ga2o3 films for electronic and optoelectronic applications · single-crystal gd,ga,olz source for...

Ga203 films for electronic and optoelectronic ap M. Passlack,a) E. F. Schubert, W. S. Hobson, M. Hong, N. Moriya, S. N. G. Chu, K. Konstadinidis, J. P. Mannaerts, M. L. Schnoes, and G. J. Zydzik AT&T Bell Laboratories, 600 Mountain Avenue, Murray Hill, New Jersey 07974

(Received 9 August 1994; accepted for publication 4 October 1994)

Properties of GazOs thin films deposited by electron-beam evaporation from a high-purity single-crystal Gd,Ga,Otz source are reported. As-deposited Ga,Os films are amorphous, stoichiometric, and homogeneous. Excellent uniformity in thickness and refractive index was obtained over a 2 in. wafer. The films maintain their integrity during annealing up to 800 and 1200 “C on GaAs and Si substrates, respectively. Optical properties including refractive index (n = 1.84- 1.8 8 at 980 nm wavelength) and band gap (4.4 eV) are close or identical, respectively, to GazO, bulk properties. Reflectivities as low as 10e5 for Ga,O,/GaAs structures and a small absorption coefficient (-100 cm-’ at 980 nm) were measured. Dielectric properties include a static dielectric constant between 9.9 and 10.2, which is identical to bulk Ga,O,, and electric breakdown fields up to 3.6 MV/cm. The GazOJGaAs interface demonstrated a significantly higher photoluminescence intensity and thus a lower surface recombination velocity as compared to Al,O,/GaAs structures. Q I995 American Institute of Physics.

I. INTRODUCTION

Dielectric films find a wide range of applications from passivation of surface states in various types of conventional and low-dimensional devices to metal-insulator- semiconductor field-effect devices. The lack of stable dielectric films providing a low interface state density is still a serious drawback of III-V semiconductors.‘-” Different dielectric materials including Si3N4, SiO, , Al,O,, and GazO, have been used in combination with dry, liquid, and photo- chemical semiconductor surface treatments.‘*-*5 Recently, Aydil et a1.‘6T’7 achieved passivation of surface states during a NH, or H, plasma treatment at room temperature by re- moval of excess As and As,O, and subsequent formation of a Ga,O, film (a few monolayers thick) on the GaAs surface. The observed increase in photoluminescence intensity indicates a decrease of surface recombination velocity, which is related to the midgap interface state density, by several or- ders of magnitude.” Furthermore, GaZO, films are chemi- cally stable on GaAs as demonstrated by thermochemical phase diagramsi This paper reports on properties of dielectric GazOs films deposited by electron-beam evaporation from a high-purity single-crystal Gd,Ga& source. Thin- film properties including stoichiometry, microstructure, phase, homogeneity, uniformity, and temperature integrity as well as optical and electrical characteristics have been investigated and a preliminary photoluminescence characterization of a Ga,Os/air exposed (100)GaAs interface is pre- sented. Possible semiconductor processing schemes using our deposition technique for Ga,O, films are outlined.

II. FILM FABRICATION

Electron-beam deposition imparts only a small amount of energy into the surface” and is compatible with existing semiconductor processing schemes. Other Ga,Os thin-film fabrication techniques using oxidation of evaporated Ga in

‘)E-mail: [email protected]

an O2 plasma may cause severe surface damage prior to deposition and did not produce dielectric GazO, films-l4 Splattering of GazO, , which occurs due to rapid sublimation during electron-beam heating, can be eliminated by using a single-crystal Gd,Ga,Olz source for electron-beam evaporation.21 The Gd,Ga,012 source material represents a chemical combination of the relatively covalent oxide Ga,O,, which volatilizes near 2000 K, and the pretransition oxide Gd,O,, which has a boiling point of 4000 K. The compound GdjGa,O,, decrepitates during heating by electron beam, slowly releasing GazO,. The deposition rate is kept low at about 0.5 &s in order to maintain a high purity of the deposited Ga,O, film. The background pressure in the vacuum chamber was 1-2X10m6 Ton-. We investigated both cases of (i) no excess oxygen and (ii) introduction of additional oxygen into the evaporation chamber. The maximum partial O2 pressure was 2X 10s4 Torr when additional 0, was introduced. Using this method, we have deposited high- purity, uniform, homogeneous, and stoichiometric dielectric Ga20s amorphous films with thicknesses between 40 and 4000 A on GaAs, Si, and fused silica substrates.

III. FILM CHARACTERIZATION

The GazOs films were analyzed by ellipsometry and Au- ger depth profiling for uniformity and homogeneity, respectively. Furthermore, the Ga,O, films were characterized by Rutherford backscattering spectroscopy (RBS) for area1 densities and atomic composition including stoichiometry as well as impurities. X-ray photoelectron spectroscopy (XPS) of GaZOPs films was performed in order to determine the oxidation state of the Ga,O, films. The microstructure and the phase of Ga20s films was analyzed by transmission electron microscopy (TEM). The etching behavior and the temperature integrity of Ga,Os films was investigated up to annealing temperatures of 1200 “C. Spectroscopic ellipsometry measurements were done in order to determine the refractive index n and the absorption coefficient a. Optical transmission and reflection measurements provided further optical

686 J. Appl. Phys. 77 (2), 15 January 1995 0021-8979/95/77(2)/686kV$6.00 Q 1995 American Institute of Physics

Downloaded 18 Feb 2002 to 132.68.1.29. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/japo/japcr.jsp

DEPTH (nm) 8 25 50 75 100 125

B I I

‘E 7 3 f; 6 a ;5 55 4 z I- 3 z a 2 b z 0

0 3 6 9 12 15 18 SPUTTER TIME (min)

FlG- 1. Auger depth profile of a 955 %, thick Ga,?O, l i lm deposited at a substrate temperature of 125 “C with an 0s partial pressure of 2X 10M4 Torr in the deposition chamber. The analysis was done by using 4 keV Ar ions and the sensitivity factors for Ga and 0 were calibrated against pressed GasOs powder. The nominal sputter rate is 65 &in.

FIG. 2. Refractive index and thickness uniformity of a Ga,O, fi lm over a 2 in. wafer. This fi lm was deposited at T,=40 “C without introduction of additional oxygen into the deposition chamber. Thickness and refractive index (measured at X=830 nm) are (85424) 8, and 1.916CO.001, giving a relative standard deviation of 0.5% and 0.05%, respectively.

properties including band gap and reflectivity, respectively. Dielectric properties of Ga.@s films were evaluated on metal/Ga,Os/semiconductor structures. Finally, the Ga,O,/air exposed (100)GaAs interface was evaluated by photoluminescence measurements.

A. Homogeneity and uniformity

Figure 1 shows the Auger depth profile of a 955 A thick Ga?Os film deposited at a substrate temperature of 125 “C with an O2 partial pressure of 2X10m4 Torr in the deposition chamber. This result demonstrates the excellent depth homogeneity of the Ga,O, films. The area1 uniformity of a Ga,O, film deposited at T,=40 "C without introduction of additional oxygen is illustrated in Fig. 2. The standard deviations of the thickness and of the refractive index over a 2 in. wafer are 4 A (0.5%) and 0.001 (O.OS%), respectively. Similar results were obtained when oxygen was introduced into the chamber during deposition.

B. Atomic composition and density

The GazOs films were characterized by Rutherford backscattering spectroscopy for area1 densities and atomic composition including stoichiometry. The measurements were performed using He+ ions at 1.8 MeV in random direction and at a 175” detection angle. Experimental results were compared to theoretical simulationsZ2 in order to obtain the area1 densities of the oxide layers and the respective O/Ga atomic ratio. Measured RBS spectra and the corresponding simulations are shown in Fig. 3. Experimental results indicate homogeneous &d stoichiometric gallium oxide films with 40% Ga and 60% 0 for moderate substrate deposition temperatures and introduction of additional oxygen into the deposition chamber.’ Nonstoichiomktric films were obtained for elevated substrate deposition temperatures (e.g., 44% Ga and 56% 0 for Ts=350 “C). The oxidation state of these

films is further investigated by XPS in the next section. The film thicknesses separately measured by ellipsometry were used to compare the volume density of the deposited oxide films to the bulk GazOs density. For introduction of oxygen into the deposition chamber (p,,=2X10W4 Torr), the calculated film densities for lower T, were between 60% and 68% of the bulk density using the bulk density of o+Ga& (6.44 g/cm3, Ref. 23). For higher substrate temperatures T, between 200‘and 350 “C, the film density increases to approximately 80% of the bulk density. If no excess oxygen is introduced, film densities between 70% and 74% of the bulk density were obtained for T,=40 "C.

200 300 400 500 Channel

FIG. 3. Rutherford backscattering spectra measured on Ga,Os fi lms deposited with introduction of additional oxygen into the deposition chamber. An excellent agreement between simulated and experimental curves indicates homogeneous and stoichiometric galIium oxide fi lms with 40% Ga and 60% 0 for substrate temperatures of 40 and 125 “C.

J. ‘Appl. Phys., Vol. 77, No. 2, 15 January 1995 Passlack et al. 687


Ino A” 100 200 300 4ao 500

Channel

FIG. 4. Rutherford backscattering spectrum measured on a Ga,03 film deposited at Ts=40 ‘C with introduction of additional oxygen into the deposition chamber. The relative concentration of Gd, which is considered to be the main impurity in our Ga203 films, was estimated to be, in general, lower than 0.1 at. %.

We consider Gd to be the dominant impurity in our Ga203 films and the existence of Gd was detectable. We estimate the relative concentration of this element to be, in general, lower than 0.1 at. % (Fig. 4). The two measured spectra shown in Fig. 5 (solid and dotted lines) were taken at different normal-to-beam angles (tilt angle 8) ih order to stimulate different geometry -configurations. Only small dif- ferences were detected in the relative energy positions between the two different tilt angles. The dashed line in Fig. 5 simulates the case of homogeneous. distribution of Gd throughout the entire oxide film. The actual elemental distribution, however, is sharp and its peak position does not shift

40 “C, 2x1 Od Torr 0,

- Experiment, 6 = 10’ -.--.-. Experiment, 6 = 57’ ---- Simulation, += lo’,

uniform distribution

400 450 Channel

500

FIG. 5. Rutherford backscattering spectra were taken at different normal-to- beam angles 4= 10” (solid line) and 57” (dotted line) in order to stimulate different geometry configurations. The dashed line simulates the case of

homogeneous distribution of Gd throughout the entire oxide film.

G@PJR [Gal (111,7.4av)

420.4eV =WWh

1128 1124 11'20 1118 1112 440 430 420 410

(4 Binding Energy(eV) 03

FIG. 6. XPS depth profile of a Ga,Os film deposited on GaAs at a substrate temperature of 125 “C and with no introduction of additional oxygen into the chamber. The Ga2~~,~ core levels for Ga,O, and elemental Ga are shown in the left figure. The Auger peak position (right plot) refers to the Ga L,M,,M,, line. The low binding-energy shoulder of the Ga 2~~12 peak in the Gaz07 film indicates the presence of elemental gallium in this tilm. The estimated mole fraction of elemental Ga is about 5%.

with the tilt angle IY, reflecting the existence of a narrow region, probably the surface region, where the Gd tends to segregate during film deposition.

C. Oxidation state

X-ray photoelectron spectroscopy of Ga201 films was performed in order to determine the oxidation state of the Ga,O, films. A Perkin-Elmer 5600 series XPS spectrometer equipped with a monochromatic Al Ka x-ray source was used. High-resolution spectra were acquired using an x-ray power of 350 W. The aperture was set at 800 ,UII and the pass energy at 29 eV, giving a 0.72 eV full width at half maximum for the Ag 3d,,, line. Depth profiling was done in situ by Ar sputtering using an ion gun at 4 keV. Ain area of 2X2 mm2 was sputtered at a rate of -100 Aimin. The sur- facecomposition at each step was calculated from the integrated areas of the core-level peaks using known sensitivity factors. The elemental gallium to Ga,03 ratio was calculated by curve fitting the Ga2p 312 peak using predominantly Gaussian component peaks. Since the difference between the Ga 2p3,2 and the Ga Auger lines is used for chemical state identification, the reported peak positions were not compen- sated for surface charging. The Auger peak position refers to the Ga L3M45M35 line.

Figure 6 shows the XPS depth profile of a Ga,O, film on GaAs deposited at a substrate temperature of 125 “C and with no introduction of additional oxygen into the chamber. The positions of the Ga 2p3j2 and the Ga L3M45M45 peaks in the GaAs substrate are 1117.4 and 420.4 eV, respectively, giving a difference of 697.0 eV. These numbers correspond well with the data found in Ref. 24 for elemental Ga, which are 1116.7 and 419.0 eV for the Ga 2p3,, and the Ga L3M45M45 peaks, respectively, yielding to an energy difference of 697.7 eV. The Ga 2p3,2 spectrum of the oxide film can be decomposed into two components. The first component at lower binding energy corresponds to elemental Ga and the second shifted by 2 eV towards higher binding en-

688 J. Appl. Phys., Vol. 77, No. 2, 15 January 1995 Passlack et al.


G&P, P&l (1119&3vJ

\ I 425.7 eV

h 420.4 ev

1128 1124 1120 1116 1112 440 430 420 410

(4 Binding Energy (ev) (b)

FIG. 7. IipS depth profile of a Ga203 film on Si deposited with introduction of oxygen (P,,~=ZX 10W4 Torr) into the chamber at a substrate temperature of 40 “C. Left and right plots show the Ga 2p,,, core levels for Ga,O, and elemental Ga as well as the Ga L1M&445 Auger line, respectively. Like in all investigated stoichiometric gallium oxide films, elemental Ga was not detected.

ergy is associated with Ga,O, . 24 Suboxides such as GaO and GazO, which were reported to be unstable and may exist only far away from equilibrium,z*26 were not detected in these films. The estimated mole fraction of elemental Ga is.tibout 5%. As expected, elemental Ga was not detected in stoichiometric gallium oxide films. Figure 7 shows the XPS depth profile of a Ga,O, film on Si deposited with introduction of oxygen Cp,,-- - _ -2X 10d4 Torr) into the chamber at a substrate temperature of 40 “C. The energy difference of the Ga 2p3,2 11119.6 eV) and the Ga L3M45M45 (425.7 eV) peaks in the Ga,03 film is 693.9 eV and no low binding-energy shoulder in the Ga 2ps12 peak could be observed within the oxide film. There is, however, evidence of elemental Ga at the Ga203/Si interface with the Ga 2p3,2 peak at 1117.8 eV and the Ga L,M,,M,, peak at 420.4 eV, giving an energy difference of 697.4 eV.

FIG. 9. Cross-sectional TEM micrographs of Ga,O,lGaAs interfaces for (a) as deposited at 40 “C and (b) after annealing at 800 “C for 30 s in N?.

D. Microstructure

Gaz03 films deposited and annealed under different conditions were investigated by transmission electron microscopy. The samples were prepared by chemical thinning using a grid-masking technique27 and inspected using a Philips 420 TEM operated at 120 kV. In general, the microstructure of the films was uniform and as-deposited films were amorphous. Figure 8 shows the plan-view TEM micrographs and the corresponding electron-diffraction patterns of Ga,O, films deposited on GaAs at substrate temperatures of (a) 40, (b) 125, and (c) 350 “C. The diffraction patterns indicate the amorphous nature of all the films within the entire range of deposition conditions. As described in Sec. III B, the film deposited at 350 “C became Ga rich with a corresponding increase in the volume density. This is also reflected by a much finer morphology revealed in the corresponding TEM micrograph. TEM cross-sectional micrographs of a Ga20, film (a) after deposition at T,=40 “C on a GaAs substrate and (b) after subsequent annealing at 800 “C for 30 s in N2 are shown in Fig. 9. The GaAs/Ga,O, interface and the

MG. 8. Plan-view TEM micrographs and the corresponding electron-diffraction patterns of Gaz03 films deposited on GaAs substrate at temperatures of (a), 40, (b) 125, and (c) 350 “C.

J. Appl. Phys., Vol. 77, No. 2, 15 January 1995 Passlack et a/. 689


GazO, surfaces were observed to be planar to the nanometer scale even after annealing at 800 “C. Since the integrity of GazO, on GaAs is preserved, these films can be used for high-temperature annealing of implants in GaAs. Annealing of Ga,O, films at temperatures above 700 “C leads to a dis- tinct change in the microstructure from an amorphous to a poiycrystalline phase. As expected, the grain size increases with annealing temperatures. Figure 10 shows the plan-view TEM micrographs and the corresponding electron-diffraction patterns for GaaOs films (a) on GaAs annealed at 800 “C and (b) on Si substrate annealed at 1200 “C in ‘0, for 30 s. The continuous diffraction rings from the fine grain film break into spotty patterns for the large gram film. The measured lattice spacings are plotted in Fig. 11(a) along with the pub- lished x-ray powder patterns for (b) a-Ga20s and (c) monoclinic Ga,O,. The majority of the lines match with the a-Ga,O,, while a few lines match the monoclinic Ga*Os, - . which indicates that the dominant crystalline phase is cr-GazOs.

E. Etching behavior and temperature integrity

The etching behavior of GazO, films was investigated both as a function of deposition conditions and annealing temperature. The annealing was done in N2 or 0, atmosphere for 30 s. Figure 12 shows the etch rate of Ga,Os films in a 10% HF solution at room temperature. The etch rate de- creases with increasing substrate temperature during deposition as well as with annealing temperature. Note that the etch rate of as-deposited Ga207 films at T,=350 "C is significantly lower than the etch rate of SiOa (200 nm/min) fabri- cated by plasma-enhanced chemical-vapor deposition at the same substrate temperature. In agreement with the refractive index, absorption, and RBS measurements, the decrease in etch rate with increasing substrate temperature during deposition and with rising annealing temperatures below 700 “C! can be attributed to improved film densities. A new quality is observed for Gaz03 films annealed at temperatures equal or higher than 700 “C which may be related to the change from an amorphous to a polycrystalline phase as revealed by THM measurements. These films do not etch even in a concen- trated HF solution and the water sensitivity which is ob-

FIG. 10. Plan-view TEM micrographs and the corresponding electron- diffraction patterns of polycrystalline Gaps films obtained after annealing at (a) 800 “C for 30 s in N, on GaAs and (b) at 1200 “C for 30 s in 0, on Si.

100 L I

50 1 0 0 (4

0 80 I 2 h ‘&o * 0 : o 0, $ P 0 I I L -I 3 100

s .E 50

-2 0 3 100

50

0 ,a 20 24 28 3.2

Lattice Spacings (A)

FIG. 11. Comparison of the measured lattice spacings (a) in the crystallized GasOs film with the pubbshed x-ray powder patterns of (b) cr-GasO, and (c) monoclinic GasOs .

served for amorphous Ga,Os films disappeared. In general, the refractive index of amorphous Ga,Os films was altered within a range of approximately - 1.5% and + 1% after soaking in de-ionized water for 60 h at room temperature, whereas polycrystalline Ga,O, films were found to be stable.

Furthermore, the refractive index measured before and after annealing was used to evaluate the temperature integrity of Ga,Os films. Figure 13 shows the relative change in refractive index versus annealing temperature for a Ga,Os film deposited on a Si substrate at T,r= 125 “C and an oxygen partial pressure of 2X10e4 Torr in the deposition chamber. After a small, initial decrease in refractive index, the refractive index slightly increases with annealing temperature in an 0, or N2 atmosphere and the film maintains its integrity up to our maximum annealing temperature of 1200 ‘C. Anneal- ing in a reducing ambient (15% Hz, 85% NJ, however, leads to a drastic decrease in refractive index for annealing temperatures over 700 “C and eventually to decomposition of the

0 2 4 6 8 Annealing Temperature ( lo2 “C)

FIG. 12. Etch rate of GasOs films in a 10% HF solution at room temperature as a function of anneahng temperature with the substrate temperature during deposition as parameter. The annealing was done in Nz or Oa atmosphere for 30 s.

690 J. Appl. Phys., Vol. 77, No. 2, 15 January 1995 Passlack et a/.


I’ -*-- 15% H 29 85% N 2 (reduci&) Mm /;nm+\ h=980nm

: 1.8 23 ? g 1.7 2

1 $Gaz030nSi {,

Ud 2 4 6 8 10 12 14 Annealing Temperature (10” “C)

1.6 0 no excess O2 0 no excess 0

Substrate Temperature (‘-‘C)

FIG. 15. Refractive index n and absorption coefficient cr measured at 980 nm wavelength for GasOs films deposited at substrate temperatures between 40 and 350 “C. The absorption coefficient is very small (-100 cm-‘) and almost independent of Z’, , indicating the nonabsorbing nature of these films.

FIG. 13. Relative change in refractive index vs annealing temperature for a GazOs film deposited on a Si substrate at T,= 125 “C and an oxygen partial pressure of 2X lo-” Tom in the deposition chamber.

film. Ga20s films deposited on GaAs maintain their integrity up to annealing temperatures of 800 “C (Fig. 14). additional oxygen into the chamber may be attributed to

higher film densities. The measured absorption coefficient is very small (~100 cm-‘) and almost independent of T,, indicating the nonabsorbing nature of these films. Since the square root of the refractive index of most III-V semiconductors is between 1.76 and 1.90, Ga,O, films are we11 suited for optical antireflection coatings. As demonstrated in Fig. 16, reflectivities as low as 10M5 were measured on Ga,03/GaAs structures. The corresponding simulation as- sumes an abrupt transition of two media with refractive index nl = 1 for air and ns(X) for GaAs,28 respectively, sepa- rated by a Ga,O, film with a measured refractive index 44

F. Optical properties

Spectroscopic variable angle ellipsometry measurements were done between 400 and 17OO’mn in order to determine the refractive index n and the absorption coefficient a as a function of wavelength. Figure 15 shows y1 and Q measured at 980 nm wavelength for Ga,O, films deposited at substrate temperatures between 40 and 350 “C. The refractive index y1 is within a range of 1.841-1.885 at 980 nm wavelength for all substrate temperatures T, used during deposition. This refractive index is close to rz = 1.91 which is given in Ref. 23 for bulk Ga,O,. The slight increase in refractive index with substrate temperature for films deposited with introduction of

0 T,=4O”C q T,=35O”C

2xlO4TorrO2 J-J

Annealing time = 30 s in Nz (inert)

-10 Wavelength (nm) 2 4 6 8 10

Annealing Temperature (10’ “C) FIG. 16. Measured reflectivity vs wavelength of a GasO,/GaAs structure. The measured reflectivity minimum is <lo-’ at 885 nm wavelength, which is in excellent agreement with the simulation data. The film thickness is 1170 A.

FIG. 14. Relative change in refractive index vs annealing temperature for a GasO, fiIm deposited on a GaAs substrate at different deposition conditions.

J. Appl. Phys., Vol. 77, No. 2, 15 January 1995 Passlack et a/. 691


4

7 8

“0 3 “, 25 !d

f2 u 2 .i 1

9

0 2 3 4 5 6 7

Energy WI

FIG. 17. Absorption coefficient LY as a function of photon energy of a Ga,O, film deposited on a fused-silica disk. The band gap of this Ga,O, film was determined to be 4.4 eV, which is identical to the bulk Ga,O, band gap.

The optical transmission was measured within a wavelength range of 185-600 nm for Ga,& films deposited on fused silica disks using a Perkin-Elmer spectrophotometer. According to a definition for materials with residual absorption,” the band gap of Ga,O, films was determined to be 4.4 eV (Fig. 17). The same band gap was reported earlier for bulk Ga,OO . 28

G. Dielectric properties

Ti/Au dots of different diameters (50, 100, 200, and 500 pm) were evaporated on 40-4000 A thick Ga203 films deposited on n+-GaAs substrates. Figure 18 shows the dc breakdown field of 400-500 %, thick Ga,O, films measured by applying a voltage ramp of approximately 1 V/s to the metal-insulator-semiconductor structure at room Eempera- ture. A breakdown field of 3.6 MVIcm was measured for

4.0

3 g 3.5

a 2 r.L z $j 3.0

ii2 2.5

I ’ I * I -

-203 ti=400-500A

A = 7.85 x 10” cm2 . 0

A A

0 no excess 02 A 2~10.~ Torr O2

I , I . I .

0 100 200 300 400 Substrate Temperature (“C)

FIG. 18. Room-temperature dc breakdown field as a function of substrate temperature T, for 400-500 8, thick Gaz03 films.

I I I

Ga203 $=750A

‘0 2 4 6 8 10 Annealing Temperature (10” “C)

FIG. 19. Room-temperature dc breakdown field of a 750 %, thick Ga,G, film as a function of annealing temperature. The Ga,O, film was deposited at 40 “C without excess 0, in the evaporation chamber.

Ga,O, films deposited at 40 “C without excess O2 in the evaporation chamber. If additional oxygen was introduced into the evaporation chamber, the breakdown field is about 3.1 MV/cm within the entire range of substrate temperatures during deposition. The specific resistivity of these films is found to be of the order of 10’2-1013 a cm at room temperature. in an earlier work?’ we also reported a static dielectric constant of Ga203 films between 9.9 and 10.2, which is in close agreement with ei= 10.2 measured for single-crystal Gaz03 platelets.31 Furthermore, like for other dielectrics, the current transport mechanisms in metal/Ga203/semiconductor structures were found to be controlled by the insulator bulk properties.

In order to evaluate the stability of dielectric properties of Ga203 films on GaAs with temperature exposure during processing, TilAu dots were evaporated on annealed films and, subsequently, the dc breakdown field was measured. As indicated in Fig. 19, the dc breakdown field degrades only by 6% for typical processing temperatures up to 400 “C. The breakdown field, however, degrades rapidly for annealing temperatures above 550 “C.

H. Interface properties

The Ga?O?/air exposed (lOO)GaAs interface was inves- - tigated by a photoluminescence study using an argon-ion la- ser (X=5 14.5 nm) operated at an optical power density of 50 W/cm’. Dispersion and detection were done using a Spex 1877 triple monochromator and a Princeton Instruments liquid-nitrogen-cooled CCD camera, respectively. The inte- gration time was 120 s. A particular structure was prepared (see inset of Fig. 20) in which the photoluminescence intensity is highly sensitive to the GaAs surface recombination velocity. Subsequently, 400 A thick Ga203 and Al203 films were deposited on this structure by electron-beam evaporation. The photoluminescence intensity increased after deposition of a Ga203 film and, as indicated in Fig. 20, the pho-

692 J. Appl. Phys., Vol. 77, No. 2, 15 January 1995 Passlack et al.


Q . .4W3 $ __....._. c . . . . . . . . . . ..I . . . . . . . _.I..... . . . . --,.~ .)..__-_...

o- 860 865 870 875 880 ‘

Wavelength (nm)

FIG. 20. Photoluminescence intensity as a function of wavelength measured for samples with a 400 ?i thick Ga,O, (solid line) and Alz03 (dashed line) tilm deposited on an air exposed (100)GaAs surface. Excitation wavelength and optical power density are 514.5 nm and 50 W /cm’, respecti\iely. The inset shows the test structure consisting of a 3000 A undoped GaAs layer and a 1 m thick InGaP layer grown on an undoped GaAs subswte.

toluminescence intensity is significantly higher for an air exposed (100)GaAs surface having a deposited Ga20, film compared to the same structure with a deposited Al,O, film. Thus a lower surface recombination velocity is obtained for the Ga20, coated sample.

IV. CONCLUSIONS

Properties of dielectric Ga,O, films deposited by electron-beam evaporation from a high-purity single-crystal Gd3Ga,012 source were reported. Thin-film properties including stoichiometry, microstructure, phase, homogeneity, uniformity, and temperature integrity as well as optical and electrical characteristics make these films well suited for electronic and optoelectronic applications. Furthermore, our deposition technique is compatible with integrated vacuum processing schemes using surface plasma passivation techniques or molecular-beam epitaxy in order to produce dielectric/III-V semiconductor interfaces with low density of interface states.

ACKNOWLEDGMENTS

We would like to thank G. P. Schwartz, R. A. Gottscho, R. L. Opila, T. D. Harris, L. C. Feldman, R. Becker, H. S. Luftman, and R. J. Fischer for many useful discussions. Spe- cial thanks to U. K. Chakarabarti and R. L. Masaitis for spectroscopic elIipsometry and Auger spectroscopy measure-

ments, respectively. M. Passlack gratefully acknowledges support by the Deutsche Forschungsgemeinschaft. We are also grateful to N. K. Dutta, A. Y. Cho, and D. V. Lang for their support.

’ Reviews can be found in C. W . Wilmsen, Physics and Chemistry of III-V Compound Semiconductor Intet@ces (Plenum, New York, 1985) and W . F. Croydon and E. H. C. Parker, Dielectric Fi lms on Gall ium Arsenide (Gor- don and Breach, New York, 1981).

2W. E. Spicer, 2. Liliental-Weber, E. Weber, N. Newman, T. Kendelewicz, R. Cao, C. McCants, P. Mahowald, K. Miyano, and I. Lindau, J. Vat. Sci. Technol. B 6, 1245 (1988).

3L. S. Hung, G. H. Braunstein, and L. A. Bosworth, Appl. Phys. Lett. 60, 201 (1992).

‘P. Viktorovitch, M. Gendry, S. K. Krawczyk, F. Krafft, P. Abraham, A. Bekkaoui, and Y. Monteil, Appl. Phys. Lett. 58, 2387 (1991).

‘H. Shigekawa, T. Hashizume, H. Oigawa, K. Motai, Y. Mera, Y. Nannichi, and T. Sakurai, Appl. Phys. Lett. 59, 2986 (1991).

6J. A. Dagata, W. Tseng, J. Bennett, J. Schneir, and H. H. Harary, Appl. Phys. Lett. 59, 3228 (1991).

‘A. Paccagnela, A. Callegari, A. Camera, M. Gasser, E. Latta, M. Mu- rakami, and M. Norcott, J. Appl. Phys. 69, 2356 (1991).

‘2. Lu, M. T. Schmidt, D. Chen, R. M. Osgood, Jr., W . M. Holber, D. V. Podlesnik, and J. Forster, Appl. Phys. Lett. 58, 1143 (199 1).

‘S. Shikata, H. Okada, and H. Hayashi, J. Vat. Sci. Technol. B 9, 2479 (1991).

“0. Nakajima, K. Nagata, H. Ito, T. Ishibashi, and T. Sugeta, Jpn. J. Appl. Phys. 24, 596 (1985).

“F. Capasso and G. E Williams, J. Electrochem. Sot. 129, 821 (1982). “S. D. Offsey, J. M. Woodall, A. C. Warren, P. D. Kirchner, T. I. Chapell,

and G. D. Pettit, Appl. Phys. L&t. 48, 477 (1986). *lC. J. Sandroff, R. N. gottenburg, J.-C. Bischoff, and R. Bhat, Appl. Phys.

Lett. 51, 33 (1987). 14See for example, A. Callegari, P. D. Hoh, D. A. Buchanan, and D. Lacey,

Ap;l. Phys. Lett. 54, 332 (1989). 15J. S. Herman and E L. Terry, Appl. Phys. Lett. 60, 716 (1992). 16E. S. Aydil K. P. Giapis, R. A. Gottscho, V. M. Donnelly, and E. Yoon, J.

Vat. Sci. &chnol. B 11, 195 (1993). 17E. S. Aydil and R. A. Gottscho, Mater. Sci. Forum 148-149, 159 (1994). ‘*M. Ettenberg and C. J. Nuese, J. Appl. Phys. 46, 3500 (1975). “G. P. Schwartz, Thin Solid Fi lms 103, 3 (1983). 2oC. W . Wilmsen, Physics and Chemistry of lII-V Compound Semiconduc-

tor Inter&es (Plenum, New York, 1985), p. 185. 2’S. Singh, L. G. van Uitert, and G. J. Zydzik (unpublished). *‘L. Doolittle, Nucl. Instrum. Methods B 15, 227 (1986). ‘3CRC Handbook of Chemistry and Physics I989-1990 (CRC, Boca Raton,

1989). p. B-92 24 Handbook of X-Ray Photoelectron Spectroscopy (Perkin-Elmer Corpora-

tion, Eden Prairie, 1992). 25T. B. Massalski, Binary Alloy Phase Diagrams (ASM International, Metal

Parks, OH, 1990). Vol. 2, p. 1834. 26CRC Handbook of Chemistry and Physics 1971-1972 (CRC, Cleveland,

OH, 1971), p. D-45. 27S. N. G. Chu and T. T. Sheng, J. Electrochem. Sot. 131, 2663 (1984). %See, for example, S. Adachi, J. Appl. Phys. 58, Rl (1985). 2YE. Kauer and A. Rabenau, Z. Naturforsch. 13a, 531 (1958). “M. Passlack, N. E. J. Hunt, E. F. Schubert, G. J. Zydzik, M. Hong, J. P.

Mannaerts, R L. Opila, and R. J. Fischer, Appl. Phys. Lett. 64, 2715 (1994).

31 See, for example, B. Hoeneisen, C. A. Mead, and M.-A. Nicolet, Solid- State Electron. 14, 1059 (1971).

J. Appl. Phys., Vol. 77, No. 2, 15 January 1995 Passlack et a/. 693 Downloaded 18 Feb 2002 to 132.68.1.29. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/japo/japcr.jsp

ga2o3 films for electronic and optoelectronic applications · single-crystal gd,ga,olz source for...

Documents