Document Oz6GG3qOoDXg9x8pMoaV5mpXQ

644 With the Compliment* of Buekmen Laboratories* Inc. Swelling of Paint Films in Water, II1 Absorption and Volumetric Swelling of Bound and Free Films Before and After Weathering F. L. BROWNE Chemist, Forest Products Laboratory,' Forest Service, U. 5. Department of Agriculture Previous methods of studying swelling of paint films In water were improved by including measurements of volumetric swellings of both free films and bound films (coatings on glass). The new technique leads to more precise measurements and discloses infor mation about the structure of the paint coalings. t e c h n iq u e for measuring the vol Data were also presented on the effects A umetric swelling of bound films of on absorption and swelling of film house paint coated on glass as well as thickness, time of contact with water, of free films was developed to supple temperature and addity or alkalinity of ment the measurements of absorption the water, aging of the films before of water, swelling in area, changes on soaking in water, and exposure of the redrying, and loss of soluble ingredients paints to both natural and artificial previously described for free films. Al weathering before testing them in wa though bound films absorb and swell ter. But direct measurements of the somewhat less than free films do, films changes in film thickness were not of some paints still display great sensi precise enough to permit calculation tivity to water. Both artificial and of the volumetric swelling. natural weathering usually reduce ab Since the earlier paper was written, sorption and swelling below the values the experimental technique has been for unweathered films; but the reduc improved to measure directly the volume tion takes place during the first few of the paint films before and after months of natural weathering or the soaking in water and after redrying. equivalent and, thereafter, remains From the changes in volume and m nearly constant for at least 16 years. area, die changes in thickness can be Unweathered films swell at least 1 calculated precisely. Volumetric swelling cubic centimeter for each gram of water was measured both for free films and absorbed; weathered films swell less, for coatings on glass (bound films). sometimes much less than 1 cubic centi Volumetric swelling proves to be even meter for a gram of absorbed water. more instructive than swelling in area. Swelling and redrying or weathering Among different paints, the volumetric may develop voids within the films that and the areal swelling are not neces an hold free water that does not con sarily proportional to each other. More tribute to swelling. over, die volumetric swelling sometimes Introduction reveals changes in the internal structure of paint films that are not evident from A previous paper (3) reported meas urements of the absorption of water by free films of house paints, the swelling of the films in surface area aused by absorption of water, the shrinking on subsequent redrying, and the loss of soluble ingredients while soaking in water. Some paints were found to swell several times as much as wood does, whereas others swell less than wood. 1A contributed paper. * Maintained at Madison, Wis., in coopera tion with the University of Wisconsin. the areal swelling. The volume of a free film of paint was measured by weighing the film in air and again in distilled water (4). The difference in weight was the water displaced. Within the precision possible in this work, it was permissible to con sider that a displacement of 1 gram of water at room temperature (70 to 75) represented a volume of 1 cubic centimeter. The film of paint was weighed in air and in water immediately before immersion in water for the soak ing test. Then, at the end of the soaking period, it was weighed in water and Sain in air after removing free water nging to the surfaces. Finally, it was weighed in air and water for the third time after it had been redried in a dessicator over calcium chloride. Thus the initial volume, the swollen volume, and the volume after redrying and the densi ties in each of the three conditions could be calculated. The volume of coatings on glass was determined in a similar way by subtracting the previously determined weights of the glass in air and in water. By use of a chainomatic balance, it was possible to immerse, weigh, and re move a dry film from the water in 140 to 260 seconds. Tests showed that in 5 minutes, even the most highly absorp tive paints take up no more than 1 percent of the water they absorb in 3 days of soaking. Changes in length and width of the free films were measured before and after swelling and after redrying as in the work previously reported (3). From the volume and area of the film, the thickness at: each stage of the soaking tests was easily computed. Paints and Vehicles Tested The paints tested were chosen from the list of those used in the experiments on swelling in area previously pub lished. The composition of the paints is expressed by volume in the short notation first described by Browne (1) and more recently described in this journal (2). Two clear vehicles with out any pigment were tested; namely, raw linswa oil containing lead-manga nese naphthenate drier and bodied finseed oil of viscosity Z6 containing the same kind of drier. Single-pigment ts made in the laboratory were as S>ws: Magnesium silicate in raw linseed oil, pigment volume (p/nv) 0.30 Reprinted from the December, 1954, Journal of the Forest Products Research Society, pages 391-400 P. O. Box 2010, University Station, Madison 5, Wisconsin GLD3798 5 Basic carbonate white lead in taw linseed oil, paint L, p/nv 0.30 Anatase titanium dioxide in taw lin seed oil, paint T.,j0, p/nv 0.30 Zinc oxide in taw linseed oil, paint Z, p/nv 0.30 Of the single-pigment paints, only the white lead paint was a practicable house paint; the others are of scientific interest only. Multiple-pigment paints, all of them practicable house paints but made in the laboratory, were as follows: Paint TL10, p/nv 0.30, anatase ti tanium dioxide, basic carbonate white lead, magnesium silicate, and calcium carbonate in raw linseed oil Paint LZ,,, p/nv 0.30, basic carbon ate white lead and zinc oxide in raw linseed oil Paint TL00Z30, p/nv 0.30.) Paint TL30Z25, p/nv 0.30,) Paint TLaaZ^0, p/nv 0.30,) Paint TL10Z10, p/nv 0.30,) anatase titanium dioxide, basic car bonate white lead, zinc oxide, and magnesium silicate in raw linseed oil; same ingredients but differing proportions Paint TZ20, p/nv 0.30, anatase and rutile titanium dioxide, zinc oxide, magnesium silicate, and calcium carbonate in raw linseed oil Paint TZ,0, p/nv 0.3d, the same pig ments as paint TZ20, p/nv 0.30, but in a mixture of raw and bodied linseed oil Commercial paints of open formula were tested as follows: No. 1---TL.Zj j , p/nv 0.35, leaded zinc oxide, titanium-calcium, ti tanium dioxide, and magnesium silicate in raw and bodied linseed oil No. 2--(TL53Z18)l3t, p/nv 0.31, basic carbonate white lead, basic silicate white lead, zinc oxide, ti tanium dioxide, and magnesium silicate in linseed oil No. 3--(TZiy)12, p/nv 0.30, zinc oxide, titanium dioxide, and mag nesium silicate in linseed and soya oils No. 4--(TZl?)122, p/nv 0.38, zinc oxide, titanium dioxide, and mag nesium silicate in bodied linseed oil No. 5--T112(re), p/nv 0.39, ti tanium dioxide, mica, magnesium silicate, and aluminum stearate in tung oil varnish and bodied lin seed oil Commercial paints 1 to 4 were chosen because they are widely sold in all parts of the United States and are reasonably representative of the range in composi tion of the commercial first-grade paints with which the public is familiar. Com mercial paint No. 5, although not yet 2 so well known, was of particular inter est for this study because it is a socalled breather-type of paint that is un usually resistant to moisture blistering. Methods of Experiment Unpigmented oils were applied by brushing in successive coats until enough had been applied, determined by weigh ing, to form a coating of the desired thickness. Pigmented paints were ap plied by drawing down manually with a Bird doctor blade over a substrate held firmly on a Bird suction plate. Each paint was applied in a single coat at a predetermined wet-film thickness achieved when necessary by use of suit able shims. Coatings were allowed to dry in the laboratory near a closed window that faced north. Drying continued for 10 days, after which those to be weathered were mounted in an Atlas Weatherometer or on suitable racks placed on a roof of the Laboratory, where the coat ings were exposed to natural weather, facing south and sloped back at an angle of 45 from the vertical. Coatings that were not weathered were tested imme diately after the 10-day drying period. Bound Films (Coatings on Glass): In order to compare the absorption of water and swelling of bound films with free films, a rigid substrate that is not measurably affected by water was needed. Glass was chosen for the pur pose. Microscope slides 45 mils thidc, 11/2 inches wide,' and 3 inches long were measured for exact thickness with an Ames dial gage and sorted into matched sets actually 46, 45.5, or 44.5 mils thick, plus or minus 0.2 mil. Close control of slide thickness was necessary to permit control of wet-film thickness while applying paint with the doctor blade. Each of the selected microscope slides was marked with an identification number, carefully cleaned and dried in a dessicator over calcium chloride, and then weighed in air and again when submerged in distilled water at the temperature of the room in order to determine its volume by displacement. After die coatings on glass had been applied and had dried for 10 days, those for testing before weathering were promptly subjected to the soaking test. The others were weighed to determine the initial coating weight. They were then fastened by suitable dips to aluminum-painted plywood backboards of a size to fit the slots in the drum of an Atlas Weatherometer, and the arti ficial weathering was started. For the purpose of these experiments, the artificial-weathering cycle consisted of constant exposure to visible and ultra violet light from a flaming carbon arc filtered through Corex D glass, which passes only that portion of the ultra violet spectrum that is present in nor mal sunshine out of doors. As the Weatherometer drum rotated, the test specimens passed three tithes an hour under specially designed sprays that wet the painted surfaces gently with a fine mist of water in sufficient quantity to drain down and drip copiously from the bottom edges of the specimens. After passing the sprays, the water left ding ing to the paint evaporated before the specimens moved more than halfway around to the position of the water sprays. Test data arc given in tables 1, 2, 3, and 4. When the data for tables 1 and 3 were obtained, Madison city water, which is very hard, was supplied to the s in the Weatherometer. As will nscussed farther on, this led to false results with some paints for the loss in weight during weathering. For that reason, die Weatherometer was later equipped with a tank and pump to sup ply distilled water to the sprays. The data in table 2 were obtained in the revised Weadierometer. One set of specimens was exposed in the Weatherometer until the watt-hour meter registered 600 kilowatt-hours, which required about 71/2 days. A sec ond set of specimens was exposed for 1,200 kilowatt-hours, or about 15 days. On removal from the Weatherometer, the specimens were dried over calcium chloride and submitted to the soaking test. From the weight before exposure in the Weatherometer and the first weighing in air to start the soaking test, the change in weight of the paint coating during artificial weathering was computed. Free Filins: The paints and oils that were tested as films bound on glass, where the changes in volume necessarily took place entirely by changes in thick ness of coating, were tested also as free films that could change in all three dimensions. The free films were pre pared by spreading the paints as coat ings 00 guimmed paper mat had previ ously been saturated with linseed oil and allowed to dry. For any 1 paint, the coatings on glass and those on gummed paper were always made on the same day, and the 2 types of specimens were taken through $e program of exposure and testing contemporaneously. It was practicable to expose the coatings on gummed paper in the Weatherometer without loss of adhesion because die periods of exposure to water spray were too short for the dampness to penetrate through the coatings to soften the layer of gum. The free films were stripped from the gummed paper after the 10-day dry ing period for unweathered specimens or after exposure in the Weatherometer by soaking in water for a few minutes until the gum was dissolved. From the free films, test specimens IV2 by GLD37986 T inches were cut with a template and razor blade. Each specimen was marked in black drawing ink with an identifica tion mark and suitable reference lines for measuring length and width under a traveling microscope, The specimens were then dried over calcium chloride before starting a soaking test. Coatings on gummed paper have the disadvantage that the changes in weight of the coatings while undergoing weath ering cannot be measured reliably be cause the paper, the linseed oil in the paper, and perhaps the gum undergo changes in weight at the same time. In later work, therefore, coatings for ex posure either in the Weatherometer or to natural weathering were spread on tinplate. The tinplate specimens could be weighed before and after exposure with reasonable assurance that any change in weight occurred in the coat ing. After exposure and reweighing, the coatings were stripped from the tin- 3 GLD 37 987 Tebie 2.-IXU' for fw f.liii of 2 reMclei.___1_1__l_a_S_o__re_t_p_r_y-_M_Ad_e, e_n__d 5 c_ trclel pelnti before end after artificial weatpcrlM for IS d` ay* and1 a`fter na'tura'l vcaU` ic*rlna`for 6 kap_a_tha wd for 18 ncnthe--loeece in weight wile weawrlM taistllleq eater Med* In Weetheroneter vtttr eiilng in valve, area, mA thlckneee evening efll j pPeiimrinncyiy.).wchwe*mioJe atotTfcM.UJytl,flebdewrptcttFt.fo0rf.3eaatemr.x cWgeein vol\e. eree. and thlckncM on redrylne. and aleht of aojuble^* lente ioet whUt PtM : paint or clear vehicle : Weathering Weight lose (density of nonvolatile:*----------- while liquid) : Kind '.Time ln SA> Percent : i r: 1s :Before eoaklcgsAbiorp*: BtftUlnc Inerttnee tOvtlUnfi Retried change : iVslght Io m PlLr > wttr In -.thick-! A/V. trohm : neat 1 ! Vl Xn In clency Xn 1 In eren stfelek-iAV.Vo/A 1 Tolia. : i m MSS A*. Ar r1 1 Xn 1 t thick- x 1 MU t *2^. : : Mils {Percent: per- Per- Per- tPercent {Percent: Percent:Percent: Percent : :: cent cent ia5 t:: Raw linseed oil (0.933) Bodied Itneeed oil, Z6 vlicoeity (0.970) Munealus silicate* p/nv 0,30 (1.515) :Xtuzl ...do.. :, 6 12 <i) w.k 92.0 1.11 ! 1.17 I 1.16 : 1.22 : 3-1 ! 3-7 ! 2.7 : 2.5 > 21.0 < 20.k t 2k.2 x 25.0 :Nooe t0 :Artificial: l/2 :Natural : 6 ..........do....: 12 (1) kT.k *1.1 i 1.10 i 1.16 : 1.19 s 1.19 I 3.6 : k.l : 3.2 ! 3.S 1 9.0 : fifi.l i 7.5 : 6.2 :Kone 1 0 t. ...................., :Artificial: l/2 19.0 :Natural 1 6 33.8 :....do...12 39.6 1.60 : 1.89 t 1.80 : 1.81 : 2.0 s 2.k > 3.9 : 3*8 t 56.2 1 9.9 1 6.k x 6.7 22.0 20.5 16.6 28.3 9.2 27.5 8.1 7.7 56.1 9-8 6.7 6.k 9.8 k.k 8.7 33 6.1 10.1 k.l 2.4 15.9 3.0 2.2 1.6 10.8 15.3 9.7 lk.k 1.4 lk.9 1.3 2.1 31.5 6.7 2.5 2.6 Bade carbonate white lead, 1, p/nv 0.30 (2.705) jlfone :0 :Artificial: l/2 : Natural : 6 :____do....: 12 Tltaalua-lead, TL,n, p/nv 0.30 (2.010) In c m t0 : Artificial: :Natural 1 1/2 6 1 j....do.: 12 : Tltar.lua dioxide, anateae type, TUS0> pl(>i.a8ra0o,3)0 Zinc oxide, Z, p/nv 0.30 (2.345) Lcad-ilnc, p/nv O.36 12.37' (2.563) :Ncne :0 :Artificial: 1/2 :Natural : 6 : :____do....: 12 : :)Vone :O :Artificial: X/2 Natural : 6 . :....do.12 :Natural :....do... : 6* 12 : 21.0 26.5 20.3 (2) 51.6 53.7 55-1 85.2 107 12.1 16.3 15.8 7.7 20.7 21.8 s 2.8k i 3.21 : 3.28 : 3.28 t 3.9 : 5.5 ! 3-9 : k.l 1 10.7 1 12.1 1 11.0 1 10.8 10.8 12-3 8.0 6.9 . 2.22 : 2.67 ! 2.65 :, k.O : 3.1 : 3.3 x 29.3 : 27.2 : 28.2 29.9 26.3 27.0 : 2.12 : 2.32 ! 2.39 : 2.77 > 2.5k t 2.58 > 5-6 : 2.9 : *.8 i 25.6 : 10.2 : 20.9 x 12.8 ! 7k. 6 2.2 1 ko.o : k.l -1 kk.k : k.O i 33.8 25.3 9.5 18.1 9.1 7k.6 32.k 38.3 29-7 i 2.7k : 2.8k : 2.9k : 2.90 x 3.9 i k.2 i k.O : k.l ; 68.0 c 37.6 x k2.5 ! kl.3 81.0 34.6 ; 37.8 36.6 6.0 2.0 1.9 3-0 12.7 9.1 B.k 11.5 2.5 7-k 32 k2.7 11.7 Ik.5 Ik.6 12.6 20.0 19.5 19.3 1.6 10.0 5.1 3.2 u,e 17.0 16.k 6.9 6.2 8.8 3-5 20.0 16.9 20.9 12.6 2k. 0 12.7 lk.9 12.3 TltuiluM-leid-slnc, 0.55^0^(20.-3p7/0>)* -3 : ^Natural 6: :....do.... 12 : 13.0 16.3 20.2 2.52 ! 2.7k 1 2.69 : 2.68 s 3.9 s 3k : 3.9 ! k.O x 61.2 x 36.6 t 38.1 1 36.6 60.5 ik.e 37.0 3k.5 11.5 11.7 17-2 18.0 19.3 3k.O Ik. 6 11.4 101 6,0 1 4*9 : -1.31 100 2.2 t -1.71 .3 * 69 -2.0 t -3.8 X +1.9 * 113 -lO.k i: -2.5 X -8.0 X 102 -k.9 -2.6 ! -2.2 : no98 -.7 11 +2.1 : +.2 : -3.3 11 -.1 : -2.8 * 9k r 3.2 X 1.1 t -1.5 x 100 -3.8 j .2.2 X .1 t 99 -2.1 1! -3.k 1 .1 1 105 -2.5 -3 x -2.3 X 96 -2.6 1 0 : -2.6 1 101 -3-3 1.0 : -3.6 X 99 1.2 1 8 : +.2 : 80 i +2.0 l +.k 1 4l.5 X 6k -91 0 1 1.0 x 102 -3.9! +1.6 : -5.3 t 9T -.3 I 2.7 1 -2.6 : 96 -.8 1 1.31 2.1 ; 99 -k.l \ .1.01 -k.7 1 193 87 -1.3 x -2.4 : --.8 1 0; -1.4 : 71 -2.3 t -.9 : -1.11 100 -3-9 1 +9.4 : 12.0 : 61 +.6 : +2.2 : -.51 86 +29.k t +2.6 : 426.5 x 88 1- -.8 * Tk.3 1 -k.9 x 119 -2.8! 46. k ! -8.5 : 92 -2.0 : 7.6 1 -1.7 X 8? +22.5 : 42.6 1 +S0.0 S 89 -3.11 <i) x 99 2.0 x .47.7 1 9.0 1 95 1.0 % +4.2 : -.?. 97 -2.9 t 5.0 t -T.k 1 9k -3.01 5.9 X 8.2 1 5.8 2-9 6.1 6*8 5.1 1.8 2.8 3.1 k.5 3k 2.1 1.7 k.8 1.8 2.7 3.5 5.1 1.8 2.3 5.2 2.3 2.9 6.7 8.2 7.5 7.3 k.6 1.7 7.1 8.3 3.8 2.5 6.9 7.1 plate by floating them face down on mercury until the tin became amalga mated and the coatings could be stripped off. The free films were then cut to size, marked, and equilibrated over calcium chloride in preparation for soaking tests. Methods of Expressing Results During exposure to natural or arti ficial weathering, there is normally a loss of weight of coatings by leaching of water-soluble ingredients and some times by erosion (chalking). Some paints reported in table 1 gained in weight during artificial weathering for a reason that is explained farther on. In either case, the change in weight, which was the difference in weight of coating in grams immediately before ex posure and afterward, is represented by the symbol Sr There was a further loss in weight from all coatings and films during the 3-day soaking test that was revealed by the difference in weight in air just before immersion in water and the weight in air after redrying over calcium chloride. This loss in weight in grams from leaching of soluble ingre dients is represented by the symbol S,. 4 All coatings and films absorbed water during the 3-day soaking test. The ab sorption of water in grams, represented by A, was the difference between the weight in air on removal from water and the weight after redrying over calcium chloride. If the initial rather than the final dry weight were taken for the subtrahend, the absorption would be underestimated by the weight of soluble ingredients lost during the soak ing. By the method used, there is still a small error from loss of volatile in gredients other than water during the period of redrying in a dessicator, but such loss of volatile matter is much less than the loss of soluble substances while soaking. The great variation in density among the paints and vehicles studied raises a problem of finding a meaningful method of expressing comparative re sults for absorption of water. A, and the tosses, S, and S,. In the previous paper (3), A and S, were reported b percent by weight of the oil present b the film calculated on the assumption that the film at die time of test retained the same percentage of oil as the non volatile portion of the liquid pabt from which the film had been made. The assumption is at best a reasonable -ap proximation for unweathered films and pecomes decreasingly reliable as the films are weathered. Volumetric measurements offer a much better basis of expression; namely, the volume in cubic centimeters of the film or coatbg at the beginnbg of the soaking test, represented by the symbol V.. Accordingly, the absorption of water is now reported as (A/v,,) x 100, which is the percent by volume if 1 gram of water is taken as 1 cubic centimeter. The losses Ss and S, are expressed similarly; (S/V,,) x 100, or the grams of substance of unknown composition lost per 100 cubic centi meters of film. During the 3-day soaking b water, the initial volume V. increased to a larger volume V,; and when the film was redried, the volume shrank to Vr As a rule, V, was less than V., presum ably by the volume of soluble ingre dients leached out while soakbg. In such cases, the percent swelling in volume,AV,, was calculated as 100 (V,Vr)/V.; and it was observed that the density of the redried film returned nearly to the density before soakbg or a GLD37988 l*ble 2.--Pat* for fret film* of g vMcUi, ll labor*tory-id* r mnA y . and ait+r artificial wathcrUm for d*y and after natural watWlna*for 6 ont: --loto. In vtljht yfatl. ueatbe-rJln~g (<11b tilled wa.te..r u*d la W. eatbcrr----ter vater -w~yT). cl _ in volume, area. and tklcknce* and ave.Lllnj> efficiency n&en aoakefl ll m rptiocof vater. w Et 'iofft <*y t change* in voluac, area, and thickoeee o p redryina. ****** m Giu aoakiag^coatlDucd Paint or clear vehicle : (denalty of nonvolatile vhlle liquid) ^feathering Kind tlae Weight lo*a {Before loading Ab*orp- BuellInf lacrMM BwUlac Msdrltd chuf. tfclfht lea* vfcile veather* loe at m- while *oak> * jBenetty; nift **t*r In t In la el.scr In i la la : tthick: mn *A,, voltaic: area. thick' *V<A volvnc i ana ithick' AV, : AO. MSB AVr i 4_ t Mis t > *,, AT, -- ...........*..A..*.r... : Klla I f.r- Mr- P.rc.p PercenttPsrcsattPcrcsnt Percent 1 cni SE : Tltaniua-lead-tinc, None 0 ...i 8.85 :3.? <9.2 i 33.3 8A.B 98 t: -15 t +6.6 : -9.A 4.8 $I p/nv 0.30 3(r?5(2.UO) Artificial 1/2 Vatural 6 9-2 11.1 : 2.40 t> A.A 32.1 s 8.38 5.3 37.7 [ 15.0 30.7 > 16.6 1A.8 t +4.6 : -.8 -1*6 t (3> . .9 6.5 ....do.... 12 84.3 : 2.1*0 3 3.9 39.8 1 15.1 16.0 90 1-7 * +5.4 : 6.9 7.4 :: Titaaiua-lead-elnc, None 0 3.7 71.0 88.6 99 -1.6 +6.0 : -8.9 3-0 'B-rX?-w p/nv 0.30 1.960) Artificial 1/2 Vatural 6 10.3 19.1 : 2.27 A,3 A2.A ; 2.20 3.6 A9.0 AO.O 95 9 * 7-3 1 +? 85.8 90 Ifc.J { +2.9 : +11.6 i.e 5.7 ... .do,... 12 15.6 : 2.22 : A.l Al.A 19.3 93 -2.1 X +4.6 : -6. A 6.3 I Titanlua-lead-sinc, None 0 ........... ...1 1.98 sI A.l 8T.3 8A.0 100 X -2.4 t +6.?; -8.7 3.7 It~io-ioap9/0nv0)-3 Artificial 1/2 natural 6 12.7 8T.5 : 8.15 { : 2.03 ; 54.6 ft! S15.6 IB.2 +1,2 : +6.6 : 1.0 +2?,,2 : +6.0 : +88.5 1.5 7.6 ... .do,... 12 27.3 i 2.06 : 3.B A3-A tX 13.8 88.1 9^ -2.1 x +4.3 : 6.0 x: 5.8 Titaal-tlnc, TZX, p/nv 0.30 ^ (1.726) None Artificial Natural 0 1/8 6 8.2 16.6 1.66 A.O : 1-93 A.6 1.95 ; A.A UB A8.3 69.0 V, 93 95 .J +2, 2 * : +17.4 +10.6 : $ -17.8 +8.8 31.6 99 30,9 t 5.3 1.6 6.4 ....do.... 12 20.0 i 1.95 : A.l 61.B 83.8 97 -3 . 6 -13.5 7-4 Cconercial paint Wo. 1, Wane tLjZoot*p/nv -30 ' (2.005) Artificial Ratbral ....do.... 0 1/2 6 12 (i> 1?.5 18,1 : 2.26 : 8.84 : 2.25 8.8 $ A.9 t 5.2 98.0 A69 :? 36.3 18.3 87.5 19.8 31.3 88.A 85.0 88.8 99 97 8 J: -3. 4 : +10.5 i -18. A -.9 1 +5.8 s +6.4 1.3 t +13.A : -10.6 -1.1 t +9-9 t -9.6 7.5 4.8 5.8 5.0 Conerclal paint Vo. 2, Wone fyAahiw J>/Y Artificial (1.960) Natural ....do.... o 12 CoBoerclal paint Vo. 3 Wooe (ral9'lS6' T/nY -3 Artificial natural (1-745) ..do.... 12 w ll.3 22.0 (|)t ij.* 16.5 ...: 2.10 3.6 : 2.19 : 2.21 : 1.95 * 1.95 :`?*6 t 4.7 . 3.1 : 4.8 ! 4.9 72.A A7.2 A6.0 90.5 82.0 38.8 85.8 ,33.9 45.8 : 15.1 48.7 15.A t 92.0 t 88.1 86.5 83-4 53.8 77.7 1 7.5 65.5 36.8 i 6.7 85-3 99 9 93 108 2 -3.0 1 +8.0 1 -9.9 ,+-1.,43 , +5.8 s 7-5 -3.4 -7.4 -4.7 , +4.8 1 -9.1 +15.7 : +1.0 15.1 -2,7 1 +1.8 -4.0 4.9 8.4 3-4 9.5 B.l 9.0 Commercial paint Vo. A,:Vone 0 ........... ...; 8.i8 ; (T2i6)122, p/nv O.38 1Artificial ? (2,000) Vatural .. ..do.... 12 lB.4 21.2 : 2.23 J : 8.26 s 84.8 41.6 42.5 .3 : 44.1 27.3 8.4 40.3 8.5 88.0 18.3 84.1 100 66 95 -8,4 1 +13.5 1 -18.5 +8.5 6.0 -.8 1 *8.7 -.5 I -5.4 19.9 *:? *7.4 Commercial paint Vo. 5, Hone T1n1o2(re)(1.?66/n8v) *39 Artificial Natural .... do.... 1/2 6 12 (i) 48.5 62.6 : 2.08 : 2.06 : 2.06 : 1.6 3 1.6 ; 1.9 83.8 U.O 10.3 10.1 84.1 4.0 9.4 9-7 5-5 14.5 2.0 J-3 8.5 5:1 8.0 104 36 96 -3.7 i -.4 -4.0 +1.8 0 +1.8 -8.5 +.3 -8.6 -3-3 -.6 -8.6 4.8 6.4 8.9 3.8 Wet ittcrxlced. -TRw.t psint. were net toted fter artificial vwatberin* vlth distilled *ur la th. VMtlwrcB.t.r *prep*. 3s.t lacking because the redried file, were too tadljr curled and too brittle to permit *e.ureaent,. became slightly greater. But with films volume of the film by 1 cubic centi List direction of flow of the paint dur of some paints, the redried volume Vr meter. After weathering, the swelling ing application, than with the grain. was greater than the initial volume V,,, efficiency tended to decrease somewhat, After computing the percentage change and the density after redrying remained and for a few paints it decreased in area of the marked portion of the significantly less than the density before markedly. There are paints not included films, the change in total area was cal soaking, despite loss of soluble ingre in these tests for which the swelling culated on the assumption that the dients. To avoid gross underestimation efficiency is less than 100 percent even whole film changed in the same propor in such cases, the volumetric swelling before weathering. Low values of swell tion as its marked central portion. Tak- was calculated by the formula AV = ing efficiency indicate that the film con ing Q0, Q,, and Qr as the areas before 100 (Va - V0)/V0. The percentage tains voids that can take up water with soaking, after soaking, and after redry I change in volume after redrying was out causing any swelling, a condition ing, respectively, the areal swelling, always calculated from the formula that was predicted from theoretical con AQ, and the change in area on redry AVr = 100(Vr - V0)/y,,. It was siderations by MacGregor (7). ing, AQr. were computed analogously \ negative or positive accordingly as Vr Changes in area of free films were to ffie changes in volume from the for i was greater or less than V0. The relation between the percent ab sorption, A/V0, and the volumetric swelling, AV,, was conveniently stud ied by computing the swelling efficiency, calculated from the changes in length and width between lines ruled on the surface of the films about an inch apart in the central portion of the films. Curling at the edges of films that swell mulas: AQ. = 100(Q, - Q,)/Q0 if Qr was less than Q,, or AQ, -- 100 (Q. - Q.)/Q. if Q, was greater than Q,, and AQr = 100(Q, - Q*)/Q.- which was AV, divided by A/V0 times greatly made it impracticable to make The thickness of the film in mils (1 100, expressed in percent. As a rule, measurements between edges. It was mil == 0.001 inch) before soaking was the swelling efficiency was not far from shown in the previous paper that films calculated from the volume and area, 100 percent, which meant that each of many paints swell more across the T0 = 61 V,yQ0 where V0 is in cu. cms., gram of water absorbed increased the grain of the paint, determined by the Q,, in square inches, and 61 is a con- 5 GLD 37989 tabJ* 3. tilled vtitr for \ day, cl--, VoiiibTe ingredient# Io*t whtT aoaklnn Mat or clear vehicle itlae expoeed tBefare soaking (density of nonvolatile:to ertlflcla vblle liquid) : weathering : Density: film : :thic* : : nee* : : t0 8 .teorp->8 Swelling increase loa o) water : In : In In */t,, :v o 1u m: area thick: *V, : 49. net* 88 Iwelllni effi Vedrled ehanf* 8 xtf*ld>t lea* Ul. oak- ciency t In > In s In s W.V* :voluw: area : thick- t : 4V : t MU J * `* s sa 8 8 8 : Kile :Percent: ter- : Per- Per- tPercent Par- Par- 8 t cent t cent . gut cent Per- t Parent cent 1 Rev lineced oil (0-933) X0 t0 * 7-1/2 7-1/2 lj 15 : l.U : 1.11 : 1.20 : 1.17 8 (1) : 1.T7 : 3.1 : 5.6 : 1.7 : 3.9 : 3-7 21.8 10.8 15.6 16.1 i 22.0 1 9.8 1 11.0 : 6.5 t 38.3 > 25.8 > 16.5 : 6.2 20.1 1 20.$ t 1.1 10.8 1.1 21.2 6.6 15.5 101 102 81 103 100 -6.0 -5.2 eh -3.5 -2.2 -6.9 L6 -16.6 +1.2 -1.7 -1.3 -3.61 .17.01 1.6 x t3 5-8 5.2 9A 5.0 2.9 Bodied lineeed oil. z6 vlscoalty (O.J70) Hegnesiun aillcate. p/nr 0.30 (1.515) X 8 ; : I X * 8 Baele carbonate vblte lead, L, p/nv 0-30 (2.705) ; : : Titaaiuo dioxide, aaataae type. T,__, p/nv 0.30 llS0 (1.020) 1 : : Zinc oxide, 2, p/nv 0.30 (2.365) * : s 0 0 7-1/2 7-1/2 15 15 0 0 7-1/2 7-1/2 15 15 0 0 7-1/2 7-1/2 15 15 0 0 7-1/2 7-1/2 15 15 0 0 7-1/2 7-1/2 15 15 : 1.11 : 1.10 : 1.18 !*$ 3.6 1 5.1 : 3-1 : 1.7 : 1.T6 ! 1.68 : 1.65 > 1.82 : 1.71 : 1.8? 1 1.71* : 1.1 : 2.9 : 6.7 : 2.6 : 6.5 : 2.1 : 6.3 : 2.98 : 2.98 : 3.20 : 3.10 3A7 s 3.21 : 2.8 : 6.1 : 2.5 : 6.1 > 2.1 : 5-5 : 2.02 : 1.97 : 2.36 : 2.07 : (1) : 2.12 : 2.7 1 6.7 : 2.1 : 6.0 8 5.6 2.16 : 2.16 : 2.70 : 2.51 : 2.77 : 2.53 : 2.5 : 6.2 : 2.1 : 6.2 : 2.2 : 5.7 9.0 7.5 15.2 18.5 t 9.2 6.1 8.0 : 5.1 > 11.8 1 16.1 x 10.6 s 6.2 1.1 1.3 23.1 9.5 20.1 t 27.5 t 10.1 16.9 56.2 33.0 7.2 5.1 9.9 5.8 1 56.1 : 15.9 t 33.1 1 11.8 t 7.0 1 3.2 i 5.3 1.7 1 9.8 1 5.0 1 5.2 : 1.9 31.5 16.8 1.9 3.1 6.7 2.9 10.0 6.5 19.1 7.5 13.1 12. k 25.6 12.2 8.8 1.3 10.2 9.2 : i 6.6 : : 19.1 1 7.9 : t 10.7 : > 12.3 * 1.7 3.6 2.0 l.g 1.8 2.0 25.3 ! 13.5 1 12.5 1 1 9.6 1 > 1.2 : a : 9.5 > 7.5 3.6 2.2 2-'5 1.1 2.3 15.7 1.9 10.0 10.0 6.9 3.1 5.3 1.7 6.2 96.7 16.1 16.2 9.2 10.0 30.0 : 96.5 : 65.9 : 16.2 : 21.9 : 11.0 : 13.7 : 9.1 1 1.3 : 32.I : U.7 : 8.8 : 3.0 16.8 16.0 21.6 1.7 18.9 5.3 102 -1.9 -2.6 -2.2 1 107 -3-7 -2.2 -1.5 99 -1.5 43.7 -6.9 j 100 -2.3 -3 -2.3 1 96 -.7 42.1 4.2 : 100 -3.8 .2.2 -6.1 t 100 2.8 41.1 -6.0 t 97 -2.1 -.6 -1.9 10! -.6 0 -.6 $ 99 -2.1 *3.6 4.1 t 90 -.5 0 4.3 1 92 -3-9 ; -.2 -3.6! 101 2.0 -1.0 1.1 t 100 -1.8 -.8 -1.2 : 105 -1.7 -3 -1.5 t 80 -.1 ...8 4.6. i 99 -1.2 ...8 4.2 x 99 4*1 102 -2.6 109 -1.7 98 5 -.5 41.0 ; -1.0 -1.0 -.3 A.7 \ -1.9 -1.0 : -3 1 93 -1.3 -.7 0 s 100 -3.5 42k. 6 -22A > 100 -1.5 +!>.! -6.0 > 89 .2.6 .2.8 0 X 102 0 4A -.5 : -.181 4.6 +2.2 -5 : 88 4.1 -.2 x 5.1 3+ (.0 3.2 1.8 6.5 2.9 H 6 6 6.1 2.6 2.9 1.8 3- 1.8 5.2 2.5 3.8 1.1 2.3 6.3 1.9 6.1 .6 81..12 Tit*nlua-lead-*inc , 50 *(2.3p/7d0t) -3 \0 0 * 7-1/2 : 7-1/2 : 15 15 tita&lua-Bine, XZoo, 1 p/nv O.36 20 0 0 (1.870) 7-1/2 1 7-1/2 : 15 : 15 2-53 : 2.51 : 2.65 : 2.57 : 2.71 : 2.58 1 2.9 : 7-1 : 2.8 : 7.0 : 2.6 : 6.6 : 2.00 1-97 : 2.08 t 2.00 : 2.07 : 2.03 t 2.5 : 5.6 : 2.3 8 5*6 : 2.2 : 5.6 61.5 30.8 27.3 10.0 33.2 8.5 90.3 H.9 60.8 16.9 71.0 18.1 : 61.1 1 38.O : 30.8 : 15.2 < 32.2 : 11.1 9.7 : 5.6 : 30.9 : 17.1 : 7.9 : 3.3 ! 86.8 ! 58.5 : 11.1 : 23.8 : 58.8 : 18.8 : 16.1 5-8 : 69.3 : 21.1 i 17.6 : 6.7 15.6 12.k 17.7 3.6 8.5 5.0 18.0 15.3 33.8 10.0 39.8 10.8 99 -2.1 410.2 -10.7 j 100 -1.3 +1.6 -3.1 118 43.1 4.0 j 97 -.1 +1.0 -1.3 93 4.8 45.6 -2.3 : 93 0 *1 4.5 1 96 -.3 420.1 -17.0 t 99 -1.6 46.6 -5.7 : 97 el5.3 46.6 410.0 > 95 -.2 -.5 4.2 X 98 4*.7 4(5.0 4.5 : 96 .5 41.3 0X 3.7 1.3 2.5 .5 2.8 .5 5.0 2.3 2.9 1.1 2.9 1.0 Coaaerdal paint lo. 1,! IL-ZTM, p/nv 0.30 ' 23 ( 2.005) j * Cemerclal paint Vo. 5,1 T,,,(re), p/nv O.39 112 (1.668) 5 : t 8 0 0 7-1/2 7-1/2 15 15 0 0 7-1/2 7-1/2 15 15 2.15 : 2.15 : 2.25 : 2.18 : 2.26 : 2.18 1 2.5 : 5.6 : 2.3 : 5.8 : 2.2 : 5.1 : 1.92 : 1.81 : 2.0J : 1.89 : 2.08 8 1.92 8 1.6 : 3-9 : 1.6 : 1.3 : 1.6 : v.e 8 71.2 39.1 63.0 12.2 16.9 11.8 : 67.5 i 36.6 : 38.8 : 15.3 : 13.6 : 20.3 : 12.0 : 1.6 > 15.1 : 18.3 : ll.O : 6.2 23.2 19.9 7.6 n7..o0 7.6 : 21.1 : : 20.5 : : .6 : : 6.1 :' > 1.0 : t V,5 t 8X 5.5 1.0 2.3 1.6 2.0 2*0 22.T 19.0 18.7 6.7 22eh 7.2 11.5 ll.l 0 3.7 1.3 2.1 95 99 101 98 97 95 101 103 8 91 36 59 +1.6 -1.6 -.3 -9 -.5 -3.7 -2.5 -3-5 -1.5 41.2 41.0 +10.1 f.5 .7A 4.5 45.6 +.9 -A -.k -.2 -3 0 0 -7.5 1 4.1 x -7.7 -.? 1 -6A : 4,& x -3.6 1 -2.3 -3.2 1 -1.6 1 41.2 : 41.0 X t 6.9 2.8 2.1 .8 6.2 1.3 6.2 2.3 3.5 2.0 6.6 2.5 "Iheee film# were too badly <Ua*ed by veatberin* to be stripped trm the yi--il paper eubetrete a* fraa filM mltabla for caklnc teeta. version factor to give the answer in mils. For films bound on glass, the thickness necessarily changed during soaking and redrying by the same per centage as the change in volume, pro vided the coating remained attached to the glass substrate. But free films usually change in area as well as in thickness. The thickness when swollen was there fore calculated from the formula T,, = 61 V./Q., and the thick ness after redrying from the formula T, = 6l Vr/Q,, The percent change in thickness on swelling was AT, = tOO(T, - T0)/ro, and the percent change on redrying was ATr = 100(Tr - T,,)/T.. 6 GL037990 Comparison of Bound and Free Films Before and After Arti ficial Weathering Table 1 records the results of experi ments with bound films (coatings on glass), and table 3 the contemporane ous results with free film of the same vehicle and paints. Each vehicle or paint was tested both bound and free in two different thicknesses of film because thin films change more rapidly or to greater extent than thick films. In mak ing comparisons, therefore, both thick and thin films should be considered and allowance made for variations in thickness. Changes in Density: The 10-dayold coatings of all of the vehicles and paints were markedly higher in density than the nonvolatile portion of the liquid coating material from which they were made. Part of the increased density came from absorption of oxygen while drying, but others (4, 5) have shown that most of it came from contraction in volume, which for raw linseed oil after 10 days may amount to nearly 13 percent (4). On weathering, the density of the coatings continued to increase materially. Part of the increase in den sity on weathering of the pigmented coatings may be attributed to loss of volatile and soluble products of decom position of the linseed oil vehicle, and to the leaving behind of a higher propor tion of the pigments, which are much denser than the vehicle. But there was also some loss of pigment by chalking or by leaching of soluble salts formed by reaction of some pigments with de composition products o/ the vehicle. After allowance for such losses, how ever, it is clear that much of the in crease in density was due to further contraction in volume of the coatings. The increase in density while drying was often slightly more for thin coat ings than for thick ones. The difference was greater for coatings on glass than for coatings on gummed paper. Such effect of coating thickness was to be ex pected because the oxygen necessary for drying enters the coating at its surface, affects the top of the coating sooner than the bottom, and takes longer to penetrate to the bottom of a thick than of a thin coating. The gummed paper substrate may have been somewhat per meable to oxygen, whereas the glass substrate was not. On weathering, thin coatings always gained in density more rapidly than thick coatings; particularly so when the coating contained opaque pigments to keep the light of the Weatherometer from penetrating readily to the bottom of the coating. Absorption of Water: Free films nearly always absorbed more water dur ing the 3-day soaking test than bound films on glass of the same paint and feble k.~fetnrofloiracahaipa* tt hmeTt Vl. MUBtalMdof naturally loowwd paint tren e^tinga ** srn%t l Tt~b.-->----ilJiaMt f!U,* 1936* feintn: t Indoor* iictl feint SIS ss sfennltys riln 1d m sthlek-:tlott ofi In ffi- tehftxuM nfeiln ionl- k m : wt*r i voluw Dinner t In l *0 S *Ao S AT. s AWjAirolvm i s Lend-line, 12^ lead-line, 13.^1 * 2.80 8.87 Mlln sferccntifercent : fercent :Percent; ss s 17.5 s 23.7 s 2L.0 109 1 40.2 * 13.k ! 83.k ! 18.1 s 77 s -1.5 X fereot 1.6 2.2 Lend-slnc-inert, <*Veo Leai-iinc-inert, <"l5>33 8.7k l 2.32 s1 10.3 : 86.5 : SI.3 ss 10.1 I 18.8 t 17.3 to sX t -1.3 t I 98 i +.8 8.1 ,k THnnlun-lend-tlne, 8.67 ss 9.3 t 89.9 : 86.9 X 90 -7 1.6 Tltnnlun-zlnc, Titaniw-xlnc, 8.33 X ; 2.20 ss 11.3 s : ss :s 10.0 s 30.6 : kl.k 86.6 X 91 X 41.1 S X 87 < -.8 8.9 1.7 Llthopone-lend-sinc >5 (8LUZ23^75 1 a.31 ss 15.1 s A.5 s EL.7 st X 89 > -.6 1.7 i- JL comparable thickness. The few excep tions to this rule were with paints of relatively low absorption, and the differ ences between the free and bound films were small. Although both faces of die free films were exposed to water, whereas water had access to 1 face only of the bound films, the period of ex posure, 3 days, was so long in con sideration of the thinness of films tested that it is very difficult to attribute the greater absorption by free films to the greater area of surface exposed. There was ample time for water to penetrate to the center of the thickest films. Moreover, the differences between free and bound films are as striking for the thinner films of each paint as they are for the thicker films. The more probable explanation, therefore, is that restraint of the bound films by their attachment to the glass substrate developed internal stresses of sufficient magnitude to reduce the absorption of water below the quan tity taken up when the film was free to swell in all three directions. Nevertheless, even bound films of some paints absorb surprisingly large quantities of water. The greatest ab sorption recorded was nearly 63 percent by volume for the unweathered thinner film of zinc oxide paint. The corre sponding free film absorbed 90 percent. The smallest absorption by a bound film was 4.4 percent by the thicker un weathered film of white lead paint for which the corresponding absorption by the free film was 6.5 percent. Un weathered films of raw linseed oil ab sorbed 17.5 percent for the thicker and 10.3 percent for the thinner films when bound, compared with 21.8 percent for the thinner, and 10.8 percent for the thicker film when free. The absorptions by different kinds of paints and by thick and thin films of any one paint when in die bound condition ran parallel, at least quali tatively, to the corresponding absorp tions Dy the free films. Thus, if one paint took up more water than another when bound, it did so also when free. Thin films always absorbed more water than thick films of the same paint whether the films were bound or free. Artificial weathering increased the absorption of both bound and free films of unpignoented raw linseed oil and bodied linseed oil at comparable film thickness. White lead paint exhibited the same effect of weathering, but at a lower level of absonstion. All of the other paints absorbed much less water after weathering than they did before. The first 7Vi days of artificial weather ing usually effected a greater change in absorption than the second 7|^ days did; in fact, there was often a minimum of absorption after 7Vi days of weather ing and a slight increase again after 15 days. Swelling in Water: The volumet ric swelling, AV,, of all of the un weathered vehicles and paints in tables 1 and 3 paralleled and approximated reasonably closely the absorption of wa ter. For unweathered, bound coatings, the swelling efficiencies were between 99 and 120 percent, and for free coat ings between 92 and 107 percent. The exceptionally high and exceptionally low values were for films of very low ab sorption and swelling, for which the ex perimental errors were therefore great est. All of the water absorbed by the unweathered films, therefore, resulted in swelling, presumably because there were no voids in the Elms that could GLD37991 7 become tilled with water without in crease in film volume. The fact that the swelling efficiency often slightly ex ceeded 100 percent may be attributed for the present to underestimation of the absorption because of evaporation of absorbed water while the soaked films were being weighed in air; al though the possibility that the volumet ric swelling docs at times actually ex ceed the volume of water absorbed will be considered in a subsequent publica tion in this series. After weathering, the swelling effi ciency of bound films was always less than 100 percent. For commercial paint No. 5, the efficiency dropped as low as 51 percent. A similar trend was evident for the free films, but it was not so marked, except for commercial paint No. 5, and it was not so consistent. Some of the water absorbed by weath ered films often fails to cause swelling, presumably because voids have been de veloped within the weathered films that take up water without changing the volume. Apparently, bound films, which can swell or shrink in thickness only, reveal the porosity developed by weath ering more fully than do the free films, which can swell and shrink in three dimensions. The approximate equality in density of bound and free films indi cates that they are about equally porous before immersion in water for the soak ing test; but when the wet films, plas ticized by the absorbed water, dry out again afterward, the free films may be able to draw together and collapse the pores more nearly completely than the bound films can. For free films, the volumetric swell ing w3s distributed between areal swell ing and swelling in thickness. The dis tribution, however, was not in fixed proportions. Thus, for the thin, un weathered film of magnesium silicate paint, the volumetric swelling was 56.1 percent, the areal swelling 15.9 per cent, and the swelling in thickness 31-5 percent; whereas, for the thin, un weathered film of titanium-lead-zinc paint, the corresponding figures were 6l.l, 38.0, and 15.6 percent, respec tively. Such variations appear through out table 3 not only among paints but between weathered and unweathered films of the same kind of paint. In the previous paper (3) it was shown that many films swell much more across than with the grain of the paint. Recovery on Redrying: On redry ing after the soaking test, both bound and free films usually shrank to a vol ume slightly less than their initial vol ume before soaking. The loss in volume may be attributed to the loss of soluble ingredients by leaching during the soak ing period. Quantitative comparison of the volumetric shrinkage with the leach ing loss is not possible in the absence of 8 knowledge of the density of the soluble ingredients. Usually the leaching loss in grams per 100 cubic centimeters was slightly greater than the percent shrink age in volume, which agrees with the presumption that the soluble ingredients had a density greater than 1. One of the 60 bound films in table 1, and 10 of the 60 free films in table 3 retained a larger volume after redry ing than the volume before soaking, al though all of them experienced loss of soluble ingredients during the soaking. Of the 11 films that remained some what swollen after redrving, 9 were films that had been weathered, and all 11 of them were films that exhibited less than 100 percent swelling efficiency in the soaking test. Eight of the 11 films were of paints that contained zinc oxide and were more brittle and less distensi ble than the other paints. It may be concluded at least tentatively, therefore, that the residual swelling left after re drying indicates voids left within the films after evaporation of the water with which they had been filled while the films were wet. Failure of free films to return on redrying to their initial dimensions before soaking was much more frequent in the measurements of area than in those of volume. Of the 60 free films in table 3, there were 28 for which the area, when redried, exceeded the area before soaking. Twenty-two of the 28 were among the 24 films of paints that con tained zinc oxide. The paints of highest absorption and swelling on soaking ap pear to be the ones most likely to retain swollen area after redrying. Tne harder, less distensible paints of high absorption and swelling thus seem to shrink more readily in thickness than in area when they dry out again. The results agree with phenomena that have long been observed in the moisture blistering of paints. Formation of blisters necessarily involves an increase in area of film. After the wet condition passes and the blistered paints dry again, the more distensible kinds of paint often draw down to a surface so level that the blis tered area can no longer be seen; whereas, the less distensible paints often leave a rumpled surface, a sort of "double-chin" effect that clearly reveals the blistered area. Similarly, in normal weathering of paints when cracking and curling occur, it can often be seen that, if the curled paint could be rolled out flat again, the parts on either side of the crack would have to overlap. Both in swelling and in redrying of free films, measurements of changes in area alone prove inadequate ana may lead to conclusions at variance with the understanding obtainable when the vol umetric changes are measured also. When both volumetric and areal changes are observed, significant infor mation is revealed about the internal structure of the films and the nature of the stresses; that must be set up within them. Loss in Weight While Soaking: The loss in weight of paint films during the 3-day soaking period were discussed in the previous paper (3). The loss arises from leaching of water-soluble substances probably formed by decom position of the paint oils, chiefly during the soaking period. They can be re*, covered from the soaking water by evaporation. They consist of organic matter, together with inorganic con stituents if the paint contained pigments partly soluble in water or capable of forming soluble salts with organic acids. In this paper, it is Sufficient to point out that bound films yielded solu bility losses of the same order of mag nitude as those from free films. Thick films lost less in proportion to their volume than did thin films of the same kind of paint, whether the films were bound or free. Comparison of Artificial and Natural Weathering Table 2 reports results with free films of 2 unpigmented oils, 11 laboratory- made paints, and 5 commercial paints when tested before weathering, after artificial weathering for 15 days, and after natural weathering for 6 months and for 12 months. The 6-month period of natural weathering was from May to October 1953, inclusive, which in cluded the summer months; the 12month period was the extension through April 1954, which included the winter season. > Weight Losses While Weather ing: Losses in weight during weath ering, either artificial or natural, were substantial. Paint T4J0, in which anatase titanium dioxide was the sole pigment, lost 107 grams per 100 cubic centimeters of weathered film after 12 months of natural weathering. But much of die loss by paint T,,0 was pigment because it was chalking very freely by the end of the exposure period. Although other paints chalked observably, none of them was chalking freely enough to attribute much of their weight loss to erosion of pigment. For most paints, the losses are believed to consist chiefly of volatile organic products of oxidation and to leaching of soluble decomposition prod ucts of the paint oils and soluble metal salts of sum products. Raw linseed oil lost 92 grams per 100 cubic centimeters of weathered film after 12 months' natural weathering; bodied linseed oil lost a little less than half as much. Paint TL40 and commercial paint No. 5, which was pigmented with titanium dioxide and extending pigments, lost somewhat more than bodied linseed oil but less than raw linseed oil. Magne- GLD37992 sium silicate paint lost a little less than bodied linseed oil; and pure white lead paint, paint L, lost less than magnesium silicate paint. All paints that contained zinc oxide lost less than pure white lead paint, but the lowest loss after 12 months' natural .weathering was 15.6 grams per 100 cubic centimeters of weathered film of paint TLt0Z!0. These findings agree with measurements of the rate loss of coating thickness in practical exposure tests (2). In natural weathering, most of the loss in weight took place during the first 6 months. After 12 months, the losses were usually greater but only slightly so. This is strong evidence that the weight losses consist chiefly of leachings of soluble decomposition and reaction products, rather than mechani cal attrition by chalking or erosion. Oxi dation takes place most rapidly in young films and slows down as they get older. Erosion, on the other hand, is negligible in young films and speeds up as they age. To be sure, the first 6-month pe riod of natural weathering included the summer period during which paint deterioration proceeds more rapidly than in winter, but the 6-month losses ap proach the 12-month losses too closely to be attributed largely to the difference between summer and winter seasons. In artificial weathering for half a month, the weight losses were smaller, usually much smaller, than the losses in natural weathering. The shorter pe riod of time for artificial weathering may be partly responsible for the differ ence, but another and perhaps more important factor is the brevity of the periods of exposure to water in artificial weathering. Each spraying with water, during which soluble ingredients could be leached out, lasted only a few min utes, after which the films had time to dry thoroughly before again reaching the sprays. The time was insufficient for the water to affect more than a thin, superficial layer of the paint films. Other workers (6) also have pointed out that customary cycles of artificial weathering fail to provide long enough periods of exposure to wet conditions. In table 1, bound films of paints containing zinc oxide usually gained in weight during exposure to artificial weathering. Losses by other paints were usually much less than the losses for the same paints recorded in table 2. The gains by the zinc oxide paints were dearly abnormal and were traced to deposits from the hard water used in the Weatherometer sprays. The metal clips by which the test specimens were held in the Weatherometer drum be came coated with the deposits. This source of error was corrected by equip ping the Weatherometer with a tank and pump to supply distilled water to the' sprays for the experiments recorded in table 2. Absorption of Water: Weathered films of pigmented paints, except pure white lead paint, absorbed less water when soaked than the unweathered films of the same paints, whether the weather ing was artificial or natural. On the whole, artificial and natural weathering did not differ greatly in their effects on absorption. If there was any real differ ence between them, after making reason able allowances for differences in film thickness, the difference was that ab sorption by artificially weathered films was more often slightly less than slightly more than the absorption by similar naturally weathered films. At any rate, the results indicate that half a month of artificial weathering ages the films fully as much as a year of natural weathering, at 45 facing south, though it does not follow that the chemical and physical effects of the 2 methods of weathering are the same in all respects. Swelling and Swelling Efficiency: No significant differences were observa ble in the effects of artificial and natural weathering on the swelling and swell ing efficiency of paint films. Before weathering, the swelling efficiency was usually dose to or slightly greather than 100 percent; and after weathering of either kind, the efficiency as a rule was slightly less, sometimes markedly less than 100 percent. Likewise, the distri bution of the volumetric swelling be tween area and thickness changes was about the same for artificial and for natural weathering. Shrinking on Redrying: When re dried after soaking, all unweathered films returned to less than the volume before soaking, presumably because of leaching of soluble Ingredients. After artificial weathering, films of 5 paints, chiefly paints that contained zinc oxide, retained after redrying slightly more than their initial volume, All films naturally weathered for 12 months re turned after redrying to less than the initial volume; but of the films naturally weathered only 6 months, those of 10 paints had larger-than-initial volume after rediying. The residual swelling was small for 3 zinc-containing pints and pure white lead paint; biff it was large, from 14.5 to 30.9 percent, for 6 paints, namely, paints Z, LZ,,, TLI0Z,0, TLj02^0, TZ,,,, and commercial paint No. 3. Large residual swelling after re drying was usually associated with low swelling efficiency during soaking. The area of films of all zinc-containing pints, except commercial paint No. 4, was always larger after redrying than it was before soaking, even for unweath ered films. Thus, the behavior of zinc-containing paints on redrying after soaking sug gests that their physical condition after 6 months of natural weathering is farther away from the unweathered condition than is the condition after 12 months of weathering. On the other hand, it'has already been pointed out that the data for loss in weight while weathering suggest that the chemical composition changes greatly during the first 6 months, but not much more dur ing the next 6 months. But it may be that the physical condition that deter mines the ability to shrink again after becoming swollen in water, oscillates beteween limits during weathering, or that it depends more on the weather conditions for a few days immediately before sampling than on the total time of exposure to weathering. Solubility While Soaking: Artifi cially weathered films lost less weight from leaching of soluble ingredients during soaking than the corresponding unweathered films did, except for the pure zinc oxide paint and commercial pint No. 5. Naturally weathered films of bodied linseed oil, all zinc-free paints, and commercial pints Nos. 1, 2, and 3 also lost less soluble material than the corresponding unweathered films. But naturally weathered films of raw linseed oil, all laboratory-made pints that contained zinc oxide, and commercial pint No. 4 had more solu ble material than the corresponding unweathered films. (The laboratorymade pints were made with raw lin seed oil, but the commercial pints con tained substantial proportions of bodied drying oils.) Solubility was higher in naturally weathered than in artificially weathered films of all paints for which the data are available for such compari son, except pure zinc oxide pint and commercial pint No. 5. It appears, then, that artificial weath ering tends to reduce solubility below that of unweathered films. Natural weathering, on the other hand, seems to increase the solubility of films of raw linseed oil and of pints contain ing zinc oxide unless they also contain much bodied oil. For bodied linseed oil and zinc-free pints, the solubility after natural weathering, though less than that of unweathered films, is not so low as it is in artificially weathered films. More should be known about the chemistry of pint weathering before an explanation is offered for these facts. Chip of 16-Year-Old Paint An opportunity to measure the volu metric swelling of some very old pint films was afforded by a test-fence study of pint maintenance programs started at Madison, Wis., in 1936, the results of which have been published (2). The pint coatings were renewed by repintrng at intervals, always with the same kind of paint, until 1948. After that, they were all allowed to stand without CLD37993 9 further repainting until 1952 to provide a period for observing the outcome critically. The thicker coatings of many of the paints were then scaling badly enough to permit collection of samples of free films large enough to test for absorption of water and volumetric swelling. The chips were collected in the spring of 1952 and kept indoors until the measurements were made in January 1954. The results are reported in table 4. AH of the paint chips available for study were more than 9 mils thick, one of them 18.1 mils thick. The absorption and swelling measured, then, are for films of excessive thickness and pre sumably are much lower than the ab sorption and swelling of similar films of normal thickness, say 5 mils. All chips available were of paints that con tained zinc oxide and mat were there fore high in. absorption and swelling. Zinc-free paints, even in excessively thick films, did not loosen and scale in pieces large enough to afford samples suitable for measurement. The absorption of water ranged from 18.8 percent for paint (LZ15)3J to 45.2 percent for paint (TZ,?)1(,,. Absorption by the 2 paints in which there was no white lead was greater than the absorp tion by any of the 6 paints in which there was white lead. Although the ab sorptions were lower than for compara ble paints in table 2 that received nat ural weathering for 1 year, the differ ences were no greater than might be accounted for by the much greater thick ness of the 16-year-old films. Thus, it appears that most of the reduction in absorption caused by weathering of paint films occurs in the first year, or perhaps even in the first 6 months. Volumetric swelling of the 16-yearold films was always less than the ab sorption of water, and the swelling efficiency was therefore less than 100 percent. The efficiencies ranged from 77 to 92 percent. The 16-year-old paints did not differ greatly from 1-year-old paints in swelling efficiency. Moreover, the shrinkage of the swollen films on redrying and the loss of soluble ingre dients during the 3-day soaking period were about the same for the 16-year-old as for the 1-year-old films. Conclusions l. A technique was developed for measuring the volumetric swelling both of free films and of coatings of house paints bound on glass when soaked in distilled water for 3 days, and the shrinkage of the free and bound films when dried again after soaking. The volumetric measurements in conjunc tion with measurement of the absorp tion of water, the changes in area, and the loss of soluble ingredients by leach ing previously reported add materially to the understanding of the action of water on house paints of different kinds. Paints were tested before exposure to weathering and after exposure to arti ficial and to natural weathering. 2. Bound films absorbed somewhat less water and swelled correspondingly less than free films, although the ab sorption and swelling of bound films of some paints were surprisingly large. The bound films, which could swell in thick ness only, were perhaps unable to ad just their volume to the full extent of swelling and remained under greater internal stresses than the free films. Free films changed in length, width, and thickness. 3. Thick films, whether bound or free, absorbed less water and swelled less than otherwise similar thin films. 4. Unweathered films, either bound or free, usually swelled at least 1 cubic centimeter in volume for each gram of water absorbed. Artificial weathering in a Weatherometer reduced the absorption and swelling. Weathered films tended to swell less, sometimes much less than X cubic centimeter for each gram of water absotbed. It may be that the weathered films become porous and are able to hold some free water that causes no swelling in addition to the swelling water. 5. When redried after soaking, most films shrank to a somewhat smaller vol ume than they had before soaking. The loss in volume is attributed to loss of soluble ingredients by leaching during the soaking. Practically all films suf fered loss in weight from this source. But a few films, chiefly of highly ab sorptive paints in which there was zinc oxide, were slightly greater in volume after redtying despite loss of soluble ingredients. This seems to be further evidence of the development of voids within some films. 6. All films lost weight substantially during either artificial or natural weath ering. The losses were much larger in natural than in artificial weathering. In natural weathering, the major portion of the loss took place during the first 6 months. The loss in weight is attrib uted chiefly to leaching of soluble in gredients, rather than to mechanical erosion or chalking, because few paints were chalking very freely within the first 6 months of natural weathering. 7. In artificial weathering, the use of hard water in the Weatherometer sprays led to low measurements of weight loss; in fact, some paints that contained zinc oxide actually gained in weight. Evidently, there were gains in weight from deposition of the salts present in the hard water, which partly or wholly overbalanced the losses of soluble ingredients from the paint films. Such abnormal results were corrected by supplying distilled water to the Weath erometer sprays. 8. Absorption and swelling of films were reduced fully as much by 15 days of artificial weathering as by 12 months of natural weathering. 9- Chips of paint films from a testfence study of paint-maintenance pro grams were available for examination. The paints had been maintained by re painting homogeneously at intervals throughout the period from'1936 to 1952, with the last repainting in 1948. The absorption and swelling of the 16year-old paint chips, when reasonable allowance was made for their excessive thicknesses, was comparable with the absorption and swelling of similar paints when weathered for only 1 year. It seems, then, that the reduction in absorption and swelling of paint films caused by weathering takes place during the first few months and thereafter remains nearly constant for at least the next 15 years. References 1. Browne, F.L. A Proposed System of Classification for House Paints. (1937). Industrial and Engineering Chemistry 29:1018-1026. 2. Browne, F. L. and Laughnan, D. F. How Often Should a House be Painted--An Experimental Study of Programs of Paint Maintenance. (1952). J. Forest Products Research Society 2(No. 5):173-193. 3. Browne, F. L. The Absorption of Water, Swelling, and Solubility of Free Films of Paint. (1953). J. Forest Products Research Society 3(No. 5):108-125. 4. Carrick, L. L. and Permoda, A. J. Shrinkage of Some Organic Film Forming Materials During Aging. (1951). Official Digest, Federation of Paint and Varnish Production . Clubs, No. 322, p. 692-700. 5. Clark, G. L. and Tschentke, H. L. Physio-Chemical Studies on the Mechanism of the Drying of Linseed Oil. I. Changes in Density of Films. (1929). Industrial and Engi neering Chemistry 21, 621-7. 6. Gay, P. J. Measurement of "Wetness" of Exposed Paint Films. (1948). J. Oil and Colour Chemists' Association, 31:481-496. 7. Mac Gregor, J. R. The UnexploredField .of Exterior Paint (1952). Of ficial Digest, Federation of Paint and Varnish Production Clubs, No. 335, 869-880. ; , . i ' > | j j j ' j `! j j i , 10 GLD37m