Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Light activated Cl₂ etching of GaAs and optical holographic pattern formation Whitwick, Michael Brian 2003

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2003-0324.pdf [ 5.09MB ]
Metadata
JSON: 831-1.0091083.json
JSON-LD: 831-1.0091083-ld.json
RDF/XML (Pretty): 831-1.0091083-rdf.xml
RDF/JSON: 831-1.0091083-rdf.json
Turtle: 831-1.0091083-turtle.txt
N-Triples: 831-1.0091083-rdf-ntriples.txt
Original Record: 831-1.0091083-source.json
Full Text
831-1.0091083-fulltext.txt
Citation
831-1.0091083.ris

Full Text

Light Activated CI2 Etching of GaAs and Optical Holographic Pattern Formation by Michael Brian Whitwick B.Sc. in Physics, The University of Northern British Columbia, 2000 A THESIS=SUBMITTED IN P A R T I A L F U L F I L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E OF M A S T E R OF SCIENCE in T H E F A C U L T Y OF G R A D U A T E STUDIES (Department of Physics and Astronomy) We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y OF BRITISH C O L U M B I A April 24, 2003 © Michael Brian Whitwick, 2003} In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Physics and Astronomy The University Of British Columbia Vancouver, Canada Abstract The photochemical reaction of CI2 gas with (100) GaAs was studied in this work. Thermal effects due to light induced heating of the GaAs substrate were isolated from photochemical effects. Light induced heating was calcu-lated and the corresponding photothermal CI2 etching of GaAs was accounted for. • The temperature dependence of the photochemical etch rate was exam-ined. The light-induced etch rate potentially follows an Arrhenius temper-ature dependence with an activation energy of 0.161 eV. Near 200°C the light-induced etch rate does not increase with increasing temperature. The photochemical etch rate was found to depend linearly on intensity. A possible explanation for the temperature dependence of the photochem-ical etch rates is presented. It is proposed that the desorption of GaCl3 is the important mechanism in the photochemical C^-GaAs etch. GaCl3 is photodesorbed with illumination, which increases the rate of etching. i i i Contents Abstract i i Contents i i i List of Figures v Acknowledgements . xii 1 Introduction 1 1.1 Chemistry of Chlorine Etching of GaAs 7 1.1.1 Thermochemistry 7 1.1.2 Photochemistry . . . 10 2 Etch Rate Measurements and Pattern Etching 16 2.1 In-situ Holographic Patterning 16 2.2 Surface Grating Reference 19 2.3 Optical Etch Rate Monitoring 24 2.4 Differences in Etch Rate Measurements 29 3 Photothermal Etching . 31 3.1 Light Induced Temperature Rise 31 3.1.1 Radiative Cooling Approximation 31 3.1.2 Thermal Diffusion Approximation 33 3.2 Photothermal Pattern Etching . 48 3.3 Etch Rates 52 4 Photochemical Etching 54 4.1 Temperature Dependence of Etch Rates 54 4.2 Surface Morphologies 56 5 Conclusions 59 6 Future Work . 61 Appendix Mercury Arc Lamp Intensity 63 Bibliography 66 List of Figures 1.1 The temperature dependence of the dark etch rate for a GaAs etch with CI2 gas. Left, the temperature dependence of the production of Ga and As reaction products. Right, the tem-perature dependence of the total etch rate. This figure was taken from Su [1] 1.2 Absorption spectra of the halogen gases: (1) F2 at 25°C; (2) C l 2 ' at 18°C; (3) B r 2 at 25°C; (4) I 2 at 70-80°C (5) I 2 at 70-80°C mixed with air at an pressure of 1 atm. This figure was taken from Calvert [2] 1.3 Potential energy curves for the diatomic C l 2 molecule. This figure was taken from Browne [3] vi 1.4 The photoexcitation of electrons to the conduction band in a semiconductor. In this figure the symbol W is the depth of the depletion layer, Lp is the Debye length of the semiconductor, and a - 1 is the light penetration depth. This figure was taken from Rajeshwar [4] 13 1.5 Three situations for a semiconductor - adsorbent interface at equilibrium conditions. The adsorbents surface states over-lap with (a) the semiconductor conduction band, (b) a sur-face state in the semiconductor, (c) the semiconductor valence band. This figure was adapted from Rajeshwar [4]. . . . . . . 14 2.1 Experimental setup for in-situ holographic pattern etching of GaAs with high vacuum CI2 • 17 2.2 A S i 0 2 grating of 5 yum pitch deposited on a (100) GaAs sur-face. This image was taken with an optical microscope at a magnification of x 50. . 20 2.3 A 25x25 /mi 2 A F M tapping mode image (top) and a 25 /im A F M line scan (bottom) of a CI2 etched GaAs surface. The average height of the grating was 163 ± 3 nm 21 1 vii 2.4 Temperature dependence of photochemical and dark C^-GaAs etch rates which were measured with the Surface Grating Ref-erence (SGR) method. The photochemical and dark etch rates were measured simultaneously to reduce any errors in the mea-surement 23 2.5 The GaAs /AlAs /GaAs structure used for the Multiple Layer Reflection (MLR) method of measuring etch rates. . . . . . . 24 2.6 Intensity of the reflected He-Ne Laser light as a function of time during a C l 2 etch of GaAs. The structure of the sample is shown in Figure 2.5, and the etched sample is shown in Figure 3.12. This is an example of data used in the multiple layer reflection method 25 2.7 Experimental setup used to measure Cb-GaAs etch rates with the M L R method 26 2.8 Temperature dependence of photochemical and dark etch rates for CI2 etching of GaAs, measured with the M L R method. For comparison the dashed line is the fit to the light assisted etch rate measured with the SGR method shown in Figure 2.4. . . 27 2.9 Dependence of the GaAs-Cl2 etch rate on the intensity of the U V radiation. A mercury arc lamp was used as the light source. The GaAs substrate was held at a temperature of 165°C and the pressure was 0.1 mTorr. This data was mea-sured using the M L R method 28 viii 3.1 Schematic of the sample holder used for the etch experiments. The holder contacts the sample with three indented platforms and three clips, which are located in the corners of the sub-strate. 32 3.2 The temperature dependence for the thermal conductivity of GaAs. This figure was taken from Adachi [5] 34 3.3 Schematic of the source and sink function used in the thermal diffusion approximation 36 3.4 Calculated temperature rise profiles for a 500 /im thick GaAs substrate illuminated with two interfering laser beams. In (a) the surface (z=y=0) temperature profile is shown. In (b) the bulk (x=y=0) temperature profile is shown 39 3.5 Calculated surface temperature profiles for a 500 /mi thick GaAs substrate illuminated with two interfering laser beams. The optical absorption coefficient is (a) 2.04 x l O 5 c m - 1 and (b) 6.09 x l O 3 c m " 1 40 3.6 Calculated surface temperature profiles for GaAs surfaces i l -luminated with two interfering laser beams. For these cal-culations the substrate thicknesses are (a) 5 /im, (b) 25 /mi, (c) 50 /mi, and (d) 5000 /im. See Figure 3.4 for a 500 /im substrate 41 ix 3.7 Calculated steady state surface temperature rise for a 500 /im thick GaAs substrates illuminated with two Gaussian shaped laser beams. The laser beam radii are (a) 5 //m and (b) 500 /zm 42 3.8 Calculated steady state surface temperature rise for a 500 fj,m thick GaAs substrates illuminated with two Gaussian shaped laser beams. The substrate width are (a) 1.5 cm by 1.5 cm and (b) 6 cm by 6 cm 43 3.9 Calculated surface temperature profiles for a 500 jum thick GaAs surfaces illuminated with two interfering laser beams. In (a) the two beams have an angle of 6° between them resulting in a pitch of 5 /im. In (b) the two beams have an angle of 1° between them resulting in a pitch of 30 /xm 44 3.10 Calculated temperature rise for a semi infinite GaAs substrate illuminated with two Gaussian shaped laser beams. In (a) the surface (z=y=0) temperature profile is shown. In (b) the bulk (x=y=0) temperature profile is shown 45 3.11 Calculated temperature rise for a two dimensional GaAs sub-strate illuminated with two Gaussian shaped laser beams for 16.6 ms. This is the surface profile (z=0) 47 3.12 A n optical image of a C l 2 etched substrate. This sample was illuminated with a single beam from a He-Ne Laser operating at 632.8 nm. Airy rings were caused by the laser beam pass-ing through an aperture. Differential etching throughout the sample resulted in the observed optical pattern 3.13 The positions of various illuminated areas on the GaAs sub-strate used for the experiments conducted in this work. The M L R method of etch rate measurement typically used these positions of the mercury arc lamp beam and He-Ne laser beams. 3.14 Optical micrograph of a GaAs /AlAs /GaAs sample which was photochemically etched with CI2 gas. U V light from a mercury arc lamp was used for the photochemical etch. The etch was stopped after the dark etch had removed the uppermost GaAs layer thus exposing the AlAs layer. The photochemical etch in the U V light spot has removed the uppermost GaAs and AlAs layers exposing the GaAs substrate 4.1 Temperature dependence of the photochemical etch rate mea-sured with the surface grating reference method and the mul-tiple layer reflection method. The arc lamp had a total power of 31.4 mW for these etch rates xi 4.2 Two 2x2 /mi 2 A F M images of CI2 etched GaAs surfaces. A GaAs substrate was etched for 30 minutes The right image is a thermally etched surface; the left image a photochemically etched surface. The RMS roughness of the illuminated and dark surfaces are 1.05 nm and 1.25 nm, respectively 57 A . l A schematic of the optical setup located after the mercury arc lamp. The iris aperture was typically set to have a 3 mm diameter opening. The light beam from the arc lamp is focused in a spot with a diameter of 10 mm at the GaAs substrate. . 63 A.2 The output spectrum of a mercury arc lamp at 200 W. This figure was taken from The Book of Photon Tools [6] 64 Acknowledgements I would like to thank my supervisor Dr. Tom Tiedje for his patience and insight. Credit is due to my friends and colleagues for there help and enter-tainment over the years. Thank you Sebastien Tixier, Anders Ballestad, Jens Schmid, Erin Young, Eric Nodwell, Martin Adamcyk, Dan Beaton, Richard Mar, Jim MacKenzie, and A l Schmalz. And finally I give my gratitude to the lovely Jane X u . Chapter 1 Introduction Etching patterns into semiconductor surfaces is an important process for nanofabrication applications. A l l electronic devices have two and three-dimensional structures within them. The focus of this work is to develop techniques which will etch patterns of a few hundred nanometers in size into semiconductor surfaces, which are compatible with ultra high vacuum (UHV) crystal growing apparatus such as molecular beam epitaxy. A tech-nique compatible with U H V systems would eliminate the exposure of devices to atmosphere during their creation, which would significantly reduce the density of impurities and oxides in the device. This would be useful in such applications as the regrowth on AlGaAs where the removal of A l oxides is difficult. Coherent laser beams can be used to form interference patterns on the surface of a semiconductor. These interference patterns in combination with a light activated chemical (photochemical) reaction can be transferred directly to a semiconductor surface without the use of a photoresist. CI2 gas reacts spontaneously with GaAs at elevated temperatures. This Cl2-GaAs etch can be enhanced with the addition of light on the GaAs surface. CI2 has the potential to etch inference patterns into GaAs. There have been other studies into the GaAs-Cl2 reaction. Furuhata [7] examined the GaAs-Cl2 reaction in the pressure range from 1 0 - 4 Torr to I O - 3 Torr, and in the temperature range from 100°C to 730°C. Furuhata observed a temperature dependence of the etch rate that is similar to Fig-ure 1.1, with an activation energy of 16.0 kcal-mol - 1 (0.69 eV) above 450°C and 10 kcal-mol - 1 (0.43 eV) below 150°C. The etch rate was found to in-crease linearly with pressure. Furuhata proposed that the GaAs-Cl2 reac-tion is limited by GaCl desorption, CI2 arrival rate and GaCl3 desorption in the temperature ranges above 450°C, between 150°C and 450°C, and below 150°C respectively. The work of Su [1] provides more detail on the GaAs-Cb etch mecha-nism. The temperature dependence of the etch rate was measured over a temperature range of 300°C to 1000°C. The temperature dependence of the etch rate is shown in Figure 1.1. Significantly, Su [1] identified the different reaction products and gives their flux as a function of temperature, which is shown in Figure 1.1. Works such as Tejedor [8] studied the GaAs-Cl2 photochemical reaction with a excimer laser. In Tejedor an ArF excimer laser (A = 193 nm) with laser fluence ranging from 10 to 150 mJ-cm 2 was used to etch (100)GaAs at sub-strate temperatures ranging from 25°C to 200°C and CI2 pressures between 1 Torr to 5 x l 0 - 3 Torr. It was observed that the etch rate increased linearly with increasing CI2 pressure and laser fluence. The etch rate was found to be thermally activated with an activation energy of 0.338 eV. Tejedor observed that the etch rate decreased with increasing n-type carrier concentration. The surface morphology was also studied. Mirror like surface were observed for laser fluences below 23 m J - c m - 2 . Tejedor concluded that the overall etch-ing process is kinetically controlled by the evaporation of GaCl3 and that the addition of 193 nm light excites the GaCl3 to a energetic state with a higher desorption rate. The work of Foulon [9] is another study on the photochemical GaAs-Cl2 etch. In this work a K r F excimer laser (A = 248 nm) was used to cre-ate a projected pattern on the (100) GaAs surface. The laser was passed through a quartz photo-mask to create a pattern. The etch rate in Foulon increased linearly with increasing laser fluence. The light induced GaAs-Cl2 etch was determined to be caused by laser induced heating of the substrate. Calculations in Foulon showed that the GaAs surface was heated to a tem-perature above 690 K with illumination. It is believed that G a C l x products are desorbed by laser induced heating. An activation energy of 0.35 eV was measured for the GaAs-Cl2 reaction. Similar studies can be found in the works of Takatani [10] and Maki [11]. In Takatani A r F and K r F excimer lasers were used to etch (100)GaAs and (100)AlGaAs. For this study the etch was conducted at room temperature with l ( T 5 - l ( r 3 Torr of C l 2 gas pressure. The etch rate increased linearly with increasing laser fluence. The ArF laser (A = 193 nm) had an etch rate 10 times greater than the etch rate of the K r F laser (A = 248 nm). In Maki an A r F excimer laser (A = 193 nm) was used to etch (100) GaAs with C l 2 . Mirror like surfaces were observed for laser fluences below 15 m J - c m - 2 . The study focused on the effects of pulse repetition rate on the etch rates. Maki [11] concluded that the GaAs surface only increased in temperature by 25 K and that the etch could only be caused by a photochemical process. Some studies such as Haase [12] and Shih [13] examined the GaAs-C l 2 reaction at low substrate temperatures. For Haase [12] a (HO)GaAs substrate was cooled to 85 K and exposed to a C l 2 flux and light from a ArF(193 nm), KrF(248 nm), or XeF (351 nm) excimer laser. The reaction products that remained on the GaAs surface were then measured. The A r F excimer laser was 20 times more effective in the creation of Ga-Cl reaction products. Haase proposed that the photochemical reaction is caused by the light induced transfer of electrons to an adsorbed C l 2 layer. The C l 2 is then dissociated by the addition of the electron, which results in the formation of highly reactive chlorine radicals on the surface. Shih etched (100) GaAs at temperatures between 120-150 K and 1-40 Torr of C l 2 gas pressure with illumination from an ArF(193 nm) excimer laser. The etch rate was found to increase linearly with increasing pressure and laser fluence up to 30 m J - c m - 2 . It was observed that the etch rate decreased with increasing substrate temperature. Shih proposed that the physisorbed CI2 is photodissociated creating a local source of CI radicals. The G a C l x products are then desorbed by illumination with the laser. Ashby [14] examined the photochemical etch of (100) GaAs with CI radi-cals. A continuous wave A r + laser (A = 514.5 nm) was used as a light source. The etch rate increased linearly with laser power. Ashby concluded that the light induced etch was photochemical in nature. In the work of Ishii [15] the third harmonic of a Thsapphire laser (200-290 nm) was used to probe the GaAs-Cl2 surface reaction. A GaAs sample at room temperature was irradiated with U V light. The amount of C l 2 adsorbed onto the surface was then detected. CI2 is chemisorbed into the surface by the formation of GaCl3 or GaCl . When light that was illuminating the GaAs surface had a wavelength of 250 nm the amount of chlorine adsorbed on the GaAs surface significantly increased compared to other wavelengths. Also at wavelengths below 200 nm the amount of chlorine adsorbed on the GaAs surface was greater than above 200 nm. Ishii concluded that the GaAs-Cb had the following mechanisms. First physisorbed CI2 is photodissociated with illumination. Second the CI radicals are chemisorbed into the GaAs surface via an electron transfer from the GaAs. Finally the GaCl^ products are thermally desorbed. Any halogen gas has the same electronic structure as chlorine and could potentially have a light induced reaction with GaAs similar to CI2. Brewer [16] observed a light induced reaction for GaAs with CH^Br and CFsBr. Brewer [17] observed a photochemical etch with HBr. In both works the light source was an A r F excimer laser (193 nm). Ehrlich [18] reported light assisted etching of GaAs with CHsBr, CF3I, and CH3CI, and a A r + laser frequency doubled to 277.2 nm. The etching of GaAs with CCI4 has been reported in Takai [19] Despite the extensive study on the photochemical GaAs-Cl2 etch the de-pendence on temperature, pressure, intensity and wavelength is not com-pletely understood. This work will characterize the photochemical CI2 etch-ing of GaAs to find ideal nanofabrication conditions for the Cl 2 -GaAs reac-tion. The characterization of the Cl2-GaAs photochemical reaction will be con-ducted by studying the photochemical etch rates and surface morphology as functions of substrate temperature, wavelength and intensity of the incident light. The photochemical etch rate's dependence on wavelength can indicate which wavelength dependent photochemical reactions are important. Most of the studies on the CI2 photochemical etch have been conducted in the ultraviolet (UV) region. The dependence of the etch rate on light intensity has been throughly examined, though there has been little attempt to isolate photochemical effects from light induced thermal effects. This work will confirm that a photochemical reaction does occur. A study of the etch rate's temperature dependence will remove confusion between light-induced thermal effects and photochemical effects. Also the temperature dependence might reveal the mechanism of the photochemical reaction. This thesis has the following structure. An introduction to C^-GaAs chemistry will be given in the remainder of this chapter. Chapter 2 will describe the experimental methods used to measure the etch rate of GaAs with CI2. In Chapter 3 the effect of laser induced heating on GaAs will be distinguished from the CI2 photochemical etch. In Chapter 4 the effects of the substrate temperature and light intensity on the photochemical Cl2-GaAs reaction will be discussed. 1.1 Chemistry of Chlorine Etching of GaAs 1.1.1 Thermochemistry The temperature dependence of the GaAs-Ci2 etch rate shown in Figure 1.1 can be separated into four separate temperature regions. The first region, Region I ranges from room temperature to 200°C. In this region the etch rate follows an Arrhenius dependence on temperature: R(T) = A-PChe-& (1.1) where Pci2 is the CI2 pressure, A is an etching constant, and Ea is the activation energy of the reaction. For Region I the rate limiting step is 300 400 500 600 700 800 900 400 500 600 700 800 900 1000 SURFACE TEMPERATURE (K) SURFACE TEMPERATURE (K) Figure 1.1: The temperature dependence of the dark etch rate for a GaAs etch with CI2 gas. Left, the temperature dependence of the pro-duction of Ga and As reaction products. Right, the temperature dependence of the total etch rate. This figure was taken from Su[ l ] . believed to be the thermal desorption of GaCl3 for the GaAs surface, which suggested by such works as Furuhata [7]. The activation energy measured in Furuhata [7] is similar to the desorption energy of GaCl3. From such works as Su [1] and Simpson [20] it is known that GaCl3 is the primary Ga reaction product in Region I. The primary As reactant products are AS4 and ASCI3 for this temperature region. Arsenic species are known to be highly volatile and would not slow the C^-GaAs reaction. Several chemical reactions for Region I have been proposed by such works as Maki [11], Tejedor [8], and Foulon [9]. Presented here is a summary of these works. The Cl2-GaAs reaction that occurs starts with the physisorption of CI2 gas onto the GaAs surface. After a period of CI2 surface diffusion the CI2 is chemisorbed into the GaAs surface: GaAs(s) + Cl2(g) -»• GaCl(s) + AsCl(s) (1.2) Reaction 1.2 involves the reaction of physisorbed CI2 with dangling GaAs surface bonds. Additional chlorine is needed to break the internal GaAs bonds: GaCl{s) + AsCl{s) + 2Cl2 -> GaCl3{s) + AsCl3{g) (1.3) As shown in Figure 1.1 AS4 production beings in the upper temperature range of Region I (near 200°C). Thus the following reaction must also occur: 4GaAs(s) + 6Cl2{g) -> 4GaCl3{s) + As4(g) (1.4) The final step is the thermal desorption of GaCl3: GaCk(s) -> GaCh(g) (1.5) Reaction 1.5 is the rate limiting step for the thermal Cl 2 -GaAs reaction. Region II ranges from 200°C to 325°C. In this region the etch rate has a weak dependence on temperature. The major reaction products are GaCl3 and AS4. The production of A S C I 3 has decreased to zero in this region. As the temperature approaches Region III (325°C) GaCl production begins. In Region II the rate limiting step is believed to be the arrival rate of CI2 at the surface, which was suggested by Furuhata [7]. Region III starts from about 325°C and continues until 400°C. The etch rate in this region follows an Arrhenius dependence similar to Equation 1.1. The major reaction products are AS4, A s 2 and GaCl . AS4 creation is im-portant at the lower temperatures of this region (325°C) and A s 2 creation dominates at higher temperatures (400°C). The thermal desorption of GaCl is the rate limiting process in Region III. Supporting evidence for this is that the activation energy is near the GaCl desorption energy. At 400°C Region IV starts. Here the etch rate depends weakly on tem-perature. As in Region III this reaction rate is believed to be limited by the C l 2 transport to the GaAs surface. Region IV continues until 600°C where Ga begins to evaporate. 1.1.2 Photochemistry Energetic photons can interact with the Cl 2 -GaAs system in many different ways. The gas phase photodissociation of C l 2 to CI radicals would have a significant effect on the Cl 2 -GaAs etch. Chlorine radicals are highly reactive. Photodissociation occurs when a C l 2 molecule which has potential energy curve shown in Figure 1.3 absorbs a photon of energy Ephoton, which is the energy difference between the C l 2 ground state ( £ 3 ) and a dissociated CI + CI state (2P3/2 + 2-F*3/ 2)- From Figure 1.2 gas phase C l 2 photodissociates strongly at 330 nm. Higher energy transitions to a CI + CI excited state are also possible but have much lower probabilities of occurring. The transition 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500 Wavelcngth(A) Figure 1.2: Absorption spectra of the halogen gases: (1) F 2 at 25°C; (2) CI2 at 18°C; (3) B r 2 at 25°C; (4) I 2 at 70-80°C (5) I 2 at 70-80°C mixed with air at an pressure of 1 atm. This figure was taken from Calvert [2]. at 330 nm has an absorption cross section, C\ of 2 . 4 6 8 x l 0 - 1 8 cm 2 . Light absorbed by the CI2 gas will follow an exponential absorption law: I(x)=I0e-ax (1.6) Where I0 is the initial intensity and a is the absorption coefficient at wave-length A. The absorption coefficient is related to the cross section by the following expression: (1.7) Where R is the ideal gas constant, and T is the temperature of the gas. Physisorbed CI2 molecules on the GaAs surface can also be photodisso-ciated. The CI2 molecular state will be altered by the GaAs surface, thus the exact properties for the photodissociation of the physisorbed CI2 are not known. The photodissociation of CI2 is not necessarily the leading photo-chemical effect. Ashby [14] reported the.etch of GaAs with CI radicals and illumination with 514 nm light. Ashby's etch would not be affected by CI2 photodissociation. X 24 1 r \ 1— Ou 2 p + 2 p 1/2 3/2 • 1 20 1 16 A /•' Su -j JO 2 p l 2 p a/2 3/2 12 -8 4 ~/'° -it 1 1 0 10 20 30 40 r(A) Figure 1.3: Potential energy curves for the diatomic C l 2 molecule. This figure was taken from Browne [3]. Photon can directly affect the chemical reaction in the Cl2-GaAs system. In the Cl2-GaAs reaction GaCl and GaCl3 can be thermally desorbed from the GaAs surface. Photons of greater energy than the thermal desorption energy can in principle photodesorb the Ga-Cl molecules off the GaAs surface, thus increasing the C^-GaAs reaction rate. Conduction Band Valence Band Adsorbent Figure 1.4: The photoexcitation of electrons to the conduction band in a semiconductor. In this figure the symbol W is the depth of the depletion layer, LD is the Debye length of the semiconductor, and a"1 is the light penetration depth. This figure was taken from Rajeshwar [4]. In a semiconductor photons can excite valence electrons across the band gap to the conduction band. These photons must have an energy greater than the semiconductor's band gap as shown in Figure 1.4. Illumination effectively increases the Fermi energy at the GaAs surface since there are more electrons in the conduction band. A change in the Fermi surface could have an effect on the GaAs-Cb reaction. Figure 1.5 illustrates three different configurations between the semiconductor electronic states and the states of the C l 2 adsorbed on the GaAs surface. Note the adsorbed CI2 surface states are more complex that 1/a LD+W-•* W Semiconductor Conduction Semiconductor Adsorbent Semiconductor Adsorbent Band Fermi Energy 0) c LU cn Band Gap Surface State Valence Band \ \ \ \ \ \ : (a) Semiconductor Adsorbent (b) \ \ \ \ \ \ \ (c) W W W Figure 1.5: Three situations for a semiconductor - adsorbent interface at shown in Figure 1.5. In a situation where the adsorbent surface states are located at the conduction band raising the Fermi surface will increase the transfer of electrons to the adsorbed chlorine. A l l wavelengths with energy greater than the semiconductor's band gap will photoexcite electrons. However, light with a wavelength in the U V range will have a greater effect on the semiconductor's surface then light with a longer wavelength. A layer that is depleted of electrons will form at the equilibrium conditions. The adsorbents surface states overlap with (a) the semiconductor conduction band, (b) a surface state in the semiconductor, (c) the semiconductor valence band. This figure was adapted from Rajeshwar [4]. surface as the adsorbed molecules will consume electrons for any chemical reactions. A depletion layer is shown in Figure 1.4. U V light has a short absorption length in GaAs, and thus more U V light will be absorbed in the depletion layer which will have more effect on the reaction. Chapter 2 Etch Rate Measurements and Pattern Etching This chapter describes the experimental methods and apparatus used to etch periodic gratings in GaAs and to measure the Cl2-GaAs etch rate. The measurement of etch rate is not a simple matter. A reference surface must be added to the system which adds an extra degree of complexity. There are two methods of etch rate measurements discussed in this work: the Surface Grating Reference (SGR) method, and the Multiple Layer Reflection (MLR) method. 2.1 In-situ Holographic Patterning The creation of nanometer scale patterns on a GaAs surface is required in many industrial and academic applications. The distributed feedback laser is one example where periodic patterns are needed. Most pattern etching in-volves lithographic techniques. Conventional lithography does not integrate easily with M B E processes. Lithography is conducted under standard at-mospheric pressure, and would require moving a sample in and out of the U H V M B E system. Lithographic techniques may add difficult to remove con-tamination to the sample since it involves multiple chemicals and deposited layers. Dry chemical etching techniques are more suited to M B E growth, as such systems are compatible in principle with U H V . However, pattern creation is difficult with dry chemical etching alone. UHV Etch Charter Mirrors Spatial Flter ^ ' ^ Mrrors View Ports Figure 2.1: Experimental setup for in-situ holographic pattern etching of GaAs with high vacuum C l 2 . By combining two or more coherent laser beams one can create an inter-ference pattern on a semiconductor is surface. A light induced reaction can transfer this holographic pattern into the semiconductor. Consider the situa-tion where two Gaussian shaped laser beams of wavelength, A and intensity, I0 are incident to a surface at angles fa and -fa from the surface normal. Assuming that the two beams are coherent and have parallel polarizations a simple superposition of the light's electric fields can be used to determine the intensity profile on the surface. After some mathematical manipulation the intensity profile on the surface is: / ( x , y ) = 27 0[l + c o s ( ^ ^ ) ] (2.1) The pitch of the interference pattern in Equation 2.1 is: A ~ 2sin{<f>i) ^'^ For the experimental setup in Figure 2.1 two sets of equipment were used to produce a holographic pattern. The first setup used a Lexel 3500 Ar Ion Laser operating at 488 nm as the coherent light source. The beamsplitter consisted of a 110 grooves-mm - 1 diffraction grating with a mask to isolate the ± 1 diffraction orders. A reflecting 45° wedge was placed 7 cm after the diffraction grating. The two beams leave the beamsplitter with an angle of 60° between them. The two beam are then reflected towards the sample in the etch chamber with mirrors in front of two high vacuum viewing ports. Each beam has an incident angle to the sample of 45° from the sample's surface normal. According to Equation 2.2 this apparatus produces an interference pattern with a pitch of 345 nm. The second apparatus consisted of a Spectra Physics BeamLok 2060 Ar Ion laser at 488 nm doubled with a Spectra Physics Wavetrain system to 244 nm. The beam diameter is increased to 30 mm using a pinhole spatial filter. A U V 50/50 plate beamsplitter was used to split the beam. The plate beamsplitter is a 5° wedge designed to eliminate multiple reflections from inside the beamsplitter. Mirrors outside the etch chamber's viewing ports reflect the two laser beams towards the sample. The laser beams are directed to the sample at an angle of 45°. The mirrors are coated with U V enhanced aluminum, and the viewing ports on the etch chamber constructed of U V fused silica. This apparatus will create a grating with a pitch of 172.5 nm. This system was not completed and such a pattern has not been observed. 2.2 Surface Grating Reference The first technique used to measure the CUj-GaAs etch rate involve the de-position of a S i 0 2 grating on a GaAs surface using lithographic techniques. The S i 0 2 grating is a zero depth reference during the CI2 etch. After the CI2 etch is complete the etch rate can be measuring from the depth of the CI2 etched GaAs grating (without the S i02 grating) and the time of the etch. This section will describe the experimental methods and apparatus used to measure etch rate with the SGR method. First a SiC>2 grating is deposited on a clean (100) S.I. GaAs substrate, which has a thickness of 500 fim. Before deposition a GaAs quarter wafer is cleaned with a successive acetone, isopropanol, and methanol rinse (5-10 s Figure 2.2: A SiC>2 grating of 5 /mi pitch deposited on a (100) GaAs surface. This image was taken with an optical microscope at a magnifi-cation of x 50. each). Between each rinse the wafer is blown dry with nitrogen gas. 1 /mi of A Z 4210 photoresist is spun onto the GaAs wafer at a rotation speed of 5000 rpm for 50 seconds. The photoresist is left to dry for 30 minutes. After that the sample is exposed to AZ 300 MIF for 3.5 minutes, and softbaked at 100°C for 10 minutes. AZ 300 MIF hardens the photoresist surface so that undercuts in the photoresist will occur when the sample is developed. Undercuts help the Si02 to form sharp edges in the grating pattern. A 5 /mi pitch pattern mask is transferred to the photoresist with a Karl Suess M J B 55 Mask Aligner which exposes the sample to U V light. The U V exposed photoresist is removed with a 4:1 distilled water:AZ 400K developer solution. The sample is developed for 3 to 5 minutes. After a 10 second rinse in a HC1 bath, 40 nm Si02 is deposited on the sample using electron beam evaporation. j 163 nm I r i Figure 2.3: A 25x25 /mi 2 A F M tapping mode image (top) and a 25 /mi A F M line scan (bottom) of a C l 2 etched GaAs surface. The average height of the grating was 163 ± 3 nm The remaining AZ 4210 photoresist is removed in a ultrasonicly agitated acetone or A Z 400K developer bath. Figure 2.2 shows a GaAs sample at this stage in the SGR method, with a Si02 grating on a GaAs surface. After the construction of a S i02 grating on a GaAs substrate the sample is etched with CI2. The CI2 etch experiments were performed in an ultra high vacuum (UHV) etch chamber attached to a V80H molecular beam epitaxy (MBE) system. Prior to etching the sample was outgassed at 200°C then etched with atomic hydrogen to remove the native oxide. A description of the atomic hydrogen etch can be found in Ritchie [21]. The GaAs substrate was held between 23°C and 200°C at a constant pressure of 10~ 4 Torr of C l 2 . The etch chamber's base pressure was on the order of 10~ 9 Torr. A 150 W mercury arc lamp was used as the U V light source. The U V light was focused onto a 1 mm diameter spot with the optics shown in Figure A . l . The incident beam was angled 65° from the surface normal of the substrate. The mercury arc lamp spot was directed onto an identifiable area of the sample. This enabled photochemical and dark etch rates to be measured at the simultaneously. After the CI2 etch the S i02 grating is removed with a 5-10 min rinse in a bath of 47% F£F in H2O. The S i 0 2 free GaAs surface is measured ex-situ with an Digital Instruments Nanoscope atomic force microscope (AFM) in tapping mode. Figure 2.3 shows an example of an A F M image of a patterned GaAs surface. The etch rate can be easily determined from the measured depth of the etched GaAs grating and the time of the etch. The photochemical etch rate is defined for this work as the total etch rate of the U V illuminated spot minus the etch rate of the dark area. The SGR method provides a straightforward method for the determination of etch rates. Figure 2.4 shows the dependence of the experimental etch rate on tem-perature measured with the SGR technique. The dark etch rate seen in Figure 2.4 follows an Arrhenius dependence as shown in Equation 1.1. A least squares fit of the dark etch rate in Figure 2.4 yields Ea — 0.29 ± 0.03 eV and A = 4.7 ±0.3 x 10 8 nm-min- 1 Torr- 1 . i I I | I I U 2.2 2.4 2.6 2.8 3 3.2 3.4 1/T(1000/K) Figure 2.4: Temperature dependence of photochemical and dark C V G a A s etch rates which were measured with the Surface Grating Ref-erence (SGR) method. The photochemical and dark etch rates were measured simultaneously to reduce any errors in the mea-surement. The photochemical etch rate in Figure 2.4 has a weaker temperature dependence then the dark etch rate. A least squares fit of Equation 1.1 yields Ea = 0.16 ± 0.01 eV and A = 1.98 ±0.33 x 106 nm-min^Tor r - 1 . However, the temperature dependence of photochemical etch rate could potentially follow a linear or power law relationship with temperature. 2.3 Optical Etch Rate Monitoring A faster and more accurate method for measuring etch rates is the multiple layer reflectance (MLR) method. In the M L R technique the interference be-tween laser light reflected off the front surface and additional buried interfaces of a semiconductor is measured. The intensity of the reflected beam depends on the layer thickness, and the optical wavelength. As the uppermost layer is etched away the change in the reflected beam intensity is measured over time. The etch rate can be inferred from the time dependence of the reflec-tivity. A GaAs /AlAs /GaAs structure as shown in Figure 2.5 was chosen for the experiments performed in this work. Figure 2.5: The GaAs /AlAs /GaAs structure used for the Multiple Layer Re-flection (MLR) method of measuring etch rates. The reflected beam is a combination of the multiple reflections from vari-ous interfaces. These multiple reflected beams will interfere with each other. 0.12 0.11 l§ 0.08 Q. < 0.07 c ' I o 0.06 o 0.05[ 0.04 Noise Due to Operator I Alignment End Etch 10 20 30 40 time (min) 50 60 Figure 2.6: Intensity of the reflected He-Ne Laser light as a function of time during a C l 2 etch of GaAs. The structure of the sample is shown in Figure 2.5, and the etched sample is shown in Figure 3.12. This is an example of data used in the multiple layer reflection method. The total intensity of the reflected beam will depend on the thicknesses of the layers which reflect the laser beam. Decreasing the uppermost layer at a constant rate will cause the reflected beam's intensity to oscillate as shown in Figure 2.6. The period of this oscillation is: A r sin(6i)^ A = cos\arcsin( ) 2ns 1 K ns n (2.3) Where ns is the index of refraction for GaAs, f?j is the angle of incidence, and A is the wavelength for the laser. 26 Figure 2.7: Experimental setup used to measure Cl 2 -GaAs etch rates with the M L R method. The GaAs /AlAs /GaAs structure shown in Figure 2.5 is grown with the following procedure. First a semi-insulating GaAs substrate is outgassed at 200°C, then it is heated to 600°C with a constant As2 pressure of 6x 10~ 6 Torr to remove the native oxides. The oxide-free substrate is lowered to 550°C and the growth of a GaAs buffer layer is started at this point. During this growth the sample is heated to 600°C where it remains for the rest of the growth. This growth procedure buries impurities and smoothes the rough GaAs substrate. The buffer layer is 50-100 nm think. Now 50 nm of AlAs is ground followed by 300 nm of GaAs. Again the CI2 etch experiment was performed in a U H V chamber, with a setup shown in Figure 2.7. A He-Ne (632.8 nm) laser was used to measure X I I I I I I 2.1 2.2 2.3 2.4 2.5 2.6 1/T(1000/K) Figure 2.8: Temperature dependence of photochemical and dark etch rates for CI2 etching of GaAs, measured with the M L R method. For comparison the dashed line is the fit to the light assisted etch rate measured with the SGR method shown in Figure 2.4. the etch rate with the M L R method. A 110 groove/mm diffraction grating was used to split the Ffe-Ne laser beam. Each beam had a power less than 5 mW. The two beams were directed to the sample so that photochemical and dark etch rates could be measured simultaneously. Beam intensities were measured with two Si photodiodes. A chopper and lock-in amplifier system isolated the Ffe-Ne laser beams from any background light. The intensity of the mercury arc lamp was adjusted by moving the final lens of the arc lamp's optics shown in Figure A . l . 10 ~ i 1 1 i E y -(nmy -CD I 8 7 i i i 0 20 40 60 80 Intensity (mW/cm2) Figure 2.9: Dependence of the GaAs-Cl2 etch rate on the intensity of the U V radiation. A mercury arc lamp was used as the light source. The GaAs substrate was held at a temperature of 165°C and the pressure was 0.1 mTorr. This data was measured using the M L R method. The etch rate shown in Figure 2.8 and Figure 2.9 were measured with the M L R method. The Appendix describes the method used to measure the arc lamp's intensity. Again as expected the dark etch rate follows an Arrhenius dependence temperature. A least-squares fit of the thermal etch rate to Equa-tion 1.1 yields Ea = 0.21 ± 0.01 eV and A = 2.4±0.1 x 10 7 nm-min^Torr - 1 . The etch rate's dependence on intensity is shown in Figure 2.9. A least squares fit to a linear function yields a slope of 0.039 c m 2 n m - m W _ 1 m i n _ 1 . 2.4 Differences in Etch Rate Measurements There are significant differences in the etch rates measured with the SGR method and the M L R method. The first critical difference is that the M L R method measures a dark etch rate that is approximately half of etch rate measured with the SGR method. The activation energy measured with the SGR method is E a = 0.29 ± 0.03 eV which is slighty different from the acti-vation energy measured with the M L R method at 0.21 ± 0.01 eV. Likewise the etching constant, A differs between the two methods. The SGR method measured 4.7 ±0.3 x 10 8 nm-min - 1 Torr - 1 , while with the M L R method mea-sured 2.4 ± 0 . 1 x 107 n m m i n _ 1 T o r r _ 1 . One possible explanation for this difference is presented here. The physisorption of CI2 on the S i02 mask is a potential cause for the difference between the SGR and M L R measurements. In the SGR method 50% of the substrate's surface is covered with Si02- In the temperature range at which the etching experiments are conducted CI2 molecules might physisorb onto the Si02- If the surface diffusion length for physisorbed CI2 on S i 0 2 was on the order 5 /mi the CI2 could migrate from the inert S i 0 2 areas to the GaAs surface where chemisorption could occur. Thus the SGR method could effectively double the amount of CI2 on the GaAs surface compared to the M L R method. The desorption of GaCl3 could also cause a difference in the two mea-surements. If the GaCl3 has a large diffusion length greater than 5 fxm on an GaAs the GaCl3 could move to the S i02 where no new GaCl3 is being formed. Thus the surface area available from which GaCl3 can evaporate is effectively doubled. This is the most likely case for the difference between the two measurements, as GaCl3 desorption is the rate limiting step in the temperature region of the experiments. If the diffusion lengths of the CI2 or the GaCl3 were significantly less than 5 /im the CI2 etched GaAs grating would develop trenches next to the Si02 strips. These trenches would be caused by the excess of CI2 diffusion from the S i 0 2 strips or the removal of the GaChj to the Si02- However, no trenches have been observed. The second potential difference between the M L R and SGR methods is that the photochemical etch rate measured with the two methods follow different temperature dependences as shown in Figure 4.1. This difference could be due to an unaccounted scatter in the measurement of the light assisted etch rate. The light assisted etch rate measured with the SGR method varies significantly from a clear pattern and could potentially be equivalent to the etch rate measured with the M L R method. Chapter 3 Photothermal Etching A significant temperature increase due to illumination would cause a notice-able increase in the etch rate since the Cl2-GaAs dark etch rate has a strong dependence on temperature. An increase in the etch rate due to light in-duced heating could be confused with a photochemical effect. This chapter presents two different calculations for determining the temperature rise on a GaAs sample under illumination and will discuss the the effect of an increase in temperature. Finally the photochemical effects will be distinguished from any photothermal effects. 3.1 Light Induced Temperature Rise 3.1.1 Radiative Cooling Approximation The first calculation used to determine the temperature rise on the GaAs surface due to illumination considers only radiative cooling of the substrate and assumes that the temperature of the substrate is uniform. The radiative cooling approximation is not completely valid as the sample is in thermal contact with the sample holder, as shown in Figure 3.1. However the radiative cooling approximation does provide an upper bound on the temperature rise. Sample Holder Figure 3.1: Schematic of the sample holder used for the etch experiments. The holder contacts the sample with three indented platforms and three clips, which are located in the corners of the substrate. The radiative cooling approximation simplifies the heat equation to: f = ^[Gtot-eeA(T*-T2)] (3.1) Where C is the heat capacity, e is the emissivity which was taken from Timans [22] as 0.09, V is the volume, A is the surface area of the GaAs substrate, a is the Stefan Boltzmann constant, TQ is the temperature of the surroundings, and Gtot is the total power absorbed in the substrate. The substrate's temperature rapidly reaches a steady state temperature of: Teq = [ T A 0 + ^ (3.2) Typically the mercury arc lamp had a total absobed power of 31.4 mW as estimated in the Appendix. The two He-Ne laser beams used in the M L R method had a combined power of 5 mW. With a power of 37.4 mW and a initial temperature of 125°C the equilibrium temperature will be 146°C. The total temperature rise is 21°C. The temperature rise would range from 14°C (T 0 = 190°C) to 67°C (T 0 = 23°C), depending on the initial temperature of the substrate and it's surroundings. These temperature increases are upper limits as conduction cooling is neglected. 3.1.2 Thermal Diffusion Approximation Another calculation of the light-induced temperature rise considers a non-uniform substrate temperature in perfect thermal contact with the sample holder while ignoring radiative cooling. For this thermal diffusion approxi-mation the following heat equation must be solved: Where K is the thermal conductivity of GaAs, and G(x,y,z) is the energy absorbed per unit volume in the substrate. The temperature dependence of the thermal conductivity is shown in Figure 3.2. The thermal diffusion approximation assumes that the rise in temperature is not significantly large such that the thermal conductivity remains constant throughout the sub-strate. Since the thermal diffusion constant is large for GaAs the solution to Equation 3.3 will quickly approach a steady state = 0). 100 10 e u 1 ' 0.1 0.01 1 10 100 1000 T ( K ) Figure 3.2: The temperature dependence for the thermal conductivity of GaAs. This figure was taken from Adachi [5]. The boundaries of the system are located at the surfaces of the GaAs substrate and the areas where the sample is in thermal contact with the sample holder. At the GaAs surface it is assumed that the heat transferred to the CI2 gas is significantly less than the heat transferred to the sample holder. The first boundary condition is that no heat is transferred to the CI2 gas: FfT I = 0 (3.4) dT ~dZ \z=0,D Where D is the thickness of the substrate. It is assumed that the thermal conductivity and surface area of the sample holder are large enough such that the temperature of the sample holder is constant. The second boundary condition is that T = T 0 at the areas of thermal contact, which are taken to be at x = and y = ± y where L is the width of the substrate. It is useful to write T(x,y,z) as a three dimensional Fourier series that satisfies that boundary conditions of the system: 00 00 00 T T . . v — \ — \ y — \ , ,TVK . L . . . ,777.71". L . . . S7T . T(x,y,z) = 2^1^ 2 ^ A n ^ 9 S m ^ ^ X ~ 2 ^ S m ^ ^ y ~ 2 ^ C O S ^ I J Z ^ 0 n=0 m=0 s=0 (3.5) Where Ant7ntS will be determined from Equation 3.3. Note the origin in Equation 3.5 is located at the center of the substrate. For illumination with two interfering Gaussian shaped coherent laser beams G(x,y,z) is: G(x, y, z) = 4 a / 0 ( l - i ^ e ^ e - ^ ^ cos2(xA; sin Oi) (3.6) Where I0 is the intensity at the center of a single laser beam, w is the beam ra-dius, 0i is the half angle between the two laser beams, and k is the wavenum-ber; R is the reflectivity of the GaAs surface, and a is the optical absorption coefficient of GaAs. Let 7 = 2aIo^~R^ for convenience. Heat Source i A . 1 / \ | 1 ' J 1 ; Real Substrate \ : \ I ! / •« / . •( \ I i / Extended Substrate \ 1 ^ ^ \ i "~ — — I ( " Heat Sinks -"" • 1 1 \ / / \ V Figure 3.3: Schematic of the source and sink function used in the thermal diffusion approximation. Consider the one dimensional heat equation with a heat source of G(y) — and the boundary condition that T = 0 at y = ±j. The solution to this will be equivalent to the solution of a heat equation which has the same heat source plus two additional mirror heat sinks that are located at the edges of an extended substrate. The extended substrate has a width that is twice the width of the real substrate, as shown in Figure 3.3. G(y) - e t? - e - e (3.7) By symmetry the boundary condition that T ( ± ^ ) = 0 will be satisfied. This method of extending the substrate and adding heat sinks can be applied to 37 three dimensions. The x component of G(x,y,z) can be written as: G(x) = cos2 (xk sin Oi) - e~^{x+L)2 cos2((x + L)ksm9i) cos2 ((a: — L)k sin 9i) (3.8) - e - M x - L y p n s i ( ( r _ For the z direction the boundary condition shown in Equation 3.4 is applied. It is useful to extend the substrate to range from -D to D and allow G(z) to be symmetric around z = 0 so that T(x,y,0) satisfies Equation 3.4. G(z) = { (3.9) jKe~az if z > 0 jKeaz if z < 0 Now is convenient to write G(x,y,z) as a three dimensional Fourier series: 00 00 oo G(x,y,z) = }^anbmcssin(^(z-^))sin(^(y-^))cos(^)] n=0 m=0 s=0 (3.10) Where the Fourier coefficients an, bn, and cn are given by: L 2sjHw sin(^f) (wKns2 _r K-AJ- e I 2L J 2 a D ( l - (-l)ne~aD) (aD)2 + (Trn)2 (3.11) Thus using Equation 3.10 and Equation 3.5 in Equation 3.3 A „ ) m j S is deter-mined to be: 7anbmcs An,m, s (™)2 + ( ™ ) 2 + (™)2 (3.12) For n+m+s ^ 0 as A),o,o = 0. Thus the temperature profile can be deter-mined. Some examples of calculated temperature rises due to illumination with two interfering laser beams are shown in Figures 3.4, 3.5, 3.6, 3.8, and s 3.9. Unless stated otherwise the calculations shown in this work used the following parameters. Each laser beam had a Gaussian shape with a beam radius of 50 /mi, a wavelength of 514.5 nm, and a power of 14.4 mW. The two beams had an angle of 3° between them resulting in a interference pattern with a pitch of 10 /mi. The GaAs substrate had the dimensions of 3 cm by 3 cm by 500 /im. The temperature of the substrate holder was 100°C. The adsorption coefficient and reflectivity for light with a wavelength of 514.5 nm in GaAs are 9.01 x l O 5 c m - 1 and 0.326. The thermal diffusion approximation gives significantly lower temperature rises than the radiative cooling approximation. This indicates that radiative cooling of the GaAs substrate is not significant compared the to diffusion of heat into the sample holder. A n example of the dependence of the temperature rise on the optical absorption coefficient, a is shown in Figure 3.5. The optical absorption coefficients of GaAs are 2.04 x lO cm" _ 1 and 6.09 x l O 3 cm for light of wavelengths 244 nm and 826 nm respectively. The maximum temperature rise is slightly larger with decreasing absorption depth. In Figure 3.5a the maximum temperature rise is 2.5°C, while in Figure 3.5b the maximum tem-perature rise is 2.6°C. The oscillations in the temperature profile also increase -5 0 5 10 15 0 0.1 0.2 0.3 0.4 0.5 (a) X-Axis (mm) (b) Z-Axis (mm) Figure 3.4: Calculated temperature rise profiles for a 500 /im thick GaAs substrate illuminated with two interfering laser beams. In (a) the surface (z=y=0) temperature profile is shown. In (b) the bulk (x=y=0) temperature profile is shown. with decreasing absorption depth. For a large optical absorption coefficient more light will be absorbed near the GaAs surface, as a result the interference oscillations of the illumination have more effect on the temperature. Figure 3.6 shows several examples of how the temperature rise depends on the substrate thickness. For substrates that are significantly thicker than the incident beam radius the maximum temperature rise will have a weak dependence on the thickness. For substrates that are thinner than the in-cident beam radius the maximum temperature rise depends significantly on thickness. However, the dependence on the thickness is not trivial. The tem-perature rise will go to infinity as the thickness goes to zero. The oscillations in the temperature rise increase as the thickness increases. As the thickness is increased more heat can flow into the bulk GaAs along the z direction. -lob • -50 0 50 100 -loo -50 0 50 100 (a) X-Axis (nm) (b) X-Axis (nm) Figure 3.5: Calculated surface temperature profiles for a 500 /.mi thick GaAs substrate illuminated with two interfering laser beams. The optical absorption coefficient is (a) 2.04 x l O 5 c m - 1 and (b) 6.09 x l O 3 cm" 1 . With more heat flowing away from the surface thermal diffusion will wash out less of the interference oscillations. The effect of the laser beam radius on the temperature rise is shown in Figure 3.7. An increase in the beam radius by a factor of 10 results in a decrease in the maximum temperature rise by a factor of 5. While an decrease in the beam radius by a factor of 10 results in a increase in the maximum temperature rise by a factor of 12. The size of the beam radius has little effect on the temperature oscillations. The effect of the substrate width on the temperature rise is shown in Figure 3.8. The width has little effect on the temperature rise. Increasing the width of the substrate moves the heat sinks at the boundaries away from the light source resulting in a slight increase in temperature. (c) X-Axis (Mm) (d) X-Axis (|Jm) Figure 3.6: Calculated surface temperature profiles for GaAs surfaces illumi-nated with two interfering laser beams. For these calculations the substrate thicknesses are (a) 5 /mi, (b) 25 /mi, (c) 50 /mi, and (d) 5000 /mi. See Figure 3.4 for a 500 /tm substrate. Figure 3.9 gives an example of the calculated surface temperature pro-file is effected by the pitch of the interference pattern. The oscillations in the temperature profile decrease as the interference pattern pitch decreases. Thermal diffusion washes the oscillations out for smaller pitched interference patterns. Figure 3.2 shows the thermal conductivity as a function of temperature. The temperature rise is inversely proportional to the thermal conductivity. -100 -50 0 50 100 5 I . . 1 1 I • • • 1 -10 -5 0 5 10 -1000 -500 0 500 1000 (a) X-Axis(ujn) (b) X-Axis(um) Figure 3.7: Calculated steady state surface temperature rise for a 500 fim thick GaAs substrates illuminated with two Gaussian shaped laser beams. The laser beam radii are (a) 5 //m and (b) 500 nm. Raising the substrate temperature (above 10 K) will lower the thermal con-ductivity and increase the temperature rise on the surface. The temperature rise has a dependence that is directly proportional to I0 and (1-R). For the situation where the substrate is semi infinite (L^> 1, D ^> 1) Equation 3.3 can be solved with a different method. In Lax [23] the case of a Gaussian shaped laser beam heating a semi infinite substrate has been dealt with. Other beam shapes can be solved following similar procedures with varying degrees of difficulty. To start consider a solution of the form T(x, y, z) = rye~azV(x, y), this will transform Equation 3.3 to: (V 2 + a2)V(x, y) = - e _ £ ¥ cos2{xksin9i) (3.13) -100 -50 0 50 100 -100 -50 0 50 100 (a) X-Axis dim) (b) X-Axis (jim) Figure 3.8: Calculated steady state surface temperature rise for a 500 /im thick GaAs substrates illuminated with two Gaussian shaped laser beams. The substrate width are (a) 1.5 cm by 1.5 cm and (b) 6 cm by 6 cm. The two dimensional Fourier transform of Equation 3.13 is: /oo roo / dVdfe-i^+y^V(r1,O^2-V2-e) + G(T1,O} = 0 (3.14) -oo J —oo Where V and G are the Fourier transforms of V(x,y) and G(x,y). Now G is: G(V,0 = ^ e - ^ ^ e - ^ ^ e - ^ ^ ^ ^ V e - T ^ s i n ^ O ) 2 ] ( 3 . 1 5 ) To satisfy Equation 3.14 we need: V(V,t) = - 2 G { T 1 2 0 C 2 (3-16) Equation 3.16 is the special solution to the heat equation, and does not satisfy the boundary conditions of the system. A solution to the Laplace equation (V 2 T=0) must be added and can be written in the following form: -loo -50 0 50 100 -fOO -50 0 50 100 (a) X-Axis (nm) (b) X-Axis (urn) Figure 3.9: Calculated surface temperature profiles for a 500 fj,m thick GaAs surfaces illuminated with two interfering laser beams. In (a) the two beams have an angle of 6° between them resulting in a pitch of 5 fj,m. In (b) the two beams have an angle of 1° between them resulting in a pitch of 30 ^m. T h = S I ( 3 ' 1 7 ) Where S(r),£) can be found from the boundary condition of the system. The boundary condition in the semi infinite substrate case is that no heat is transferred to the C l 2 gas across the GaAs surface at z =0. Thus S(?7, £) becomes: « ( , , { ) — ( 3 . 1 8 ) The calculation of the temperature rise profile is reduced to an integral which when written in polar coordinates is: '0 Jo Jo  a ~ P (3.19) A n example of Equation 3.19 is shown is Figure 3.10. Equation 3.19 predicts similar temperature profiles to the finite substrate calculations at the location of the light source. However away from the light source the finite substrate calculation predicts higher temperatures. The semi infinite approximation predicts greater oscillations in the temperature profile due to illumination with an interference pattern than the finite thickness calcula-tions. In the finite substrate calculation heat must diffuse towards the edge of the substrate which will wash out most of the temperature oscillations on the surface. However, in semi infinite approximation heat diffuses isotrop-ically into the bulk of the infinite substrate reducing the heat flow in the lateral directions, which maximizes the effect of the interference oscillations on the surface temperature profile. \ 2.65 \ \ 2 5 5 \ \ 2 4 5 \ N ' 0 (1.2 0.4 0.6 0.8 1 N. Z-Axis (nm) (a) 0 5 10 X-Axis (mm) 0.1 0.2 0.3 0.4 Z-Axis (mm) 0.5 Figure 3.10: Calculated temperature rise for a semi infinite GaAs substrate illuminated with two Gaussian shaped laser beams. In (a) the surface (z=y=0) temperature profile is shown. In (b) the bulk (x=y=0) temperature profile is shown. The oscillations in the temperature rise can be increased by changing some of the parameters of the finite substrate calculation. The most effec-tive method for maximizing the temperature oscillations is to increase the pitch of the illuminating interference pattern However smaller pitched grating are more desired which limits this method of increasing the thermal oscil-lations. Increasing the thickness of the substrate also increases the thermal oscillations. The semi infinite substrate calculation predicts the largest oscil-lations. Of coarse an thick substrate is not always experimentally practical. The adsorption depth also effects the oscillations. With U V light, which has a short adsorption depth more light is absorbed near the surface increasing the effect of the interference oscillations on the surface temperature. The effect is not that significant though and produces only marginal gains. A short pulse of light could produce significant oscillation in surface tem-perature without heating the substrates bulk. A single pulse followed by a period of cooling would reduce the general heating that eliminates tem-perature oscillations in the case of a continuous beam. The heat equation (Equation 3.3) can be solved numerically by iteratively calculating the tem-perature: Ti = Ti-i + At^ V 2 r i _ ! + ^G(x, y, z) (3.20) Figure 3.11 shows an example of an two dimensional GaAs substrate illumi-nated is a 16.6 ms pulse of light. The substrate had a width of 300 jxm and -50 0 50 X-Axis dim) 150 Figure 3.11: Calculated temperature rise for a two dimensional GaAs sub-strate illuminated with two Gaussian shaped laser beams for 16.6 ms. This is the surface profile (z=0). thickness of 5 /j,m. The illuminating spot had a beam radius of 50 //m, a pitch of 10 fj,m, and supplied a power of 28.8 mW. The adsorption coefficient and reflectivity of the substrate were taken as 9.01 x l O 5 c m - 1 and 0.326 respec-tively. In Figure 3.11 the temperature oscillations are still independent of each other providing a large variation in surface temperature. The difficulty that occurs by using short pulses of light to thermally etch patterns of light in a substrate is that a thermal etch typically cannot be stopped quickly and would continue to etch during any cooling periods. 3.2 Photothermal Pattern Etching With a light induced temperature rise similar to that of Figure 3.9 periodic patterns can be etched into GaAs with thermal etching alone. However, ther-mal etching can only create gratings with a pitch larger than a few microns. If the temperature rise profile seen in Figure 3.9b on a substrate with a base temperature of 125°C that was etched in 0.1 mTorr of CI2 the etch rate at the temperature rise peaks that is 0.20 nm-min - 1 greater than the dark etch rate. A 30 min etch would have a depth of 10 nm, which is observable with instruments such as A F M . For smaller pitch diffraction gratings in the sub-micron range the periodic part of the temperature rise will not be sufficiently large enough to etch a noticeable grating pattern in a reasonable time. Gratings with a smaller pitch will be washed out with thermal conduction. As shown in Figure 3.9 the more closely spaced (5 /mi period) interference pattern has it's temperature rise pattern effectively washed out by thermal diffusion. A n example of laser induced heating in the CI2 etching of GaAs is shown in Figure 3.12. For this sample a GaAs /AlAs /GaAs structure was etched with CI2 gas and illuminated with a Ne-He laser which had a position on the GaAs substrate shown in Figure 3.13. Note that no U V light was on the sample for this run. The He-Ne laser beam had an output wavelength of 632.8 nm, a spot diameter of 2 mm, and a power of 5 mW. The sample Figure 3.12: An optical image of a C l 2 etched substrate. This sample was illuminated with a single beam from a Ffe-Ne Laser operating at 632.8 nm. Airy rings were caused by the laser beam passing through an aperture. Differential etching throughout the sample resulted in the observed optical pattern. was processed with methods described in Section 2.3. The CI2 etch was conducted at a substrate temperature of 135°C, and a pressure of 0.1 mTorr. Unlike most other runs the CI2 etch was stopped soon after the uppermost GaAs layer was removed at the laser spot. At the center of the laser spot the 300 nm GaAs layer and an estimated 30 nm of a 100 nm AlAs layer has been removed. Over the rest of the sample most of the uppermost GaAs layer has been removed. For Figure 3.12 the illumination on the substrate caused differential heat-ing on the substrate's surface. This difference in temperature caused a dif-ference in the etch rate over the surface and thus different surface heights. Interference oscillations due to the GaAs /AlAs /GaAs or A lAs /GaAs struc-Figure 3.13: The positions of various illuminated areas on the GaAs sub-strate used for the experiments conducted in this work. The M L R method of etch rate measurement typically used these po-sitions of the mercury arc lamp beam and He-Ne laser beams. ture gives the surface a coloration which depends on the material composition and thickness of the layers. Oxidation of the exposed AlAs where the laser illuminated the surface gives that area a dark blue tint. The pattern displayed in Figure 3.12 could only have been caused by a laser induced thermal heating during the CI2 etch. A photochemical etch would etch evenly where illuminated, ie the laser spot and the Airy rings. Each Airy ring should be etched evenly throughout the ring which is not the case. The pattern of Figure 3.12 does not match the illuminated area of a single He-Ne laser beam. This nonuniform etch of the Airy rings is an indication that a thermal CI2 etch is the cause of the pattern shown in Figure 3.12. Figure 3.14: Optical micrograph of a GaAs /AlAs /GaAs sample which was photochemically etched with CI2 gas. U V light from a mercury arc lamp was used for the photochemical etch. The etch was stopped after the dark etch had removed the uppermost GaAs layer thus exposing the AlAs layer. The photochemical etch in the U V light spot has removed the uppermost GaAs and AlAs layers exposing the GaAs substrate. A n example of a GaAs /AlAs /GaAs sample that is photochemically etched with CI2 gas is shown in Figure 3.14. The substrate was processed with the methods described in Section 2.3. During the etch the substrate was held at 160.8°C, and the pressure of the CI2 gas was 0.1 mTorr. A mercury arc lamp produced a light beam with 31.4 mW of total power and a beam radius of 1 cm. The beam was incident on the sample at 65° from the surface normal. In the area illuminated with the U V light beam a 240 nm GaAs layer, a 50 nm AlAs layer and an estimated 100 nm of the GaAs substrate were removed. The exposed GaAs in the U V illuminated area gives this area a shiny metallic colour. In the dark area of this substrate a 240 nm GaAs layer and an estimated 40 nm of the AlAs layer were removed. The exposed AlAs in the dark area gives it a dark blue tint. The dark area in Figure 3.14 has an uniform colour throughout the surface, which indicates that the dark etch was uniform. A uniform dark etch indicates that light induced heating was minor. 3.3 Etch Rates The temperature dependence of a light induced photothermal C l 2 etch will yield a distinct behavior which is different from a photochemical effect. The temperature dependence of the dark CI2 etch rate is described in Equa-tion 1.1. For a photothermal reaction the measured light induced etch rate will have the following form: R = APCi[e~*6(*+AT) _ e - 7 ^ ) ] (3.21) where the first term gives the rate of the reaction in the presence of light and the second term corrects for the dark reaction. A T is the temperature rise due to the illumination. From Equation ?? A T is directly proportional to the intensity of the illumination. The light assisted etch rates which are shown in Figure 4.1 have a weak dependence on temperature. The temperature dependence of the etch rate obtained by the SGR and M L R methods can be fitted to Equation 3.21. To obtain the correct etch rate from a thermal effect alone the mercury arc lamp would need to heat the substrate by 28°C. The estimates for the experimental conditions used in this work predict a temperature rise no greater than 10°C. The etch rate has a linear dependence on intensity which does not match Equation 3.21. However, in the case of low level heating (T » A T ) , equa-tion 3.21 predicts an etch rate of: Equation 3.22 does not fit the photochemical etch rate shown in Figure 2.9. With the measured parameters A = 2.4 x 107 nm-min _ 1 Tor r _ 1 , and E a = 0.209 nm, and the experiment values Pci2 = 0.1 mTorr, T = 165°C, w = 0.5 cm and I0 = 33.1 m W - c m - 2 Equation 3.22 predicts that: However a slope of 0.039 cm 2nm-mW 1 min 1 was measured from intensity dependence of the etch rate shown in Figure 2.9. R = A T (3.22) R = (0.085cra2nm • mW~lmin~l)I0 (3.23) Chapter 4 Photochemical Etching The temperature dependence of the photochemical etch rate shown in Fig-ure 4.1 has a weak dependence on the substrate's temperature. The light assisted etch rate does generally increase with increasing temperature. The intensity dependence of the etch rate shown in Figure 2.9 has a linear de-pendence on the incident intensity. This chapter discusses the origin of the temperature and intensity dependence and the effects of illumination on the surface morphology. 4.1 Temperature Dependence of Etch Rates The temperature dependence of the etch rate can be explained with the chem-istry of the thermally activated C^-GaAs reaction. Near 200°C the thermal reaction undergoes a transition in the nature of it's chemistry. Initially the thermal reaction is limited by GaCl3 thermal desorption however after 200°C the reaction is limited by CI2 availability. In this work it is believed that the photochemical etch is caused by the light assisted desorption of GaCl3. c 'E 1 1 & rr o ' LU 0.1 o o o x O x A l A s Method o S i O Grat ing Method 2 5 3 1/T(1000/K) O 3.5 Figure 4.1: Temperature dependence of the photochemical etch rate mea-sured with the surface grating reference method and the multi-ple layer reflection method. The arc lamp had a total power of 31.4 mW for these etch rates. There are two photochemical processes in which light can desorb GaCl3 molecules from the GaAs surface. The first mechanism involves the absorp-tion of a photon by a GaCl3 molecule on the surface. Adsorbed GaCl3 can be directly photodesorbed from the surface or the GaCl3 molecule can be photoexcited to an energetic state which has a higher thermal desorption rate than the GaCl3 ground state. However, a direct photon absorption by the surface layer is unlikely, as only a small fraction of the incident light will be absorbed in the thin GaCl3 monolayer. The second and more probable process is for the GaCl3 molecule to absorb a photoexcited electron or hole from the semiconductor. As described in Section 1.1.1 electrons can be ex-cited to the conduction band through photon absorption. With illumination the desorption of GaCl3 is increased due to the addition of photoexcited elec-trons in the semiconductor. In this interpretation illumination reduces the importance of the thermal desorption of GaCl3 in the C^-GaAs reaction. A photochemical Cl2-GaAs reaction that is a result of photoexcitation in the semiconductor would depend on density of electron hole pairs. The variation in electron hole pair density was not accounted for in the etch rate measurements, which may account for the variation in the etch rate shown in Figure 4.1. Unfortunately the temperature dependence cannot be clearly determined. 4.2 Surface Morphologies q The following procedure was used to measure the surface morphology of a CI2 etched surfaces. First a (100) GaAs substrate was outgassed at 200°C then the native oxides were thermally desorbed at 600°C. With a M B E sys-tem a GaAs cap was then grown at 580°C such that the GaAs cap would be much greater than the depth etched during the CI2 etch. In a U H V etch chamber the GaAs substrate was etched for 10 or 30 min at a substrate temperature between 130 and 260°C with a CI2 gas pressure of 0.1 mTorr. Approximately 30% of the surface was illuminated with U V light from a mer-cury arc lamp. The light beam had a total power of 31.4 mW and a beam diameter of 10 mm. The illuminated and dark areas of the GaAs surface were then measured with an A F M in tapping mode. The scan areas ranged from 2x2 / im 2 to 20x20 / im 2 . Figure 4.2: Two 2x2 / im 2 A F M images of C l 2 etched GaAs surfaces. A GaAs substrate was etched for 30 minutes The right image is a thermally etched surface; the left image a photochemically etched surface. The RMS roughness of the illuminated and dark surfaces are 1.05 nm and 1.25 nm, respectively. The surface morphologies of the CI2 etched GaAs surface were not effected by illumination. Little to no effect was observed even under conditions when the photochemical etch rate was comparable to the dark etch rate. An ex-ample of the similarities between the thermally etched and photochemically etched surfaces is shown in Figure 4.2. The only difference between the two surfaces in Figure 4.2 is that the illuminated surface is slightly smoother with an RMS roughness of 1.05 nm compared to the dark surface which has an RMS roughness of 1.25 nm. This is not a significant difference. The sample shown in Figure 4.2 had the largest deviation between dark and i l -luminated areas in the temperature ranges examined. The similarity of the photochemical and dark surface morphologies is not surprising as they are both potentially produced by the same chemical mechanisms. Chapter 5 Conclusions In this work the photochemical reaction of CI2 gas with crystalline GaAs has been investigated. The photochemical etch rate was found to display a weaker temperature dependence than the dark etch rate. The addition of an S i 0 2 surface grating to the Cl2-GaAs system increased the etch rate, nearly doubling the dark etch rate. The photochemical etch rate increased linearly with light intensity. The temperature rise caused by light induced heating and the correspond-ing increase in dark etch rate were investigated. An estimate of the light-induced temperature rise was calculated by solving the heat equation. With this estimate of the temperature rise and the measured dark etch rate the photochemical etch rate could be distinguished from thermal effects. The photochemical etch rate observed could not have been caused by thermal effects. It is proposed that the photochemical etch rate is due to the GaCl3 pho-todesorption. The dark reaction rate may be enhanced by photon induced electron or hole transfer from the GaAs substrate to the physisorbed GaCl3 molecules, resulting an overall increase in the etch rate. Calculations of the temperature rise due to illumination with interference patterns demonstrate that the surface temperature of the GaAs substrate can varies periodically with position. It is possible to etch periodic patterns into a GaAs surface with photothermal etching alone. However, the amplitude of the temperature oscillations on the surface decrease as the pitch of the interference patterns decrease. It is not efficient to etch nanometer patterns with a photothermal reaction driven by an interference pattern. However, it should be possible to etch nanometer scale patterns with the photochemical effect. Chapter 6 Future Work Mpre work is needed to truly understand the subject of Cl2-GaAs photo-chemical etch. A study of the etch rate's dependence on wavelength would be one way to increase the understanding of this subject. A mercury arc lamp is an ideal tool for an investigation of the wavelength dependence. A mercury arc lamp has a continuous spectrum from the infrared to the U V re-gion with many large peaks in intensity within the U V region. With various cutoff filters one could isolate specific wavelengths and estimate the effect that wavelength has on the Cl2-GaAs photochemical etch. A system to create periodic patterns in the surface on GaAs within a high vacuum chamber is needed. Such a system is described in section 2.1. The characterization of the C^-GaAs etch is incomplete; ideal conditions needed to create nanometer scale patterns are not well known. CI2 may not be the ideal gas for the photochemical GaAs etch. The rate of the Cl2-GaAs dark etch is often greater that the photochemical etch rate for the powers of the mercury arc lamp used. A strong thermal etch increases the difficulty of etching patterns with high intensity illumination. As discussed in Chapter 3 thermal diffusion washes out small scale patterns in a light induced temperature rise. To etch nanometer scale pattern a direct photochemical reaction is ideal. Bromine is a potential candidate for pattern etching gas. Dark etching of GaAs with bromine does not occur until 400°C according to such works as Patrin [24], Brake [25], qqand Cha [26]. Since bromine has similar chemistry to that of chlorine a similar photochemical reaction with GaAs might occur. Appendix Mercury Arc Lamp Intensity Figure A . l : A schematic of the optical setup located after the mercury arc lamp. The iris aperture was typically set to have a 3 mm diam-eter opening. The light beam from the arc lamp is focused in a spot with a diameter of 10 mm at the GaAs substrate. The intensity of the mercury arc lamp's illumination is not easily deter-mined. The spectrum shown in Figure A.2 is difficult to measure accurately since it ranges from the U V to the infrared. As shown in Figure A . l , the optics used to focus the arc lamp's beam are not trivial. This appendix will describe the method used to estimate the power of the arc lamp's beam Figure A.2: The output spectrum of a mercury arc lamp at 200 W. This figure was taken from The Book of Photon Tools [6]. The intensity was measured with a Newport 818-UV Si photodiode which was filtered with a Newport RG850 cutoff filter. The Si photodiode in combi-nation with the RG850 cutoff filter measures the intensity in the wavelength range of 850-1050 nm. There is little change in the spectrum's intensity with respect to wavelength between 850 and 1050 nm which makes this a conve-nient region to measure. The Si photodiode was placed 440 mm away from the last lens shown in Figure A . l . This is the position that the substrate would be located during an etch experiment. The transmission through the cut off filter is 90% and the Si photodiode has a reponsivity of 0.55 A - W - 1 . The arc lamp's spectrum in Figure A.2 was adjusted so that it matched the measurement made by the Si photodiode and cutoff filter. The mercury arc lamp spectrum was integrated to obtain the total intensity of the beam over all wavelengths until the absorption edge of GaAs. The reflectance of the GaAs surface was also accounted for at this point. The reflectances were taken from Palik [27]. The total power of light incident to the GaAs surface from a Gaussian shaped beam with a diameter of 1 cm and an measured intensity of 40 W - c m - 2 in the center of the beam is then 31.4 mW. Bibliography [1] Chaochin Su, Hui qui Hou, Gang Ho Lee, Zi-Guo Dai, Weiang Luo, Matthew F Vernon, and Brian E. Bent. " Identification of the Volatile Reaction Products of the C^+GaAs Etching Reaction". Journal of Vacuum Sci Technology B, 11(4):1222-1242, 1993. [2] Jack G. Calvert, and James N . Pitts Jr. "Photochemistry". John Wiley and Sons Inc, 1966. [3] Robert James Browne. "Absolute Emission Intensity Studies on the Halogen Afterglows and Excited Molecular Oxygen". PhD thesis, Uni-versity of British Columbia, June 1969. Department of Chemisty. [4] K . Rajeshwar, L. M . Peter, A . Fujishima, D. Meissner and M . Tomkiewicz (Eds.). "Photoelectrochemistry". The Electrochemical So-ciety, Pennington, NJ , 1997. [5] Sadao Adachi. "Physical Properties of III-V Semiconductors Com-pounds". John Wiley and Sons, Inc, 1992. [6] "The Book of Photon Tools". 150 Long Beach Blvd, Stratford, Con-necticut, Unitied States of America, 06615-0872. Sales Manual. [7] N . Furuhata, H. Miyamoto, A . Okamoto, and K . Ohata. "CI2 Chemical Dry Etching of GaAs Under High Vacuum Conditions-Crystallographic Etching and its Mechanism". Journal of Electronic Materials, 19(2):201-208, 1990. [8] P Tejedor and F. Briones. "On the Mechanism and Surface Morphol-ogy of Gallium Arsenide Laser-Assisted Etching by Chlorine at 193nm". Journal of Chemical Physics, 101(3):2600-2605, 1994. [9] F. Foulon, Mino Green, F. N . Goodall, and S. De Unamuno. "Laser-Projection-Patterned Etching of GaAs in a Chlorine Atomsphere". Jour-nal of Applied Physics, 71(6):2898-2907, 1992. [10] Shinichiro Takatani, Seiji Yamamoto, Hiroyuki Takezawa, and Kozo Mochiji. "Excimer Laser Assisted Etching of AlGaAs and GaAs". Jour-nal of Vacuum Science and Technology B, 13(6):2340-2343, Nov/Dec 1995. [11] P. A . Maki and D. J. Ehrlich. "Laser Bilayer Etching of GaAs Surfaces". Applied Physics Letters, 55(2):91-96, 1989. [12] G. Haase, V . Liberman, and R. M . Osgood Jr. "Ultraviolet Laser-Induced Interaction of CI2 with GaAs(llO)". Journal of Vacuum Science and Technology B, 10(1):206-215, Jan/Feb 1992. [13] M . C. Shih, M . B. Freiler, G. Haase, R. Scarmozzino, and R. M . Osgood Jr. "Condensed Chlorine Etching of GaAs Induced by Excimer Laser Radiation". Applied Physics Letters, 61(7):828-830, August 1992. [14] C. I. H . Ashby. "Photochemical Dry Etching of GaAs". Applied Physics Letters, 45(8):892-894, October 1984. [15] M . Ishii, T. Meguro, T. Sugano, K . Gamo, Y . Aoyagi. "Surface Reaction Control in Digital Etching of GaAs by Using a Tunable U V Laser Sys-tem: Reaction Control Mechanism in Layer-by-Layer Etching". Applied Surface Science, 86:554-558, 1995. [16] Peter Brewer, Scott Halle, and R . M . Osgood Jr. "Photon-Assisted dry Etching of GaAs". Applied Physics Letters, 45(4):475-477, August 1984. [17] P.D. Brewer, D. McClure, and R . M . Osgood Jr. "Dry, Laser-Assisted rapid HBr etching of GaAs". Applied Physics Letters, 47(3):310-312, August 1985. [18] D.J Ehrlich, R . M . Osgood Jr., and T. F. Deutsch. "Laser-Induced M i -croscopic Etching of GaAs and InP". Applied Physics Letters, 36(8) :698-700, Apri l 1980. [19] Mikio Takai, Jun Tokuda, Hiroyuki Nakai, Kenji Gamo, and Susumu Namba. " Laser Induced Local Etching of Gallium Arsenide in Gas At-mosphere". Applied Physics Letters, 45(4):475-477, August 1984. [20] W. C. Simpson, W. M . Tong, C. B. Weare, D.K. Shuh, and J.A. Yarmoff. "The Temperature Dependence of the Cl2/GaAs(110) Surface Product Distribution". Journal of Chemical Physics, 104(1):320-325, 1996. [21] S. Ritchie, Shane R. Johnson, Christien Lavoie, J. A . Macken-zie, Thomas Tiedje, and R. Streate. "Semiconductor Substrate Cleaning and Surface Morphology in Molecular Beam Epitaxy". Surfure Science, 374:418-426, 1997. [22] P. J . Timans. "The Experimental Determination of the Temperature Dependence of the Total Emissivity of GaAs Using a New Temperature Measurement Technique". Journal of Applied Physics, 72(2):660-670, July 1992. [23] M . Lax. "Temperature Rise Induced by a Laser Beam". Journal of Applied Physics, 48(9):3919-4003, September 1977. [24] J. C. Patrin, Y . Z. L i , M . Chander, and J. H. Weaver. "Atomic Layer Etching of GaAs(llO) with B r 2 Studied by Scanning Tunneling M i -croscopy". Applied Physics Letters, 62(11):1277-1279, March 1993. [25] J . Brake, C. Y . Cha, B . Y . Han, D.W. Owens, and J. H. Weaver. "Coverage-Dependent Etching Pathways for Br-GaAs(l lO)". Journal of Vaccuum Science Technology B, 15(3):670-674, May/June 1997. [26] C. Y . Cha and J. H. Weaver. "Layer-by-layer Removal of GaAs(llO) by Bromine". Journal of Vaccuum Science Technology B, 14(6):3559-3562, Nov/Dec 1996. [27] Edward D. Palik, editor. "Handbook of Optical Constants of Solids", pages 429-444. Acedemic Press Inc, 1985. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0091083/manifest

Comment

Related Items