tt tt 4.9 t 14.0 1 23.7 4.0 t 17.4 l 31.4 4.2 t 17.4 t 29.8 5.9 > t 12.5 14.6 * jjgy t 31.0 tt 7.7 t 7.6 t 39.0 n 5.2 t 12.9 t 41.2 u 6.3 t 11.5 t 21*8 tt 8.4 1 t yszx tt t t 13.0 t 23.2 t 20.6 t t t 17.8 t 9.1 t 7.8 t 3.6 t l8.2 t 20.2 t 16.2 t 16.2 1 16.8 t 29.5 t 29.0 t 13.4 t 197? T( are) T(are) Ware) ffro.HF.J I .30 tt 3.4 t 5.6 t 7.8 t 11.8 tt 6.9 t 1.? t 1.4 t t .36 It 4.3 t 5.1 t ?.4 t 7.3 tt 6.2 t 3.6 I 5.3 t l .40 tt 4.2 t 5.3 t 8.3 t 10.3 tt 6.1 t 4.0 t 5.6 t Oroup 4 - Llneeed oil palnte containing alnc oxide and of p/nv 0.30 1.9 6.6 6*6 Z LZ37 IL50Z30 TL30Z25 TL20Z20 TL10Z10 TZ20 Average for group 4 .30 tt 2.81 8.9 1 19.9 1 26.5 tt 6.8 t 3.0 t 5.8 s 8.1 .30 it 4.61 -3.9 t 8.4 t 11.6 tt 8.2 t 1.3 t 4.0 t 5.1 .30 tt 5.01 4.8 t 6.1 t 7.6 tt 8.4 t 2.3 t 2.6 t 3.1 .30 u 5.41 4.8 t 7.0 t 9.3 tt 7.6 t 2.8 t 4.0 t 5.5 .30 n 4.71 5.3 t 6.8 1 8.9 tt 7.9 3.2 t M t 3.9 4 fit .30 tt 3.9 * 8.1 < 11.5 t 14.6 tt 6.5 * .30 tt 4.11 t 14,4 tt 8.2 t tt t t t 1377 it * t 5.8 t t t t t Oroup 5 - Llneeed oil palnte containing alnc oxide and of p/nv 0.36 or O.W 5*30^25 2-20*20 5;10z10 W20 TZ20 Average for group 5 .36 u 2.9 S 11.0 t 19.8 t 25.6 lie It It 4.1 t 3.6 t 2.1111.0 6.9 1 12.8 113.5 I 16.0 It 4.0 t 6.1 11.3 * 14.0 11 4.0 I 6.0 l 12.0 I 14*9 It I 578 t 1178 l 15V8 Oroup 6 - Ooaaerolal palnte 1:5 li fe* Mo . 2 - p&ttllBllSM Mo No Mo l " (TZl6)122 * Mo. 5 tl39UrJ .35 .31 .30 tt 11 tt 3.8 s 4.6 s 3.6 t 6.1 t 12.9 5.0 1 8.6 8.0 t 8.3 I 18.8 t 12.1 I 12.5 tt St ts 6.5 t 7.0 t 6.4 t 2.8 1.8 2.4 t t t 5.5 2.5 3.5 .38 tt 4.2 t 4.9.* 9.8 I 14.5 tt 7.0 t 2.3 t 4.5 .31 u tt 2.3 t t 4.5 t 14.2 t t 18.3 t tt 4.0 t 4.9 ts 9 t t 9.8 s 8.1 t 5.6 t 5.5 i .1.9. t 11.4 X 13 GLD37980 area, a Q, divided by the quantity of water absorbed per unit weight of oil in the film, A/W0, that is, a QW0/A. Examination of the tabulated data then showed that the coating materials fell into certain groupings within which the swelling efficiencies were essen tially the same. The groupings of coat ing materials were those that already appear in tables 6 and 9, except that the pure white lead paints in group 2 and commercial paint No. 5 differed from the other paints in their groups. Accordingly, table 8 lists the average swelling efficiencies for each group with the two exceptions omitted and listed separately. All materials except commercial paint No. 5 and the alkyd-resin paints were swelled more efficiently by water when the films were freshly prepared than after weathering. Most of the re duction in efficiency from weathering was accomplished in 400 to 800 kilo watt-hours of exposure; often the effi ciency after 1,200 kilowatt-hours was slightly higher than it had been after 800 kilowatt-hours. For the alkydrcsin paints, the initial swelling effi ciency was lower than that of other paints except commercial paint No. 5, but the swelling efficiency for the alkyd-resin paints increased slightly after weathering. For thick films, the swelling efficiency was nearly always slightly lower than it was for thinner films, but the diffetence largely dis appeared after weathering. Unpigmented films were subject to higher swelling efficiency than other films when freshly prepared and, in the thicker films, throughout the weathering periods. For zinc-contain ing paints of pigment volume 0.30, the efficiency was slightly higher than for otherwise similar paints of pig ment volume 0.36 or 0.40, but the differences in swelling efficiency were small among the zinc-containing paints of the higher pigment volume, the commercial zinc-containing paints, and the zinc-free paints other than weath ered pure white lead paint. Conspicuously unusual results were obtained with pure white lead paint and with commercial paint No. 5. For the latter, the swelling efficiency of water was slight at any stage in life. For the former, the swelling efficiency was as high as for other paints when freshly prepared, but on weathering the swelling efficiency for the thinner films of paint L quickly dropped to values nearly as low as those for com mercial paint No. 5. The thicker films of white lead paint on weathering were swelled more efficiently than the thinner films but were far less swellable than other paints. "Breather-type" paints, such as commercial paint No. 5, are said to be physically porous by reason of voids in such pigments as diatomaceous silica or ot interstices left unfilled by the vehicle. If so, they have a capacity for holding free water or for taking up part of the swelling of the oil internally without change in external dimensions. It is conceiv able that white lead paint becomes physically porous on weathering be cause it is known that once it has passed well into the chalking stage the surface can no longer be restored to a glossy condition-by rubbing as can be done with most other paints. Mac Gregor (22) postulated the develop ment of such porosity during tne weathering of some paints, although he did not believe that it occurs in white lead paint. At any rate, the low swelling efficiency of commercial paint No. 5 and of weathered paint L indi cates clearly that such paints take up water in a manner that causes less swelling than is brought about in other paints. Thickness of Free Filins: In calcu lating the swelling efficiencies in table 8 it is tacitly assumed either that the thickness of free films docs not change during swelling or that the thickness always changes proportionately to the area. The first assumption is certainly incorrect. The second may be true within reasonable limits but must be considered uncertain. For that reason some of the smaller differences in swelling efficiency recorded in table 8 may be misleading because they may be due chiefly to variations in the participation of thickness change in the total volumetric swelling. It was hoped that it would be pos sible to measure changes in thickness *as well as in length and width in these experiments, but all techniques tried so far for measuring swelling in thickness have proved too inexact for quantitative purposes. Plans for future work therefore include direct measure ments of volumetric swelling of at least some of the paints ana unpig mented vehicles. All film thicknesses reported in the tables were calculated from the weight of the film, its area, and the density of the nonvolatile ingredients of the coating material while still in the liquid condition. Such computation in volves an underestimation of the den sity of the coating material when dried and an overestimation of the film thickness because air-drying mate rials become measurably more dense during the process of drying (9, 10). Existing data, however, are inade quate for calculating the densities of tne films after drying and weathering. For films of unpigmented raw linseed oil, the correction probably amounts to several tenths of a mil, but for the 14 pigmented films, the corrections would he smaller. Actual measurements of thicknesses of the films before immer sion in water confirmed the expected order of magnitude of the discrep ancies. Solubility of Free Films in Water: The quantity of soluble materials leached from the films during the de terminations of absorption of water and swelling is recorded in table 9. AU films tested yielded soluble mate rials. Thin films yielded more, in pro portion to their oil content, than truck films of the same kind, as might well be expected from the greater specific surface of the thinner films. On the * average, the solubility remained about as great after weathering as it was before, even though the weathering cycle involved enough spraying with water to remove much soluble mate rial before the weathered films were subjected to immersion tests. This fact, together 'with die steady increase in solubility during the prolonged im mersion tests reported in table 1, sug gests that most, if not all, of the sol uble substances reported in tables 1 and 9 .were formed by reactions that took place while the films were im mersed in water rather than before immersion. The soluble ingredients leached from the unpigmented oils and from paints with insoluble, chemically inert pigments such as magnesium silicate and titanium dioxide, clearly originate in decomposition of the oxidized lin seed oil in the films. The extract can be recovered from the soaking water by evaporation to dryness and is then seen to be an amorphous, glossy, amber-colored deposit that looks mud: like the clear material found under paint blisters of the "glossy-backed" type. When the paint contains white lead, die extract contains no more than a faint trace of lead, but the extract from paints containing zinc oxide has a substantial content of zinc, presum ably as zinc salts of organic acids. The differences in solubility among the various; kinds of films recorded in table 9 were not large enough to re quire discussion at this time, except to point out that commercial paint No. 4 proved exceptionally high in solubil ity both initially and after weathering, and commercial Mints Nos. 1 and 3 were unusually high when tested in the thinner films. No reason for such peculiar behavior is evident in the formulas printed on the labels of the paints. Leaching of Coatings During Arti ficial Weathering: Much larger pro portions of soluble substance were leached from many of the coatings by the water spray while in the weather- GLD37981 ing machine than were extracted from the free films subsequently when they were immersed in water. Leaching during weathering was determined by exposing coatings applied on tared tinplate in the weathering machine side by side with the coatings on gummed paper, and by drying and weighing the coated tinplate speci mens after each period of exposure. The leaching losses, expressed in per centage of the linseed oil in tire coat ing, are reported in table 10. Leaching losses were greater from thin than from thick coatings of the same kind of paint. Since the quan tity of soluble material increased mark edly during each successive interval of weathering, it is evident that the process of solubilization went on dur ing the weathering, for all soluble material in the coatings when weather ing began would surely have been ex tracted within the first 400 kilowatthours of exposure. Higher losses in the weathering machine than in 3-day immersion periods indicates that solu bilization goes on faster when ultra violet light and water act together than when water acts alone. (During one period of 400 kilowatt-hours, or 5 days, in the weathering machine the total time of subjection of the coatings to running water could not have ex ceeded half a day.) Leaching losses from coatings free from zinc oxide, including white lead paint and unpigmented raw linseed oil, were notably greater than the losses from paints that contained zinc oxide. The result offers an explanation of the observation by Browne and Laughnan (6, 7) that white lead paint and titanium-lead paint waste away in coating thickness during the intervals between repaintings more rapidly than mixed-pigment paints that contain zinc oxide. Tests of 10 Paints After Natural Weathering Ten paints, the five commercial paints and five of those made in the laboratory, were applied in three-coat work by brushing on substrates of tin plate and of paper coated with sili cone and were then exposed to natural weathering for 5I/2 months, from Sep tember 1952 to February 1953, facing south and sloped back at an angle of 45 from the vertical. At the end.of the weathering period samples were cut, stripped from the substrates, and tested by immersion in distilled water for 72 hours. The results are reported in table 11. The absorption, swelling, and solu bility of the films exposed to natural weathering were reasonably similar to those obtained with the same paints by artificial weathering, tables 6, 7, Table 11.--Absorption of water, meUlna, and solubility of free films or /lye eQumeroial iinai five laboratory-axle paints after natural weathering .for 5-1/2 months i The films were l.$ by 2.5 inches in slxe, of the thickness indicated, for each paint averages are reported for four films, of which two were exposed on tinplate, and two on silicomd paper. Free films were tested by immersion in distilled water at 69.5* F. for 72 hours. Kind of paint p st 1 1 iPigment1Thick-tAbsorptionISwelling Swelling(Solubility, (Volume ness of water i(Increase! effi- 1 based on t -in of < based on tin area) t eiennr ((night of 1 paint film (weight of 1 p/nv i oil in AQ I im^fk soil in film t s/va 1 t film * r A/W0 1 t ltil s Percent Feroent (Percent s Percent Commercial paints No* l-TLrj^2j t 0.35 ! 5.0 t 163 51 t 31 1 JJo. 2-(TL^g .31 1 4.8 1 130 1 33 ft 25 ft Ho. 3-(ra,.9)x26 t .30 t 4.0 1 128 11 24 ft 18 s No. 4-(72^22 ft .38 < 4.5 i 104 ft 30 29 1 No. S-T-^fr*) ft .31 t 2.5 t 99 11 51 4` . ft laboratory--de_p*:Lnte L ft 30 1 3.3 1 23 1 2 ft -8 t LZ37 TZ20 **20 K20 ft .30 t 3.6 ( 91 f 35 ft 24 1 1 .30 1 3.9 1 126 S 46 ft 36 1 t .36 1 4.7 1 112 ft 32 ft 29 ft ft .40 1 3.6 1 100 ft- 26 ft 26 ft 13.1 9.4 15.2 20.8 6.7 6*4 7.4 12.7 10.8 11.0 and 9. All of the paints except pure white lead and commercial paint No. 5 were "still swelling much more than southern yellow pine does even after weathering. At least for the present, it may be assumed that artificial weath ering brings about changes in paint films of much the same kind as oocur during natural weathering. Further re sults of natural weathering will be ob tained when a longer period of time has elapsed. It is not possible to estab lish the relative rate of change by arti ficial and natural weathering, but the results suggest that 5J/^ months oyer a winter period are equivalent to no more than 400 kilowatt-hours in the artificial-weathering cycle used. Tests of Certain Paints After 16 Years of Natural Weathering In 1936 a series of exposure tests was started at Madison in which wood anels were painted with 12 kinds of ouse paint and each paint was re painted at intervals, always with the same kind of paint used originally, until 1948, when painting ceased for ' 4 years while the results were being observed. As pointed out In other pub lications, (6, 7), most of the painting programs led to coatings of excessive thickness, which, if the paint con tained zinc oxide, produced cross grain cracking, curling, and scaling in relatively large pieces. Late in 1952 large chips of eight kinds of paints were collected for test by immersion in Water. No paint free from zinc oxide yielded paint chips large enough for such study. ' The chips collected were of various sizes and of irregular shapes. It there fore was not possible to cut them into the 1.5- by 2.5-inch or 2- by 2-inch specimens usually preferred; instead, chips were used very nearly In the sizes and shapes in which they were available. The chips selected were marked, dried, weighed and meas ured, immersed in distilled water for 72 hours, again weighed and meas ured while wet, dried over calcium chloride, and weighed and measured once mote. Hie results are reported in table 12, In which the paints are listed in the order of decreasing content of white lead. Although part of the paint in each of the chips was 16 years old and all of it was at least 4 years old, the chips still absorbed about as much water as 15 GLD37982 roughly similar paints did after mod erate exposure to artificial weathering, as recorded in table 6. The swelling and swelling efficiency of water for the 16-year old paints were somewhat lower than the values in tables 7 and 8 for much younger paints; but when due account is taken of the great thickness of the old chips, their swell ing cannot be considered much less striking than that of the younger paints after moderate weathering. Even after 16 years and in excessive thickness of film, some paints still swell more than flat-grain surfaces of southern yellow pine. Conclusions Free films of house paints when soaked in distilled water absorb sur prisingly large amounts of water, often more than 11/2 times the weight of lin seed oil that the films contain. As a rule, the absorption of water causes marked swelling, which in free film finds expression in expansion of the surface area that for many paints may amount to more than 50 percent when the films are still young. Such swelling is several times as much as that of flatgrain surfaces of wood between the air-dried and the water-soaked condi tions. Coatings firmly attached to wood must, of course, conform to the sur face dimensions of the wood; the larger swelling of the coating must be accomplished by increase in coating thickness. Such accommodation of the dimensions of the swollen coating to the requirements of the substrate nec essarily gives rise to large stresses within the coating, which, no doubt, have direct bearing on the phenomena of cracking, curling, and flaking in the normal weathering of coatings and on such abnormalities of coating be havior as blistering and peeling. Be sides the absorption and swelling, water causes a leaching of water-soluble ingredients from paint films, appar ently developed, at least in large part, by chemical reactions occurring while the paint films are in contact with water. Still larger losses of soluble in gredients occur during the weathering of paint coatings through the com bined action of ultraviolet light and spraying with water. For paint of any one kind, absorp tion and swelling are greatest when the free films have been freshly pre pared. They decrease slightly with age even when the films are kept indoors in subdued light under dry conditions. Absorption and swelling usually de crease markedly on weathering either naturally or artificially, but even after 16 years of weather exposure some paints may still swell in water more than wood does. Table 12.--Absorption of water, availing, and solubility of paint chips itan coatings exposed for 16 years on a tost fence at Madison, Wla. The paint chips were collected in 1952 from naturally loosened, coatings of the *1936 Paint Maintenance Teats,* which have been previously published (6, 7). The coatings on wood had been maintained by paintii^; in 1936 and repainting at intervals until 1948, when the last costs of paint had bean applied. The chips were of varying sites and of. the measured thicknesses indicated. They were tested by Immersion in distilled water at 6?.5* P for 72 hours. it i t i, t Kind of paint*PigmentiThickness libaorptioniSwelllng iSvsllingjSolubility, ivolume t of film < of water,i(increasei effi- t based on t in (measured)* baaed on iln area) i cienoy iweight of , paint * *weight ofi AQ i AQV.A toil in film s p/nv * * oil in * * * S/Vo *t * film * tt ti t A/rf0 * ** *J- *J- : J Mila Percent t Percent i Peroenti Percent U23 i 0.29 1 17.5 1 36 t 1.7 t 5 t 2.7 L237 i .29 s 13.4 1 18 % 1.6 t 9 1.7 *L230*80 i .29 t 10.3 1 35 t 3.2 8 9 1 2.7 fTt28Z28*ll6 s .29 * 9.3 t 23 f 5.5 1 24 t 1.2 aZ15>33 i .36 s 18.1 % 31 t 4.4 s 14 8 2.7 (SljiZj^j i .29 s 15.1 t 48 i 12.9 s 27 S 2.7 *TZ23)107 i .29 1 11.3 s 54 * 7.6 8 17 % 2.9 *TZ66*207^e* t .19 I 10.0 t 26 8 4.2 3 19 f 1.5 16 Paints in which there are markedly acicular (rod-shaped) pigments swell anisotropically and, in that sense, pos sess gram. The grain is produced by orientation of the acicular particles during the, last stroking of the paint when it is being applied. As in wood, swelling is greater across the grain of the paint than it is along the grain. Grain in paint often expresses itself in the pattern of cracking or other de fects mat show up later when the paint becomes embrittled with age. Although the absorption and swell ing of free paint films dearly origi nate in the oil vehide, they are mark edly affected by rite pigments, espe cially if the pigments are chemically reactive. Thus zinc oxide imparts greatly increased absorption and swell ing, much greater than the absorption and swelling of films of unpigmented linseed oil. Such chemically inert pig ments as titanium dioxide and mag nesium silicate alter the absorption and swelling of oil only moderately. On the other hand; white lead mark edly reduces both absorption and swell ing, especially after the paints have begun to weather. When the paint contains both zinc oxide and white lead, the augmenting effect of the zinc oxide predominates even when the proportion of zinc oxide is relatively low. There is evidence that paints of the so-called breather type, which are claimed to be unusually resistant to blistering and peeling by reason of a porosity of their coatings, may in fact make coatings with internal void spaces capable of holding free water without a proportionate swelling or, perhaps, of accommodating much of the swell ing internally rather than externally. One of the commercial paints, which dried entirely without gloss, was found to absorb a substantial amount of water without undergoing much swelling in consequence. Likewise, pure white lead paint, after it had be come weathered, swelled much less than would normally result from the amount of water it absorbed. It may well be that the resistance of such paints to moisture blistering comes from the low swelling rather than from the alleged ability to breathe by letting moisture readily pass out from the substrate through the coating. Further studies of the absorption of water, swelling, and solubility of free films of house paints offer promise of explaining many of the phenomena of paint behavior that have so far been subject to great diversity of opin ion, and of furnishing a tool for the search for better paints. But until such further studies lave been made, and particularly until a connection between swelling of free films and blistering GL 037983 of paint coatings has been established by direct experiment, it would be un wise to evaluate paints for practical use on the basis of the absorption of water and swelling of their free films. Literature Cited 1. Blom, A. V. Swelling and Per meability of Paint Films. J. Oil & Colour Chemists' Assoc. 22, 104 (1939). 2. ------ Swelling and Permeability of Paint Films. Oil Colour Trades J. 95, 851, 864 (1939). 3. -------The Swelling and Porosity of Paint Films. Farben-Zeitung 44, 161, 188 (1939). 4. Browne, F. L. Effectiveness of Paints in Retarding Moisture Ab sorption by Wood. Ind. Eng. Chem. 25, 835 (1935); 28, 798 (1936). 5. ------- A Proposed System of Clas sification for House Paints. Ind. Eng. Chem. 29, 1018 (1937). 6. ------- and Laughnan, D. F. How Often Should a House Be Painted? J. Forest Products Research So ciety 2 (5), 173 (1952); Paint, Oil & Chem. Rev. 115 (16), 16 (17), 26 (July 31 and Aug. 14, 1952). 7. -------and Laughnan, D. F. Effect of Coating Thickness on the Per formance of House Paints under Different Programs of Mainte nance. Official Digest, Paint & Varnish Production Clubs No. 338, 137 (1953). 8. Buser, Karl. Swelling of Oil Films. Farbe and Lack 1932, 256. 9. Carrick, L. L. and Permoda, A. J. Shrinkage of Some Organic Film Forming Materials during Aging. Official Digest, Paint & Varnish Production Clubs, No. 322, 692 (1951). 10. Clark, G. L. and Tschenke, H. L. Physico-chemical Studies on the Mechanism of the Drying of Lin seed Oil. I. Changes in Density of Films, Ind. Eng. Chem. 21, 621 (1929). 11. Dunbar, R. E. Chemical Changes in Films with Aging. Official Di gest, Paint & Varnish Production Clubs, No. 323, 857 (1951). 12. Dunlap, M. E. Moisture Resistant Finishes for Airplane Woods. Natl. Advisory Comm, for Aero nautics, Report 85 (1920). 13. ------- Protecting Wood, from Moisture. Ind. Eng. Chem. 18, 1230 (1926). 14. Eibner, A. Das Oeltrocknen, ein kolloider Vorgang aus chemischen Ursachen. Allgemeiner IndustrieVerlag G.M.B.H. (Berlin) pp. 115 to 118. 15. Gay, P. J. Measurement of Wet ness of Exposed Paint Films (Cli matic Factors in Paint Break down). J. Oil & Colour Chemists' Assoc. 31, 481 (1948). 16. Gordon, Victor. The Effect of Temperature and Humidity on Paint Films. Special Circ., Natl. Paint, Varnish & Lacquer Assoc. Papers presented before annual meeting Federation of Paint Var nish Production Clubs. Oct. 1939, p. 211. 17. Kendall, C. E. Water Absorption of Varnish and Paint Films. J. Oil & Colour Chemists' Assoc. 26, 35 (1943). 18. Kittelberger, W. W. and Elm, A. C. Water Immersion Testing of Metal Protective Paints. Role of Osmosis in Water Absorption and Blistering. Ind. Eng. Chem. 38, 695 (1946). 19. -------and Elm, A. C. Water Im mersion Testing of Metal Protec tive Paints. Role of Electroendosmosis in Water Absorption and Blistering. Ind. Eng. Chem. 39, 876 (1947). 20. Ledwith, R. J., Lewin, J. B. G., and Newell, G. A. Some Notes on the Water Absorption of Paint and Varnish Films. J. Oil & Col our Chemists' Assoc. 29, 67 (1946). 21. Long, J. S. Film Formation, Film Properties, Film Deterioration. Official Digest, Federation of Paint & Varnish Production Clubs, No. 297, 648 (1949). 22. MacGregor, John R. The Unex plored Field of Exterior Paint. Official Digest, Federation of Paint & Varnish Production Chibs, No. 335, 869 (1952). 23. Morrell, R. S. The Water Absorb ing Properties of Varnishes and Paints. J. Oil & Colour Chemists' Assoc. 4, 130 (1921). 24. Peters, Frite-Jurgen. The Swelling of Rust-Protective Paints in Wa ter. Farben-Zeitung 38, 1609, 1633, 1658 (1933). 25. Rinse, J. and Wiebols, W. H. Swelling of Drying Oil Films in Water. Ind. Eng. Chem. 29, 1149 (1937). 2d.-------and Wiebols, W. H. G., and Takes, H. V. Swelling of Oil Varnish Films in Water, Ind. Eng. Chem. JO, 1043 (1938). 27. ------ and Takes, H: V. The Wa-' ter Absorption of Linseed Oil Paints. Verfkronick 12, 26 (1939). 28. Serb-Scrbin, P. W. Water Ab sorption by Varnish and Lacquer Coats and Its Relation to Their Composition. Malyarnoe Delo. 1931 (5/6), 62: Chem. Zentr. 1932,1. 2388. 29. Stem, E. The Micrography of Paint Films. Korrosion und Metallschute 3, 153 (1927). 30. Stewart, Alex. Lead Pigments and Paints. Official Digest, Federation of Paint & Varnish Production Clubs, No. 311, 1100 (1950); Am. Paint J. 35 (16), 71, (17), 66 (1951). 31. Thomas, A. Morris and Gent, W. L. Permeation and Sorption of Water Vapor in Varnish Films. Proc. Phys. Soc. (London) 57, 324 (1945). 32. Wagner, Hans. Iron Oxide Pig ments and Rust Resistance. Farb en-Zeitung 35, 1151 (1929). 33. Wagner, Hans. Detection of the Water Permeability of Coatings. Korrosion und Metallschute 8, 66 (1932). 34.-------and Hoffman, G. Manga nese Soaps. Farben-Zeitung 36, 1214 (1931). 35. ------ and Zipfel, M. Black Pig ments and Corrosion I and II. Farben-Zeitung 37, 876, 912 (1932). 36--------- The Swelling of Films Con taining Pigments^ Farben-Zeitung 38, 266 (1932). * 57. -------and Pabst, Edgar. The Eval uation of Rust Protection from the Swelling Number. Farben-Zeitung 40, 789, 811, 1199 (1935). 38. Wolff, H. Testing and Evaluation of Water Resistance of Coatings. Korrosion und Metallschute 7, 191 (1931). 39. Wormum, W. E. Colloidal Be haviour. J. Oil & Colour Chem ists' Assoc. 17, 132 (1934). 40. Zeidler, G. Swelling Capacity and Permeability of Paint and Varnish Films. Paint 8c Varnish Produc tion Mgr. 19, 76 (1939). 17 GLD37984