Journal of Colloid and Interface Science 423 (2014) 48–53
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Stabilization of Silicon Carbide (SiC) micro- and nanoparticle dispersions in the presence of concentrated electrolyte Annamaria Vilinska ⇑, Sathish Ponnurangam, Irina Chernyshova, Ponisseril Somasundaran, Damla Eroglu, Jose Martinez, Alan C. West Department of Earth and Environmental Engineering, Columbia University, 500 West 120th Street, New York, NY 10027, USA
a r t i c l e
i n f o
Article history: Received 14 November 2013 Accepted 5 February 2014 Available online 13 February 2014 Keywords: Dispersion stability High ionic strength SiC Polyethyleneimine
a b s t r a c t Achieving a stable and robust dispersion of ultrafine particles in concentrated electrolytes is challenging due to the shielding of electrostatic repulsion. Stable dispersion of ultrafine particles in concentrated electrolytes is critical for several applications, including electro-codeposition of ceramic particles in protective metal coatings. We achieved the steric stabilization of SiC micro- and nano-particles in highly concentrated electroplating Watts solutions using their controlled coating with linear and branched polyethyleneimines (PEI) as dispersants. Branched polyethyleneimine of 60,000 MW effectively disperses both microparticles and nanoparticles at a concentration of 1000 ppm. However, lower polymer dosages and smaller polymers fail to disperse, presumably due to insufficient coverage and bridging flocculation. Dispersion stability correlates well with the adsorption density of PEI on microparticles. We discuss the results in the framework of DLVO theory and suggest possible dispersion mechanisms. However, though the dispersion is enhanced with extended adsorption time, the residual PEI in solution adversely affects electroplating. We overcome this drawback by precoating the particles with the polymer and resuspending them in Watts solution. With this novel approach, we obtained robust dispersions. These results offer new possibilities to control dispersion at high electrolyte concentration, as well as bring new insights into the dispersion phenomenon. Ó 2014 Elsevier Inc. All rights reserved.
1. Introduction In many practical applications, it is critical to have stable dispersions and to prevent particle aggregation, such as for the processing of paints, pigments, paper and pulp, high performance ceramics, minerals, electronics and pharmaceutical and personal care products. Dispersion stability is especially critical for electro-codeposition of particles. Recent approaches for producing wear-resistant coatings include codeposition of ultrafine sized particles into metal matrix, such as nickel films [1–3]. In case of electrocodeposition, particles dispersed in a plating bath are attracted towards the electrode and embedded into the growing metal film. Such polycomposite coatings can have improved material properties and mechanical resistance [4–6]. In electroplating applications, it is critical to have stable dispersions, to prevent particle aggregation and uneven particle incorpo⇑ Corresponding author. Fax: +1 2128548362. E-mail addresses: [email protected] (A. Vilinska), [email protected] (S. Ponnurangam), [email protected] (I. Chernyshova), [email protected] (P. Somasundaran). http://dx.doi.org/10.1016/j.jcis.2014.02.007 0021-9797/Ó 2014 Elsevier Inc. All rights reserved.
ration. The incorporation of particles is dependent on the particle concentration [7] and size, stirring speed and particle aggregation [1], surface hydrophobicity [8–12] and surface charge [13]. Synergy between surfactants and polymers can be beneficially used for dispersion. Palla and Shah [14] used surfactant and polymer mixtures to stabilize slurries, but with low ionic strength, when compared with plating applications. Combination of a copolymer (Acrylic Acid-2-Acrylamido-2-Methylpropane Sulfonic Acid Copolymer) and polyvinylpyrrolidone has been shown to produce a synergisitic effect on the rheology and the stability of SiC dispersions in deionized water [15]. Fluorosurfactants have also been reported to stabilize nanoparticle dispersions of SiC and SiN [16]. However, most of these studies were performed in deionized water or at much lower ionic concentrations than required for electroplating. Only a few studies have been focused on stabilization under plating conditions; cetyltrimethylammonium bromide [17,18], azo-cationic polymer [6] and carbon modification [8] were used to disperse and to support incorporation of ceramic particles during electroplating. In the case of nickel electroplating investigated here, the medium in which the particles are dispersed is extremely concentrated, with an ionic strength of around 2 moles, which makes conventional electrostatic stabilization of particles difficult [19].
A. Vilinska et al. / Journal of Colloid and Interface Science 423 (2014) 48–53
In the case of electro-codeposition of ceramic particles with Ni(II), the positive charge of the particles can improve their incorporation and, therefore cationic dispersive agents can be expected to be beneficial. Among such dispersants is PEI, a cationic polymer (Scheme 1). PEI has been tested as a dispersant for various ceramic particles, but under low ionic stregth conditions [20–22]. The stability of the dispersions is determined by the adsorption density and surface saturation of PEI [21,23], pH [23] and electrolyte concentration [22]. A strong concentration dependence of flocculation has also been reported for kaolinite suspensions stabilized with PEI [24] and silica suspensions with polyacrylic acid–polyethylene comb polymer [25]. Linear and branched PEIs of different molecular weight were tested by Kakui et al. [26] for the dispersion of alumina in ethanol. Since most of the stabilization studies of PEI were done at a low ionic strength those results are not applicable to Watts solutions at extremely high electrolyte content. Therefore the dispersion of ceramic particles by PEI in Watts solution is studied in this work. Herein, we adress the dispersion properties of PEI in Watts bath, with a perspective of its possible application as a dispersant in highly concentrating plating solutions.
49
Branched polyethyleneimine of 60,000 average molecular weight was purchased from Acros Organics and is later referred to as PEI 60. Linear polyethyleneimine of a lower 1200 molecular weight (PEI 1.2) was purchased from Sigma Aldrich. Both polyethyleneimine samples were obtained as 50% aqueous solutions. 2.2. Dispersion preparation All particulate dispersions were prepared at 20 g/L SiC content in Watts solution at pH 3.5. Specifically, the dry powders were added to a Watts solution containing polyethyleneimine and were stirred at 700 rpm for one hour. The dispersions were sonicated for 30 min after stirring to maximize dispersion. In order to study the effect of dissolved polymer in Watts solution on dispersion, the particles were first precoated with PEI and then redispersed in Watts bath. The precoating protocol involves (1) dispersion preparation step mentioned above, (2) centrifugation and supernatant decantation and (3) redispersion of solids in Watts solution without PEI. This method keeps the solid content constant but eliminates the presence and interference of polymer in the solution. To resuspend the particles, sonication was applied for additional 30 min.
2. Materials and methods
2.3. Settling test
2.1. Materials
Settling tests were performed and evaluated according to [14] for 120 min and the settling was documented at discrete time intervals. The tests were performed right after sonication at laboratory temperature of 23 °C. The dispersion stability is expressed as dispersed volume (=100% – settled volume %).
Silicon carbide particles of two sizes were obtained from Alfa Aesar, both b-phase SiC, 99.8% (metal basis). The larger particles were of 1 lm average size and are called microparticles henceforth, while the smaller particles were of 45–55 nm size and are referred to as nanoparticles. Both particle types were characterized using TEM, XPS and Raman microscopy. TEM was performed on a Jeol JEM 100CX Transmission Electron Microscope operated in the bright field mode at 100 KeV. X-ray diffractograms were recorded using a Scintag Model X2 X-ray powder diffractometer. A Cu Ka ( = 0.154 nm) radiation source operated at 45 kV and 35 mA was used. The scan step size was 0.05 deg. XPS spectra were collected with a Perkin Elmer 5400 instrument using monochromatic Al Ka X-rays (1486.6 eV) with pass energies of 17.6 eV at a take-off angle of 45° at pressures of less than 4 10–9 Torr, calibrated using the Au(4f7/2) peak at 84.0 eV. Regional XPS scans were collected at 0.1 eV steps. The atomic concentration ratios were evaluated using the PHI atomic sensitivity factors. The Shirley function was used to subtract the background. The XPS samples of initial particles were prepared by spreading a thin layer of their thick slurries on a UHV metallic holder followed by air-drying. Before measurements, particles were washed with water (two cycles of resuspension-–centrifugation). Raman spectra were measured using a LabRam Aramis microscope (Horiba) equipped with a 532-nm laser. The zeta potential of SiC microparticles was estimated measuring electrophoretic mobilities of the particles in 0.01 M NaCl at a pH of 3.5 with Zeta-Meter 3.0 + (Zeta-Meter Inc.). Particles were dispersed in Watts plating solution of the following composition: 300 g/L NiSO46H2O, 35 g/L NiCl26H2O, 40 g/L H3BO3.
Scheme 1. Molecular structure of polyethyleneimine.
2.4. Absorption test The total organic carbon (TOC) technique was employed to determine the amount of adsorbed PEI. Suspensions of SiC in deionized water and in 0.1 M NaCl solutions with PEI 60 were prepared, stirred at 700 rpm for one hour, sonicated for 30 min and centrifuged. The amount of total organic carbon in the supernatant was measured with a TOC-5000 A Shimadzu. Due to the limitations of this method, the Watts solution cannot be used and the adsorption was estimated for deionized water and 0.1 M NaCl solution. The adsorbed amount of PEI was estimated after 1, 2, 3, 5 and 7 days of adsorption time. 3. Results The effect of polyethyleneimine on the stability of SiC dispersions in plating solution is first discussed and then, the quantification of the polymer adsorption and DLVO calculations are employed to explain the phenomena of particulate stabilization at high ionic strength. 3.1. Particle characterization Morphology and size of the SiC micro- and nanoparticles were characterized using TEM. The microparticles are cubes of about 1 lm size (Fig. 1a), while nanoparticles have spherical shape with size of about 50 nm (Fig. 1b). The surface areas of particles specified by the supplier were 11.5 m2/g and 70–90 m2/g for microparticles and nanoparticles respectively. Raman spectroscopic analysis shows that the SiC phase of both the particles is a zinc-blende 3C (ß) phase. This phase is characterized by a peak at 780–800 cm 1 (Fig. 2). In addition, the Raman spectra display peaks at 1350 and 1590 cm 1 due to polycrystalline graphite [27]. The relative intensities of the graphite peaks
50
A. Vilinska et al. / Journal of Colloid and Interface Science 423 (2014) 48–53
Fig. 1. TEM micrographs of (A) SiC microparticles and (B) SiC nanoparticles.
25000 SiC Microparticles SiC Nanoparticles
15000
10000
5000
0 100
600
1100
1600
2100
2600
Raman shift (cm-1) Fig. 2. Raman spectra of SiC microparticles (dashed) and SiC nanoparticles (full).
Regime I
Dispersed volume (%)
100%
Regime II
Regime III
95% 90% 85% 80% 75%
3.3. Dispersion of nanoparticles vs microparticles
70% 0
5
10
25
50
100
150
200
500
1000
PEI concentration (ppm) Fig. 3. Effect of PEI 60 concentration on dispersion of SiC microparticles in Watts solution after 3 min of settling.
are higher in the case of nanoparticles as compared to microparticles, suggesting a higher content of graphite in the former. XPS reveals oxidized SiC species on the surface of both particle types such as SiO2 and SiOxCy with a Si 2 p binding energy of 104.4 eV and 103.1 eV, respectively. The C 1s photoemission contains a peak at 285 eV of graphite [9,28]. 3.2. Effect of polymer concentration and molecular weight High electrolyte concentration of the Watts solution shields the electrostatic repulsion between microparticles and allows their quick settling without added polymer (Fig. 3). When 60,000 MW polyethyleneimine is added to the system, three regimes are observed depending upon the polymer concentration. The first one,
The electro-codeposition of nanoparticles can lead to superior film properties [29,30]; therefore nanoparticles of SiC were included in our studies. Dispersion tests of nanoparticles, were
100%
Dispersed volume (%)
Intensity (cnt)
20000
at low PEI concentrations, shows a slight increase in the dispersion stability. In the second regime, at intermediate PEI concentrations, dispersion stability is worse and the settled volumes were higher. The third regime, above 500 ppm of PEI, stabilization is high with the dispersed volumes increased significantly. The first two regimes are considered to be due to insufficient coverage of particles with the polymer. In fact, if the polymer can form sufficiently long tails extending from particle to particle, it may induce bridging flocculation [25], as seen at intermediate concentrations. However at the lowest concentrations of PEI, polymer does not cause bridging flocculation leaving the system more dispersed. At the highest concentration (regime III), the system is possibly stabilized due to complete coverage of particles with the polymer, with the polymer extending into the solution and creating a steric barrier between the particles. In addition to PEI concentration, another important parameter in achieving sterically stabilized dispersions is the molecular weight of the polymer. Settling experiments were therefore performed using a linear 1200 MW polyethyleneimine (Fig. 4). The stabilizing effect of PEI 1.2 for microparticle systems is lower than that of PEI 60. Dispersion and settling are unaffected up to 500 ppm, above which a slight improvement can be seen. The effect is however much weaker than with the PEI 60, with the particles dispersed after 3 min at 1000 ppm concentration. The PEI 1.2 either does not adsorb on the particles, or the adsorbed polymer does not extend sufficiently into the bulk solution to cause steric stabilization.
95% 90% 85% 80% 75% 70% 0
10
50
100
500
1000
5000
PEI concentration (ppm) Fig. 4. Effect of PEI 1.2 concentration on dispersion of SiC microparticles in Watts solution after 3 min of settling.
51
A. Vilinska et al. / Journal of Colloid and Interface Science 423 (2014) 48–53 100%
Dispersed volume (%)
Dispersed volume (%)
100% 95% 90% 85% 80% 75%
90%
80%
70% 0
5
10
25
50
100
150
200
500 1000 2000 5000
70% 0
Concentration (ppm)
10
50
100
500
1000
PEI concentration ppm
conducted with samples prepared identically to microparticle dispersions with PEI 60, and the results after 120 min are given in Fig. 5. The prepared nanoparticulate dispersions are naturally more stable than the microparticle dispersions, due to a much smaller particle size of the former. Some settling and clearing is observed after 120 min for all systems below 200 ppm PEI 60. The nanoparticulate dispersions are completely stabilized at 1000 and 5000 ppm. In contrast, low molecular weight PEI has a destabilizing effect on SiC nanoparticle dispersion as shown in Fig. 6. Within the concentration range studied no significant improvement of stability is found compared to the sample without any polymer. At concentrations above 500 ppm, the PEI 1.2 acts as a flocculant. It is possible that the stabilization concentration is over 5000 ppm, but since such high concentrations of PEI 1.2 showed a detrimental effect on electroplating [31] those concentrations were not studied.
Fig. 6. Effect of PEI 60 concentration on dispersion of SiC nanoparticles in Watts solution after 120 min of settling.
100%
Conventional (PEI in solution) Precoated with PEI (resuspended)
Dispersed volume
Fig. 5. Effect of PEI 60 concentration on dispersion of SiC nanoparticles in Watts solution after 120 min of settling.
95%
90%
85%
80% 0.04
1
2
3
5
7
Aging time (days) Fig. 7. Effect of aging time and coating protocol on SiC microparticles dispersions stabilized with 1000 ppm PEI 60 in Watts solution after 5 min of settling.
3.4. Effect of adsorption time and particle precoating in Watts solution 3.5. Polymer adsorption To understand the mechanisms of dispersion in Watts solution, we studied the effect of ionic strength and polymer concentration. From the electroplating perspective, it is important, that only a small portion of (5–10%) of the loaded PEI 60 amount is adsorbed on the SiC microparticles both with water and 0.1 NaCl. The adsorption of PEI increases with PEI concentration (Fig. 8), with coverage of microparticles reaching a maximum of 0.2–0.3 mg/ m2 at 500–1000 ppm of PEI in solution. The adsorption of PEI on SiC is higher in 0.1 M NaCl than in deionized water. Surface saturation is reached at between 500 and 1000 ppm of PEI 60. The Langmuir linear regression gives a straight line indicating monolayer
0.5
Microparticles water
0.45
PEI adsorbed (mg/m2)
Electroplating tests show that large concentrations of PEI can lead to undesired plating conditions, presumably because PEI also adsorbs on Ni, impacting charge-transfer mechanisms [31]. Therefore a scheme was developed to remove the residual polymer in solution by centrifugation, decantation and replenishment of the solution. This process is henceforth called precoating. The method improved the efficiency of plating [31]. To test the effect of the precoating methodology on the dispersion, adsorption for a prolonged time as well as the aging of redispersed precoated particles in Watts solution was tested. Fig. 7 compares the dispersed volume of SiC microparticles stabilized with 1000 ppm PEI 60 conventionally and with precoating: the first data set is the result of dispersion of SiC in Watts bath and PEI in solution and the second data set is for a dispersion of SiC particles precoated with PEI and resuspended in fresh Watts solution. The aging effect was tested on both dispersions for duration of 7 days. Initially the settled volumes of both dispersions are similar, which suggests that the centrifugation and solution replacement does not have a deleterious effect on stability. With an increase in time, the dispersion volume of the conventional dispersion increases while that of the precoated one remains stable. As there is some residual polymer present in the solution, in the former case some additional adsorption may take place and further improve the stability. However, such a process is slow. The precoated dispersion exhibits constant stability of 91–92% over a period of several days, suggesting the precoating protocol to be a viable option for use in the codeposition of microparticles. However, when the same protocol is applied to a dispersion of nanoparticles, the resuspended dispersion settles quickly (data not shown), thus this protocol can be concluded not to be suitable for processing nanoparticles.
Microparticles 0.1 M NaCl
0.4
Nanoparticles water
0.35
Nanoparticles 0.1 M NaCl
0.3 0.25 0.2 0.15 0.1 0.05 0 1
10
100
1000
Equilibrium PEI concentration (ppm) Fig. 8. Effect of NaCl on adsorption of 60,000 PEI onto SiC microparticles and nanoparticles relative to equilibrium concentration in water. Lines are drawn for the convenience of the reader.
A. Vilinska et al. / Journal of Colloid and Interface Science 423 (2014) 48–53 0.30
40
0.25
30
Electrostatic interaction LVDW SiO2 LVDW ZrC
20
LVDW Carbon
0.20 10
0.15
G (kT)
PEI adsorbed (mg/m 2)
52
0.10
0 0
2
4
6
8
10
12
14
16
18
20
-10
0.05
-20 -30
0.00 0
1
2
3 4 5 Adsorption time (days)
6
Attraction
7
Fig. 9. Effect of adsorption time on the adsorption of 60,000 PEI onto SiC microparticles. Line is drawn for convenience of the reader.
coverage with a Cmax of 0.3 mg/m2 and 0.17 mg/m2 in 0.1 M NaCl and deionized water respectively. The leveling off the curves indicates close to complete monolayer coverage at 500–1000 ppm initial concentration of PEI 60, and 20–45% surface coverage at intermediate initial PEI concentrations of 50–100 ppm. The adsorption of PEI 60 on nanoparticles is found to be negligible in the concentration range studied (Fig. 8). At the highest concentration of PEI 60 of 1000 ppm the adsorption density is below 0.01 mg/m2. According to literature, PEI preferentially adsorbs onto oxidized surface of SiC [32] with available silanol and siloxane groups. The strong H-bonding interaction between amino groups and surface silanes is well known [33]. Since SiO-like species are present on both surfaces, PEI was expected to adsorb on both micro and nanoparticles. However, the comparatively higher graphite content of nanoparticles is possibly responsible for the reduced number of available adsorption sites and the lower adsorption capacity. Interestingly, the amount adsorbed onto microparticles increases with time by factor of 1.5 after 3 days (Fig. 9). We attribute the higher stability of these particles in Watts solution at 1000 ppm PEI after 3 days to higher adsorption of the polymer.
-40
Separation distance (nm)
Fig. 10. Theoretical evaluation of interparticle (electrostatic and LVDW) forces between two SiC microparticles as a function of separation distance at the ionic strength of Watts solution.
tive LVDW forces. A rough estimate of polymer size based on molecular weight can be done with the following formula: 0:06M 1=2 w , which gives 14.69 nm for the PEI 60 and 2 nm for PEI 1.2 respectively. Assuming monolayer coverage, the PEI 1.2 cannot create a sufficient separation distance between the microparticles and is not suitable as a dispersant under high ionic strength conditions. PEI 60 can create the required 10–15 nm thick monolayer on the surface of SiC microparticles and prevent aggregation, if the coverage is complete. In the case of partially covered particles, attractive H-bond interactions between the particles and PEI can cause aggregation as observed for regime II in Fig. 2. For the nanosized particles the repulsive electrostatic forces can extend to 0.5 nm and the LVDW forces are attractive up to a distance of 2 nm from the nanoparticle surface (results not shown). Nanoparticles can be effectively dispersed with a much thinner adsorbed steric layer of 1–2 nm. To summarize, the theoretical calculations suggest that steric stabilization of silica like surfaces in high ionic environment requires the thickness of adsorbed polymer of 10 nm and more.
3.6. Inter-particle forces evaluation 4. Discussion To quantify the dominant interparticle forces between two SiC particles, DLVO-based calculations were performed. Electrostatic and Lifshitz van der Waals interaction energies were calculated from electrophoretic data and published data of particles with similar composition. The zetapotential of SiC particles was measured with the electrophoretic method in 0.01 M NaCl and calculated using the Smoluchowski equation to be around 6 mV (±3 mV) at pH of 3.5 and this value is used in our calculations. The electrostatic interactions are typically strong and long range for charged particles at low ionic strength. In 2 molar solutions the calculated double layer thickness is less than 1 Å and the electrostatic repulsive force is close to zero at distances higher than 1 nm from the surface (Fig. 10). Lifshitz van der Waals (LVDW) interactions were calculated using a Hamaker constant calculated from the dispersive components of surface energy [34] of the surface species found on SiC particles by XPS (carbon, SiO2) or particles with similar composition [35–37]. For all materials, the LVDW interactions below the distance of 10 nm are sufficiently attractive to predominate over the electrostatic repulsion to cause aggregation. Therefore, to prevent aggregation in this case the steric stabilizing layer has to be thicker than the distance below which the LVDW interactions dominate. Based on the interaction energy values calculated, SiC microparticles have to have a continuous coverage with steric stabilizer of a 10–15 nm thickness in order to prevent aggregation due to attrac-
In order to keep SiC particles dispersed in Watts plating solution, steric stabilization is necessary, and this can be achieved using a cationic polymer. PEI was chosen for the present study based on the positive results found in the literature and our initial screening tests with different cationic and nonionic polymers. The stabilization effect of PEI depends on polymer concentration, molecular weight as well as the particle size and surface composition. The higher molecular weight polymer PEI 60 was more successful in stabilization than the lower molecular weight PEI 1.2. The reason that, the 60,000 MW PEI can create a polymer layer thicker than the separation distance required to overcome the LDVW attractive forces, while the 1200 MW polymer can create a steric repulsion layer of only a few nm. In contrast, to achieve similar stabilization with PEI 1.2 as with PEI 60, a 5–10 multilayer adsorption would be required, which is impractical. Adsorption of PEI 60 was found to be the main cause of stabilization of dispersions with microparticles with two important parameters: surface coverage and thickness of adsorbed layer. In particular, stabilization occurred only in the concentration ranges with recognizable surface saturation and close to a complete surface coverage. PEI 60 at intermediate concentrations of 50– 200 ppm covers the particles only partially, which can lead to bridging between two particles causing aggregation. The second important characteristic of the steric layer is the thickness neces-
A. Vilinska et al. / Journal of Colloid and Interface Science 423 (2014) 48–53
sary to overcome the LVDW attraction between the particles. We estimated PEI 60 coverage, assuming that the polymer species have spherical shape on the basis of Analytical centrifugation of polymer solution (data not reported). As an estimate of the polymer species diameter we took the value of 14.7 nm which was calculated as described in Results section. Under these assumptions and using the adsorption data, we obtained polymer surface coverage of 23% and 49% for microparticles in deionized water and NaCl, respectively. Since both systems exhibited surface saturation, it can be proposed that, the polymer species presumably adsorb in a flatter arrangement rather than as spheres, leading to a thinner polymer layer. The adsorption amount from NaCl solution was larger than that from deionized water suggesting that the background electrolyte favors a thicker polymer arrangement on the surface of particles. The presence of electrolyte can thus influence the adsorbed amount as well as the alignment of PEI polymer coils on the interface similarly as pH [38]. It was also observed that, PEI adsorption from Watts solution is strong enough for the dispersion of microparticles to remain stable even after redispersion in Watts solution without any PEI. The dispersion stability of the resuspended PEI precoated microparticles was constant over the period of 7 days, indicating a stable polymeric layer adsorbed on the particles. On the other hand, the nanoparticle dispersions are destabilized after polymer removal from solution and resuspension. The instability of nanoparticle dispersions upon resuspension is attributed to an insufficient steric layer as a result of low PEI adsorption. This may be due to a different mechanism such as depletion stabilization being operative for stabilization, or weak physisorption of polymer on the nanoparticle surface and desorption after resuspension. In the case of the depletion stabilization [39], the polymer has to be in solution in order to disperse the particles and prevent aggregation. Adsorption on nanoparticles was found to be low and 1/20 in magnitude of microparticles. The thickness of such layer is then less than the required 1 nm, or the particles are not covered completely and, hence the particles can aggregate and settle after resuspension. Importantly, the stability of microparticle dispersions is enhanced if the adsorption time is extended to several days with the increase in the adsorption densities correlating with the increase in the dispersion stability. Extended adsorption time may allow the adsorption of additional polymer coils on the particle surface or replacement of the shorter polymeric chains with the longer ones [40], both leading to an increase of steric layer thickness and thus efficient particle dispersion. 5. Conclusions We studied PEI as a dispersant for SiC microparticles and nanoparticles in a high electrolyte Watts plating solution at pH 3.5. PEI of 60,000 MW is an excellent dispersant for both the particle types. We determined dosage of the polymer required to disperse these particles. We demonstrated that the dispersion stability is controlled by the adsorption density of the polymer, polymer layer thickness, and the extent of Lifshitz van der Waals (LVDW) interaction between the particles. Therefore, to retard particle aggregation due to (LVDW) interactions when repulsive electrostatic forces are screened out, it is necessary to have a close to complete coverage of the particles with a thick enough polymer layer. We found that the adsorption density and surface saturation level of PEI depend on the ionic strength of the solution. We revealed that the ceramic particles are dispersed by two different mechanisms depending on the surface properties of SiC particles and their size. The microparticles are dispersed due to the buildup of a steric polymeric
53
layer of a sufficient thickness. In contrast, PEI does not adsorb on the nanoparticles suggesting the dispersion stabilization by the free polymer in solution, known as depletion stabilization. We also found that the adsorbed layer of PEI on SiC microparticles is stable when the particles are coated with the polymer and then redispersed in Watts solution. This finding is important for attaining dispersion in systems sensitive to polymer content such as electroplating. Acknowledgment Financial support from The Boeing Company is gratefully acknowledged. References [1] H.-K. Lee, H.-Y. Lee, Surf. Coat. Technol. 201 (2007) 4711–4717. [2] L. Benea, P.L. Bonora, A. Borello, S. Martelli, F. Wenger, P. Ponthiaux, J. Galland, J. Electrochem. Soc. 148 (2001) C461–C465. [3] C.T.J. Low, R.G.A. Wills, F.C. Walsh, Surf. Coat. Technol. 201 (2006) 371–383. [4] E. García-Lecina, I. García-Urrutia, J.A. Díez, M. Salvo, F. Smeacetto, G. Gautier, R. Seddon, R. Martin, Electrochim. Acta 54 (2009) 2556–2562. [5] N.K. Shrestha, K. Sakurada, M. Masuko, T. Saji, Surf. Coat. Technol. 140 (2001) 175–181. [6] N.K. Shrestha, M. Masuko, T. Saji, Wear 254 (2003) 555–564. [7] T. Borkar, S.P. Harimkar, Surf. Coat. Technol. 205 (2011) 4124–4134. [8] D. Soccol, J. Martens, S. Claessens, J. Fransaera, J. Electrochem. Soc. 158 (2011) D515–D523. [9] R.P. Socha, K. Laajalehto, P. Nowak, Colloids Surf., A 208 (2002) 267–275. [10] R.P. Socha, K. Laajalehto, P. Nowak, Surf. Interface Anal. 34 (2002) 413–417. [11] R.P. Socha, P. Nowak, K. Laajalehto, J. Väyrynen, Colloids Surf., A 235 (2004) 45–55. [12] L. Stappers, J. Fransaer, J. Electrochem. Soc. 153 (2006) C472–C482. [13] M. Sarret, C. Muller, A. Amell, J. Nanopart. Res. 9 (2007) 1073–1080. [14] B.J. Palla, D.O. Shah, J. Colloid Interface Sci. 256 (2002) 143–152. [15] C. Xiao, L. Gao, M. Lu, H. Chen, L. Guo, L. Tao, Colloids Surf., A 355 (2010) 104– 108. [16] A. Amell, C. Muller, M. Sarret, Surf. Coat. Technol. 205 (2010) 356–362. [17] M.-D. Ger, Mater. Chem. Phys. 87 (2004) 67–74. [18] K.H. Hou, M.D. Ger, L.M. Wang, S.T. Ke, Wear 253 (2002) 994–1003. [19] B.S. Necula, I. Apachitei, L.E. Fratila-Apachitei, C. Teodosiu, J. Duszczyk, J. Colloid Interface Sci. 314 (2007) 514–522. [20] J. Zhang, Q. Xu, F. Ye, Q. Lin, D. Jiang, M. Iwasa, Colloids Surf., A 276 (2006) 168– 175. [21] X. Zhu, T. Uchikoshi, T.S. Suzuki, Y. Sakka, J. Am. Ceram. Soc. 90 (2007) 797– 804. [22] G. Xu, L. Xu, S. Pan, G. Song, China Particuol. 2 (2004) 182–184. [23] F. Tang, T. Uchikoshi, K. Ozawa, Y. Sakka, J. Ceram. Soc. Jpn. 113 (2005) 584– 587. [24] W.K. Mekhamer, N. Al Andis, M. El Shabanat, J. King Saud Univ. (Sci.) 21 (2009) 125–132. [25] C.P. Whitby, P.J. Scales, F. Grieser, T.W. Healy, G. Kirby, J.A. Lewis, C.F. Zukoski, J. Colloid Interface Sci. 262 (2003) 274–281. [26] T. Kakui, T. Miyauchi, H. Kamiya, J. Eur. Ceram. Soc. 25 (2005) 655–661. [27] M. Yoshikawa, N. Nagai, Vibrational spectroscopy of carbon and silicon materials, in: Handbook of Vibrational Spectroscopy, John Wiley & Sons, 2006, pp. 2593–2620. [28] C. Onneby, C.G. Pantano, J. Vac. Sci. Technol., A 15 (1997) 1597–1602. [29] L. Benea, P.L. Bonora, A. Borello, S. Martelli, Wear 249 (2001) 995–1003. [30] X.J. Sun, J.G. Li, Tribol. Lett. 28 (2007) 223–228. [31] D. Eroglu, A. Vilinska, P. Somasundaran, A.C. West, J. Electrochem. Soc. 160 (2013) D35–D40. [32] J. Sun, L. Gao, J. Eur. Ceram. Soc. 21 (2001) 2447–2451. [33] I.V. Chernyshova, Langmuir 16 (2000) 8071–8084. [34] A. Vilinska, H.K. Rao, Miner. Metall. Process. 28 (2011) 151–158. [35] H.-J. Jacobash, K. Grundke, E. Mader, K.-H. Freitag, U. Panzer, Application of surface free energy concept in polymer processing. Contact Angle, Wettability and Adhesion, in: K.L. Mittal (Ed.), 1993, pp. 921–936. [36] C. Rulison, Two component surface energy characterization as a predictor of Wettability and Dispersability, Kruss Application/Technical, Notes, AN213, 2000, pp. 1–22. [37] K. Bobzin, Benetzungs und Korrosionsverhalten von PVD-beschichteten Werkstoffen fur den Einsatz in umweltvertraglichen Tribosystemen, Thesis (Werkstoffwissenschftl. Schriftenreihe, Bd. 39), 3-8265-7414-1, 2000. [38] B.P. Singh, J. Jena, L. Besra, S. Bhattacharjee, J. Nanopart. Res. 9 (2007) 797–806. [39] R.I. Feigin, D.H. Napper, J. Colloid Interface Sci. 74 (1980) 567–571. [40] P. Frantz, S. Granick, Phys. Rev. Lett. 66 (1991) 899–902.
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Stabilization of Silicon Carbide (SiC) micro- and nanoparticle dispersions in the presence of concentrated electrolyte Annamaria Vilinska ⇑, Sathish Ponnurangam, Irina Chernyshova, Ponisseril Somasundaran, Damla Eroglu, Jose Martinez, Alan C. West Department of Earth and Environmental Engineering, Columbia University, 500 West 120th Street, New York, NY 10027, USA
a r t i c l e
i n f o
Article history: Received 14 November 2013 Accepted 5 February 2014 Available online 13 February 2014 Keywords: Dispersion stability High ionic strength SiC Polyethyleneimine
a b s t r a c t Achieving a stable and robust dispersion of ultrafine particles in concentrated electrolytes is challenging due to the shielding of electrostatic repulsion. Stable dispersion of ultrafine particles in concentrated electrolytes is critical for several applications, including electro-codeposition of ceramic particles in protective metal coatings. We achieved the steric stabilization of SiC micro- and nano-particles in highly concentrated electroplating Watts solutions using their controlled coating with linear and branched polyethyleneimines (PEI) as dispersants. Branched polyethyleneimine of 60,000 MW effectively disperses both microparticles and nanoparticles at a concentration of 1000 ppm. However, lower polymer dosages and smaller polymers fail to disperse, presumably due to insufficient coverage and bridging flocculation. Dispersion stability correlates well with the adsorption density of PEI on microparticles. We discuss the results in the framework of DLVO theory and suggest possible dispersion mechanisms. However, though the dispersion is enhanced with extended adsorption time, the residual PEI in solution adversely affects electroplating. We overcome this drawback by precoating the particles with the polymer and resuspending them in Watts solution. With this novel approach, we obtained robust dispersions. These results offer new possibilities to control dispersion at high electrolyte concentration, as well as bring new insights into the dispersion phenomenon. Ó 2014 Elsevier Inc. All rights reserved.
1. Introduction In many practical applications, it is critical to have stable dispersions and to prevent particle aggregation, such as for the processing of paints, pigments, paper and pulp, high performance ceramics, minerals, electronics and pharmaceutical and personal care products. Dispersion stability is especially critical for electro-codeposition of particles. Recent approaches for producing wear-resistant coatings include codeposition of ultrafine sized particles into metal matrix, such as nickel films [1–3]. In case of electrocodeposition, particles dispersed in a plating bath are attracted towards the electrode and embedded into the growing metal film. Such polycomposite coatings can have improved material properties and mechanical resistance [4–6]. In electroplating applications, it is critical to have stable dispersions, to prevent particle aggregation and uneven particle incorpo⇑ Corresponding author. Fax: +1 2128548362. E-mail addresses: [email protected] (A. Vilinska), [email protected] (S. Ponnurangam), [email protected] (I. Chernyshova), [email protected] (P. Somasundaran). http://dx.doi.org/10.1016/j.jcis.2014.02.007 0021-9797/Ó 2014 Elsevier Inc. All rights reserved.
ration. The incorporation of particles is dependent on the particle concentration [7] and size, stirring speed and particle aggregation [1], surface hydrophobicity [8–12] and surface charge [13]. Synergy between surfactants and polymers can be beneficially used for dispersion. Palla and Shah [14] used surfactant and polymer mixtures to stabilize slurries, but with low ionic strength, when compared with plating applications. Combination of a copolymer (Acrylic Acid-2-Acrylamido-2-Methylpropane Sulfonic Acid Copolymer) and polyvinylpyrrolidone has been shown to produce a synergisitic effect on the rheology and the stability of SiC dispersions in deionized water [15]. Fluorosurfactants have also been reported to stabilize nanoparticle dispersions of SiC and SiN [16]. However, most of these studies were performed in deionized water or at much lower ionic concentrations than required for electroplating. Only a few studies have been focused on stabilization under plating conditions; cetyltrimethylammonium bromide [17,18], azo-cationic polymer [6] and carbon modification [8] were used to disperse and to support incorporation of ceramic particles during electroplating. In the case of nickel electroplating investigated here, the medium in which the particles are dispersed is extremely concentrated, with an ionic strength of around 2 moles, which makes conventional electrostatic stabilization of particles difficult [19].
A. Vilinska et al. / Journal of Colloid and Interface Science 423 (2014) 48–53
In the case of electro-codeposition of ceramic particles with Ni(II), the positive charge of the particles can improve their incorporation and, therefore cationic dispersive agents can be expected to be beneficial. Among such dispersants is PEI, a cationic polymer (Scheme 1). PEI has been tested as a dispersant for various ceramic particles, but under low ionic stregth conditions [20–22]. The stability of the dispersions is determined by the adsorption density and surface saturation of PEI [21,23], pH [23] and electrolyte concentration [22]. A strong concentration dependence of flocculation has also been reported for kaolinite suspensions stabilized with PEI [24] and silica suspensions with polyacrylic acid–polyethylene comb polymer [25]. Linear and branched PEIs of different molecular weight were tested by Kakui et al. [26] for the dispersion of alumina in ethanol. Since most of the stabilization studies of PEI were done at a low ionic strength those results are not applicable to Watts solutions at extremely high electrolyte content. Therefore the dispersion of ceramic particles by PEI in Watts solution is studied in this work. Herein, we adress the dispersion properties of PEI in Watts bath, with a perspective of its possible application as a dispersant in highly concentrating plating solutions.
49
Branched polyethyleneimine of 60,000 average molecular weight was purchased from Acros Organics and is later referred to as PEI 60. Linear polyethyleneimine of a lower 1200 molecular weight (PEI 1.2) was purchased from Sigma Aldrich. Both polyethyleneimine samples were obtained as 50% aqueous solutions. 2.2. Dispersion preparation All particulate dispersions were prepared at 20 g/L SiC content in Watts solution at pH 3.5. Specifically, the dry powders were added to a Watts solution containing polyethyleneimine and were stirred at 700 rpm for one hour. The dispersions were sonicated for 30 min after stirring to maximize dispersion. In order to study the effect of dissolved polymer in Watts solution on dispersion, the particles were first precoated with PEI and then redispersed in Watts bath. The precoating protocol involves (1) dispersion preparation step mentioned above, (2) centrifugation and supernatant decantation and (3) redispersion of solids in Watts solution without PEI. This method keeps the solid content constant but eliminates the presence and interference of polymer in the solution. To resuspend the particles, sonication was applied for additional 30 min.
2. Materials and methods
2.3. Settling test
2.1. Materials
Settling tests were performed and evaluated according to [14] for 120 min and the settling was documented at discrete time intervals. The tests were performed right after sonication at laboratory temperature of 23 °C. The dispersion stability is expressed as dispersed volume (=100% – settled volume %).
Silicon carbide particles of two sizes were obtained from Alfa Aesar, both b-phase SiC, 99.8% (metal basis). The larger particles were of 1 lm average size and are called microparticles henceforth, while the smaller particles were of 45–55 nm size and are referred to as nanoparticles. Both particle types were characterized using TEM, XPS and Raman microscopy. TEM was performed on a Jeol JEM 100CX Transmission Electron Microscope operated in the bright field mode at 100 KeV. X-ray diffractograms were recorded using a Scintag Model X2 X-ray powder diffractometer. A Cu Ka ( = 0.154 nm) radiation source operated at 45 kV and 35 mA was used. The scan step size was 0.05 deg. XPS spectra were collected with a Perkin Elmer 5400 instrument using monochromatic Al Ka X-rays (1486.6 eV) with pass energies of 17.6 eV at a take-off angle of 45° at pressures of less than 4 10–9 Torr, calibrated using the Au(4f7/2) peak at 84.0 eV. Regional XPS scans were collected at 0.1 eV steps. The atomic concentration ratios were evaluated using the PHI atomic sensitivity factors. The Shirley function was used to subtract the background. The XPS samples of initial particles were prepared by spreading a thin layer of their thick slurries on a UHV metallic holder followed by air-drying. Before measurements, particles were washed with water (two cycles of resuspension-–centrifugation). Raman spectra were measured using a LabRam Aramis microscope (Horiba) equipped with a 532-nm laser. The zeta potential of SiC microparticles was estimated measuring electrophoretic mobilities of the particles in 0.01 M NaCl at a pH of 3.5 with Zeta-Meter 3.0 + (Zeta-Meter Inc.). Particles were dispersed in Watts plating solution of the following composition: 300 g/L NiSO46H2O, 35 g/L NiCl26H2O, 40 g/L H3BO3.
Scheme 1. Molecular structure of polyethyleneimine.
2.4. Absorption test The total organic carbon (TOC) technique was employed to determine the amount of adsorbed PEI. Suspensions of SiC in deionized water and in 0.1 M NaCl solutions with PEI 60 were prepared, stirred at 700 rpm for one hour, sonicated for 30 min and centrifuged. The amount of total organic carbon in the supernatant was measured with a TOC-5000 A Shimadzu. Due to the limitations of this method, the Watts solution cannot be used and the adsorption was estimated for deionized water and 0.1 M NaCl solution. The adsorbed amount of PEI was estimated after 1, 2, 3, 5 and 7 days of adsorption time. 3. Results The effect of polyethyleneimine on the stability of SiC dispersions in plating solution is first discussed and then, the quantification of the polymer adsorption and DLVO calculations are employed to explain the phenomena of particulate stabilization at high ionic strength. 3.1. Particle characterization Morphology and size of the SiC micro- and nanoparticles were characterized using TEM. The microparticles are cubes of about 1 lm size (Fig. 1a), while nanoparticles have spherical shape with size of about 50 nm (Fig. 1b). The surface areas of particles specified by the supplier were 11.5 m2/g and 70–90 m2/g for microparticles and nanoparticles respectively. Raman spectroscopic analysis shows that the SiC phase of both the particles is a zinc-blende 3C (ß) phase. This phase is characterized by a peak at 780–800 cm 1 (Fig. 2). In addition, the Raman spectra display peaks at 1350 and 1590 cm 1 due to polycrystalline graphite [27]. The relative intensities of the graphite peaks
50
A. Vilinska et al. / Journal of Colloid and Interface Science 423 (2014) 48–53
Fig. 1. TEM micrographs of (A) SiC microparticles and (B) SiC nanoparticles.
25000 SiC Microparticles SiC Nanoparticles
15000
10000
5000
0 100
600
1100
1600
2100
2600
Raman shift (cm-1) Fig. 2. Raman spectra of SiC microparticles (dashed) and SiC nanoparticles (full).
Regime I
Dispersed volume (%)
100%
Regime II
Regime III
95% 90% 85% 80% 75%
3.3. Dispersion of nanoparticles vs microparticles
70% 0
5
10
25
50
100
150
200
500
1000
PEI concentration (ppm) Fig. 3. Effect of PEI 60 concentration on dispersion of SiC microparticles in Watts solution after 3 min of settling.
are higher in the case of nanoparticles as compared to microparticles, suggesting a higher content of graphite in the former. XPS reveals oxidized SiC species on the surface of both particle types such as SiO2 and SiOxCy with a Si 2 p binding energy of 104.4 eV and 103.1 eV, respectively. The C 1s photoemission contains a peak at 285 eV of graphite [9,28]. 3.2. Effect of polymer concentration and molecular weight High electrolyte concentration of the Watts solution shields the electrostatic repulsion between microparticles and allows their quick settling without added polymer (Fig. 3). When 60,000 MW polyethyleneimine is added to the system, three regimes are observed depending upon the polymer concentration. The first one,
The electro-codeposition of nanoparticles can lead to superior film properties [29,30]; therefore nanoparticles of SiC were included in our studies. Dispersion tests of nanoparticles, were
100%
Dispersed volume (%)
Intensity (cnt)
20000
at low PEI concentrations, shows a slight increase in the dispersion stability. In the second regime, at intermediate PEI concentrations, dispersion stability is worse and the settled volumes were higher. The third regime, above 500 ppm of PEI, stabilization is high with the dispersed volumes increased significantly. The first two regimes are considered to be due to insufficient coverage of particles with the polymer. In fact, if the polymer can form sufficiently long tails extending from particle to particle, it may induce bridging flocculation [25], as seen at intermediate concentrations. However at the lowest concentrations of PEI, polymer does not cause bridging flocculation leaving the system more dispersed. At the highest concentration (regime III), the system is possibly stabilized due to complete coverage of particles with the polymer, with the polymer extending into the solution and creating a steric barrier between the particles. In addition to PEI concentration, another important parameter in achieving sterically stabilized dispersions is the molecular weight of the polymer. Settling experiments were therefore performed using a linear 1200 MW polyethyleneimine (Fig. 4). The stabilizing effect of PEI 1.2 for microparticle systems is lower than that of PEI 60. Dispersion and settling are unaffected up to 500 ppm, above which a slight improvement can be seen. The effect is however much weaker than with the PEI 60, with the particles dispersed after 3 min at 1000 ppm concentration. The PEI 1.2 either does not adsorb on the particles, or the adsorbed polymer does not extend sufficiently into the bulk solution to cause steric stabilization.
95% 90% 85% 80% 75% 70% 0
10
50
100
500
1000
5000
PEI concentration (ppm) Fig. 4. Effect of PEI 1.2 concentration on dispersion of SiC microparticles in Watts solution after 3 min of settling.
51
A. Vilinska et al. / Journal of Colloid and Interface Science 423 (2014) 48–53 100%
Dispersed volume (%)
Dispersed volume (%)
100% 95% 90% 85% 80% 75%
90%
80%
70% 0
5
10
25
50
100
150
200
500 1000 2000 5000
70% 0
Concentration (ppm)
10
50
100
500
1000
PEI concentration ppm
conducted with samples prepared identically to microparticle dispersions with PEI 60, and the results after 120 min are given in Fig. 5. The prepared nanoparticulate dispersions are naturally more stable than the microparticle dispersions, due to a much smaller particle size of the former. Some settling and clearing is observed after 120 min for all systems below 200 ppm PEI 60. The nanoparticulate dispersions are completely stabilized at 1000 and 5000 ppm. In contrast, low molecular weight PEI has a destabilizing effect on SiC nanoparticle dispersion as shown in Fig. 6. Within the concentration range studied no significant improvement of stability is found compared to the sample without any polymer. At concentrations above 500 ppm, the PEI 1.2 acts as a flocculant. It is possible that the stabilization concentration is over 5000 ppm, but since such high concentrations of PEI 1.2 showed a detrimental effect on electroplating [31] those concentrations were not studied.
Fig. 6. Effect of PEI 60 concentration on dispersion of SiC nanoparticles in Watts solution after 120 min of settling.
100%
Conventional (PEI in solution) Precoated with PEI (resuspended)
Dispersed volume
Fig. 5. Effect of PEI 60 concentration on dispersion of SiC nanoparticles in Watts solution after 120 min of settling.
95%
90%
85%
80% 0.04
1
2
3
5
7
Aging time (days) Fig. 7. Effect of aging time and coating protocol on SiC microparticles dispersions stabilized with 1000 ppm PEI 60 in Watts solution after 5 min of settling.
3.4. Effect of adsorption time and particle precoating in Watts solution 3.5. Polymer adsorption To understand the mechanisms of dispersion in Watts solution, we studied the effect of ionic strength and polymer concentration. From the electroplating perspective, it is important, that only a small portion of (5–10%) of the loaded PEI 60 amount is adsorbed on the SiC microparticles both with water and 0.1 NaCl. The adsorption of PEI increases with PEI concentration (Fig. 8), with coverage of microparticles reaching a maximum of 0.2–0.3 mg/ m2 at 500–1000 ppm of PEI in solution. The adsorption of PEI on SiC is higher in 0.1 M NaCl than in deionized water. Surface saturation is reached at between 500 and 1000 ppm of PEI 60. The Langmuir linear regression gives a straight line indicating monolayer
0.5
Microparticles water
0.45
PEI adsorbed (mg/m2)
Electroplating tests show that large concentrations of PEI can lead to undesired plating conditions, presumably because PEI also adsorbs on Ni, impacting charge-transfer mechanisms [31]. Therefore a scheme was developed to remove the residual polymer in solution by centrifugation, decantation and replenishment of the solution. This process is henceforth called precoating. The method improved the efficiency of plating [31]. To test the effect of the precoating methodology on the dispersion, adsorption for a prolonged time as well as the aging of redispersed precoated particles in Watts solution was tested. Fig. 7 compares the dispersed volume of SiC microparticles stabilized with 1000 ppm PEI 60 conventionally and with precoating: the first data set is the result of dispersion of SiC in Watts bath and PEI in solution and the second data set is for a dispersion of SiC particles precoated with PEI and resuspended in fresh Watts solution. The aging effect was tested on both dispersions for duration of 7 days. Initially the settled volumes of both dispersions are similar, which suggests that the centrifugation and solution replacement does not have a deleterious effect on stability. With an increase in time, the dispersion volume of the conventional dispersion increases while that of the precoated one remains stable. As there is some residual polymer present in the solution, in the former case some additional adsorption may take place and further improve the stability. However, such a process is slow. The precoated dispersion exhibits constant stability of 91–92% over a period of several days, suggesting the precoating protocol to be a viable option for use in the codeposition of microparticles. However, when the same protocol is applied to a dispersion of nanoparticles, the resuspended dispersion settles quickly (data not shown), thus this protocol can be concluded not to be suitable for processing nanoparticles.
Microparticles 0.1 M NaCl
0.4
Nanoparticles water
0.35
Nanoparticles 0.1 M NaCl
0.3 0.25 0.2 0.15 0.1 0.05 0 1
10
100
1000
Equilibrium PEI concentration (ppm) Fig. 8. Effect of NaCl on adsorption of 60,000 PEI onto SiC microparticles and nanoparticles relative to equilibrium concentration in water. Lines are drawn for the convenience of the reader.
A. Vilinska et al. / Journal of Colloid and Interface Science 423 (2014) 48–53 0.30
40
0.25
30
Electrostatic interaction LVDW SiO2 LVDW ZrC
20
LVDW Carbon
0.20 10
0.15
G (kT)
PEI adsorbed (mg/m 2)
52
0.10
0 0
2
4
6
8
10
12
14
16
18
20
-10
0.05
-20 -30
0.00 0
1
2
3 4 5 Adsorption time (days)
6
Attraction
7
Fig. 9. Effect of adsorption time on the adsorption of 60,000 PEI onto SiC microparticles. Line is drawn for convenience of the reader.
coverage with a Cmax of 0.3 mg/m2 and 0.17 mg/m2 in 0.1 M NaCl and deionized water respectively. The leveling off the curves indicates close to complete monolayer coverage at 500–1000 ppm initial concentration of PEI 60, and 20–45% surface coverage at intermediate initial PEI concentrations of 50–100 ppm. The adsorption of PEI 60 on nanoparticles is found to be negligible in the concentration range studied (Fig. 8). At the highest concentration of PEI 60 of 1000 ppm the adsorption density is below 0.01 mg/m2. According to literature, PEI preferentially adsorbs onto oxidized surface of SiC [32] with available silanol and siloxane groups. The strong H-bonding interaction between amino groups and surface silanes is well known [33]. Since SiO-like species are present on both surfaces, PEI was expected to adsorb on both micro and nanoparticles. However, the comparatively higher graphite content of nanoparticles is possibly responsible for the reduced number of available adsorption sites and the lower adsorption capacity. Interestingly, the amount adsorbed onto microparticles increases with time by factor of 1.5 after 3 days (Fig. 9). We attribute the higher stability of these particles in Watts solution at 1000 ppm PEI after 3 days to higher adsorption of the polymer.
-40
Separation distance (nm)
Fig. 10. Theoretical evaluation of interparticle (electrostatic and LVDW) forces between two SiC microparticles as a function of separation distance at the ionic strength of Watts solution.
tive LVDW forces. A rough estimate of polymer size based on molecular weight can be done with the following formula: 0:06M 1=2 w , which gives 14.69 nm for the PEI 60 and 2 nm for PEI 1.2 respectively. Assuming monolayer coverage, the PEI 1.2 cannot create a sufficient separation distance between the microparticles and is not suitable as a dispersant under high ionic strength conditions. PEI 60 can create the required 10–15 nm thick monolayer on the surface of SiC microparticles and prevent aggregation, if the coverage is complete. In the case of partially covered particles, attractive H-bond interactions between the particles and PEI can cause aggregation as observed for regime II in Fig. 2. For the nanosized particles the repulsive electrostatic forces can extend to 0.5 nm and the LVDW forces are attractive up to a distance of 2 nm from the nanoparticle surface (results not shown). Nanoparticles can be effectively dispersed with a much thinner adsorbed steric layer of 1–2 nm. To summarize, the theoretical calculations suggest that steric stabilization of silica like surfaces in high ionic environment requires the thickness of adsorbed polymer of 10 nm and more.
3.6. Inter-particle forces evaluation 4. Discussion To quantify the dominant interparticle forces between two SiC particles, DLVO-based calculations were performed. Electrostatic and Lifshitz van der Waals interaction energies were calculated from electrophoretic data and published data of particles with similar composition. The zetapotential of SiC particles was measured with the electrophoretic method in 0.01 M NaCl and calculated using the Smoluchowski equation to be around 6 mV (±3 mV) at pH of 3.5 and this value is used in our calculations. The electrostatic interactions are typically strong and long range for charged particles at low ionic strength. In 2 molar solutions the calculated double layer thickness is less than 1 Å and the electrostatic repulsive force is close to zero at distances higher than 1 nm from the surface (Fig. 10). Lifshitz van der Waals (LVDW) interactions were calculated using a Hamaker constant calculated from the dispersive components of surface energy [34] of the surface species found on SiC particles by XPS (carbon, SiO2) or particles with similar composition [35–37]. For all materials, the LVDW interactions below the distance of 10 nm are sufficiently attractive to predominate over the electrostatic repulsion to cause aggregation. Therefore, to prevent aggregation in this case the steric stabilizing layer has to be thicker than the distance below which the LVDW interactions dominate. Based on the interaction energy values calculated, SiC microparticles have to have a continuous coverage with steric stabilizer of a 10–15 nm thickness in order to prevent aggregation due to attrac-
In order to keep SiC particles dispersed in Watts plating solution, steric stabilization is necessary, and this can be achieved using a cationic polymer. PEI was chosen for the present study based on the positive results found in the literature and our initial screening tests with different cationic and nonionic polymers. The stabilization effect of PEI depends on polymer concentration, molecular weight as well as the particle size and surface composition. The higher molecular weight polymer PEI 60 was more successful in stabilization than the lower molecular weight PEI 1.2. The reason that, the 60,000 MW PEI can create a polymer layer thicker than the separation distance required to overcome the LDVW attractive forces, while the 1200 MW polymer can create a steric repulsion layer of only a few nm. In contrast, to achieve similar stabilization with PEI 1.2 as with PEI 60, a 5–10 multilayer adsorption would be required, which is impractical. Adsorption of PEI 60 was found to be the main cause of stabilization of dispersions with microparticles with two important parameters: surface coverage and thickness of adsorbed layer. In particular, stabilization occurred only in the concentration ranges with recognizable surface saturation and close to a complete surface coverage. PEI 60 at intermediate concentrations of 50– 200 ppm covers the particles only partially, which can lead to bridging between two particles causing aggregation. The second important characteristic of the steric layer is the thickness neces-
A. Vilinska et al. / Journal of Colloid and Interface Science 423 (2014) 48–53
sary to overcome the LVDW attraction between the particles. We estimated PEI 60 coverage, assuming that the polymer species have spherical shape on the basis of Analytical centrifugation of polymer solution (data not reported). As an estimate of the polymer species diameter we took the value of 14.7 nm which was calculated as described in Results section. Under these assumptions and using the adsorption data, we obtained polymer surface coverage of 23% and 49% for microparticles in deionized water and NaCl, respectively. Since both systems exhibited surface saturation, it can be proposed that, the polymer species presumably adsorb in a flatter arrangement rather than as spheres, leading to a thinner polymer layer. The adsorption amount from NaCl solution was larger than that from deionized water suggesting that the background electrolyte favors a thicker polymer arrangement on the surface of particles. The presence of electrolyte can thus influence the adsorbed amount as well as the alignment of PEI polymer coils on the interface similarly as pH [38]. It was also observed that, PEI adsorption from Watts solution is strong enough for the dispersion of microparticles to remain stable even after redispersion in Watts solution without any PEI. The dispersion stability of the resuspended PEI precoated microparticles was constant over the period of 7 days, indicating a stable polymeric layer adsorbed on the particles. On the other hand, the nanoparticle dispersions are destabilized after polymer removal from solution and resuspension. The instability of nanoparticle dispersions upon resuspension is attributed to an insufficient steric layer as a result of low PEI adsorption. This may be due to a different mechanism such as depletion stabilization being operative for stabilization, or weak physisorption of polymer on the nanoparticle surface and desorption after resuspension. In the case of the depletion stabilization [39], the polymer has to be in solution in order to disperse the particles and prevent aggregation. Adsorption on nanoparticles was found to be low and 1/20 in magnitude of microparticles. The thickness of such layer is then less than the required 1 nm, or the particles are not covered completely and, hence the particles can aggregate and settle after resuspension. Importantly, the stability of microparticle dispersions is enhanced if the adsorption time is extended to several days with the increase in the adsorption densities correlating with the increase in the dispersion stability. Extended adsorption time may allow the adsorption of additional polymer coils on the particle surface or replacement of the shorter polymeric chains with the longer ones [40], both leading to an increase of steric layer thickness and thus efficient particle dispersion. 5. Conclusions We studied PEI as a dispersant for SiC microparticles and nanoparticles in a high electrolyte Watts plating solution at pH 3.5. PEI of 60,000 MW is an excellent dispersant for both the particle types. We determined dosage of the polymer required to disperse these particles. We demonstrated that the dispersion stability is controlled by the adsorption density of the polymer, polymer layer thickness, and the extent of Lifshitz van der Waals (LVDW) interaction between the particles. Therefore, to retard particle aggregation due to (LVDW) interactions when repulsive electrostatic forces are screened out, it is necessary to have a close to complete coverage of the particles with a thick enough polymer layer. We found that the adsorption density and surface saturation level of PEI depend on the ionic strength of the solution. We revealed that the ceramic particles are dispersed by two different mechanisms depending on the surface properties of SiC particles and their size. The microparticles are dispersed due to the buildup of a steric polymeric
53
layer of a sufficient thickness. In contrast, PEI does not adsorb on the nanoparticles suggesting the dispersion stabilization by the free polymer in solution, known as depletion stabilization. We also found that the adsorbed layer of PEI on SiC microparticles is stable when the particles are coated with the polymer and then redispersed in Watts solution. This finding is important for attaining dispersion in systems sensitive to polymer content such as electroplating. Acknowledgment Financial support from The Boeing Company is gratefully acknowledged. References [1] H.-K. Lee, H.-Y. Lee, Surf. Coat. Technol. 201 (2007) 4711–4717. [2] L. Benea, P.L. Bonora, A. Borello, S. Martelli, F. Wenger, P. Ponthiaux, J. Galland, J. Electrochem. Soc. 148 (2001) C461–C465. [3] C.T.J. Low, R.G.A. Wills, F.C. Walsh, Surf. Coat. Technol. 201 (2006) 371–383. [4] E. García-Lecina, I. García-Urrutia, J.A. Díez, M. Salvo, F. Smeacetto, G. Gautier, R. Seddon, R. Martin, Electrochim. Acta 54 (2009) 2556–2562. [5] N.K. Shrestha, K. Sakurada, M. Masuko, T. Saji, Surf. Coat. Technol. 140 (2001) 175–181. [6] N.K. Shrestha, M. Masuko, T. Saji, Wear 254 (2003) 555–564. [7] T. Borkar, S.P. Harimkar, Surf. Coat. Technol. 205 (2011) 4124–4134. [8] D. Soccol, J. Martens, S. Claessens, J. Fransaera, J. Electrochem. Soc. 158 (2011) D515–D523. [9] R.P. Socha, K. Laajalehto, P. Nowak, Colloids Surf., A 208 (2002) 267–275. [10] R.P. Socha, K. Laajalehto, P. Nowak, Surf. Interface Anal. 34 (2002) 413–417. [11] R.P. Socha, P. Nowak, K. Laajalehto, J. Väyrynen, Colloids Surf., A 235 (2004) 45–55. [12] L. Stappers, J. Fransaer, J. Electrochem. Soc. 153 (2006) C472–C482. [13] M. Sarret, C. Muller, A. Amell, J. Nanopart. Res. 9 (2007) 1073–1080. [14] B.J. Palla, D.O. Shah, J. Colloid Interface Sci. 256 (2002) 143–152. [15] C. Xiao, L. Gao, M. Lu, H. Chen, L. Guo, L. Tao, Colloids Surf., A 355 (2010) 104– 108. [16] A. Amell, C. Muller, M. Sarret, Surf. Coat. Technol. 205 (2010) 356–362. [17] M.-D. Ger, Mater. Chem. Phys. 87 (2004) 67–74. [18] K.H. Hou, M.D. Ger, L.M. Wang, S.T. Ke, Wear 253 (2002) 994–1003. [19] B.S. Necula, I. Apachitei, L.E. Fratila-Apachitei, C. Teodosiu, J. Duszczyk, J. Colloid Interface Sci. 314 (2007) 514–522. [20] J. Zhang, Q. Xu, F. Ye, Q. Lin, D. Jiang, M. Iwasa, Colloids Surf., A 276 (2006) 168– 175. [21] X. Zhu, T. Uchikoshi, T.S. Suzuki, Y. Sakka, J. Am. Ceram. Soc. 90 (2007) 797– 804. [22] G. Xu, L. Xu, S. Pan, G. Song, China Particuol. 2 (2004) 182–184. [23] F. Tang, T. Uchikoshi, K. Ozawa, Y. Sakka, J. Ceram. Soc. Jpn. 113 (2005) 584– 587. [24] W.K. Mekhamer, N. Al Andis, M. El Shabanat, J. King Saud Univ. (Sci.) 21 (2009) 125–132. [25] C.P. Whitby, P.J. Scales, F. Grieser, T.W. Healy, G. Kirby, J.A. Lewis, C.F. Zukoski, J. Colloid Interface Sci. 262 (2003) 274–281. [26] T. Kakui, T. Miyauchi, H. Kamiya, J. Eur. Ceram. Soc. 25 (2005) 655–661. [27] M. Yoshikawa, N. Nagai, Vibrational spectroscopy of carbon and silicon materials, in: Handbook of Vibrational Spectroscopy, John Wiley & Sons, 2006, pp. 2593–2620. [28] C. Onneby, C.G. Pantano, J. Vac. Sci. Technol., A 15 (1997) 1597–1602. [29] L. Benea, P.L. Bonora, A. Borello, S. Martelli, Wear 249 (2001) 995–1003. [30] X.J. Sun, J.G. Li, Tribol. Lett. 28 (2007) 223–228. [31] D. Eroglu, A. Vilinska, P. Somasundaran, A.C. West, J. Electrochem. Soc. 160 (2013) D35–D40. [32] J. Sun, L. Gao, J. Eur. Ceram. Soc. 21 (2001) 2447–2451. [33] I.V. Chernyshova, Langmuir 16 (2000) 8071–8084. [34] A. Vilinska, H.K. Rao, Miner. Metall. Process. 28 (2011) 151–158. [35] H.-J. Jacobash, K. Grundke, E. Mader, K.-H. Freitag, U. Panzer, Application of surface free energy concept in polymer processing. Contact Angle, Wettability and Adhesion, in: K.L. Mittal (Ed.), 1993, pp. 921–936. [36] C. Rulison, Two component surface energy characterization as a predictor of Wettability and Dispersability, Kruss Application/Technical, Notes, AN213, 2000, pp. 1–22. [37] K. Bobzin, Benetzungs und Korrosionsverhalten von PVD-beschichteten Werkstoffen fur den Einsatz in umweltvertraglichen Tribosystemen, Thesis (Werkstoffwissenschftl. Schriftenreihe, Bd. 39), 3-8265-7414-1, 2000. [38] B.P. Singh, J. Jena, L. Besra, S. Bhattacharjee, J. Nanopart. Res. 9 (2007) 797–806. [39] R.I. Feigin, D.H. Napper, J. Colloid Interface Sci. 74 (1980) 567–571. [40] P. Frantz, S. Granick, Phys. Rev. Lett. 66 (1991) 899–902.
