Article Cite This: ACS Appl. Polym. Mater. 2019, 1, 2213−2223
pubs.acs.org/acsapm
Cure-Dependent Morphology of Acrylic/Alkyd Hybrid Latex Films via Nanomechanical Mapping Elodie Limousin,† Daniel E. Martinez-Tong,‡,§ Nicholas Ballard,†,∥ and Jose ́ M. Asua*,†
Downloaded via NOTTINGHAM TRENT UNIV on August 13, 2019 at 13:52:59 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
†
POLYMAT and Departamento de Química Aplicada, Facultad de Ciencias Químicas, University of the Basque Country UPV/EHU, Joxe Mari Korta zentroa, Tolosa Hiribidea 72, 20018 Donostia, San Sebastián, Spain ‡ Departamento de Física de Materiales, University of the Basque Country UPV/EHU, P Manuel Lardizabal 3, 20018 Donostia, Spain § Centro de Física de Materiales (CSIC-UPV/EHU), P. Manuel Lardizabal 5, 20018 Donostia, Spain ∥ Ikerbasque, Basque Foundation for Science, E-48011 Bilbao, Spain S Supporting Information *
ABSTRACT: Acrylic/alkyd hybrid latexes offer the potential to obtain hard films from waterborne polymer dispersions without using coalescing agents. However, despite being widely used commercially, the curing process of the alkyd resin used in such formulations, and its influence on the final film morphology, remain poorly understood. In this work, we explore the curing process and its influence on film morphology in alkyd/acrylic hybrid latexes by nanomechanical mapping of the film during curing. It is shown that, although acrylic domains aid oxygen diffusion and facilitate curing, the rate of curing was substantially lower in the interior. Nevertheless, homogeneous curing was achieved at long times. The role of primary and secondary catalysts in the process is also explored, finding that the variation of the rate of curing of the alkyd plays an influential role in the development of film structure due to phase migration of the two polymers as curing progresses. It is hoped this work goes some way to unraveling the mystery behind the curing process in acrylic/alkyd hybrid latex films. KEYWORDS: hybrid latex, film structure, curing, nanomechanical mapping
1. INTRODUCTION Because of growing environmental concerns, protective and architectural outdoor coatings are progressively being changed from traditional solvent-based formulations to waterborne dispersions.1 However, formation of a film from a waterborne dispersion requires satisfying two seemingly contradictory requirements, with regard to their mechanical properties: the polymer must be soft enough to form a film at relatively low temperatures, but it must also be hard under the same conditions.2,3 Because of this, matching the performance of solvent-based coatings with waterborne dispersions is not trivial, and often additives such as coalescing agents must be added to the formulation.4,5 An alternative way to do this is by using a blend of two latexes: one should have a low glass-transition temperature (Tg) (50 °C), to provide film hardness. However, blending often leads to phase separation, which negatively affects the final properties of the coating.6−14 A more uniform distribution of the two phases can be obtained by using hybrid latex particles in which both phases are contained in the same particle.10,15−17 One important commercial example of hybrid polymer coatings, and the focus of this work, is the combined use of alkyd resins and (meth)acrylic polymers.10,18−20 The alkyd resin is in a liquidlike state before film © 2019 American Chemical Society
formation; therefore, relatively hard (meth)acrylic polymers can be used while still maintaining a low minimum film formation temperature (MFFT). After film formation, the double bonds present in the resin are cured by an autoxidative process that leads to hardening.21 The synthesis of acrylic/alkyd hybrid particles,10,18,19,22−26 as well as the morphologies of films cast from these hybrid particles latexes,10,20,27 has been widely studied. For example, Goikoetxea et al.10 synthesized acrylic/alkyd hybrid particles having a core− shell morphology with a hydrophobic alkyd resin core and an acrylic polymer shell. Films were cast from the hybrid latex particles without using curing catalysts. Under these conditions, in which modest or no curing is expected, they observed a significant degree of migration of the discrete polymer domains, resulting in a polymer film containing domain sizes significantly larger than the original latex particles. It was shown by a simulation model for film formation that phase migration was heavily influenced by the compatibility between the acrylic polymer and the free alkyd resin, which, in turn, was determined by the presence of an alkyd-acrylic graft copolymer.10 The graft Received: May 30, 2019 Accepted: July 11, 2019 Published: July 11, 2019 2213
DOI: 10.1021/acsapm.9b00507 ACS Appl. Polym. Mater. 2019, 1, 2213−2223
Article
ACS Applied Polymer Materials copolymer can form as a result of the presence of radicals, either by radical addition to the double bonds of the resin or, alternatively, by abstraction of allylic hydrogen in a chain transfer reaction,10,28 and subsequent propagation or termination on the alkyd during the radical polymerization process. Another critical feature not studied previously that can influence the film structure of acrylic/alkyd hybrids is the hardening of the alkyd resin itself. When the rate of hardening of the alkyd resin is much faster than that of phase migration and coalescence, it is expected that small domains sizes are obtained and the film morphology is determined by the initial particle morphology. Alternatively, if the rate of hardening is slow, then significant phase migration will occur, leading to large domain sizes. Given that the film structure is of paramount importance in determining the physical properties of the film,15,16,29−32 understanding the curing process of the alkyd resin in such hybrids is essential to enhance product performance. The curing of the alkyds is a complex autoxidative process, in which oxygen reacts with unsaturated fatty acid side chains of the alkyd resin (Figure 1). The process is divided in several
(also called surface catalysts) accelerate the decomposition of peroxides34 and, in addition, have been reported to increase the initial rate of peroxide formation.35 Although, for many years, the most commonly used commercial catalysts have been cobaltbased,36 because of its potential carcinogenic properties,37,38 cobalt catalysts are now often replaced by iron- or manganesebased catalysts.39−41 When used alone in pure alkyd resins, primary catalysts have a tendency to cure the film inhomogeneously.34,42 A highly cross-linked region is soon formed near the film/air interface, while the interior remains fluid. This crosslinked region advances very slowly toward the interior of the film, such that the interior may remain uncured, even after several days.43,44 This has been attributed to the low rate of oxygen diffusion across the highly cross-linked region, which, since oxygen is required for the propagation step (see Figure 1), retards curing. The secondary catalysts (zirconium, strontium, aluminum, calcium, potassium, lithium) are claimed to enhance homogeneous curing.44,45 Although they are universally applied in alkyd coating formulations, the mechanism by which improvements in the curing process is achieved remains unclear.44 Mallegol et al.21 reported that, using Co as the primary catalyst, the secondary catalyst accelerated the initial phases of curing and, in addition, they led to a more homogeneous curing. These are conflicting findings, because, if diffusion limitations are the reason for the heterogeneous curing, faster curing should lead to a more heterogeneous curing. Mechanisms involving an “oxygen-induced” hydroperoxide decomposition have been suggested, but this mechanism requires radicals of unknown origin.21 All of the works on curing discussed above involve pure alkyd resins. However, there are no studies on the curing of acrylic/ alkyd hybrids, where curing is complicated, because of the additional effects of phase migration. In this work, we study the curing process of alkyd resins in films cast from hybrid acrylic/ alkyd latex particles and explore how it affects the development of the film morphology. First, the effect of concentration of the Mn-based primary catalyst on the curing of the alkyd and the morphology of the polymeric film is investigated using an atomic force microscopy (AFM)-based nanomechanical mapping technique. In the second part, a secondary catalyst was included in order to understand its influence on the curing process and the film morphology of the hybrid material.
Figure 1. Radical reactions involved during the autoxidation of alkyd resin. (Reproduced with permission from ref 35. Copyright Elsevier, Amsterdam, 2017).
2. EXPERIMENTAL SECTION 2.1. Materials. Methyl methacrylate (MMA) (Quimidroga, technical grade), butyl acrylate (BA) (Quimidroga, technical grade), stearyl acrylate (SA) (Aldrich, technical grade), tert-butyl hydroperoxide (TBHP) (Panreac, >98%), NaHCO3 (Aldrich, >99.5%), and FF7 (Bruggolite) were used as received. 2,2′-Azobis (2-methylpropionitrile) (AIBN) (Sigma−Aldrich, 98%) was used as a thermal initiator. Dowfax 2A1 (alkyldiphenyloxide disulfonate, Dow Chemical Co.) was used as a surfactant. The alkyd resin (SETAL 293, hydrophobic) was kindly provided by Allnex. GPC-grade tetrahydrofuran (THF) (Scharlau) was used as received. Nuodex DRYCOAT (Huntsman, Mn primary catalyst) and Durham Nuodex Zirconium 12 (Huntsman, Zr secondary catalyst) were used as catalysts. Deionized (DI) water was used throughout the work. 2.2. Latex Synthesis and Characterization. Acrylic/alkyd hybrid latex particles were synthesized in batch by miniemulsion polymerization, following a method previously reported in the literature.23,46 The solids content was 50 wt %, and the alkyd/acrylic ratio was 50/50 (wt/wt). Residual monomers were measured by gas chromatography (GC), using a Hewlett−Packard GC system equipped with a headspace
stages: (1) initiation, (2) hydroperoxide formation, (3) hydroperoxide decomposition, and (4) cross-linking.33 In the first step, abstraction of the double allylic hydrogen of the unsaturated fatty acid occurs. The resulting radical reacts instantaneously with oxygen to form a hydroperoxide radical. Abstraction of another H atom from an unsaturated fatty acid leads to a hydroperoxide. Next, the hydroperoxides decompose into radical species. This step is slow and is aided by the addition of catalyst. The last step is the cross-linking, which occurs through termination of the radical species by recombination, resulting in a range of cross-linking points, including ether (COC), peroxide (COOC), and carbon−carbon (CC) bonds.34,35 Two types of catalysts are used to accelerate the curing process: primary and secondary catalysts. The primary catalysts 2214
DOI: 10.1021/acsapm.9b00507 ACS Appl. Polym. Mater. 2019, 1, 2213−2223
Article
ACS Applied Polymer Materials
For AFM measurements, films were cut with a microtome that was equipped with a diamond knife at 6 °C. A sample holder designed for cross-section analysis was used. This fixed the sample in the same position throughout the entire process (cutting + imaging). Considering that once the sample cross-section is exposed to the air conditions, cross-linking will start to occur, PF-QNM analysis was performed immediately after microtoming. Also, following these considerations, it is impossible to use the same sample for a continuous analysis as a function of time. In order to perform measurements with time, each AFM measurement was performed on a nominally equivalent, but separate film. Finally, we note that all of the samples were thick enough to allow disregarding the possible impact of the supporting substrate on the nanomechanical properties. Briefly, PF-QNM allows one to capture force−distance curves at each pixel, at high speed and with high resolution. Every force curve records the force on the probe as it approaches and retracts from a point on the sample surface. When the force reaches a user-specified trigger value, or peak force, the system records the height for that pixel and the tip retracts. This allows one to record both topography information and mechanical properties simultaneously. In our work, we used a fixed peak-force set-point value of 15 nN for all the probed samples. We calculated the mechanical modulus from the force curves using the Derjaguin−Müller−Toporov (DMT) model:50
sampler. Particles sizes were determined by dynamic light scattering (DLS) spectroscopy, using a Malvern Zetasizer nano, which provides a Z-average diameter. The gel fractions of the dried latexes were determined by Soxhlet extraction, using technical-grade THF as a solvent. A glass fiber disk was impregnated with latex (a few drops) and allowed to dry in a vacuum oven at room temperature overnight in order to avoid cross-linking of the alkyd resin with oxygen. The gel content was calculated based on a previously reported method.47 The molecular weights of the soluble fraction of the polymers were determined by gel permeation chromatography (GPC). The soluble part of the polymers from the Soxhlet extraction was dried and redissolved THF at a concentration of ∼1 mg/mL. It then was filtered (polyamide Φ = 45 μm) before being injected into the GPC via an autosampler (Waters, Model 717). A pump (Shimadzu, Model LC20A) controlled a THF flow of 1 mL/min. The GPC was composed of a differential refractometer detector (Waters, Model 2410) and three columns in series (Styragel HR2, HR4 and HR6, with pore sizes ranging from 102 to 106 Å). Measurements were performed at 35 °C. Molecular weights were determined using a calibration curve that was based on polystyrene standards. The mass fraction of alkyd grafted to the acrylic polymer (represented by the resin degree of grafting (RDG)) in the sol fraction were determined by GPC, the method of which has been explained elsewhere.23 The double-bond content of the resin (RDB) was determined by iodine titration, based on the Wijs method,48 and calculated as the relationship between the iodine value of the alkyd in the hybrid particle latex after polymerization and the iodine value of the neat alkyd resin. Particle morphology was determined by transmission electron microscopy (TEM), using a Jeol Model TM-1400 Plus series 120 kV electron microscope. A drop of the diluted latex was deposited on a copper grid and allowed to evaporate. The sample then was stained with OsO4 for 2 h. With this staining, the alkyd resin will appear darker through reactions of the double bonds of the alkyd resin with the OsO4.49 2.3. Film Preparation and Nanomechanical Mapping. Before film formation, the catalyst was added to the latex under stirring in a vortex mixer (VELP Scientifica, Model ZX3) for 30 s at 2500 rpm and then 1 h under magnetic stirring using 0.25, 1, and 2 wt % (based on alkyd resin content) of Nuodex DRYCOAT, an oil-based manganese primary catalyst. The secondary catalyst (Nuodex Zirconium 10−3.5 wt %) was added under the same conditions. Table 1 summarizes the
F − Fadh =
Composition (%, Based on Alkyd Resin) name code
DRYCOAT
Durham Nuodex Zirconium 10
0.25 1 2 1
/ / / 3.5
(1)
where F is the force, Fadh the adhesion force that occurs during tip retraction from the surface, E the mechanical modulus, R the tip radius, δ the indentation, and υ the Poisson ratio of the sample. The mechanical modulus fit, following eq 1, was performed, using 90% and 20% of the peak-force set point as maximum and minimum boundaries, respectively. In this region of the force curve, the trace and retrace were superimposed, indicating an elastic response. This approach has been used extensively in recent literature to study the nanomechanical properties of polymers.51−53 PF-QNM images were captured using a 256 × 256 resolution, and a frequency modulation (f m) of 2 kHz. All samples were probed using Multi75-G probes, by BudgetSensors (tip radius ≈ 15 nm, as estimated from a tip-checker sample provided by Bruker). The spring constant was determined using the Sader method,54 giving a value of ∼3−5 N/m. The resonance frequency (f 0) of the probe was in the 80−90 kHz range, a value that allows one to avoid possible filtering effects.55,56 The ratio between f m and f 0 was in the 40−45 range, which is a value close to that proposed in the literature for proper mechanical moduli quantification via nanomechanical mapping.57,58 PF-QNM quantitative images were acquired following the relative calibration method proposed by Bruker and recently used in the literature.56,59 The relative method of calibration uses a sample of known modulus to obtain the ratio of spring constant to the square root of the tip end radius. In our work, we used as a reference an acrylic sample prepared by us (MMA/BA: 50/50 wt/wt), with a known modulus of 70 MPa (υ = 0.4),60 for imaging the mechanical maps of the hybrid latex films. Using this reference, we obtained maps with good mechanical contrast between soft/hard areas. However, we observed that mechanical modulus values of
