ARTICLE pubs.acs.org/Langmuir
Force Interactions of Porous Silica Glass Microspheres against Mirror-Polished Stainless Steel in Nonaqueous Solvents Karran V. Woan† and Wolfgang M. Sigmund*,†,‡ † ‡
Department of Materials Science and Engineering, University of Florida, Gainesville, Florida 32611-6400, United States WCU Department of Energy Engineering, Hanyang University, Seoul 133-791, Korea ABSTRACT: Force interactions of porous silica particles against mirror-polished stainless steel surfaces were quantified in the presence of various solvents to facilitate processing of ceramics with less reliance on organic aids which subsequently need to be burned off. The results were compared to and found to be in good agreement to idealized models of van der Waals force interactions. Significantly, van der Waals attractive forces between steel surfaces and silica surfaces were minimized through the use of tetrahydrofuran and enhanced using methanol. The solvent selections were further extended to settling behavior and were found to follow the general trends determined by Stokes law. The methods presented herein can be extended to other real-world systems.
’ INTRODUCTION Processing of ceramics and tailoring of interparticle surface forces still heavily depends on established use of dispersants, processing under acidic or highly basic conditions, and/or the additions of other organics such as binders and plasticizers. Burnoff of the organics or wastewater removal has often incited great environmental concerns in regions near ceramics processing facilities, as they have the potential to make exhaust gases more corrosive.1,2 Shifting to nonaqueous high vapor pressure solvents has the benefit of facilitating recovery and decreasing the environmental impact of the burnoff process. A great deal of research has been conducted using atomic force microscopy (AFM) to gauge interactions of colloid probes with atomically smooth surfaces such as cleaved crystals or deposited/ coated surfaces.314 Using colloid probes, force interactions between particle and surface can be measured and compared with theoretical calculations of electric double layer, van der Waals (vdW), capillary, electromagnetic, or hydrostatic forces in regards to relatively simple geometries such as spheres and flat surfaces.15,16 Of unique interest is the potential ability to design solvents which would impart repulsive vdW forces between two surfaces.1723 Under industrial processes, surfaces coming into contact frequently deviate broadly from atomic smoothness and exhibit a high degree of roughness even after polishing. Mirror-polished stainless steel or solgel-derived particles having submicrometerlevel roughness might hinder or impede solvent transport, leading to multiple interfacial interactions. Though there are a multitude of methods for quantifying the particleparticle interactions such as using rheology,24 sedimentation studies,25 zeta potential measurements,2629 or using light-scattering techniques to identify agglomerate sizes,30 there are very few techniques to actually quantify the interactions of a particle with a vessel or container wall. r 2011 American Chemical Society
Experimentally, other than colloid probe AFM, which has challenging sample preparation and only single-particle interaction, the main methods of characterizing particlewall interactions present in the literature include the centrifugal method,31 an airflow method,32,33 and a vibration method.34 All of these methods operate on determining the force required to counteract adhesive forces of entrained particles on a surface. Recent developments with total internal reflection microscopy to characterize the trajectories of particles as they are released from optical traps and interact with a substrate show promise and are under development.35 Still, much of the research of particlewall interactions is still in the domain of modeling. Beyond the usual hydrodynamic forces modeled for particlewall interactions,36 Chein and Liao addressed various theoretical constructs of particlewall interactions incorporating Brownian and colloidal forces (van der Waals attraction and electric double layer repulsion) under aqueous conditions.37 Furthermore, particle wall interactions were also investigated by electrophoretic mobility within microchannels also on a theoretical basis.38,39 Though the colloid probe AFM technique lacks a large sampling population provided by the aforementioned techniques, it allows direct measurement of both interactions as a particle nears a surface as well as when the particle is removed from the surface. Specifically, this research seeks to provide a measure of force interactions with a mirror-polished stainless steel surface with porous amorphous silica microspheres in the presence of various common solvents. Such interactions are especially pertinent in applications such as column filling for chromatography or as Received: January 13, 2011 Revised: February 16, 2011 Published: April 11, 2011 5377
dx.doi.org/10.1021/la200157v | Langmuir 2011, 27, 5377–5385
Langmuir
ARTICLE
Figure 1. Calculated dielectric response function curves for the two surfaces and solvents of interest.
model systems of slurry transport through steel tubes in various industrial processes. The first part of the present paper deals with the dielectric response function (DRF) and solvent selection for the interactions of the two solid phases, i.e., sphere vs plate, in the presence of the solvent medium. The response function is determined from optical properties found in the literature and result in the determination of Hamaker coefficients and calculation of the expected vdW force contribution for the interaction. We use the Parsegian methods to estimate the pertinent regions of DRF which is dominant in determining the force interactions. The Hamaker coefficients are then used in the idealized model of sphere and plate interactions to facilitate solvent selection. The second part consists of the measured force interactions for the particleplate system in various solvents and quantifying the degree of attraction, adhesion, and repulsion. Deviations and variations from idealized models are presented. Finally, the third part discusses the particleparticle interaction effects in various solvents through sedimentation and settling studies.
’ PART 1: THEORY AND SOLVENT SELECTION The vdW interaction arises from the presence of dipoles which induces polarization on one another and is expected to be the dominant force at short ranged distances on the order of a few nanometers. The vdW force between a sphere (A) and an infinite plate (B) configuration in a medium (m) can be represented by the following equation consisting of the particle radius (R), the separation distance of the particle to the surface (d), and the Hamaker coefficient (AHam). FAB ¼
AHam R 6d2
frequencies.41 The Hamaker coefficient term can be represented as:42 AHam ¼
3kT εA ðiξn Þ εm ðiξn Þ εB ðiξn Þ εm ðiξn Þ relðlÞ 2 sampling frequencies εA ðiξn Þ þ εm ðiξn Þ εB ðiξn Þ þ εm ðiξn Þ
∑
ð2Þ where k is Boltzmann’s constant, T is temperature in Kelvin, ε are dielectric responses, and rel(l) deal with the relativistic or retardation term. iξ represents the following eigenfrequencies: ξm ¼
ð3Þ
where m is an integer, k is Boltzmann’s constant, T is temperature in Kelvin, and h is Plank’s constant. Note that the m = 0 term provides only a half contribution to the Hamaker coefficient. The NinhamParsegian oscillator model was simplified by Hough and White43 who demonstrated that the dielectric response functions for nonconductors can be estimated from the IR and UV contribution and allows for 1 order-of-magnitude estimation from limited spectral data. εðiξn Þ ¼ 1 þ
CIR CUV 2þ 1 þ ðξn =ωIR Þ 1 þ ðξn =ωUV Þ2
ð4Þ
where CIR and CUV are the absorption strengths in the IR and UV region, and ωIR and ωUV are the corresponding characteristic absorption frequencies for the IR and UV region. If unknown, the absorption strengths can be estimated by the following relationships:
ð1Þ
40
From the treatment by Lifshitz, the Hamaker coefficient is a term describing the polarization dependence of each of the materials on one another: particle, substrate, and intervening medium. The parameter is a difficult term to evaluate because it requires the dielectric response, ε, behavior of a material across the entire frequency spectrum. Parsegian and Ninham were able to provide a simplification of the dielectric response by establishing that the primary contributions to the term arise from the relaxations occurring at microwave, infrared, and ultraviolet
4π2 kT m h
CIR ¼ ε0 n2
ð5Þ
CUV ¼ n2 1
ð6Þ
where ε0 is the dielectric constant and n is the refractive index. For two solid phases made of the same materials interacting with one another, minimization of the attractive vdW interaction is achieved by selection of a solvent medium with similar DRF as the solid phase. For the interaction of two different solid phases, and especially interesting from a scientific standpoint, is when the dielectric response of the fluid medium is in between the 5378
dx.doi.org/10.1021/la200157v |Langmuir 2011, 27, 5377–5385
Langmuir
ARTICLE
Table 1. Properties Used in the Calculation to Determine Hamaker Coefficientsa component silica
chromic oxide toluene
tetrahydrofuran methanol
1-propanol
2-propanol
1-butanol
2-butanol
benzene
a b
nD 1.46
2.5 1.4961
1.404 1.3288
1.385
1.3776
1.3993
1.3978
1.5011
ε0
CUV
ωUV (rad s1)
3.82
1.1316
2.034 1016
11.9 2.38
7.6 33.64
20.8
20.18
17.84
17.26
2.28
ωIR (rad s1)
CIR 0.829
8.670 1013
0.095
1.508 1014
0.798
2.026 1014
5.2500
2.000 10
5.65
1.000 1014 b
1.2383
6.633 10
0.0642
2.183 1013
0.02295
4.485 1013
0.0545
7.780 1011
c
16 b
15
c
6.000 10 9.189 1015
c
16 b
0.9712 0.7657
8.970 10
15
0.9182
9.189 1015
0.8978
8.761 1015
0.9580
7.245 10
15
0.9538
6.776 10
15
1.2533
5.6287 0.254
1.000 1015 b 3.058 1013
0.123
4.149 1013
1.042
8.682 1013
2.801
1.007 1014
0.2134
3.148 1013
0.1778
4.317 1013
0.3808
8.802 1013
0.70977 0.0198
9.977 1013 2.434 1013
0.05927
2.842 1013
0.1441
3.466 1013
0.1801
4.125 1013
0.2208
8.874 1013
0.53462
9.917 1013
0.13545
3.178 1013
0.13233 0.26674
4.287 1013 8.784 1013
0.50743
9.917 1013
0.2582
2.998 1013
0.2185
4.113 1013
0.4767
8.904 1013
0.9931
1.004 1014
0.00969
2.015 1013
0.00561 0.0114
4.437 1013 9.144 1013
b
Solvent information from ref 44, silica properties from ref 19, tetrahydrofuran properties from ref 45, and chromic oxide properties from refs 46 and 47. Estimated parameters. c Calculated from Hough and White simplification.
responses of the two phases. When this occurs, the Hamaker coefficient becomes a negative value, which makes the vdW force equation into a positive value corresponding to a repulsive force interaction. This would provide conditions whereby two surfaces are dispersed without surface charging mechanisms or by the use of adsorbed organic dispersants. Our system of interest involves the surfaces of a silica particle and that of a stainless steel plate, which from surface analysis (see Materials and Methods) exhibited a chromium(III) oxide or chromic oxide passivation layer. The dielectric response function calculations are presented in Figure 1 for the solid phases as well as for various solvents. Most of the spectral parameters were obtained largely from reports in the literature19,44 or estimated. The values used are presented in Table 1, and the corresponding Hamaker coefficients are presented in Table 2. The calculated force response was determined for all the solvents and normalized to the force from methanol; see Figure 2. Thus, a force factor of 1 indicates an attractive methanol-type force interaction. Of the eight solvents selected, only tetrahydrofuran
Table 2. Calculated Hamaker Coefficients for Silica Interacting with Steel Surface in Various Solvent Media solvent
Hamaker coefficient (J)
methanol
8.1914 1020
toluene
7.4646 1020
tetrahydrofuran 1-propanol
2.4108 1020 7.2765 1020
2-propanol
7.2668 1020
1-butanol
7.1561 1020
2-butanol
8.3097 1020
benzene
7.2636 1020
(THF) was determined to facilitate repulsive vdW interactions between the steel plate and silica sphere. Often times, methanol and THF are used in the packing of spheres in the chromatography industry, so these solvents were selected to represent the attractive and repulsive conditions, respectively. 5379
dx.doi.org/10.1021/la200157v |Langmuir 2011, 27, 5377–5385
Langmuir
ARTICLE
Figure 2. Force factor determined by using the calculated Hamaker coefficients in the van der Waals force equation at 10 nm from the 7 μm particles. Results are normalized to methanol’s force.
Figure 3. Electron microscopy of porous silica spheres.
’ PART 2: FORCE CURVE MEASUREMENTS IN NONAQUEOUS SOLVENTS Materials and Methods. The AFM colloidal probe technique allows the surface interactions between different materials to be directly measured.5 Briefly, a particle of one material is adhered to a cantilever tip and precisely brought into contact with a fixed surface of another material. As the two objects interact, the cantilever experiences a deflection either toward the surface or away, depending on attractive or repulsive forces, respectively. These movements are accentuated by the deflection of a laser off the back of the cantilever and measured quantitatively by a series of photodetectors. The cantilever mount can be placed within a liquid cell and used to measure interactions in liquids.4 For the colloid probe, well-characterized Zorbax silica from Agilent Technologies was examined due to their monomodal particle size of 7 μm in diameter. Scanning electron microscopy of the particle is shown in Figure 3 and indicates the particles to have significant variations from a smooth spherical surface. Particles were mounted to tipless AFM cantilevers (Mikromash CSC12 Al BS) with epoxy resin (Loctite, 5 min) using micromanipulators (OptoSigma MB-L-65C-UNC) atop of an optical microscope (Zeiss Axiovert 100A). The epoxy was mixed, and a fine metal syringe tip was used to obtain a very fine droplet on the sharp end. Using the micromanipulator, another syringe tip was brought into contact with the droplet to acquire a much smaller amount of the adhesive at the tip. The chip was then mounted and the adhesive deposited along the central axis of the tipless cantilever. Another syringe which was already dipped in some particles is quickly manipulated to deposit a single particle to the epoxy. Note that the two micromanipulators were electrically coupled with a copper wire to prevent electrostatic attraction of the silica particle. The cantilever was allowed to rest overnight in
an enclosed container while the epoxy cured. Gentle methanol rinses and dips were performed on the cantilever to ensure the particle was adhered well and that other contaminants were removed. The contact regions of the colloid probes were characterized using a reverse imaging grid (NT-MDT TGT1) in the manner described in the literature.48 The colloid probe was used to scan the test grid which is covered by an array of sharp tips. The resulting image is that of the contact region of the probe and can be used to determine the diameter of the contact region. The contact radius average and standard deviation were determined from six spheres and was found to be 2.32 ( 0.20 μm for the hydrophilic silica probe. These values were used to normalize the force curve interactions. Mirror-polished 316 L stainless steel was purchased from McMaster-Carr, cut into squares with a rotary tool, and affixed with epoxy to an AFM specimen substrate. Chemical composition was determined by Auger electron spectroscopy. Three points were examined at 52 cycles of 6 s sputtering with 3 kV Arion sputter. The sectional profile is represented in Figure 4. Note the results are not specifically a depth profile, because the ablation is not uniform or consistent, depending on the chemical composition of the material being ablated. The counts furthermore are relative due to the variations in the number of electrons which can be emitted from each material. Instead, the profile indicates compositional changes for specific elements, and to interpret the data, limited a priori knowledge of the specimen’s layers is required. From the literature we find that the expected structure of stainless steel layers is composed of a contamination region of organic materials, a passivating film, and finally the bulk composition,49 which is expected to be the standard 316 composition of Fe:
