The atomic force microscope (AFM) is an essential tool for imaging surfaces in
applications in cell biology and biomaterials science, measuring surface topography on a scale from angstroms to 100 microns. The key component of the AFM is its cantilever. At the end of the cantilever is a tip that is used to sense a force between the sample and tip. The tip is held several nanometers above the surface using a feedback mechanism that measures surface-top interactions on the scale of nano-Newtons. The tip is brought into continuous or intermittent contact with the sample and raster-scanned over the surface. Continuous contact is referred to as Contact Mode, while intermittent contact is referred to as Tapping Mode. The height of the sample is measured by continuously scanning the sample and recording the deflection of the cantilever. Three-dimensional topographical maps are constructed by plotting the local height against the horizontal probe tip position.
Nano-scale force interactions represent a developing area of study within the life sciences. The AFM can record the amount of force felt by the tip when brought close to and/or indented into the sample surface and then pulled away. The force is measured as a function of the deflection of the tip on the cantilever. This technique can be used to measure the attractive or repulsive forces between the tip and the sample surface, revealing chemical and mechanical properties like adhesion and elasticity.
The diagram of AFM instrument is shown in Fig. 13.
Chapter 3
Results and Discussion
3.1 The Analysis of Etch Depth 3.1.1 Effect of KOH
As the Fig. 14(a) shown, the etch depth has a peak appearance versus the KOH concentration (Bias: 0V, K2S2O8: 0M). We observed a peak etch depth of 1.25μm in the KOH 0.5M solutions. The OH- in both PEC oxidation and oxide dissolution reaction plays a reactant role, and the 1st reaction product (Ga2O3(s)) is just the 2nd reaction reactant. Generally speaking, if the amount of chemistry reactant is greater than the amount of product, it is tend to the positive reaction. And according the mass transport rule, the diffusion rate changes into fast in higher concentration and can accelerate the 1st reaction. But experimental results seem be not as we expect.
However, the observation of a peak etch depth indicates that both of the solute (KOH, K2S2O8) and the solvent (H2O) play an important role in the PEC etching of GaN. In the 0.01M and 0.05M K2S2O8 conditions(0V bias) also shows the same results, but the KOH concentration of maximum etching depth in 0.05M K2S2O8 condition reduced to 0.05M. So we suppose that it may be related to H2O molecular about the result.
Indeed, peaking in the wet chemical etching is a well-known phenomenon [29]. It has been reported in an electrochemical etching [30] or a thermal etching process [31].
Before a recent report by Quinlan which indicates a competing effect due to the
photo-reduction of hydrogen ions could result in similar effects [32], it is less noticed in the PEC etching process. Such an effect is attributed to a hydration effect in which competing effects from free water molecules (H2O)f and hydroxyl ions(OH-) produce a peak in the etch depth.
The hydration effect model for analyzing PEC etching of GaN has been reported. In this model, both the free water molecules (H2O)f and hydroxyl ions(OH-) is needed to accomplish the etching[33]. The hydration effect means that if the more OH -concentration, the less (H2O)f concentration. And so does sulfate (K2S2O8). The two competing effects therefore produce a peak in etch depth whose location is very sensitive to the mean hydration number of solute. A general reaction rate equation is proposed by PENG et al. [34] to account for the GaN PEC etch rate ie.,
R=C。[H2O]rf[OH-]s
In the equation, the factor C(λ, I, Ea, T) is a function of the incident photon wavelength, illumination intensity, activation energy of the reaction, and temperature, respectively. Hence we can infer the shift of peak location at 0.05M K2S2O8 that is due to water hydrate the sulfate and reduce amount of free water molecular. Due to oxidation and oxide dissolution reaction in PEC wet etching, the slow etch rate at low KOH concentration is caused by [OH-] that not enough to dissolve the oxide. On the contrary, the high KOH concentration makes a few [H2O]f and holes which is generated from UV light also become insufficient. The first oxidation reaction can not drive the second oxide dissolution reaction. The factor brings about the less etch rate at high KOH concentration.
Fig. 14(b) shows etch depth of applied bias voltage. It is clearly detected that the maximum etch depth at 0.05M K2S2O8 is 3.02μm, but it is not located at 0.5M
KOH. It means that the dominated mechanism changes from photo-assist into bias-assist etching (holes injection). And the photo-generated holes will not govern etching actions in the conditions. Owing to the reactant (holes) is sufficient, the etch rate becomes proportional to the KOH concentration.
3.1.2 Effect of Sulfate (K
2S
2O
8)
The photochemistry of peroxydisulfide has been extensively studied [35~37], and the relevant reactions for UV photoenhanced etching are summarized in Table.
2[38~39]. As Table. 2 shown, as time goes on, the K2S2O8 will be consumed as well as KOH .In order to consider the influence of K2S2O8 in etching reaction, the dose of K2S2O8 in the experiment were set in none, not saturation, over-saturation. At wavelengths shorter than ~310 nm, peroxydisulfate absorbs photons, resulting in the initial production of the sulfate radical (SO4-*), reaction (4), Table. 2. Depending on the pH of the solution, this radical either recombines, or results in production of the hydroxide radical (OH*) [in alkaline deaerated solutions, reaction (5)], or production of theoz onide (O3-) [ inalkalineaerated solutions, reaction (8)]. Both the sulfate radical ion and the hydroxide radical are very strong oxidizing agents, and thus could be involved in the photoenhaced wet etching of GaN, through reactions(6) and (7), Table. 2. The presence of peroxomonosulfate or of any other strong oxidizing product was checked with a test of iodide oxidation rate [40]. It should be noted that S2O8
2-does not absorb at 365 nm. Figure. 15 shows results of the etch rate versus K2S2O8
concentrations. The low etch rate at low KOH concentration which is caused by
diffusion-limited can be enhanced with K2S2O8. The generated hydroxyl radical and sulfate ion radical assist the first oxidation reaction and then promote overall etch rate.
The etch rate at 0.5M KOH decreases from 0.01 M to 0.05 M K2S2O8 is due to the lower free water molecular (hydration effect) and not enough alkaline to solve the Ga2O3(s). On the other hand, although the free water molecular of 2M KOH becomes less, but the plenty of alkaline can overcome the disadvantage of dissolving Ga2O3(s)
at long period of etch. Thus the etch rate will be increased finally.
3.1.3 Effect of Bias Voltage
In the study, a negative bias voltage has been applied to the GaN samples, but
no bubble is appeared. For this reason, 3 kinds of different positive bias voltage were chosen to compare in this work (0V, 1.5V, 3V). The etch depth results of these experiments are showed in Fig. 16. When the bias increased, the etch depth increased whether the electrolyte is mixed with the K2S2O8 or not. There are 2 possible explanations.
First, it is believed that GaN etching process through the oxidative
decomposition of GaN into its component elements and subsequent dissolution of the semiconductor into the solution. Ultraviolet illumination is used to generate electron-hole pairs at semiconductor surface, which enhance the oxidation and reduction reactions within an electrochemical cell. So we can upward or downgrade the band bending of GaN at the interface between GaN and electrolyte by applying bias voltage. If a positive bias is applied to the GaN film, the chemical potential of the
surface is reduced, which flattens the bands [41~42]. The upward band of n-GaN forms a potential well for holes at the surface, and then accelerates the reaction rate.
The surface energy band diagram which changed with different bias is shown in Fig.
17. By raising the positive bias voltage, the well becomes deeper and the more holes center on the surface of GaN. This means that the accumulated holes at the surface of the n-GaN will contribute to the PEC etching process.
Another explanation is holes injection provided by power supply. Because the experimental samples are patterned by Ti metal mask, the Ti mask can provide the current a convenient conducting path under the n-GaN surface. Thus the reactant holes can be provided not only by the UV but also power-supply. But a much larger etch rate is observed near the periphery of the etched region. Figure. 19 shows the etching morphology and present the non-uniform etch depth. Because the overall etching rate at low concentration is slow (due to diffusion-limited), the etching mechanism will be dominated by holes injected (not photo-induced) and current crowding will be near the Ti film. Etching process at low bias voltage, the electric-field above the wafer may be not enough strong and cause obvious potential gradient, so holes injected by power supply will not affect the central etching of bare GaN. But on high voltage condition, the non-uniform phenomenon (shown in Fig) will be disappeared due to the large electrical-field.
3.1.4 Effect of H
3PO
4Etchant
Figure.18 (a) displays the etching depth versus the ratio of H3PO4 in H2O.
Owing to the etch depth at no applied bias in 10minutes is lower than 0.1μm, it is
difficult to detect their difference so that we take the etch depth results of long time (30minutes). It is obviously found that the etch rate in the KOH solution was faster than it was in H3PO4 solution. At no bias voltage condition, it also reveals the appearance of PH dependent, mass transport effect is also same with KOH. The assumed reaction of dissolution of GaN in H3PO4 is [43]
GaN + H3PO4 → Ga3+ + PO43- + NH3
The reaction is accompanied with the formation of a soluble acidic salt GaPO4。H3PO4。H2O in the solution.[44] The fastest etch depth in 10 mins is 0.067 μm, but still lower than KOH condition. The peak location shifts right at the condition of adding bias voltage, it may be due to the band bending change. It reveals that PEC etching of GaN ceases in the pure H3PO4 solutions even if it applied bias voltage. The result can be attributed to no free water molecular and no photo-generated electron-hole pair on the GaN surface. Combined with information presented above, it concludes that the importance of UV light in wet etch reaction is more than that of bias voltage. Figured.18 (b) shows the SEM images after 10 minutes PEC etching in phosphoric acid. The whisker formation also reveals the relation of threading dislocation and reactant charges generated from UV light. The PEC etching mechanism of H3PO4 etchant may be similar to that of KOH etchat on the basis of mention above.