Comparison of Surfactant Distributions in Pressure-Sensitive Adhesive

Research Article www.acsami.org

Comparison of Surfactant Distributions in Pressure-Sensitive Adhesive Films Dried from Dispersion under Lab-Scale and Industrial Drying Conditions S. Baesch,*,†,⊥ D. Siebel,†,⊥ B. Schmidt-Hansberg,‡ C. Eichholz,‡ M. Gerst,§ P. Scharfer,† and W. Schabel† †

Institute of Thermal Process Engineering, Karlsruhe Institute of Technology, Kaiserstr. 12, 76131 Karlsruhe, Germany Process Research and Chemical Engineering, Coating and Film Processing and §Global Research Advanced Materials & Systems, Polymers for Adhesives, BASF SE, Ludwigshafen, Germany

‡

S Supporting Information *

ABSTRACT: Film-forming latex dispersions are an important class of material systems for a variety of applications, for example, pressuresensitive adhesives, which are used for the manufacturing of adhesive tapes and labels. The mechanisms occurring during drying have been under intense investigations in a number of literature works. Of special interest is the distribution of surfactants during the film formation. However, most of the studies are performed at experimental conditions very different from those usually encountered in industrial processes. This leaves the impact of the drying conditions and the resulting influence on the film properties unclear. In this work, two different 2-ethylhexyl-acrylate (EHA)-based adhesives with varying characteristics regarding glass transition temperature, surfactants, and particle size distribution were investigated on two different substrates. The drying conditions, defined by film temperature and mass transfer in the gas phase, were varied to emulate typical conditions encountered in the laboratory and industrial processes. Extreme conditions equivalent to air temperatures up to 250 °C in a belt dryer and drying rates of 12 g/(m2·s) were realized. The surfactant distributions were measured by means of 3D confocal Raman spectroscopy in the dry film. The surfactant distributions were found to differ significantly with drying conditions at moderate film temperatures. At elevated film temperatures the surfactant distributions are independent of the investigated gas side transport coefficients: the heat and mass transfer coefficient. Coating on substrates with significantly different surface energies has a large impact on surfactant concentration gradients, as the equilibrium between surface and bulk concentration changes. Dispersions with higher colloidal stability showed more homogeneous lateral surfactant distributions. These results indicate that the choice of the drying conditions, colloidal stability, and substrates is crucial to control the surfactant distribution. Results obtained under lab-scale drying conditions cannot be transferred directly to the industrial application. The results were similar for both tested adhesive material systems, despite their different properties. This indicates that other properties, such as the particle size distribution and glass transition temperature, have surprisingly little effect on the development of the surfactant distribution. KEYWORDS: drying of film-forming dispersions, pressure-sensitive adhesives, surfactant distribution, drying conditions, upscaling, confocal Raman spectroscopy chain interdiffusion13,14 and the drying rate.15 Therefore, they are an important factor in the film formation process. The influence of uneven surfactant distributions might impact the process and make it even more complicated. Furthermore, they are also known to influence the function of the final film. They can cause adhesive failure and influence the water resistance.6,16,17 If located at the surface, high surfactant concentrations are known to cause poor gloss and tackiness.2,6 Therefore, it is not surprising that the distribution of surfactants in films dried from film-forming dispersion has been intensively investigated with different methods. Raman spec-

1. INTRODUCTION Latex dispersions have been in the focus of research for many decades as shown by multiple reviews and books1−4 on this topic. They are an important class of systems, e.g., for adhesive applications.5 The replacement of solvent-based adhesive systems is desirable as it allows reduction of volatile organic components (VOCs). Even though the research effort of the community has been focused on the improvement of waterborne adhesives in comparison to their solvent-borne counterparts, there are still some drawbacks like water resistance.6 Many of the disadvantages are believed to be caused by surfactants and their uneven distribution throughout the film.7−9 Surfactants are known to influence the ordering of particles10,11 and particle deformation,12 as well as polymer © XXXX American Chemical Society

Received: January 21, 2016 Accepted: March 8, 2016

A

DOI: 10.1021/acsami.6b00830 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Table 1. Overview of Mass Fractions of Surfactants in the Adhesive Polymer, Initial Water Content in Mass Percent, and Summary of Additional Properties of the Material Systems Dowfax 20A42

Disponil FES 77

Lumiten ISC

Initial water content

Tg

dp

adhesive system 1

1.2%

2%

-

-

1.5%

44.5%

−42 °C

adhesive system 2

-

-

1.5%

1.5%

1.4%

44.5%

−67 °C

150 nm 800 nm 310 nm

Disponil LDBS20

Texapon NSO

Figure 1. Schematic of the batch coater and dryer with temperature-controlled plate in coating operation (solid lines) and in drying operation (dashed lines) (left) in comparison with the device image (right).

troscopy,18−21 ATR-FTIR,22−24 Rutherford back scattering,25 AFM,20,21 and XPS26 were used. One of the first studies that investigated the surfactant distribution and provided a distribution mechanism was made by Zhao et al.26 They interpreted the surfactant enrichment at the interfaces as the result of the surfactants minimizing the interfacial energy. Additionally, the insolubility of surfactant in the polymeric phase combined with water flux as driving force to transport surfactant to the surface was proposed. Many researchers invoked very similar theories in other studies. The first hypothesis was affirmed by Evanson et al.27 They found that increasing surface energy of the substrate increases surfactant enrichment at the substrate. Also the lateral distribution of surfactants both in the bulk and at the surface possibly induced by increased capillary transport to the edge was investigated.18,21,28 Surfactants were found to be enriched at the edges. This indicates that surfactant transported with water due to capillary flow does play a role in the surfactant distribution. Kinetz and Holl24 studied the evolution of anionic, cationic, and nonionic surfactant gradients with FTIR spectroscopy. All surfactants were insoluble in the polymeric phase. They were able to show that the polymer surfactant distribution is already developed during the water evaporation stage. Only morphological changes occurred after the particles reached close packing. This finding was later supported by Belariou et al.19 Gundabala et al.23 was able to model measured surfactant distributions at the film−air interface with a model only including diffusion of particles and surfactant, as well as adsorption of the surfactant on the particles. It suggests that these are the determining physical mechanisms for surfactant distribution in latex films. The ongoing research in the area and the different proposed mechanisms show that there are still many open questions. Furthermore, most experiments are conducted at very low drying rates, where films are dried uncontrolled for hours or even days.29,30 Most industrial processes, however, only take seconds to minutes to dry dispersion films. Studies at

industrially relevant drying rates and temperatures are scarce.22 It was shown that the drying conditions have a large impact on the film-forming process in general.30−32 Depending on the drying rate and temperature the dominating physical mechanisms might change. Studies that deal directly with the surfactant distribution also state that the drying rate has a large impact for the distribution.19,23 The aim of this study is therefore to investigate the influence of different drying rates and temperatures on the resulting surfactant distribution in films for pressure-sensitive adhesive (PSA) applications. Hereby, the drying rate was varied independent of the film temperature by varying gas side heat and mass transfer coefficients, which are linked by Lewis law. Both were adjusted by changing the intensity of the forced airflow. Very high drying rates were obtained by using a dedicated batch coater and dryer that emulates the coating and drying conditions in an industrial belt dryer. Additionally, drying was performed at typical lab-scale conditions as used in other studies. The choice of drying conditions allowed the investigation of the influence of temperature and drying rate as independent parameters. Two 2-ethylhexyl-acrylate (EHA)-based dispersion formulations for adhesive applications were investigated. Both dispersions contained multiple types of surfactants. The surfactant distribution was investigated by 3D confocal Raman spectroscopy spatially resolved through the film. The influence of the very different drying conditions (industrial in comparison to lab-scale conditions) on the surfactant distribution was evaluated and discussed with respect to the models proposed in the literature.

2. EXPERIMENTAL SECTION 2.1. Material System. Two material systems, provided by BASF SE, were used. Both dispersions were polymerized by emulsion polymerization of EHA. For the stabilization of each dispersion, two of the following four different surfactants were used: Dowfax 20A42 (sodium salt of dodecyl diphenyl ether-disulfonic acid), Disponil FES 77 (Poly(oxy-1,2-ethanediyl), α-sulfo-ω-(dodecyloxy)-, sodium salt), B


Research Article


Two different substrates were used in this study: 140 μm thick glass slides with a size of 11.5 cm × 11.5 cm and siliconized polyethylene terephthalate (PET) foil with a thickness of 40 μm. The coated area on the substrates was approximately 11 cm × 6 cm in the case of glass substrates and 30 cm × 10 cm in the case of siliconized PET. For subsequent spectroscopic analysis the samples coated on glass were measured as coated, whereas the samples prepared on siliconized PET were transferred to glass slides before measuring by sticking a glass substrate to the free surface and removing the PET foil. During the measurements the glass slides were always at the bottom. The procedure is schematically shown in Figure 2.

Disponil LDBS20 (sodium salt of dodecyl benzenesulfonate), and Texapon NSO (sodium lauryl ether sulfate). In the first dispersion, further specified as adhesive system 1, Dowfax and Disponil FES 77 were used. In the second dispersion, further specified as adhesive system 2, Disponil LDBS 20 and Texapon NSO (sodium lauryl ether sulfate) were used. Both dispersions additionally contained Lumiten ISC (diethylhexyl ester of sulfosuccinic acid) for an improved wetting behavior. Adhesive system 1 has a bimodal particle distribution with particle diameters of dp = 800 nm and dp = 150 nm and a glass transition temperature of Tg = −42 °C. Adhesive system 2 has a particle size distribution with its maximum at a particle diameter of dp = 310 nm. The width of the size distribution is characterized by (d90 − d10)/d50 = 0.15. The glass transition temperature of adhesive system 2 is Tg = −67 °C. Due to its different surfactants, adhesive system 1 exhibits better dispersion stability in comparison to adhesive system 2. The surfactant concentrations of all surfactants in the dispersions as well as the initial water content are shown in Table 1. More information about the material system can be found in Kimber et al.22 2.2. Coating and Drying. The coating and drying step was intended to be as close as possible to industrial conditions. Therefore, the dispersions were coated by a knife coating procedure and both dried in a dedicated batch coater and dryer either under industrially relevant conditions or without forced convection to simulate lab-scale drying conditions. A schematic and image of the dedicated batch coater and dryer are shown in Figure 1. The batch coater and dryer is described in more detail elsewhere.33 The used coating and drying device consists of a temperaturecontrolled aluminum plate, which can be moved underneath a slot nozzle dryer with a defined velocity. For the coating procedure the adjustable doctor blade (Zehntner ZUA 2000 Universal Applikator) is mounted in front of the dryer, while the tempered plate with the substrate on top moves in coating direction below the doctor blade into the dryer. The coating speed was set to vcoat = 12.5 mm/s. The coating gap was varied, depending on the temperature and coating solution, to obtain the same dry film thickness of hdry = 25 μm. After coating, the temperature-controlled plate was moved periodically underneath the slot nozzle dryer with a velocity of vplate = 12.5 mm/s for 20 min. This simulates an industrial belt dryer in large detail. The temperature of the airflow and the temperature-controlled plate were both set to uniform values of 25 or 60 °C, respectively, which in the following are referred to as drying temperature. These temperatures correspond to the wet bulb temperature which are reached in a typical belt dryer with nozzles blowing from the top and bottom at an air temperature of Tdryer = 45 °C for a wet bulb temperature of 25 °C and Tdryer = 250 °C for a wet bulb temperature of 60 °C. For this calculation heat transfer coefficients of α = 35 W/(m2·K) and α = 95 W/(m2·K) at both sides of the belt and a dew point of Tdew = 5 °C have been assumed. The average heat transfer coefficient was varied by changing the air flow rate. Additional experiments with no air flow were carried out to simulate drying conditions in an oven as typical lab-scale conditions. An average heat transfer coefficient was estimated to be approximately α ≈ 5 W/(m2·K). The correlating drying rates for the different experimental settings are listed in Table 2. The experimental conditions were chosen in a way that it is possible to distinguish the influence of increasing drying rate and film temperature during drying.

Figure 2. Schematic of the differences of the coating and measurement process for films coated on glass (left) and siliconized PET foil (right). 2.3. 3D Confocal Raman Spectroscopy. Semiquantitative surfactant concentration profiles were measured with customized 3D confocal Raman spectroscopy. A Raman microscope with motorized microscope stage (Märzhäuser, Tango3D), a DPSS laser with a wavelength of 532 nm, and a commercially available spectrometer were used (Jobin Yvon Labram 8/178 IM). The prepared films were measured from the glass side with an immersion oil objective with 100× magnification. A pinhole with 200 μm diameter and a slit of 100 μm were used. This results in a measured spatial resolution of 2 μm. With a spacing of 2 μm in coating direction and 1 μm in the film thickness, Raman spectra were measured along a 150 μm long cross section in the middle of the films. Quantitative information about the local composition in the focal point is obtained by modeling of the measured spectra. The measured spectra are remodelled by superposition of the pure component spectra analogous to Alsmeyer et al.34 and Scharfer et al.35 The weighting factors ωi, which are obtained by minimizing a least-squares sum of deviations between measured and superposed spectrum, correlate with the intensity ratios of the Raman spectra of the individual components n

Imix =

i=1

ωsurf, i ωpolymer

belt dryer gas temperature

250 °C

45 °C

60 °C

25 °C

drying rate in g/(m2·s) at

0.7 4.6 12.54

0.079 0.6 1.5

α ≈ 5 W/(m2·K) α = 35 W/(m2·K) α = 95 W/(m2·K)

(1)

i=1

n denotes the number of components. The ratio of the weighting factor of each of the surfactants ωsurf,i and the weighting factor of the pure polymer is proportional to the surfactant mass loading in the focal point34

Table 2. Drying Rates at Various Wet Bulb Temperatures and Average Heat Transfer Coefficients and Corresponding Temperatures in Industrial Belt Dryers

wet bulb temperature/experimental film temperature

n

∑ Ii = ∑ Ii ,pureωi

= K ′i

Ii ,surf Ii ,polymer

= Ki

msurf, i mpolymer

(2)

The proportionality factor of the surfactant i, Ki, can be obtained by a calibration.35 Even though the weighing factors themselves render no information about the absolute local surfactant concentration, they give information about the relative surfactant concentration and therefore the surfactant distribution in the coating. The displacement and distortion of the focal point due to the change of refractive index from glass to the film were taken into account by means of d.o.f. (depth of focus) and c.o.g. (center of gravity) corrections.36 The required refractive indices were measured at a temperature of T = 25 °C with an automated refractometer (Dr. C


Research Article


Figure 3. Pure component Raman spectra of the surfactants and of adhesive system 1 (left) and the comparison of the spectra of adhesive system 1 and the dialyzed adhesive system 1. Kernchen Abbemat Digital Automatic Refractometer) and were quantified to be nadhesive 1 = 1.474 and nadhesive 2 = 1.470. The influence of different surfactant distributions on the refractive index was neglected. Pure component spectra of surfactants were obtained from dry films that were prepared from pure aqueous surfactant solutions without adhesive polymers. To obtain pure component spectra for the polymers of the adhesives, dry films were prepared from the dialyzed dispersion. The resulting normalized spectra of the surfactants, as well as a comparison of dialyzed and nondialyzed adhesive mixture 1, are exemplarily shown in Figure 3. The surfactant spectra are clearly overlapping. However, every spectrum has its own characteristic peaks, which is crucial for avoiding multiple minima in the square sum during superposition and thereby guaranteeing unambiguous relative mass fractions. In comparison to the intensity of the dialyzed adhesive film, the measured signal intensity of the surfactants is low. In addition to the low signal intensity of the surfactants, the surfactant concentration in the polymer is very low. Therefore, the difference in the spectra of adhesive system 1 and the dialyzed adhesive system 1 is very small. It is also possible that the surfactants were not completely removed by the dialysis. Since overall surfactant concentrations were not measured and calibrated, the potentially remaining surfactant in the dialyzed dispersion only lowers the least detectable surfactant gradient. The glass substrate has a low Raman activity but shows two peaks in the evaluated Raman shift range (750−1800 l/cm). Not including the glass spectra in the evaluation leads to a false detection of surfactant concentrations due to the low surfactant content. This was tested by measuring the same spot of a film with and without glass cover. Only by evaluating the spectra measured with glass cover including a glass spectrum in the evaluation, the same results as without glass cover and without including glass spectra could be obtained. To determine the least detectable surfactant gradient, double layers with addition of different amounts of surfactants were prepared. Aqueous surfactant solutions were added to adhesive systems 1 and 2. Subsequently the mixtures were coated on a 140 μm thick glass substrate to produce 40 μm thick films and stuck to a 40 μm thick film of the original adhesive system, which was also coated on a glass substrate. Subsequently Raman spectra were taken in both films. A plot of the resulting surfactant intensity ratios of a double-layer film of adhesive system 1, as a function of the measurement depth, is shown in Figure 4. The surfactant intensity ratios directly correlate to the surfactant mass concentration. The generated gradient at the interface of the two coatings and the different surfactant levels are clearly visible for all surfactants. The least detectable difference of surfactant between both layers was determined to be Δxsurfactant = 0.25%.

Figure 4. Surfactant distribution over a double layer with a surfactant enriched film (left) and a layer with low surfactant content (right) in adhesive system 1. Dowfax is denoted with red squares, Disponil FES 77 with green triangles, and Lumiten ISC with blue diamonds. In the enriched layer 1% of Dowfax, 1% of Disponil, and 2% of Lumiten ISC each relative to the nonvolatile content was added.

Each dot depicts one measurement point and the color of its surfactant fraction. For all three surfactants a gradient with decreasing surfactant fraction from top to bottom is visible. The surfactant distribution also varies locally along the measurement plane. This indicates a nonuniform surfactant distribution, most likely due to agglomeration, as reported in other articles.7,18,20,21 Since the measurement resolution is limited to 1−2 μm, potentially existing surfactant agglomerates can only be resolved if their size is above the resolution limit. All measurements showed comparable results in terms of lateral distribution, independent of adhesive and drying conditions. The distribution over the film thickness, however, correlates with different drying temperatures and air flows. For the quantification of the drying process influence, all intensity ratios at the same measurement depth along one film were averaged for each surfactant, and the measurement variance was calculated. A resulting plot of the relative surfactant distribution is exemplarily shown in Figure 6. The film was measured through the glass substrate at the bottom, which is always located at a measurement depth of dm = 0 μm. The free surface of the film is at the maximum measurement depth for films coated on glass. According to the previously described experimental procedure, for films coated on PET the former free surface, adhered to glass, is located at a measurement depth of 0 μm, and the released substrate side is located at maximum measurement depth. Each measurement point is averaged over 75 measurements at different lateral

3. RESULTS AND DISCUSSION In Figure 5 the distribution of all three surfactants of adhesive system 1 in a cross section of a dried film is exemplarily shown. D


Research Article


concentration profile in the bulk of the sample varies significantly with the drying conditions. In the case of Dowfax and Disponil FES 77, which are stabilizing agents and therefore initially located at the particle surfaces, the gradients in the bulk of the film are influenced by the drying conditions. Whereas drying without forced convection (α ≈ 5 W/(m2·K)), representing lab-scale drying conditions, leads to a constant surfactant concentration in the bulk of the film, drying with forced convection, representing industrial conditions, at heat transfer coefficients of α = 35 and α = 95 W/(m2·K), leads to a steeper gradient in the bulk. However, no difference was found for the samples dried at α = 35 and α = 95 W/(m2·K), indicating that there is a critical heat transfer coefficient beyond which the gradients do not increase further. We suspect that high heat transfer coefficients lead to a “freezing” of the nonequilibrium gradients, since both diffusion and adsorption kinetics are time-dependent processes. Accordingly, gradients which are caused by transport mechanisms (e.g., capillary transport) to the surface have less time to be leveled out by back diffusion of the surfactants because coalescence of the latex particles occurs faster. The concentration in proximity of the interface seems to depend on the nearby bulk concentration, as a result of a local equilibrium. This is either caused by increased adsorption on the particles or by adsorption on the free surface and desorption at the substrate. Since the vast majority of studies in the literature is conducted at very low heat and mass transfer coefficients in the gas phase (corresponding to α ≈ 5 W/(m2·K) in our study), their transferability to industrial manufacturing conditions has to be evaluated critically. While Dowfax and Disponil FES 77 show very similar behavior, Lumiten exhibits different characteristics. Especially, the depletion near the substrate is less significant for this surfactant. This is in accordance with its function as wetting agent functioning at the film−substrate interface and therefore interacting less with particles in accordance to Gundabala et al.23 Figure 7 (bottom row) depicts the variance s2 of the relative surfactant distribution, which is defined as

Figure 5. Relative surfactant distributions in a cross section of adhesive system 1. Each dot is one measurement point. The cross sections show agglomerations of each surfactant at the surface. Measurement resolution in coating direction Δx = 2 μm, in height Δz = 1 μm.

si 2 =

1 n

n

∑ (ωi ,k − ωi̅ )2 k=1

(3)

with n being the number of measurement points and ω̅ i the average relative intensity ratio. In all cases the variance increases with increasing local surfactant content. We attribute this behavior to the formation of inhomogeneously distributed surfactant agglomerates. In the case that more surfactant is present, more or bigger agglomerates may form. In the case the distance between the agglomerates is higher than the resolution of the measurement setup, this leads to higher variance of the data. At high drying temperatures of 60 °C (see Figure 8 (top row)) the effect of the varying heat transfer coefficients on the bulk gradient is much less significant or not existent at all. In contrast to the case of high heat transfer coefficients at a drying temperature of 25 °C which leads to gradients in film, there are almost no gradients in the bulk of the film at 60 °C for all boundary conditions, except for Lumiten. We suspect that either the elevated diffusion speed of the surfactant at 60 °C leads to a faster leveling of gradients and equilibration of the bulk-interface equilibrium or faster film formation occurs at the surface, sealing the film while the bulk is still wet. Accordingly

Figure 6. Exemplary relative surfactant distributions in a film coated on glass with adhesive system 1. The measured spectra ratios are laterally averaged at the same measurement depth over 75 measurement points. The substrate is located at measurement depth h = 0 μm, the free surface at maximum measurement depth. Dowfax is depicted in red, Disponil FES 77 in green, and Lumiten ISC in blue.

positions and the same measurement depth within the range of 1 μm. In the following section the influence of different drying conditions in terms of heat transfer coefficient α and film temperature is shown. Figure 7 (top row) depicts the surfactant distribution measured after drying at a film temperature of 25 °C. In all cases layers at both interfaces and a bulk zone can be distinguished with significantly varying gradients. Depletion of the surfactants at the substrate can be observed, whereas accumulation occurs at the film−air interface. The slope of the E


Research Article


Figure 7. Surfactant distribution (top row) and the variance of the measurement data (bottom row) measured by Raman spectroscopy for the various surfactants at 25 °C in adhesive system 1 coated on glass. The gas side heat transfer coefficient has been set to 5 W/(m2·K) (no air flow) (blue triangles), 35 W/(m2·K) (green diamonds), and 95 W/(m2·K) (red squares). The profiles are chosen exemplarily from a set of experiments. All measured profiles are depicted in Figure S1. The substrate side of the film is at the position 0 μm.

the surfactants are able to reach an equilibrium distribution in the film, independent of the heat transfer coefficients. The resulting homogeneous surfactant distribution in the bulk of the film also results in an almost uniform variance at drying temperatures of 60 °C (Figure 8, bottom row). Even though Lumiten shows a slightly different behavior, with linear detectable gradient in the bulk, the different drying conditions at drying temperature of 60 °C do not have a clear impact on the gradients. Figures 9 and 10 show the respective results for adhesive system 2 at 25 and 60 °C. The observations are comparable to those in adhesive system 1. Depletion at the substrate and accumulation at the film−air interface of the surfactants can be observed, as well as a bulk region. Again, the influence of the external heat transfer coefficients on the gradients is similarly noticeable at 25 °C. At 60 °C adhesive system 2 tends toward a more gradient-free bulk of the film for surfactants Disponil LDB520 and Texapon NSP. However, this difference is less distinct compared to adhesive system 1. The level of variance is generally higher in adhesive system 2. This indicates a less uniform surfactant distribution in the system, caused by larger agglomerates in comparison to adhesive system 1. However, we observed the variance only to increase with locally increasing surfactant concentration. The average variance of the surfactant distribution for all surfactants was independent of the drying rate. Therefore, the drying rate either has no effect on the agglomerate size or the agglomerates are about a magnitude smaller than the microscope resolution.

The differing variance in both systems can be explained by the drying mechanism proposed in the literature30 and the different colloidal stability of the two dispersions. A schematic of the steps influencing the surfactant distribution is shown in Figure 11. Initially the free surfactant is in equilibrium with adsorbed surfactant on the particles. With proceeding solvent loss due to drying, both particle and surfactant concentrations increase. The increasing free surfactant concentration provokes more surfactant adsorption onto the particles and absorption into the particles. In the case of sufficiently high colloidal stability, the particles form a close packing and subsequently coalesce to form a film. The surfactant is released from the vanishing surface and is distributed in the remaining voids between the particles. The voids have the tendency to contain surfactant agglomeration when the film is dried. A better stability therefore leads to a more equal distribution of voids, hence a more equal lateral distribution of surfactant. Adhesive system 1 has a higher colloidal stability in comparison to adhesive system 2. Consequently the particles are less ordered in adhesive system 2 when first agglomeration occurs, resulting in higher variance of the surfactant concentration. The very different particle size distributions and colloidal stability of adhesive system 1 and adhesive system 2, which lead to different capillary sizes and transport, have little influence on the surfactant distribution in the dry film and could not be identified as important parameters in this study. The surfactant distributions are more influenced by drying kinetics and the equilibrium between interfaces and the bulk. This indicates that F


Research Article


Figure 8. Surfactant distribution (top row) and the variance of the measurement data (bottom row) measured by Raman spectroscopy for the various surfactants at 60 °C in adhesive system 1 coated on glass. The gas side heat transfer coefficient has been set to 5 W/(m2·K) (blue triangles) (no air flow) and 95 W/(m2·K) (red squares). The profiles are chosen exemplarily from a set of experiments. All measured profiles are depicted in Figure S2. The substrate side of the film is at the position 0 μm.

Figure 9. Surfactant distribution measured by Raman spectroscopy for the various surfactants at 25 °C in adhesive system 2 coated on glass. The gas side heat transfer coefficient has been set to 5 W/(m2·K) (no air flow) (blue triangles), 35 W/(m2·K) (green diamonds), and 95 W/(m2·K) (red squares). The profiles are chosen exemplarily from a set of experiments. All measured profiles are depicted in Figure S3. The substrate side of the film is at the position 0 μm.

coating processes.37 Figure 12 exemplarily depicts the distribution of the surfactants in adhesive system 2 on siliconized PET at drying temperatures of 25 and 60 °C and a heat transfer coefficient of 95 W/(m2·K). For adhesive system 1 comparable behavior was observed. The film−air interface of the drying process is now depicted at the position of 0 μm. In

capillary forces are not the dominating mechanism for the formation of gradients. To evaluate the substrate influence on the surfactant distribution, a siliconized PET which has a lower surface energy compared to glass was used as substrate. Siliconized PET is a common substrate used as release liner in transfer G


Research Article


Figure 10. Surfactant distribution measured by Raman spectroscopy for the various surfactants at 60 °C in adhesive system 2 coated on glass. The gas side heat transfer coefficient has been set to 5 W/(m2·K) (blue triangles) and 95 W/(m2·K) (red squares). The profiles are chosen exemplarily from a set of experiments. All measured profiles are depicted in Figure S4. The substrate side of the film is at the position 0 μm.

Figure 11. Schematic of the different stages influencing the lateral homogeneity of the surfactant distribution. Stage I: initial stage, surfactant concentration below critical micelle concentration. Free surfactant is in equilibrium with adsorbed surfactant; stage II: surfactant and particle concentration increase due to evaporation. Due to the increased free surfactant concentration, the surfactant adsorption onto the particle surface increases; stage III (a): at sufficient colloidal stability the particles order into a close packing. The particles are well-ordered; stage III (b): due to insufficient colloidal stability particle agglomeration occurs. Particles are not well-ordered; stage IV: particles compact and coalesce. The surfactant released from the vanishing surface accumulates in the voids between the particles.

surfactant gradient formation, as done by Gundabala et al.,23 we have to compare the diffusion times of all components in relation to the drying times. This is commonly done by calculating the Peclet number, which is the ratio of the surface descending velocity E0, initial film thickness h0, and diffusion coefficient D38

contrast to the results on glass, there are almost no gradients in the film at low temperatures. Only a small increase of the surfactants toward the film−air interface is seen, again indicating a local equilibrium between bulk concentration and surface concentration. The absence of depletion at the substrate side of the film is notable and very different from the behavior on glass. This difference can be attributed to the difference in surface energy of glass and siliconized PET and hence the different underlying driving force for the distribution mechanism. The results clearly indicate that the choice of the substrate has strong influence on the resulting surfactant distribution. Properties of the coating which are dependent on the surfactant content near the film−substrate interface, such as adhesion, therefore have to be tested and optimized on the actual substrate or a substrate with very similar surface energy. Results obtained on other substrates cannot be transferred. If the diffusion and adsorption of the surfactants on the particles are considered as main mechanisms provoking the

Pe =

E 0h 0 D

(4)

Gundabala et al. assume strong particle accumulation at the top and varying Peclet numbers for the surfactant. With the size of the particles ranging from dp = 310 nm to dp = 800 nm and the different drying rates in this work, the particle Peclet numbers range from Pep = 2−300 and Pep = 10−1600, respectively, employing a Stokes−Einstein diffusion coefficient. Due to the low density gradients of water and polymer, sedimentation is negligible. This leads to a particle accumulation at the top according to Cardinal and Jung39 for all drying conditions. This H


Research Article


Figure 12. Surfactant distribution measured by Raman spectroscopy for the various surfactants at 25 °C (filled squares) and 60 °C (half-filled squares) in adhesive system 2 coated on release liner. The gas side heat transfer coefficient has been set to 95 W/(m2·K). The profiles are chosen exemplarily from a set of experiments. The substrate side of the film is at the maximum measurement position.

was the first assumption of Gundabala et al., which is also valid in our case. If we assume a surfactant diffusion coefficient of Dsurfactant = 4 × 10−10 m2/s, surfactant Peclet numbers in the range of Pesurf = 0.1 to Pesurf = 1.56 are calculated for the different drying rates. To account for the adsorption of surfactants to particles, Gundabala et al. use a Langmuir isotherm. The constants of the adsorption isotherms for surfactant adsorption on the particle interfaces are not known in our case. However, according to the model calculations of Gundabala et al., to obtain a surfactant profile as measured here, very high ratios of adsorption to desorption constants are needed to obtain such distinct profiles at relatively low surfactant Peclet numbers. If the adsorption of the surfactants on the particles leads to these surfactant profiles, they should also be present on PET substrates, which is not the case in our measurements. This invokes that the substrate plays an important role for the surfactant distribution, and the gradients are not likely to be formed due to pure adsorption on the particles and diffusion in the water phase. A schematic plot summarizing the main results is depicted in Figure 13. Figure 13. Summarizing schematic depicting the general influence of varying drying conditions and substrates for all three surfactants in both dispersions.

4. SUMMARY AND CONCLUSION In this study the influence of the drying rate and film temperature on the surfactant distribution in two different pressure-sensitive adhesive material systems was investigated. Hereby the different drying conditions in the lab and at industrial scale were in the focus of our study. High drying rates were obtained by using a self-designed batch coating and drying device which emulates an industrial large-area belt dryer. The surfactant distribution was measured by means of 3D confocal Raman spectroscopy after drying in two dispersions each containing three surfactants. The different formulations of the two investigated systems resulted in surprisingly similar surfactant distribution at high reproducibility. The surfactants were always enriched at the free surface and depleted at the glass substrate. A clear distinction between bulk and interface near area in terms of surfactant distribution could be seen. On siliconized PET substrates, commonly used for indirect coating and subsequent transfer, no depletion was found at the substrate interface, whereas the enrichment at the air interface was present. At low drying rate and low temperature (25 °C film temperature equivalent to 45

°C gas temperature of convective drying), the surfactant distribution reached an equilibrium state with no gradients in the bulk. Keeping the film temperature low, but increasing the drying rate, due to more intensive convection, results in a surfactant gradient in the bulk. At high temperatures the influence of different heat transfer coefficients disappeared. We could show in this work that drying conditions play an important role for the surfactant gradient formation. Higher colloidal stability leads to an increase of homogeneity of the lateral surfactant concentration, as shown by the measurement variance. Even though the literature proved that adsorption of the surfactants onto the particles might be an important mechanism, it was shown that the substrate surface energy has a distinct impact on the surfactant distribution even on the micrometer length scale. For future works in this field, it should be considered that results from experiments carried out at labscale conditions are suggested to be designed with controlled I


Research Article


Macromonomers in Methanol-Water. Comparison of Experiment with Theory. Macromolecules 1995, 28 (4), 1159−1166. (15) Erkselius, S.; Wadsö, L.; Karlsson, O. J. Drying rate variations of latex dispersions due to salt induced skin formation. J. Colloid Interface Sci. 2008, 317 (1), 83−95. (16) Mulvihill, J.; Toussaint, A.; De Wilde, M. Onset, follow up and assessment of coalescence. Prog. Org. Coat. 1997, 30 (3), 127−139. (17) Butler, L. N.; Fellows, C. M.; Gilbert, R. G. Effect of surfactant systems on the water sensitivity of latex films. J. Appl. Polym. Sci. 2004, 92 (3), 1813−1823. (18) Arnold, C.; Thalmann, F.; Marques, C.; Marie, P.; Holl, Y. Surfactant Distribution in Waterborne Acrylic Films. 1. Bulk Investigation. J. Phys. Chem. B 2010, 114 (28), 9135−9147. (19) Belaroui, F.; Hirn, M. P.; Grohens, Y.; Marie, P.; Holl, Y. Distribution of water-soluble and surface-active low-molecular-weight species in acrylic latex films. J. Colloid Interface Sci. 2003, 261 (2), 336−348. (20) Xu, G. H.; Dong, J.; Zhang, J.; Severtson, S. J.; Houtman, C. J.; Gwin, L. E. Characterizing the Distribution of Nonylphenol Ethoxylate Surfactants in Water-Based Pressure-Sensitive Adhesive Films using Atomic-Force and Confocal Raman Microscopy. J. Phys. Chem. B 2008, 112 (38), 11907−11914. (21) Arnold, C.; Klein, G.; Maaloum, M.; Ernstsson, M.; Larsson, A.; Marie, P.; Holl, Y. Surfactant distribution in waterborne acrylic films: 2. Surface investigation. Colloids Surf., A 2011, 374 (1−3), 58−68. (22) Kimber, J. A.; Gerst, M.; Kazarian, S. G. Fast Drying and Film Formation of Latex Dispersions Studied with FTIR Spectroscopic Imaging. Langmuir 2014, 30 (45), 13588−13595. (23) Gundabala, V. R.; Zimmerman, W. B.; Routh, A. F. A Model for Surfactant Distribution in Latex Coatings. Langmuir 2004, 20 (20), 8721−8727. (24) Kientz, E.; Holl, Y. Distribution of surfactants in latex films. Colloids Surf., A 1993, 78, 255−270. (25) Lee, W. P.; Gundabala, V. R.; Akpa, B. S.; Johns, M. L.; Jeynes, C.; Routh, A. F. Distribution of Surfactants in Latex Films: A Rutherford Backscattering Study. Langmuir 2006, 22 (12), 5314− 5320. (26) Zhao, C. L.; Dobler, F.; Pith, T.; Holl, Y.; Lambla, M. SurfaceComposition of Coalesced Acrylic Latex Films Studied by Xps and Sims. J. Colloid Interface Sci. 1989, 128 (2), 437−469. (27) Evanson, K. W.; Urban, M. W. Surface and interfacial FTIR spectroscopic studies of latexes. I. Surfactant−copolymer interactions. J. Appl. Polym. Sci. 1991, 42 (8), 2287−2296. (28) Schulz, M.; Römermann, H.; Pohl, K.; Chindawong, C.; Johannsmann, D. Latex Films with In-Plane Composition Gradients Caused by Lateral Drying. Soft Mater. 2015, 13 (3), 138−143. (29) Belaroui, F.; Grohens, Y.; Boyer, H.; Holl, Y. Depth profiling of small molecules in dry latex films by confocal Raman spectroscopy. Polymer 2000, 41 (21), 7641−7645. (30) Chen, X.; Fischer, S.; Men, Y. Temperature and Relative Humidity Dependency of Film Formation of Polymeric Latex Dispersions. Langmuir 2011, 27 (21), 12807−12814. (31) Routh, A. F.; Russel, W. B. A Process Model for Latex Film Formation: Limiting Regimes for Individual Driving Forces. Langmuir 1999, 15 (22), 7762−7773. (32) Carter, F. T.; Kowalczyk, R. M.; Millichamp, I.; Chainey, M.; Keddie, J. L. Correlating Particle Deformation with Water Concentration Profiles during Latex Film Formation: Reasons That Softer Latex Films Take Longer to Dry. Langmuir 2014, 30 (32), 9672−9681. (33) Baunach, M.; Jaiser, S.; Schmelzle, S.; Nirschl, H.; Scharfer, P.; Schabel, W. Delamination Behavior of Lithium-Ion Battery Anodes: Influence of Drying Temperature During Electrode Processing. Drying Technol. 2016, 32 (4), 462−473. (34) Alsmeyer, F.; Koss, H. J.; Marquardt, W. Indirect spectral hard modeling for the analysis of reactive and interacting mixtures. Appl. Spectrosc. 2004, 58 (8), 975−985. (35) Scharfer, P.; Schabel, W.; Kind, M. Mass transport measurements in membranes by means of in situ Raman spectroscopy - First

boundary conditions and need to be validated before their usage in industrial applications.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.6b00830. Fig. S1: Additional measurement at 25 °C in adhesive system 1 coated on glass; Fig. S2: Additional measurement at 60 °C in adhesive system 1 coated on glass; Fig. S3: Additional measurement at 25 °C in adhesive system 2 coated on glass; Fig. S4: Additional measurement at 60 °C in adhesive system 2 coated on glass (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Author Contributions ⊥

Susanna Baesch and David Siebel contributed equally to this paper. Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Keddie, J.; Routh, A. F. Fundamentals of latex film formation: processes and properties; Springer Science & Business Media: 2010. (2) Keddie, J. L. Film formation of latex. Mater. Sci. Eng., R 1997, 21 (3), 101−170. (3) Steward, P. A.; Hearn, J.; Wilkinson, M. C. An overview of polymer latex film formation and properties. Adv. Colloid Interface Sci. 2000, 86 (3), 195−267. (4) Winnik, M. A. Latex film formation. Curr. Opin. Colloid Interface Sci. 1997, 2 (2), 192−199. (5) Jovanović, R.; Dubé, M. A. Emulsion-Based Pressure-Sensitive Adhesives: A Review. J. Macromol. Sci., Polym. Rev. 2004, 44 (1), 1−51. (6) Yang, Y.-K.; Li, H.; Wang, F. Studies on the water resistance of acrylic emulsion pressure-sensitive adhesives (PSAs). J. Adhes. Sci. Technol. 2003, 17 (13), 1741−1750. (7) Xu, G. H.; Dong, J.; Severtson, S. J.; Houtman, C. J.; Gwin, L. E. Modifications of Surfactant Distributions and Surface Morphologies in Latex Films Due to Moisture Exposure. J. Phys. Chem. B 2009, 113 (30), 10189−10195. (8) Staicu, T.; Micutz, M.; Leca, M. Electrostatically and electrosterically stabilized latices of acrylic copolymers used as pressuresensitive adhesives. Prog. Org. Coat. 2005, 53 (1), 56−62. (9) Charmeau, J. Y.; Berthet, R.; Gringreau, C.; Holl, Y.; Kientz, E. Effects of film structure on mechanical and adhesion properties of latex films. Int. J. Adhes. Adhes. 1997, 17 (2), 169−176. (10) Isaacs, P. K.; Edelhauser, H. A. Influence of polymer solubilization on latex behavior. J. Appl. Polym. Sci. 1966, 10 (1), 171−183. (11) Okubo, M.; Ando, M.; Yamada, A.; Katsuta, Y.; Matsumoto, T. Studies on suspension and emulsion. XLVII. Anomalous composite polymer emulsion particles with voids produced by seeded emulsion polymerization. J. Polym. Sci., Polym. Lett. Ed. 1981, 19 (3), 143−147. (12) Eckersley, S. T.; Rudin, A. The effect of plasticization and pH on film formation of acrylic latexes. J. Appl. Polym. Sci. 1993, 48 (8), 1369−1381. (13) Farinha, J. P. S.; Martinho, J. M. G.; Kawaguchi, S.; Yekta, A.; Winnik, M. A. Latex Film Formation Probed by Nonradiative Energy Transfer: Effect of Grafted and Free Poly(ethylene oxide) on a Poly(nbutyl methacrylate) Latex. J. Phys. Chem. 1996, 100 (30), 12552− 12558. (14) Kawaguchi, S.; Winnik, M. A.; Ito, K. Dispersion Copolymerization of n-Butyl Methacrylate with Poly(ethylene oxide) J


Research Article

ACS Applied Materials & Interfaces results of methanol and water profiles in fuel cell membranes. J. Membr. Sci. 2007, 303 (1−2), 37−42. (36) Everall, N. J. Confocal Raman microscopy: why the depth resolution and spatial accuracy can be much worse than you think. Appl. Spectrosc. 2000, 54 (10), 1515−1520. (37) Singha, K. A review on coating & lamination in textiles: processes and applications. Am. J. Polym. Sci. 2012, 2 (3), 39−49. (38) Routh, A. F.; Zimmerman, W. B. Distribution of particles during solvent evaporation from films. Chem. Eng. Sci. 2004, 59 (14), 2961− 2968. (39) Cardinal, C. M.; Jung, Y. D.; Ahn, K. H.; Francis, L. F. Drying regime maps for particulate coatings. AIChE J. 2010, 56 (11), 2769− 2780.

K


Comparison of Surfactant Distributions in Pressure-Sensitive Adhesive

Recommend Documents