- Email: [email protected]
In situ USAXS measurements of titania colloidal paint films during the drying process
Journal of Colloid and Interface Science 336 (2009) 612–615
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
In situ USAXS measurements of titania colloidal paint films during the drying process Bridget Ingham a,b,*, Scott Dickie c, Hiroshi Nanjo d, Michael F. Toney b a
Industrial Research Limited, P.O. Box 31-310, Lower Hutt, New Zealand Stanford Synchrotron Radiation Lightsource, 2575 Sand Hill Road, Menlo Park, CA 94025, USA c Resene Paints Limited, P.O. Box 38-242, Lower Hutt, New Zealand d National Institute of Advanced Industrial Science and Technology, 4-2-1, Nigatake, Miyagino-ku, Sendai 983-8551, Japan b
a r t i c l e
i n f o
Article history: Received 3 February 2009 Accepted 10 April 2009 Available online 21 April 2009 Keywords: USAXS Small-angle X-ray scattering Synchrotron Titania Paint
a b s t r a c t Real-time ultra-small-angle X-ray scattering (USAXS) was used to follow the flocculation of titania particles in two paint systems as the paint films were drying. The inter-particle scattering was extracted by comparing the time series with diluted titania dispersion (having negligible inter-particle interaction). The paint system with pigment affinic groups showed considerably less flocculation than a pure acrylic emulsion. The results were confirmed by scanning electron microscopy (SEM) images of the dried film. The likely cause of this difference and utility of USAXS for such measurements are discussed. Ó 2009 Elsevier Inc. All rights reserved.
1. Introduction Titania (TiO2) is a wide band-gap semiconductor with a high refractive index, which lends it to be used as a whitening agent. In powder form it is an effective opacifier. Titania colloids are used in many manufacturing industries, such as cosmetics, paper, coatings, foods, toothpastes and paint. The scattering power of individual titania particles for visible light (300–700 nm) is maximised when the particles have a diameter of approximately 300 nm [1]. One problem encountered in the paint industry is the flocculation or agglomeration of the titania particles, both in the colloidal liquid and while the paint film is drying. The flocculates can be microns in size. This is no longer optimal for scattering visible light and can cause loss of gloss and inferior durability. Flocculation of titania pigments can be retarded by applying moderate shear, although the flocculates typically reform quickly when the shear is removed. In the paint industry flocculation is commonly overcome by adding an excess of surfactant, although this is undesirable as it has a negative impact on other areas of paint performance, such as reduced gloss, tackiness (through leaching of the surfactants), and whitening or blooming of the film surface. The mechanism of flocculation is not well understood [2]. It is suspected that the surfactant molecules preferentially bind to certain facets of the titania particles, and as the solvent is removed * Corresponding author. Address: Industrial Research Limited, Applied Mathematics, P.O. Box 31-310, Lower Hutt, New Zealand. E-mail address: [email protected] (B. Ingham). 0021-9797/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2009.04.035
from the mixture (through evaporation), rearrangement of the titania particles occurs which presses the facets with fewer surfactants bound to them against each other [3]. The titania particle size is between 250 and 300 nm and the flocculates are up to several microns. While dynamic light scattering probes the same size range, this technique is not appropriate for the paint system due to its high opacity. There has been some recent work using atomic force microscopy (AFM) to investigate the surface of dried paint films, and transmission electron microscopy (TEM) on cryo-frozen samples to obtain images of the wet state as well as dried paint films [4,5]. These were able to directly observe the presence and size of flocculates in the wet and dry states for various systems. Other techniques such as disc centrifugation and remission light spectroscopy (RLS) are able to imply the presence of different sized scattering objects [5] but are unable to provide further information such as the distribution of sizes. While microscopy techniques are useful in that they give a direct picture of the actual film, they have limitations in that they only sample a very small area of the paint film and cannot be used in real time. AFM gives a picture of the surface only, and TEM requires the use of thin cross-sections. Both of these can sometimes give misleading results. Ultra-small-angle scattering is a technique that can probe the size range of interest (200– 2000 nm). This samples much larger areas (the size of the beam, which is hundreds of microns to millimetres in size) and is a transmission measurement through the entire film, thereby sampling from the whole volume.
B. Ingham et al. / Journal of Colloid and Interface Science 336 (2009) 612–615
Previous ultra-small-angle neutron scattering (USANS) experiments have been performed on the wet paint system under varying rates of shear [6]. This study showed the degree of flocculation was dependent on the shear rate, and on the polymer co-species. However due to the low signal, USANS measurements cannot be used to perform real-time experiments. Ultra-small-angle X-ray scattering (USAXS), on the other hand, has an advantage of high scattering signals and hence the potential for real time measurements. Previous time-resolved SAXS studies have been performed on silica–titania/polymer composites during formation [7], and on silica/polymer gel systems during flocculation [8]. However, the inorganic particles in these systems are significantly smaller than in commercial paints (10 nm for Ref. [7] and 20 nm for Ref. [8]). Even so, these studies show that SAXS is useful for real-time measurements of flocculation and aggregation. Here we report USAXS measurements on titania colloidal paint films to follow the drying process in two paints with the same grade titania, but different polymer systems. The goal is to explore and better understand the effect of the polymer on the degree of flocculation that occurs during drying.
2. Experimental Two paint systems were studied, using the same ingredients (titanium dioxide, water, polyacrylic dispersant, non-ionic surfactant, coalescing solvent, antifoam, alkali swellable and associative thickeners, humectant and biocides) but different acrylic polymer systems. The millbase is a commercial grade titania with alumina and zirconia coatings and organic polyol treatment, combined with a commercial hydrophobically modified polyacrylate dispersant. The two paint systems using this millbase (henceforth called sample A) are an acrylic emulsion with pigment affinic groups (sample B) and a pure acrylic high solids emulsion (sample C). The pigment volume concentration (PVC) for both samples B and C was 20%. Beam line BL-20-XU at the SPring-8 synchrotron facility, Hgoyo, Japan, was used for the USAXS experiments [9]. The sample-detector distance was 160 m and the wavelength was 0.539 Å (23 keV). The detector is a CCD, which enables scans to be taken in rapid succession, while the paint film dried (typically 1.5–2.5 h). Typical exposure times were 5–10 s. A simple controlled humidity cell was constructed, with X-ray transparent Kapton film windows, one of which acted as the substrate for the film. The paint sample was applied to the Kapton film at an average thickness of approximately 200 lm (wet), the humidity cell sealed, and the film dried under a constant airflow. In addition, a series of data were collected on the paint samples in the liquid state, at various dilutions with aqueous solvent. These were held in place on the beam line in pouches with a well-defined thickness (approximately 0.5 mm). The raw CCD USAXS data images are radially averaged to obtain the intensity I(q) vs. q, the magnitude of the scattering vector. An appropriate background signal was subtracted from all of the datasets: for the time series, the Kapton window signal, and for the dilution experiments, a pouch filled with distilled water. Scanning electron microscopy (SEM) images were obtained of the dried samples using a Leo 440 microscope at 10 kV and 30,000 magnification.
613
is the form factor, and S(qr) is the structure factor. The form factor is an expression describing the shape of the scattering objects (in this case, spheres of the titania particles). This term is assumed to be constant throughout the time series, since the individual titania particles do not change. The structure factor is an expression describing the X-ray scattering interference between neighbouring particles, including flocculation (particles coming together). The dilution experiments are a system where inter-particle scattering is negligible (and hence S(qr) in Eq. (1) is unity). During the drying process, solvent is lost due to evaporation which will lead to two strong effects in the time series: (a) particles moving closer together, but still separated by the surfactant shell, causing changes in the inter-particle scattering S(q); and/or (b) flocculation. There are two ways to model flocculation. One is to add an additional term to Eq. (1) to account for the aggregated particles [10]. The other is to consider flocculation as a continuous variation in S(q) as the individual titania particles come together. Given that: (a) we have a good measure of n(r)[f(qr)]2 from the dilution measurements, (b) the data have a limited dynamic range, and (c) for the sake of simplicity (fewer terms), we have chosen to use the latter approach. Attempts were made to fit Eq. (1) to the data (using a log-normal size distribution, f(qr) for spheres [11], and S(qr) according to Pedersen’s local monodisperse approximation for hard spheres [11]). However the results of the fitting were poor. This is most likely due to a broad and non-monotonic size distribution and the limited q-range of the data. It is less likely to be due to an incorrect choice for f(qr) (e.g. for rods or core–shell structure), since the broad size distribution will smear out any effects due to particle shape/morphology. Fig. 1 shows the dilution series for sample A (millbase) and sample B (paint) samples. As the systems become more dilute, the curves eventually develop the same shape (at dilutions below 12.5%). This shows that the inter-particle scattering is negligible for concentrations below 12.5%. We have chosen the 3% sample A curve as the ‘reference’ curve which shows the colloid SAXS when S(qr) = 1. This is because this is the lowest dilution and it is the same at the 6% dilution. The other sample curves can be divided by this reference dataset to obtain an effective S(qr) for those samples. The assumptions are that the particle size distribution, n(r), and the form factor, f(qr), are the same for all dilutions. Fig. 2, then, shows S(qr) for the dilution series on the paint system.
3. Results and discussion The scattering intensity I(q) is given by the equation
IðqÞ ¼ c
Z
1
nðrÞ½f ðqrÞ2 SðqrÞdr
ð1Þ
0
where c is a constant related to the sample properties and incident intensity, n(r) is the size distribution of the scattering objects, f(qr)
Fig. 1. Reduced data I(q) vs. q for dilution series: (a) sample A (millbase), (b) sample B. The concentrations are shown in the figure legends.
614
B. Ingham et al. / Journal of Colloid and Interface Science 336 (2009) 612–615
Fig. 2. Structure factor for sample B, obtained by dividing the data in Fig. 1b by the 3% data in Fig. 1a.
Each of the curves in Fig. 2 has a peak at around q = 0.015 nm1. This corresponds to constructive interference of inter-particle scattering with a particle–particle spacing of about 420 nm (2p over peak q). The position of the peak moves to slightly lower q at lower dilutions, following a power law of C1/3 where C is the dilution, a common result for hard spheres in dilute solutions [12,13]. The magnitude of S(q) at the peak drops linearly with the particle concentration. In the paint system, this is consistent with the polymer units already being partially bound to the titania particles; recall that for sample B, the polymer units have pigment affinic groups which reduce the need for additional surfactant species. These results show that for sample B, there is minimal flocculation in the wet state, which likely results from the polymers bound on the titania particle surfaces. The time series data of the drying paint film can be thought of as a continuation of the dilution curves, but progressing to concentrations higher than 100% as the solvent evaporates. The first point in the time series corresponds to the 100% ‘dilution’ curve (i.e. the wet state). From this point, we may expect to observe the effective inter-particle scattering S(q) change as the film dries, becomes more compact and the particles move closer together. This will occur either through flocculation or a decrease in the particle-particle spacing as the solvent evaporates. Fig. 3 shows the reduced data (I(q) vs. q) for the sample B film time series during drying. As is apparent the changes are rather subtle. As for the dilution series, each curve is divided by the 3%
Fig. 3. Reduced data I(q) vs. q for sample B, time series. The arrow shows the direction of increasing time.
sample A (millbase) data to yield a plot of effective S(q) and this is shown in Fig. 4. As for Fig. 2, Fig. 4 shows a peak at q = 0.015 nm1. This peak moves to higher q as a function of time, and becomes broader. The peak corresponds to the average interparticle separation decreasing from about 420 to 370 nm over the course of the drying process as is shown in Fig. 5. The change in the separation distance is initially fast and then slows down and reflects the kinetics of the drying process. At the low-q side of Fig. 4 there is relatively little change, indicating that there is minimal flocculation in this system. (The presence of large flocculates would be evidenced by a rise in intensity at low q.) This is confirmed from the SEM images taken of the dried film (Fig. 6a) which shows little flocculation of the titania particles. Fig. 7 shows the reduced data (I(q) vs. q) for the sample C film time series during drying. The curves are quite different to those in Fig. 3. Again the time series data is divided by the 3% sample A (millbase) data to give the effective S(q), which is shown in Fig. 8. This shows significantly different behaviour compared to Fig. 4. Most notably, S(q) increases at low q (0.0025–0.005 nm1), indicating the presence of scattering objects that are larger than 2 lm (the upper size limit of the instrument). This is also evidenced by the lack of any low-q plateau in Fig. 7. The SEM image in Fig. 6b shows a high degree of flocculation; the USAXS curves (Fig. 7) indicate that it is likely that this is present from the initial, wet state. The presence of flocculates mean that the effective S(q) contains
Fig. 4. Structure factor for sample B time series, obtained by dividing the data in Fig. 3 by the 3% data in Fig. 1a. The arrow shows the direction of increasing time.
Fig. 5. Inter-particle scattering distance obtained from the correlation peak in Fig. 4, as a function of time.
B. Ingham et al. / Journal of Colloid and Interface Science 336 (2009) 612–615
615
Fig. 6. Scanning electron micrographs of two dried films: (a) sample B, (b) sample C.
The contrast between Figs. 4 and 8 show the benefit of using USAXS to follow the flocculation process in real time. Traditional methods of assessing the degree of flocculation in wet paint (e.g. measuring viscosity and opacity) are not as sensitive to the absolute flocculate size as USAXS. In addition, these other methods can only be applied in the dry or wet state, not in situ as a function of time. In contrast, the USAXS patterns of the dispersed and flocculated systems are significantly different that flocculation can be readily detected in situ and in real time. 4. Summary
Fig. 7. Reduced data I(q) vs. q for sample C, time series. The arrows show the direction of increasing time.
Ultra-small-angle X-ray scattering (USAXS) has been used to follow the drying process of titania paint films in real time. For two systems studied, we have interpreted changes in terms of the inter-particle scattering. The paint system with pigment affinic groups showed considerably less flocculation than the pure acrylic emulsion. This is likely the result of strong binding of the pigment affinic groups to the titania particles. Acknowledgments The authors thank the SPring-8 beam line scientist, Yoshio Suzuki, for technical support. The synchrotron radiation experiments were performed at BL-20-XU at SPring-8 with the approval of the Japan Synchrotron Radiation Research Institute (JASRI) (Proposal No. 2007A1020). Portions of this research were carried out at the Stanford Synchrotron Radiation Laboratory, a national user facility operated by Stanford University on behalf of the U.S. Department of Energy, Office of Basic Energy Sciences. References
Fig. 8. Structure factor for sample C time series, obtained by dividing the data in Fig. 7 by the 3% data in Fig. 1a. The arrow shows the direction of increasing time.
contributions from both the actual S(q) and the size distribution of the flocculates. We believe that the difference in behaviour of the two systems is due to the strong binding of the polymers in the paint with pigment affinic groups to the titania particles. This prevents the particles from getting close enough to aggregate and hence prevents the flocculation seen in the paint with pure acrylic emulsion. This is consistent with our conclusion that this paint is flocculated even in the wet state (see Figs. 7 and 8).
[1] J. Ferber, J. Luther, Solar Energy Mater. Solar Cells 54 (1998) 265. [2] J.L. Keddie, Mater. Sci. Eng. 21 (1997) 101. [3] Y.W. Jun, M.F. Casula, J.H. Sim, S.Y. Kim, J. Cheon, A.P. Alivisatos, J. Am. Chem. Soc. 125 (2003) 15981. [4] S. Farrokhpay, G.E. Morris, D. Fornasiero, P.G. Self, J. Coat. Technol. Res. 3 (2006) 275. [5] F. Tiarks, T. Frechen, S. Kirsch, J. Leuninger, M. Melan, A. Pfau, F. Richter, B. Schuler, C.L. Zhao, Prog. Org. Coat. 48 (2003) 140. [6] A.I. Nakatani, A. Van Dyk, L. Porcar, J.G. Barker, NIST Center for Neutron Research Annual Report, 2006, p. 26. [7] J.E. Mark, Mater. Res. Soc. Symp. Proc. 628 (2000) CC8.4. [8] V.L. Alexeev, J. Colloid Interface Sci. 206 (1998) 416. [9] N. Yagi, K. Inoue, J. Appl. Crystallogr. 36 (2003) 783. [10] J. Teixeira, J. Appl. Crystallogr. 21 (1998) 781. [11] J.S. Pedersen, J. Appl. Crystallogr. 27 (1994) 595. [12] I. Bodnar, A.S. Silva, R.W. Deitcher, N.E. Weisman, Y.H. Kim, N.J. Wagner, J. Polym. Sci. B Polym. Phys. 38 (2000) 857. [13] V. Castelletto, L.Q. Amaral, Macromolecules 32 (1999) 3469, and references therein.
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
In situ USAXS measurements of titania colloidal paint films during the drying process Bridget Ingham a,b,*, Scott Dickie c, Hiroshi Nanjo d, Michael F. Toney b a
Industrial Research Limited, P.O. Box 31-310, Lower Hutt, New Zealand Stanford Synchrotron Radiation Lightsource, 2575 Sand Hill Road, Menlo Park, CA 94025, USA c Resene Paints Limited, P.O. Box 38-242, Lower Hutt, New Zealand d National Institute of Advanced Industrial Science and Technology, 4-2-1, Nigatake, Miyagino-ku, Sendai 983-8551, Japan b
a r t i c l e
i n f o
Article history: Received 3 February 2009 Accepted 10 April 2009 Available online 21 April 2009 Keywords: USAXS Small-angle X-ray scattering Synchrotron Titania Paint
a b s t r a c t Real-time ultra-small-angle X-ray scattering (USAXS) was used to follow the flocculation of titania particles in two paint systems as the paint films were drying. The inter-particle scattering was extracted by comparing the time series with diluted titania dispersion (having negligible inter-particle interaction). The paint system with pigment affinic groups showed considerably less flocculation than a pure acrylic emulsion. The results were confirmed by scanning electron microscopy (SEM) images of the dried film. The likely cause of this difference and utility of USAXS for such measurements are discussed. Ó 2009 Elsevier Inc. All rights reserved.
1. Introduction Titania (TiO2) is a wide band-gap semiconductor with a high refractive index, which lends it to be used as a whitening agent. In powder form it is an effective opacifier. Titania colloids are used in many manufacturing industries, such as cosmetics, paper, coatings, foods, toothpastes and paint. The scattering power of individual titania particles for visible light (300–700 nm) is maximised when the particles have a diameter of approximately 300 nm [1]. One problem encountered in the paint industry is the flocculation or agglomeration of the titania particles, both in the colloidal liquid and while the paint film is drying. The flocculates can be microns in size. This is no longer optimal for scattering visible light and can cause loss of gloss and inferior durability. Flocculation of titania pigments can be retarded by applying moderate shear, although the flocculates typically reform quickly when the shear is removed. In the paint industry flocculation is commonly overcome by adding an excess of surfactant, although this is undesirable as it has a negative impact on other areas of paint performance, such as reduced gloss, tackiness (through leaching of the surfactants), and whitening or blooming of the film surface. The mechanism of flocculation is not well understood [2]. It is suspected that the surfactant molecules preferentially bind to certain facets of the titania particles, and as the solvent is removed * Corresponding author. Address: Industrial Research Limited, Applied Mathematics, P.O. Box 31-310, Lower Hutt, New Zealand. E-mail address: [email protected] (B. Ingham). 0021-9797/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2009.04.035
from the mixture (through evaporation), rearrangement of the titania particles occurs which presses the facets with fewer surfactants bound to them against each other [3]. The titania particle size is between 250 and 300 nm and the flocculates are up to several microns. While dynamic light scattering probes the same size range, this technique is not appropriate for the paint system due to its high opacity. There has been some recent work using atomic force microscopy (AFM) to investigate the surface of dried paint films, and transmission electron microscopy (TEM) on cryo-frozen samples to obtain images of the wet state as well as dried paint films [4,5]. These were able to directly observe the presence and size of flocculates in the wet and dry states for various systems. Other techniques such as disc centrifugation and remission light spectroscopy (RLS) are able to imply the presence of different sized scattering objects [5] but are unable to provide further information such as the distribution of sizes. While microscopy techniques are useful in that they give a direct picture of the actual film, they have limitations in that they only sample a very small area of the paint film and cannot be used in real time. AFM gives a picture of the surface only, and TEM requires the use of thin cross-sections. Both of these can sometimes give misleading results. Ultra-small-angle scattering is a technique that can probe the size range of interest (200– 2000 nm). This samples much larger areas (the size of the beam, which is hundreds of microns to millimetres in size) and is a transmission measurement through the entire film, thereby sampling from the whole volume.
B. Ingham et al. / Journal of Colloid and Interface Science 336 (2009) 612–615
Previous ultra-small-angle neutron scattering (USANS) experiments have been performed on the wet paint system under varying rates of shear [6]. This study showed the degree of flocculation was dependent on the shear rate, and on the polymer co-species. However due to the low signal, USANS measurements cannot be used to perform real-time experiments. Ultra-small-angle X-ray scattering (USAXS), on the other hand, has an advantage of high scattering signals and hence the potential for real time measurements. Previous time-resolved SAXS studies have been performed on silica–titania/polymer composites during formation [7], and on silica/polymer gel systems during flocculation [8]. However, the inorganic particles in these systems are significantly smaller than in commercial paints (10 nm for Ref. [7] and 20 nm for Ref. [8]). Even so, these studies show that SAXS is useful for real-time measurements of flocculation and aggregation. Here we report USAXS measurements on titania colloidal paint films to follow the drying process in two paints with the same grade titania, but different polymer systems. The goal is to explore and better understand the effect of the polymer on the degree of flocculation that occurs during drying.
2. Experimental Two paint systems were studied, using the same ingredients (titanium dioxide, water, polyacrylic dispersant, non-ionic surfactant, coalescing solvent, antifoam, alkali swellable and associative thickeners, humectant and biocides) but different acrylic polymer systems. The millbase is a commercial grade titania with alumina and zirconia coatings and organic polyol treatment, combined with a commercial hydrophobically modified polyacrylate dispersant. The two paint systems using this millbase (henceforth called sample A) are an acrylic emulsion with pigment affinic groups (sample B) and a pure acrylic high solids emulsion (sample C). The pigment volume concentration (PVC) for both samples B and C was 20%. Beam line BL-20-XU at the SPring-8 synchrotron facility, Hgoyo, Japan, was used for the USAXS experiments [9]. The sample-detector distance was 160 m and the wavelength was 0.539 Å (23 keV). The detector is a CCD, which enables scans to be taken in rapid succession, while the paint film dried (typically 1.5–2.5 h). Typical exposure times were 5–10 s. A simple controlled humidity cell was constructed, with X-ray transparent Kapton film windows, one of which acted as the substrate for the film. The paint sample was applied to the Kapton film at an average thickness of approximately 200 lm (wet), the humidity cell sealed, and the film dried under a constant airflow. In addition, a series of data were collected on the paint samples in the liquid state, at various dilutions with aqueous solvent. These were held in place on the beam line in pouches with a well-defined thickness (approximately 0.5 mm). The raw CCD USAXS data images are radially averaged to obtain the intensity I(q) vs. q, the magnitude of the scattering vector. An appropriate background signal was subtracted from all of the datasets: for the time series, the Kapton window signal, and for the dilution experiments, a pouch filled with distilled water. Scanning electron microscopy (SEM) images were obtained of the dried samples using a Leo 440 microscope at 10 kV and 30,000 magnification.
613
is the form factor, and S(qr) is the structure factor. The form factor is an expression describing the shape of the scattering objects (in this case, spheres of the titania particles). This term is assumed to be constant throughout the time series, since the individual titania particles do not change. The structure factor is an expression describing the X-ray scattering interference between neighbouring particles, including flocculation (particles coming together). The dilution experiments are a system where inter-particle scattering is negligible (and hence S(qr) in Eq. (1) is unity). During the drying process, solvent is lost due to evaporation which will lead to two strong effects in the time series: (a) particles moving closer together, but still separated by the surfactant shell, causing changes in the inter-particle scattering S(q); and/or (b) flocculation. There are two ways to model flocculation. One is to add an additional term to Eq. (1) to account for the aggregated particles [10]. The other is to consider flocculation as a continuous variation in S(q) as the individual titania particles come together. Given that: (a) we have a good measure of n(r)[f(qr)]2 from the dilution measurements, (b) the data have a limited dynamic range, and (c) for the sake of simplicity (fewer terms), we have chosen to use the latter approach. Attempts were made to fit Eq. (1) to the data (using a log-normal size distribution, f(qr) for spheres [11], and S(qr) according to Pedersen’s local monodisperse approximation for hard spheres [11]). However the results of the fitting were poor. This is most likely due to a broad and non-monotonic size distribution and the limited q-range of the data. It is less likely to be due to an incorrect choice for f(qr) (e.g. for rods or core–shell structure), since the broad size distribution will smear out any effects due to particle shape/morphology. Fig. 1 shows the dilution series for sample A (millbase) and sample B (paint) samples. As the systems become more dilute, the curves eventually develop the same shape (at dilutions below 12.5%). This shows that the inter-particle scattering is negligible for concentrations below 12.5%. We have chosen the 3% sample A curve as the ‘reference’ curve which shows the colloid SAXS when S(qr) = 1. This is because this is the lowest dilution and it is the same at the 6% dilution. The other sample curves can be divided by this reference dataset to obtain an effective S(qr) for those samples. The assumptions are that the particle size distribution, n(r), and the form factor, f(qr), are the same for all dilutions. Fig. 2, then, shows S(qr) for the dilution series on the paint system.
3. Results and discussion The scattering intensity I(q) is given by the equation
IðqÞ ¼ c
Z
1
nðrÞ½f ðqrÞ2 SðqrÞdr
ð1Þ
0
where c is a constant related to the sample properties and incident intensity, n(r) is the size distribution of the scattering objects, f(qr)
Fig. 1. Reduced data I(q) vs. q for dilution series: (a) sample A (millbase), (b) sample B. The concentrations are shown in the figure legends.
614
B. Ingham et al. / Journal of Colloid and Interface Science 336 (2009) 612–615
Fig. 2. Structure factor for sample B, obtained by dividing the data in Fig. 1b by the 3% data in Fig. 1a.
Each of the curves in Fig. 2 has a peak at around q = 0.015 nm1. This corresponds to constructive interference of inter-particle scattering with a particle–particle spacing of about 420 nm (2p over peak q). The position of the peak moves to slightly lower q at lower dilutions, following a power law of C1/3 where C is the dilution, a common result for hard spheres in dilute solutions [12,13]. The magnitude of S(q) at the peak drops linearly with the particle concentration. In the paint system, this is consistent with the polymer units already being partially bound to the titania particles; recall that for sample B, the polymer units have pigment affinic groups which reduce the need for additional surfactant species. These results show that for sample B, there is minimal flocculation in the wet state, which likely results from the polymers bound on the titania particle surfaces. The time series data of the drying paint film can be thought of as a continuation of the dilution curves, but progressing to concentrations higher than 100% as the solvent evaporates. The first point in the time series corresponds to the 100% ‘dilution’ curve (i.e. the wet state). From this point, we may expect to observe the effective inter-particle scattering S(q) change as the film dries, becomes more compact and the particles move closer together. This will occur either through flocculation or a decrease in the particle-particle spacing as the solvent evaporates. Fig. 3 shows the reduced data (I(q) vs. q) for the sample B film time series during drying. As is apparent the changes are rather subtle. As for the dilution series, each curve is divided by the 3%
Fig. 3. Reduced data I(q) vs. q for sample B, time series. The arrow shows the direction of increasing time.
sample A (millbase) data to yield a plot of effective S(q) and this is shown in Fig. 4. As for Fig. 2, Fig. 4 shows a peak at q = 0.015 nm1. This peak moves to higher q as a function of time, and becomes broader. The peak corresponds to the average interparticle separation decreasing from about 420 to 370 nm over the course of the drying process as is shown in Fig. 5. The change in the separation distance is initially fast and then slows down and reflects the kinetics of the drying process. At the low-q side of Fig. 4 there is relatively little change, indicating that there is minimal flocculation in this system. (The presence of large flocculates would be evidenced by a rise in intensity at low q.) This is confirmed from the SEM images taken of the dried film (Fig. 6a) which shows little flocculation of the titania particles. Fig. 7 shows the reduced data (I(q) vs. q) for the sample C film time series during drying. The curves are quite different to those in Fig. 3. Again the time series data is divided by the 3% sample A (millbase) data to give the effective S(q), which is shown in Fig. 8. This shows significantly different behaviour compared to Fig. 4. Most notably, S(q) increases at low q (0.0025–0.005 nm1), indicating the presence of scattering objects that are larger than 2 lm (the upper size limit of the instrument). This is also evidenced by the lack of any low-q plateau in Fig. 7. The SEM image in Fig. 6b shows a high degree of flocculation; the USAXS curves (Fig. 7) indicate that it is likely that this is present from the initial, wet state. The presence of flocculates mean that the effective S(q) contains
Fig. 4. Structure factor for sample B time series, obtained by dividing the data in Fig. 3 by the 3% data in Fig. 1a. The arrow shows the direction of increasing time.
Fig. 5. Inter-particle scattering distance obtained from the correlation peak in Fig. 4, as a function of time.
B. Ingham et al. / Journal of Colloid and Interface Science 336 (2009) 612–615
615
Fig. 6. Scanning electron micrographs of two dried films: (a) sample B, (b) sample C.
The contrast between Figs. 4 and 8 show the benefit of using USAXS to follow the flocculation process in real time. Traditional methods of assessing the degree of flocculation in wet paint (e.g. measuring viscosity and opacity) are not as sensitive to the absolute flocculate size as USAXS. In addition, these other methods can only be applied in the dry or wet state, not in situ as a function of time. In contrast, the USAXS patterns of the dispersed and flocculated systems are significantly different that flocculation can be readily detected in situ and in real time. 4. Summary
Fig. 7. Reduced data I(q) vs. q for sample C, time series. The arrows show the direction of increasing time.
Ultra-small-angle X-ray scattering (USAXS) has been used to follow the drying process of titania paint films in real time. For two systems studied, we have interpreted changes in terms of the inter-particle scattering. The paint system with pigment affinic groups showed considerably less flocculation than the pure acrylic emulsion. This is likely the result of strong binding of the pigment affinic groups to the titania particles. Acknowledgments The authors thank the SPring-8 beam line scientist, Yoshio Suzuki, for technical support. The synchrotron radiation experiments were performed at BL-20-XU at SPring-8 with the approval of the Japan Synchrotron Radiation Research Institute (JASRI) (Proposal No. 2007A1020). Portions of this research were carried out at the Stanford Synchrotron Radiation Laboratory, a national user facility operated by Stanford University on behalf of the U.S. Department of Energy, Office of Basic Energy Sciences. References
Fig. 8. Structure factor for sample C time series, obtained by dividing the data in Fig. 7 by the 3% data in Fig. 1a. The arrow shows the direction of increasing time.
contributions from both the actual S(q) and the size distribution of the flocculates. We believe that the difference in behaviour of the two systems is due to the strong binding of the polymers in the paint with pigment affinic groups to the titania particles. This prevents the particles from getting close enough to aggregate and hence prevents the flocculation seen in the paint with pure acrylic emulsion. This is consistent with our conclusion that this paint is flocculated even in the wet state (see Figs. 7 and 8).
[1] J. Ferber, J. Luther, Solar Energy Mater. Solar Cells 54 (1998) 265. [2] J.L. Keddie, Mater. Sci. Eng. 21 (1997) 101. [3] Y.W. Jun, M.F. Casula, J.H. Sim, S.Y. Kim, J. Cheon, A.P. Alivisatos, J. Am. Chem. Soc. 125 (2003) 15981. [4] S. Farrokhpay, G.E. Morris, D. Fornasiero, P.G. Self, J. Coat. Technol. Res. 3 (2006) 275. [5] F. Tiarks, T. Frechen, S. Kirsch, J. Leuninger, M. Melan, A. Pfau, F. Richter, B. Schuler, C.L. Zhao, Prog. Org. Coat. 48 (2003) 140. [6] A.I. Nakatani, A. Van Dyk, L. Porcar, J.G. Barker, NIST Center for Neutron Research Annual Report, 2006, p. 26. [7] J.E. Mark, Mater. Res. Soc. Symp. Proc. 628 (2000) CC8.4. [8] V.L. Alexeev, J. Colloid Interface Sci. 206 (1998) 416. [9] N. Yagi, K. Inoue, J. Appl. Crystallogr. 36 (2003) 783. [10] J. Teixeira, J. Appl. Crystallogr. 21 (1998) 781. [11] J.S. Pedersen, J. Appl. Crystallogr. 27 (1994) 595. [12] I. Bodnar, A.S. Silva, R.W. Deitcher, N.E. Weisman, Y.H. Kim, N.J. Wagner, J. Polym. Sci. B Polym. Phys. 38 (2000) 857. [13] V. Castelletto, L.Q. Amaral, Macromolecules 32 (1999) 3469, and references therein.