Effect of Molecular Weight on Packing during Latex Film Formation

Journal of Colloid and Interface Science 234, 72–78 (2001) doi:10.1006/jcis.2000.7280, available online at http://www.idealibrary.com on

Effect of Molecular Weight on Packing during Latex Film Formation ¨ Ertan Arda∗ and Onder Pekcan† ∗ Department of Physics, Trakya University, 22030 Edirne, Turkey; and †Department of Physics, Istanbul Technical University, Maslak 80626 Istanbul, Turkey Received February 29, 2000; accepted October 16, 2000

concentration. As the solvent evaporates a uniform shrinkage of the interparticle distance occurs and the voids are gradually filled by particle sliding until a dense packing of spheres is obtained. Evaporation of solvent leads to the second stage in which the particles form a close packed array. Here, if the particles are soft they are deformed to polyhedrons. Hard latex, however, stays undeformed at this stage and annealing the hard latex system deformation of particles first leads to viscous flow and void closure (3, 4). This stage is also called coalescence. After void closure the process is completed and the mechanism of film formation, by the annealing of hard latex films, is known as the interdiffusion of polymer chains, which leads to the third stage. The process of coalescence of latex particles into a film is thermodynamically favorable because of the decrease in free energy achievable with a minimization of the total surface. All contiguous particles would flow into deformed spheres representing a minimization of surface and gravitational energy. In the first stage of film formation, as the solvent evaporates from the surface, the particles’ centers approach each other but the particles remain until they are forced into contact by spatial limitations. The second stage begins when the particles can no longer slide each other into new positions. The particles are brought into close proximity so that their stabilizing layers may collapse, resulting in polymer–polymer contact. The particle coalescence resulted from the viscous flow of polymers induced by the surface tension between solvent and polymer particles, where coalescence is driven by capillary forces. Once the second stage has completed, the third stage can begin. In this final stage, interdiffusion of polymer chains of adjacent particles across particle boundaries causes further coalescence. Several studies have been published in this matter. Small angle neutron scattering (SANS) has been used to examine deuterated particles in a protonated matrix (5). More extensive studies have been performed using SANS by Sperling and co-workers (6, 7) on compression molded PS latex films. These works covered the interdiffusion process during film formation. Alternatively, the process of interparticle polymer interdiffusion has been studied by direct nonradiative energy transfer (DET) using fluorescence decay measurements in conjunction with particles labeled with appropriate donor and acceptor chromophors (8–11). This transient fluorescence technique has been used to examine latex film formation of 1-µm diameter high-T polymethyl methacrylate (PMMA) particles (8) and of 100-nm

A UV-visible technique is used to study the evolution of transparency during film formation from latex particles. Latex particles with high and low molecular weight (HM and LM) polymethyl methacrylate (PMMA) are used to prepare films. Two sets of films with different latex content were prepared from HM and LM particles separately, by annealing PMMA particles above the glass transition temperature. Transmitted photon intensity, Itr , from HM and LM films increased as the annealing temperature was increased. The increase in the transmitted photon intensity is attributed to the latex content (film thickness) for the annealed film samples. It is suggested that as the latex particles are packed (film thickness is increased) fewer voids or cracks are formed in the films. Positive and negative absorption coefficients are measured below and above 210 and 180◦ C annealing temperatures for the HM and LM films. Packing coefficients are obtained for films in various latex contents. It is observed that LM particles are packed much easier than HM particles. °C 2001 Academic Press Key Words: molecular weight; photon transmission; latex content; packing coefficient; absorption coefficient.

INTRODUCTION

In the past two decades new coating technologies such as high solids, powder, water-borne, and radiation-curable coatings have been developed to meet the requirements of governmental regulations in ecology and the increasing costs of petroleum-based solvents. The process of latex film formation is the basis for these new coating technologies and the topic has received extensive attention in the past decade in particular. The process consisting of driving a latex from its colloidal state to a continuous film needs to be well understood because of its great importance in the coating industry. Latex film formation is a critical aspect of all applications that involve coating a surface or forming a layer with good cohesive properties. Consequently, great efforts have been devoted to studying the processes of latex film formation. The term “latex film” normally refers to a film formed from soft particles where the forces accompanying the evaporation of water are sufficient to compress and deform the particles into a transparent, void-free film (1, 2). However, hard latex particles remain essentially discrete and undeformed during the drying process. Film formation from these dispersions can occur in several stages. In both cases, the first stage corresponds to the wet initial state which consists of solvent evaporation and colloidal 0021-9797/01 $35.00

C 2001 by Academic Press Copyright ° All rights of reproduction in any form reserved.

72

73

PACKING DURING LATEX FILM FORMATION

diameter low-T polybutyl methacrylate (PBMA) particles (9, 10). These studies all indicate that in the particular systems examined, annealing the films above Tg leads to polymer interdiffusion at the particle–particle junction as the particle interface heals. Mazur (12) has written an extensive review on the coalescence of polymer particles, where he mainly discussed the neck growth mechanism and its several geometrical approximations before the interdiffusion of polymer chains takes place. Recently, the steady state fluorescence and photon transmission techniques have been used in this laboratory to study interdiffusion processes at the particle–particle junction during film formation by latex particles (13–19). In this work films formed from two different hard latex particles of two different molecular weights. These two different films were prepared with various latex contents and annealed in equal time intervals at elevated temperatures above Tg . Transmitted photon intensities, Itr , from annealed latex films were measured by the UV-visible (UVV) technique to study the molecular weight dependent packing effect on latex film formation. It is observed that below 180◦ C, Itr obeyed Lambert’s law in low molecular (LM) latex films with positive imaginary absorption coefficient, κ. Above 180◦ C, however, negative κ is detected for LM samples. High molecular (HM) latex films present positive and negative κ below and above 210◦ C, respectively. A phenomenological equation is used to interpret the dependence of Itr on annealing temperature and packing coefficients; α was measured for LM and HM films and quite different behaviors were observed with respect to latex content. EXPERIMENTAL

Two different batches of PMMA-polyisobutylene (PIB) polymer particles were prepared separately in a two-step process. First MMA was polymerized to low conversion in cyclohexane in the presence of PIB containing 2% isoprene units to promote grafting. The graft copolymer so produced served as a dispersant in the second stage of polymerization, in which MMA was polymerized in a cyclohexane solution of the polymer. Details have been published elsewhere (20). In both batches a stable dispersion of spherical polymer particles was produced, ranging in radius from 1 to 3 µm. A combination of 1 H-NMR and UV analysis indicated that these particles contain 6 mol% PIB. (These particles were prepared by B. Williamson in Professor M.A. Winnik’s laboratory in Toronto.) In the first and second batches of particles, molecular weights of graft PMMA were measured as Mw = 2.15 × 105 and Mw = 1.10 × 105 , respectively. These particles are used to prepare HM and LM samples respectively by redispersing them in heptane. The polydispersity of the corresponding PMMA were 1.49 and 2.33 for the HM and LM particles. Two different sets of films were prepared from the dispersions of HM and LM particles by placing different number of drops on glass plates of size 0.9 × 3.2 cm2 . Each set of samples contained seven different films. The heptane was allowed to evaporate and the LM and HM films were annealed

TABLE 1 Latex type

Film thickness (µm)

α (◦ C−1 )

LM

2.70 4.05 5.36 6.75 8.09 9.44 11.70

0.01614 0.02537 0.03125 0.04168 0.04392 0.04735 0.04941

HM

2.37 3.55 4.73 6.31 7.88 10.25 13.40

0.00534 0.01235 0.01689 0.01966 0.02173 0.02042 0.01787

above the glass transition temperature of PMMA for 30 min time intervals at elevated temperatures from 140 to 210◦ C and 130 to 270◦ C, respectively. During annealing, the temperature was maintained within ±1◦ C. Samples were weighed before and after film casting to determine the latex contents and film thicknesses. These values are summarized in Table 1. Average particle size was taken as 2 µm to calculate the thickness of the films. UVV experiments were carried out with the Lambda 2S UV-Visible spectrometer of Perkin–Elmer and transmittance of films was detected between 300 and 400 nm. Film samples were placed perpendicular to the incident light direction during UVV experiments. The transmitted light intensities from latex films were measured during the latex film formation process. TURBIDITY OF LATEX FILMS

When the latex film is formed and annealed it goes from a turbid to a transparent state. Usually the turbidity of a medium is created by the scattering of light. If the intensity of incident light is Io , and on passage of the light through a medium of thickness, d, the incident intensity is reduced to I , as a result of scattering and the turbidity of the medium is determined by I /Io = e−τ d ,

[1]

where turbidity, τ , is defined as the fraction of the primary light beam scattered in all directions on passage through a medium with thickness of 1 cm, i.e., τ = − ln(I /Io ).

[2]

The expression I /Io has been known as the light transmission or relative transparency. In general turbidity is given by τ = (ε + κ)c, where ε and c are the molar absorption coefficient and molar concentration, κ is known as the imaginary

74

ARDA AND PEKCAN

absorption coefficient, and ε usually depends on the wavelength of the light being observed, temperature, and the nature of the medium. ε presents the true absorption of light; i.e., when light is observed by a system its energy is transformed into thermal energy. The molar absorption coefficient, ε, does not include the size of the particles if they exist in the medium. It is known that the size of colloidal particles affects light absorption indirectly through either light scattering or reflection. In general light is scattered only when its wavelength is greater than the size of a particle in the medium. If the wavelength of the light is much smaller than the particle diameter, light is reflected. When the size of particles becomes considerably larger than λ, light is no longer scattered but reflected, regardless of the light wavelength. If particles are too large, light reflection from them increases which cause the reduction of the intensity of scattered light. The fact is that, as a result of light scattering, the white light which passes through a medium still loses the same radiation in the short wavelength region which is not the true absorption. Such absorption is known as the imaginary absorption which is identified with κ, which is a function of particle size. If ε = 0 the system has no true absorption; i.e., the system is white. In our experiments the wavelength of the light is ten times smaller than the latex particle size and is comparable to the void size. In our case it is believed that latex films do not reflect light or make true absorption. A decrease in the transmitted light intensity most probably caused by the imaginary absorption occurred in the latex system.

RESULTS AND DISCUSSION

Transmitted photon intensities from HM and LM films were obtained and plotted versus annealing temperature in Figs. 1a and 1b for seven different latex films in various thicknesses. It is seen that all Itr intensity curves increase as the annealing temperature is increased. When the Itr intensities are compared for HM and LM samples, it is seen that the HM film needs annealing temperatures to reach the same transparency as that of the LM film for the same film thicknesses. Figure 2 compares the Itr curves for the HM and LM samples of similar thicknesses. The behavior of Itr suggests that latex films become transparent to photons as they are annealed. Relatively small Itr intensities were observed in low latex content films by indicating that some photons dissipate, i.e., cannot reach to photodiode after they pass through these films. If the latex content is increased, Itr presents larger values as the annealing temperature is increased. This behavior of Itr from both HM and LM samples intuitively predicts that samples that have larger latex content create more transparent films. In Figs. 3a and 3b Itr is plotted versus film thickness for HM and LM samples where it is seen that, below 210 and 180◦ C, Itr decreased as the film thickness is increased; however, above 210 and 180◦ C, Itr increased as film thickness is increased in HM and LM samples. In other words, below 210

FIG. 1. Plot of Itr versus annealing temperature for seven different latex films of (a) HM and (b) LM annealed at 30 min time intervals. Numbers on each curve indicate the film thickness.

and 180◦ C films obey Itr = Io e−κd ,

[3]

where κ is the imaginary absorption coefficient and d is the film thickness. Above 210 and 180◦ C Eq. [3] is obeyed with negative κ. The data in Fig. 3a are fitted to the logarithmic form of Eq. [3] and plotted in Figs. 4a and 4b below and above 210◦ C. Similar analyses are done in Figs. 5a and 5b for the LM samples below and above 180◦ C. The slopes of the linear relations in Figs. 4a

75

PACKING DURING LATEX FILM FORMATION

TABLE 2 Latex type

Annealing temperature T (◦ C)

κ (◦ m)−1

LM

140 150 160 170 180 190 200 210

0.2705 0.1919 0.1482 0.0917 0.0662 −0.0522 −0.0726 −0.0973

HM

130 150 170 190 210 230 250 270

0.3882 0.3255 0.2148 0.1398 0.1212 −0.0196 −0.0811 −0.1707

and 4b and Figs. 5a and 5b produced positive and negative κ values, which are plotted versus annealing temperature in Fig. 6 for HM and LM samples and are listed in Table 2. Below 210 and 180◦ C, the behavior of κ values are understandable; i.e., as the thickness of the film is increased, absorption of light decreased in both HM and LM samples. Above 210 and 180◦ C, an increase in film thickness amplifies the transmitted light intensity. Most probably an increase in latex content decreases the imperfections in HM and LM film such as voids and cracks by annealing them above 210 and 180◦ C. The fit of the data in Fig. 6 obey the

FIG. 3. Plot of Itr versus film thickness for (a) HM and (b) LM samples. Numbers on each curve indicate the annealing temperature.

empirical relation as κ = κ1 − κo T,

FIG. 2. Comparison of Itr versus annealing temperature for HM and LM film samples with the thicknesses of 6.31 and 6.75 µm, respectively.

[4]

where κ1 and κo are the parameters depending on the material that are used. The slopes of the linear relations in Fig. 6 produce κo values as 4.0 × 10−3 and 5.4 × 10−3 (µm ◦ C)−1 for LM and HM film samples, respectively. The differences in κo values predict that transparency in HM films evolves slower than in LM films. This picture confirms the behavior of Itr in Fig. 2. For example, the HM film annealed at 170◦ C produces less transparent film than the LM sample annealed at the same temperature. This

76

ARDA AND PEKCAN

on the latex content (see below) it is called a packing coefficient. The logarithmic form of Eq. [5] is fitted to the data in Figs. 1a and 1b for HM and LM latex films and the results are plotted in Figs. 7 and 8 respectively for the films with 3.55- and 4.05-µm thicknesses. The slopes of the linear relations in Figs. 7 and 8 produced the packing coefficients, α, for HM and LM films, respectively, which are listed in Table 1 and plotted versus film thickness in Fig. 9. As seen in Fig. 9, α values increase in both HM and LM samples as the film thickness is increased and saturated above a certain latex content. In Fig. 9, linear relations

FIG. 4. Logarithmic plots of the data in Fig. 3a versus film thickness, d, for (a) below and (b) above 210◦ C. Numbers on each curve indicate the annealing temperature. Data are fitted to Eq. [3] to produce κ values.

behavior can be explained by knowing that the larger polymer chains need higher energy than the smaller chains to move across the junction border. HM film reaches the same transparency as the above LM film when it is annealed at 230◦ C. In order to quantify the behavior of Itr in Figs. 1a and 1b an empirical equation is offered as follows Itr = AeαT ,

[5]

where α is called the packing coefficient, T is the annealing temperature, and A is the Itr intensity at T = 0. Since α is dependent

FIG. 5. Logarithmic plots of the data in Fig. 3b versus film thickness, d, for (a) below and (b) above 180◦ C. Numbers on each curve indicate the annealing temperature. Data are fitted to Eq. [3] to produce κ values.

PACKING DURING LATEX FILM FORMATION

77

FIG. 6. Plot of absorption coefficient (κ) versus annealing temperature for HM and LM films.

obey

FIG. 8. Logarithmic plots of the data in Fig. 1b for 4.05-µm thick LM film. Data are fitted to Eq. [5] to produce α values.

α = αo d,

[6]

where αo is the slope of the curves which are obtained as 3.6 × 10−3 and 6.1 × 10−3 (µm ◦ C)−1 for the HM and LM films, respectively. Here αo indicates the degree of packing which states that LM samples packed much easier than HM samples by placing the same amount of latex particles on the surface of the glass substrate. Relatively high packing coefficients of LM samples compared to HM may be explained

by the shorter chains of LM samples, which may cause easier void–closure processes due to the viscous flow of the polymeric material in the sample. Complete wetting in LM samples can be reached much easier than in HM samples in the same amount of latex packing which then results in the quicker interdiffusion of chains across the polymer–polymer interface in LM samples. Quite recent work by Wilkinson et al. (21) showed that as surface loading is increased in silica gel samples

FIG. 7. Logarithmic plots of the data in Fig. 1a for 3.55-µm thick HM film. Data are fitted to Eq. [5] to produce α values.

FIG. 9. Plot of packing coefficient (α) versus film thickness for HM and LM films.

78

ARDA AND PEKCAN

emission intensity of anthracene decreased, which may indicate an increase in transmitted light intensity. In conclusion, this paper presented the effect of molecular weight on packing during latex film formation. Here, by using a photon transmission technique, it is observed that as the latex content is increased latex particles form more transparent, void free films by annealing, and it can be stated that low molecular weight films packed much easier than high molecular weight films. REFERENCES 1. Eckersley, S. T., and Rudin, A., J. Coatings Technol. 62, No.780, 89 (1990). 2. Joanicot, M., Wong, K., Maquet, J., Chevalier, Y., Pichot, C., Graillat, C., Linder, P., Rios, L., and Cabane, B., Prog. Colloid Polym. Sci. 81, 175 (1990). 3. Sperry, P. R., Snyder, B. S., O’Dowd, M. L., and Lesko, P. M., Langmuir 10, 2619 (1994). 4. Mackenzie, J. K., and Shutlewort, R., Proc. Phys. Soc. 62, 838 (1949). 5. Hahn, K., Ley, G., Schuller, H., and Oberthur, R., Colloid Polym. Sci. 66, 631 (1988). 6. Yoo, J. N., Sperling, L. H., Glinka, C. J., and Klein, A., Macromolecules 23, 3962 (1990).

7. Kim, K. D., Sperling, L. H., and Klein, A., Macromolecules 26, 4624 (1993). ¨ Winnik, M. A., and Croucher, M. D., Macromolecules 23, 2673 8. Pekcan, O., (1990). 9. Zhao, C. L., Wang, Y., Hruska, Z., and Winnik, M. A., Macromolecules 23, 4082 (1990). 10. Wang, Y., Zhao, C. L., and Winnik, M. A., J. Chem. Phys. 95, 2143 (1991). 11. Boczar, E. M., Dionne, B. C., Fu, Z., Kirk, A. B., Lesko, P. M., and Koller, A. D., Macromolecules 26, 5772 (1993). 12. Mazur, S., in “Polymer Powder Processing” (N. Rosenweig, Ed.). Wiley, New York, 1996. ¨ and Canpolat, M., J. Appl. Polym. Sci. 59, 277 (1996). 13. Pekcan, O., ¨ Polymer 36, 4433 (1995). 14. Canpolat, M., and Pekcan, O., ¨ Polymer 36, 2025 (1995). 15. Canpolat, M., and Pekcan, O., ¨ Kemero˘glu, F., and Arda, E., Eur. Polym. J. 34, 1371 16. Pekcan, O., (1998). ¨ and Arda, E., Polym. Int. 47, 231 (1998). 17. Pekcan, O., ¨ Arda, E., Kesenci, K., and Pi¸skin, E., J. Appl. Polym. Sci. 68, 18. Pekcan, O., 1257 (1998). ¨ and Arda, E., J. Appl. Polym. Sci. 70, 339 (1998). 19. Pekcan, O., ¨ Winnik, M. A., and Egan, L., Macromolecules 16, 702 20. Pekcan, O., (1983). 21. Wilkinson, F., Worrall, D. R., and Williams, S. L., J. Phys. Chem. 99, 6689 (1995).