Photoelectrochemical Etching of Gallium Nitride for Solar Applications

 

Solar panels are most efficient when they reflect the least light energy. This is often accomplished with an antireflective surface. The goal of this project is to understand how various 3D nanostructures can be imparted on gallium nitride (GaN) surfaces by photoelectrochemical etching to determine the least reflective surface. In this experiment, GaN was used as the semiconductor because of its durability and stability, both physically and chemically. This gives it great use in areas where replacing panels is difficult, such as on mountains or satellites. Although silicon based semiconductors are often used, GaN is potentially more cost effective and suitable for use in harsh environments. The surface structures of GaN wafers were modified using a wet chemical etching process which included either 0.3 M nitric acid or tribasic phosphate.  In this process, the wafers were treated with both UV light and either 2.0 or 3.0 Volts, (between the etching threshold and the band gap). The surface structure was dependent on both the etchant and the voltage. The results show that nitric acid combined with 3.0 V produced the lowest reflectivity (9.3%) compared with trisodium phosphate at 3.0 V (24.0%) and an unetched GaN wafer (30.4%). The low reflectivity produced by he GaN was comparable to that of etched silicon, which has been shown to be approximately 10-15%. These results indicate GaN is a potential alternative for certain solar applications.

Close

I have always loved physical sciences because they explain the world in a way that makes sense to me. Solar panels have seemed particularly interesting because they combine chemistry, physics, and geometry with huge real world applications.

In 10th grade, I took a Research Design course where I analyzed scientific papers in an effort to create an original experiment. I loved the papers that altered the microscopic geometric shapes and heights on solar panels.  My experiment represented an extension of this scientific literature. The project was too difficult to perform at my school as most of the nanoscale equipment is exceptionally expensive, so I emailed some local labs and found an amazing lab in Yale’s Department of Chemical engineering. I met these researchers and discovered many common interests. I have been working on my project in their lab for the past year. While I have a heavy course load, play ultimate frisbee, and enjoy Quiz Bowl, my experience in the lab is completely different and amazing.  Original research adds the exciting element of trying totally new things where the results are unknown.             

Currently, I live in Stamford, CT, and I am a junior at Greens Farms Academy. In college, I plan to major in engineering (possibly Materials Science). Ultimately, I would like to go to graduate school and continue to do research. Wining Google would enable me meet and learn from an incredible group of like minded people and accelerate my path in engineering research.

 

 

Close

The goal of this project is to understand how various three dimensional nanostructures can be imparted on gallium nitride (GaN) surfaces by photoelectrochemical etching to determine the least reflective surface for solar applications.

What would be the optimal combination of voltages and etchants to create the least reflective surface morphology on a GaN wafer?

Hypothesis:

Acidic conditions under higher voltages combined with UV light, should create the least reflective surface because these variables will eliminate more point and line disjunctions than other conditions.

Close

Solar energy is promising because it is abundant renewable energy. Photovoltaic cells work optimally when they have antireflective coatings and modified surface structures that allow them to retain light energy[1]. Traditionally, silicon has been the primary material in photovoltaic cells, in large part because of the material’s success in the electronics industry. Gallium Nitride (GaN) shows promise as an alternative material to silicon for some applications because of its durability and stability both physically and chemically [1]. It is possible that GaN panels will prove effective in particularly harsh weather environments or on satellites where the material’s stability under higher radiation conditions will be useful. In general, unetched GaN has a reflectiveness of about 31% so 31% of the light that hits the wafer is reflected off of it. Etched silicon has a reflectiveness from 10-15% [2]. In order for GaN to be a viable alternative to silicon, its reflectiveness must be reduced

To improve the energy conversion efficiency of a semiconductor material in a solar cell, the surface structure is modified by a wet etching process which removes material and leaves behind a specific surface morphology. In the etching process, an unfavorable surface layer is removed and dissolved into an ionic fluid. Further, low applied voltage is used along with an intense UV light source [3] and etchants to remove the surface layer from the wafer. Each combination of light, electricity, and etchants creates a distinct three dimensional surface morphology with antireflective properties [3]. Typically, photoelectrochemical etching creates a productive anti reflective coating.

Different surface morphologies have been studied at the molecular level for reflectivity. Many different microstructures have anti reflective properties [4, 5] but hexagonal pyramids are generally optimal on GaN wafers compared to triangular or rectangular pyramids or a honeycomb shaped surface [5, 6]. Previous studies have not used voltage as a method of controlling the heights of the pyramids.

Various etchants have been used to remove the surface of the semiconductor. These etchants can be acids, bases, or neutral. Nitric and hydrofluoric acid has been shown to etch away point and line defects in single crystalline and is known to create optimal pores sizes [7, 8].  Nitric acid is usually the acid of choice as hydrofluoric acid is highly dangerous.

This study will use trisodium phosphate as a base, which has not been heavily studied. In previous studies, sodium hydroxide have been shown to produce surface morphologies of hexagonal pyramids [9], which are an ideal antireflective surface [6]. In addition, Jung et. al. examined photoelectrochemical etching of GaN with both acids and bases. They found that both bases and phosphoric acid can create hexagonal pyramids independently of each other [10]. This current study builds upon Jung’s work by using trisodium phosphate which is a base that has phosphate ions.

 

 

Close

Independent variable: Voltage: 2.0V, 3.0V, Etchant: Trisodium phosphate, nitric acid

Dependent variable: Reflectivity

Experiments were conducted at Dr. Lisa Pfeferle’s lab, Department of Chemical Engineering at Yale University.  

The wafers used had a silicon-doped GaN thickness of 2 microns with a doping level of 2x1019Si atoms/cm^3. This doped layer was grown on top of an unintentionally-doped GaN layer, and both GaN layers were grown by metalorganic chemical vapor deposition. The GaN wafers were cut into strips 1.5cm by 0.5cm. The strips were then prepared by rinsing them in hot water and sonicating them in a beaker filled with hexane to remove organic residue.

When preparing wafers for etching, a piece of metallic tape was cut and placed on the front side of one end of each wafer.  An electrical wire was attached to the metallic tape. The wafers were then wrapped in acid resistant electrical tape.  

A portion of each wafer was covered with electrical tape so that 0.25 cm^2 of each wafer was exposed to etchant. The electrical tape was wrapped tightly to prevent etchants in the petri dish from contacting the wire.

The wafers were then taped down onto the right side of the petri dish, and a platinum wire counter electrode was taped to the left side of the dish. The counter electrode was used to conduct the current through the wafer and solution of water and etchant. The petri dish was then placed in an enclosure which was designed with a removable barrier to prevent the UV light from illuminating the wafer prematurely. A 0.3M solution of specific volume of either trisodium phosphate or nitric acid was poured into the petri dish. The UV lamp was given 5 minutes to warm up after which, a positive voltage of 2.0V or 3.0V was applied to the wafer with respect to the counter electrode, and the barrier in the enclosure in front of the lamp was removed so the UV light could illuminate the wafer. The wafers were etched for 10 minutes. The current was read with the voltage controller every two minutes.

After 10 minutes, the UV lamp and applied voltage were turned off. The petri dish with the wafer and the counter electrode was removed and placed in a fume hood, where the etchant was disposed of into a waste container. The wafer was washed in deionized water for 3 minutes, then dried and placed in a vial, and kept for analysis. Gloves were worn throughout the procedure and handling of acids was done in a fume hood.

An Atomic Force Microscope was used to observe the surface morphology of the wafers. To measure the reflectivity, the GaN wafers were exposed to a 1.0mW continuous laser. A handheld laser beam profiler was used to measure the intensity before the laser hit the sample (incident laser power) and the intensity of the spectrally reflected light (reflected laser power).  Three readings were taken for each sample.

 

Close

 

FIGURE 1: Spectral Reflectiveness of Etched And Unetched GaN Wafers

In order to calculate the average spectral reflectiveness, the average reflected laser power was divided by the average incident laser power.  

All conditions had less reflectivity than the unetched wafer. The wafer samples etched with nitric acid at 3.0 V had the lowest average spectral reflectiveness at 9.31% (Figure 1). Nitric acid at 3.0 V were over 3 times less reflective than the unetched GaN, (Table 3). The trisodium phosphate 3.0 V was somewhat less reflective than the unetched wafers (Tables 2 and 4), however, the trisodium phosphate 2.0 V was similar to the unetched wafer.  Overall, nitric acid was more antireflective than trisodium phosphate and 3.0 V was more antireflective than 2.0 V.

Table 2: Reflectivity of Wafers - Unetched GaN

 

Reading 1

Reading 2

Reading 3

Average

Incident laser power

0.996 mW

1.03 mW

1.01 mW

1.01 mW

Reflected laser power

0.310 mW

0.303 mW

0.307 mW

.307 mW

Average Spectral Reflectiveness

     

30.4%


 

Table 3: Reflectivity of Etched - Wafers Nitric Acid, 2.0 and 3.0 Volts

Nitric Acid 2.0 V

 

Reading 1

Reading 2

Reading 3

Average

Incident laser power

1.00 mW

1.03 mW

0.999 mW

1.01 mW

Reflected laser power

0.286 mW

0.293 mW

0.286 mW

288 mW

Average Spectral Reflectiveness

     

28.5%

 

Nitric Acid 3.0 V                 

 

Reading 1

Reading 2

Reading 3

Average

Incident laser power

0.988 mW

1.01 mW

0.999 mW

.999 mW

Reflected laser power

0.0922 mW

0.0924 mW

0.0944 mW

.0930 mW

Average Spectral Reflectiveness

     

9.31%

 

Table 4: Reflectivity of Etched Wafers - Trisodium Phosphate 2.0 V and 3.0 V

Trisodium Phosphate 2.0 V

 

Reading 1

Reading 2

Reading 3

Average

Incident laser power

1.01 mW

.998 mW

1.00 mW

1.00 mW

Reflected laser power

.299 mW

.301 mW

.302 mW

.301 mW

Average Spectral Reflectiveness

     

30.1%



 

Trisodium Phosphate 3.0 V

 

Reading 1

Reading 2

Reading 3

Average

Incident laser power

1.02 mW

1.07 mW

0.989 mW

1.03 mW

Reflected laser power

0.249 mW

0.247 mW

0.245 mW

.247 mW

Average Spectral Reflectiveness

     

24.0%

 

 

The surface morphology of the etched wafers differed from the unetched surface.  Nitric acid 3.0 V had a rough, needle like surface morphology compared to the unetched wafer (Figure 2). While less needlelike that nitric acid 3.0 V, tribasic phosphate 2.0 V (Figure 3) showed large differences in height compared to the unetched wafer which was almost completely flat. (Figure 4).

FIGURE 2: Nitric Acid 3.0 V Atomic Force Microscope

Nitric 3.0 V.jpg

  

FIGURE 3: Trisodium Phosphate 2.0 Atomic Force Microscope

Tribasic 3.0 Dark 3D angled (for comparison).jpg

  

 

FIGURE 4: Unetched Wafer as Seen Through the Atomic Force Microscope

Unetched angle view flattish.jpg

(All photos taken by Matthew Lichtenberg)

Close

This study used a wet etching process and four modified GaN samples to determine potential techniques for improving the antireflective properties of a GaN wafer. Specifically, the conditions included nitric acid at 2.0 V and 3.0 V and trisodium phosphate at 2.0 V and 3.0 V. The study also used a consistent amount of UV light to modify the GaN surfaces. After the surface layer was modified, the antireflectiveness of the resulting wafers was studied.

The condition that produced the most promising result for antireflective properties was the nitric acid at 3.0 V condition. In this case, the resulting reflectivity was 9.31%.  This is a considerable improvement compared to unmodified silicon (approximately 30%) and unetched GaN (measured at 30.4%). In addition, this result is broadly consistent with etched silicon which has displayed reflectiveness between 10-15% in previous studies[2]. This suggests that GaN has further potential for low reflectivity.

The condition using nitric acid and 2.0 V produced a reflectiveness of 28.5%, which is slightly better than the unetched GaN or silicon but not better than nitric acid at 3.0 V.  It is important to note the decrease in antireflectiveness created by going from 2.0 V to 3.0 V which suggests that future experiments refine the changes in voltage to increments of 0.1 V or 0.2 V.

When trisodium phosphate was used as an etchant at 2.0 V, the resulting reflectivity was 30.1%, consistent with the unetched GaN wafer.  In the condition using trisodium phosphate and 3.0 V, the resulting reflectivity was 24.0%. This condition was slightly better that the unetched GaN wafer or nitric acid etched with 2.0 V condition but not as good as nitric acid condition etched with 3.0V. Overall, nitric acid 3.0 V was much less reflective than trisodium phosphate at either voltage. It is possible that smaller voltage increments in future studies of trisodium phosphate would produce improved reflectivity results, but it appears that nitric acid is a superior etchant for reflectance.

Using nitric acid and 3.0 V as an etchant was the most favorable condition in terms of antireflectiveness. The etched GaN wafers performed comparably to modified silicon wafers in terms of reflectiveness. Further studies should explore the use of different concentrations of nitric acid as an etchant for GaN.  In particular, while this study used nitric acid at a molar concentration of 0.3 M, it is possible that stronger concentrations will produce greater changes in surface morphology and better antireflective surfaces. It would also be interesting to analyze more voltage conditions to determine if a voltage higher 3.0 V (but less than the band gap of 3.4 V if using UV light) further lowers spectral reflectiveness. These extensions could lead to even more efficient GaN wafers for photovoltaic cells. GaN is more stable and durable than silicon and could have significant implications for use in harsh environments, such as satellites and mountain tops.

 

 

 

Close

REFERENCES

1. PVEducation." PVEducation. N.p., n.d. Web. 15 Dec. 2014.Singh, P K, R. Kumar, S N Singh, and B K Das. "Effectiveness of Anisoptropic Etching of Silicon Aqueous Alkaline Solutions." Solar Energy Material and Solar Cells 70.1 (2001): 103-13. Web.

2. Singh, P K, R. Kumar, S N Singh, and B K Das. "Effectiveness of Anisoptropic Etching of Silicon Aqueous Alkaline Solutions." Solar Energy Material and Solar Cells 70.1 (2001): 103-13. Web.

3. Radzali, R., Z. Hassan, N. Zainal, and F. K. Yam. "Nanoporous InGaN Prepared by KOH Electrochemical Etching with Different Light Sources." Microelectronic Engineering 126.25 (2014): 107-12.

4. Tao, Meng, Weidong Zhou, Hongjun Yang, and Li Chen. "Surface Texturing by        Solution Deposition of Omnidirectional Antireflection." Applied Physics Letters 91 (2007): 081118-081120.

5. Huang, C. K., K. W. Sun, and W. L. Chang. "Efficiency Enhancement of Silicon Solar Cells Using Nano-scale Honeycomb Broadband Anti-reflection Structure." Optics Express 20 (2012): A85-A93.

6. Shi-Ying, Zhang, Xiu Xiang-Qian, Hua Xue-Mei, Xie Zi-Li, Liu Bin, Chen Peng, Han Ping, Lu Hai, Zhang Rong, and Zheng You-Dou. "GaN Hexagonal Pyramids Formed by a Photo-assisted Chemical Etching Method." Chinese Physics B 23.5 (2014): 058101.

7. Schwab, Mark J., Danti Chen, Jung Han, and Lisa D Pfefferle. "Aligned Mesopore Arrays in GaN by Anodic Etching and Photoelectrochemical Surface Etching." The Journal of Physical Chemistry C 117 (2013): 16890-6895.

8. Chen, Danti, Hongdi Xiao, and Jung Han. "Nanopores in GaN by Electrochemical Anodization in HF: Formation and Mechanism." Journal of Applied Physics 112 (2012): 064303.

9. Shi-Ying, Zhang, Xiu Xiang-Qian, Lin Zeng-Qin, Hua Xue-Mei, Xie Zi-Li, Zhang Rong, and Zheng You-Dou. "The Formation and Characterization of GaN Hexagonal Pyramids." Chinese Physics Letters 30 (2013): 056801.

10. Jung, Younghun, Jaehui Ahn, Kwang Hyeon Baik, Donghwan Kim, Stephen J. Pearton, Fan Ren, and Jihyun Kim. "Chemical Etch Characteristics of N-Face and Ga-Face GaN by Phosphoric Acid and Potassium Hydroxide Solutions." Journal of The Electrochemical Society 159.2 (2012): H117.

 

ACKNOWLEDGEMENTS

I would like to acknowledge the assistance provided by Dr. Mark Schwab at the Yale School of Engineering, Department of Chemical Engineering.  Dr. Schwab mentored me throughout this project.  While I conducted all conditions of the experiment on my own, he helped me obtain the equipment and trained me on how to use the equipment employed in this study. Further, he helped ensure that all materials were dealt with safely. This research was conducted in Professor Lisa Pfefferle’s Chemical Engineering Lab at Yale University, and I would like to thank her for the use of her lab and facilities. Finally, I would like to thank Dr. Mattieu Freeman at Greens Farms Academy for supervising me while I conducted my research.

All photographs were taken by the author

Close