Water whitening of polymer films: Mechanistic studies and comparisons between water and solvent borne films

Progress in Organic Coatings 105 (2017) 56–66

Contents lists available at ScienceDirect

Progress in Organic Coatings journal homepage: www.elsevier.com/locate/porgcoat

Water whitening of polymer films: Mechanistic studies and comparisons between water and solvent borne films Bo Jiang a,1 , John G. Tsavalas a,b , Donald C. Sundberg a,∗ a b

Nanostructured Polymers Research Center, Materials Science Program, University of New Hampshire, Durham, NH, 03824, United States Department of Chemistry, University of New Hampshire, Durham, NH, 03824, United States

a r t i c l e

i n f o

Article history: Received 9 October 2016 Received in revised form 7 December 2016 Accepted 7 December 2016 Keywords: Water Whitening Blushing Polymer Opacity Solvent borne Latex

a b s t r a c t Water whitening of polymer films derived from solution, bulk and emulsion polymerization processes was studied by the use of UV–vis-NIR spectroscopy, differential scanning calorimetry (DSC) and scanning electron microscopy (SEM), as well as visual observations. In addition to quantifying the wavelength dependent light scattering of the films over time, the different physical forms of water present in blushed films were quantified by DSC. SEM was used to observe sections of the films and characterize the scattering domains responsible for the whitening phenomenon. We studied the same polymers with and without the surfactants and salts used in emulsion polymerization, and compared the blushing of water borne and solvent borne films. We have found that all of the wide variety of (co)polymers we used water whiten under the right conditions of time and temperature. Residual surfactants and salts in latex derived films make the blushing process more rapid and more extensive than for the same polymer without them, but they are not the principal cause for water whitening. Neither is the particulate nature of the starting point for latex films, as the same whitening process occurs in solvent borne films of the same polymer. Both absorbance measurements and SEM images show that there is water domain size growth within the polymeric matrix over time. The size and number of the water domains are responsible for the water whitening effect and both can be restricted by the stiffness of the polymeric matrix. Mechanistic modeling of the time dependence of whitening has led to the prediction that the extent of whitening of non-latex based polymer films is directly proportional to the inherent water solubility in the polymer as well as the diffusivity of water within the polymer at the temperature of testing. © 2017 Elsevier B.V. All rights reserved.

1. Introduction and background It has long been known that polymers are plasticized by water. Indeed many reports exist for the distribution of water in polymers used in the textile, food, and coatings industries [1 and references therein]. Cellulosic materials have been the most widely studied, typically by calorimetry, with respect to the physical behavior of absorbed water [2–6]. Hatakeyama et al. [7,8], have coined nomenclature distinguishing each category of water associated with the material as “freezing free water”, “freezing bound water”, and “nonfreezing bound water”. The last of these is closely associated with hydrophilic functionalities in the polymer and contributes to plasticization. We have previously reported on both the prediction [9]

∗ Corresponding author. E-mail address: [email protected] (D.C. Sundberg). 1 Present address: Itaconix Corporation, Stratham, New Hampshire, 03885, United States. http://dx.doi.org/10.1016/j.porgcoat.2016.12.027 0300-9440/© 2017 Elsevier B.V. All rights reserved.

and the experimental measurement [10], by differential scanning calorimetry (DSC), of this plasticization due to the non-freezing bound water content. Beyond plasticization by water, for some systems and under some circumstances, polymer films also whiten, or blush. The opacity comes from water scattering centers of appropriate size within the film. Indeed, these occluded domains of water are comprised of the freezing free and freezing bound water content in the film [11]. While there have been many reports of blushing of water borne films, especially those containing residual surfactants and salts, some reports on similar events in bulk polymers and solvent borne films are also found. The papers by Brown [12] and Johnson et al. [13,14] are typical of water sorption studies in which polymer films were suspended in water vapor at various partial pressures. The weight gain versus partial pressure curves have two distinctly different sections with the second being described as “anomalous water uptake” [12]. This section was analyzed by assuming that a dual site adsorption mechanism was at play where the water sorption in the second part of the data set was associated with clusters of

B. Jiang et al. / Progress in Organic Coatings 105 (2017) 56–66

“associated” water molecules in the near vicinity of “unassociated” water molecules that are “bound” to the more polar constituents along the polymer chain. The Cluster Integral analysis of Zimm and Lundberg [15] was often used to determine the average number of water molecules clustered together in the specific neighborhood of a bound water molecule. Often the number of molecules in the cluster was calculated to be of the order of a few molecules, depending on the polarity of the polymer. Prausnitz and co-workers [16] wrote a wonderful paper along these lines by comparing the water sorption characteristics of four polymers with varying polarity and applied the Zimm-Lundberg cluster integral analysis as a function of the thermodynamic activity of the water (partial pressure/pure water vapor pressure) up to 0.9. The results from these collected authors (not meant here to be an exhaustive list) suggest that before any water whitening occurs, the water in the polymer already exists in two, distinctly different forms, one in very close proximity to a constituent on the polymer chain and the other clustered around the first. Although there are some comments in these papers about the effect of both forms of water on polymer properties, neither form has been associated with film whitening. Some time ago Johnson and co-workers [13,14] reported that when molten polyethylene (PE) was saturated with water and then temperature quenched, it turned white. Subsequent evaluation via SEM showed that there were “domains” of 1–3 ␮m in diameter in the PE. In experiments with polycarbonates (PC) they found that water whitening only occurred at temperatures above its glass transition temperature (Tg ). These results suggest that the apparent “domains” created during water whitening were restrained in size by the stiffness of the polymer matrix at the experimental temperature. The other important result came from the use of DSC to measure the water content of the polymer as a function of water immersion time. Their data for poly(vinyl acetate) (PVAc) show the differentiation between the water absorbed as “bound” and “clustered”, the latter responsible for water whitening. Further, their data clearly demonstrate that whitening is only seen after the plasticizing water is at saturation. Thus the water whitening process appears to require two, sequential steps. There are a large number of literature reports on the water sensitivity of latex derived polymer films but only a few of them specifically report on water whitening, even though it was likely to have happened in many of the cases reported. We cite a few of these reports here [17–28] without the intent to provide an exhaustive list. Among the earliest papers that specifically describe water whitening of latex films are those by Wheeler [29], Wilkes [30], Cote [31] and Bindschaedler [32]. These authors primarily used PVAc latices containing some poly(vinyl alcohol) (PVOH) and found that the films quickly became white and opaque upon immersion in water. Much or all of the blame went to the PVOH surfactant creating continuous phases and/or pockets within the films. In a patent by Wood [33] it was claimed that by deionizing the latex, the blushing of acrylic films could be significantly reduced, and that increasing the pH of these vinyl acid containing latices to 6–7 made further, important improvements. Feng and Winnik [34] found a similar effect of neutralization, this time for poly(n-butyl methacrylate) (PBMA) latices made with methacrylic acid (MAA). In addition they found that when whitened films were thoroughly dried at T > Tg , the film regained clarity, only to blush again when re-immersed in water. Bassett [35] reported on the rate and extent of blushing of latex films made from VAc and VeoVaTM monomers (branched vinyl esters), relating the extent of whitening to the oxygen content of the copolymer – he called this correlation the “hydrophilic budget”. SEM observations of freeze fractured surfaces of water whitened, latex based, films were studied by Agarwal and Farris [36]. They used films created from blends of two, acrylic copolymer latices (without deionization) having different Tg ’s; one above

57

and one below the water immersion test temperature. SEM photos of the freeze fractured surfaces showed domains of several microns in diameter in the films. Okubo et al. [37] worked with polystyrene (PSt) latices made with 8 mol% MAA (dry state polymer Tg of 112 ◦ C) and did similar SEM investigations. However they separated the P(St-co-MAA) from the serum phase, dissolved the polymer in THF and then cast solvent borne films, and then immersing them in water at pH = 13 for one hour. At room temperature (RT) there was no blushing but at 150 ◦ C the film was white and opaque. Fractured surfaces of these treated films were observed in the SEM and domains of several microns appeared in the sample treated at 150 ◦ C. No domains were found in the sample treated at RT. Within the past few years Leiza and colleagues [38] have shown reductions in water whitening of latex films when replacing common, anionic surfactants with certain polymerizable surfactants. In addition, they reported that water whitening was quite sensitive to the final pH of latices stabilized by carboxylic acid groups, much in line with the earlier observations of Wood’s patent claims [32]. Recently, Liu et al. [39] offered a very detailed set of analytical characterizations (particularly NMR relaxometry) of the whitening features of films cast from a single copolymer, P(St-co-2EHA-coBMA-co-AA) with 8 wt% carboxylic acid content, produced via three different methods (emulsion polymerization, solution polymerization, and as a secondary dispersion). They concluded that the total amount of sorbed water is not necessarily a good indicator of water whitening, but that together, the amount and location of water regions within the film determined the extent of water whitening. As a result of reviewing the above literature it is clear that most, if not all, polymers will water whiten under certain conditions. Further, for non-latex based systems, there appears to be a sequential process of water absorption into polymers that, in the first step, has water hydrogen bound to specific constituents on the polymer chain, then additional water becomes clustered in close proximity to the bound water [11], and finally much larger domains (> 1 ␮m) of water are formed in a final act. The purpose of our study has been to examine the mechanism by which the water domains grow with time, as well as the rate at which whitening increases in films composed of a wide variety of (co)polymers. In addition, we sought to contrast the differences in the rates and extents of whitening between latex and solvent borne films of the same polymer composition.

2. Experimental aspects 2.1. Materials Most materials considered in this work were prepared by us via simple emulsion or solution polymerization techniques. We produced latices at 20% polymer solid contents and used sodium dodecyl sulfate, SDS, (99%, Acros) as the surfactant (ca. 1% of polymer weight), potassium persulfate (99.99%, KPS, Alfa Aesar) as the initiator (ca. 0.1–0.2% of water weight), and bicarbonate of soda as the buffer (ca. 0.1% of water weight) when desired. All of the polymerization reactions were conducted at 70 ◦ C in 250 mL jacketed, glass reactors. The monomers styrene (St), n-butyl acrylate (BA), n-butyl methacrylate (BMA), methyl methacrylate (MMA), methyl acrylate (MA), 2-ethylhexyl acrylate (2EHA), and methacrylic acid (MAA) were all obtained from Acros and were cleaned of inhibitor by passing them through activated alumina columns prior to reaction. Some latices were cleaned of salts, ionic surfactants and water soluble oligomers by first diluting to 10% solids, then mixing them with a mixed-bed ion-exchange resin (Dowex MR-3, Aldrich) and stirring them overnight on a shaker table. Solution polymerizations were also conducted. Monomers and benzoyl peroxide (97%, BPO, Aldrich) as initiator (ca. 0.1% of total weight) were dissolved in

58

B. Jiang et al. / Progress in Organic Coatings 105 (2017) 56–66

tetrahydrofuran (THF, HPLC grade, EMD Chemicals Inc.) added to glass vials, and kept at 70 ◦ C in a water bath for 5 h to affect batch polymerizations. Due to the reactivity ratios between some of the comonomer pairs used, there would be some expected copolymer compositional distribution within the final polymer. A commercial sample of polyurethane dispersion (PUD, NeoRez R967, DSM) was used to prepare surfactant free polyurethane films. A commercial sample of poly(methyl acrylate), PMA, dissolved in toluene (Scientific Polymer Products) was also used as received. 2.2. Film preparation Polymer films were prepared by slowly drying latices or polymer solutions in a fume hood at room temperature for 1 day, so as to avoid a possible skinning effect. The films were then dried at a higher temperature (above their glass transition temperature) in an air-circulating oven for another day to make sure of good film formation. Sample films were put in a room temperature desiccator containing DI water vapor at 100% relative humidity or, alternatively, immersed directly into liquid DI water at a variety of temperatures. Just before weighing or DSC measurements, extra water on the film surfaces was removed with Kimwipes (Kimtech). 2.3. Thermal analysis The instrument used to obtain the thermal transition profiles was a TA Instruments Q2000 temperature-modulated DSC. We used an overall heating rate of 3 ◦ C/min, amplitude of ±2 ◦ C, and a period of 60 s. The glass transitions and water melting transitions were obtained from the first heating cycle. 2.4. Scanning electron microscopy (SEM) Selected polymer films were prepared for SEM evaluation by embedding dry polymer films in epoxy resin, cryo-microtoming at −50 ◦ C for a smooth block surface, and sputter coating with platinum. These surfaces were viewed in an Amray 3300FE SEM at various magnifications. 2.5. Ultraviolet-visible-near-infrared spectrophotometry and color ranking We used a Cary 500 UV/Vis/NIR spectrophotometer to obtain a quantitative record of the blushing process as a function of time. The wavelength range used was 3300 nm (near infrared, NIR) to 175 nm (ultraviolet, UV) with an accuracy of 0.1 nm in the UV/Vis range and 0.4 nm in the NIR range. Scans were made from 100 nm to 1000 nm at different time intervals, with scan rates up to 2000 nm/min (UV/Vis). Clear strips of samples films (∼0.5 mm in thickness) were immersed in a glass cuvette filled with water, and the absorbance was recorded as a function of time as the film whitened. We converted from absorbance to transmittance values (absorbance = −Log10 (transmittance)). Another spectrophotometer (Hitachi U-2000) was used to measure the absorbance of the films at a fixed wavelength (500 nm, near the green light that is most sensitive to the human eye). For a convenient, visual perspective, we adopted a ranking system for our samples to complement the instrumental data. The computer-generated images shown in Fig. 1 represent specific examples of our ranking. We chose a value of 0 to represent complete transparency (through which the black background is visible), 1 for “cloudy”, 2 for moderate whitening, 3 for severe whitening, and 4 for complete opacity. We made judgments of those values for the films as they were in water at the test temperature. Others [33–35,40] have reported visual rankings but chose different

Fig. 1. Visual ranking of polymer film showing variable extents of blushing. 0 denotes a transparent film through which the black background is visible. From left to right the transparency of the film decreases and the background becomes less and less apparent. The pictures are computer generated for illustration purposes by changing the transparency in a linear fashion.

numerical scales to represent the same gradations in degrees of blushing. 3. Results and discussion We begin this section by stating what appears to be general knowledge, but sets the stage for our discussion below. Water diffuses through polymers and plasticizes them to lower their effective (wet) glass transition temperature [9]. This occurs even when the water immersion temperature is lower than the dry Tg . Water permeation creates domains of freezable water when T > effective Tg . Polymer films do not whiten when the test temperature is well below the effective Tg . Highly crosslinked films do not whiten but water does lower their effective Tg . Polymer films whitened at T > effective Tg and dried again at the same T, regain clarity. Conversely, films whitened at T > effective Tg and dried at T < effective Tg remain whitened. The latter two points offer us useful experimental conditions under which we can document the microscopic features that occur during the whitening process. 3.1. Thermal analysis of polymer films The DSC is an extremely useful tool to investigate the sequential process of film hydration. Fig. 2 shows the reversing heat capacity (rev Cp , left y-axis, top three profiles) vs. temperature curves for a poly(methyl methacrylate −co-methyl acrylate), P(MMA-co-MA), film (latex derived, but cleaned of surfactants and salts with ion exchange resin prior to film formation) with a dry Tg of 55 ◦ C, before and after water immersion. The derivative of these profiles (right yaxis, bottom darker grey region of the figure) affords those profiles in the form of peaks where it is apparent that when the overnight immersion was at room temperature (glassy relative to the dry Tg), the effective Tg was depressed to 42 ◦ C, while no water whitening occurred. Consistent with this, there was also no melting transition observed for water near 0 ◦ C in that case. Alternatively, when the immersion took place at 70 ◦ C (the uppermost profile in Fig. 2), we see that the wet Tg was further lowered to 39 ◦ C (labeled as the hydroplasticized glass transition) and, importantly, a freezable water peak appeared very near 0 ◦ C. This condition led to an opaque film. In a separate experiment (not shown here) for a solution polymerized and solvent borne film of poly(methyl methacrylate −cobutyl acrylate), P(MMA-co-BA), at a dry Tg of 45 ◦ C, the weight gain of the film was measured gravimetrically prior to this same DSC analysis. The combined mass of water responsible for plasticizing the film [9] and that responsible for whitening (area under the water freezing peak relative to the enthalpy of fusion for pure water) was computed to be equivalent to the overall weight gain of the hydrated, opaque film. These results demonstrate the quantifiable, sequential process of water permeation of polymer films

B. Jiang et al. / Progress in Organic Coatings 105 (2017) 56–66

59

Fig. 2. DSC thermal transitions for cleaned P(MMA-co-MA) latex films in three temperature conditions: dry, and then in water overnight at either RT or 70 ◦ C. Reversing heat capacity is plotted in the top section of the plot, while the derivative of reversing heat capacity is plotted in the bottom section.

leading to water whitening. This sequence is in accordance with the ideas suggested by Berthold et al. [41] in their study of the water vapor permeation of cellulosic polymers; water first associates strongly with the polar groups on the polymer chain (non-freezable bound water, accounting for plasticization of the polymer) and then additional water forms clusters around those locations, eventually leading to freezable water domains. It is the latter, we believe, that leads to water whitening.

3.3. Mechanistic model The combination of Figs. 2 and 3, and the mechanistic concepts of water hydration of Berthold [41] and others, has led us to consider the hydration process from a mass transport of water perspective. It is well established that the water vapor diffusion through polymer films can be quantified by establishing a flux [gwater /(gpoly ,sec)] equation [42] as follows; Water Flux = −D · S

3.2. SEM evaluation of water whitened films The P(MMA-co-MA) cleaned latex films described in Fig. 2 were removed from the immersion water and dried at RT in a fume hood until the film weights returned to their pre-immersion values. Importantly, the film that was immersed at the higher temperature remained opaque during the drying period. Both of the dried films were embedded in epoxy resin and that was cured at room temperature into hard blocks. These embedded samples were then cyro-microtomed at −50 ◦ C to provide smooth surfaces, and then coated with platinum prior to SEM observations. Fig. 3a shows the surface of the P(MMA-co-MA) film that was immersed in water at RT and it is obvious that the surface is free from any defect. Fig. 3b shows the microtomed surface of the film immersed at 70 ◦ C. Here one can see the epoxy-acrylic interface and that there are “domains” within the acrylic polymer near the edge of the embedded film. At the sampling time (length of time immersed in water), there were only a few small “domains” in the center of the acrylic film, but they develop throughout the film given enough time (i.e. water continues to diffuse into the films after the chosen sampling times and existing domains grow larger and new domains are created). Fig. 3c shows the domains at greater magnification and one can see that sizes of the cavities are on the order of several microns, similar to those reported by Johnson [14], Agarwal [36] and Okubo [37] for very different polymers. These figures offer great visual evidence that the whitening we see by eye is related to domains of water that have formed within the film during water contact.

 dp  dx

= D·S



p film thickness



(1)

where D is the diffusion coefficient of water in the polymer, S is the solubility of water in the polymer, p is the upstream vapor pressure and x is distance within the film. The product D·S is often referred to as the permeability coefficient, P, and is an intrinsic property of the polymer-water pair. dp/dx is the “driving force” for transport. If we assume that the rate of water whitening in a polymer film is limited by the rate of water transport into the film, we can write;

t Extent of Water Whitening =

(Water Flux) dt

(2)

0

Using the flux expressed in Eq. (1), we can recast Eq. (2) as



Extent of Water Whitening = D · S

t

(Driving Force)dt

(3)

0

Here we see that the extent of water whitening may be directly proportional to both the diffusion coefficient of water in the polymer and the water solubility. The latter will undoubtedly be strongly related to the polarity of the polymer, and the former to the difference between the water immersion test temperature and the (wet) Tg of the polymer. This already begins to conform to experimental observations. But what about the “driving force”? Here we go back to the images suggested by Berthold [41] in which water molecules strongly associate (via hydrogen bonding) with polar constituents (e.g. carbonyl groups) along the polymer chain and clusters of water (freezable water as noted in the DSC trace) form around those associated water molecules. We imagine that these clusters develop

60

B. Jiang et al. / Progress in Organic Coatings 105 (2017) 56–66

Fig. 3. SEM images of microtomed surfaces of P(MMA-co-MA) cleaned latex films (dry Tg = 55 ◦ C) embedded in cured epoxy. (a) Film immersed at RT and dried at RT; (b) Film immersed at 70 ◦ C for 20 h and dried at RT; (c) Higher magnification of (b).

into “water domains” at the nanometer scale and that the internal pressure of the water within the domain can be quantified as Pi = 2 ␥/R, where ␥ is the interfacial tension between the polymer and the water, and R is the radius of the domain. It is clear that for R values in the submicron rage, the internal pressure can be very large, even capable of overcoming the stiffness (modulus) of the hydroplasticized matrix polymer under the right conditions. Thus we envision a process by which the water first hydrates the film in the form of molecular associations (leading to plasticization, but not whitening), and then further forms water clusters near those molecular associations. Under the right conditions of water immersion temperature and polymer wet Tg , the internal pressure in these clusters is able to overcome the polymer’s resistance to deformation (via its hydroplasticized modulus) and expand, providing that water continues to diffuse into the film. We imagine that the dynamics of this process may well be limited by the rate at which water can diffuse into the film, and that as the water domains grow in size, their internal pressure decreases accordingly (Pi = 2 ␥/R), thereby eventually limiting the size of the domains when the internal pressure is countered by the external resistance to further deformation (external pressure on the domain, Pext ). As we report in the discussion below, this is consistent with our experimental observations. Fig. 4 displays a schematic that shows this concept. 3.4. Experiments and observations Our experimental operational space encompassed both solution polymerized and emulsion polymerized homo- and copolymers.

Fig. 4. Schematic representation of the sequence of steps that ultimately lead to water whitening. Starting from a fully dry and transparent film (against a black background) and with continuous exposure to water influx, first the polar functional groups on the polymer are hydrated (leftmost inset), followed by additional water clustering beyond saturation and the onset of turbidity from small domains (middle inset), followed by growth of those domains (rightmost inset) and film opacity with continued influx of water. The internal domain pressure (Pi ) acts against the resistance of the polymer matrix to deformation in the form of an external pressure (Pext ), via the hydroplasticized modulus of the film.

The emulsion polymers were used “as is” (i.e. containing both surfactants and residual persulfate initiators) and were also cleaned with mixed bed ion exchange resins prior to water borne film formation. Some emulsion polymer samples were thoroughly dried, dissolved in solvent and then cast as films. The films from solution polymerized materials were all solvent cast. 3.4.1. Solution polymerized polymers and solvent borne films The first experiments to be discussed are based upon a solution polymerized P(MMA-co-MA) with a dry Tg of 67 ◦ C (wet Tg = 45 ◦ C) [9] and their films cast from solvent. Inadvertently, these samples were dried a bit too quickly and the resulting films were slightly cloudy to the eye. SEM evaluation showed that there were “solvent evaporation holes” throughout the films with a size dimension of around a micron (Fig. 5a). When these films were immersed in water at RT, the slight cloudiness did not increase nor did the original “solvent holes” change in size (Fig. 5b). But when the water immersion temperature was increased to 40 ◦ C and the film was in that condition for 2 days, the film became opaque. The SEM image of that film in Fig. 5c (hydrated film dried at RT, embedded in epoxy, microtomed at − 50 ◦ C) shows the emergence of domains ∼10 ␮m in size, and we interpret these to have caused the water whitening. It is of further interest to note that this happened at a temperature 5 ◦ C below the measured wet Tg (the temperature range over which the thermal transition took place (measured by DSC) was ∼20 ◦ C). With that set of results, the original films were immersed in water at several different temperatures for variable lengths of time, and then dried at RT. Fig. 6 shows the SEM results for a number of these experiments, all (except for the RT immersion) of which produced opaque films. It is quite clear that the apparent water domains varied in size depending upon the experimental conditions. In Fig. 7 we have plotted the temporal features of these domain sizes as the immersion temperature varied. To be sure, these sizes are only approximate as relatively few domains (10 s of domains) were measured and counted to obtain averages, but the results are still rather striking. It is obvious that the apparent growth rate of the water domains increases at higher immersion temperatures and that the ultimate size of those domains increase with temperature. These results are consistent with the ideas expressed by Eq. (3) and Fig. 4 in that with higher temperatures (above the wet Tg of 45 ◦ C), the diffusion coefficient of water in the polymer (D) increases significantly (as this temperature range passes through the polymer’s Tg ) and the deformation resistance of the polymer matrix (modulus) decreases significantly (commonly, the elastic modulus decreases by 3 orders of magnitude in changing from a glass to a melt). The data set in the lower portion of Fig. 7 represents the measured sizes of the “solvent holes” with time during water immersion at RT (∼20 ◦ C). In addition, observing the time dependent size of these same “solvent holes” at the higher immersion temperatures (as viewed in Fig. 6) shows that these “holes” do not expand during water immersion. We interpret this to mean that these domains do not contain water at any time during the immersion test.

B. Jiang et al. / Progress in Organic Coatings 105 (2017) 56–66

61

Fig. 5. SEM images of solution polymerized P(MMA-co-MA) solvent cast films. (a) dried film showing “solvent holes” resulting from too rapid a drying step; (b) same film immersed in water at RT for 12 h and then dried at RT; (c) same film immersed in water at 40 ◦ C and then dried at RT.

Fig. 6. SEM images of solution polymerized P(MMA-co-MA) solvent cast films after water immersion at the temperatures and times indicated. Scale bars are all 10 ␮m except for the second image from the left in the lower row; this one is 1 ␮m. All films were dried at RT after water immersion.

The temporal aspects of film whitening were also studied by using a segmented polyurethane (Tg = −70 ◦ C) film cast from a commercial, surfactant free, aqueous dispersion (NeoRez R967). A dried film was fixed in a cuvette filled with water at RT and placed

Fig. 7. Water domain average diameters for solution polymerized P(MMA-co-MA) solvent cast films after water immersion at various temperatures and for various times. Data points derived from the images shown in Fig. 6 above. All films were dried at RT after water immersion. As a visual aid, the dashed curve is a simple fit to the data for the sample immersed at 70 ◦ C.

in a Cary 500 UV/Vis/NIR spectrophotometer. As described in the Experimental Section, scans were made over a wide range of wavelengths at many time intervals and the absorbance values recorded. Fig. 8 shows a plot of the associated transmittance values as time increased during the immersion period. Separately, we placed an identical film in water at RT and noticed a slight, visual change of transparency after 2 or 3 h. The film turned cloudy (visual ranking of 2 from Fig. 1) after about 6 h, and became moderately to severely whitened (visual ranking of 3 or 4) after 24 h. The human eye is most sensitive to green light, so visual observations of water whitening are best reflected by the curve of 500 nm wavelength in Fig. 8. First

Fig. 8. Wavelength dependent UV–vis normalized transmittance versus time for a polyurethane film immersed in water at RT. The relevant wavelengths are shown above each data curve.

62

B. Jiang et al. / Progress in Organic Coatings 105 (2017) 56–66

Fig. 9. Decreased transmittance of solvent borne films of solution polymerized copolymers with greatly varying polarities vs. their theoretical water uptakes (water solubility). Transmittance (absorbance) values were measured at 500 nm after immersion in water at 70 ◦ C for 12 h (all polymers had the same wet Tg of 10 ◦ C). The apparent linearly dependence on solubility (at constant D) is indicated by the dashed line.

we note that the transmittance of light at the 250 nm wavelength (not shown here) goes to zero nearly immediately. This appears to be the film itself and is not a result of the film whitening. The data in the 300–500 nm wavelength range responds more gradually. This suggests that the size and/or number of the scattering centers (domains of water) responsible for scattering light have a time variation. Specifically, it appears that these scattering centers begin as small domains (probably in the nm range) and grow from there, resulting in light scattering at the higher wavelengths at later times. Our next probe was to create solution polymerized, linear copolymers with widely varying polarities (from hydrophobic P(Stco-2EHA) to hydrophilic P(MA-co-MAA)), but at the nearly the same wet Tg ’s, (near 10 ◦ C). These were cast as films from solvent and then immersed in water at 70 ◦ C for 12 h. Absorbance values were recorded at the end of that period at a wavelength of 500 nm and then converted to transmittance values. In Fig. 9 we show the% reduction in transmittance (100% reduction represents an opaque film) after 12 h for 5 different polymer samples plotted as a function of their relative polarities; the latter was expressed quantitatively as the “wt.% theoretical water uptake” [9]. These water uptake values are equivalent to the solubility of water in the polymers (S in Eqs. (2) and (3)). In addition, we have noted the visual characteristics of the films (such as “cloudy” or “opaque”) at the time of measurement. What is clear from the data in Fig. 9 is that all of the films blushed with the degree of whitening very much dependent on the polarity of the polymer. Further, the very polar samples were completely opaque after 12 h, while the less polar samples had lost less of their transmittance characteristics, at equivalent time. Clearly these non-polar polymers would have whitened further if we had continued the tests for longer periods of time. Additionally, for those samples not completely opaque (<100% loss of transmittance), the slope of that data set is essentially linear with the solubility of water in the polymers, S (as suggested by the dashed line). If we take the decreased level of transmittance to be a good measure of the extent of whitening, then data plotted in this manner allow us to test the usefulness of Eq. (3) that suggests the degree of whitening should vary linearly with the water solubility of the polymer. This equation also suggests that the degree of whitening is linearly dependent upon the diffusion coefficient of water in the

polymer, D. Since we used copolymers with the same 10 ◦ C wet Tg and the immersion test temperature was always about 60 ◦ C above the glass points, all of the films were equally soft during the tests. Under these conditions it is to be expected that the diffusion coefficients, D, were approximately the same for all of the samples [43]. Thus it appears quite likely that the data in Fig. 9 complies nicely with the expectations from Eq. (3) and that the rate of whitening is linearly dependent upon S. We note here that Bassett [35] proposed that the degree of whitening should be strongly related to the “oxygen content” of the VAc-VeoVa copolymers that he studied. When we plot (not shown here) our data in this manner (% loss in transmittance vs. wt.% oxygen in the copolymer), the data are scattered and the two isomers (PMA and P(MA-co-MAA)) cannot be distinguished, although their polarities are quite different. We now return to the solution polymerized P(MMA-co-MA) solvent borne films discussed in association with Figs. 5 and 6. The “solvent holes” in these films fortuitously became markers for the displacement of matrix polymer during the expansion of the water domains as the immersion time increased. We have shown this in Fig. 10 where the arrows point to several solvent holes in the immediate vicinity of a water domain that had grown to 16 ␮m in diameter. Assuming that we are correct in our proposed mechanism of water domain growth during water whitening, these particular solvent holes would have been displaced in a radial manner as the adjacent water domain grew in size, and further, due to the nature of the stress field around the expanding water domain, the originally spherical solvent holes would deform into ellipses. In addition, the aspect ratio of ellipses would be greatest for those solvent holes closest to the expanding water domain surface, and decrease with distance away from that surface. At distances far enough away from the expanding domain, one should expect the solvent holes to remain spherical. These features are readily seen in Figs. 5, 6 and 10. In order to gain more assurance that these interpretations of the SEM images are correct, we have applied a stress field analysis to this situation, assuming that the expanding water domain can be treated as a “dilation center”. We have used Eshelby’s solution for the stress field surrounding a growing bubble (dilation center) in elastic materials [44] to approximate the distorted shapes of ‘solvent holes’ adjacent to the growing domain. As shown

B. Jiang et al. / Progress in Organic Coatings 105 (2017) 56–66

Fig. 10. SEM image of a solution polymerized P(MMA-co-MA) solvent cast film after water immersion at 70 ◦ C for 12 h. The film was dried at RT after water immersion. The large cavity is the remains of a water domain that grew to a size of ∼16 ␮m. The arrows point to “solvent holes” that are in the immediate vicinity of the water domain.

earlier in Fig. 4, the internal pressure within the curved surface of a spherical domain is written as Pi = 2 ␥/R, where ␥ is the interfacial tension between polymer matrix and water, and R is the radius of the domain at any point in its growth, as shown in Fig. 11 below. Assuming the polymer matrix is homogeneous and elastic, Eq. (4) (below) describes the radial displacement, (u(R)), of the original point (A) in the matrix to its new position (A’) after dilation. R is the original distance from the water domain center to the original point (A) within the polymer matrix. G is the shear modulus of a purely elastic matrix. Inserting the pressure equation into Eq. (4) one sees that the displacement falls off as 1/R2 , meaning that the distortion of the shape of the “solvent holes” will decrease rapidly the further their distance from the expanding water domain. Furthermore, since the stress in the polymer matrix is higher on the side of the “solvent hole” that is nearest to the expanding water domain, the originally spherical “solvent holes” should appear as ellipses after the water domain has expanded. u (R) =

Pi R03 4GR2

(4)

We have measured the aspect ratios of the elliptical “solvent holes” highlighted in Fig. 10. They are 4.0, 2.3 and 1.6, respectively, moving outward from the surface of the water domain. The center for the first ellipse (aspect ratio = 4.0) is located at 9.3 ␮m from the center of the water domain. Measuring the distances of the other two ellipses from the water domain center, and then applying the 1/R2 function from Eq. (4) to calculate the remaining aspect ratios (using the 4.0 value for the first ellipse), we get values of 2.4 and 1.6; these values are extremely close to those measured. These results provide strong, corroborative evidence that there is growth of water

Fig. 11. Displacement vector, u(R), emanating from a dilation center expanding under an internal pressure, Pi.

63

Fig. 12. % Reduction in transmittance after 12 h of water immersion at (a) RT and (b) 70 ◦ C. The left and right bars in each plot refer to P(St-co-BA) and PMA, respectively, while the solid green portion refers to solvent borne films and the patterned blue portion refers to the water borne films. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

whitening domains that push away and deform the matrix polymer as they expand. The only driving force for the expansion is the internal pressure of the water in the domain, which can be enormous within nanometer sized domains, and is only countered by an external pressure acting on the growing domain in the form of the hydrated polymer modulus. 3.4.2. Comparisons between water borne and solvent borne films It is well known that water borne films containing residual surfactants, salts and water soluble oligomers can water whiten quite rapidly. This is the reason that reactive surfactants were developed for emulsion polymerization (and the incentive for their use in [38]). Given the evidence presented above for the whitening of solvent borne films, it is instructive to compare the rates and extents of blushing for water and solvent borne films of the same polymer composition. We do this by first comparing the % loss in transmittance during water immersion of non-polar and polar polymer films of the same composition, prepared from emulsion and solution polymerization (as described in the Experimental Section). The polymers were P(St-co-BA) and PMA, both with a wet Tg ∼ 0 ◦ C (they will have different dry Tg ’s, yet it is the wet Tg that is relevant during whitening). All films were immersed in water at both RT and 70 ◦ C for a period of 12 h and immediately thereafter the absorbance (converted to transmittance) values were measured at 500 nm. The degree of whitening is taken to be directly proportional to the % reduction in the transmittance of the films (100% reduction in transmittance means a completely opaque film, 0% is completely transparent). In the left bar of Fig. 12a, the solvent borne and water borne results are compared for the P(St-co-BA) system immersed at RT. Here it is quite clear that the water borne film whitens much more extensively (88% reduction in transmittance) in the 12 h period than does the solvent borne film (31% reduction in transmittance). In the left bar of Fig. 12b we then compare the same P(St-co-BA) samples immersed at 70 ◦ C. The latex derived film lost nearly 100% of its transmittance in the 12 h immersion, but the solvent borne system has whitened only modestly more than its counterpart at RT immersion (45% reduction at 70 ◦ C versus 31% reduction at RT). This is clear indication that residual surfactant, salts (residual initiator), and water soluble oligomers in the latex derived film have led to an increase in the rate of whitening, but it is also clear that the solvent borne film whitens as well. Moving

64

B. Jiang et al. / Progress in Organic Coatings 105 (2017) 56–66

of whitening increased markedly over the cleaned latex derived films. We perceive that such surfactants and salts reside in channels (perhaps lenses) between the coalesced polymer particles and/or in nano-sized domains within the film. The latter must be relatively infrequent within the film given that the surfactants and salts usually constitute only ∼3% of the weight of the dried film. Liu et al. [39] have suggested that water fills these surfactant nano-domains and can scatter light enough to cause film whitening. Perhaps so. But we suspect that the residual salts and surfactants serve as water permeable conduits to get the water to interior portions of the film (then to solvate the polymer and form water domains within it) much more quickly than can be accomplished by simple molecular diffusion of water through polymer, as is necessary in solvent borne films.

4. Concluding remarks Fig. 13. RT weight gain and water whitening data for water borne PMA films prepared with “cleaned” and “as is” latex. Films were suspended in 100% relative humidity chambers for various lengths of time. Degree of visual whitening was gauged by using the template on the right representing the same opacity scale described in Fig. 1.

to the very much more polar PMA system, comparisons between solvent and water borne films immersed at RT and 70 ◦ C are shown in the right bars of Figs. 12a and b, respectively. Here one sees that at RT there is only a slight difference in the results and essentially no difference is observed at 70 ◦ C. So in this case, for this particular set of circumstances, little, if any, difference can be seen between the water and solvent borne systems − both whiten extensively during the time of the experiment. A final note on the results of these experiments is that when the reduction of transmittance is less than 100%, water transport into the polymer within the films has been limited during the experiment. This supports the supposition in our proposed model that water transport is rate limiting, as expressed in Eq. (3). At the last, we report on the results obtained from experiments in which polymer films were suspended in humidity chambers (at 100% relative humidity) in order to slow the whitening process down and to also obtain weight gain data vs. time. Here we have chosen to contrast the differences between whitening of water borne films derived from “as is” latex and from “cleaned latex” (surfactant and residual salts removed via mixed bed ion exchange resins). The same PMA latex that generated the data in Figs. 12a and b was used to make such films for use in the vapor phase, water penetration experiments. Fig. 13 shows the time dependent weight gain for the “cleaned” and “as is” PMA latex samples cast as water borne films. In addition to the weight gain data, we have superimposed notations of the visual whitening, using the metrics of Fig. 1. In contrast to the rate at which the latex film whitened when immersed in liquid water at RT (Fig. 12a, right bar), the same film in the humidity chamber whitened much more slowly, as expected. It is quite interesting to note that the cleaned PMA latex film showed no signs of water whitening after as much as 20 h in the RT humidity chamber, even with 4% weight gain. When we go back to the expected water solubility in PMA (as in the “theoretical water uptake” values used in Fig. 9), we find that its value is just in excess of 4% [9]. Thus it appears that this film had not yet absorbed enough water to create the domains of water that scatter light (graphically represented in Fig. 4). In contrast, the “as is” film showed a 5% weight gain and moderate whitening after 4 h, and after 20 h it had severely whitened and gained more than 10% in weight. These differences in rate of whitening must clearly be related to the residual initiator salt and surfactant in the “as is” latex derived film. Indeed, when we added back initiator and surfactant to the cleaned latex and made films (not shown here), the rate

All polymers water whiten, given the right conditions of time and temperature. Those conditions first require the polymer to deform to allow water domains to grow. This is satisfied if Pext < Pi (Fig. 4), where Pext is a function of (T-Tg) and macromolecular architecture, such as whether or not the polymer is crosslinked, and if so, how strongly. Simultaneously, the rate of water influx into the film must be sufficient during the timescale of water exposure (the D portion of P, in Eqs. (1) and (3)) to provide enough water in the film to scatter sufficient quantities of light so that the whitening is obvious. If those conditions are met, the rate at which whitening is observed is further increased when the polymer is more polar (the S portion of P) and whether there are additional percolating heterogeneities in the film (such as surfactants, salts, or residual interstitial space from incomplete particle coalescence). It is important to note that discontinuous and nonionic defects in the film (such as in Fig. 5a and b) do not appear to strongly influence this rate of water transport. Considering all of the varied experiments and modeling discussed above, we conclude that the water whitening process in solvent borne films can be divided into four stages. At the beginning the dry films are clear, the DSC can measure the dry Tg , and crosssections of the films are free of defects under SEM observation. After placing the films in water for a short time, the films remain clear but the DSC shows a decrease of Tg as in Fig. 2, regardless of how glassy they are at the immersion temperature. At this first stage water molecules have diffused into the polymer matrix, associated with polar entities along the polymer chain in a homogeneous fashion, thus plasticizing the material. This portion of water will not freeze even at subzero temperatures. In the second stage, once the water has saturated the polymer, additional water diffuses into the matrix and forms nanometer sized domains of heterogeneous and freezable water. Such small water domains have a different refractive index from the polymer matrix and they scatter light of short wavelengths. In the third stage, the droplets grow larger due to high internal pressure, but this growth rate may be limited by the rate at which water can diffuse into the polymer. This rate depends upon both the diffusion coefficient (glass point dependent) and the water solubility in the polymer (composition dependent). Early in this process, only short wavelength light is strongly scattered. As time proceeds, light of longer wavelengths becomes more strongly scattered as the droplet grows, as evidenced by Figs. 3 and 4. The film then begins to look cloudy to the eye. DSC results for such moderately cloudy films show a melting peak near zero degrees; this is the “heterogeneous” water in the growing domains. During the last stage, the water domains have grown much larger in size (and probably more numerous), and the films become whiter to the eye. DSC results for such films show that the melting peak has become significantly larger, but the measured glass transition

B. Jiang et al. / Progress in Organic Coatings 105 (2017) 56–66

temperature does not change from its previous wet Tg value. This process appears to happen in all polymer films. The difference between the temperature, T, of the water immersion test and the wet Tg (not the dry Tg ) of the polymer plays a critical role in the water whitening process. Experience tells us that when T is well below the wet Tg , the film will not whiten. Our mechanistic understanding tells us that the reason for this is that the polymer matrix is so stiff (high modulus) that it can restrict the influx of water that otherwise would be needed to accommodate the internal pressure in and expand the many nanometer sized water domains. But as the testing temperature approaches to within perhaps five Celsius degrees of the wet Tg , the hydroplasticized polymer modulus decreases enough to allow water domains to develop and expand (Figs. 3 and 4) to the point that scattered light becomes visible to the naked eye. Water borne films cast directly from emulsion polymerized polymers whiten much more rapidly than those from solvent borne films of the same polymer composition. Further, water borne films cast from “cleaned” latices whiten much more slowly than those cast from “as is” latex. Obviously the surfactants and residual salts in the “as is” latex must be responsible for these differences. It appears to us that the role of the surfactants and residual salts is either to provide nanoscopic channels, or pathways, for the water to reach polymer within the film without it having to molecularly diffuse through the polymer to reach the inner portions of the film, or simply to provide a larger osmotic driving force (caused by nano-domains of residual salts and surfactants). We are not convinced that visible light is markedly scattered by nanometer sized domains of surfactant (and salt) that may reside in the interstices between the deformed polymer particles that were created during the latex film drying process, especially at the small volume of salts and surfactants relative to polymer in typical latex systems. Instead we think that the polymer of the coalesced latex particles receives water that has moved more quickly (Figs. 12 and 13) through nanoscopic surfactant channels to first plasticize the polymer and then create heterogeneous water whitening domains, just as in solvent borne films. The basic difference between whitening of solvent and water borne films is due to the percolating paths of defects in the latter (either from incomplete coalescence and/or surfactants and salts at that interface) providing a unique mechanism by which the water more rapidly penetrates the polymer film. But the reason that either film whitens is the same.

Acknowledgment The authors are grateful for the financial support provided by the Latex Particle Morphology Industrial Consortium at the University of New Hampshire.

References [1] H. Levine, L. Slade, Water as a plasticizer: physio-chemical aspects of low-moisture polymeric systems, in: Water Science Reviews 3, Cambridge University Press, Great Britain, 1988, pp. 79–185, Chapter 2. [2] E.L. Andronikashvili, G.M. Mrevlishvili, G.S. Japaridze, V.M. Sokhadze, K.A. Kvadze, Thermal properties of collagen in helical and random coiled states in the temperature range from 4◦ to 300 K, Biopolymers 15 (1976) 1991–2004. [3] M.F. Froix, R. Nelson, The interaction of water with cellulose from nuclear magnetic resonance relaxation times, Macromolecules 8 (1975) 726–730. [4] S. Nomura, A. Hiltner, J.B. Lando, E. Baer, Interaction of water with native collagen, Biopolymers 16 (1977) 231–246. [5] F.C. Magne, H.J. Portas, H. Wakeham, A calorimetric investigation of moisture in textile fibers, J. Am. Chem. Soc. 69 (1947) 1896–1902. [6] F.C. Magne, E.L. Skau, Nonfreezing water and nonfreezing benzene capacities of cottons and modified cottons, Text. Res. J. 22 (1952) 748–756. [7] K. Nakamura, T. Hatakeyama, H. Hatakeyama, Studies on bound water of cellulose by differential scanning calorimetry, Text. Res. J. 51 (9) (1981) 607–613.

65

[8] T. Hatakeyama, M. Tanaka, A. Kishi, H. Hatakeyama, Comparison of measurement techniques for the identification of bound water restrained by polymers, Thermochim. Acta 532 (20) (2012) 159–163. [9] J.G. Tsavalas, D.C. Sundberg, Hydroplasticization of polymers −model predictions and application to emulsion polymers, Langmuir 26 (2010) 6960–6966. [10] B. Jiang, J.G. Tsavalas, D.C. Sundberg, Measuring the glass transition of latex based polymers in the hydroplasticized state via differential scanning calorimetry, Langmuir 26 (12) (2010) 9408–9415. [11] Y. Lei, J.R. Child, J.G. Tsavalas, Design and analysis of the homogeneous and heterogeneous distribution of water confined within colloidal polymer particles, Colloid Polym. Sci. 291 (1) (2013) 143–156. [12] G.L. Brown, Clustering of water in polymers, ACS Symp. Ser. 127 (1980) 441–450. [13] H.E. Blair, G.E. Johnson, R. Merriweather, Water sorption of polycarbonate and its effect on the polymer’s dielectric behavior, J. Appl. Phys. 49 (1978) 4976–4984. [14] G.E. Johnson, H.E. Blair, S. Matsuoka, E.W. Anderson, J.E. Scott, Water sorption and its effect on a polymer’s dielectric properties, ACS Symp. Ser. 127 (1980) 451–468. [15] B.H. Zimm, J.L. Lundberg, Sorption of vapors by high polymers, J. Phys. Chem. 60 (1956) 425–428. [16] O. Rodriguez, F. Fornasiero, A. Arce, C.J. Radke, J.M. Prausnitz, Solubilities and diffusivities of water vapor in poly(methylmethacrylate), poly(2-hydroxyethylmethacrylate), poly(N-vinyl-2-pyrolidone) and poly(acrylonitrile), Polymer 44 (2003) 6323–6333. [17] L.-T. Lim, I.J. Britt, M. Tung, Sorption and transport of water vapor in Nylon 6,6 films, J. Appl. Polym. Sci. 71 (1999) 197–206. [18] C. Wu, W. Xu, Atomistic simulation study of absorbed water influence on structure and properties of crosslinked epoxy resin, Polymer 48 (2007) 5440–5448. [19] Z.H. Ping, Q.T. Nguyen, S.M. Chen, J.Q. Zhou, Y.D. Ding, States of water in different hydrophilic polymers −DSC and FTIR studies, Polymer 42 (2001) 8461–8467. [20] P.R. Sperry, B.S. Snyder, M.L. O’Dowd, P.M. Lesko, Role of water in particle deformation and compaction in latex film formation, Langmuir 10 (1994) 2619–2628. [21] M. Joanicot, K. Wong, B. Cabane, Interdiffusion in cellular latex films, Macromolecules 29 (1996) 4976–4984. [22] K.L. Keddie, Film formation of latex, Mater. Sci. Eng. 21 (1997) 101–170. [23] N.L. Butler, C.M. Fellows, R.G. Gilbert, Effect of surfactant systems on the water sensitivity of latex films, J. Appl. Polym. Sci. 92 (2004) 1813–1823. [24] E. Aramendia, M.J. Barandiaran, J. Grade, T. Blease, J.M. Asua, Improving water sensitivity in acrylic films using surfmers, Langmuir 21 (2005) 1428–1435. [25] Y. Reyes-Mercado, F. Vazquez, F.J. Rodriguez-Gomez, Y. Duda, Effect of acrylic acid on the permeability and water uptake of poly(styrene-co-butyl acrylate) latex films, Colloid Polym. Sci. 286 (2008) 603–609. [26] C.H. Lei, K. Ouzineb, O. Dupont, J.L. Keddie, Probing particle structure in waterborne pressure-sensitive adhesives with atomic force microscopy, J. Colloid Interface Sci. 307 (2007) 56–63. [27] G.H. Xu, J. Dong, S.J. Severtson, C.J. Houtman, L.E. Gwin, Modifications of surfactant distributions and surface morphologies in latex films due to moisture exposure, J. Phys. Chem. 113 (2009) 10189–10195. [28] C. Arnold, F. Thalmann, C. Marques, P. Marie, Y. Holl, Surfactant distribution in waterborne acrylic films. 1. Bulk Investigation, J. Phys. Chem. 114 (2010) 9135–9147. [29] O.L. Wheeler, H.L. Jaffe, N. Wellman, A mechanism of film formation of polyvinyl acetate emulsions, Off. Digest 26 (1954) 1239–1241. [30] W.A. Cote, A.C. Day, G.W. Wilkes, R.H. Marchessault, Particle packing in latex films, J. Colloid Interface Sci. 27 (1968) 32–37. [31] G. Wilkes, R.H. Marchessault, Light-scattering characterization of poly(vinyl acetate) latex films, J. Appl. Phys. 37 (1966) 3974–3980. [32] C. Bindschaedler, R. Gurny, E. Doelker, Influence of emulsifiers on film formation from cellulose acetate latexes: experimental study of phase separation phenomena due to sodium dodecyl sulfate. I, J. Appl. Polym. Sci. 34 (1987) 2631–2647. [33] T.G. Wood, Process for improving water-whitening resistance of pressure sensitive adhesives, US Patent 5,286,843, February 15 (1994). [34] J. Feng, M.A. Winnik, Effect of water on polymer diffusion in latex films, Macromolecules 30 (1997) 4324–4331. [35] D.R. Bassett, Hydrophobic coatings from emulsion polymers, J. Coat. Technol. 73 (2001) 43–55. [36] N. Agarwal, R.J. Farris, Water absorption by acrylic-based latex blend films and its effect on their properties, J. Appl. Polym. Sci. 72 (1999) 1407–1419. [37] M. Okubo, A. Ito, M. Okada, T. Suzuki, Variation of the morphology of a carboxylated polymer film by alkali treatment, Colloid Polym. Sci. 280 (2002) 574–578. [38] Z. Aguirreurreta, J. Dimmer, I. Willerich, J.C. de la Cal, J.R. Leiza, Water whitening reduction in waterborne pressure-sensitive adhesives produced with polymerizable surfactants, Macromol. Mater. Eng. 9 (2015) 925–936. [39] Y. Liu, A.M. Gajewicz, V. Rodin, W. Soer, J. Scheerder, G. Satgurunathan, P.J. McDonald, J.L. Keddie, Explanations for water whitening in secondary dispersion and emulsion polymer films, J. Polym. Sci. Part B: Polym. Phys. 54 (2016) 1658–1674. [40] D. Horwat, R. Szuchyt, Blushing resistant adhesives used in bottle labeling, US Patent 7,361,713 B2, April 22 (2008).

66

B. Jiang et al. / Progress in Organic Coatings 105 (2017) 56–66

[41] J. Berthold, M. Rinaudo, L. Salmen, Association of water to polar groups; estimations by an adsorption model for ligno-cellulosic materials, Colloids Surf. A Physicochem. Eng. Aspects 112 (1996) 117–129. [42] R.M. Barrer, Permeation, diffusion and solution of gases in organic polymers, Trans. Faraday Soc. 35 (1939) 628–643.

[43] O.J. Karlsson, J.M. Stubbs, L.E. Karlsson, D.C. Sundberg, Estimating diffusion coefficients for small molecules in polymers and polymer solutions, Polymer 42 (2001) 4915–4923. [44] M.L. Kachanov, B. Shafiro, I. Tsukrov, Handbook of Elasticity Solutions, Springer, 2016.