Published in Journal of Adhesion (2006) 82(3): pp 217-238 Skin Development during the Film Formation of Waterborne Acrylic Pressure-Sensitive Adhesives containing Tackifying Resin J. Mallégol,1+ G. Bennett,1 O. Dupont,2 P. J. McDonald1 and J. L. Keddie1* 1. Department of Physics, School of Electronics and Physical Sciences, University of Surrey, Guildford, Surrey GU2 7XH, United Kingdom 2. Cytec Surface Specialties, 33 Anderlechtstraat, Drogenbos B1620, Belgium Abstract. Tackifying resins (TR) are often used to improve the adhesive properties of waterborne pressure-sensitive adhesives (PSAs) derived from latex dispersions. There is a large gap in the understanding of how, and to what extent, the film formation mechanism of PSAs is altered by the addition of TR. Herein, magnetic resonance profiling experiments show that the addition of TR to an acrylic latex creates a coalesced surface layer or “skin” that traps water beneath it. Atomic force microscopy of the PSA surfaces supports this conclusion. In the absence of the TR, particles at the surface do not coalesce but are separated by a second phase composed of surfactant and other species with low molecular weight. The function of the TR is complex.. According to dynamic mechanical analysis, the TR increases the glass transition temperature of the polymer and decreases its molecular mobility at high frequencies. On the other hand, the TR increases the molecular mobility at lower frequencies and thereby promotes the interdiffusion of latex particles to create a skin layer.In turn, the skin layer is a barrier that prevents the exudation of surfactant to the surface. The TR probably enhances the coalescence of the latex particles by + Current address: Arcelor Research Liège SCRL, Bd. de Colonster, B57, B-4000 Liège, Belgium 1 Published in Journal of Adhesion (2006) 82(3): pp 217-238 increasing the compatibility between the acrylic copolymer and the solids in the serum phase. Running Title: Film Formation of Tackified Waterborne Adhesives Key words: tackifier; latex; film formation; atomic force microscopy; magnetic resonance profiling; coalescence Introduction Pressure-sensitive adhesives (PSAs) constitute a distinct category of adhesives that instantly wet and firmly adhere to a variety of dissimilar surfaces, when applied with only light pressure, without activation by water, heat or solvent.[1] A major contribution to the energy that is required to debond two surfaces joined by a PSA is from its viscoelastic energy dissipation.[2,3] Recent work by Brown and co-workers [4] has vividly demonstrated the influence of polymer mobility on the adhesion force between solids and polymers. The adhesive performance of a PSA depends strongly on the balance of viscoelastic properties, and hence a considerable amount of research has aimed to correlate the two. [1,5,6,7] PSAs are often manufactured from acrylic ester copolymers that have a glass transition temperature, T , as low as -60°C. These polymers are inherently tacky g without any additional compounding, and their adhesive properties can be modified through a variation of the copolymer composition. It is a common practice, however, to incorporate a tackifying resin (TR), also known as a tackifier, in acrylic formulations to enhance their properties further.[6,8,9] A TR is usually a bulky, low molecular weight molecule, such as n-butyl ester of abietic acid [10] or * Corresponding author: [email protected] 2 Published in Journal of Adhesion (2006) 82(3): pp 217-238 pentaerythritol rosin ester derived from abietic acid.[11] The properties of a PSA can be finely tuned through the addition of an appropriate TR.[12] The effects of tackifiers on viscoelasticity and adhesion have been thoroughly studied [13,14,15,16]. The effect of a compatible TR on polymer viscoelastic properties is twofold. First, the elastic modulus E’, at the temperature of the PSA’s use, is lowered. Second, blending with a compatible tackifier causes a shift of the glass transition temperature (T ) to a higher value.[8] The lower E’ promotes polymer flow and bond formation g and, coupled with the higher T , enhances the resistance to bond rupture.[17] In terms g of the strain rate, a tackified PSA is stiffer at high strain rates, such as during debonding, but it flows more easily at low strain rates, such as when wetting a substrate.[8] In acrylic PSAs, the most noticeable effect of the addition of TR is the enhancement of tack and peel strengths to substrates with a low surface energy, such as polyolefines and carton board.[6] Problems presented by the use of TR in acrylic PSAs are a lower resistance to the ageing of TR/PSA mixtures and a lower shear holding power.[6] Although the two phases can be soluble in a common solvent, when the solvent leaves the PSA, phase separation can occur. Tackified PSAs can also undergo slow phase separation and surface segregation, resulting in a change in the adhesive and mechanical properties.[10] Hence, the degree of compatibility between the polymer and TR is a very important factor in formulating a PSA. This fact explains the great interest in TR miscibility with polymers and hence the large number of publications on the miscibility in such systems.[18,19,20,21,22,23,24] 3 Published in Journal of Adhesion (2006) 82(3): pp 217-238 Tighter environmental regulations in the production of PSAs have led to a shift away from solvent-cast formulations to aqueous dispersions of polymer colloids, i.e. latex. Waterborne PSA technology has increasingly more often been a subject of research, as indicated in a recent review article.[25] Generally it has been found that the performance of waterborne acrylic PSAs is inferior to that of their solvent-based analogues. They exhibit lower water resistance,[26] a tendency to whiten in moist atmospheres, and lower tack, adhesion and shear strengths.[6,27,28,29,30] This relatively poor performance of waterborne PSAs has been correlated with a heterogeneous film morphology[27] and with the discontinuity of the molecular network structure[30] of dry films. The distribution and migration of small molecules, especially surfactants,[27,29,31] has also been correlated with poor waterborne adhesive performance. In our previous research,[32,33,34] we have determined the morphology of waterborne PSAs as a first step in improving their performance. In formulating waterborne PSAs, the use of TR dispersions in water is preferred for obvious reasons. The inferiority of waterborne PSA shear strength, however, is even greater when waterborne, acrylic PSAs are compounded with waterborne dispersions of TR.[9] Furthermore, in waterborne, tackified PSAs, an additional level of complexity is added by the requirement for compatibility of latex and tackifier emulsifiers with each other and with the polymer.[35] There has been only limited published research on the characteristics of waterborne, tackified PSAs. [9,17,35,36] The work of Tobing and Klein[36] is particularly significant because it considers the mechanism by which tackifiers improve waterborne PSA performance. These authors commented, however, that the 4 Published in Journal of Adhesion (2006) 82(3): pp 217-238 development of tackified PSA technology was being limited by a poor understanding of the film formation process. The present work was motivated by this knowledge gap and by an appreciation of the importance of film formation in determining waterborne PSA properties.[25] In our previous research,[33] we found with magnetic resonance (MR) profiling that during the drying of an acrylic waterborne PSA film, a water concentration gradient develops. There was less water near the interface with the atmosphere, however the water level never receded from this interface. Unlike the drying of silicone[37] and alkyd[38] emulsions, studied elsewhere, there was no evidence for a dry surface layer or “skin”. Atomic force microscopy (AFM) images of dry PSA surfaces revealed that the particles were not coalesced but were separated from their neighbors by a liquid- like medium. With complementary use of Rutherford backscattering spectrometry (RBS), we deduced that surfactant (possibly in addition to other water-soluble species) was present in excess at the surface, where it stabilizes the particles against coalescence. We speculated that because particles did not coalesce during this stage of drying, the distribution of surfactant (and other water-soluble species) along the particle/particle boundaries was enabled. In the present work, we consider the drying and film formation mechanisms in tackified, waterborne acrylic PSAs. We provide the first report of how the presence of a waterborne TR modifies the water concentration profiles and particle coalescence in a waterborne acrylic PSA. We show that the TR has a major impact. There are no thorough studies of phase stability and ageing in tackified, waterborne acrylic PSAs. Therefore, this issue was also investigated in the current work. 5 Published in Journal of Adhesion (2006) 82(3): pp 217-238 Experimental Materials and Methods Materials A model acrylic PSA latex (referred to in previous publications [32,33] as PSA A) was investigated. This latex, with a solids content of 60 wt.%, has a bimodal particle size distribution (weight-averaged particle sizes of 180 nm and 350 nm determined by dynamic light scattering (Nicomp 370 particle sizing systems)) and a solids content of 60 wt.%. A random copolymer that makes up the latex particles consists mainly of 2- ethylhexyl acrylate, an acrylic ester that yields a soft and tacky material with a low glass transition temperature (T ), with additions of methyl methacrylate and polar g monomers (acrylic acid and methacrylic acid). DSC analysis of the copolymer obtained a T of -45 °C. The latex was prepared by standard techniques of semi-batch g emulsion polymerization. Dissolution of the latex polymer in organic solvent has revealed a low insoluble fraction, which indicates the presence of polymer molecules with very high molecular weight and/or a partially cross-linked network. The loop- tack strength of the PSA film, determined according to Finat Test Method No. 9 (FTM9), is 12.8 N/inch (on stainless steel substrate). The TR is a stabilized rosin ester (Tacolyn3189 from Eastman Chemical, Kingsport, Tenn., USA) dispersed in water at a solids content of 50 wt.% and with a weight- average particle size of 220 nm According to DSC, the T of the dry TR is 29 °C, but g this value is as low as 20 ºC when it is not fully dried, as a result of plasticization by water. The softening point, determined by the Hercules drop method, is 70 °C. The latex was blended with the TR dispersions at various concentrations: 1%, 5%, 10%, 6 Published in Journal of Adhesion (2006) 82(3): pp 217-238 18%, 25%, 33%, 41%, 50%, and 75% tackifier (by weight). The blended dispersions were stirred for several hours to ensure complete mixing. Dynamic Mechanical Analysis The dynamic tensile moduli (E’ and E”) of the acrylic PSA and the tackified PSA films were determined by dynamical mechanical analysis in tensile deformation using a 2980 Dynamic Mechanical Analyzer (TA Instruments, New Castle, USA). A strain oscillation with an amplitude of 0.1 % was applied at a frequency of 1 Hz. The temperature was increased from -60 to 120 °C at a heating rate of 3 °C/min. The linearity condition was verified at selected temperatures. Free-standing PSA films (1 mm thick) were prepared by de-laminating films that had been cast on release paper, dried for 48 hr. at ambient temperature, and then heated for 16 hr. at 50 °C. Before analysis, the films were submitted to an additional heating at 110 °C for 2 hr. Lateral dimensions of the rectangular film samples were 10 mm x 8 mm. Atomic force microscopy of PSA surfaces Films were cast onto silicone-coated paper release liners (30 cm x 20 cm) using a 40 µm hand-held bar coater. The films were dried under laminar air flow for 3 min. on heated plates at 60 °C in a controlled humidity chamber maintaining a relative humidity of 40 %. The dried films for AFM analysis were about 20 µm thick. Small pieces (1cm x 1cm) of the cast PSA were cut from the large-area films and were analyzed with an atomic force microscope (Nanoscope IIIa, Digital Instruments, 7 Published in Journal of Adhesion (2006) 82(3): pp 217-238 Santa Barbara, CA, USA) within three hours of casting. All measurements used a silicon cantilever (NT-MDT, Moscow, Russia) equipped with an ultrasharp, conical silicon tip having a radius of curvature of about 10 nm. The nominal resonant frequency f of the cantilever is 320 kHz and its spring constant k is 48 N/m. o AFM analysis was performed on the original interface with air. Images were recorded simultaneously in the topographic (height) mode and in the phase mode, with scan sizes ranging from 5 µm to 30 µm. The optimum method for obtaining images of latex PSA surfaces has been reported previously.[32] Parameters needed to describe the tapping conditions are the "free" amplitude A and the setpoint value d . o sp The high tack of the PSA surface makes it necessary to use high tapping amplitudes to impart enough energy to the tip to "pull off" of the adhesive surface. A high setpoint ratio (d /A ) is required to minimise indentation of the tip and thereby to reduce sp o deformation of the soft surface. To obtain values of A and d in metric units, a o sp systematic calibration of the cantilever was obtained from amplitude-distance curves on a clean silicon wafer, assuming no deformation of the silicon surface and no bending of the cantilever during tapping.[39] All AFM images presented here were obtained with very similar tapping conditions. Typically, d was between 90 and 100 nm for all measurements. A was typically 20 sp o nm above the d for the original air interface of the PSAs and 30 to 40 nm above the sp d for the face delaminated from the release liner. These tapping conditions ensure sp that the indentation of the AFM tip into the PSA surface is small enough to avoid significant distortion of the morphology. It is, however, important to realise that the 8 Published in Journal of Adhesion (2006) 82(3): pp 217-238 the AFM tip does indent into the polymer surface, as shown previously, but that the structure is not permanently altered.[32,34] NMR Spectroscopy 1H NMR spin-spin relaxation time (T ) distributions were measured for dried films 2 with varying concentrations of TR on a low resolution, 20 MHz spectrometer (Maran, Oxford Instruments, Abingdon, Oxon., UK). Specifically, free induction decay (FID) and Carr-Purcell-Meiboom-Gill (CPMG) echo train signals were acquired. T is a 2 measure of molecular mobility; a higher value indicates greater mobility. The free induction decays were recorded with a sampling rate of 2 MHz following a short 90° excitation pulse of 3.9 µs duration, enabling the observation of faster relaxing components. However, the instrument's dead time of ca. 25 µs prevented the very fastest components of the dried TR from being detected. The CPMG trains were used to observe the slowly relaxing components better. The basic 90° - 180° pulse gap was 64 µs. Alternate echoes were recorded out to 16.4 ms. In both experiments, a total of 64 averages was collected at a repetition interval of 2 s. To prepare samples, aqueous dispersions were cast on silicone-coated paper (using an 80 µm bar coater), dried at 60 °C for 10 min., de-laminated from the surface, and then rolled and inserted into glass NMR tubes (1 cm diameter) for analysis at 23 ºC. The samples were all prepared on the same day under identical conditions and analysed approximately 3 hr after casting. An additional experiment was carried out to examine the rate of mixing of the TR and PSA. Two films of pure TR and pure PSA were cast separately on silicone paper. 9 Published in Journal of Adhesion (2006) 82(3): pp 217-238 After they were dry, they were pressed together face-to-face, then peeled from the silicone-coated paper substrates, and finally inserted into an NMR tube for analysis. Initially the two substances are completely separate, and there is no molecular interaction, except perhaps at the interface. The FID was obtained at various times after bringing the two materials into contact. MR Profiling A small permanent magnet, which was designed specifically for obtaining magnetic resonance profiles of 1H in planar films, was used to probe the water concentration in drying latex layers in the direction normal to the substrate. The details of this magnet, called Gradient At Right-angles to the Field (or GARField), have been reported previously.[40,41] In the experiments performed here, samples were placed in the magnet at a position corresponding to a magnetic field strength of 0.7 T and a field gradient strength of 17.5 T m-1. In experiments, latex films were cast onto clean glass coverslips (2 cm x 2 cm) using either a 120 µm or a 250 µm applicator. Immediately after casting, the film was placed in the magnet. MR profiling was commenced with the sample in the open atmosphere at an average temperature of 23 ºC within the instrument. Signals were obtained using a quadrature echo sequence[42]: 90 - τ - x (90 - τ - echo - τ -) for n = 32 echoes and a pulse gap of τ = 95.0 µs. To obtain a y n profile, the echoes were Fourier-transformed and then summed, thus giving the NMR signal intensity profile as a function of depth with a pixel resolution of 8.8 µm. Profiles were normalized by an elastomer standard in order to correct for the sensitivity decline over the film thickness. Results and Discussion 10