Formulation effects on the distribution of pigment

Progress in Organic Coatings 48 (2003) 140–152

Formulation effects on the distribution of pigment particles in paints F. Tiarks a,∗ , T. Frechen a , S. Kirsch b , J. Leuninger a , M. Melan b , A. Pfau a , F. Richter c , B. Schuler a , C.-L. Zhao d b

a BASF Aktiengesellschaft, Emulsion Polymers, 67056 Ludwigshafen, Germany BASF Aktiengesellschaft, Product Development Adhesives and Construction, 67056 Ludwigshafen, Germany c BASF Aktiengesellschaft, New Business Development, 67056 Ludwigshafen, Germany d BASF Corporation, Charlotte Technical Center, 11501 Steele Creek Road, Charlotte, NC 28273, USA

Abstract Modern water-borne paints are widely used in different areas of applications ranging from high-gloss lacquers to flat, scrub-resistant interior paints. From this point of view, the pigment volume concentration (PVC) is one key-parameter adjusting the desired application properties. In high-gloss paints, for example, a low PVC is required to accommodate the proper surface roughness to achieve a high gloss. Consequently, a high concentration of TiO2 is needed to obtain a good hiding power at the same time. Flat paints nonetheless are highly filled due to cost reasons preferentially by CaCO3 and the pigment binding capacity of the binder is crucial. In this work, paint formulations differing in PVC, and the type of binder or dispersing agent were investigated by various techniques concerning the distribution and aggregation of pigment particles, e.g. TiO2 . To get a detailed insight into the structure of the liquid paints and the corresponding dried paint films, suitable analytical tools were applied for characterization. The structure of the liquid paints was analyzed by remission light spectroscopy (RLS), disc centrifugation, cryogenic-replication transmission electron microscopy (Cryo-replica TEM) and cryogenic-scanning electron microscopy (Cryo-SEM). The pigment distribution in the corresponding dried paint films was examined by means of atomic force microscopy (AFM), TEM and RLS. The tendency of the TiO2 -pigments to form aggregates was found to depend on both: first on the type of binder used in the formulation and second on the employed dispersing agent. It is shown that only by adjusting the properties of the binder in combination with common dispersants, it is possible to get well-distributed TiO2 particles within the paint. Correlation of application properties, e.g. gloss and blocking to the microscopic structure is presented. © 2003 Elsevier B.V. All rights reserved. Keywords: Water-borne paint; Pigment volume concentration; Binder; Film properties; Gloss; Dispersing agent; Dispersion

1. Introduction The composition of modern water-borne architectural coatings strongly depends on the desired application properties and therefore on the pigment volume concentration (PVC). In general, paints consist of water, a polymeric binder, pigments and filler particles. Additionally, additiveslike coalescents, thickeners, dispersants and defoamers are added to enable a sufficient stability and good application properties (Fig. 1). As already mentioned, the PVC is one key-parameter for describing paints. The higher the PVC, the lower is the content of polymeric binder within the paint and the higher is the portion of pigment and filler particles. The PVC thus strongly determines application properties such as gloss, scrub resistance, tensile strength, etc. (Fig. 2). Besides the ∗ Corresponding author. Tel.: +49-621-60-43208; fax: +49-621-60-40855. E-mail address: franca.tiarks@basf-ag.de (F. Tiarks).

0300-9440/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0300-9440(03)00095-X

PVC, one would expect that the pigment distribution in a paint is as critical and influences the properties: gloss and hiding power at low PVC as well as scrub resistance at high PVC. Whereas in flat paints, the high PVC is reached by using a combination of cheap filler material such as CaCO3 and small amounts of TiO2 , which is needed to achieve sufficient hiding power, one is limited to the use of TiO2 in low PVC paints. This is due to the fact that a high portion of filler would not contribute to the desired opacity but increases the surface roughness and hence reduces the gloss. In order to obtain maximum hiding power with a minimum amount of expensive TiO2 , a homogeneous distribution of the pigments is essential and one of the main requirements in formulating water-borne coatings. The pigment distribution is determined by the colloidal interaction of the TiO2 particles, which tend to aggregate and form clusters in the paint. These clusters can result in a higher surface roughness of the film as well as a decreased hiding power due to reduced light scattering efficiency.

F. Tiarks et al. / Progress in Organic Coatings 48 (2003) 140–152

141

Lacquer paints

Indoor paints

Low PCV (20%)

High PVC (85%)

11% Water 53% Dispersion

9% TiO2

2% Additives

40% CaCO3

3% Additives 23% TiO2

27% Water 5% Extender 12% Thickener

16% Dispersion

Fig. 1. Typical composition of a low PVC high-gloss lacquer compared to a high PVC interior flat paint.

The influence of the surface roughness on the gloss of a paint film is obvious and has been described in [1]. For gloss paints, a much lower PVC must be applied to prevent particles disrupting the surface and hence reducing the gloss of the film. PVCs below 20% are for that reason typical for gloss paints. At this concentration, the scattering efficiency of the pigment is crucial to achieve a fully opaque coating. TiO2 is the favored pigment for producing opacity in paint films due to its exceptionally high refractive index (nD = 2.5–2.7). Light scattering is maximized when the pigment particle size ranges between 200 and 250 nm (Fig. 3). However, if the pigment particles form clusters, the scattering efficiency decreases because each cluster scatters approximately like a single, larger particle. Accordingly, the scattering efficiency strongly depends on the dispergation state of the titanium dioxide particles within the paint film. Clusters may be produced because of poor dispersion techniques. In the process of paint formulation/production, it is important to breakdown primary structures like agglomerates into fine particles and to stabilize this state [2]. However, it has been shown that most clumps are dispersed in a short time by the milling process [3]. Due to this, pigment clustering in emulsion paints must involve other factors [4].

Properties

Gloss

Scrub resistance Tensile strength

Hiding

0

10

20

30

40

50

60

70

80

PVC Fig. 2. Dependence of characteristic application properties on the PVC of a paint formulation.

Pigment clustering is more pronounced in water-borne gloss paints than in solvent-borne paints [5]. This is due to the necessary step of pigment-wetting that is generally quite easy in solvent-borne systems because of the low surface tension of ordinary solvents and resins. Stabilization of pigment particles in aqueous systems functions by a different mechanism. While in solvent-borne systems, a sterical barrier prevents re-agglomeration, water-borne systems require the ionic mechanism of electrostatic repulsion, which is mainly solved by the usage of dispersants and prevents pigment settling [6]. Stabilizing the well-dispersed state and preventing re-agglomeration is the designated task of the dispersing agent [7]. To conclude: uneven distribution and clustering of pigment particles have a detrimental effect on application properties such as gloss and hiding power. To influence the pigment distribution, several parameters namely the nature of the latex, type and amount of dispersants, etc. can be varied. In this work, we elucidate the influence of the binder and the dispersing agent onto the pigment distribution and identified appropriate analytical tools for characterization of the dispergation state. A further crucial question to be answered is, whether the pigment distribution in the final film is predetermined in the liquid paint or whether additional agglomeration might occur during film-formation and loss of water. The following samples were formulated and examined Table 1. The main difference between paste 1 and pastes 2–4 is the higher PVC (40% versus 18%) and the used dispersant.

Table 1 Survey of the different formulations to examine the influence of either binder or dispersing agent onto the pigment distributiona

Paste Paste Paste Paste a

1 2 3 4

Binder 1

Binder 2

P1

P2

For details see Appendix A.

Binder 3 P3 P4 P5


size: 300 nm

size: 250 nm

400

45 40 35 30 25 20 15 10 5 0

450 500 550

500 450

600

400 350

650

Wavelength [nm]

scattering coefficient Ks [1/µm]

142

300

700

250 200

750

150 100

800

50

particle size [nm]

10

Fig. 3. Dependence of the scattering efficiency at different wavelengths on the particle size of TiO2 employing the scattering factor according to Mie-theory [17].

Pastes 2–4 differ only in the type of dispersing agent and are optimized for high-gloss formulations. Binders 1 and 2 are straight acrylics and differ in their comonomer system. Binder 3 is a styrene–acrylic specifically developed for high-gloss paints. Concerning the different dispersants employed, paste 1 contains pigment dispersant A® , a low molecular polyacrylic acid in the form of the ammonium salt. Paste 2 comprises pigment dispersant MD 20® , a copolymer of diisobutene and maleic anhydride and paste 3 contains pigment dispersant Collacral LR 8954® , a solution polymer consisting of an acrylate and acrylic acid ammonium salt. The fourth dispersant Orotan 681® consists of an acrylate–methacrylic acid copolymer and is added to paste 4. Overall, the dispersants differ in their hydrophilicity and functional groups. For details about the formulations, see Appendix A. 2. Experimental 2.1. Synthesis of the latexes The latexes were prepared by a standard, seeded, semibatch emulsion polymerization process employing anionic and/or non-ionic emulsifiers and sodium persulfate as initiator. The polymerization was carried out at 85 ◦ C according to standard procedures [8]. 2.2. Paint formulation and sample preparation The binders mentioned above were formulated in pigmented, interior gloss paint formulations with small amounts

of coalescent in the standard recipe. The PVC was set to ca. 18%, except P1 and P2 (40%). Commercial TiO2 was used as white pigment (Kemira 675® , diameter of 0.2–0.25 ␮m) and different types and amounts of dispersing agents were employed. As paints P3–5 were supposed to be of high-gloss quality, no CaCO3 was used as filler material. The binder was added in a final step to the pigment slurry. For a detailed paint formulation recipe, see Appendix A. 2.3. Electron microscopy To gain information about particle size and distribution in the wet paint, TEM was used. In principle, with electron microscopy as a vacuum technique, only dried specimen can be investigated. For dispersions and suspensions, Cryo-techniques are widely applied for preparation and imaging [9]. 2.3.1. High-pressure freezing of aqueous latex paints and high resolution cryogenic-scanning electron microscopy (Cryo-SEM) A droplet of the paint is quickly frozen by pressurized liquid nitrogen in a high-pressure freezing machine at approximately 2100 bar. The frozen sample is then broken off with a knife cooled in liquid nitrogen. Afterwards, the sample is transferred into a high-vacuum planar magnetron-sputtering device for controlled etching by sublimation and coating with conducting platinum. The specimen is sublimed for 20 min at 178 K in vacuum better than 10−6 mbar to etch ice from the fracture sample surface. In a following step, the fracture surface is sputter-coated with approximately 2–3 nm of platinum in a 0.02 mbar argon atmosphere.


The metal-coated surface is then examined under an in-lens Hitachi S-900 FESEM (Nissei Sabgyo America Ltd., Rolling Meadows, IL). The sample is inserted inside the pole pieces of a highly excited objective lens so that it is possible to have 0.7 nm secondary electron image resolution operating at 30 kV operating voltage. During the SEM examination, the sample temperature is kept at 143–163 K. A more detailed sample preparation and examination procedure is described elsewhere [10]. Due to the analysis via back-scattering (compositional contrast) and secondary electron imaging (topological contrast), the electron-rich TiO2 -particles appear as white dots in the micrographs and a topological contrast is superimposed. 2.3.2. Cryogenic-replication transmission electron microscopy (Cryo-replica TEM) For particle sizes down to about 10 nm, we used a Cryo-replication technique for preparation in combination with conventional TEM for imaging (Cryo-replica TEM). The Cryo-replication TEM equipment consists of a rapid freezing system, a specimen transfer holder, a vacuum cryo-replication unit (Balzers BAF 400 D) with the possibilities for freeze-fracturing, freeze etching and two evaporation sources and a conventional TEM, in our case a Zeiss EM 902 which is operated at 80 kV. In the most simple case, this rapid freezing systems is a dewar vessel filled with a cryogen, e.g. supercooled nitrogen so-called slush-nitrogen, 63 K, cooling rate some 1000 K/s or liquid ethane or propane with a higher heat capacity and cooling rate up to 15,000 K/s. A small droplet of the dispersion or paint formulation is shock-frozen in the cryogen and transferred into the Cryo-replication unit. After freeze-fracturing to get a fresh and clean surface, the specimen is freeze-etched by ice-sublimation at elevated temperature (173 K), leaving the solid particles at the surface. In the next step, this profiled surface is coated by oblique evaporation of Pt/C (3–5 nm). As a result of the shadowing effect, the thickness of the Pt/C layer varies according to the topography, thus containing the whole information about particle size and shape and thereby giving rise to a contrast in the TEM. For better mechanical stability, a 20 nm thick carbon coating is evaporated additionally. The specimen is then thawed at room temperature and the replica is cleaned from adhering dispersion material. The replica is transferred into the TEM and imaged. The interpretation of the TEM microgaphs is relatively simple, as the TiO2 is still present in the sample and of high contrast (black) due to its high electron density. Yet the binder is removed due to the sample preparation and the Pt/C that was deposited on the sample under a certain angle induces the topology contrast. These areas are therefore of relatively low contrast and appear as gray to white regions.

143

2.4. Atomic force microscopy (AFM) The AFM data presented here were recorded at ambient air in tapping mode with a Nanoscope dimension 3000 SPM (digital instruments) capable of recording phase images and using Si-cantilevers (35 N/m, ν0 approximately 300 kHz, Nanoprobe). The tapping mode is a dynamic measurement mode in which the cantilever performs a driven oscillation at or near its resonance frequency with amplitudes typically between 20 and 100 nm [11]. After approaching the sample, the cantilever comes into contact with the sample surface during its vibration. Therefore, the amplitude of its vibration is reduced. In the tapping mode, the effect is used to map the surface by moving the cantilever in the z-direction while rastering the surface to keep the damping constant. Recording the necessary z-movement of the piezo-actuator during rastering yields topography. Simultaneously, in phase-imaging, the phase shift between driving force for the lever vibration i.e. the driving voltage for the piezo that vibrates the lever, and the optically detected motion of the lever is recorded while rastering the sample. As this phase shift depends on the tip-sample contact and hence on the viscoelastic and the adhesive properties of the material present at the surface or in other words, on the energy dissipation into the sample, it can be used to map distributions of different materials [12]. Phase-imaging gives images with so-called “material contrast” in which patches, for example soft and hard regions—as, for example, pigment and binder—can be identified, which are routinely correlated to the height pictures [13]. 2.5. Remission light spectroscopy (RLS) The analysis of remission spectra enables to calculate the pigment distribution within dried paint films or liquid paint formulations. By assuming that in a white, TiO2 -pigmented coating, the particular remission values of the spectrum are proportional to the number of pigment particles within a distinct particle size range, the spectrum mirrors a distorted image of the particle size distribution within the matrix [14]. A probe for measuring the remitted light of a sample is in principle a bundle of glass fiber filaments, which is immersed into the sample. In case of these measurements, this bundle consists of 14 filaments in a concentric array, each fiber with a diameter of 200 ␮m. Seven of these filaments are connected to a xenon lamp (emits higher portion of blue light than a halogen lamp) to radiate light into the sample. The other filaments are connected to a diode-array spectrometer (wavelength range 200–1000 nm). For better mechanical stabilization and protection against aggressive media, the glass fibers are fitted within a steel tube that is capped by a sapphire window. Light that radiates from the glass fibers into a dispersion will be scattered by the dispersion particles in all directions.

144


Back-scattered, that means remitted light, gets with a distinct probability in the seven receiver filaments and will be analyzed in the spectrometer. The spectrum of the remitted light is dependent on the particle size (including the particle distribution in space), concentration and optical constants (such as refractive index and adsorption constant) of the scattering particles (see Fig. 3). For diluted systems, this correlation is exactly described by the Mie-theory [17]. Even in concentrated turbid samples there exists a correlation between the remission spectrum and the listed parameters, however because of the missing quantitative description, the measurement is only useful for qualitative comparison of different samples. The optical remission probe was dipped into the undiluted liquid paint formulations and the spectral remission, i.e. the intensity of the back-scattered light was normalized to a white-standard.

The gloss measurements were performed on a Haze-Plus (Byk/Gardner; Geretsried, DIN 67530) under an observation angle of 20◦ and 60◦ and the average of three measurements was taken.

3. Results and discussion The effect of (i) binder and (ii) dispersing agent on the distribution of TiO2 -pigments was investigated. The latter was studied on both the liquid paint formulation and the dried paint film. Via Cryo-replica TEM and Cryo-SEM, it was possible to characterize the pigment distribution in the wet paint formulation. The samples of the dried films were investigated by means of AFM and RLS. By comparing the results gained from analysis of the wet and dried state, it should be possible to get an insight into the effect of film-formation on the clustering of TiO2 .

2.6. Disc centrifugation Employing disc centrifugation, the hard sphere diameter of pigment pastes and corresponding paint formulations can be determined. The formulations and pastes were diluted to a pigment concentration of 0.025% and conditioned by ultrasound for 30 s. The particle size range covered by the disc centrifuge DCF A ranges from 50 to 1000 nm and the centrifuge is operated at 3000 rpm. ZnNO3 was used as buffer and an ethanol: water solution (1:19) employed as sedimentation liquid. 2.7. Measurement of gloss Gloss of the paint film was measured as follows: the wet paint was doctor bladed with a thickness of 240 ␮m (wet) onto a freshly cleaned glass plate and dried for 3 h at ambient temperature resulting in a film thickness of about 50 ␮m dry. Subsequently, the plates were dried at 50 ◦ C in an oven for 72 h to remove all solvents and residual coalescents.

3.1. Distribution of titanium dioxide pigments in the liquid state The Cryo-replica technique enables to measure directly in the undiluted paint or pigment paste. As a reference, the structure of a pigment paste with a PVC of 40% without addition of a polymeric binder was investigated. In Fig. 4, it is shown that in the pigment paste P1 already large aggregates of TiO2 are formed. The dilution of the sample by different amounts of water affects the pigment distribution (Fig. 4b and c). With increased dilution, the size of TiO2 agglomerates is decreased. However, it is clear from the image given in Fig. 4c, that even at high dilution clusters are formed. In order to further separate the clusters into single TiO2 particles, additional interactions have to be introduced. By formulating a paint from a paste, both the dilution effect and additionally specific interactions are introduced by the binder. Thus, a good dispersion should not only be able to break-up the pigment agglomerates by quasi-diluting

Fig. 4. Influence of dilution onto the pigment distribution within a TiO2 -paste examined by Cryo-replica TEM: (a) undiluted pigment paste; (b) dilution 1:20; (c) dilution 1:50. The reversible agglomerates seen in the undiluted paste are broken off by dilution with water.


but additionally stabilize the small clusters by electrostatic forces against re-agglomeration in the paint and meanwhile film-formation. In addition to these results, it is apparent that fast and easy analytical tools like dynamic light scattering that generally rely on diluting a sample to very low concentrations are restricted in their information, as the dilution effect is obviously important in terms of clusters size. Therefore, it is a major task to find analytical tools, which are able to detect pigment clusters in the wet undiluted samples. In the next two sections, the effect of binder and dispersing agent onto the distribution of TiO2 is investigated. 3.1.1. Influence of the binder onto the pigment distribution in the liquid paint As we concentrate on the pigment distribution and formulation effects in gloss lacquers, PVC was not much varied in this paper. Pigment distribution at high PVC (>40%) is described elsewhere [15]. The focus of this work is • effect of different types of binder; • influence of additives (e.g. type of dispersing agent) used in the formulation. For the first topic, two straight acrylic binders were used and compared. Binders 1 and 2 differ only concerning their comonomer compositions. Fig. 5 displays results from Cryo-electron microscopy. The formulated paints at PVC 40% are named P1 and P2, respectively. TiO2 is due to its high electron density of high contrast (black). The binder is removed by the sample preparation and the Pt/C, that was deposited on the sample under a certain angle induces the topology contrast. These areas are therefore of relatively low contrast and appear as gray to white regions. The replica of individual latex particles is clearly detectable indicating that no film-formation occurred. As is apparent from this preliminary testing, the nature of the binder has an influence on the pigment distribu-

145

tion. Whereas P1 shows only small and well-distributed TiO2 -clusters, P2 gives much bigger aggregates in the paint. Apparently, the differences in the comonomers of the binder are essential. Similar results were obtained by Zhao et al. analyzing the effect of functional groups onto the dispergation state [16]. The second conclusion drawn from this preliminary testing is that the dispersing agent alone is not able to provide complete separation of all TiO2 -pigment particles. As an effect of water-soluble oligomers on the pigment distribution is unlikely, as already being shown by [15], it is clearly the latex-particle itself, which has a strong interaction with the surface of the pigment resulting in such big differences in pigment distribution. After confirming that the employed binder influences the dispergation state, the effect of different dispersants was examined. As a combination of functional comonomers seemed to be promising for good pigment dispergation, a similar latex system to binder 1 but optimized for high-gloss applications was employed in the following testing (binder 3). 3.1.2. Effect of dispersing agent on the pigment distribution The binder optimized for high-gloss applications used in this testing was kept constant and the type of dispersant was varied. For details of the paint formulations P3–5, see Appendix A. Fig. 6 displays the results by Cryo-SEM (upper row) and Cryo-replica TEM (bottom line). The difference in contrast (TiO2 as white or black dots) is due to the different imaging technique. In the SEM, the electron-rich TiO2 -particles appear as white dots in the micrographs due to the analysis via imaging the back-scattered electrons. On the other hand in the TEM microgaphs, TiO2 is of high contrast (black) due to its high electron density. For these samples with low PVC, sophisticated Cryo-techniques have to be used. Whereas in flat paints, the large

Fig. 5. Use of different dispersions as binders in paint formulations and their influence on the pigment distribution analyzed by Cryo-replica TEM: (a) binder 1 leading to a fairly good TiO2 dispergation; (b) binder 2 resulting in highly agglomerated pigment particles.

146


Fig. 6. Cryo-replica TEM and Cryo-SEM micrographs of formulations possessing the same type of binder but different dispersing agents: (a) dispersing agent MD 20® (P3); (b) CollacraLR 8954® (P4); (c) Orotan 681® (P5). The use of different dispersants influences the dispergation state of titanium dioxide.

CaCO3 particles would behave like spacers and thus can hinder structural changes caused by water crystallization. These lowly pigmented paints are more sensitive towards ice-crystals. The small TiO2 particles could easily be pushed around and pressed together giving a completely different appearance of the dispergation state. Therefore, ice-crystallization has to be avoided and is realized by shock-freezing. As a highly reliable method to avoid ice-crystal formation even by shock-freezing, the high-pressure technique is well known [10]. As this technique is even more complex to the plunge-in cryo-technique, we were curious to know, whether both techniques give the same results concerning the pigment distribution. As is maintained from the Cryo-replica TEM as well as from the Cryo-SEM measurements, the three samples exhibit differences in the dispergation state of the titanium dioxide. Whereas the dispergation by use of pigment dispersant MD 20® in P3 is not perfect but relatively homogeneous and with small clusters, the use of Orotan 681® in P5 leads to pronounced clustering of the pigment particles. Assuming a better interaction with the pigment of the maleic acid groups in pigment dispersant MD 20® compared to acrylic acid in Orotan 681® , the effect of a better pigment distribution is understandable. This positive interaction seems to enable the de-agglomeration of the pigment clusters.

The Cryo-micrographs of P4, especially the SEM, display relatively small agglomerates but compared to the other samples the number of detected clusters is too small, as all samples possess the same PVC. It has to be assumed that large pigment clusters should be observed somewhere else in the sample, which were not visualized in this image. However, no distinctive difference between the Cryo-SEM and the Cryo-replica TEM technique regarding the aggregation behavior of TiO2 is observed. This indicates that ice-crystal formation via plunge-in freezing or high-pressure freezing is comparable and has no influence on the dispergation state visualized by the TEM and SEM techniques. The imaging techniques by electron microscopy are obviously a nice tool for getting an impression of the pigment distribution in the liquid paint. However, these techniques are on one hand rather complicated and cost and time-consuming and on the other hand give only images of small volumes. These are not necessarily representative for the whole sample and therefore possibly misleading. Hence, it is important to have an integral testing method which allows the measurement of larger volumes. Consequently the paints are compared by disc centrifugation and RLS. Employing disc centrifugation, the hard sphere diameter of the pigment pastes as well as the corresponding paint formulations were determined.


147

Paste 4

7

6

frequency

5

4

Paint P 4 3

2

Paint P 3 Paste 3

1

0 0

100

200

300

400

500

600

700

800

900

1000

hard sphere equivalent diameter [nm]

Fig. 7. Results of disc centrifugation of the diluted pigment pastes and corresponding paint samples P3 and P4 possessing different dispersants (MD 20® and Collacral LR 8954® , respectively). The shift of the curve of P4 to larger hard sphere equivalent diameters indicates the formation of clusters.

The formulations and pastes were diluted to a pigment concentration of 0.025% and conditioned by ultrasound for 30 s (Fig. 7). Pastes 3 and 4 and the paint P3 exhibit an almost identical particle size distribution (about 270 nm) that comes close to the primary grain size of the pigment particles (0.25–0.27 ␮m). Yet paint P4 exhibits a reproducible and thereby significant shift to larger diameters (ca. 310 nm). This shift can only be explained by partly agglomerated titanium dioxide particles. Interestingly, the agglomeration process does not occur in the paste, which already contains the dispersant, but happens after adding the binder. This effect clearly demonstrates that both dispersant and binder have a strong influence on the dispergation state of TiO2 . These findings support the assumption of better specific interactions of functional groups in P3 with the pigments as stated above. The result corroborates the hypothesis that P4 contains larger TiO2 clusters that were not detected by imaging techniques due to the small picture area. However, the disc centrifugation is based on measurements of diluted samples. As described previously, the dilution affects the pigment agglomerates as shown by Cryo-replica TEM (Fig. 4). Therefore, we have to backup this assertion by a second method. This second non-imaging technique is RLS. Advantageously, RLS works in concentrated opaque samples and can be employed without diluting the paint formulations. The qualitative extent of different particle sizes onto the remission spectrum can be drawn from the scattering factor according to Mie-theory [17]. The wavelength of monochromatic light and the particle size leading to optimal back-scattering of this light depend on each other (Fig. 3).

Thus for a given particle size distribution, a large extent of small particles or clusters leads to a high remission in the short-wave region and a high extent of larger particle sizes or clusters yields to pronounced remission in the long-wave region. The obtained remission spectrum is a distorted image of the original grain size distribution of the pigments in the paint formulation. Concerning the results of the RLS (Fig. 8), one has to keep in mind that in a pigment suspension, the back-scattering decreases with an increasing pigment concentration. Paints P3 and P4 exhibit the same PVC and consequently their remission behavior can be compared. P4 displays more remission in the long-wave region of the spectrum than in the short-wave, as compared to P3. This effect is small but reproducible and is a hint of a higher content of agglomerated particles in the undiluted paint, which is in agreement with the results of disc centrifugation within the diluted state. As both methods give only qualitative information, no answer to the question about the size and the volume fraction of the agglomerates can yet be provided. As can be concluded from all data presented here, it is the PVC of a formulation and the employed binder that influences the dispergation state of the pigment particles. Furthermore, the type of dispersing agent is critical for the aggregation behavior and dispergation of the pigment particles and therefore has to be chosen carefully. Using sophisticated imaging techniques as Cryo-SEM or Cryo-replica TEM and additionally relatively easy methods as RLS and disc centrifugation, a broad insight into the undiluted liquid state of a paint formulation is accessible. The question arises if the clustering seen in the liquid state is

148

F. Tiarks et al. / Progress in Organic Coatings 48 (2003) 140–152 450

P4 400

rel. Remission [%]

350

P3

300 250 200 150 100 50 0 300

350

400

450

500

550

600

650

700

750

800

Wavelength [nm] Fig. 8. RLS on paints P3 and P4 in the undiluted state. P4 displays more remission in the long-wave region of the spectrum, which can be seen as a hint for weakly agglomerated particles.

also found in the dried paint film and if so, how does the dispergation state influence the macroscopic paint properties. 3.2. Pigment distribution in the dried paint film Coming from the fluid paint formulation to the dried paint film, suitable characterization methods determining the pigment distribution have to be evaluated. One of them is RLS already proven to be useful in the liquid state. Paint films of P3 and P4 were measured with this technique. These films were obtained by doctor bladeing the

paints (thickness of 50 ␮m wet on acetate foil) and then drying at ambient temperature. The remission was measured by using an Ulbricht-probe. Regarding Fig. 9, one recognizes that the shape of the remission spectra of the paint films are much different than for the liquid paint. Whereas in the wet paint, an increase in remission in the long-wave region is observed, this behavior is not detected in the spectrum of the films. This can be explained by the fact, that in the liquid paint film the measurement only reflects the pigment distribution of the film surface, whereas in the paint the bulk is

90

Film P 3

80

rel. Remission [%]

70

Film P 4

60 50 40 30 20 10 0 300

400

500

600

700

800

900

1000

Wavelength [nm] Fig. 9. Analysis of the dried paint films of P3 and P4 by RLS. The stronger remission of P3 in the short-wave region compared to P4 assumes the existence of a larger amount of smaller particle clusters.


149

Fig. 10. AFM of the paint films in the amplitude mode (20 ␮m × 20 ␮m): (a) P3; (b) P4; (c) P5. The differences of the surface smoothness affected by TiO2 -clustering are obvious.

detected. Thus, it is assumed that during film-formation, the larger aggregates tend to settle down and hence only smaller clusters or isolated TiO2 -pigments stay at the surface. This would explain the overall lower remission in the long-wave region of both paint films P3 and P4 versus their liquid state. In the short-wave region, the remission of P3 is higher than that of P4, indicating a higher amount of small clusters or single titanium dioxide particles at the surface. However, this finding has to be proven by an imaging technique. One well-established method for surface analysis is AFM. In the following, this technique has been employed to study the surface properties of P3 and P4 (Fig. 10a and b). One clearly sees none of the paint films possesses a completely smooth surface but interestingly strong differences in the pigment clustering. Whereas P3 displays at least pigment particles that are relatively homogeneously distributed on the film surface, the analysis of P5 reflects the inhomogeneity as already revealed

by Cryo-techniques in the wet paint. One also finds clustered particles that lead to a more disturbed surface of the paint film. The surface analysis by AFM of a film of P4 displays only a small amount of pigment particles. One possible explanation for this result could be the already mentioned tendency for aggregation. This predisposition leads to a pronounced segregation process meanwhile the film drying process. The pigment particles gather at the bottom side of the paint film due to their higher density whereas the less dense binder particles form a polymer film on top. It can be concluded that the surface analysis by AFM fits very well the data obtained by RLS. Another result is that AFM measurement also mirrors the image obtained by Cryo-techniques. Sample P5 contains strongly agglomerated TiO2 clusters due to insufficient stabilization of the titanium dioxide. The large clusters of P4 observed by RLS and disc centrifugation have not necessarily to be detected by imaging techniques.

Fig. 11. TEM of cross-cuts of the dried films of P3 (a) and P4 (b). The size of holes created by the cutting reflects the original size of pigment clusters in the polymeric matrix.

150


Whereas AFM analysis gives an impression of the surface of the paint samples, cross-cuts of dried paint films and subsequent analysis by TEM enables a view into the bulk structure (Fig. 11). The cross-sections of the film reveal the diversity of the particle distribution. Whereas P3 contains relatively well-distributed pigment particles, the cross-section of P4 displays larger clusters or holes that resulted from the cutting process by pigment clusters that fell out of the polymeric matrix. This finding funds the assumption that larger TiO2 clusters are existing in P4. However, from the image it is getting quite obvious that these large clusters have not necessarily to be located at the interface to air. Due to this, surface analysis techniques alone are not sufficient to get all information about the dispergation state of pigments. After proving the hypothesis that both binder and dispersing agent strongly affect the clustering of TiO2 in paint formulations (neglecting dilution effects) and after demonstrating that the agglomeration in the liquid state is also seen in the dried paint film, it remains to ascertain if these microscopic findings correlate to macroscopic paint properties. 3.3. Correlation to macroscopic paint properties The correspondence to macroscopic paint properties is gained by correlating the evaluated dispergation state and the gloss measured under an observation angle of 20◦ and 60◦ (see Table 2). For high-gloss applications, the level of the 20◦ gloss is more predicative than the 60◦ gloss. By comparing the paints P3 and P4, the manifested difference within the dispergation state is reflected by a lower gloss of P4. The surface roughness of P3 and P4 were determined by AFM, showing that P3 has a much higher roughness than P4. As P3 has a higher gloss, obviously the surface roughness is not the only parameter being responsible for this application property. It is in addition the

Table 2 Application properties of the presented high-gloss formulations: gloss values (observation angle 20◦ and 60◦ ) and haze, blocking after 1 day drying at room temperature and stressing with 10 kg for 24 ha Sample

Gloss 20◦

Gloss 60◦

Haze

Blocking

Defects

P3 P4 P5

62 43 47

84 76 80

81 191 124

1 4 2

0 4 0

a Defects means film defects noticed after testing the blocking resistance. Marks for blocking and defects ranging from 0: excellent to 5: worse.

pigment distribution as a result of the high refractive index of TiO2 compared to the polymer. The haze of a film, which is often evoked by microstructures due to a bad dispergation of pigments, of P4 is much higher than that of a P3 film and might only partly be due to a larger surface roughness. Similar is valid for P5, as its worse agglomeration was established by Cryo-techniques and is seen again in the gloss and haze values. Another important paint property is blocking. Comparing the results of blocking of the high-gloss formulations, one has to note that P4 fails the blocking test and additionally shows defects in the paint film after the test. Compared to this, the blocking of P3 and P5 was without failure and can be attributed to the more homogeneous pigment particle distribution at the film surface. It can be concluded, that the dispergation state obviously strongly affects the macroscopic paint properties. Therefore, the binder and the dispersant of a high-gloss formulation should always be chosen carefully. 3.4. Controlling the dispergation state by tailored dispersants and binders By now we learnt how to analyze the state of dispergation of TiO2 either in a liquid paint formulation, e.g. by

Fig. 12. Cryo-SEM micrographs of (a) model latex SL0, and (b) the latex SL1 with a special surfactant. The difference in TiO2 distribution depending on the surfactant system is obvious.


151

Fig. 13. Effect of functional groups of the model latex onto the pigment distribution within a paint formulation measured by Cryo-SEM. Parts (a)–(c) resulted from latexes modified by comonomers with different functionalities.

Cryo-techniques and RLS or a dried paint film by AFM, TEM cross-cuts and RLS. In addition, we figured out which parameters affect the pigment distribution in a high-gloss formulation: mainly the binder and dispersing agent. The question arose, whether it would be possible to tailor the polymeric binder in order to overcome the aggregation phenomenon of pigments completely. Fig. 12 displays Cryo-SEM images of paints with tailored binders by using special surfactants differing in their functional groups. The distribution of TiO2 in Fig. 12a is relatively homogeneous but the pigment particles agglomerated and only pigment clusters are detected by Cryo-SEM. This means that the binder employed was not able to guarantee specific interactions needed for the break-up of pigment agglomerates. The system failed to stabilize smaller clusters or even single pigment particles. In contrast to this result, the addition of a special surfactant leads to the absence of pigment clusters. The TiO2 particles are finely divided within the paint and no agglomerates are detected. This diversity in pigment distribution is again reflected by macroscopic paint properties. The gloss of SL0 is higher than that of SL1 (see Table 3), supporting our findings in section 3.3.

Table 3 Gloss values of the presented formulations under an observation angle of 20◦ and 60◦ Sample

Gloss 20◦

Gloss 60◦

SL0 SL1

10 27

55 74

Apart from tailoring the binders by the use of special surfactants, the copolymerization with functionalized comonomers is another suitable tool to break-up pigment clusters. The obvious improvement caused by optimizing the binder is reflected in the Cryo-SEM micrographs in Fig. 13. The dispergation state of TiO2 as perfectly homogeneously distributed pigments throughout the whole sample can be achieved by choosing an appropriate comonomer. The dependence of the TiO2 distribution on the binder functionalization is clearly seen in this section. In the best case, the tailored binder completely eliminates TiO2 clusters, which is even possible without any change of formulation additives.

4. Conclusion To conclude the results presented in this work, one can state that the aggregation behavior of pigments in wet and dried paints displays parallels. Pigment agglomeration caused either by binder properties or the functionality or the dispersing agent chosen is already visible in the liquid state and subsequently found in the dried film. The influence of the employed dispersing agent is obvious and an effect of clustering on corresponding macroscopic parameters as e.g. gloss was established. The structure of paint films and liquid paints has been investigated by different techniques. The distribution of TiO2 particles in the paint was analyzed, for example, by Cryo-SEM, Cryo-replica TEM, RLS and disc centrifugation, respectively. By AFM and RLS, the pigment distribution on the surface of the dried paint films was examined. Cross-cuts allowed the visualization of pigment distribution in the bulk via TEM.

152


Acknowledgements

P3

The support of L.E. Scriven (Department Chemical Engineering and Materials Science, University of Minnesota) regarding the Cryo-SEM measurements is gratefully acknowledged. The authors would also like to thank V. Koch, T. Frey and A. Rennig (BASF AG, Performance Chemicals Physics) for their help concerning RLS and disc centrifugation.

Appendix A Paint formulations for use of different polymeric binders (PPW)

Water Natrosol 250 HR® Herkules GmbH Preservative Pigment disperser A® BASF AG AMP 90® BASF AG Sodium polyphosphate (10% solution) Defoamer TiO2 , Kronos 2043® Kemira Omyacarb 5 GU® Omya GmbH Coalescent Talcum AT 1® Omya GmbH Binder 1 (s.c. 48%) Binder 2 (s.c. 48%) Defoamer Collacral PU 85® BASF AG Water

P1

P2

180 2 2 1 1 3 3 190 180 20 50 320

180 2 2 1 1 3 3 190 180 20 50

3 5 40

320 3 5 40

Paint formulations for use of different dispersants (PPW) P3 Water Defoamer Preservative

96 3 2

Pigment disperser MD 20® BASF AG Collacral LR 8954® BASF AG Orotan 681® Rohm & Haas Collacral LR 8990® BASF AG TiO2 , Kemira 675® Kemira Defoamer Lusolvan FBH® BASF AG Collacral LR 8989® Binder 3 (s.c. 48%) Water

P4

P5

44 37 10 230 2 15 15 560 23

10 230 2 15 15 560 23

31 10 230 2 15 15 560 23

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10]

[11] [12] [13] [14] [15]

P4

P5

[16]

103 3 2

109 3 2

[17]

J.H. Braun, D.P. Fileds, J. Coat. Technol. 66 (1994) 93. J. Scröder, Farbe and Lack 91 (1995) 11. J. Balfour, D. Huchette, Paint Ink Int. 8 (1995) 52. S. Fitzwater, J.W. Hook, J. Coat. Technol. 57 (1985) 39. A. Brisson, A. Haber, J. Coat. Technol. 63 (1991) 59. F. Haselmeyer, J. Water-Borne Coat. 2 (1984) 2. M. Liphard, W. von Rybinski, L. Schieferstein, Farbe and Lack 97 (1991) 18. P.A. Lovell, M. El-Aasser (Eds.), Emulsion Polymerisation and Emulsion Polymers, Wiley, New York, 1997. R.A. Steinbrecht, K. Zierold (Eds.), Cryo-Techniques in Biological Electron Microscopy, Spinger, Berlin, 1987. E. Sutano, U. Dittrich, C.L. Zhao, H.T. Davis, L.E. Scriven, in: Proceedings of the 10th International Coating Science and Technology Symposium, 2000, p. 63. Q. Zhong, D. Innis, K. Kjoller, V.B. Ellings, Surf. Sci. L385 (1997) 3775. J. Tamayo, R. Garcia, Appl. Phys. Lett. 73 (1998) 2926. S.N. Magonov, V. Ellings, M.-H. Whangbo, Surf. Sci. L385 (1997) 375. J. Schmelzer, Farbe and Lack 107 (2001) 337. S. Kirsch, A. Pfau, T. Frechen, W. Schrof, P. Pföhler, D. Francke, Prog. Org. Coat. 43 (2001) 99. C.-L. Zhao, U. Dittrich, W. Heckmann, Y. Ma, V. Thiagarjan, E. Sutanto, L.E. Scriven, in: Proceedings of the 222nd ACS National Meeting, Division of Polymeric Materials: Science and Engineering, Chicago, 2001. G. Mie, Ann. Phys. 25 (1908) 377.