Chapter 3 Experimental
3.5 Instrument
3.5.2 Characterization
Energy Dispersive X-ray (EDX)
EDX attached on a SEM was used to perform the semi-quantatative composition report and elemental mapping for the inverse opals.
High Resolution X-ray Diffractometer (HRXRD)
A HRXRD (Bede D1) was used to determine the crystallinity and relevant composition for the nickel inverse opals before and after thermal treatments.
Four-Point Measurement (4pp)
Keithaley 4200 to obtain the conductivity for the planar nickel inverse opals. For cylindrical nickel inverse opals, samples were fixed on a non-conductive glass with two silver bumps serving as the probe contact points, and a forward bias of 0~400 mV was applied during the measurement.
UV-Vis-IR Spectrum
Optical response for the colloidal crystals and their inverse opals were acquired through a Fourier Transform Infrared (FTIR) microscope (Hyperion 2000, Bruker) in a nitrogen atmosphere. During the measurement, the samples were illuminated with a selective light source (CaF2), and the reflectance was obtained by normalizing the measured signals detected from the samples to that from a flat aluminum mirror.
Chapter 4
Results and Discussion on EPD Process
This chapter provides the results and discussion on fabrications of colloidal crystals by the EPD in various processing parameters. Variables under studies were suspension concentration, applied electric field, and deposition time. In general, the deposit weight PS microspheres in suspension, S is the area of substrate, ε and ξ are the dielectric constant and zeta potential of PS microspheres, η is the viscosity of solvent in suspension, V is the applied voltage, L is the distance between working and counter electrodes, and t is the deposition time. In equation [4.2], there are three more variables in cylindrical EPD system;
l and a are the length and radius of the working electrode, and b is the radius for the coaxial
counter electrode (b > a). In this work, suspensions made of 460 and 660 nm PS microspheres (PS-460 and PS-660) were prepared in different Cs and examined to achieve a smooth EPD coating. By driving the EPD process with suitable applied voltage (electric field), the planar and cylindrical colloidal crystals (PCC-460/660 and CCC-460/660) were successfully fabricated. The following sections provide detailed discussion on individual variables and their effects on the quality of colloidal crystals.
4.1 Suspensions and Substrates
The as-synthesized PS microspheres are demonstrated in Fig. 4.1. The larger ones exhibited a diameter of 660 nm with a standard deviation of 16.4 nm while the smaller ones revealed a diameter of 460 nm with a standard deviation of 10.1 nm. Assembly of PS microspheres could be easily achieved simply by solvent evaporation. As shown, we obtained reasonable uniformity in size distribution and this attribute was critical for our EPD process. The PS microspheres were further mixed with ethanol to prepare the suspensions, PS-460 and PS-660.
Figure 4.1 SEM images of PS microspheres with diameter of (a) 460 nm and (b) 660 nm.
The scale bars are 2 μm.
Figure 4.2 presents the plots of electrophoretic mobility in PS-460 and PS-660. Their concentrations were 2.5 g/L and 10 g/L, respectively. The mobility for the PS microspheres in each suspension appeared in a single peak with corresponding zeta-potential was above 60 mV. Both results suggested that our suspensions were in excellent stability and the size of PS microspheres was reasonably mono-dispersive.
(a) (b)
Figure 4.2 Electrophoretic mobility of suspension (a) PS-460 and (b) PS-660.
4.1.1 Microsphere Concentration in Suspensions
While the PS microspheres were driven to deposit on the substrate from bulk suspension, it took some time to assemble them into a desirable close-packed arrangement. It is necessary for the first layer to be close-packed before the second layer can be deposited. Therefore, a reduced deposition rate was considered critical in fabricating high-quality colloidal crystals.
After rearranging those variables in equation 4.1, we could derive that the Cs is proportional to the deposition rate (Y/t). Thus, the Cs plays significant influence over the quality of the colloidal crystals.
Figure 4.3 presents the SEM images of EPD layers fabricated in PS-460 with Cs of 10 g/L and 2.5 g/L under 10 V/cm. As clearly shown in Fig. 4.3(a), the PS microspheres were in random arrangement with approximate domain size in 2~3 μm. In addition, a number of voids was observed among particles. When we reduced the PS microsphere concentration to 2.5 g/L, the domain of colloidal crystals became larger, as exhibited in Fig 4.3(b). These behaviors suggested that the EPD fabrication with an excess Cs could not deliver high-quality colloidal crystals. Moreover, increase in the Cs enhanced the volume fraction of PS microspheres, which raised op the viscosity of suspension and zeta potential. Thus, a larger Cs not only rendered a faster deposition rate but also affected the Brownian motion among
(a) (b)
Figure 4.3 SEM images of EPD layers prepared in PS-460 with various microsphere concentrations of (a) 10 g/L and (b) 2.5 g/L. The electrical field was fixed at 10 V/cm for 3 min. The scale bars are 3 μm.
Identical experiments were also performed in PS-660 suspensions, with their results presented in Fig. 4.4. Similarly, the suspension of 10 g/L demonstrated a better quality in colloidal crystals over 25 g/L one. Howerer, this great difference in the optimized PS microsphere concentration between PS-460 and PS-660 suspension was resulted from their distinct volume fraction.
Figure 4.4 SEM images of EPD layer made in PS-660 with microsphere concentrations of (a) 25 g/L and (b) 10 g/L. The electrical field was fixed at 10 V/cm for 10 min. The scale bars
(a) (b)
(a) (b)
the values for electrophoretic mobility in PS-460 (2.5 g/L) and PS-660 (10 g/L) were almost the same, as shown in Fig. 4.2. This suggested that the stability of suspension was more relevant to the volume fraction but not mass fraction. Therefore, in following experiments we chose the 2.5 g/L of PS-460 and 10 g/L of PS-660 to fabricate colloidal crystals.
4.1.2 Shape of Substrates
Figure 4.5 exhibits the SEM images of EPD layers on planar ITO glass and carbon fiber prepared under 10 V in 0.25 g/L PS-460. The EPD layer on the ITO glass revealed a (111) plane in FCC lattice. In contrast, a defective structure was observed in the EPD layer on CF.
We surmised that linear defects and curved surface of the CF might convert the state of arrangement from close-packed structure into a non-close-packed one. In the insets of Fig.
4.5, the thickness of ITO layer was 200 nm, and the diameter of CF was about 7 μm. In cylindrical case, the diameter size of the CF (7 μm) was only 10~15 times larger than that of PS microspheres (460 nm and 660 nm). Namely, EPD process was carried out on this curved surface (CF) with PS microspheres assembly, leading to a disordered EPD layer.
Figure 4.5 SEM images of EPD layer on (a) ITO glass and (b) carbon fiber. The insets provide the SEM images of each substrate. The scale bars are 5 μm in the figure and 2 μm in the inset.
(a) (b)
4.2 Fabrication of Planar Colloidal Crystals (PCCs) 4.2.1 Electric Field in EPD process
Figure 4.6 provides the thickness of PCC-660 made of various electric fields for 15 min on ITO glasses. The deposit thickness was linearly increased when the electric field was larger than 10 V/cm. However, a reduced deposition rate was present at 5 V/cm, which suggested that there was a transition regime for EPD rate between 5 V/cm and 10 V/cm.
Figure 4.6 Thickness versus electric field for the colloidal crystals made of PS-660. The EPD fabrication time was 15 min.
Figure 4.7 displays the SEM images for the colloidal crystals from 5 V/cm, 10 V/cm, and 20 V/cm, respectively. The PS microspheres were in close-pack structure at 5 V/cm, and became disordered as the applied electric field was increased. This suggested a larger driving force would engender a defective assembly. It is also known that a large electric
Figure 4.7 SEM images of PCC-660s fabricated in PS-660 under (a) 5 V/cm, (b) 10 V/cm, and (c) 20 V/cm for 15 min. The scale bars are 2 μm.
The thickness of EPD layer as a function of electric field using the 460 nm microspheres is displayed in Fig. 4.8. It was observed that the thickness of PCC-460 was linearly increased with the electric field. Figure 4.9 provides the SEM images for the samples made of 10 V/cm, 15 V/cm, and 25 V/cm, respectively. As expected, a larger electric field rendered a relatively disordered structure. Therefore, to improve the quality of PCCs, the electric field should be moderately reduced. Hence, we selected 15 V/cm in the EPD process using PS-460. Figure 4.10 demonstrates the relationship between current density (j) and deposition time (t) under the selected electric fields in PS-660 and PS-460. Both j-t curves reveal a decreasing current density at extended deposition time. This behavior indicated that the PS microspheres were self-assembling on the ITO glass, producing a mask to shield the
(a) (b)
(c)
Figure 4.8 Thickness versus electric field for the colloidal crystals made of PS-460. The EPD fabrication time was 10 min.
(a) (b)
(c)
Figure 4.10 Evolution of current density vs. time during the EPD process under (a) 10 V/cm using PS-660 microspheres and (b) 15 V/cm using PS-460 microspheres.
4.2.2 Deposition Rate for Planar EPD
Figure 4.11 demonstrates the plot of thickness versus time for the samples prepared in the EPD process.
Figure 4.11 Thickness versus EPD time for the PCC-660 (■) and PCC-460 (☆). The applied electric field was 10 V/cm in PS-660 and 15 V/cm in PS-460.
The estimated deposition rates were 1.064 μm-min-1 and 1.337 μm-min-1 for the PCC-660 and PCC-460, respectively. As clearly shown, the growth of PCCs was slowing down with
(a) (b)
the EPD time. We attributed this phenomenon to the decreasing electric field as the PS microspheres assembled on the substrate. Figure 4.12 and 4.13 exhibit the SEM images for the samples prepared in the EPD process with the optimized parameters. As expected, the microspheres were assembled properly to form colloidal crystals in multiple layers.
(a) (b)
(c)
(e)
(d)
Figure 4.13 SEM images of PCC-460s fabricated under 10 V/cm for (a) 2.5, (b) 5, (c) 7.5, (d) 10, and (e) 15 min, respectively. All scale bars are 5 μm.
(a) (b)
(c) (d)
(e)
4.3 Fabrication of Cylindrical Colloidal Crystals (CCCs) 4.3.1 Applied Voltage in EPD Process
As shown in equation 4.2, the yield amount from the cylindrical EPD process was modified by additional three parameters, a, b, and l, and it was further reduced by a constant about 8.79 (ln(a/b)) from that of planar EPD process. On the other hand, the applied voltage (V) was present in the equation as an independent parameter. Hence, we only controlled the voltage in the cylindrical EPD process.
Figure 4.14 display the SEM images of CCC-460s made from two different voltages in PS-660 for 30 sec. The morphology for the CCCs revealed a defective packed structure, which was resulted from the curved CF surface, as shown in Fig. 4.14(a). The diameter of CCC made of 5 V was about 9 μm, which was equal to 3 EPD layers. At this stage, the color of substrate remained dark, which can be easily observed by naked eyes.
In contrast, as shown in Fig. 4.14(b), the diameter of sample made of 10 V was about 60 μm, which was almost 30 layers. Because the size of CF was in micron range, its dimension was relatively small compared with the planar substrates. Therefore, under identical yield amount, the EPD layer prepared on the CF would be 103-104 times thicker as opposed to that of planar substrate. As a result, the effect of voltage during the EPD became extremely critical in the cylindrical case.
Figure 4.15 provides the diameter of CCCs under different applied voltages in the PS-460 and PS-660. Obviously, a voltage of 5 V was not able to deposit large amount PS microspheres in a short time, but 10 and 20 V were capable of delivering a much enhanced deposition rate for both PS-460 and PS-660 suspensions.
Figure 4.14 The SEM images of CCC-460s made of (a) 5 V, and (b) 10 V. The fabrication time was 30 sec. The scale bars are 5 μm in (a) and 10 μm in (b).
Figure 4.15 Diameter versus voltage for the CCC-460s (☆) and CCC-660s (■). Their fabrication time was 30 sec.
Figure 4.16 presents the SEM images of CCCs made of 10 V and 20 V by PS-660 and PS-460. Among them, the applied voltage of 20 V rendered a poor packing quality in the EPD layer because unnecessary larger EPD rate. In contrast, the 10 V exhibited a relatively dense packing in both suspensions. Because the diameter for the CCC-460 made of 10 V was much larger than that of CF, its fabrication could be approximated to a planar EPD case.
Hence, the crystal quality of CCCs made of 10 V was still reasonable even the EPD rate of 10
(a) (b)
V was much larger than that of 5 V. In addition, the CCC-460 revealed a better quality than that of 660 nm, as shown in Fig. 4.16(a) and Fig. 4.16(c). This suggested that the ratio of microsphere size to the diameter of substrate plays a critical role in desirable ordered arrangement. Hence, difference in the diameter for the CCCs indicated the importance of applied voltages. However, it was challenging to control the layer numbers with a large voltage because the deposition rate for the cylindrical EPD was taking place too fast.
Figure 4.16 SEM images on the surface morphology of CCC-660 made of (a) 10 V and (b) 20 V, as well as CCC-460 made of (c) 10V and (d) 20V. The fabrication time was 30 sec, and the scale bars are 5 μm.
(a) (b)
(c) (d)
and the self-assembling of PS microspheres was easy to be achieved. Hence, we selected the 10 V in following experiments
Figure 4.17 Time dependence of current density for (a) CCC-660 and (b) CCC-460 during cylindrical EPD process. The applied voltage was fixed at 10 V.
4.3.2 Deposition Rate for Cylindrical EPD
Figure 4.18 presents the plot of diameter versus time plots for the cylindrical colloidal crystals in the EPD process with PS-460 and PS-660. The diameter of CCC was linearly increased with deposition time, and a higher deposition rate was obtained in PS-440 than that in PS-660. Their respective deposition rates for the CCC-460 and CCC-660 were 0.322 μm-s-1 and 0.301 μm-s-1 intially, following by a slightly decrease to 0.201 μm-s-1 and 0.145 μm-s-1 after 6 min. These reduced rates indicated that those microspheres were arranged in proper packing order, reducing the pathway for the current on the substrate. According to the modified Hamaker’s equation [51] for cylindrical EPD, the linear dependence indicated our experiment was still in the primary stage of EPD. This suggested that the deposits could be further thickened in a short time.
(a) (b)
Figure 4.18 Diameter versus EPD time for the CCC-660 (■) and CCC-460 (☆) fabricated in suspensions under a voltage of 10 V.
Figure 4.19 displays the SEM images for the CCC-660 made of 10 V in PS-660 for different EPD time. Microspheres in size of 660 nm were not stacking well on the CF first, but ordered layers were observed later and became FCC (111) plane, as shown in Fig. 4.19(a) and 4.19(b). It suggested that as the diameter of CCCs increased, the disorder region could be compressed and became the grain boundaries distributed in the close packed structure. If the deposition time was further prolonged, there were some BCC grains obtained in the EPD layers, as exhibited in Fig. 4.19(c). This indicated that the most stable state was not present by self-assembly on the carbon fiber substrate. There might be two reasons to explain the appearance of planes such as FCC (100) and BCC (100).
The first plausible cause is the undesirable high EPD rate. It is known that the EPD rate increased with deposition time within 10 min. Hence, a higher EPD rate would raise the
planes would reveal a non-close packed structure, or the BCC (100). This implied the EPD on a curved surface might be useful to achieve a defective packing structure. Figure 4.20 presents the SEM images of CCC-460 made of 10 V for various time in PS-460. Similarly, the size of ordered domains was becoming larger with the EPD time. However, the BCC plane was not found in the EPD layers.
Figure 4.19 SEM images for the surface morphology of CCC-660s made of 10 V in PS-660 for (a) 2 min, (b) 4 min, and (c) 8 min, respectively. All scale bars are 5 μm.
(a) (b)
(c)
Figure 4.20 SEM images of the surface morphology of CCC-460s made of 10 V in PS-460 for (a) 2 min, (b) 4 min, and (c) 8 min. All scale bars are 5 μm.
Figure 4.21 presents the time dependence plot for the diameter of CCCs within 1 hr. From this figure, the increase in diameter of CCCs tended to reach a stable state after certain time.
As pointed with arrows in the figure, a higher growth rate of CCCs was occurring after 30 min. The abnormal increase might be accidentally induced by “dip-coating” during sample removal.
(a) (b)
(c)
Figure 4.21 Diameter versus EPD time for the CCC-660 (■) and CCC-460 (☆) fabricated in suspensions under a voltage of 10 V.
Figure 4.22 and 4.23 present the photographs of CCCs taken by an optical microscope.
Both CCC-460 and CCC-660 exhibited smooth surfaces along the longitudinal axis, and the diameter for each sample revealed rather impressive uniformity. Therefore, for every sample, a direct reflectance was observed in the middle of CCC, and its intensity was enhanced when its diameter was increased under identical light source.
In optical response, the CCC-460 displayed a set of light diffracting fringes as the EPD process was extended longer, as shown in Fig. 4.22. These diffracting lines demonstrated a distribution of the reflective light in illumination with various angles on FCC (111), and the wavelength of reflectance increased from substrate to the surface due to increasing lattice spacing. However, there is no obvious light reflecting when the CCC-660 was illuminated by a white light, as exhibited in Fig. 4.23(a) and 4.23(b). The self-arranged layer of PS-660 was essentially distorted by a curved substrate and the size incoherence between substrate and spheres, as well as those linear defects on the CF.
Figure 4.22 Optical microscopy images of CCC-460 self-assembled in PS-460 for (a) 10 min and (b) 50 min. The scale bars are 200 μm.
Figure 4.23 Optical microscopy images of CCC-660 self-assembled in PS-660 for (a) 10 min, (b) 30 min, and (c) 50min, respectively. The scale bars are 200 μm.
(a) (b)
(c)
(a) (b)
So far, we have demonstrated that in fabricating the CCCs in a fixed diameter, smaller microspheres were easier to assemble. We observed that the CCC-660 was strongly affected by the substrate. For example, a pale rainbow color appeared on the CCCs with a longer EPD time, as shown in Fig 4.23(c). In short, the EPD process could deliver cylindrical colloidal crystals in a short period, and the applied voltage was optimized with the time dependence of current density. Furthermore, with a high EPD rate and curved surface, self-assembling of PS microspheres on a cylindrical electrode was capable of producing colloidal crystals with defective arrangements. More details from optical observation will be discussed in following sections.
Chapter 5
Results and Discussion on NED and PS Removal Processes
This chapter provides the results and discussion for NED and PS removal process in nickel inverse opals fabrication. In the NED process, a potentiostatic electrodeposition technique was carried out on the colloidal crystals to fabricate nickel inverse opals (NIO-460 and NIO-660) and cylindrical inverse opals (CIO-460 and CIO-660). Figure 5.1 provides the SEM images for those inverse opal structures. To remove colloidal PS templates, the as-prepared samples were treated under different temperatures in chemical dissolution or heat treatment.
Figure 5.2 presents the cross-sectional views for the as-prepared PCCs on ITO glass. The thickness of EPD layer for the PCC-460 and PCC-660 were 11.25 and 8.97 μm, respectively.
As clearly shown in the figure, the PCCs were in good quality of proper stacking, and the PS microspheres demonstrated a three-dimensional close-packed arrangement.
Figure 5.3 displays the surface morphology for the as-prepared CCCs prior to the NED process. Clearly, the microspheres in the CCC-460 were in good order. In contrast, in the CCC-660, there revealed a defective packed structure. The insets exhibit the OM images for the as-prepared samples. The diameters were about 67.31 μm for the CCC-460 and 52.94 μm for the CCC-660, and these values were rather uniform along the longitudinal axis for both cases. It is noted that these colloidal templates were rather robust as immersing them in electrolytes did not produce unwanted detachments or structural damage.
Figure 5.2 Cross-sectional SEM images for the (a) PCC-460 and (b) PCC-660 before NED process. The scale bars are 5 μm.
(a) (b)
Figure 5.3 SEM images on the surface morphology of (a) CCC-460 and (b) CCC-660. The scale bars are 5 μm. Insets display optical microscopic images for CCCs with scale bars of 100 μm.
(a) (b)
5.1 Fabrication of Planar Nickel Inverse Opals 5.1.1 Current Density in NED Process
Figure 5.4 demonstrates the time dependence of current density during the NED process on planar colloidal templates. The value for current density exhibited a sudden jump initially
Figure 5.4 demonstrates the time dependence of current density during the NED process on planar colloidal templates. The value for current density exhibited a sudden jump initially