Document Gmv3Or7555pkmM8jr53qGKgpN

FOREST PRODUCTS LABORATORY (Madison 5, Wis. ) FOREST SERVICE, U. S. DEPARTMENT OF AGRICULTURE Approved Technical Article 1521 Disintegration of paint by checking and crumbling, .after it has served its useful life. Note how the "crumbs'* fall bom the bands of summerwood long before they come from the springwood under* neath. This form of disintegration is characteristic of low-swelling paints. Disintegration of point by cracking, curling, and flaking, after It has served its useful life. Note how the "flakes'* fall from the bonds af summerwood before the springwood begins to be laid bare. This form of disintegration is characteristic of high-swelling points. wmsTANtm m mec han is ms or Deterioration of House Paint Al t h o u g h ma r k e d imp r o v e F. L. Br o w n e men t s have been made in the Forest Products Laboratory*, Forest appearance and some aspects of the Service U. S. Department wear of house paints, they remain of Agriculture prone to deterioration by undesirable crumbling or Baking from bands of Wafer, especially in summerwood, especially on such widely available softwoods as Douglasfir and southern yellow pine; and, un conjunction with ultraviolet light, accelerates chemical der some conditions of service, to un expectedly early failures by blistering or other abnormalities. For some time deterioration. The products of the decomposition cause die Forest Products Laboratory has therefore been conducting research on normal and abnormal failure. the mechanisms of paint deterioration Shrinking and swelling by more fundamental methods than die predominandy empirical procedures caused by wafer also stress that have been customary in die past and disrupt paint films. Chemical Changes During Drying and Deterioration The chemical processes by which paint coatings deteriorate are continua tions of much the same oxidative proc esses by which the coatings are formed in the first place. When paint is ex posed to air in a thin film, oxygen re acts at the unsaturated carbon atoms of the drying oil to form hydroperoxides. The Aulkne: F. I- Browne holds * PhD in colloid chem istry from the University of Wisconsin. He joined the FPL in 1918, ana since has been a frequent contributor to the Joum<S' The hydroperoxides initiate chain re actions of two lands (20, 24, 26, 30, 35)J. One kind leads to cross-linked polymers or gels that make the hard ened paint film. The other kind pro ceeds in the opposite direction to split molecules of drying oil and subsesquently of the polymerized gel into a succession of smaller oxygenated frag ments. The ultimate products of the disrup tive oxidation, which begin to appear even during the initial drying of coat ings, are gases or vapors that escape from the film, such as water, carbon dioxide, carbon monoxide, formalde hyde, and formic add (20, 21, 24, 23, 26). Intermediate products are larger fragments that are less volatile or non volatile. They indude propionic, caproic, nonylic, and azelaic adds, still higher saturated and unsaturated adds (22, 39), hydroxyadds (33, 38), aldehydes (2), ketones (20, 22, 24, 26), and glycerol (39), although the ester linkage to the glycerol is stable enough to persist through aging and weathering (22). The nonvolatile fragments remain in die film until it comes in contact with water, when those that are soluble in water may be extracted. The insoluble oxidation products and the oxygenated groups formed in the polymerized gel are highly polar and hydrophilic, and therefore increase die capadty of the film for absorbing water (22, 24, 49). Pigments materially alter toe course of both polymeric and disruptive oxi dation and therefore toe drying and deterioration of coatings. The con sumption of oxygen and toe evolution of volatile products are increased by some pigments and decreased by others (41). Such effects are exerted by pigments that are as inert chemically as titanium dioxide, silica, and barium sulfate as well as by chemically active pigments. Basic pigments, such as zinc oxide and white lead, in addition react with add decomposition products to * Presented t the Short Course In Print Technology at the Unir. of Florida, Geinesrilie. Feb. 1939. * Maintained at Madison. Wis., In coopera tion with the University of Wisconsin. * Numbers in parentheses refer to literature Cited at the end of this report. Reprinted from the November, 1959, Forest Products Journal (Vol. IX, No. 11), pages 417-427 Forest Products Research Society, 417 North Walnut Street, Madison 5, Wisconsin Th beards on lha left were painted with three coats of a highswelling, titanium-lead-zinc paint, which blistered badly ender con ditions of cold-weather condensation. Boards on Hie right were primed with low-swelling aluminum house paint and finished with two casts of the seme titanium-lead-zinc paint. form soaps that may be more or less Young coatings that are highly plas- rioration by chalking and erosion, soluble in water than the correspond tidzed by oxidation products and by whereas longer waves promote less ing free acids (33, 38, 39). incomplete cross linking of the poly drastic cleavage with volatilization of The nonvolatile decomposition prod meric gel can adjust their internal more hydrogen (as water) than car ucts that remain in the film signifi stresses. Old coatings that are more bon, increased cross linking, and fur cantly affect the properties of coatings completely cross linked and have lost ther shrinking and embrittlement of (32, 45, 49). Liquids or soft solids of plasticizers necessarily develop great the coating (25, 26, 37). low molecular weight act as plasticizers internal stresses (1, 35, 36, 46), be Less attention has been paid to the to impart flexibility and distensibility cause two of their three dimensions action of water in deterioration, al ) and perhaps to ketp the degree of cross linking of the polymeric gel from be must continue to conform to the re quirements of their substrate. This though its importance has been recog nized (26, 35, 42, 50). It has even coming excessive. Loss of the plasticiz leaves only the thickness free to take been claimed that water causes much ing products by leaching with water up the changes in volume. Further the same deterioration as light, but at a impairs flexibility and distensibility changes in volume imposed by periodic slower rate (43, 45). Certainly paints and promotes embrittlement. On the absorption of water and redwing dur exposed outdoors are repeatedly other hand, hard, crystalline solids, ing weathering eventually build up brought in contact with water (28), such as are formed by some of the enough internal stress to disrupt the which, from the knowledge already zinc soaps, act like pigments to make coating. reviewed, can be expected to affect the film harder and less flexible. Light speeds both the polymeric and materially the composition and proper Water-soluble or strongly hydrophilic decomposition products make the film absorb more water. Although the ab the disruptive oxidation in the drying and deterioration of coatings (20, 21, 25, 26). Light Is more effective the ties of the coating and contribute to faster deterioration outdoors than in doors. sorbed water acts helpfully to plasticize shorter its wave length (24, 35, 42). the film, it also has the harmful effect The ultraviolet components of sun Experimental Procedure of causing swelling. light with wave lengths from 280 to Previous publications reported the In the first stages of drying, there is 400 millimicrons are more effective behavior of paint coatings and free a gain in weight despite loss of volatile than visible light with wave lengths films (detached from their substrate), products, because the weight of oxygen from 400 to 800 millimicrons. Wave both before and after weathering, retained in the film exceeds the weight lengths shorter than 280 millimicrons, when brought in contact with water of volatile products lost (18). The although still more effective, are of (4 to 14). The absorption of water, oxygen retained soon reaches a max little practical importance because they changes in density, volumetric swell imum. Linseed oil, for example, con are absorbed in the atmosphere before ing, changes in dimensions, recovery tains 11.5 percent of oxygen to begin sunlight reaches the earth's surface. on redtying, extraction of water-solu with, but may gain more than 13 per Window glass absorbs much of the ble ingredients, and losses by volatili cent to reach a maximum of 25 per ultraviolet from sunlight zation and leaching during weathering cent. But loss of volatile products con On the other hand, the longer were measured. The coatings were tinues during aging or weathering, so waves of light penetrate to the full weathered by exposure either outdoors that the weight of the film declines depth of paint films of ordinary thick or in a Weatherometer so that dete from its maximum. The volume of ness, whereas the shorter ultraviolet rioration was accomplished by the com the film, on the other hand, begins to light penetrates only part way (25). bined action of sunlight and water, to shrink from a very early stage in the Pigments of high opacity to ultra gether with oxygen. Changes during drying (19), partly from loss of vola violet light, such as zinc oxide (40), the first 10 days after the liquid paints tile products and partly from increase further restrict the action of short were applied, that is while the paints in density of the oxidized film. Thus waves to a very thin superficial layer. dried and hardened, were not meas the density of linseed oil increases The shorter waves stimulate partic ured. from 0.93 to more than 1.1 within 2 ularly the production of volatile com The experiments described here af days. h pounds of carbon, and result in dete ford information about changes in 2 weight, volume, and density of some linseed oil paints during the drying period and about the effects on these properties of subsequent exposure un der various conditions for 13 days and their behavior toward water. In par ticular, the purpose was to learn mote about the relative action of water and light in paint deterioration. All of the 8 paints used were made at 30 percent pigment volume in a vehicle of raw linseed oil containing lead-manganese naphthenate drier. The composition was always 0.261 gal lon of pigment, 0.609 gallon of lin seed oil and 0.130 gallon of min eral spirits and liquid drier to the gallon of paint. The paints differed only in pigment composition. In paint L the pigment was entirely basic car bonate white lead, in paint Z lead-free zinc oxide, in paint T rutile titanium dioxide, in paint A antimony oxide, in paint X magnesium silicate. The other 3 paints contained mixtures of 3 or 4 of these pigments from the same lots. In punt TLX the pigment con sisted of 14 percent titanium dioxide, 40 percent white lead, and 46 percent magnesium silicate by volume. In paint TZX the pigment contained 19 per cent titanium dioxide, 20 percent zinc oxide, and 61 percent magnesium sili cate. In paint TLZX, the pigment con tained 14 percent titanium dioxide, 20 percent white lead, 20 percent zinc oxide, and 46 percent magnesium sili cate. The paints were spread on clean tin plate by means of a suction plate and doctor blade, at a thickness of coating calculated to yield 4 mils after the volatile thinner evaporated. The weight of paint was determined by weighing each tinplate before and immediately after application. Hie initial weight, volume, and density of the coatings after evaporation of the volatile thin ner were calculated from the composi tion of the paint, the bulking values of the ingredients, and the weight of paint applied. The coated tinplates were allowed to stand 10 days in light filtered through a glass window facing north. Temperature and humidity were not controlled, but the temperature did not depart much from 80F. and the relative humidity was always less than 40 percent. Specimens of each paint were weighed daily for the first 10 days and then at suitable intervals up to 83 days to follow the changes in weight. On the 10th day, a set of specimens, one for each paint, was selected and the coatings stripped from the tinplates by floating them on mercury until die tin was amalgamated and the paint film set free. The free films were submitted at once to test for their behavior to ..~Chang|i*t la weight, volume, mod density of paint coating* while drying and aging indoor* la dry 2 mst- Beostty 1 t : 1st : day r tt : 10th 25th ; A(S St : day day Hiss i t weight L (vbttc lead) 2 (tlw addi) 1 i ; 2.10 s 2.90 i 2.94 ! * 2.32 : 2.VT : 2.52 t Ho t 6 8 1 (tltanim dioxide) t 1.91 : 2.0V : 2.10 i A (matlooey oxide) ! 2.3& : 2.53 . 2-5V i X (--pnlte : 1.51 : 1.62 1,69 { lUet*) tt : tU (tltanlu*-lad- : 2.04 x 2.2} t 2.24 ; aillcate) TOC (tltanlus^xlae* t 1.74 : 1.92 1.92 t silicate) * is: TUX (Utmaluo-lMd- 1.91 : 2.15 2.14 : tine-*ilie*te) 2 2 6 6 6 0 Changed in weight $ Decline from maxima* Maximal t after t gala : 10 day*; 25 day*: 63 days: af er 20 days 25 days Ctc/m* Safer Sate3 Ste3' 6.75 0.50 1.13 3.18 *T-T 63.0 9-lfl .46 .63 1.79 24.4 25.6 10.75 1-29 2.56 4.95 1T.1 51.2 12.25 1.76 3.02 5.00 14.9 24.6 7.90 .55 1.70 4.1>6 59-1 68.5 1.63 .38 .80 2.68 51.3 54.0 9-72 .33 .68 2.56 46.6 46.2 8.50 .17 .20 1.89 44.8 39.4 Coating* oa tinplat* were approximately 4 n thick when 10 day* old. Change* calculated for 1 cubic centiawter of liquid palat free froe TolatU* thinner. 5 UBwet ell p*lat at 30 percent pipMot volute. ward water by the methods described in previous publications (9). Also on the 10th day, 6 other sets of specimens were taken, each set to be submitted for 15 days to one of die following conditions of exposure: 1. Indoor light, dry air: This was a continuation of the conditions already described for the initial 10-day drying period. 2. Darknesj, dry air: The specimens were kept in a totally dark room in which temperature and relative hum idity were the same as in condition 1. 3- Darkness, damp air: The speci mens were kept in a totally dark room held constantly at 80 F. and 97 per cent relative humidity. 4. Repeated water: From 7 to 10 a.m. and again from 2 to 5 p-m. daily the specimens were submerged in rap idly running tap water at about 50 F. Between immersions, die specimens were allowed to drain and dry under condition 1. 5. Sunlight without water: The spe cimens were irradiated constantly with artificial sunlight from a flaming car bon arc enclosed in a Corex D glass globe. 6. Sunlight with water: The speci mens were irradiated constantly with the artificial sunlight and sprayed gen erously with distilled water for ap proximately 2 minutes in each period of 20 minutes. Tests under conditions 5 and 6 were conducted simultaneously in a Weatherometer with die specimens to be kept dry arranged in the upper tier of the rotating dram and protected by suit able baffles from the water sprayed on the specimens of die lower tier at each rotation of the drum. At the end of the 15-day exposure etiods, when the specimens were 25 ays old, the exposures were discon tinued, the specimens were allowed to dry, were weighed, and the films stripped from the tinplate for test for behavior toward water. Aging Indoors in Dry Air Within 24 hours after application, all paints were gaining in weight Evidently most of the volatile thinner had evaporated, a minimum in weight had been passed, and absorption of oxygen was greater than the combined losses of volatile decomposition prod ucts and any residue of volatile thinner still left, with further passage of time, all paints reached a maximum weight and then declined when losses of vola tile products began to exceed further absorption of oxygen. The changes in weight, volume, and density of the 8 paints during drying and aging indoors in dry air are sum marized in Table 1. It is to be remem bered (hat, in Table 1, the changes are computed from the starting point of the weight, volume, and density of the unoxidized paint free from volatile thinner, and that all paints contained the same volumetric proportion of lin seed oil at the outset. All paints readied maximum weight within the 10-day diving period, but they differed strikingly in the magni tude of the maximum and in the time required to reach it. Paints A and T in 2 days reached maxima of 12.25 and 10.75 centigrams per cubic centi meter; L, X, and TLX took 6 days to reach maxima of 6.75 to 7.90, and paints Z, TZX, and TLZX required 8 days to reach 8.50 to 9.18. After they passed the maximum, paints A and 2* declined in weight more rapidly than the other paints, and the 3 paints that contained zinc oxide declined less rap idly than the others. Although all of die paints still exhibited a net gain in weight even after 240 days, all expe rienced a substantial loss in volume with consequent increase in density within the first 10 days. There usually was further loss in volume and increase 3 Table 2.--Behavior of Lo-day-old paint fiias Auric* er*Hn< and after redrytcuj.* Pniat- .^S5lS5J.22!i25-25!ii2?-- tOeaalty: Water Voluae Am iThick- ;absorbed uaa aolubU 1 ,, Oaogq after.rediylog naterlal Otfuitr Voluae : Area (Thick. extracted : nose L (whit* lead) *s 1 t <0.07 ^ W Percent Percent:Percent 5.6 5.9 +1*0 +2.U Percent: Perceat: Percent : 0.0k -2.6 \ -0.2 : -2.it 2 (zinc oxide) t -.69 t 91.0 *90-5 -33.S Al.0 3.3 t *.02 -2*3 : 8.9 :-10.k T (tttanlus dioxide) : -.15 : 22.3 22.6 10,0 6.3 2.9 j +.0b -2.9 I 1.0 I -3-9 A (antimony oxide) ! -.03 1.8 *2.2 *.5 .5 1-3 *.01 1.2 -.5 i -.7 2 (*oga9l\et : illeats) TLX (tltaaiUB-le&d- : tllicat*) TZX (tltanlw-zinc- ^ fllllcatft) TLZX (tltaalua-lesd- : zlnc-stlicats) -.20 i -.15 : -.38 i -.39 ! 12.3 H.5 71.9 50.S *3.7 *11.9 *72.0 50.7 15.0 M 33-5 *17-3 23.2 T.7 27.1 2T.0 3-1 i +.02 -3.1 : *1.9 : -b.8 .'t * 3.5 : +.02 -2.5 -.1 : -2-3 3.1 *.008 -2.0 : 9-5 1-10.7 2.6 t *.005 -l.H t 1.1 -5.5 i Soaked in distilled water far 3 days sad then aHawed to dry out for J days la a desiccator. Coatings,, approximately k nils thick, stripped froa their substrate before starting the test. Percent change calculated on the basis of the volute of the dried flla before soaking la water. Linseed oil at 30 percent piffoeot voluae. in density during the next 15 days ex cept for paint TLZX, which appar ently recovered a small part of its ini tial loss in volume with a slight de crease in density. Table 1 shows clearly that none of the pigments used can'be considered "chemically inert." AU exert profound effects on the course of oxidation of die drying oil. In particular they affect the extent of disruptive oxidation, which produces volatile products that escape from the coating and presum ably also somewhat larger fragments that remain behind but are extractable by solvents (36). Pigments likewise affect the shrinkage in volume during drying and aging. Still further effects of pigments on the behavior of paint films toward water are therefore to be expected. When 10-day-old coatings were stripped from their substrates and die free films soaked in distilled water for 3 days and then allowed to dry again for 3 days in a desiccator, they behaved as reported in Table 2. The changes are computed on the basis of the vol ume of uie 10-day-old films rather than that of the volume of nonvolatile in the liquid paint as in Table 1. All films absorbed water and swelled in volume but in widely differing amounts. Paint A absorbed only 1.8 centigrams per cubic centimeter, where, as paint Z absorbed 91.0 centigrams. Volumetric swelling was nearly but not necessarily exactly equal to the volume of water absorbed. Film density de creased in proportion as the absorption increased. When redried, all films were lighter in weight than they were immediately before soaking by the weight of sol uble matter extracted by the water. Most of the soluble material can be recovered by evaporating the soaking water to dryness (8). From the data of Table 1 it can be shown that loss of volatile material during the 3 days in the desiccator in no case was as much 4 as 0.2 centigram per cubic centimeter. The redried films shrank below their volume before soaking by amounts that were comparable to the solubility losses. The density of the redried films was greater than it was before soaking. The free films swelled in volume when soaked in water by increasing both in area and in thickness. The paints differed widely in the distribu tion of their volumetric swelling be tween swelling in area and swelling in thickness. Thus paint Z increased 33.4 percent in area and 41.0 percent in thickness, whereas paint TZX in creased 33.5 percent in area but only 27.1 percent In thickness. Shrinkage on rearying was always accomplished more by contraction in thickness than in area; in fact, the redried films of 5 of the paints remained larger in area although much thinner than they were before soaking. The paints containing zinc oxide, which are harder and less flexible than the others, contracted least in area when redried. ' Deterioration under Different Conditions of Exposure Changes in weight, volume, and density of the paint coatings under the 6 different conditions of exposure al ready described are reported in Table 3. Table 3 gives also the behavior of free paint films after exposure when submitted to the soaking water test. Daring 15-day exposure in dry air with or without indoor fight (condi tions 1 and 2), the changes in weight, volume, and density were small, as was to be expected in the relatively brief exposure period. Much longer exposure and closer control of the exposure con ditions would be needed for satisfac tory quantitative measurements to re veal any effect of indoor light on the rates of deterioration. Nevertheless, measurable deterioration dearly oc curred both in indoor light and in darkness, as is indicated by loss in weight usually accompanied by loss in volume and gain in density. Paint TLZX, however, increased slightly in volume and decreased slightly in den sity despite its loss in weight, and paint TZX likewise decreased in den sity. Table l shows that these 2 paints declined from their maximum weight less rapidly than the other paints up to the 25th day, but thereafter they lost weight about as fast as the others. The behavior of the free films toward water after the 15-day exposure in dry air with and without indoor illumination did not differ greatly from the beha vior of the 10-day-old films reported in Table 2. Exposure to damp air in darkness for 15 days, condition 3, markedly altered the thavior of the free films toward water by increasing the ab sorption of water and swelling. Thus paint L, which swelled 5.9 percent in water after exposure to dry air in darkness, swelled 11.5 percent after exposure in damp air. Pamt Z swelled 91.5 percent and 121.5 percent in water after exposure to dry and to* damp air, respectively. The effect of damp air on weight, volume, and den sity does not appear to be consistent. Its effect may be complicated by chem ical reactions of moisture with decom position products formed in the pres ence of some pigments but not in the presence of others. Thus paint Z, which in previous tests (6) was found to absorb 20 centigrams of moisture per cubic centimeter in 7 days in 97 percent relative humidity, showed (Table 3) an anomalous gain after redrying of 1.4 centigrams per cubic centimeter whereas paint L, which in the previous tests absorbed only 2.9 centigrams of moisture per cubic cen timeter, exhibited greater loss in weight and volume in damp than in dry air. Similar differences between paints Z and L have been reported by others (40). On the whole, the data of Table 3 support the conclusion that dampness is more potent than indoor light in altering the properties of paint coatings, at least in the early stages of deterioration. Repeated immersion of die paint coatings in water, condition 4, pro duced greater losses in weight ana in volume, usually with greater increase in density, than any of the other con ditions of exposure without sunlight. These greater losses were due in part, but perhaps only in part, to extraction of soluble materials by the water, be cause the free films nearly always ex hibited a diminished content of watersoluble substances when they were tested after the exposure period. In the test of the free films for behavior to ward water after exposure, the films exposed to repeated immersion in water always swelled much more than I j) I ! \ those exposed to dry air with or with out indoor light, but they swelled somewhat less than those exposed to damp air. This effect was probably caused by the water-soluble products extracted during the repeated immer sion treatments. Exposure to artificial sunlight with out water in the Weatherometer (con dition 5), produced much greater loss in weight, loss in volume, and increase in density than did any of the exposure conditions without sunlight But water acting in conjunction with sunlight (condition 6), caused still further changes that were often as great as those produced by sunlight alone. Thus for paint T sunlight alone increased the weight loss from 1.1 centigram per cubic centimeter in dry air and darkness to 7.2 centigrams, whereas water with sunlight increased the weight loss to 21.9 centigrams. In practice, water may be even more im portant than the data indicate, because the test conditions were such that each exposure to water was very brief, 2 minutes of spray, possibly 5 or 6 minutes to drain and dry, followed by 12 or 13 minutes of dryness and warmth. Water could scarcely pene trate the coatings very deeply. In nat ural weathering the periods of wetness, although less frequent, usually last long enough to permit thorough pene tration of the coatings (28). The 3 paints that contained zinc oxide (Z, TZX, and TLZX) suffered smaller losses of weight and of volume during exposure to sunshine, either with or without water, than any of the other paints. On the other hand, the losses were greater for the magnesium silicate paint X than for any of the others. Zinc oxide is highly absorp tive of ultraviolet light, whereas mag nesium silicate is mote transparent than any of the other pigments to both ultraviolet and visible light (40). After exposure to sunlight, free films of paints L and A absorbed more water in the soaking test than they did before exposure (Table 2) or after ex posure to dry air in darkness. These are paints of exceptionally low absorp tion; in fact, they absorb less water than films of unpigmented linseed oil (9, 11). After exposure to sunlight alone, the swelling in water was nearly as much percentagewise as the absorp tion of water but after exposure to water and sunlight together the swell ing lagged behind the absorption. This is attributed* to development of poros ity within the paint coating when water extracts soluble matter after the coating has become rigid enough to resist collapse. Tnble of condition* of exposure on the change* in paint coatings iuid on the behavior of their (roe film* during nocking and after redryingi. Change during exposure^ Change during noaklng- : Change on ; drying* Exposure condition*- i : Water ; Soluble * : : Density: Weight: Volute Penalty:absorbed:Voltne:material ; Density: Voluaa :t :extracted: $ tCg/cn^ ;percent i 5 SfilSTy .tPerc*ent: Cg/--cm,JJ PACT* Indoor lieht, dry air : 40 .o; Darkness, dry air : 4*033 : lead) -0.076 : 46.5 : -1.5 -.067 t *9.8 Darkness, damp air s *.Oho s -.9 : -1.8 -.162 : +11.5 Repeated voter : +.06i : -2.7 : -3-0 -.109 : 47.4 Sunlight vltbout water i 4.260 : -10.6 : -11.6 -.166 : *10.7 Sunlight with voter : 356 : -1T.3-1 -16.3 -.116 : 46.6 46.6 : 5.9 : 4*11.5 : 7*2 : 429..05 t2 2.6 2.9 i:* k.i 2.8 . tpercent 40.050 : 4.0U2 .02h ; 4.02k : 4.012 : -.090 : 2.6 -2,k -l.k -l.k *412..70 Indoor light, dry air t 40.05* [sine oxide) "T35VT52 498.5 : 498^0 % 1.8 : 40X15 t -1.8 Darkness, dry air 1 4.053 : -.3 : -2.1 : -.720 : 492.0 : 491.5 * 2.9 : 4.01k : -1.7 Berimes*, damp air Repeated water 4X10j 4.0al : 4l,U : -1.2 : -.622 *.4122.0 :4l21.5 : 3.0 : 4.025 : 2.2 +.067 : -l.k : -3.2 : -.782 :4l05.0 :+10b.o : 1.3 : : -1.3 Sunlight without water ; 4.19k : -6.0 : -9.2 : -.kl6 : 435.2 : 43k.k : 1.5 : 4.006 : -.9 8unlight with water 4.259 1*-11-2 t -13.5 : -.392 : 431.2 : 430.lt : .9 : 4.009 * -.7 PABit T (tltsniuc dioxide) Indoor light, dry air I <o.oSS i T&J'i' -3.5 3 -0.261 : 43k .2 Darkness, dry air 3 4Xk$ 3 -1.1 3 -2.5 I -.263 ! 35.6 Darkness, damp air : 4.057 t 4.1 -2.6 1 -,365 1 452.7 Repeated water : ..018 : -5.8 s -5*S : -3Q3 : 439*8 (fanlight without water l ..136 : -7.2 : -9.5 : -.062 1 9.6 Sunlight with water 3 4,332 ? -21.9 * -23-0 t -.091 : 4.3 3 36.5 ! : 35.5 : 452.7 1 : 439*2 * 1 9-8 1 1 48.6 3 k.2 k.7 1.5 3-1 5.* 1.3 : .0W> : t ..0.5 s t 4.0kl 3 3 4.027 S 1 +.05* ! 3 4.023 t -k.2 -6.3 -8.6 -3.1 -k.6 -1.5 PAIR k (antimony caridel Indoor light, dry air s ).b io t T.rr -1.0 : -6.661 : Darkness, dry air 3 4.001 t -1.3 3 -.5 ; -.001 j Darkness, damp air Repeated water : Sunlight without water Sunlight with water J 4.5 t ..026 : -2.9 : -2.1 s 10.3..197 j -a.i t - t 4 57 22.1 22.7.k t - *- 1 02 -.010 : -.010..026 : : 41.3 41.3 J 42.0 3 41.3 J 42.243.3 : 1 41.6 3 41.3 * 42.1 3 4413..513t 41.9 * 2.2 8.1 21..5k 2.5 6.0 t 40.027 t 3 4.019 t 3 3 : 3 4444....000021728796 3 : t : -1.9 1.6 ----8151....0307 Indoor light, dry air Darkness, dry air Darkness, damp air s Repeated water : Sunlight without water 3 Sunlight with water s PAIR X (magnesium silicate) 1.0' - : -O.it? t 43k.0 : .36.1 1 -.7 -.193 0.7: 1 4*0.5 1t *k * 1.8 99*3--019 : - t .3 * -35a i 499.0*5; + * 4.009 2 3 --250 457.8-k.k .3.0 t s 57.8 1: : .74.066 3 -lk : -17.0 ! --206 ! 38.6 1: .38.3 1 29.2 0 7*.299 : -27.6 s - : -. k : . 6.9 :: 4 5.2 3 5.2 5.k 53.8.T .0lk 1-T :3 t : 44-40....000051210191 3; t t -k.2 -k.k 2.8*,k - 1.13 4.066 3 -10.8 t ..006 : - 40 5 6.113 412.0 412.3 3.3 40.030Indoor light, dry air 3 4 412.0 3 8 4.035Darkness, dry air 17 --5 .6 15-8 3 4.016 3Darkness, damp air : 3 4.020 2.1 -.139 13.1 1.8 ..018Repeated water 4.179 9 11.5 -.126 411.3 411.2 2.9 4.031Sunlight without water 3 261 16.3 -163 4 01Sunlight with water : PAIR .to i -X .. t : : ,: TtX (tltesi.ua dioxide, white lend, magnesium silicate) --* : -J ! - 1 -.1 3 --HO : :: ll.k t: :: : .k : t : -.158 t . :i .16.1 ` -3.0 : - : < 1: .13.6 t : - .k 3 -16.1 : - : i : :: ` 46J. :: 48.3 : 3 2.4 3 . : t 1 : k{ -23..08 ---1l21....k555 PACT* TZX (titanium dioxide, cine oxide, megnetluo ULleste) 0 66 0.2 3 468.2 3.8indoor light, dry air t - . k s -O.k ;: - -6.371 1: 488.0 t 1 : 40.019 3 --007 --2 465.7 3 466.0 3 3.5 4.083Darkness, dry air 3 t i1 0 s -.360 :: :t --019 -.2 .6 2.9Darkness, dnq? air : : : ,.7 : -.638 !1 49k : <96.9 : 3 4.019 1 470.5 2.9 3 4.011 3Repeated water 3 +.071 5.0 5 2 337 433*3 2.0 4 01Sunlight without wateir : 091 5-1 -.298 12.2 4 2.3 1.2 4.007Sunlight with water ! -.011 ; 3 .. . t -1.0 :1 - :: -5-6 11 -3 J - -k 3 -* --377 !: -. kk ;: ;: * . 3 470.8 : : :k : t : . k: s* "----22213-......950902 Indoor light, dry air ? Darkness, dry air $ Darkness, tap air : Repeated water 3 Sunlight without water t Sunlight with water : PACT* TLZX (titanium dioxide, white lead, tine oxide, magnesium silicate) -0.00? : -O.k : VJk : 436.6 t j 4d."6itf j 3ET3 -X08 : -X : 4.k s -.335 : *k8.3 : 4k2.T * 3*3 * 4.022 : -2.5 --01k : -.3 j 4*5 j -*k72 s 472.2 t 472.5 : 30 j 4X2k : -2.5 --010 : -1.3 : 0 : --358 : 4^6.2 s 4k6.e : 2.8 3 4.018 $ -2.1 4.067 < -k.k : -6.0 : -.250 3 +26.5 : 425.5 t 2X 3 --007 t --7 4.U9 : -7-k : -8.5 : -.289 : 49.k : 429-2 : 1.5 t -X06 : -.k 1 Coatings on tinplate dried indoors for 10 days (approximately k mils thick when dry), then exposed* to the conditions indicated, after which they vera stripped, from tha tinplate and tested for behavior toward water. 2 hts eleventh to twenty-fifth day inclusive. J Rased on volume of film on tenth day. k Based 00 voduee of fils on twenty-fifth day. Soaking tine 3 days. 2 Based ea volume of fUm on twenth-fitth day. Dried 3 days in desiccator after seeking. Ait of the other paints, after expo sure to sunlight with or without water, absorbed less water and swelled less than they did before exposure or after exposure to dry air in darkness. The difference was especially striking for those paints that were highest in swell ing before exposure, namely, the 3 paints that contained zinc oxide. The zinc-containing paints differed from all the others also in that the absorption of water and swelling after exposure to water and sunlight together were nearly as great or even greater than they were after exposure to sunlight alone. The other paints were less ab sorptive and swelled less after expo sure to water and sunlight than after exposure to sunlight alone. Paint films after exposure to water and sunlight always yielded less sol uble material to the soaking water than corresponding films after exposure to sunlight alone and also less soluble matter than films, with the unex plained exception of paint A, before exposure or after exposure to dry air in darkness. On the other hand, films of paints containing no zinc oxide, with die unexplained exception of paint TLX, yielded more soluble mate rial to the soaking water than the cor responding films before exposure or after exposure to any of the other conditions. Paint X after exposure to sunlight alone yielded the remarkable amount of 14.0 centigrams of soluble material per cubic centimeter. The 3 S JL -V PAtHT TL2X e. - -a/ Pig. 1- Chonget in waigfit and in volume of films under various con* ditfofl* of exposure. Films contained 1 cubic centimeler of non-volalile . when applied. zinc-containing paints after exposure to sunlight alone gave up less soluble material than their films before expo sure or after exposure to dry air in darkness. Apparently sunlight, but not indoor light, acts to increase the for mation of soluble decomposition prod ucts that remain in the coating unless extracted by water. Zinc oxide, because of its high absorption of ultraviolet light, shields nearly all but the outer surface of the coating from such ac tion. The low opacity of magnesium silicate, on the other hand, offers even less shelding than the other pigments studied. The changes in weight, volume, and density recorded in Table 3 are those that occurred during the 15-day .expo sure periods, calculated on die basis of the volume of the dry coatings when 10 days old. Fig. 1 is a graphic presentation of the changes in weight and volume that occurred, first during die 10-day drying period, and then during the entire 25 days for drying and exposure to each test condition, omitting the two exposures in dark ness. The changes are calculated on die basis of the volume of unoxidized paint free from volatile thinner applied to begin with. After drying for 10 days, the coatings still retained a net gain in weight, the magnitude of which varied with the kind of pigment 6 in the paint. There was, however, al ready a loss in volume. During the succeeding 15`days of exposure, both weight and volume decreased, but to differing extents ac cording to the conditions of exposure and the kind of pigment present. The decreases were least rapid for exposure to dry air indoors, were distinedy faster for repeated immersion in water, still faster for exposure to sunlight, and fastest for exposure to sunlight and water together. Among the 8 paints, the decline in weight and volume when exposed to sunlight was slowest for the 3 paints that contained zinc oxide, but no explanation appears for the greater decline in the paint with 100 percent than in the paints with only 20 percent of zinc oxide. The magnesium silicate paint lost weight and volume faster than any of the other paints. Significance of die Experiments The import of the experiments is that deterioration proceeds in much the same way under all conditions of ex posure in which oxygen is present, but that the rate of deterioration can be speeded greatly by such conditions as exposure to sunlight and exposure to water. In darkness or in indoor light, water in the form of either vapor or liquid makes paint films more sensitive to absorption of water and swelling. and in the liquid form it accelerates loss of weight and volume by extract ing soluble decomposition products. Indoor light has little stimulating effect on deterioration. Sunlight, which contains ultraviolet ' wave lengths, accelerates formation of decomposition products, both volatile products that escape at once and sol uble products that remain in the film unless extracted by water; No doubt sunlight also increases polymeric oxida. tion and cross linking of the gel struc ture, because it makes coatings more resistant to absorption of water and swelling--except, of course, those coat ings that are highly resistant before exposure. When water acts together with sunlgiht, deterioration attains its greatest speed, not only because the soluble decomposition products are extracted but because volatile products form and escape more rapidly. Pig ments of high absorption for ultra violet light, however, retard the action of sunlight, presumably by keeping the ultraviolet light from penetrating much beyond the exposed surface. Coatings of house paint are called upon to accommodate themselves to substantial changes in volume from two sources, 1) the continual loss in volume from deterioration of the coat ing, and 2) repeated absolution of water with swelling, redrying with shrinkage, and extraction of soluble components by water. To maintain adhesion to the substrate and remain unbroken, the coatings must make the adjustments entirely by changes in thickness, except as accompanying movement of the substrate may permit some change in area. The soaking tests reveal that free films, unrestrained by substrates, dis tribute any change in volume among all three dimensions. Attachment to a substrate, therefore, imposes restraint that necessarily distorts the internal structure. Only a liquid free to flow caa'submit to such restraint without de veloping internal stress. As a coating dries and acquires rigidity, internal stresses come into play that increase in magnitude as the rigidity increases. Very young coatings can relax the stresses largely by plastic deformation (23, 32). Older coatings that remain sufficiently distensible can withstand the stresses by elastic deformation. But such adjustments become increasingly difficult as further weathering increases the rigidity of the coating and deprives it of the simpler compounds that act as plasticizers. Eventually the internal stresses exceed the cohesion of the coating or its adhesion to the substrate, whereupon the coating breaks and be gins to come loose. Disruptive stresses in coatings on wood have often been attributed to movement of the wood surface with change in moisture content (29) Per haps there are circumstances under which this ran occur, but the disinte gration of house paints under normal conditions of service cannot be ex plained in that way (15). Applications to Normal House Paint Behavior Although present knowledge of the mechanisms of paint deterioration is far from complete, an attempt to apply it to the behavior of paints as com monly observed on houses and test fences may suggest directions for pro mising further research. The Forest Products Laboratory recognizes six stages in the normal wear of house paints: the soiling, flatting, chalking, fissure, disintegration, and advanced break-up stages (27). The Soiling Stage: Newly applied paint gradually becomes dirty. Car bonaceous dirts ding to paint more firmly than do siliceous dirts, glossy paints soil more seriously than other wise similar flat paints, and the pre sence of moisture greatly facilitates the attachment of dirt (31, 44). Ap-' parently dirts that are wet by oil more readily than by water stick to the oil vehide still present in generous amount in the surface of young glossy paint. The actions of water to swell, soften, render the oily surface more tacky, and displace cushions of ab sorbed air may bring paint and dirt into doser contact. The joining of paint and dirt partides by a film of water, and the subsequent evaporation of the water at the periphery of the junctions, will draw dirt and paint togather by the force of surface tension. The water-soluble substances extracted from paint by water may play a part, and conceivably account for the fact that some paints gather less dirt than do other paints that differ only in kind of pigments. Growth of fungi on paint, com monly called mildew, occurs only when there is moisture, because the micro organisms require water for their meta bolism. The. organic material extracted from paint by water may provide food for the fungi: Zinc oxide in paint is strongly inhibitive to fungi. Its effect is dearly due to the toxicity of sol uble zinc soaps, because zinc has been found in generous amounts in the aqueous extracts from paints contain ing zinc oxide (4, 8). Soluble salts of lead are also poisonous to fungi, but white lead in paint, although not without prohibitive effect, affords less protection than zinc oxide. The aque ous extracts from paints may contatn a little lead (38), but they contain much less lead than zinc (4, 8). The Flatting Stage: House paints gradually lose their initial gloss and become dull or flat. Losses of volatile products and leaching of soluble products, which proceed most rapidly at the exposed surface, destroy the superficial layer of oil vehide and lay bate the granular bed of pigments (46). In addition, contraction in vol ume from deterioration of deeper parts of the coating and distortion from repeated swdling and shrinking contribute to the roughening of the surface until its initial ability to reflect light specularly gives way to markedly diffuse reflection. Flatting occurs ex ceedingly slowly in dry air indoors. It is hastened by contact with water, by the action of sunlight, and particularly by joint action of sunlight and water in die order indicated by die changes in weight and volume recorded in Fig. 1. The part played by water should not be overlooked. House paints and auto motive finishes that dull rapidly in the climate of Florida retain their gloss for a much longer time in the equally sunny but very dry climate of south ern Arizona (48). The Chalking Stage: Continuation of the processes that cause flatting eventually destroys the binding me dium by which those particles of pig ment near the surface are held in place. When the loose pigment ran be rubbed off, the coating is said to be chalking. Chalking is so easily observed that many have been led to attribute to it the gradual erosion by which paint coatings diminish in thickness (16). But the normal service of house paints, unlike that of automotive finishes, af fords little opportunity for mechanical removal of chalk by wiping or abra sion. The shrinkage in thickness of coatings of house paint begins long before chalking b evident, and pro ceeds primarily by decomposition of the vdiide with release of volatile products and extraction of soluble products by water. Mechanical loss of pigment is secondary. When chalking sets in, some of the dirt previously acquired b usually sloughed off, a process commonly called self-cleaning. The change from an oil-rich glossy composition to a pigment-rich flat composition makes the surface less able to retain dirt. Dirt dislodged by disappearance of the oil to which it was attached b not readily replaced by fresh dirt. The most suc cessful self-cleaning paints nearly always contain zinc oxide, and there fore yield zinc soaps rather dun free organic adds to their aqueous extracts (4, 8). It may well be that the zinc soaps have a detergent action that is lacking in the mote acidic extracts. Paints that contain colored pigments fade when they begin to chalk. This has long been recognized as an optical effect of a porous condition of the superficial chalking layer from which oil vehide has disappeared. Voids left between partides of pigment make highly reflective optical interfaces that restrict penetration of visible light and reduce the extent of selective absorp tion by the colored pigment. Other evidence of the development of voids in weathered coatings b found In studies of their swelling in water. Be fore they are weathered, glossy films swell at least as much, sometimes a little more than, the volume of the water absorbed, whereas films weath ered sufficiently by exposure to sun light and water usually swell less than the volume of water absorbed (5, 7, 8, 14). Weathering develops voids that can hold water without adding to the volume of the coating. It would be highly desirable for house paints to wear entirely away by erosion without any disruption of the coating or loss of adhesion to the sub strate. Repainting would then be re quired when the coating became too thin to continue to hide the substrate, and could be done without trouble some problems with broken and loos ened paint Although such wear by erosion b now obtained with some paints under favorable conditions, no paint can be relied upon to wear that way under all conditions of practical service. Zinc oxide retards the erosion of paint coatings. On vertical test-fence surfaces facing south at Madison, Wb., Forest Products Laboratory re search has shown that linseed oil house paints containing zinc oxide lose about 0.4 to 0.5 mil a year in thickness, whereas paints free from zinc oxide lose nearly 0.7 mil a year (17). These observations are explained in the data of Table 3 and Fig. 1, which show that sunlight is less effective in accelerating the production of both volatile and water-soluble decomposition products when zinc oxide b present than when it is absent The Fissure Stage: Young coatings are able to withstand the interna] stresses caused by shrinkage in volume and by swelling in water followed by redrying because they ran undergo a high degree of both elastic and plastic deformation (23). On aging, such distensibility diminishes at a rate deter mined by the progress of deterioration under die conditions of exposure (23). Loss by volatilization and leach ing of low-molecular decomposition products which have plasticizing ac tion, and increased cross linking of the vehide gel embrittle die coating (1, 32, 45). The internal stresses from changes in volume become ever greater 7 until the cohesion of the coating is ex ceeded. Fissures then occur in the coat ing to relieve the stresses. The fractures occur when the coat ing is dry. Even very brittle coatings are strongly placticized by absorbed moisture, whidi restores a considerable degree of distensibility (23). Fissures develop more often in dry summer weather than in winter, ana are more severe in dry than in damp climates. The following experiment is illustra tive. An exceedingly brittle paint com posed of titanium dioxide in sodium silicate was applied to matched edgegrain and nat-grain specimens of Douglas-fir and southern yellow pine. The smooth coatings remained intact and unbroken for a long time when kept at reasonably constant relative humidity and temperature. They also remained unchanged when exposed to damp air for 3 days, during which the wood gained 3 percent in moisture content and swelled accordingly. But within an hour after the specimens were restored to lower relative humid ity, the coatings cracked severely in a reticulate pattern entirely unrelated to the grain of the wood and long before the wood had lost much of the mois ture it had taken up. The cracking clearly originated in stresses within the coating rather than in movement of the wood under the coating. Fissures may take (he form either of checking or of cracking. In check ing, the fissures begin in a superficial layer of die coating but may work their way to the bottom of the coating later on. In cracking, die fissures ex tend at once throughout the depth of the coating. Checking usually occurs, if at all, at an earlier age than crack ing. Checking suggests a condition in which a superficial layer of die coat ing has become embrittled arid shrunken at a time when the deeper portions are still distensible and less contracted. Cracking suggests a coating of high strength in tension that has be come embrittled throughout before it yields at all. Coatings of white lead paint, which soon develop checking, retain distensibility for a long time even when exposed to sunshine and water, have only moderate strength in tension, and swell very little in water. Paint coatings containing zinc oxide, which imparts cracking tendencies, be come embrittled even in the absence of sunlight, have high strength in ten sion, and become greatly swollen in water. Paints in whidi early checking relieves internal stresses either may re main free from cracking or, if they do eventually crack, the cracking may occur in a smaller, less conspicous pat tern than in paints in which checking is absent and cracking is long delayed. Coatings applied by brush or doctor 8 blade possess grain that is absent when the coatings are applied by spray gun, especially if the paint contains acicular' pigments. The grain comes from orien tation of the adcular particles in the last direction in which the paint flows before it hardens. The grain of the paint is revealed in a preferential orientation of cracks in the direction of the last strokes of the brush (23, 32). It appears also in greater swelling across than with the last brush strokes when the films are stripped and soaked in water (4, 8) and in greater strength in tension along than across the grain when the films are tested mechanically (23). The mistaken belief that the direction of cracking is determined by the grain of the wood underneath arises from the fact that paint is cus tomarily applied with the last strokes in the direction of the wood grain. Checking, which begins in super ficial fissures, usually shows little effect of paint grain, develops with random orientation, and rapidly forms an in conspicuous reticulate pattern of fine meshes. The pattern of cracking is usually much coarser and more con spicuous. The. Stage of Disintegration: When checking penetrates to the bot tom of the coating, water gains direct access to the wood underneath. Adhe sion between coating and wood is greatly weakened by water (34). Moreover, the internal stresses within the coating reach a maximum at the in terface between coating and wood, which is the seat of external restraint on the free movement of the coating. There is soon a loosening of the coat ing along the edges of the checks. The coating then begins to crumble by de tachment, one by one, of the tiny poly gons outlined by the reticulate fissures. Paint cracks likewise admit water directly to the wood, and result in loosening along the edges of the cracks. Loose edges of paints that swell greatly in water curl outward mark edly, sometimes to the extent of form ing scrolls. Free films of such paints curl similarly when soaked in water and even more sharply when redried. The curling may be understood as a response to greater internal stresses in the superficial layer of the coating, where greater exposure has resulted in greater contraction in volume, than in the deeper layers of the coating. Curl ing tendencies are more marked with high-swelling than with low-swelling paints, but with high-swelling paints the curling can be minimized by for mulating the paint as closely as pos sible to its critical pigment volume. The loosened areas of coating, now free to change in area as well as in thickness, are often seen to be larger in area than the patch of wood from which they have been detached. This finds explanation in the fact, recorded in Table 2, that high-swelling paints tend to remain distended in area after they have swelled in water and'redried. Loose edges of coating eventually break away from the rest of the coating or are broken off by the elements and fall away as flakes. When disintegration sets in after checking or cracking, the wood begins to affect the course of further deterio ration. Loosening occurs much sooner over summerwood than over springwood, particularly so when the bands of. summerwood are relatively wide. There is reason to believe that young paint coatings ding to wood by spe cific adhesion--that is, by a bonding that involves the same polar, oxygen ated groups of the vehicle get that are responsible also for cross linking of the gel itself. The progress of weather ing may gradually release such woodpaint bonds as the bonds become in creasingly engaged in cross linking within the gel, and as plasticizers are lost and die adhesion repeatedly strained by swelling and shrinking of the coating. Specific adhesion may then be lost and the coating left dinging wholly by mechanical adhesion or em bedding in the pores of the wood (3). Springwood bands, by reason of their relatively thin-walled wood cells with large cavities, offer much greater play for purely mechanical adhesion than do summerwood bands, which have thick walls and small cavities. The Stage of Advanced Break-up: The progress of crumbling, flaking, or scaling depends largely on the nature of the wood surface under the paint and only partly bn the nature of the paint. As a rule, however, paints that ' tend to swell little in water and wear by checking and crumbling exhibit slower progress of sloughing from the summerwood and less unkempt appear ance in die stage of advanced break up than paints that wear by cracking in a coarse pattern followed by curl ing and flaking or scaling. Applications to Abnormal Paint Behavior Under some conditions of service, paints depart from their normal pat tern of behavior. Among the condi tions that lead to abnormal behavior are unsuitable spacing of paint coats or repaintings, unsuitable thickness of coating, incompatibilities between paints, and blistering. Unsuitable Spacing of Paint Coats or Repaintings: Much experience has shown that oil paints applied while the previous paint was still in the soiling or flatting stage often fail to perform normally. The condition frequently occurs when houses are completed late in the fall or winter, are given a prim ing coat of paint only, and the rest of the painting is deferred for several months until the weather moderates. Such jjaint jobs seem to be partic ularly inclined to early failure, often by intercoat peeling or by moisture blistering. The relatively high content of water-soluble substances and greater swelling capacity of young coatings may well be responsible. In addition jt has been found (12) that paint coat ings applied at unsuitably low tempera tures contain more water-soluble prod ucts and swell more in water than similar coatings applied at higher tem peratures. Much the same explanation may apply to early failures of repaint jobs on parts of a house that, by their posi tion, are sheltered from the full im pact of sunshine and rain, such as ceilings of porches and overhangingeaves. It is not unusual for new paint to come loose and scale from older paint or for coatings to become blis tered on the well-sheltered areas while the same paint is still performing nor mally on the fully exposed areas. Ab sence of the leaching effect of full rainfall, together with longer contact with still water in the form of dew, allows soluble decomposition products to accumulate in the protected coatings to a harmful extent. Unsuitable Thickness of Coating: Practical experience shows that most good house paints serve best when the coating is between 4 and 5 mils thidc. Modem paints will hide satisfactorily at half that thickness, but such thin coatings are more quickly penetrated by water, their content of soluble mate rial is greater, sunlight readies a greater portion of the film, and they have less reserve to waste away than do films of more suitable thickness (7). It is therefore not surprising that they erode unnecessarily rapidly.' On the other hand, unduly thidc coatings are prone to bad behavior (16, 17). Cracking often appears ear lier, is coarser in pattern and more conspicuous than is normal for the kina of paint, and very often is ori ented across rather than with the grain of die paint coating. The coarser crack ing leads to unsightly scaling that is unrelated to the stimmerwood pattern of tiie wood surface. Such scaling oc curs about as early on the woods that hold paint best as it does on the more exacting woods. For each kind of paint there seems to be a critical thickness of coating above which the abnormal behavior sets in. The critical thickness is as little as 6 to 8 mils for the paints that swell greatly in water and form brittle coatings of high strength in tension, whereas it is as great as 15 mils for low-swelling paints that retain plasticity longer and have moderate strength in tension. Modem practices in paint mainten ance tend strongly to build unduly thick coatings after a few repaintings. The paints in predominant use, which erode at the slower rate of 0.4 mil a year when fully exposed to the weathet, require 5 years or more to wear suffidently to accept one good coat of new paint without adding appreciably to tiie total thickness of coating, yet they are often repainted at intervals as short as 3 years of even less, some times with two new coats at a time. A few repaintings suffice to bring the coating thickness above the low critical thickness of such paints. The curious anomaly exists that, in former times, when the paints in predominant use lost 0.7 mil a year and had a higher critical thickness, the customaiy inter vals between paintings were consider ably longer and the abnormalities of unduly thick coatings were seldom en countered. Incomptabitities Among Paints: Abnormalities often develop when houses ate repainted with a kind of paint that differs too much in some of its properties from the paint used pre viously. The reason for such incom patibilities were obscure until the swelling effect of water on different aints became known. The junction etween coatings of markedly disim ilar swelling capacity necessarily be comes the seat of very high stresses when the coating absorbs water and dries out again. Moreover, the adhe sion of one coat to the other is im paired by water, just as is the case with the adhesion of paint to wood. Hetero geneous free films that consist of a coat of high-swelling and a coat of low-swelling paint curl into tight scrolls when soaked in water, and after somewhat longer soaking they will often blister and come apart easily. Such heterogeneous coatings on houses may crack, curl, and scale from the wood interface abnormally soon or, in more severe cases, the outer coating may crack and scale from the inner coating. Still another variant is alligatoring, in which tiie outer coat cracks in a reticulate pattern and then shrinks in area, so that the cracks open into relatively wide gaps and the polygon shaped islands of topcoat may become wrinkled. In at least one case, dissimilarity be tween coats proves advantageous rather than harmful. It is now recognized that, when suitable house-paint primers are used under finish coats that contain zinc oxide, the improved performance obtained comes in part from the fact that the primers are usually made with out zinc oxide and at relatively high pigment volume and therefore have tow swelling capacity. The low-swell ing priming coat buffers the swelling tendencies through the coating by min imizing tiie strain at the interface be tween wood and paint without too greatly increasing the strain at the in terface between paint layers. Blistering: Paint blistering can oc cur in several different ways. A neces sary condition that they snare in com mon is the development of pressure in a gas or liquid at the interface where the blisters arise. In temperature blistering, the pres sure comes from thermal expansion of air in a porous substrate under tiie coating. A fairly frequent example occurs when houses are repainted in weather when days are warm and sunny but nights are cool. New paint applied perhaps in the afternoon be comes blistered next morning soon after warm sunshine strikes tiie sur face. Air in the wood , expands with warmth and pushes through tiny cracks In the old paint to press against the new coating at a time when its outer skin has hardened but its deeper parts are still soft. Even well-weathered paint may be blistered if the surface is struck by a sudden blast of heated air. Practical use is made of temperature .blistering when old paint is removed with a painters' blow torch. In "glossy-back" blistering, the pres sure comes presumably from the gases evolved by deterioration of old paint that has been barred from access of oxygen and ultraviolet light by coat ings subsequently applied. Such blis ters are. characterized by the presence of a transparent, glossy, sometimes sticky material that looks like old var nish on the back of the blisters and on the surface underneath. The glossy ma terial is at least partly soluble in water, and seems to be identical with the material extracted from paint film by water. Events that finally result in glossy-back blisters begin with a yel low or brown discoloration of a layer of the old paint, followed by forma tion of holes with smooth, glossy walls that look under a microscope like the holes in Swiss cheese. The holes grad ually grow larger, join together, and finally form large cavities that become blisters.' The glossy material accumu lates in a relatively deep portion of die film, perhaps because the thickness of the coating makes its extraction by tain difficult The deposits may then prevent the ready escape of decomposi tion gases. There are times, however, when the glossy material exudes from paint coatings and discolors the ex posed surfaces or even collects in beads or droplets that can be easily mistaken for wood resin. The source of pressure in moisture 9 blistering has not yet been established satisfactorily. It may not be the same in all eases. Pressure may develop in the air in the wood under the coating from thermal expansion or when a wet line in the wooa advances toward the coating under a temperature gradient, thereby squeezing the trapped air against the hack of the coating. Soluble substances in the paint next to the wood may give rise to osmotic pres sure, with the coating acting as a semipermeable membrane. The source of pressure may differ with circumstances, and there may be more than one source in a given case. But unless water in the form of liquid reaches the wood-paint interface, the pressure required to blis ter well-dried paint coatings is usually very great, often of the order of 300 pounds per square inch (34). Water at die interface is necessary as a rule (47) for moisture blistering, because it weakens the adhesion so much that pressure as low as 15 pounds per square inch (34) may suffice. Water can reach the interface read ily if it enters in sufficient quantity from the unpainted edges or backs of boards. Movement toward the interface between the wood and paint is hast ened if the painted face is much colder than the back of the board, but mois ture blistering can develop without such temperature gradient (47). On houses, moisture may gain access be hind painted woodwork by coldweather condensation of vapor that originates inside the house. It may also enter from the outside by penetra tion through joints. The latter is prob ably the more common cause, but it is often overlooked because coldweather condensation has been so widely publicized. Blistering seldom occurs from water that reaches the paint-wood interface through the paint coating alone. Water moves through paint so much less rap idly than it does through wood, and there is usually so much more wood than paint present, that the dry wood keeps the paint at the interface reason ably dry for a longe time. Water does blister paint on non-absorptive sub strates such as glass or metal and prob ably does so on wood if there is time enough for the wood to come to its liber saturation point Even free paint films may blister when soaked for sev eral days. Tests on laboratory blistering boxes, in agreement with long observation of experience on houses, prove that some paints are more sensitive to moisture blistering than others. There is a close parallel between die degree to which paints absorb water and swell and their sensitivity to moisture blistering. The sensitive paints are the highswelling paints. Resistant paints are 10 low in swelling. When high-swelling paint is used for the finish coat, the sensitivity to blistering fa reduced if a tow-swelling priming coat fa inter posed between wood and finish coat. Conclusion Study of the mechanisms of paint deterioration reveals that water must be considered a major destructive agent along with ultraviolet light. Water alone, but even more in conjunction with ultraviolet light, accelerates dete rioration chemically by stimulating pro duction of volatile and water-soluble decomposition products that have much to do with the various stages of both normal and abnormal behavior of house paints. Water also attacks phys ically by swelling coatings and then causing severe internal stresses as the coatings dry and shrink again. The stresses disrupt coatings in one way or another. Water is also involved in the self-cleaning action of paints that has been so greatly emphasized in recent years. In the effort to make modern paints as dean as possible, they have been made increasingly sensitive to the harmful attacks of water. A current trend to the manufacture of paints that are more resistant to water is highly encouraging. But there is still need for more research on the mechanisms by which water attacks paint and means of controlling them, so that more reli able paints can be produced. Literature Cited I. Drying and aging of paint films. FATIPEC May 18-23, Noordwijk aan Zee, Holland (Paint & Varnish Pro duction, July 1954, p. 72). 2. Bishop, R. R., and Butler, P. 1958. Oxidative degradation of drying oils. Chem. and Indus. (8) 219. 3. Browne; F. L. 1931. Adhesion in the painting and in the gluing of wood. Indus, A Eng. Chem. 23: 290. I. Browne. F. i- 1953. Tire absorption of water, swelling, and suliihilitv of free films of paints. lout. FPRS III (5) 108. 5. Broune. I. I. 1951. Swelling of paint films in water. II. Absorption and voluirrtric swelling of bound and free hlms before and after weathering. Jour. FPRS, IV (6): 391. 6. Browne, F. L. 1955. Swelling of paint films in water. III. Absorption of bound and free films from air of dif ferent relative humidities. Forest Prod. Jour. 5:92-6. 7. Browne, F. L. 1955. Swelling of paint films in water. IV. Effects of film thick ness and of pigment volume. Forest Prod. J. 5: 142-6. 8. Browne, F. L. 1955. Swelling of paint films in water. V. Effects of different pigments. Forest Prod. J. 5: 192-200. 9. Browne, F. L. 1956. Swelling of paint films in water, VI. Effects of different oil and oleoresinous vehicles. Forest. Prod. J. 6: 152-8. 10. Browne. F. L. 1956. Swelling of paint films in water. VII. Effects of different latex vehicles. Forest Prod. J. 6: 235-40. 11. Browne, F. L. 1956. Swelling of jaint films in water. VIII. Swelling of lin seed oil paints in different liquids. Forest Prod. J. 6: 312. 12. Browne, F. t. 1956. Swelling of paint films in water. IX. Effects of tempera ture of soaking water, temperature during formation of films, and re peated swelling and drying. Forest Prod. J. 6: 453. 13. Browne, F. 1. 1957. Swelling of paint films in water. X. Permeability to water, water vapor, and air. Forest Prod. J. 7: 145. 14. Browne, F. L. 1957. Swelling of paint films in water. XI. Mixed-pigment paints. Forest Prod. J. 7: 248. 15. Browne, F. L. 1957. Swelling of springwood and summerwood in softwood. Forest Prod. j. 7: 416. 16. Browne, F. L. and D. F. Laughnan. 1952. How often should a house be painted--an experimental study of programs of maintenance. Jour. FPRS. II (5) :173. 17. Browne, F. L, and D. F. Laughnan. 1953. Effect of coating thickness on the performance of house paints under dif ferent programs of maintenance. Offi cial Digest, Federation of Paint 8c Varnish Production Clubs, (No. 338): 137. 18. Carrick, L. L, and W. J. Snoddon. 1951. Depth to which oxygen pene trates drying oil films. Official Digest, Federation of Paint A Varnish Produc tion Clubs, (No. 322): 682. 19. Carrick, L. L., and A. J. Permoda. 1951. Shrinkage of some organic film ' forming materials during aging. Fed eration of Paint 8c Varnish Production Clubs, (No. 322): 692. 20. Crecetius, S. B., R. E. Kagarise, and L. L. Alexander. 1955. Drying oil oxidation mechanism, film formation and degradation. Indus. 8c Eng. Chem. 47: 1643. 21. Crecelins, C. B., and L. L, Alexander. 1956. Mechanism of linseed oil film degradation under ultraviolet irradia tion: Naval Research Laboratory, Re port PB 121330. (Abstracted, Amer ican Paint Jour. 41 (No. 17): 34). 22. Elm, A. C. 1949. Deterioration of dried oil films. Indus. 8c Eng. Chem. 41: 319. 23. Elm. A. C. 1953. Some mechanical properties nf paint films. Official Digest. Federation of Paint A Varnish Production Clubs. fNo. 346): 751. 21. -Elm, A. C 1957. Recvalnation of the function of pigments in paint. Federa tion nf Paint A Varnish Production Clubs. fNo. 387): 351. 25. Fitzgerald, E. B. 1953. Deterioration of alkyd resin films. Indus. A Eng. Chem. 45: 2545. 26. Fitzgerald, E. B. 1955. Photooxidative degration of alkyd films. American So ciety for Testing Materials, Bull. No. 207: 65. 27. Forest Prod. Lab. 1955.."Wood Hand book", Agriculture Handbook No. 72: 355. U. S. Dept of Agric. 28. Gay, P. J. 1948. Measurement of wet ness of exposed paint films (Climatic factors in paint breakdown). Jour. Oil A Colour Chemists Association 31: 481. 29. Haslam, J. H. and S. Werthan. 1931. Studies in the painting of wood. Indus. A Eng. Chem. 23: 226. 30. Hutchinson, G. H. 1958. Some recent advances in the chemistry and tech nology of drying oils. Jour. Oil A Colour Chemists Association 41: 474. 31. Jenkins, J. D., and P. R. Croll. 1924. 38. MaynejJ. E, O., and D. Van Rooycn. 45. Talen, H. W. 1956. Effects of radia Some observations on the dirt collec 1954. The mechanism of the corrosion- tion humidity, and atmospheric condi tion by painted surfaces. Paint Manu inhibitive action of paints, with special tions on the mechanical properties of facturers Association of the United reference to basic pigments. Jour. vamish and paint films. Veifinistituut States, (now National Paint, Varnish Applied Chem. 4: 384. T.N.O. Circular No. 83, 14 pp. 8c Lacquer Association), Scientific Sec tion Circular 219: 137-43. 32. Jordan, L. A. 1936. Structural features of paint films. Paintindia 6 (1): 63. 33. Knudsen, K. E. 1956. The mechanism of the specific rust-inhibitive action of plumbates in linseed oil. Jour. Oil & 39. Nelson, H. A. 1933. Metal priming paints. Inhibiting qualities and in fluence of reactions within the paint film. Indus. 8c Eng. Chem. 27: 33. 40. Nelson, H. A., and G. W. Rundle. 1923. Further studies of the physical properties of drying oil, paint and var 46. Twiss, S. B,, W. L. Weeks, and F. O. Thomas. 1958. The role of pigments in the weathering of automotive paints. Offidal Digest, Federation of Paint 8c Vamish Production Clubs. 30 (403): 788. Colour Chemists Association 39:909. nish films. Proceedings, American So 47. Veer, J. J. G. 1957. Miosture blistering 34. Kuzraak, J. ML. and P. J. Sereda. 1935. ciety for Testing Materials 23 (H): of paints on wood. Foiest Prod. Jr. The blistering of paint in the presence . 356. (10): 342. of water. Canadian Jour, of Tech. 33: 67. 35. Long, J. S. 1949. Film formation, film properties, film detrioration. Official Digest, Federation of Paint 8c Vamish Production Cubs, No. 297: 648. 36. Long, J. Si A. E. Rheineck, and G. L. 41. Rhodes, F. H.. and A. E. Van Wirt 1923. Effect of various pigments upon the rate of oxidation of linseed oil. Indus. 8c Eng Chem. 13: 1133. 42. Saur, R. L 1955. Photochemical de gradation of automobile lacquers. American Society for Testing Ma- 48. Wirshing R. J. 1937. A new approach to the degradation of organic films. Presented at Gordon Research Confer, ence on Organic Coatings at Kimball Union Academy, Metidan, N. Hi Aug. 22. Ball. 1933. Studies in the drying oils. rials, Bulletin No. 207 61: 49. Bullett, T. R. and A. T. S. Rudram. XVII. Influence of several factors on 43. Schierholtz, O. J. 1943. Test results 1959. Polymer structure and film for 3 the mechanism of drying of oil films. Indus. 8c Eng. Chem. 23: 1086. of films exposed to extreme moisture. Canadian Paint 8c Varnish Msg. 17 mation. Jour. Oil 8c Colour Chemists Assodaiton. 42: 778-97. 37. Marshall, Nissim J. 1937. The effect (11): 10, 12, 39. 50. MacGregor, John R. 1939. New con of light of different wave lengths on 44. Shelberg, W. E, J. L. Madrid, and cepts of fundamental causes of paint the degradation of clear coatings. Offir R. K. Fuller. 1934. Artifidai surface perishing. Official Digest, Federation of ciat Digest, Federation of Paint 8c dirts for detergency studies with painted Paint 8c Vamish Production Clubs, No. Vamish Production Clubs ( 391): 792. surfaces. Ind. 8c Big. Chem. 46: 2572. 173: 77-8. j h sir u