Relating laboratory and outdoor exposure of coatings: IV. Mode and mechanism for hydrolytic...

Relating Laboratory and Outdoor Exposure Of Coatings: IV. Mode and Mechanism For Hydrolytic Degradation of Acrylic- Melamine Coatings Exposed to Water Vapor In the Absence of UV Light T i n h N g u y e n , J o n M a r t i n , a n d E r i c B y r d N a t i o n a l Ins t i t u te o [ S t a n d a r d s a n d Tech r~o logy *

INTRODUCTION

T hermoset acrylic-melamine resins are widely used for automobile exterior coatings. These materials are furmed by reacting an acrylic polyol with an

alkylated melamine. However, the ether crosslinks of acrylic-melamine coatings are known [o be susceptible to hydrolysis when exposed to moist enviromnents. Under outdoor and artificial acid rain environments, acrylic-melamine coatings undergo etching, which is primarily a result of' acid hydrolysis of the crosslinks. ~-5 A particular attribute of etching is the localized loss of material, resulting in pi tdng of the coating surface. In the presence of ultraviolet (UVI light, the hydrolysis process becomes more complex. For example, the hy drolysis in Miami, FL, has been observed to be faster than that in Phoenix, AZ. 6 Numerous studies using controlled environments also sho~ved that. in the presence of UV light, the hydrolysis rate is greater than the sum of dark hydrolysis and photolysis combined. ~ ~z The enhanced degradation, which follows closely wi th the hydrolysis rate and relative humidi ty (RH) levels, not on ly occurs at the act.vile melamine crosslink and in the melamine chains but also on the acrylic resin structure. r' Enhanced degradation in UV/humid i tv conditions has been variously attributed to melamine excited state chemistr S oxidized products of formaldehyde, ~ catalysis by carbox~lic acids produced by the photooxi- dation, 6 and chromophoric activi%, of formaldehyde molecules released f'~onr the h2drolysis reactions. ~:~

The most comprehensive s tudy of the hydrolysis of both fully and partially alkylated melamine coatings in the absence of UV light ~,as conducted by Bauer. ~ He showed that the rate of hydrolysis depends on the type and concentration of acid curing agents, hydrolysis temperature, and type of melamine resin. During hy drolysis, the ether crosslinks are broken and the melamine-melamine linkages are formed, and coatings

Te, :hn : , ! : ,Zv . ,~ ~:,~emb,_-]" !, 7 .2( , ( ,L iu / : , , l la ] l t ,~ . ',_-'/-\ § B ] i ' , -~u T , r . h/[aft St. . I, ;-~l " ~ .~a ~ ~ , b , r : . ML. , 2nb~:lg.

bff ~," ..., ,!!',ff'/',~,'~,5',' ',,~'~.~;~"',":,q "~';,":~'~'~,,~;~",,".;'"~':~il ~!i' ,'),~!~"s~:':d,4,~ ',' }#`I~%~.~.:@`~`~.~`~..``~`.~..!1s:`~..t`?`~y`~1`~`.::`~\~id`~.~` ~ ~,~L~!'~,~':'~:,':,,~ ~/~ f~":~,~ ",r ~,~: =', :,~,.'~ ,,~,<.~,~'~!~:.~,,,~:'.~,' ;'.. :.,~,.~.,,~:~,'~ !,It~ ~",,<~<"'~"41 ~' '~ ~!i ''~' ~: "'~;~ ~ ,' >'~':"~.'!',',,,,',~::' '~", L,,~m'.

Ac,71ic melamine coatings are known to be susceptible to hydrolysis when exposed to water or humid environments. The mode and specific pathways for hydrolytic degradation of acrylic- melamine coadngs exposed to water vapor in the absence of ultraviolet light are presented. Samples of a partially methylated melamine-acrylic coating applied to CaF~ substrates were subjected to five different relative humidity ]eveJs ranging from approximately 0 to 90% at 50~ Coating degradation was measured with transmission Fourier transform infrared spectroscopy (FTIR) and tapping mode atomic force microscopy (AFM). In hunfid environments, partially methylated melamine acrylic coatings undergo hydrolysis readily, causing considerable material loss and formation of mainly primary amines and carboxylic adds. The rate of hydrolysis increases with increasing RH. Hydrolytic degradation of acrylic-melamine coatings is an inhomogeneous process in which pits form, deepen, and enlarge with exposure. Such localized degradation mode suggests that hydrolysis of this material is an autocatalytic progression where add& degrad& don products formed in the pits catalyze and accelerate the hydrolysis reactions.

Vol. 75, No. 941, June 2003 37

T. Nguyen, J. Martin, and E. Byrd

. oc~ [I /

~ N Where R is:

N N N H < C ~ "-y -.,

W" ./~'-,~ ~oCH,O.~,ocr--%,

C||3OCI.[ 2 [~

the partia/(y methy/ated melamine-acrylic | coatino used in thg study, l

based on part ial ly a lkylamd melamines hydro lyze much faster than do fully alkylated coatings. In mild acid conditions, the crosslinks of methylated melamine based coatings hydrolyze faster than the unreacted methoxy groups of the mela~q~ine.l~ t-towever, in strong acid solutions, the melamine rnethox~ has been report ed to hydro l ) ze first, fol lowed by hydrolysis of the crosslinks. 4 Hydrolysis rate dependence of melamine alkox) ~ ether on the chemical structure and pI-{ is clearly revealed fiorn solution studies by Berge and cowork ers. 1'H5 In mild acid solution, methylated tr imethylol melamine (a partially alky[ated melamine) hydrolyzes approx imate ly 45 t imes faster than hexamethy lo l melamine (a fully alkylated melamine) while, in strong acids, the latter hydrolyzes almost 30 times faster than the formen

Despite such extensive research on the etching and hydrolysis of acrylic-melamine coatings, little information exists on the mode for the hydrolyt ic degradat ion of these materials. Further, a l though general hydrulysis

, [o schemes for acrylic n~elamine coaiings are available, " specific hydrolysis pathways for these materials are not known. This s tudy presents the rnode and pa thways that form carbo,~lic acids and pr imary amines ['or the degradat ion of coatings based o~ a par'dally methylated melamine and a hydroxy- terminated acrylic resin exposed to water vapor in the absence of UV light. (By mode, it is meant whether the degradat ion is uni form across the surface or is concentrated at specific locations so as to result in pitting.) Specific hydrolysis pa thways are presented based on Fourier t ransform infrared (FTIR) spectral evidence, and the degradat ion mode is proposed based on atomic force microscopy (AFM) nanoscale images of the coating microstructure and its smface morphological changes dur ing hydrolysis. In addition, a water vapor sorption experiment was car- ried out to provide information on the water concentration in the film and the nature of interactions between water and acrylic melamine coatings at different RH levels. Information on the mode and mechanism of the hydrolytic degradat ion is essential for unders tanding the controlling factors responsible for coating failures dur ing service arid for helping to develop more durable

coatings. This research is part of a larger effort at the National Institute of Stasdazds and Technology to develop metrologies for relating outdoor exposure performance with accelerated testing results and to devel op improved rnethodologies for predicting the service lives of polymeric coatings.

EXPERIMENTAL CONDITIONS AND

PROCEDURES ~

Materials Materials and sample preparat ions procedures were ment ioned previously.u Brie fly, samples were prepared from a mixture of a hydroxy-.terminated acrylic resin and a partially methyla ted melamine resin. The acrylic material contained 68% norrnal bubqmethacrylate , 30% hydroxy ethylacrylate, and 2% acrylic acid (all are ex pressed in mass fraction), and the melamine resin was Cymel 325 (Cytec Industries). The ratio of acrylic resin to melamine resin was 70:30. It should be noted that, other than the 2% organic ac~l ic acid, no strong acid catalyst w, as added to this model coating formulation. Calcium fluoride (CaF z) plates wi th a diameter of 100 rmn and a thickness of 2 rnrn were used as the substrate for the exposure s tudy due to their t ransparency [n the 0.13 to 11.5 ~m optical region and their resistance to moisture and heat. te Coatings were applied to CaF z plates by spin casting. AcD'lic and melamine resins in respective solvents were mixed at the required ratio, de- gassed, f looded onto the substrates, and spun at 2000 rprn for 30 see. Coated samples were then cured at 130':C for 20 rain. All samples were felly cured, as evidenced by the lack of further reactions after prolonged heating at the same ternperamr~. The cured film had a thickness of ]0 p.m • 1.2 ~rn (the number after [he • sign indicates one s tandard deviation) and a glass transition temperature (T a) of 45':C _+ 2:C (by differential scanning calorimetw). Figure l shows the postulated chemical structure of the cured act)'|ic melamine films used in this study:

Instrumentation and Exposure Conditions Complete details on the exposure cell, instrumenta

t[on, and controlled systems are given [n reference 17. This section briefly describes the instruments and experimental procedures pert inent to this study: that is, hydrolysis at different relative humidit ies in the absence of UV light.

The exposure cells were designed to simultaneously expose different sections of the same film to different conditions of RH, UV, and temperature (T). Nominal R [-I levels of << l, 20, 40, 70, and 90% and at a tempera- rare of 50'-'C were employed for this study. Figure 2a dis plays a pho tograph of three exposure cells along wi th the tubing system for carrying air at the desired RH and temperature, and F~gure 2b schematically shows the

~i-,- ] '~di l l i:i: ~lI]Ini~iIi[ia! ~liIiIii-] U I : ~ IY i}l~ljj}Iiin-ii~]]~ j r ' i } IjI~OI:YI~,I~I" !lJ [ h l ' ]:l~}-ii~l" i~'i i:iri~l-]" ~i: I S~II}I[ -

i.[.,. , Ol~lel~i ]o[ci~ [!,l~ : '~h. ierhq] ie! , td] i : . r : , . :e , :Lur~ h , no, .:='.,:, ,:!.:-:c cu,- ' l , bJ , : h t f f 1 , : ~d , : n iu~ i . ] ) , [~:,:,_.[r! nn ,H id ]<h: .n ,:.r ,? i l . / . : . rs,Hn>?nt b y t ]h : r ' ]ad , : . i la ] [ n s t i h J t e ,[ S t3r i ,%)r , - s ar,,q T}, : !J, l . : , ] , : .~, ; u : : J,:.,% i~ i m p l y th ,~ t i l E l~.~.:,~ssaii l 3' I ln~ ~.,?st a-esi la~.l,~ h:,," t h ~ ] ,d r l : , : , sm

38 Journal of Coatings Technology

cross-section of the exposure cell. Each cell consisted of a 12-window aluminum fiame, a quartz plate, and a 100-toni diameter CaF.? plate wRh a film elf" the coating applied to It. One window was completely covered wltIt an alunllnum plate to prevent radiation from impinging upon the coating Cdeslgnated as no-UV or covered), one exposed to full UV spectrum, and 10 col~tained Interfer- ence filters transmitting different spectral ranges and In- tensities of UV light. Care was taken to ensure that no radiation strayed onto the no-UV {covered) coating area. '[he space between the quartz plate and the coated-C aF. plate forms the salT, pie exposure chamber. Each exposure cell contained an Inlet and an outlet that al- lowed fiesh air of the deshed ~ and RH to continuous- ly enter the sample exposure chamber and let. the outgo- Ing air vent m the outside of the chamber. Each exposure cell was equipped with a t_l-,ermocouple and a chilled- mirror hygrometer to monitor T and RH. respectively. It should be noted here that. during the earl>" stage of exposure, essentINly no photodegradatlon was observed In the spedmens beneath the filters. The results of these specimens were included in the analysis of hydrolysis rates. For that reason, the description of specimens under the interference filters ts Included tn this section.

"[wo tern peraa~re--humldlty generators, based on the principle of dry.,molsture-saturated a~r mixture, supplied the desired relative hurotditles. I,: Each tempera- t_ure-hun-ddl .W generator was capable of providing air at one temperature and up to ~'our different RH levels. The Rt-{ and T in each cell were controlled by a tbedback system. which tracked Rt-{ and T three times per second. Rt-[ and ~[ in each exposure cell could be Independently controlled and maintained to within • 0.5~ and + a%. respectively, of the preset values.

W a t e r S o @ r i c h M e a s u r e m e n t

Free standing flln-t spedmens from the same coating formulation having a dimension of approximately 0.15 . 3 �9 6 mn-t were employed for water sorptlon measurement at different relative buntiditles. Free films were prepared by drawdown on poly(ethylene tereph- thalate) sheets attached to an aluminum substrate. Spacers were used to control the film thlck~ess. ~[he films were cured at 130=C ' for 20 rain. similar to that used for the hydrolysis study. The sorptlon was performed In a H |dden ISAsorp Moisture Sorptlon Analyzer, which is equipped with a nllcrobalance hay- tug a resolution of 1 �9 10--rg. The instrument can acvu- rarely control the environmental chamber to within - i% of the preset values of RH from 0 to 90% and 4- 0.05~C of T from 5 ~ to 70 ~c . respectively, q2he sorption Isotbern-~ was generated using the Isothermal mapper mode. In this niode, the Instrument Is programmed to record the wa~er uptake at the higher specified RH after reaching ~ e equilibrium uptake of the lower specified RE[. In the Isotherm mode. only o~e specimen Is re-- qulred for the whole RH range. Be{'ore starting kite ad- sorption process, specln-,ens were dried at 6WC In the analyzer using the dwlng mode until no further decrease tn the sample mass was observed. Ligless other- wise stated, the results were the average of the three spec|mens.

P e l a t ~ n g I_at:_,,:,rat,z,r\~ a n d O u t d o , : , r Exl :_, ,z ,sure , : , f C , : , a t ~ n g ~ - : IV

Atomic Force Microscopy (AFM) Tapping mode AFM was used to provide information

on nanoscale surface mlcrostructure and morpholog-y of the coating betbre arid during exposures, in AFM. a sharp tip located at the end of a cantilever beam Is scanned across the sample surface. For tapping mode AFM. the cantilever Is oscillated at a frequency near its resonar, ce. typically, a few hundred kHz. so that the tip makes contact with the sample only for a short duration In each oscillation cycle. As the tip approaches the sam- pie. the tip-sample interactions change the amplitude. resonance frequency, and phase angle of the oscillating cantilever. Monitoring changes in the amplitude and phase angle would provide the topographic for helght) image and phase image, respectivel> Phase angle shift is sensitive to changes tn sample properties. ~fherefore. phase liT, aging oReri provides significantly more contrast than topographic Image and has been shown to be a useful technique for studying surface heterogeneity with nanoscale lateral resolution] s

AFM measurements were performed using a Dimension 3100 Scannhig Probe Microscope fDIgital

- , , ii , ~t I I

' " ' , : ~'~ ' t',',e~f ~ , _ , ; ,~ ~,a-" ': , . " . 'i, L e ,r ' ,~ i

I N ;C I ~ 1 L,~'!: ,4: t ~ ,T,, r

'1 4 , : ~ I il ~ i

,~ - ,-.,.. ,,'

: :!I,<, . . . . . . . . . ~l .......... ,, ........... ~ ,,.: ......

i .i "~ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ! . . . . . . I i, ' ,

@,

I 6,'.L~U,r-C-, ~ ' - - (9 , ) /~ C,hc~[C'qY;yLL,/? O f ::c~'J~731 @k-If','2,- ~ . '_.t ,'7~: C -C_~!,~F u,~';i~'6'[email protected]~;.17 t.~ -C_~ !,'C,'/] i_ '_.C.I. ,'j C @ ~ I?.2' { t, ) / C,,",'7,:;'_,-5, _~,_- ?_,',7,[7 ,_Jr 3[-,. E'$7,0:;t//c:' C d,,'i ?]qG1V~'[7~ ~ 217C: ,_~/Jglr2" .s ~./7-_--o ,L,L~st~=o,_-J ~U~g.-.., ~r,'o." ~ C ~

'v',-,i. 7~_ r . j : ' n 4 I . . ,' ,u ,~ ,_- - 'C,C.,~ c,,-,.. =,

l , N g u y e n , l , M a r t i n , a n d E, B y r d

Instruments) unde r ambien t condit ions (24':'C and 45% RH). AFM images taken dur ing exposure were obtained f iom the same specimens used for FTIR analysis. A spe ciai holder was fabricated so that the 100 m m d iamete r coated sample could be secured repeatedly in the same orientation. AFM images ~evere taken approx imate ly at the same sample area for every measur ing tirne by using a reference mark and controlling the x-y stage, More details on this procedure are described elsew, here. I~ Both topographic and phase images were taken using a t app ing force of between 50 and 70% of the free amplitude. Silicon tips having a drive f requency of approxi-- mate ly 300 kHz and a radius of approx ima te ly 5 n m were employed.

Fourier l ransform InFrared Spectroscopy Analysis Coating hydrolys is was fo l lowed by FTIR spec

t roscopy in the t ransmission m o d e using an au tosam piing accessory, At each specified time, coated CAP,, plates w.'ere r emove d f rom the exposure cell and fitted into a demountab le 150 rnm diameter ring of the auto sampl ing device. The ring was computer-control led and could be rotated and t ranslated to cover the entire sarn- pling area. Spring loaded Delrin clips ensured that the specimens were precisely located and correctly regis tered, This a t t tomated sampl ing sys tem al lowed unat-- tended, efficient, and speedy, recording of the KI'IR trmrsmission spectra of the coating at all 12 w i n d o w s befbre or at each exposure time, A pho tog raph of the autosampler loaded wi th specimens is g iven in reference ll. Since the exposure cell was m o u n t e d precisely in the autosampler , error due to var ia t ion of sampl ing at d i f ferent exposure t imes was essentially eliminated. The specimen-conta ined au tosample r was placed in an FTIR spect rometer c o m p a r t m e n t equ ipped wi th a liquid ni trogen cooled m e r c m ) / c a d m i u m telluride (MCT} detec tor. Spectra were recorded at a resolution of 4 cm 1 using d ry air as the purge gas. All spectra were the average of 132 scans. The peak height was used to rep resent IR intensity, which is expressed in absorbance+ A. Except for the << 1% RH exposure condit ion where four specimens were employed, only one specimen was used at other RH levels.

EXPERIMENTAL RESULTS

Stability ol Exposure Relative Humidity and l empera tu re

Relative humidi t ies p roduced by the me thod of mixing dry' air and mois ture-sa tura ted air were stable dur ing

the course of exposure, even at elevated temperatures . The maxin lun ~ c oeffic lent of variation, i.e., (standard de v ia t ion/ rnean) :.: 100, of the five R[-I levels s tudied was 3% dur ing the six months of exposure. The coefficient of var ia t ion for the 50':C tempera ture was < 1~ 5, The RI--I at 0.002% (designated as << ]% or very dry') was est imat ed fi'om the Table of Vapor Pressure of Water, 2~ given the 50':C exposure t empera ture and the -?0':'C dew point. ']?he RH sensor; i.e., the chilled mirror hygrometer , was not sensitive enough {o record this low RH.

Water Sorption Characterist ics

Figure 3 displays the equi l ibr ium wate r up take as a function of relative humid i ty (sorption isotherm) at 50:C for the model acrylic me lamine coating used in this study, In this figure, the moisture contents [% equi- l ibr ium mass of water uptake wi th respect to the d ry mass of the coating) at 50, 60, 80, and 95% RH were ob rained f rom only one specimen. [-Iowever, the da ta point for each of the other five RH levels was the average of three specimens. Figure 3 includes the one s t a r t

dard deviat ion error bars at the mult iple sample da ta points, which show that variat ion between the samples is small. Equi l ibr ium mois ture contents and water solu- bilities of the coating at the five relative humidi t ies used in this hydro lys i s s tudy are p resen ted in table 1. Solubilib, is expressed as moles of wa te r (i.e., equilibrium mass of wa te r up take d iv ided by the molecular mass of w, ater) per ]00 g d ry coating film. The water v a por adsorpt ion in these acrylic--melamine coatings is al- mos t linear be tween zero and 50% RI-{ but increases more rap id ly thereafter, reaching a value of about ].8% at 90.8% RH.

The sorpt ion isotherm shown in Figure 3 provides essential infom~ation not only on the hydrolysis kinetics (a function of water concentration) but also for under s t and ing h o w wa te r sorpt ion characteris t ics in an acrylic melamine coating varies wi th REI change. The i so therm in Figure 3 is of the Type I[I in the BET IBrunauer, Emmett , and Teller) classification, zl Except for the magni tude , the shape of this curve is similar to those of cellulose acetates, 2z aciylic polymers , ''3 and polyarnides 2~ at 50':C. For this type of sorpt ion (up~,ard curvature) , the heat of adsorpt ion of the wa te r v a p o r on an adsorbate is s imilar to or less than the heat of con- densat ion of" the water vapon ':'1 For the acrylic melamine material , this means that the interaction between wate r molecules themselves is energetically similar or greater than that bet~,een w a t e r v a p o r and the acrylic me lamine material. Based on the shape of the curve in F]gure 3, it is suggested that muir[player adsorpt ion ("clustering") of wa te r has occurred in this acrylic-

Table 1--Moisture Content and Water SNubility at DiFferent Relative Humidities el the Acrylic-Melamine Coating Used in this Study

0,002 19,2

1~,4ois[ure c o n t e n t (%) . . . . 0 .009 �9 0.01 " 0 .23 + 0 .007 So lub i l i t y . . . . . . . . . . . . . . . ...0.0005 • 0 .0007 0 .013 - 0 .0004

(Mole H.O/'IO0 g r-oatino)j

Relative Hurnidifty, %

,,~) rh~b r !u I ni-,,~l d [h,Er tJ],5 e S'.,~IT, I:,,7,i i r l ,41oat ,53 C,r vP Star ,,J~S, 4 , I F / q a t h :rl.

41,7 72,7 90,8

0.53 _~: 0 .008 1.16 _~: 0 .08 1.84 + 0.06 (%) 0.03 , 0 .0004 0.06 , 0 .005 0.01 ~ 0 .003

4 0 I o u r n a l o f C e a ~ i n g s T e c h n o l o g y

Relating Laboratory and Outdoor Exposure of Coatings: [\/

melamine coating. However: the deviation from lineari ty [i,e.. f rom the }-|en .ry's law behavior) of' the curve in Figure 3, a l though substantial, is not radical, Such b, pe of curves, i.e., small positive deviat ion fiorn the Henry ' s law,, generally indicates that the number of water layers adsorbed in the po lymer film does not exceed the dimer level.'/z."3

Atomic Fo[ce Mic[oscopy Figure 4 displays 500 nm:< 500 nm topographic and

phase AFM images of an approximately 50-gm thick coating before exposure. This sample was prepared b) spin casting of the coating in solution on a silicon sub strate. The cast film was cured using the same schedule used for FTIR analysis samples, i.e,: 130:C for 20 rain. The images o[" Figut'e 4a w, er~ taken horn the surface ex posed to the air (exterior), and those of Figure 4b were obtained from the surface that was in contact wi th the silicon substrate [interior) dur ing film formation. The interior surface sample was prepared by first initiating a cut using a razor blade: follo~ved by peeling using tweezers, It should be noted that AFM images taken at high magnification show that the silicon surface is essentially featureless, wi th no evidence of a surface pat tern. For both i"~gure5 4a and 4b: the topographic images are on the left and the phase images are on the right. The bright and dark area~ in the topographic images corre spond to the surface peaks and valleys, respectively. Except fbr the surface topographic pattern, the height images reveal little itfformation about the microstructure of" the interior or exterior surface. Ho~,ever, there are two characteristic features in the phase images that need to be discussed because they provide essential data for the analysis of the degradat ion mode.

One feature is that the acrylic melamine coating has a two-phase microstructure (/figure 4b), consisting of a matrix that appears bright and interstitial regions dis persed throughout the matrix that appear dark. The 500 nm regions that were imaged were very fiat, wi th topographic changes of less than 5 nm for both the intelior and exterior surfaces. Further: there is no direct correla tion between the topographic variations with dark or bright areas of the phase images. On the other hand: the phase angle difference between the bright and dark regions in the phase image of Fi@ure 4b is substantial (50:'). These results indicate that the proper ty differences between the matrix and the interstitial regions of the coating are significant. High resolution AFM images of the pit ted legions (i.e., microstructure beneath the outer surface laver) of this coating after exposure to UV/hu - mid enviror~ments also showed a two-phase microstructure. t9 resembling that s h o ~ n in F]ffure 4b. Further, our extensive AFM data of" crosslinked arr~ine cured epoxies also reveal that, while the exterior surface of the epo,x~ samples appeals featureless (similar to that observed for acrylic melamine), the two phase microstructure of" the interior surface is similar to that observed for the c ryogenic~ly fractured surfaces (bulk) and the UV-degraded regions of" these materials. These results suggest that the rnicrostructure of the interior surface (the sur face that was in contact with the silicon substrate dur ing film formation) more truly represents the microstruc-

2*it l

[:::1

NBI~v~ NUmld~, %

I Fi,qure 3 Sorption isotherm at 50'~C for the acfylic-meiamine coating (one .standard de-

! viation error ba~s are given for the five RH lev- el5 u.sed in thi.s .study).

ture of the bulk than that of the air surface. This is because the outermost layer of the outer surface is covered with a thin layer of" a homogeneous low surface flee en ergy layer, as explained later.

Phase differences bet~veen regions of a surface recorded by an AFM have been generally at t r ibuted to differences in the mechanical a n d / o r chemical properties of those regions} ~ Softer, more compliant phases tend to be dark w, hile harder, less viscoelastic phases tend to be bright, z5-27 Following these assignments, it is believed that the act~-lic-melamine coating used in this s tudy is a heterogeneous material consisting of softer re gions (dark) dispeised in a harder matrix [bright). A similar he terogeneous microstrucEure has been observed by AFM for crosslinked amine-cured epoxies: zs.'/s and crosslinked polyester. 2"~ The microstructure hereto geneity of coatings has a great implication on their degradat ion mode and mechanism dur ing service.

Another feature of Figure 4 is the higher contrast and more well defined rnicrostructure of the interior surface as compared to that of the exterior surface, This difference is also indicated by a difference in the phase shift, which is only ] 0': for the air side but 50: for the substrate side. Such morphological differences between the inte-- rior and exterior surfaces have also been observed in other po lymer systems. 26.~~ It should be noted [hal be cause AFM operates in the near field (i.e., very close to the surface), it detects only the top layer of rigidly bound atoms and, thus: does not provide an image of the layer beneath the outermost surface. Cher~ical analysis by X-ray photoelectron spectroscopy (XPS) of these acrylic-melamine samples shows that the exte dor surface contains 6(<7% C and 8.0% N, as corr~pared to 64.8% C and ] 2. ]% N for the interior surface. From these results, i.e,, the low phase shift and less-defined microstructure (by AFMI, a lower concentration of the po- [al N containing groups, and a higher concentration of the low electronegativi b, (related directly to polarity) C element, it is suggested that the surface exposed to the air of this acrylic melamine film is probably covered

Vol, 7b, No, 941, lune 2003 41

T r'Jguyen, J, P4artm and E, Bvrd

b

F-,,!:?,'jro .~--. 5(JO ~ ' m >" ~;00 nr,<. ,:,_sDo,yrs,.2,.~r Gn,_-~ fl..,.l?~ 5@ ,.~ i--I ,~ d lr ,::',,f ,.'21~ § :9r ,, ","!,',_'_'2-,r rr162 ~,"( r ,:r ,r162 1'3)@,','r~-~,QO,t-::UI'f&C@ &'l-,.'_l '(?') !G'- ~2@Y,'o,r '5 iFf&C@, II-~ ~_,O~,~-1 ~ @*-2 r 'E', ~5@ baE 'o - ~ &[:.,lYu- ,'n-,O~]_4@ ~ ,:~n' ~]-,.0 ,'-@f~ ~ 5 5 r LH.,@ f:-.,]?.25@ ~.rt'~3,G@ 6 '_--r f:n__s l-,].,l~ L 50,',E,~-.t_ v3.r,'&fJ,DK,. ~-Z1~ ~2,~r '_-,_--' V,J !3 _~@ ,6D. r &-# : 'OJOI : ,L , , ' , : ,'Frr r b,_3112 O 9 I ] G t ~-' S I-r " I , . f.'2-~." d-i : . e ' v--, ~'.17.3::@ "?-,<'OG@. f S )

?0'" ~!2,'2' i'E') 50".

wRh a very thin layer of a more homogeneous, lower surface-free energy maDrlal (as compared [o the Interi- or). 17his interpretation Is conslsterJt with extensive experimental data obtained by various surface analytical techniques show|ng that the air surface of a n-mltlcom- ponent polymer system Is generally enriched with a lower surface-free-ener�9 component to n-,tntn-,Ize polymer-air Interracial energy. ;''-~

The presence of a lower surface-free energy layer on the exterior surface should have an effect on the clegra- datlon behavior of polymer coatings. This is because. wh|le the bulk coating underneath degrades, the more Inert. hydrophoblc thin layer on the exterior suKace may not undergo change until much later This is the case for the hydrolysis of ~e acryltc-melan-dne coating used In this study. While the hydrolysis Is substantial after relatively short exposure tlxr, e (see the FI"IR results), AFM results do not reveal any evidence of surface degradatlou, even at nan@scale resolution, until more than 600 hr exposure {see FNu~e 5). The difference 1,7 the degradation charac tort sties betvseen the exterior' surface tas revealed by AFMi and the bulk fas measured by FllR-transmlsston) observed In this study may explain the lack of a c~rrelattou between surface gloss loss and bulk chemical degradaUou reported wevLously for acwltc-r,-~elan-,lne coatings. ~ Similarly, the delay In the hydrolysis of polyfsebacic anhydrldet (PSAi and

poly{DL-lactlc acid) (PLA} blends used to control drug delg:ertes has been verified by XPS arid ArM as due to the surface enrichment of the more hydrolytically resistant PLA hi rite blends. -'4 Work In our laboratocr also clearly showed that a layer of enriched polyivhlylldene fl uodde! (PVDF} is p~esent o~ the surface of PVDF/acetic polymer blends and that this layer is responsible for the excellent UV resistance of PVDF-based coatings, a'',:~

F*gul~ 5 presents ArM topographic images of spedmens exposed to 70% RH at 50~ for 1264 hr In the absence of UV light. Pits with lateral dimensions from 10 nan@meters to several micrometers formed on the surface of this coating. However. pits were detected only after 600 hr of exposure.

FIIR Analysis of Coating Hydromysis at Different ReNative Humidities

Partially alkylated melamine resins are generally composed of dlffereut compounds with different de- grees of substitutions. In addition to the methoxymeffq?l functionality r a partially methylated melamine resin also contains substantial amounts of amino r and methylol groups t.-Ct-[,OI-[~. ~r Many poss|ble reactions can take place during ~ e curing of a parUally alkylated melamine resin with the hydro.x T- terminated acrylic ester used in this stud'/-) ~' Therefore. a cured fthn of partially alkylated nreiamhle-acryllc coatings contains a variety of groups arid linkages that may affect the hydrolysis of the coatings when they are exposed to weathering conditions.

KUR spectra recorded betbre and after exposure to three RH levels at 50 ~c for 1536 hr are displayed In /;'~gure 63. Because the flh-ns were relatively thick f--10 Ltm). FIIR Intensities of several strong absorpUon bands are substantial, t-{owever, except for the high extinction coeftlclent r C=O band at 1730 cn-r ~ m = 0.36), our analyses at other absorption bands of films having thicknesses from 0.1 to 10 pm showed little deviation from the Beer's absorption law. Fhus. the use of relatively thick films does not affect our ability to monitor degradation quantitatively, but rather it allows us to de- tect and follow subtle changes in the low IR absorption groups and the appearance of new species produced during exposure.

The major bands of Interest In a cured, unaged coat- Inn prepared by reacting a h.gdroxT-tern-dnated acrylic reslr~ and a partially methylated melamine resin are the absorpUons peaking near 3520 cm-'. due to the OI-1 group of the acrylic resin, near 3380 cm -~. due to O1-{/ NI-[ groups of the me/amine resin.::'-' and at 1085 on-,--'. due to C--O-C grot~ps. The band near 1555 cm -'~ has been asslgaed to the contrlbt~tlons of three different groups: trlazlne ring. ( 'N attached to the ring, and CI-{,?" Another band frequently used for curing and degradation analyses of melamine-based coatings is the band at 910 cm -~. due to OCHo and semicircle stretch. 4' ][-he band at 1490 cm -~. whoseO/ntenslty decreases sub- stantially during exposure to humlci conditions with or without light, has been assigned to the CNT~ group of the --N1-[--CH..-NH- linkages.:'

Under the curing condition used In this study 1130~ for 20 rain), the samples were fully cured, as evidenced

.4,7: .I , : ,ur I', ~ji ,:, f cT , : .a i I r ; T , - .chn, : , i , : , gy

g'el~t~ng. I_at:_,,:,rat,:,r\, and Outd,:,,:,r Exl:_,,z,%~re ,:,f C,:,at~ng~: IV

by the fact that the FTIR Intensity of the bands at 910 cm 4 and 2825 cm -! {associated with the melamine C)CHj did not decrease fl~rther after prolonged heating at 130"~C. ft is obvious from F~.ff_,'are 6a (lowest c-m-ve) that the before-exposure, cured film still contained a substantial amount of unreacted OH groups of the ac~-IIc resin ('?,520 era-:' band~ and - N H / O H r cm -i band) and OCH: groups (2825 cm-' and 910 cn-c ~ bands) of the melamine resin. These groups undergo further reactions during exposure to humid environments.

Spectra recorded at a given exposme tin-,e for different Rt-[ levels (e.g.. Figure 6el may provide some l n fo f marion about the Influence of moisture oll the degradation. However. the effects are more dear ly seen in a difference spectrum where the Intensity loss or gain of cer ta in b a n d s o r the appearance of n e w b a n d s can be readily observed. Such difference FIIR spectra for samples exposed for 1536 hr at five different Pd-i levels are prese~ted in Ffgure 6b. These specua were obtained by directly subtracting the spectrum of the unexposed specimen from the spectrum of the exposed specimen. Because the spectra were collected at almost the same location of the samples by the autosampler, error due to sampling at different evposure periods was mini- mal. In F~ure 6b. the lnte~si W of numerous bands in the regions from 3050 to 2750 r and from 1600 to 900 cm-' decreases, while the Intensity- tn the regions from aloo to 340o cm-* and from 1710 to 1580 cm-" In-- creases. Several new bands and shoulders also appear. Including those near 1705 and 1630 cra-:, which are assigned to the C=O stretching of acids a,]d NH bending of primary amines, respectively. 17he acid formatio~ ca~ also be seen In the OH stretching of carboxyllc acids, which extends from 2550 to 3600 cm-'. On the other hand. fl-~e 1630 cn-c * band assignment Is consistent with those given in the literature for primary amines. ';i It should be noted that the difference ETIR spectra of the same coating exposed to condensing humidity at 50~'C for 60 days fnot showm also showed similar band tbrmation and depletion to those displayed 1,-~ b3.gure 6b: that Is. the formation of primary amines at 1630 cm -~ and carboxylic acids at 1705 cm-"

It is evident from Figure 6b that exposl~re to humid envtronraents In the absence of UV light can produce a substantial chmlge In various functional groups In an acryIlc-melanllne coating. The intensities of the bands at 1085. 1630. and 2960 cm -~ (CH suetchlng) have been cbosen to follow C-O bond loss. prima W amine formation. and CH loss. respectively, during exposm'e to different relative humidities, it should be ,~oted that the 1085 cm-' band Is due to the C-O-C bond of hod-, the acwIlc-melar,-,Ine crossl inks.-N-C-O-R-, and the unreacted -N-C-O-CH: groups h~ the rnelamlne chains. Acid-cat@tzed hydrolysis of this type of ether Is very complex. ::,'n and no attempt-was n-ode to separate the contribution of each ether' type In the coating used In this stud3;. Thus. a change In the t085 cm -~ band Intensi- ty Is assumed to be a sum of C--O-(- concentration of b,~th the acryllc-n-,elamlne crossllnk and that of the rne[amhle metboxy group.

i<'~gui~s 73-7c display- F1[IR Intensity changes with ex-- posure time at different RH levels for the C-O loss, NH: fon-nation, and CH loss. respectively, of specimens exposed to 5WC In the absence of UV light, in this figure. the total t,~tenslt'j Increase or decrease at a particular humidity ts due to the sum of post-curing reactions t changes occurrt,~g at RH < < 1%) and hydrolysis result- Ing from the reactions with water In the film at that par- titular RH. Figut:e 7 results show that the extent of all three degradation processes (I.e.. NH, formation. C-O loss, and CH Iossl Increases with Increaslng RH.

i%'gure 8 depicts the hydrolysis rates, along with d-,etr one standard deviation error bars. as a function of RI-[ t~r the three degradation processes, qThe rates, which are expressed as change In FqTIR absorbance with exposure time ~dA,' dt). were taken from the data wld-dn a five- clay- exposure. The results at << l% RH In F~ure 8 were the average of four specimens, as i n d i c t e d In the E'<perin-,ental section. On the other hand. the rate re- suits ~ r each of the relative hun-,ldlties fron-~ 19.2 to .~0.8% were the average of 11 specimens, which Includ- ed the covered mo-UV) specimen and 10 specimens under the Interference filters. The xesults of specimens un-

and the loss of ether linkages f1805 cm -* band). More complete assignments, hlciudlng the bands formed d u r i n g de~rac la t l o~ , are p r e s e n t e d e l s e w h e r e .

E v e n t rade r veD~ d l 3, e x p o s u r e conditions (<< 1':'.6 RH). this model coating underwent changes, as or|- deuced by the decreases In the Inten- sity- of the bands In the regions between 3000 and 2800 cm-: and between 600 and 900 cm-'. fhese chmlges were probably caused by the hydro[ysls reactions during exposure to this condition. The water molecules required for this hydrolysis were probably supplied by several water--produdng reactions taking place during this post-curing. In- cluding the serf-condensation reac- tlons.:.r,.x,

~:,,.::,;,-,~:,.,.-v:,,, ,',:i,: { "f~:

b

F..Ou.C-c~ 5 - - iOL~.,a.rr. '.." ?OC,#rrt r.'_D ('a., ~.nd 5D L't,~ t,-,F.o.g.,s',oh.v /~F.,%,'., ,'m- i .~o~ ~Z',~4 ,tr ~ - . . . .

{fflLd, 7 I-' , [ ' i l-I , I - ~ 4 l . IL i rW .- 'O0"[J 4 ~


der the interference filters were included for the analy sis because only hydrolysis (i.e., no photodegradat ion) was observed to occur on these specimens dur ing the first five days of exposure, which was the period used for determining the hydrolysis rates. The larger error bars for the CI-{ band are due to the small change in [he IR intensity' of this degradat ion process. The specimen exposed to the full UV condition underwen t a substan tin1 pho[odegradat ion at all RId levels, even at early exposure time, u and, thus, was not used for the analysis.

Except for the low RH levels (<< 1% and 19.2%), where the rate is nearly constant v;ith time, the hydrol ysis rates at other RH levels are rapid at the early exposure stage and then slow down at the long exposure times. Bauer is also observed a decrease in the hyd io ly sis rate of partially alkylated melamine acrylic coatings at long exposure times, a i d attr ibuted the decrease to the different reactivities of" different rnethoxy groups in the melamine resin. The most reactive methoxy groups hydrolyze first and the less reactive groups are consumed later, leading to the slowing down of the reaction rate at longer exposures. P~gmes 7 and 8 show that both the rates m~d extents of" the three degradat ion processes increase wi th increasing RI-L However, the rates of,.~ H./ formation are greatest and those of C [-[ consumpt ion are lowest, a n o n g the there processes. ']'he rate difference between the pr imary amine formation and the ether loss

........... I r~i i iaI ] - - 1 5 3 6 ~,, 7~ .7%

~.0 -~ - - - - - - 1 5 3 6 h , 4t.7r

O~5 -f ~ ,"&

400Q 35{}0 8000 ~:50~, 2000 ~,500 lO0e

W a v e r m m b e r , c m -1

a

i o o o ~!

4 0 % R N

t 9 % R I f

4000 3500 3000 2500 2000 150o 1000

W a v e n u r a b e r , era"

l Figure 6 (a) FTIR spectra of specimens be ~._ fore and after exposure to 8hr~e re@tire hu

1 midides at 50"C in the absence of I/qht and (b) differetlce H/R specffa for severTd RH /ev- eLs at 50~C for 7536 hA

is probably due to more than one reaction taking place for the latter process, in which the crosslinked ether groups were both consumed and generated at the same time dur ing hydrolysis (see Mechanism section).

DEGRADATION MODE

h has been generally- believed that the degradation of' polymeric coatings is a homogeneous process in which the loss of material is uniform across the surface of the sample. However, the AFM images shown in l~gure 5 and other AFM results in reference 19 indicate that the degradat ion of an acrylic melamine coating exposed to a humid environment in die absence of UV light is locN [zed and inhomogeneous. Hydrolysis initiated at isolated locations followed by pitting at these initially degraded areas. The pits then deepened and e~flarged with expo sure time, with essentially little increase in the number of pits. The same localized degradat ion mode is ob-- served on the same coating that was exposed to humid ity in the presence of UV light, except that the pits ar'e deeper and their radii are larger, m The same pitting phenomenon has also been obse~x'ed for acid-catalyzed hydrolysis of pol)(orthoester) and poly(anhydrides; 47,4s and base-catalyzed hydrolysis of polyester} 7 Similar to our result for aciFlic-melamine coatings, AFM results of these studies showed that pits are formed at certain sites: pits then enlarge and deepen from these sites. And finally, pit fbrmation at the nanometer to micrometer sizes detected by AFM at the earl), hydrolysis stages for this material is similar to the natural etching (pitting) re por ted for acrylic melamine coatings exposed to acid rain env i ronments or acid rain-rich ou tdoor l o c a - t io ns. 1.3.4

Localized degrada t ion seen in the AFM images shown in Figure 5 supports our belie[ that hydrolyt ic degradat ion of an ac i ) i ic -melanJne coating initiates at the hydrophil ic (as defined in the following discussion) areas of the film. The hydrolysis then expands from these areas following an autocatalytic process. ']?his pos tulation is consistent with the fact that both the water uptake and the diffusion rate of water in a highly crosslinked material or region are low and that hydrol ?,sis reactions of organic compounds in a neutral environment are very slow. This coqec ture is similar to that p roposed by Nguyen et al. 44 for the formation of con ductive pa thways in polymer coatings on steel dur ing exposure to electrolytes. Such conduct ive pathways, which are also believed to develop in the hydrophilic. water susceptible regions of the coatings, are responsi ble for the direct t ransport of ions from the erwironment to the metal surface, leading to accelerated corrosion of metals protected by an intact organic coating.

The inhomogeneous degradat ion mode may be ex plained by the heterogeneous rnicrostrucmre of this rna terial as shown in blgure 4 and of a variety of organic coatings, which have been well documented. Karyakina and Kuzmak ~s have r ev ie~ed and discussed in detail heterogenei ty of polymers and coatings and have con- c luded that polymer coatings generally consist of micro- gels connected via a spatial network. Bascom ~6 has pro posed that he te rogene i ty in po lymers is fo rmed

44 Journal of Coatings lechnology

Relating laboratory and Outdoor Exposure of Coatings: IV

because, as the high molecular mass segments p o l y m e r ize and terminate to form a network, sorne unreac[ed and part ial ly po lymer ized materials are unable to merge into the homogeneous structure and are left at the per iphery of the neWvork units. Early microscopic studies of epoxy; phenolic, and phthalate resins revealed that the microstructure of [hese materials is not homogeneous but is comprised of high densib~ nodules separated by narrow interstitial regions of lower molecular mass mate rial. 4' 50 For thermoset epox); the interstitial regions behave like a liquid material, 4~ and the vol urne frac{ions of the two regions can be est imated by modeling. 5~ Such two-phase structures for a variety of amine-cured epoxies and a polyester are clearly evident from recent high lateral resolution AFM phase imaging studies. :~5'%s~ The first direct chemical evidence of inho-- mogenel ty in a thermoset film was provided by a recent investigation using a high resolution near infrared mul tispectral imaging [echnique. 5z This s tudy sho~,ed that the reaction rates within an epoNW sample are ve ry inhomogeneous and the difference in the degree of cure at different loca[ions can be as high as 37%.

The most intensive s tudy of film heterogeneity and how it affects the protective performance of organic coatings was perfbrrned by Mayne and his coworkers. 55-55 Using microhardness and DC resistance measurements , they have identified a "conduct ing po lymer phase/" which they termed D type, in a variety of coating films. The D type regions are softer, swell more, and have a lower crosslinking density than the rest of the fihn. Corrosion of organic-coated steel has been observed directly beneath the D type regions. Mills and Mayne pos tulated tha~ the D type ~g ions are formed by partially po lymer ized or "dead" molecules, which are present in the resin before casting. During curing, these molecules congregate, and due to their low functionality, do not crosslink to the same extent as the rest of' the film. 55 The water sorption and t ransport characteristics of the D areas have been found to be similar to those of a hydro philic, ion exchanged rnembrane5657: that is, they take up a large amount of water (45-75%i and have a high ion diffusion coefficient. The existence of the low resistance~ D type regions in phenolic, polyurethane, and alkyd coatings has been confirmed by array electrode technique, 5~ A water extraction study also showed that cured films of alkyd and epoxy ester lose about/]-5% of their initial r n a s s 9 clearly indicating that these coatings contain low rnolecular mass, water-soluble materials.

All these referenced studies point to a conclusion that most. if not all, [herrnoset coatings are inhomogeneous mater ia ls consis t ing of low molecular mass, low crosslink, and soft regions (designated as hydrophilic regions) dispersed in a highly crosslinked network structure. During exposure to aggressive environments, pores or pits are believed to initiate and develop in the hydrophil ic regions of the fihns. Direc[ evidence to sup- port the \.'iew that pit formation starts at the hydrophilic regions is from a recent s tudy of h y d r o l y z a b l e / n o n h y drolyzable polymer blends exposed to hydrolytic environments. 6~ AFM images of this s tudy clearly showed that pits initiate and develop at the locations occupied by the hydrolyzable polymer. The postulation that poly meric coatings consist of hydrophil ic regions dispersed

O~

4ka

4 ~

0,a

41,tI~

-llqt~

a

b

r

tion l C in '085 CH

I Figure 7 FfiR intensity changes as a function of exposure time tot five RH levels at 50'~C the abserlce of !ight: (a) C 0 band at 1085 cm -~', Ib) NH. band at 16.I0 cm-", and (cj bat.,d at 2960 cm-'.

in a crosslinked network and that hydrolytic degrada- [ion mostly occurs in the hydrophil ic regions is also ex- perimentally demonst ra ted by scanning electron mi croscopy (SEM), small angle neutron scattering (SANS), and transmission electron microscopy (TEM) stud- [es. 4559 For example, SEM results obtained during water sorption in a number of crosslinked polymers clearl) showed that water does not diffuse uniformly in these materials, but in regions along the boundaries of the network units. ~ Similarly, SANS and TEM results clear ly revealed that the hydrophilic carboxylic acid corn pounds [n carbox}qated styrene-butadiene latex films mostly concentrate at the particle--particle interface and {hat wate r t ransport [n these latex films is mainly along the hydrophi l ic interface, and not th rough the crosslinked particles. 59

For acrylic melamine coatings, such ads heteroge neous structure is schematically illustrated in F~guce 9, which shows hydrophil ic (hydrolysis-susceptible) re-

Vol, 75, No, 941, June 2003 45


0.08

-~ 0,06

0r ~ o,o4

0~02

0.00

0 20 40 60 80

R e l a t i v e H u m i d i t y , %

100

Figure 8--Hydrolysis rate (dA/dt) as a function of relative humidi(y of a partia!!y methylated melamine acrylic: coating exposed to 50"C in the absence of UV liqht for C 0 Mss, primary amine formation, and CH io~. (The results at << 7% RH were the average of four speci mens and at other RH !evets were the average of 17 specimens; and the e~ror bars indi- ca te one standard deviation,)

gions dispersed in a highly crosslinked r~etwork. The hydrophilic regions in this coating are believed to consist of hygroscopic compounds , including melamine methylols and acrylic acid, and hydrolysis-susceptible materials, such as low molecular mass melamine me[hoxy and low molecular mass, partially crosslinked molecules. Since the lo~, molecular mass, low crosslmked molecules in the hydrophilic regions contain a variety of polar groups, such as OH and NH, it is expected that these groups exist in these regions as highly' hydrogen bonded molecules. In F~gure 9, the hy drogen bonds are represented by the dot ted lines con- necting the elec{ronegative O and N atoms with the hydrogen atoms. During exposm~ to humid environ ments, water enters the acrylic-melarnine fihns predom- inantly at the hydrophil ic sites, probably through both concentrat ion gradient diffusion and osmotic dr iven transport. The latter process is due to the presence of hygroscopic materials. Hydrolysis then takes place in the hydrophilic regions, forming the initial pit sites.

The proposed chemically heterogeneous structure of acrylic-melamine coatings shown in Figure 9 is consistent with the concept advocated by Bascorr) 6 and MayneS~ 55 that the low molecular mass, low densib,, partially crosslinked materials [hat are responsible for the loss of protection of organic coatings dur ing exposure {o electrolytes are located at the per iphery of the crosslinked units. On the other hand, the postulation that hydrolysis is more [ikely to take place in the low molecular mass, partially crosslinked, hydrophil ic re gions, and not in the highly crosslinked network, is strongly suppor ted by Berge st al. :s hydrolysis studies of alkoxy, melamine compounds in solutions. They reported that, in mild acid conditions, the ether group of a monosubst i tu ted amino (NH CHoOR}, which is simi iar to the partially crosslinked chains in an acrylic-

melamine coating, hydrolyze rapidly However, the ether groups of a disubstituted amino [ROH,,,C-N-CH.,,OR), which is similar to the fully crosslinked structure of an acrylio-melamme coating, undergo little hydrolysis m mild acid conditions.

As indicated earlier, the exterior surface of this coating is probably covered with a thin layer of a lower sur face free energy material to minimize, the po lymer air interfacial energy. The fact that AFM images reveal pit formation and deve lopment suggests that the exterior surface of this coating probably unde rwen t a rearrangement dur ing exposure to humid environment, schemat ically displayed in Figure 10, The phenomenon of a hydrophobic surface enrichment , which subsequendy undergoes radical surface rearrangement when chang ing the contact medium from air to water is cornmonly observed for many polymers, including polyurethanes and polyamines:~56z: a brief review on the subjec t is given in reference 35. The apparent driving force for the surface rearrangement is the formation of a water hy drophilic interface. According to Figure 10, a thin low surface free energy layer initially covers the outer s u r face of the ac~lic-.melanrine coating film. When this coating is exposed to a h u n f i d / w a t e r environment, the thin hydrophobic layer on top of the hydrophil ic regions becomes unstable and ruptures (dewetting). 3s The dewet t ing continues wi th exposure, gradual ly exposing the hydrophil ic regions of the film to the water vapor environment. Pits initially forth at these exposed hy drophilic regions, as i lhs t ra ted in Figure 10. It is noted that, except for a difference in the composit ion and dis tr ibution of the hydmphi l ic material, the scheme shown in I~gure I0 is similar to that proposed by' Tezuka et al, ~5 for the rear rangement of the polystyrene enriched layer on the surface of polystyrene (PS}-poly(viny[ alcohol) (PV)\) block copolymers when these materials ar~ ex posed to waten

It is easy to comprehend w h y hydrolysis reactions initiate at the hydrolysis-susceptible regions of the film. However, this could not explain w h y the degradat ion expands from these initially hydro lyzed locations, as observed in this and other studies, 4'<,~ instead of forming new' sites. We believe thai the deepening and broad ening of the pits is a result of an acid att tocatalyzed hydrolytic process taking place at the initiated pit sites, where the increased acidity is provided by the hydroly sis products. Substantiation for the pit deepening of acrylic melamine being an autocatalytic process comes &orn two sources. First. the band at 1705 cm :, due to C -O of carboxylic acids, in the spectrum shown in b)igure 5b indicates that acids were generated dur ing the hydrolysis of this coating (see the Mechanism section for the reactions). Second, the increased acidity in pores and pits formed dur ing hydrolysis of other polymers has been rneasur'ed. '~:' Using a pH sensitive dye in corn bination wi th f luorescence scanning confocal microscopy, this s tudy obtained a pH of 5.7 at a dep th of appioximate ly 200 ~an in the pits formed on the surface of polyanhydr ides , as compared to a pH of 7.5 for the exposed solution.

It is likely that the acidity within the pits of acrylio- melamine coatings is high, as evidenced by the high in

46 Journal or Coatings Technology

Relating laboratory and Outdoor Exposure of Coatings: I\/

tensity of the 1?05 cm ~ band for RH levels from 19.2 to 90,8% (F/gure 6b). VVe believe that the high acidity generated by the hydrolysis reactions in the degraded areas accelerates the rate of hydrolysis in the pits, and that this harsh environrnent attacks the nearby polymer chains, This autocatalytic activity explains the deepening and enlargement of the existing pits, rather than for mation of new pits, with exposure time, Autocatalytic hydrolysis has also been at tr ibuted as the main reason for the pit deepening and enlargement in other polymer systems, 42

Other contributors to the pit deepening and broader> ing include the penetrat ion of acids generated dur ing hydrolysis into the polymer matrix, For example, the rate of po]y(or thoester /eros ion was reported to depend strongly on the types of acids, 63 In hydrophil ic acids, such as HC], the hydrolysis reactions were observed to occur only on the sample era'face. But the hydrolysis rate [n the presence of hydrophobic acids, such as per- chloric acid, increased many fold, and the attacks were both on ~he surface and beneath the surface. Another contributing factor may be the rate of water t ransport to the reaction sites. Hydrolysis is a bimolecu]ar reaction in ~,hich water and the labile functional groups are in volved. Therefore. the reaction rate is de termined by the concentration of both reactants, i.e., the more wate r there is present in the reaction zones the higher is the re-- action rate. Since the pitting areas contain mostly polar materials, e.g., short chain starting compounds and hy drolysis products, they likely- attract a high amount of ,crater into these areas and, thus, increase the degrad& tion rates. Finally, in order ]*or pits (holes) to form in the sample, as shown in Figure 5, the hydrolysis products must decompose to fbrm gases or diffuse away from the pit sites. The overall degradat ion rate at the pits is, therefore, con trolled by a number of complex factors, including the pH in the pits, the types and molecular masses of hydro lyzab]e mole- cuds , po]yrner microst ructure , rate of water t ransport into the pits~ and the rate at which degra dat ion p roduc t s are r e m o v e d from the pits.

Information on the degrada tion mode and the nature of the regions where hydrolys is reactions occur should help to design more hydro]ytical ly stable coat ings. The mode of degradat ion proposed in this study, which is based mainly on the experimental evidence of nanoscale coating mi crostructure, extensive literature of corrosion of po lymer coated steel, and knowledge of polymer degradat ion chemistry, is prelimi- nary, Hopefully-, future research on chemical ident i f icat ion and | characterization at nanosca]e of the heterogeneous domains and degraded regions of coatings will

provide a more definitive unders tanding on the mode of hydrolyt ic and other types of coatings degradation.

MECHANISM

FTIR spectra] results given in I?i~ure 6b show that the hydrolysis of a partially rnethylated melamine acrylic coating result in, among others, loss of the ether bonds (C O at 1085 cm --1 band) and formation of" p r imary amines (NH~ at 1630 cm -1 band) and carboxy]ic acids (C.-.O at 170s crn 1 band). The pathways responsible fbr the depletion of" the C---O loss probably occur by the following reactions~:;,tg:

1) Tr--NH-CH:.-O-CH~ + I-LO ~ Tr--HN-CHt.--OH - CH:,OI--I (I)

2) Tr NH CH: O R + H t O ~ ' l r fIN CHt OH+ROH (H)

where "D is the triazine ring, R is the acrylic resin chain, and the compounds in the ]eft hand side of" reactions 1 [o 2 (structures [ and I I /a re the po lymer chains present in the cured film, as shown in Figure 1.

Reac dons ] and 2 proceed by protonation of the ether O&r, ~gen atom to form melamine methy]ols. Disubstituted amino chains (secondary ether), such as structures [II and IM

. CH: 0 CH., .CHt 0 R

'It" N. (IID 8~ld D N. CHo 0 R CH.: 0 R

(iv)

Y~hich are abundant in the fihn, as shown in Fi@ut'e 1, al so undergo hydrolysis to form two methylo] groups per

~twerk

F(gure. 9 ~ schematic ptesentatior~ showing the hydrophilic regions and cross~inked net'wo~k Mr a partially methylated meh_-~mh.,e-acryiic coadng

'9ol. 75, No, 941, June 2003 47


~ 4 Thin layer of H y d r o p h i l i c " 7 ~ ~ hvrlmnhoMe m~fe'r~l

Coaffng m a t d / I I Surface rearrangement RH/Water

. . i P i t fbrmation

I Figure 10 Schematic presentation for the re k- ]

arrangement of the thin low surface tree en

J e&gy layer; on the exterior s urf~ ce of a partially methylated melamine-acrylic coating exposed to water vapor:

melamine chain, i5 However, because of the hydroget~ on the N atom, monosubst i tu ted amino chains (primary etherl of melarriines (i.e., structures [ and II) hydrolyze 30 to 40 times fiaster ffLan disubstituted amino c h a i n s ) ~ 15

Since the hydrolysis rates of" rnonosubstituted amino chains are much greater than those of disubsti tuted amino chains, it is reasonable to conclude that the rapid loss of the C 0 bonds at 1085 crn -I bands at the early ex posure stage shown in fTgures 6 and 7 was due mainly to the reactions 1 and 2.

7'he appearance of the pr imary amine band at 1630 cm -t probably follows the pathw, ays shovvn in reaction 3, where the melamine methylols generated in reactions 1 and 2, being pr imary alcohols, deformylate readily to form pr imary amines and aldehydesiS:

,I) %--HN-CI-L-OH (from Rx 1 and 2) fast

Tr--NH~ + CI--IX) (Primary am~e, 1630 cm -~)

In this case, the methylol group is activated by the addition of a proton to the N atom in the side group to w, hich the methylol is attached. Since formaldehyde formed in the exposure chamber was either removed or readily oxidized to form acids (see reactions 4 and 5), reaction 3 is irreversible. The disubsti tuted amino methy lols (that is, the hydrolysis products of" structures III and IV) do not undergo deformylat ion to form pr imary amines, la Therefore, the intensity increase attributable to the pr imary amines formed dur ing exposure in hu mid environmerits (e.g., Figure 7) was essentially due to the defbrmylat ion of monosubst i tu ted amino melamine methylols that were generated f iom reactions 1 and 2. The formadot~ of volatile methanol in reaction I and the escape of some formaldehyde molecules produced in reaction 3 probably account for most the CH losses at 2960 cnc i dur ing hydrolysis of acwlic melamine coat ings observed in Figure 7c.

' [he methylols formed in reactions 1 and 2 can also self 'condense to form melamine melamine linkages.i~ Howevei~ we believe that reaction 3 was the main path way in the absence of UV radiation for this model coat-

ing. This is clearly demonst ra ted by the substantial in crease of the pr imary amine band at 1630 cm t(Figures 6b and 7b) and very small increase in intensity of the band at 1350 cm -~ (Figure 6b), due to melamine

, t o , melamine hnkage, o The 50':C temperature used in this s tudy should accelerate this reaction, i.e., reaction 3.

The formation of carbo>@qic acids, as observed by the band at 1705 cm i probably occur by the following re-- actionsS!

4) 2CH;O 40~ ~ 2I-tCOOH (formic acid, 1705 cm-~).

The initial oxidation product is a peroxy.-carboxylic acid (HCO OOH), which reacts readily w, lth another formal dehyde to yield t~o molecules of formic acid. In the pres ence of water, acids may be formed by the pathwayaS:

5} CHoO ~ CH~(OH} r -~ HCOOH H~O [O] fast

Reaction 5, which is fast, may be favored inside the pits because of the hygroscopic characteristics of the corn-- pounds in these locauons, which adsorb much more water than the material outside the pits. Another source of acid formation is hydrolysis of the ester groups that were [brined by the reaction between the acrylic acid and the hydrox?, terminated acrlvlic resin taking place dur ing curing. The ester formation is clearly seen by FTIR spectroscopy, n

A number of pathways for the hydrolysis of acrylic rrielamine coatings, several of which produce amines has been presented previously9 ~ Howeven the formation of pr imary amines and acids has not been discussed.

S U M M A R Y A N D C O N C L U S I O N S

Acrylic-melamine resins are widely used for automobile exterior coatings. However, these rriaterials are known to be susceptible to hydrolysis. Despite extensive re search on etching and hydrolysis, there is little information on the mode of degradation for the hydrolysis of these materials exposed to water vapor in the absence of UV light. "['his s tudy presented the mode and specific pathways for the hydrolyt ic degradat ion of a partially methyla ted melamine-ash-tic coating exposed to differ-- ent relative humidit ies in the absence of UV light. The coating was prepared by reacting a partially methylated melarnine resin arid a hydrox)~-terminated acrylic resin. Samples of cured f i lm on a CaF 2 substrate were subject ed to five different RH levels ranging i'Ion~ approxi lnately << 1% to 90.8% at 50:C for more than 60 days using specially designed exposure cells. Coating degradat ion as a function of exposure dine was measured using ~ransmission Fourier transforrn infrared spectroscopy and tapping mode atomic force microscopy. The following conclusions are made:

(a) [n the presence of moisture and wi thout UV lighL partially methy[ated melamine-acrylic coatings hydro lyze quite rapidly; resulting in considerable mate rial loss and formation of various functional groups in the film.

(b/ The rate of hydrolysis, which increases with in creasing RH, is rapid at the early stage and slows down at long exposure time.

48 Journal or Coatings |echnok~gy

Relating Laboratory and Outdoor Exposure of Coatings: I\/

(c) H y d r o l y s i s o f the c ross l inks a n d m e l a m i n e methox?- g r o u p s f rom m o n o s u b s t i t u t e d amino chains are p r imar i ly responsible for the d e g r a d a t i o n of partially rne thyla ted me lamine acrylic coat ings exposed to m o i s t e n v i r o n m e n t s in the a bse nc e of l ight. D i s u b s t i t u t e d a m i n o cha ins (i,e,, fu l ly a l k y l a t e d m e l a m i n e chains) h y d r o l y z e v e r y s lowly and contr ibute little to the hydro ly t i c degrada t ion .

(dj The p r i m a r y p r o d u c t s f o r m e d d u r i n g hydro lys i s in the absence of UV light are p r i m a r y amines and formic acid. P r ima W amines are f o r m e d by the de fo rmyla t ion o f m o n o s u b s t i t u t e d a m i n o me thy lo l s pro- d u c e d by the hydrolys is , and formic acids are gene ra t ed bv ox id iz ing f o r m a l d e h y d e s gene ra ted bl~ the me thylol d 'eformylat ion.

(e) Hydro ly t i c d e g r a d a t i o n is an i n h o m o g e n e o u s process w h e r e pits, w h i c h forn~ locally, deepen and en large wi th exposure time. Such local ized d e g r a d a t i o n is a result of an autocata lyt ic process in wh ich acidic hydrolys is p r o d u c t s a c c u m u l a t e d in the pits catalyze and accelerate the hydro lys i s reactions.

in fo rmat ion on the specific rnechan i sm and ,node fbr the hydro ly t i c deg ra da t i on o f a c r y l i c - m e h m i n e coatings is essential for unders tm~ding the control l ing fac tors responsible [or the failures a nd for deve lop ing ao- curate m e t h o d o l o g i e s to predic t the service life of these coatings. Further, a clear u n d e r s t a n d i n g of the na ture o f the regions whe re hydro lys i s occurs a nd of the ensu ing d e g r a d a t i o n process as offered in this s t u d y shou ld help to des ign more hydro ly t i ca l ly resistant coatings,

A C K N O W L E D G M E N T S

The research repor ted here is par t o f a g o v e r n m e n t / i n d u s t r y c o n s o r t i u m on Se,-v'ice Life P red i c t i on o f Coa t ings at NIST, C o m p a n i e s i nvo lved in this consor- t ium inc lude A K Z O Nobel , ATOFINA, Atlas Mater ia l Testing "]'echnolo~v LLC~ The D o w Chemical Company , D u P o n t Automot ives , Duron Inc,, F, a s tman Chemicals , Mi l l enn ium Inorganic Chemicals , PPG industr ies , Inc., and Shervvin ~,Villiarns Co. Federal Highw, ay A d m i n i s trat ion, Air Force Research Laboratory. a nd U S D A Forest Service (Forest P roduc t s Labora tory) also p rov ided addi t iona l funds for this research. Vv'e also t hank Drs. X i a o h o n g Gu and Mark V a n L a n d i n g h a m for their AFM measu remen t s , and Dr. Diep N g u y e n , o f PPG Indus- tries, for XPS analysis assistance.

R e f e [ e n c e s (1) Schmitz, R], Holubka, ].W., and Xu~ L,., "Add Etch of

Automotiw Clearcoats II. Comparison of Degradation Chemistry in Laboratory and Field T~sting," j,-,>~,'r.<~ah ,-,~ C,..,/<vr,~:; "[~.~r.:, ,~,:,,,'~, 72, No. 902, b3 (2000).

(2) Holubka~ J.W, SchnJtz, Rj., and Xu, L E, ~ ~br E,vironmental Etch of Acrylic Melamine-Based Automoti',~o Cleaccoats: Identification ot Degradatiot~ Products~" j,_,upr.>.-,_,e C,_,AFT.r,,~L, ~FE<!-_.N<,b_,q,Y, 7Z, NO. 904, 39 (2000).

(3) Schulz, U., Ttubiroha, P., Schernau, U., and Baumgart H., Prug. O, g. Coat.. 40, I51 (2000).

(4) Rogers~ W.R.~ Garner. D.R, and Cheerer. G.D., ~Study of the Attack of Acidic Solutions oft Melamit~e-Acrylic Basecoat/ Clearcoat Pait~t Systems, " J,-,U~J.~AL. ,-+" C,-,&~iNGq T~,.- Ht>-,t ,-,,-:z, 70, No 8F[, 83 (1998).

(5) W~rnstSh, K.L., Pdym. Deg. ,Stab.. 5,1, 5? (1996). (6) English, A.D. and Spinelii, .].J., "Degradation Chemistry of

Primary Crosslinks in High Solids Enamel Finishes: Solar Assisted Hydrolysis/' J,-,HF, r~/~L ,-,~ C'-,/'.r i,% 'FEC'~n,-,L,-'CT, 56, No. 711, 43 (1984).

(7) Bauer, D.R and Briggs, L,.M.~ in Characterization of Highly Lro,ss~inked Po/v~ers~ Labana, S.S. and Dickie, R.A. (Eds.), .ACS Symposium Series 243. American Chemical Society, 'vVashingtom D.C., p. 271, 1983.

(8) Bauer, D. and Mielewskt, D.F., Polym. Deg. Stab., 40,349 (1993). (9) Gedock~ J.L. \,hn Oene, H., and Bauer~ D., Euro. Folym..L, 19, 11

(i983~. (10) GeflocR. j.I_,, Dean, M.J., Korrtiski, E.J., and Bauer, D.R, Ind. Eng.

(gera. Prod. Res. Dev. 2.3,449 (1986). (11) \-guyen, T., Martin, J., Byrd, E., m~d Embre~, N., "Relating

Laborato W and Outdoor Exposure of Coatings: II. E.t~ects of Relative Humidity" on Photodegradadon and the Apparent Quantum Yield or" Acrylic-Melanfit~e Coatit~gs~" ],-,T~Rr.~AL ,-,~ C,-,/cim:s T~c>:ra,-,t.,-,,.~% 74, No. 932, 65 (2002).

(12) Nguyen, ]., Martin, j.\a, Byrd, E., and Embree, .\., Polym. Deg. Stab.. 77, 1 (2002).

(13) Bauer, D.R., J. Apld. Po~ym. Sci.. 27, 3651 (1982). (14) Beige, A. Gudmunsen, S., and Ulgelstad, J, Bur. Pdym. j.. 5, 171

(1969). (15) Bet-8e, A . I<',Jaevem and Ugelstad, ].~ Fun Po~ym. f.. 6,981 (1970). (16) Rancourt, J.D., Optica~ Thin ,~lms, User% Handbook~ McGraw-

HiLl, New York~ Chap. 6, p. 183, 1987. (17) Mar*ira J.W. N-guy~n, T., Byrd, E., Embme, N . and Dickeus, B.

Pulym. Deg. Stab. 75~ 193 (2002). (18) Magonov. S.N. and Heaton, M.C, Am. Labarato~, 30, May (1998). (19) VanLarJdingham, M., Nguyen, f., Byrd, E., and Martin, J.W.~ "On

the Use of the Atomic Force Microscopy to Monitor Physical Degradation of Polymeric Coatir~g Surfaces/ J,-,--Kr.;~vJ~ ,-,F C,_,,~Twr<,a 7m'~_rv_,L,_,~y~ 7.3, No. 923, 43 (2001).

(2O) Weast, R. (Ed.), Handbook of Chemistry and Physics, CRC Press~ 53rd ed., p. D 14K 1972.

(21) Adamson, A.W.. FhysicM Chemistqv of Surfaces. Interscience, New "~b[k. 2rid ed., pp. 584 589, 1967.

(22) Barde~ J.A. and Machin, D., T~ans T"'a,aday Soc. 67, 244 (197i). (.?.3) Toprak~ C., Agar, J.N. and ~alk~ M. J. Chem. Soc. Earaday, 75, 803

(1979). (24) Barrie, ].A. in Di~o~sion in Pdymers, Crar~k, ]. and Park~ G.S.

(Eds.), Academic Press, Now York, pp. 259-508, 1968. (25) VanLandingham, M.R., gduljee~ R.E. and Gillespie, ].W. Jr, L

.Appt. P01ym. Sci., 'fl, 669 (1999). (26) Magonov. S.N. Slings, \LB., and Papko\,, VG., Polymer~ 38, 297

(199;). (27) Sauer, B.B, McLean, R.S, and :honlas, R.R., LangmuiL [4, 3045

(1998). (28) Ciraud, M., N-guyem T., Gu~ X. and VanLar~dingham, M., Prior.

Adhesion .5ocie W Meeting, Emerson, ].A. (Ed.), p, 260,2001. (29) Ou, X. Raghavan~ D., Nguyem T., arid VanLarMingham, M.

Pdym. Deg. St~b. 74~ ~39 (2001). (30) .%guyen, I'., Cu., X., VanLandingham, M., Giraud, M., Dutruc

Rosset, R, RTvntz~ R., and Nguyem D., Proc. Adhesion Society Meeting~ Emerson, ].A. (Ed.)~ p. 68, 2001.

(31) Hearn, M.J., Ramer, B.D. and Briggs, D. ,3dacromotecuL, s, 21, 2950 (1988).

(32) Yoon, S.C. Ramer, B.D. Ivam B., at~d Kent~edy. ].R, Macromdecules, 27, 1548 (1994).

(33) Shakesheff, K.M., Evora, C, Sofiano, I, and Langer, R., d. Colloid Interlace Sd, 185, 538 (1996).

(34) Chen, X., McCurk~ S.L. Davies, M.C., Roberts, C.J., Shakeshefi-, K.M., ]endler, S.J.B, and \Mllian)s, RM., A,facromolecu&s. 31, 2278 (1998).

(35) Pienka, Z. Oike~ [rt.~ a~d Tezuka, Y, L a ~ m i n 15~ 3197 (1999). (36) Ou, X., N-guy~n, T. Sung~ L. aodjea~, ]. Broc. of the Federation of

Socie~es for Coatings 7~chnoloyy/ Annual Meeting Program, Ne-w Orleans, LA, November, 2002.

(37) Chang, T.T. Ping. O,g Coat. 29~ '15 (1996). (38) Blank, WJ. and Henslev, VV.L, "Use of Amino Crosslinking

Agents in ~'ater Based Coatings/' J,_,,_,.Rr.j&- ,_,F P.<~r.jr T~c~rv_,t;_, m, 46, No. 593~ 56 (1974).

(39) Bauer, D. and Dickie~ R..L .&ppi /~0/Sqn. Sci., 18, 2014 (1980). (40) LarMn, EJ. Mako\x, ski, M.R, Colthup~ N.B. and Flood, L.A.

VYbradonal Spectros., 17, 53 (1998).

Vol, 7,5, No, 94"I, lune 2003 49


(41) Colthup, N.B., Daly, L.H. and Wiberley. S.E., [ntroducthm to Infiared and Raraan Spectroscopy, 3rd ed, Academic Press, New York, p. 439, 1990.

(42) Leadley. S.R.~ Shakesheff. K.M., et al., Biomateriab, tg~ 1353 (1998). (43) Copferich, A. and LaIGer, R.. J. P d ~ . Sd., Part &. Pdym. Chem.,

31,245 {1993). (44) Nguyon, T., Hubbard, j.g., and Pommersheim, J.M. 'Ut~ified

biodel tbr the Degradation of Organic Coatings on Steel in a Neutral ElectrolytG' .[,-,,Rt.~/vc ,-,F C,-,qTI.'iG~ TD2>-.N,-,L,-,GY. 68, No. 855, 45 (1996).

(45) KawaMna, M.I. and Kuzmak, A.E., Prog. Or~. Coat., 18,325 (1990). (46) Bascom, W.D., .f .&dhedon, 2, 168 (1970). (47) Cuthreli, R.E., f. @pL Polym. Sd., [2, 1263 (1968). (48) Erath, E.H. and Robinson, M., Prec. Am. Che.m. Soc. Meeting.

Organic Coatfn~ :~d Plastics Division, 23,395 (1963). (49} Racich, j.L,. and Koutsky, J.E., in Chemisay and Proper~es of

C*'ossli~s Polymers, Labana, S.S. (Ed.), Academic Press, p. 303, 1977.

(50) Kontou, E., Spathis, G. and Theocaris, RS. jr. Polymer Sci.. P@/m. Chem, 23~ 1493 (198~).

(51) Gu, X. Nguyen, T., VanLaI~dingham~ M., and Raghavan, D. Prec. Adhesion SocieryMeedng. Orlando, pp. 537 539, February 2002.

(52) Fischer, M. and T, an, C.D., Anal Chem., 71,953 (1999). (53) Mayne, j.E.O, and Scandebury-. J.D, Br. Pol3caer 2, 240 (1970).

(54) Mayne, J.E.O. and Mills, DJ., jr. Oil & CoIour Chemists'Assoc., 58, 155 (1975).

(55) Mills, D.]. and Mayne~ J.E.O., in Convsion Contrut ~ Organic Coatings, H. Leiheiser, jn (Ed.), National Association of Corrosion Engineers~ Houston, 'I'X~ p. 12: and refere~c~s therein~ 1981.

(56) Fernandez Plini, R. and Corti, H., "EpoxT Coal 1at Fllras: Membrane Properties and Film Deterioration," J,-,TTR!L'.t. ,-'F C,_,,,-',ir~,,:, TE~I~U,,_,L,_,,.;,. 49~ No. 632, 62 (I977).

(57) Corti, H., Fernandez, P.R., and Gon>z, D., Pr'og. Or'g. Coat., 10, 5 (i982).

(58) Wu, C.L., Zhou, XJ., and Tan,'s grog. Or K CoaL, 25, 379 (1995). (59) Vvalker. R, OfflciaIDT,~e:>r., 32 No. 491, 11~6i (i965). (60) Raghavaa, D., Gu, X., Van Landingham, M, and Nouyon, T., J.

Polym. Sd.. Po{vmer Physics, 39, 1460 (2001). (61) Richard, J. Mignaud, C. and %bng~ K. Pu~yr~. hLter, 30, 431

(1993). (62) T~suka, "~, N-obe, S., and Shiomi, T., ?r 28, 8251

(1995). (63) N-guy~o~ T.H., Himrneistein, K.J., and Higuchi, T. f. Cont~vlled

Rel. 4, 9 (1986). (64) StreiVwieser, D. Jr. and Heatcock, C.H., b~trud~lction to Or_cg, anfc

Che.m/sr~; 2nd ed., Macmillan Publishing~ New "xbrk, p. 39K 1981.

(65) Ibid, p. 367.

50 Iournal or Coatings technology

Relating laboratory and outdoor exposure of coatings: IV. Mode and mechanism for hydrolytic...

Documents

Transcript of Relating laboratory and outdoor exposure of coatings: IV. Mode and mechanism for hydrolytic...