Effect of Processing Parameters on the Electrophoretic Deposition of

Jan 31, 2011 - ... Parameters on the Electrophoretic Deposition of Carbon Black Nanoparticles in Moderately Viscous Systems ... *Telephone: 978-934-34...
0 downloads 0 Views 2MB Size
ARTICLE pubs.acs.org/Langmuir

Effect of Processing Parameters on the Electrophoretic Deposition of Carbon Black Nanoparticles in Moderately Viscous Systems Satyam Modi, Ming Wei, Joey L. Mead, and Carol M. F. Barry* NSF Nanoscale Science and Engineering Center for High-rate Nanomanufacturing, University of Massachusetts Lowell, Lowell, Massachusetts 01854, United States

bS Supporting Information ABSTRACT: Polymer-melt-based manufacturing processes for nanostructures offer high-rate, environmentally friendly, and commercially viable alternatives to solution-based methods. In this work, electrophoresis of a model carbon black and polystyrene system with moderate viscosity was used to investigate the viability of adapting nanoassembly processes to the high viscosity environment of polymer melts. The presence of polystyrene did not prevent deposition of carbon black, but deposition rates decreased at shorter deposition times; deposition was not linear with increasing applied voltage; and greater solution concentrations reduced the critical voltages (i.e., the voltage at which the rate of deposition changed). X-ray photoelectron spectroscopy (XPS) results and comparison of experimental data with Hamaker’s model showed that about 1.6% of the available polystyrene was initially deposited with the carbon black. At voltages above the critical voltage, the deposited mass was less than the Hamaker prediction, indicating the formation of electrically insulating layers on the electrodes. The overall behavior suggests that polymer melt-based processes could be employed for high-rate fabrication of nanooptical devices, biochemical sensors, and nanoelectronics.

’ INTRODUCTION Precise assembly of nanoparticles on micrometer- and submicrometer-scale patterns is essential in a range of applications, including nanoelectronics,1 nano-optical devices,2 biochemical sensors,3,4 and high-density magnetic data storage devices.5 Nanoparticles have been assembled using drying-mediated,6,7 chemically directed,8-10 template-assisted,11,12 magnetic-fieldassisted,13 lithography and printing,14-16 and liquid interfacial assembly17 processes. Although these methods have permitted assembly of nanoscale materials, many of these processes are complex, involve multiple steps, are slow, costly, and low resolution, and are unable to provide control of interparticle spacing. In contrast, electrophoretic deposition (EPD) can be used with a wide range of materials and substrate shapes, provides short formation times, requires no binder burnout for ceramics,18 and produces uniform deposition with controlled thickness and low levels of contamination. Since EPD is fast, is cost-effective, and can be automated as a continuous process on an industrial scale,19 the process has long been used in processing of ceramics, coatings, and composite materials to deposit metals, glasses, phosphors, inorganic and organic paints, and rubber latex from both aqueous and nonaqueous media onto unpatterned substrates.20 Recently, EPD has r 2011 American Chemical Society

been used for the deposition of nanomaterials, such as colloidal gold particles,21 silica nanoparticles,22 polystyrene particles,23-25 cadmium selenide nanoparticles,26 carbon nanotubes,27,28 and polyaniline29 on patterned substrates. The EPD process has been widely used in material processing and has been the subject of theoretical and experimental research. From theoretical analysis of the movement of particles in a medium under the influence of an electric field and the mass conservation principle, Hamaker30 derived an expression for the mass, m, deposited on an unpatterned substrate: Z t2 ð1Þ m ¼ μEAC dt t1

where μ is the electrophoretic mobility of the particles, E is the electric field strength, A is the substrate surface area, C is the mass concentration of conductive particles in the suspension, and t is the deposition time. Avgustnik et al.31 correlated the deposited mass with the process parameters in the EPD process. Using zirconia and yttria-stabilized zirconia particles, Basu et al.32 and Received: October 28, 2010 Revised: December 28, 2010 Published: January 31, 2011 3166

dx.doi.org/10.1021/la1043359 | Langmuir 2011, 27, 3166–3173

Langmuir

ARTICLE

Table 1. Measured and Calculated Parameters of the Solutions 1/κ

mix

sol. conc.

[CCB]

[CPS]

(wt %)

(w/v%)

(mg/mL)

(mg/mL)

d (nm)

η (mPa 3 s)

pH

μ (10-9 m2/(V s))

(nm)

ζ (mV)

σ (μS/cm)

q (10-4 nm-2)

1

1

0.1

10

222

4.0

8.5

4.7

53

-48

21

1.8

0.2

10

336

4.1

8.5

4.9

53

-50

25

1.2

0.3

29

347

6.5

8.3

1.7

42

-15

29

0.4

0.6

29

385

6.6

8.4

1.4

48

-14

31

0.3

0.5

47

431

8.0

8.4

1.2

50

-13

30

0.3

1.0

47

432

8.2

8.3

0.8

43

-8

32

0.2

2 1

3

2 1 2

5

Chen and Liu,33 respectively, found experimentally that the deposited mass exhibited a nonlinear increase with deposition time but a linear increase with applied voltage. The same behavior was observed with hydroxyapatite particles using EPD34 and tin oxide nanoparticles using dielectrophoretic deposition.35 Wang et al.36 observed that the thickness of the deposited particles initially increased linearly with deposition time, but with additional time the deposition rate decreased. For long deposition times, the thickness remained constant. Deposition in these unpatterned systems was consistent with Hamaker’s prediction, that is, eq 1. To date, the effect of process parameters on the deposition behavior using patterned substrates has not been examined. Viscosity is one of the main controlling parameters in the EPD process because the electrophoretic mobility of the particles is inversely proportional to the solution viscosity.18 Electrophoretic mobility, however, also depends on the on the Debye length of the suspended particles;18 thus, the measured electrophoretic mobility is related to the system’s zeta potential using the Huckel equation37 and the Smoluchowski equation.38,39 In most EPD processes, the solids level has been very low and the viscosity has had little effect on deposition kinetics.40 As a result, little work has been done to understand the effect of viscosity on EPD.18 Jean41 did study the effect of adding water to a suspension of aluminum oxide particles and silicon carbide whiskers in isopropanol. He found that mass deposited during EPD decreased with increasing water concentration. This behavior was attributed to the increasing viscosity and decreasing stability of the suspension with increasing water concentration. EPD is well-suited for high-rate, high-volume template-directed assembly of nanoparticles suspended in low viscosity fluids such as solvents.22 It would be highly desirable to utilize this process to assemble nanoparticles from polymer melts (which are highly viscous at normal processing temperatures). Polymermelt-based manufacturing methods, such as injection molding, are solventless high-rate processes. Using this approach, a nanostructured surface could be obtained directly in a one-step process instead of requiring a two-step process where nanoparticles are deposited on the surface of a previously formed part. For this application, understanding the effects of viscosity on solution parameters, processing conditions, and deposition kinetics is necessary. Toward this goal, carbon black was assembled on microscale templates from moderately viscous solutions. The effect of applied voltage, deposition time, and solution viscosity were investigated for their effects on the surface coverage of the patterns and the height of the deposits. Mass deposited on the patterned surface as a function of applied voltage was determined experimentally and compared to the mass deposition predicted using the Hamaker equation. The deviation from the theoretically predicted linearity of deposited mass was also investigated.

’ EXPERIMENTAL SECTION Materials. A model system consisting of carbon black particles and polystyrene in tetrahydrofuran (THF) was employed in this study. A polystyrene with a measured weight average molecular weight of 224 000 g mol-1 (Styron 615APR, Dow Plastics), carbon black (CB N330, Cabot Corporation) with a reported average particle size of 2630 nm42 (aggregate sizes are larger), and THF (99.9% reagent grade, Sigma-Aldrich Co.) were used as received. The polystyrene and carbon black were melt compounded using a miniaturized internal batch mixer (Brabender, model PL2200) to create mixtures with 1 and 2 wt % carbon black; these systems are denoted as 1 wt % and 2 wt %, respectively. During compounding, the fill factor was 0.7, the chamber temperature was 200 °C, the rotor speed was 60 rpm, and the mixing time was 11 min. Each polystyrene/carbon black mixture was dissolved in THF to produce three solutions with concentrations of 1, 3, and 5 w/v% carbon black/polystyrene mixture. The solution was stirred for 24 h using a magnetic stirrer followed by an additional 5 min of mixing at 3540 rpm using a high-speed mixer (model SpeedMixer, Hauschild Engineering). Table 1 presents the calculated concentrations of carbon black, [CCB], and polystyrene, [CPS], for each solution. The viscosity, η, of the solutions was measured using a Brookfield Viscometer (model: LVT) with spindle number 2 at 60 rpm. The average carbon black aggregate size, d, was determined using transmission electron microscopy (model: EM400T, Philips); since carbon black particles do not exist as a discrete entity, but rather are fused into aggregates, the smallest entity is the aggregate.42 The solution conductivity, σ, and electrophoretic mobility, μ, were measured using a zetasizer (Zetasizer Nano series, model ZEN3600, Malvern Instruments) which employs a combination of laser Doppler velocimetry and phase analysis light scattering in a patented technique called M3-PALS to measure particle electrophoretic mobility.43 In order to calculate the zeta potential, the double layer thickness, 1/κ, was determined using44   1 εεo KT 1=2 ¼ ð2Þ k 2000e2 INA where ε is the dielectric constant of the solvent (7.5), εo is the vacuum permittivity (8.854  10-12 A2 s4 kg-1 m-3), K is the Boltzmann constant (1.38  10-23 J K-1), Τ is the temperature in Kelvin (298 K), e is the elementary charge (1.602  10-19 C), I is the ionic strength in molarity, and ΝA is Avogadro’s number (6.022  10-23 mol-1). The ionic strength, I, was calculated from the pH values measured for the solution,45 that is, I ¼ ½Hþ  ¼ 10-pH

ð3Þ

For the first solution (the 1 w/v % solution of the 1 wt % carbon black/ polystyrene mixture) in Table 1, the pH was 8.5, the calculated ionic strength was 3.16  10-6 mol m-3, and the calculated double layer thickness was 53 nm. The Henry function, f(κa), where a is the particle radius, was then determined using a graphical representation of the 3167

dx.doi.org/10.1021/la1043359 |Langmuir 2011, 27, 3166–3173

Langmuir

ARTICLE

Figure 1. AFM image of carbon black (CB) deposition height for a 0.1 mg/mL solution of carbon black using an applied voltage of 70 V and deposition time of 10 s. Henry function as a function of κa values.46 When the particle size was assumed to be the aggregate size, the κa value for the first solution was 2 and the Henry function was 1. The Henry function was also calculated based on the primary carbon black particle sizes; this method produced a κa value of 1 and a Henry function of 1. Assuming the particle size is the aggregate size, for all solutions the κa values ranged from 2 to 5, producing Henry functions of 1 to 1.2. Therefore, the Henry function was reduced to the Huckel limit (f(κa) = 1). As a result of these calculations and considering the apolar nature of the solvent, zeta potentials shown in Table 1 were calculated using the Huckel equation:37 2εεo ζ ð4Þ μ¼ 3η where ζ is the zeta potential. The number of charges per unit area of carbon surface, q, was calculated using the first order approximation assuming that the particles are spheres:47 2ζεεo ð5Þ q¼ ed As shown in Table 1, the conductivity of the solutions increased with carbon black concentration, whereas the measured viscosity increased primarily with polystyrene content. The electrophoretic mobilities, and the resultant zeta potentials (-8 to -50 mV), generally decreased with increasing solution concentration and were lower than the -88-mV zeta potential measured for a 0.1 mg/mL solution of carbon black in THF (the zeta potential measured for polystyrene in THF was negligible). The decrease in electrophoretic mobility was attributed to the solution viscosity, partial covering of surface of the carbon black particles with polystyrene, and competition between the carbon black and polystyrene for ions (as has been observed with polymer systems such as toners).48

Template-Directed Assembly of Carbon Black from a Moderately Viscous System. The templates used in this research were an interdigitated pattern of copper foil on a polyimide substrate (14 mm  14 mm). The width of the electrode finger and the spacing between two neighboring electrodes were 75 and 100 μm, respectively. The carbon black was assembled onto patterned surfaces (templates) using electrophoretic deposition. The template was first rinsed five times with acetone and blow dried with nitrogen. The template was then connected to the DC power source and immersed into the carbon-blackcontaining solutions using a dip coater with a vertical speed of 86 mm/ min. The template was lifted from the solution at the same rate. The applied voltage was varied from 10 to 70 V at intervals of 10 V, and for each voltage the deposition times were 10, 30, and 60 s. The template was dried in air after the assembly process. The procedure used to assemble carbon black onto the template is illustrated in Supporting Information Figure S1.

Characterization of Electrodeposited Material. The surface morphology of the deposited carbon black particles was observed using scanning electron microscopy (SEM) (model: JSM-7401F, JEOL). The surface chemical composition of the deposited electrode was investigated using X-ray photoelectron spectroscopy (XPS) (VG Scientific ESCALAB MK II). A Mg KR X-ray source was used at a power of 200 W. A Zeiss Discovery stereomicroscope (model V20) equipped with an Achromat S 1.0 objective and AxioCam HRc camera was used to obtain 2D optical images of assembled templates at 32.5 magnification. These images were analyzed using image analysis software (Scion Image, Scion Corporation) to calculate the percentage surface coverage of the electrode. The percentage area coverage, AC, of the electrode was calculated as ACB  100 ð6Þ AC ¼ Ao where ACB is the area on the electrode occupied by the assembled carbon black particles and Ao is the area of the electrode. The height of the deposited layer was determined using atomic force microscopy (AFM) (PSIA, model: XE-150). As shown in Figure 1, a scratch was made perpendicular to the long axis of the electrode to produce a sharp delineation between the deposited material and the electrode. Noncontact mode measurements of these samples were performed using a scan size of 40 μm  40 μm and a scan rate of 0.20 Hz. The reported height of the deposited layer was determined by averaging the height measured from three separate locations on a particular sample; there were five measurements for each location.

’ RESULTS AND DISCUSSION Electrophoretic Assembly. SEM micrographs of the template after EPD of carbon black from a moderately viscous solution of polystyrene in THF shows that, as expected, when the electric field was applied, the negatively charged carbon black deposited only on the copper fingers connected to the positive electrode (the SEM image is in Supporting Information Figure S2). The coverage of the electrode and the height of deposition with carbon black, as will be discussed in more detail later, increased with applied voltage, deposition time, and solution concentration. Prior to complete (100%) coverage of the electrodes, carbon black began to deposit between the electrodes; eventually the entire surface of the template was covered. Effect of Applied Voltage. Figure 2a and b presents the SEM micrographs of a template after EPD using a deposition time of 60 s, applied voltage of 10 V, of a 0.1 mg/mL solution of carbon 3168

dx.doi.org/10.1021/la1043359 |Langmuir 2011, 27, 3166–3173

Langmuir

ARTICLE

Figure 2. Effect of applied voltage of the template-directed assembly using a deposition time of 10 s: (a,b) SEM micrographs for an applied voltage of 10 V, (c,d) SEM micrographs for an applied voltage of 30 V, (e,f) SEM micrographs for an applied voltage of 60 V of a 0.1 mg/mL solution of carbon black (CB), (g) plot of area coverage versus applied voltage, and (h) plot of the height of deposition versus applied voltage for a deposition time of 10 s of a 0.3 mg/mL solution of carbon black.

black. The templates subjected to applied voltages of 40 and 70 V are shown in Figure 2c-f, respectively. The amount of material deposited increased with applied voltage, a result of the enhanced electrophoretic force created by the higher voltages. At the lowest applied voltage (Figure 2a and b), carbon black particles preferentially deposited onto some of the areas of the electrode with no interconnection between deposited particles. This pattern of deposition occurred when the area coverage was 0-55% and is defined as underdeposition. With further increases in the applied voltage (Figure 2c and d), the carbon black particles continued depositing on the uncovered areas of the electrode while making good interconnection between particles. Complete deposition, in this research, was defined at when the area coverage of the electrode with carbon black was 55-80%. Increasing the voltage to still higher levels (Figure 2e and f) caused the carbon black to deposit (bridge) between the electrodes. This bridging between the electrodes increased with further increases in applied voltage; eventually carbon black bridged the repelling electrodes, and the

entire surface of the template was covered with carbon black. This pattern was defined as overdeposition and was found when the area coverage of the electrode was greater than 80%. Deposition between the electrodes may be caused by the higher applied electrical potential. The higher applied electric potential may cause turbulence and fast movement of the particles in the suspension.32 The turbulent flow in the suspension may have interfered with the deposition pattern of the particles and caused the unwanted deposition of the particles. This unwanted deposition pattern, once started, increased with increasing applied voltage, eventually covering the entire surface of the template. Figure 2g and h plots the area coverage of the electrode and the height of deposition with respect to applied voltage, respectively. The area coverage of the electrode and the height of deposition increased with applied voltage. The electric field strength strongly depends on the applied voltage and the distance between the electrodes. The velocity of the dispersed particles is directly proportional to the electric field strength. In this case, all geometric 3169

dx.doi.org/10.1021/la1043359 |Langmuir 2011, 27, 3166–3173

Langmuir features of the electrodes were identical; thus, the electric field strength was determined by the applied voltage. When the applied electric potential increased from 10 to 30 V, the electric field strength increased three times and electrophoretic force increased nine times, facilitating greater assembly of carbon black particles on the electrode. In previous work on constant voltage EPD of nanoparticle systems,25,32,33 the surface coverage and deposition height varied linearly with applied voltage; however, in this work, inflection points in the rate of area coverage and height of deposition were observed with applied voltage. For the case of area coverage, with a deposition time of 10 s, the inflection point occurred at ∼15 V. Similar nonlinear behavior was observed for deposition height, but there were two inflection points. The first inflection point was consistent with the inflection in the area coverage plot at ∼15 V. The second inflection point occurred at ∼50 V and correlated with the onset of overdeposition in area coverage. The height of deposition approached a plateau after second inflection point, but the area coverage did not. Moreover, the rate of particle assembly before and after the inflection points was significantly different. In the case of area coverage, the slopes of the area -voltage curve were 4% per V and 0.5% per V (with respective coefficients of determination, r2, of 0.99 and 0.96) before and after the inflection point, respectively. A similar difference was observed for deposition height with the slopes being 40 nm/V (r2 = 0.99) and 9 nm/V (r2 = 0.98) before and after the first inflection point, respectively. The change in deposition rate as well as the plateauing of the deposition height at applied voltages of 50 V or greater likely resulted from three phenomena. First, the coverage of the electrode surfaces by less conductive carbon black reduces the electric field strength and so hinders the assembly of additional particles on the electrodes.49 Due to the geometry of the patterned substrate, the change in field strength also promotes deposition between the wires, thus producing the overdeposition that occurred at high applied voltages. Second, strong polymer-filler interactions, such as the well-known attraction between carbon black and styrene containing polymers,50,51 would result in the polystyrene molecules traveling onto the template surface with the carbon black under the influence of the electric field. The deposited polystyrene molecules have a higher electrical resistance than carbon black particles, causing a rapid decrease in the rate of area coverage and deposition height with increasing applied voltage. Third, the local depletion of the particles near the electrodes would change the concentration and thus the deposition kinetics.48,49 XPS was performed to compare the amount of carbon black and polystyrene deposited onto the electrode before and after the inflection point. For this purpose, samples produced with applied voltages of 10 and 30 V using a deposition time of 10 s of a 0.3 mg/mL solution of carbon black were chosen. XPS analysis was also performed on a reference sample in which carbon black was assembled onto the template from a suspension of carbon black in THF. Figure 3 presents the XPS C1s spectra of these samples. The XPS spectra of carbon black sample as well as the 10 and 30 V samples showed the C1s peak at ∼285 eV. A peak fitting analysis of the C1s peak of all three samples was performed by fixing the full width at half-maximum (fwhm) range from 2.0 to 2.5 eV. The deconvoluted C1s peaks of the 10 and 30 V samples exhibited two distinct components: the aromatic carbon peak at ∼284.6 eV and the graphitic plus aliphatic carbon peak at ∼285.0 eV. The peak at ∼291.7 eV that appeared in the 10 and 30 V

ARTICLE

Figure 3. Mg KR XPS C1s spectra of carbon black, 10 and 30 V samples using a deposition time of 10 s of a 0.3 mg/mL solution of carbon black.

samples, but not in the carbon black sample, was due to π-π* shakeup satellite (which is associated with the presence of localized π electrons in the conjugated phenyl ring of polystyrene52,53). The shakeup satellite has been observed in carbon black samples,54,55 but since it was not present in the reference sample, the peak can be used as a clear an indicator for the presence of polystyrene in the electrodeposited material. The area ratios of the aromatic to aliphatic plus graphitic carbon peaks derived from the XPS C1s scan were used to estimate the ratio of carbon black and polystyrene present in the samples. The aromatic to aliphatic plus graphitic peak area ratios were 0.84 and 1.24 for the 10 and 30 V samples, respectively. Since the aromatic carbons are present only in the polystyrene, the increased aromatic to aliphatic plus graphitic carbon ratio suggests a greater amount of polystyrene was present in the 30 V sample (as compared to 10 V samples). Effect of Deposition Time. Increased deposition time produced the same deposition behavior (under, over, and complete) as observed with increasing applied voltage (the SEM images are shown in Supporting Information Figure S3). The overdeposition pattern, however, was only observed with higher applied voltages. Figure 4 shows (a) the area coverage of the electrode and (b) the height of deposition with deposition time. At constant applied voltage, the area coverage and the height of deposition increased in a nonlinear fashion with deposition time. The area coverage and deposition height increased rapidly with short deposition times (