The patterning characteristics of the indium tin oxide (ITO) thin films having different microstructures were investigated. Several etching solutions (HC1,. HBr, and ...
Regular Issue Paper
Journal of Electronic Materials, VoL 25, No. 12, 1996
Patterning of Transparent Conducting Oxide Thin Films by Wet Etching for a-Si:H TFT-LCDs JE-HSIUNG LAN and JERZY KANICKI Center for Display Technology and Manufacturing, University of Michigan, Ann Arbor, MI 48105-2551 ANTHONY CATALANO and JAMES KEANE National Renewal Energy Laboratory, Golden, CO WILLEM DEN BOER and TIEER GU Optical Imaging Systems Inc., Northville, MI 48084 The patterning characteristics of the indium tin oxide (ITO) thin films having different microstructures were investigated. Several etching solutions (HC1, HBr, and their mixtures with HNO 3)were used in this study. We have found that ITO films containing a larger volume fraction of the amorphous phase show higher etch rates than those containing a larger volume fraction of the crystalline phase. Also, the crystalline ITO films have shown a very good uniformity in patterning, and following the etching no ITO residue (unetched ITO) formation has been observed. In contrast, ITO residues were found after the etching of the films containing both amorphous and crystalline phases. We have also developed a process for the fabrication of the ITO with a tapered edge profile. The taper angle can be controlled by varying the ratio of H N Q to the HC1 in the etching solutions. Finally, ITO films have been found to be chemically unstable in a hydrogen containing plasma environment. On the contrary, aluminum doped zinc oxide (AZO) films, having an optical transmittance and electrical resistivity comparable to ITO films, are very stable in the same hydrogen containing plasma environment. In addition, a high etch rate, no etching residue formation, and a uniform etching have been found for the AZO films, which make them suitable for a-Si:H TFT-LCD applications. K e y w o r d s : Al-doped zinc oxide, etching, films, hydrogen containing plasma, indium tin oxide
INTRODUCTION Transparent conducting oxides (TCOs) are highly degenerate doped (N -- 1021cm-3) and large bandgap (Eg_> 3.0eV) semiconductors. A high optical transmittance and low electrical resistivity make them very attractive for optoelectronic device applications such as solar cells, 1 image sensors, 2 and flat panel displays. 3 For display device applications, TCO films have to meet several requirements simultaneously, i.e. high electrical conductivity, high optical transmittance, process compatibility, and high yield processability by photolithography. In the hydrogenated amorphous silicon (a-Si:H) thin-film transistor liquid crystal display (TFT-LCD) technology, indium tin oxide (ITO) films are mainly used as common (Received November 8, 1995; accepted July 22, 1996) 1806
and pixel electrodes defined on the top and bottom plates, respectively; they can also be used for the source/drain electrode formation in the top-gate TFTs.4,~ In patterning of ITO films, a uniform etch rate and no etching residue (unetched ITO) formation over a large area are critical. The former requirement ensures that all the pixel electrodes are of the same size over the display area, while the latter is to ensure defect free displays. The residual ITO defect will cause shorts between pixel electrodes and metal lines, resulting in pixel failure and image deterioration. So far, ITO films have been patterned by various dry~-17 and wet 's-21 etching techniques. The etching properties of ITO films, such as etch rate and edge roughness, have been associated with the film preparation conditions. 22 However, no work has been devoted to the investigation of the etching residue formation. In
Patterning of Transparent Conducting Oxide Thin Films by Wet Etching for a-Si:H TFT-LCDs
1807
Table I. Characteristics of the ITO and AZO Films, Including Sheet Resistance, Film Resistivity, Average Optical Transmittance, and Film Thickness Sheet TransThick. Resistance Resistivity mittance jk) (fl/sq) (10-4~-cm) (%) ITO Film A ITO Film B ITOFilmC AZO Film C AZO Film D AZO Film E AZO Film F
1200 2800 1200 1000 1000 3000 10000
18 7.5 15.1 103.7 75.0 29.2 5.2
2.16 2.10 1.81 10.37 7.50 8.76 5.24
94.3 87.2 94.4 93.6 91.6 89.7 90.8
this paper, an ITO etching residue formation, etch rate and edge roughness studies for the films deposited at low temperatures will be discussed. It has also been mentioned in the literature t h a t ITO can be used as source/drain electrodes in the topgate TFT structures. 5 In this structure, a tapered edge profile for ITO films is important to ensure a good step coverage of the subsequent deposited Pdoped (n § and intrinsic a-Si:H layers. Therefore, in order to ensure a combination of high quality TFTs and high production yield, a process for obtaining an ITO tapered edge profile is needed. In this paper, we show, for the first time, how the etching solutions composed of different HC1 and HNO 3 ratios can be used to define the ITO tapered edge profiles. Recently, it has been reported t h a t ITO films can become dark/white in a hydrogen containing plasma environment. 23-25In the a-Si:H TFT-LCD fabrication, amorphous silicon nitride (a-SiN:H) is usually deposited by plasma-enhanced chemical vapor deposition (PECVD) technique as a final passivation layer to protect TFTs and ITO pixel electrodes. During this glow discharge process, a plasma containing hydrogen is generated, which will damage the ITO surface. It has been suggested t h a t the degradation of the ITO transparency should be associated with the chemical reduction of indium oxides to metallic indium; and the whitening phenomenon is attributed to the abnormal growth of a-SiN:H on top of the porous silicon oxide layer in the vicinity of the ITO-nitride interface. 25At the same time, the ITO film resistivity increased after such a hydrogen containing plasma treatment. 26 To avoid these problems, either the a-SiN:H deposition conditions must be adjusted, 25 or ITO has to be replaced by another TCO. In this paper, we propose, for the first time, an alternative t r a n s p a r e n t conducting oxide, aluminumdoped zinc oxide (AZO), for the a-Si:H TFT-LCD applications. AZO films are stable under the exposure to hydrogen containing plasma, 27 and they have an optical transmittance and electrical resistivity comparable to ITO filmsY~,29Although a typical resistivity of AZO films is higher t h a n t h a t of ITO films (less t h a n one order of magnitude higher for optimized AZO films), it is still acceptable for the fabrication of the
pixel-electrodes. The patterning ofundoped ZnO films can be accomplished by wet etching in most acids such as HC1, H2SO 4, BHF, and HNO32 ~ The most commonly used etching solution for ZnO films is the H3PO4:CH3COOH:H20 (1:1:30) solution. 31 In the p r e s e n t w o r k , d i f f e r e n t m i x t u r e s of a c i d s (H3PO4:CH3COOH and HCI:HNO 3) diluted with a large amount of D.I. H20 was used to study the patterning characteristics of AZO films. In summary, in this paper, we will focus on the following issues: 9 Patterning characteristics of ITO and AZO films, which includes etch rates, etching residue formations, and the edge roughness of the patterned films; 9 Optical transparency variation of ITO and AZO films induced by glow-discharge containing hydrogen atoms; and 9 Introduction of AZO as an alternative TCO pixelelectrode material to ITO for a-Si:H TFT-LCD applications. EXPERIMENTAL In this work, Corning glass code 7059 was used as substrate for both ITO and AZO films. ITO films were prepared by sputtering methods using hot-pressed sintered targets, which were provided by Leybold Material Inc. (composed of 90 wt.% of In20 ~and 10 wt.% of SnO2). Three kinds of ITO samples prepared at different sputtering conditions were used and named as films A, B, and C. Films A were sputtered at room temperature, and some of them were post-deposition thermally annealed at a temperature of 230~ in air; while films B and C were sputtered at a substrate temperature of 200 and 230~ respectively. Films A were about 1200A thick and had a resistivity of 6.99 • 10-4 and 2.16 • 104 f~-cm before and after thermal-annealing, respectively. Films B had a thickness of 2800A and a resistivity of 2.10 • 104 tl-cm. Films C were about 1200A thick and had a resistivity of 1.81 • 10~ gl-cm. Similarly, AZO films were deposited by DC magnetron sputtering using a hot-pressed sintered target, provided by Plasmaterial Inc. (composed of 98 wt.% ZnO and 2 wt.% Al203). Four different AZO samples were used and named as films C, D, E, and F. The specification and substrate preparation temperatures of these films are described as follows: films C (10.37 • 10 .4 ~-cm, 1000t thick) and E (8.76 • 10 4 tl-cm, 3000A thick) were deposited at room temperature; films D (7.50 • 104 tl-cm, 1000.~ thick), and F (5.24 • 10 4 ~-cm, 1 ~m thick) were deposited at 250~ All the AZO films were deposited at the same sputtering conditions, except the substrate temperature. The detailed characteristics of these TCO samples are given in Table I. The sheet resistance and resistivity for all of the TCO films were determined by a four-point probe method. The crystallinity and microstructure of the films were analyzed by a Rigaku x-ray diffraction (XRD) system (Cu K, radiation). The atomic concentration of the films was analyzed by a Perkin Elmer
1808
Lan, Kanicki, Catalano, Keane, den Boer, and Gu
PHI 5400 X-ray photoelectron spectroscopy (XPS) system. The optical transmittance spectra of the films were measured by a UV-visible spectrum analyzer; the transmittance shown in Table I was obtained from the average values of transmittance taken between 0.39 and 0.7 ~tm. Wet etching of the ITO and AZO films were carried out using different solutions ofHCl:HNO3:H20 , HBr, and H3PO~:CH3COOH:H20. The samples following the Shipley 1813 photoresist coating were pre-baked at a temperature of 90~ for 20 min before UV-light exposure, and post-baked at a t e m p e r a t u r e of 120~ for 30 min before etching. HMDS adhesion promoter was spinned on all of the samples before photoresist application. In order to enhance the etch rate of ITO films, the etching temperature was increased from room temperature (22~ to 40~ and was kept constant during the etching. For AZO films, the temperature was kept at room temperature to slow down the etching rate. After the etching, photoresist was removed by acetone, followed by an isopropyl alcohol cleaning and N~ drying. Etch depth of the films was measured by a Dek-tak IIA surface profilometer. Etch rate was obtained from the evolution of the etch depth as a function of etching time; the slope of this thickness-time evolution yields the etch rate. The etch profile of the patterned films was examined using a J E O L scanning electron microscope (SEM). The taper angle (0) of the edge profiles is defined as the angle between the etched ITO edge surface and the glass substrate, Fig. 5. The optical transparency variation of the TCO films before and after the PECVD a-SiN:H was also studied. The deposition conditions for a-SiN:H films were: 30 sccm for both Sill 4 and N H 3flow rates; 300~ for the substrate temperature; 100 mT for the chamber pressure; and 80 W for the R.F. power. The thickness of the a-SiN:H film was about 4000,~. In addition, an atomic hydrogen flux generated by hotwire chemical vapor deposition (HWCVD) technique was used to further investigate the chemical stability of both ITO and AZO films. The atomic hydrogen treatment was done under the following conditions: 60 sccm for H a flow rate, 100 mT for chamber pressure, 300~ for the substrate temperature, and 1 min for exposure time.
RESULTS AND DISCUSSION P a t t e r n i n g o f TCO films ITO Films
The XRD patterns obtained for ITO films A and B are shown in Fig9 1. Figures l a and lb show the XRD patterns for the as-deposited and post-deposition thermally annealed film A, respectively; while Fig. lc shows the XRD pattern for film B. From the XRD pattern in Fig. la, we can conclude that the asdeposited film A does not show very well defined peaks, indicating that the film structure is mostly amorphous. However, as shown in Fig. lb, the film microstructure changes after a post-deposition ther-
real-annealing, and a very strong peak along (222) direction can be seen. It is expected that the difference in microstructure of these films will result in a variation of etch rate. To check if the atomic composition of ITO film A before and after thermal-annealing remained the same, ITO film A was quantitatively analyzed by XPS. As shown in the Table II, the atomic concentration of oxygen, indium, and tin atoms in the as-deposited ITO films A almost did not change after thermal annealing. Accordingly, it is confirmed that only the film crystallinity has been changed after the thermal-annealing. The etching solution, HCI:HNO 3:H20 = 4:1:5, was used at room temperature to etch ITO films A. Experimental results showed that the etch rates for the 1200A thick ITO film A were about 600 and 46,~/min for the as-deposited and postdeposition thermally annealed films, respectively. It is worthwhile to mention that (a) amorphous-like films do not have a uniform etch rate over a large area because the film crystallinity (in terms of the maximum XRD peak intensity along the direction of(222)) varies over a large area in the same film, and (b) etching residue was observed after the completion of etching. In the case of post-deposition thermally annealed film, a lower etch rate can be associated with a higher crystallinity, and a higher etch rate of amorphous films can be attributed to a higher void ratio and looser atomic structure present in such films. To elucidate the relation between the etch rate and the major XRD peak intensity of ITO film A, the (222) peak intensities were changed by post-depostion thermal-annealing of the film done at different temperature with a fixed annealing time; a similar (222) peak intensity change could also be obtained by varying -
'
9
9
'
I
.
.
.
.
I
'
'
'
"
~
.
.
.
.
1
.
.
.
.
6OO -
ITO Film A
4OO
tf =1200 A
( 2 2 2 ~
(as-deposited)
2OO 0 6O0
.Z"
(222) i l
40O
ITO Film A
tf =12oo A
l/) 200 P
=
0 6O0 (222)(400) j
4O0
I
2O0 0 10
20
30
ITO Film B tf =2800 A
=
40
50
60
Diffraction angle (20) Fig. 1. XRD patterns for the (a) as-deposited ITO film A, (b) postdeposition thermally annealed ITO film A, and (c) in-situ thermally annealed ITO film B.
Patterning of Transparent Conducting Oxide Thin Films by Wet Etching for a-Si:H TFT-LCDs
1809
annealing time at a constant annealing temperature. Figure 2 shows the etch rate variation of ITO film A vs the (222) XRD peak intensity in log-log plot. It is clear from this figure that ITO films having a higher (222) peak intensity tend to have a lower etch rate and those films having a lower (222) intensity show a higher etch rate. In addition, experimental results show that etching residue, of which the mechanism is explained in the next paragraph, always exists in the post-deposition thermally annealed ITO film having low (222) XRD peak intensity (under -600 counts). This phenomenon can suggest that ITO films need to be thermally annealed to attain a certain crystallinity, in order to avoid etching residue formation. Because the as-deposited ITO films usually contain a small amount of the crystalline phase, as shown in Fig. la, it is expected that the difference in etch rate Table II. The A t o m i c C o n c e n t r a t i o n s for t h e A s - D e p o s i t e d a n d P o s t - A n n e a l e d ITO Film A O b t a i n e d from the XPS M e a s u r e m e n t XPS A t o m i c C o n c e n t r a t i o n (%) ITO Film A O In
As-deposited Post-annealed
48.29 48.43
Sn
46.40 46.39
5.31 5.19
1000
c
100
o u,I
H C I : H N O s : H 2 0 = 4:1:5
" ~
T = 22 ~
10
200
i
i
i
[
300
400
soo
soo
J
i
700 0oo
Intensity of (222) peak (a.u.) Fig. 2. Room temperature etch rate vs • (222) peak intensity for ITO film A in the HCI:HNO3:H20 (4,5:0.5:5) solution.
between the amorphous and crystalline phases will result in the ITO residue formation after the completion of etching. As shown in Fig. 3a, etching residues were observed in all areas of the amorphous films. On the other hand, clear etching without the presence of residue was found for the post-deposition thermally annealed film, Fig. 3b. However, even in apparent absence of large etching residue formation in such a annealed film, granular etching residues can still be found in some areas of the film, Fig. 3c. This granular etching residue formation is probably due to a nonuniform distribution of the crystalline phase across the film. Different areas of the film might have different degrees ofcrystallinity. As the etching proceeds, some areas etch faster than the others, and this deviation of etch rates results in the etching nonuniformity, leading to a residue formation. A nonuniform temperature distribution across the sample during the post-deposition thermal-annealing can be at the origin of this problem. In order to reduce the etching residue formation, the microstructure of ITO films prior to etching should be very well controlled; and it would be desirable to have a film that is either totally in amorphous or crystalline structure prior to etching. From a production TFT-LCD throughput point of view, films having only amorphous structure are preferred because a higher etch rate can be achieved for such films. Another alternative to reduce the etching residue formation is to find the etching solutions that have the same etch rate for both amorphous and crystalline ITO, such that etching can proceed more uniformly over the whole substrate area. Finally, it should be mentioned that there is no etching residue formation for film B. This could be attributed to a higher structural uniformity in film B achieved during the deposition of ITO at a higher temperature (in-situ thermally annealed film). Alternative explanation could be associated with the difference in microstructure of ITO films B in comparison to ITO films A. As shown in Fig. lc, film B has a main peak along (400) direction and a minor peak along (222) direction. It has been reported that the crystalline orientation of the sputtered ITO films depends on the film deposition conditions. ~2Therefore, it is possible that the residue presence or absence
a b c Fig. 3. SEM results for the patterned (a) as-deposited ITO film A, showing a residue formation, (b) post-deposition thermally annealed ITO film A, showing no etching residue formation, and (c) post-deposition thermally annealed ITO film A, showing a limited granular residue formation.
1810
Lan, Kanicki, Catalano, Keane, den Boer, and Gu
Table III. R o o m T e m p e r a t u r e E t c h R a t e s o f t h e ITO F i l m s in V a r i o u s E t c h i n g S o l u t i o n s
E t c h Rates: ~Jmin Etchants
ITO Film A ITO Film B
HC1 (37%)
267 400
H B r (46%)
HCI:H~O (1:1)
120 163
HCI:HNO3:H20 (4:1:5)
34 45
46 67
Note: Film A has been treated by a post-deposition thermal-annealing.
a b c Fig. 4. SEM results for the (a) ITO film B etched in the concentrated HCI solution, (b) thermally annealed ITO film A etched in the concentrated HBr solution, and (c) ITO film B etched in the concentrated HBr solution.
might have a certain relationship with the sputtering conditions. Table III lists the room temperature etch rates of the post-deposition thermally annealed ITO film A and ITO film B in various etching solutions. From this table, we have established that films B had higher etch rates in all kinds of etching solutions. This phenomenon could be due to film B having a lower crystallinity (lower XRD (222) peak intensity) in comparison to the annealed film A. From Table III, films A and B show the highest etch rates in the concentrated (37%) HC1 solution. However, poor pattern transferring and rough edge of the ITO films were observed in such a solution; Fig. 4a shows the SEM micrograph of the patterned ITO film B. The failure of this patterning was mainly due to the adhesion weakening ofphotoresist attacked by the concentrated HC1 acid. The poor patterning thus makes HC1 unfavorable for the micro-processing of ITO films. When the concentrated (48%) H B r solution was used, lower etch rates were obtained for both ITO samples in comparison to the etching in the concentrated HC1 solution. However, the H B r solution could define the ITO etched patterns more precisely; and in such a solution, almost no pattern lost due to the sidewall overetching was observed for films A and B. But, the edge profile was not very smooth for both patterned films, and a higher edge roughness occurred for film A than in film B. Figures 4b and 4c show the SEM pictures for the patterned films A and B, respectively. The difference in the patterning characteristics of these two films may be attributed to their difference in the film microstructures shown in Figs. lb and lc. When the concentrated HC1 solution was diluted with D.I. H20 to a volume ratio of 1:1, the etch rate decreased from 46 and 67~Jmin to 34 and 45.~/min for
350 300
L
9
~
, ~
~
.
9
,
.
.
.
l
'
'
.
i
.
'
'
z u
ITO F i l m B
.
1 9
I
18
-I
2so T= G
=~
17
15 0
5 0 0
[HCI]+[HN
../..'J
i
glass
9
i
,
f
o.=
,
,
[
i
"" ~ . . . . ~
O~]=[H~OJ
I
0.4 [HNO 3] /
16
~
t
t
~"
|
,~ ~ 1 3
t
T
,
o.6
i
t
,
12
o.8
[HCI]+[HNO 3]
Fig. 5. Etch rates and taper angle variation for ITO film B as a function of the mixing volume ratio of [HNO3]/[HCI]+[HNO3] solution, where the D.I. H20 is kept at 50% of the total volume. Etching temperature was set at 40~
films A and B, respectively; and the ragged edge profiles of the patterned films seen in the concentrated HC1 solution were not observed in this diluted solution, e.g. the edge profile of the patterned films became smoother and the severe undercutting due to photoresist adhesion failure was reduced. At the same time, a tapered edge profile of the patterned film was found in such a solution. The taper angle was about 12.5 ~ When HNO 3 was added to the diluted HC1 solution, the edge surface roughness of the patterned films was further improved. In addition, the taper angle of the ITO edge profile was also changed and etch rate increased. In order to understand the role of HNO 3 in the diluted HC1 solution, the mixing ratio of HC1 and HNO 3 was varied. Figure 5 shows the etch rate and taper angle variation for film B in various ratios of
Patterning of Transparent Conducting Oxide Thin Films by Wet Etching for a-Si:H TFT-LCDs HC1 and HNO 3 solutions, where the volume ratio of D.I. H20 was kept at 50% of the total solution. By changing the HC1 and HNO 3 mixing ratios, the etch rate and taper angle of the ITO films can be tuned. The etching temperature was increased from 22 to 40~ to enhance the etch rates; the taper angle was not affected by change of the etching temperature. The temperature effect on the variation of etch rate will be discussed in another paragraph. From Fig. 5, we can conclude that the etch rate produced by the mixture of HC1 and HzO was very small, and the etch rate can be increased by adding H N O 3 into the HCI:H20 (1:1) solution. At the same time, we have found from the SEM results that the edge roughness of the patterned films can be reduced. Therefore, the role of HNO 3 is not only to act as an oxidized agent to enhance the etch rate, but it can also improve the surface roughness of the patterned films. The optim u m etching solution in terms of etching rate was produced by the H C l : H N Q : H 2 0 solution with a volume ratio of 4.5:0.5:5. Figure 6 shows the SEM side view for the patterned ITO film in the HCI:HNO 3:H20 (4.5:0.5:5) solution. The smooth and tapered edge sidewalls are clearly visible in this figure. Beyond this optimized mixing ratio (HCl:HNQ:HzO = 4.5:0.5:5), as the HC1 volume ratio of the etching solution became smaller or the HNO 3volume ratio became larger,
Fig. 6. SEM side view of the ITO film B etched in the HCI:HNO~:H~O (4.5:0.5:5) solution at a etching temperature of 40~
a
b
1811 the etch rate decreased. This phenomenon indicates that HC1 is the major acid that etches the ITO films, and the increase o f H N O 3concentration does not help the etching. From experimental result, we have found that the H N Q : H 2 0 (1:1) solution could also yield an ITO tapered edge profile, indicating that H N Q can attack the Shipley 1813 photoresist and weaken the photoresist adhesion to ITO films. However, etch rate (12.3A/ min) in such a solution was much lower than any other solutions. It is interesting to note that the taper angle increased to a maximum and then decreased. This taper angle dependence on the HNO 3 volume ratio in the HC1/H20 mixtures can be associated with a combination of the etch rate variation and change of the photoresist adhesion to the ITO surfaces. Initially, the taper angle increased (or less taper etching effect) as the [HNQ]/([HC1]+[HNOa]) increased. This is because less volume ratio of HCl in the etching solution was available to weaken the adhesion of photoresist to ITO films. Therefore, it is expected that the taper angle will be enhanced as the [HNOa]/ ([HC1]+[HNOa]) ratio increases. In other words, there should be very small or even no taper etching effect when there is no HC1 present in the etching solution; i.e. ([HNOa]/([HC1]+[HNO3]) = 1). However, experimental results showed that the taper angle reached a maxmium of 19.3 ~ when the [HNOa]/([HC1]+[HNOa]) ratio was 0.4 (or HCl:HNOa:H20 = 3:2:5). And beyond this point, the taper angle decreased. This phenomenon could be associated with the weakening of the photoresist adhesion to ITO films by the HNO 3 and it could be explained as following: As the etch rate of etching solution became smaller, the time required to complete the etching became longer. As a result, the adhesion of photoresist to ITO films degrades with longer etching time and this degradation of the adhesion will thus lead to more clear taper etching profile. Furthermore, in order to investigate the HNO 3 influence on the variation of patterned edge profile for HBr solution, the concentrated H B r solution was mixed with 10% and 20% volume ratios of H N O 3. Unlike for the mixture of HC1 and H N Q , only isotropic etching occurred and no clear tapered edge profile was found for ITO films A and B. Thus, the tapered edge formation of ITO films observed for the HC1/HNO 3solutions was not found for the HBr/HNO 3 solutions.
c
Fig. 7. SEM micrographs for (a) ITQ film A, (b) ITO film B, and (c) ITO film C, showing the surface roughness.
1812
Lan, Kanicki, Catalano, Keane, den Boer, and Gu
To investigate the mechanism causing the tapering of ITO etched profile, we have used the same solutions to etch another ITO films (film C) having a similar polycrystalline structure as film B. However, no tapered profile could be found in such films. Therefore, the microstructure of ITO films does not affect the ITO taper etching profiles in the HCl:HNO3:H20 solution. This implies that the tapered etching profiles of ITO films is directly related to the ITOphotoresist adhesion, which can be affected by the surface roughness of ITO films. Figures 7a, 7b, and 7c show the SEM photographs of ITO films A, B, and C. Apparently, ITO film B has a rougher and more porous surface texture than the other two films. Because of the surface roughness and the porous structure in ITO film B, the etchant can penetrate through the photoresist easier than the other ITO films, resulting in the photoresist adhesion weakening to ITO film and causing tapered etching profiles. Therefore, we conclude that a tapered etched ITO profile can only be obtained for films having a certain surface roughness which will affect the adhesion durability between ITO films and photoresist in a certain etching solutions. The influence of the temperature on the etching rate in the HCl:HNO3:H20 (4:1:5) and concentrated H B r solutions have also been studied. Figure 8 shows an Arrhenius plot of the etch rate vs etching temperature for film B, and a similar result can be found for film A. The variation of the etch rate as a function of temperature can be described by: E~(T) = ERoexp(-E /kBT)
(1)
where, E R is the etch rate at temperature T, ER0 is a constant extrapolated at T -~ = 0, E is the activation energy for the etching process, and k Bis the Boltzmann constant. From the curve fitting, we have determined the activation energy for the etching of ITO film in HCl:HNO3:H20 (4:1:5) and H B r solutions, which are 0.61 and 0.58 eV, respectively. Once the activation energy is known, the etch rate at any temperature can be estimated from Eq. (1). The variation of etch rate as a function of temperature can be approximately described as: etch rate increases by about a factor of two for each 10~ increment; and this relation can be obtained from the inset of Fig. 8. AZO F i l m s The electrical and optical characteristics of the AZO films for patterning study are shown in Table I. The films sputtered at higher temperature show a lower resistivity; however, the optical transmittance of the films does not have a large variation for different substrate temperatures. We have also found that the films deposited at the same sputtered condition had different film resistivity; and thicker films had a smaller film resistivity. This can be attributed to the higher carrier concentration and mobility in the thicker films. Thicker polycrystalline films usually contain a larger volume of a larger grain size, such that the free carriers suffered from grain boundary scattering and
1000
^
,
,
,
,
i
,
,
,
,
i
]
J
,o
,
,
J
i
,
r
i
I
'
i
I
I
,
i
]
]
\E=0.,,eV
o
200
9
J~
.c,:..o.:.~O~:l:S
~
~
HBr
"
\
~-~-~
,0o HChHNO~H20
50
i
i
~
i
3
63
J
i
157
I
I
318
J
3.1
f
I
~
--
66"9
I
I
I
3,2
J
I
I
I
3,3
I
1
3.4
3.5
1000 / T (K" 1)
Fig. 8. Arrhenius plot of the etch rate vs T-' for ITO film B in the following etching solutions: HCI:HNOa:H20(4:1:5) and concentrated HBr (48%). 120
'' 'AZO'R|m C' ' ' I ..... tf'=iO'OOA' ' '
8o
Jl
Ts:RT
40 o 2ooo A
tf = 1000A
D
AZO Film
Ts=250oc
looo (b)
o .m
l
e,
e.,
tf =3000A Ts=RT
/~o Rim E 400 (c)
. . . . .
,07,,
O.,mF ~O
i
10
I
(d) .
.
.
.
t,--l m
I~ ,
2O
,
,
.
,
,
30
,
lk..
.
,
4O
.
.
.
.
,
50
.
.
.
.
60
Diffraction angle (2e) Fig. 9. XRD patterns for AZO films having two different thicknesses, deposited at room temperature (a and c) and 250~ (b and d), respectively. Film thicknesses are given on the figures.
trapping can be reduced, leading to a lower resistivity f i l m y The average grain size can be obtained from analysis of the width and intensity of the XRD peaks. The XRD patterns for AZO films C, D, E, and F are shown in Figs. 9a, 9b, 9c, and 9d, respectively. From these figures, we found that all the AZO films either deposited at room temperature or 250~ show very distinct diffraction peak along (002) direction, indicating that AZO films deposited at such sputtering condition are in crystalline phase. These results indicate that poly-crystalline AZO films can be deposited at a lower temperature than ITO films. Also, the higher substrate temperature enhances the film crystallinity, thereby leading to a lower resistivity films. The patterning of AZO films were carried out using the following etching solutions:
Patterning of Transparent Conducting Oxide Thin Films by Wet Etching for a-Si:H TFT-LCDs
1813
Table IV. Room Temperature Etch Rates of the AZO Films in Various Etching Solutions
H~PO4:CH3COOH:H20 (1:1:30)
Etchants
AZO Film AZO Film AZO Film AZO Film 1
104
C D E F 9
9
~500 353 -170 '
,
9
'
'
,
,
,
,
,
9
-
9
,
'
-
.
r
'
"
"
,
AZOFilmF 8 0 0 0
Etch Rates: ~Js HChHNO3:H20 (4:1:200)
-350 103 233 79 9
'
"
,
'
'
'
/
HChHNO3:H20=4:1:500 T=22 ~
A
\
6 0 0 0 Jr O. O
=
'10
-
.=~ ~r,ee
4 0 0 0
2 0 0 0
:
,
,
20
.
9
.
,
40
.
9
,
J
60
,
,
,
i
80
,
,
,
i
100
,
J
,
i
120
,
,,
,
I
1 4 0
,
,
,
6 0
Etching time (sec)
Fig. 10. Etch depth variation as a function of etching time for AZO film F, using a etching solution of HCI:HNO~:H20 (4:1:500) at room
temperature.
H3POjCH~COOH:H20 (1:1:30) and H C l : H N Q : H 2 0 (4:1:200 and 4:1:500). The etching temperature was kept constant at room temperature. The etch rates for different films in these solutions are shown in Table IV. In comparison to ITO films, the etch rates for AZO films were at least two orders of magnitude higher even in a more diluted etching solutions and lower etching temperature, indicating that AZO films are much easier etched than ITO films. Experimental results showed that AZO films deposited at lower temperature had a higher etch rate than those deposited at higher temperature in all etching solutions. Similar to the etching behavior of ITO films, this can be attributed to the difference in the crystallinity or crystalline grain size of AZO films. Because crystalline films have more compact atomic structure than that for the amorphous films, etch rate is usually lower for the crystalline films. The film having higher crystallinity also contains a larger volume fraction of crystalline phase. As a result, etching will proceed more slowly in the film having a higher crystallinity. In addition, we have found that etch rate depends on the film thickness, with thicker films showing smaller etch rates in all solutions. This might indicate that the films have a larger volume fraction of amorphous phase adjacent to the substrate surface; this could result from the lattice mismatch during the film nucleation near the glass substrate. After several atomic layers have been grown, the films will become more crystalline, and more uniform etching of the
HCI:HNO3:H20 (4:1:500)
~300 89 166 61
films can be achieved. This phenomena can be observed by analyzing the variation of the etch depth as a function of etching time, shown in Fig. 10. The etching process was carried out in the solution of HCl:HNO3:H20 (4:1:500) at room temperature. Initially, the etch rate was low and was kept at a rate of about 46A/s, indicating that an uniform poly-crystalline grains existed closer to the film surface or farther away from the film-substrate interface. However, as the etching proceeded closer to the glass substrate, the etch rate gradually increased to 81A/s. A similar etching has also been found for another AZO film, which was deposited at room temperature and had a thickness o f - 9 0 0 0 A . This observation implies that the film crystallinity varied along its growing direction and this variation of etch rates is in agreement with our assumption mentioned above. In short, for AZO films we have found that the films having a lower and a higher XRD peak intensities show a higher and a lower etch rates, respectively. Films deposited at higher temperature tend to have a smaller etch rate. In addition, thicker films have a smaller average etch rate than the thinner one even though they were deposited at the same conditions. We have found that when the H3PO4:CH3COOH:H20 (1:1:30) solution was used, the AZO film deposited at room temperature (film C) had a smoother etched edge profile than the film deposited at 250~ (film D). However, when the HCl:HNO3:H20 (4:1:200 or 4:1:500) solution was used, smooth edge patterns were obtained for both films C and D. Therefore, the edge roughness of the patterned AZO films would not only depend on the etching solution used, but it also depends on the preparation conditions of the films. In order to know the patterning resolution, a SEM picture of a 3 ~m line-and-space photoresist pattern is used for comparison and it is shown in Fig. 1 la. Figures 1 lb and 11c show the SEM results for the p a t t e r n e d films C and D in the HCl:HNO3:H20 (4:1:200) solution, respectively. Both patterned films show very smooth line edges and very small sidewall etching, which are desirable profiles for the a-Si:H TFT-LCD fabrication. Similar results were also observed in the HCl:HNO3:H20 (4:1:500) solution. Although isotropic etching always occurs in the wet etching process, theosidewall etching is not very severe for these 1000A thick AZO films. On the contrary, a different etching result was found for the 1 ~tm thick films in all of the above solutions; about 1 ~tm of pattern lost due to isotropic etching was observed in
1814
a
Lan, Kanicki, Catalano, Keane, den Boer, and Gu
b
c
d
Fig. 11. SEM micrographs for (a) the patterned photoresist on top of AZO film, (b) patterned AZO film C, (c) patterned AZO film D, and (d) patterned AZO film C; in Figs. (a)-(c), a 3 ~tm space-and-line is used for comparison.
such films. Extra undercutting due to over-etching for both thick and thin films can be reduced by diluting the etching solutions with D.I. H20 to a reasonable ratio, such t h a t the etch rate can be reduced and the etch time can be precisely controlled. Figure 11d shows the SEM picture of the patterned film C; clear and smooth space-and-line patterns were easily obtained. In addition, it is clear from this figure t h a t no etching residue (unetched AZO) formation was found in such film. The same result was also found for film D, E, and F. Unlike for ITO films, the AZO films deposited at either room temperature or 250~ show no residue formation after the completion of the etching. This result could be attributed to the microstructure or the crystallinity difference of the films. Because all of the AZO films are crystalline, there is no etch rate difference problem seen in the ITO film A. Therefore, the etching residue formation could be avoided for the AZO films. We have also found that, unlike ITO films, AZO films can be more easily etched by the diluted H3POjCH~COOH, HCI:HNO3, HBr, and BHF etching solutions, thereby introducing some limitation on how AZO films could be used in the a-Si:H TFT-LCD processing. In another words, etching selectivity between AZO film and its adjacent layers such as aSi:H, a-SiN:H, and metals will be a critical issue. We are currently working on this issue and further results will be published later. However, by properly designing the process flow and/or changing the TFT structure, AZO films can still be used in the TFT-LCD technology. Figure 12 shows a proposed bottom-gate tri-layer TFT structure where AZO is used as a pixel electrode material for a high aperture ratio a-Si:H TFT-LCDs. In this structure, AZO film is deposited as the final layer, such t h a t the exposure of AZO films to acid based solutions can be avoided.
Optical Properties of ITO and AZO Films Optical Transmittance and Energy Bandgap Figure 13a shows the transmittance spectra for ITO film A and AZO film C. In the near-IR and visible light region of the spectra, the transmittance for both films are almost the same. The average values of transmittance in the visible light region are 94.4 and 93.6% for the ITO and AZO films, respectively. In the short wavelength region, the transmittance of AZO film decreases faster t h a n ITO film, which indicates t h a t AZO film has a smaller optical bandgap. The
a-Si:H(i)
\\
a-SiNx:H
~
\
Ta(Gate)
TffAI
~a-/S~:H(n')
Ta205
AZO
a-SiNx:H
Fig. 12. Cross-sectional view of a proposed a-Si:H TFT structure with AZO film being used as a pixel-electrode material.
optical bandgap of TCO films can be obtained from the absorption spectra given in Fig. 13b. The absorption coefficients of TCO films at different wavelength can be obtained from transmittance and reflectance data from the following equation: T = T0exp(-a-t f)
(2)
where, a and tf are the absorption coefficient and thickness of the film, respectively; T is the optical transmittance of the film; T o is a constant t h a t represents the percentage of optical light intensity transmitted through the air-film interface and it can be expressed as T O-- 1 - Ro, where R o is the optical reflectance. If the optical reflectance between TCO film and its ambient is neglected, T Ocan be approximate to unity (T 0-- 1) and the absorption coefficient can easily be obtained from Eq. (2) once the film thickness is known. The absorption coefficient for the direct bandgap allowed transition can be given by 34'35 r
c~= (hv - Eg) '/2
(3)
where, hv is the photon energy and E is the transition energy gap. Figure 13b shows the square of absorption coefficient (a 2) vs photon energy (hv) for the ITO film A and AZO film C. By extrapolating the straight portion of the a 2 curve to the hv axis (~2 = 0), the energy gaps for these TCO films can be obtained and they are 4.04 and 3.63 eV for the ITO film A and AZO film C, respectively. The optical bandgap values for the other films can be obtained by using the same procedures, and they are listed in Table V. Because there is no direct evidence for whether ITO or AZO films are direct or indirect bandgap semiconductors, it is convenient to define an optical energy ( E ~ ) for a given a, for example, (~= 6 • 104 cm -1. A plot of(~ in logarithm scale vs hv is shown in Fig. 13c. The optical energy values obtained by this definition for all.TCO films are listed
Patterning of Transparent Conducting Oxide Thin Films by Wet Etching for a-Si:H TFT-LCDs
1815
, , , I , , , i , , , i 9 , , i , 9 , i , , , i
-
8o
~
60
rro n,rn A .....
40
r
1
AZO Rim i
101~
_
7
t
'"
ITO Film A
.....
/
9
10 4
"~~ 20 300
0 400
500
600
700
800
900
3.2
a
,I ~
1 ,,
,
10"
1000
Wavelength (nm)
3.4
3.6
..~'"~
AZO Rim C
3.8
4
4.2
I 3.2
4.4
I
,,L~,,,,,,.I,,,~,J,,.,
Photon energy (eV)
3.4
3.6
3.8
4
4.2
Photon energy (eV)
b
c
Fig. 13. (a) TransmittancespectrafortwoTCOfilms:thermallyannealed ITOfilmA (
) and AZO film C (--...), (b) plots of the square of absorption coefficients (c~2) as a function of photon energy (hv) for the TCO films, and (c) plots of a vs photon energy for the TCO films.
Table V. Energy Bandgap of TCO Films Obtained from the Transmittance Spectra of the Films Eg (eV)
E~ (eV)
4.04 3.84 3.63 3.72 3.69 3.60
3.92 3.70 3.60 3.78 3.48 3.33
9O 8o
7O
ITO Film A ITO Film B AZO Film C AZO Film D AZO Film E AZO Film F
60 "F" g~ e-
as o~= 6 x 104 cm -~)
._tf = 1 2 0 0
50 ~
1:
40
z: . . . . . 3: . . . . .
A
ITO ITO/a-SiN:H ITQ: H treatment
. . . . .., . . . . -'"
] ~, ~'l'
"
P~
20 i ,, , .... , .... I 9 ,~j .... I ....... I 300 400 500 600 700 800 g00
in Table V. The optical bandgaps of AZO films obtained by both methods are, in general, smaller t han those of ITO films, implying t h a t the cutoff wavelength of AZO films is smaller. Nevertheless, it is still suitable for t r a n s p a r e n t electrodes for optoelectronic devices.
Figures 14a and 14b show the t r a n s m i t t a n c e spect r a for ITO film A (thermally annealed) and AZO film C before and after the interaction with atomic hydrogen g en er ated either during the a-SiN:H deposition by PECVD technique or during the hydrogen dissociation by HWCVD technique. In the first case, a 4000A thick a-SiN:H was deposited on both ITO and AZO films simultaneously; while in the second case, ITO and AZO films were exposed to atomic hydrogen flux for i rain at 300~ The optical t r a n s m i t t a n c e of ITO film A was reduced from 94.3 to 91.3% and from 94.3 to 29.0% for cases I and 2, respectively. Similar results showing degradation of optical t r a n s m i t t a n c e were also found for the other ITO films. In contrast, the optical t r a n s m i t t a n c e of AZO film C changed from 93.6 to 95.0% and from 93.6 to 91.5% for cases 1 and 2, respectively. Similarly, almost no optical t r a n s m i t t a n c e variation has been found for the other AZO films. These results indicate t h a t AZO films are not only stable during the a-SiN:H film deposition, but it is also very stable in the presence of atomic hydrogen flux generated by HWCVD technique. It is interesting to note
~'~ i
30
Note: Optical gap of TCO films. Eg: direct transition; E~(defined
Optical T r a n s m i t t a n c e V a r i a t i o n of TCO F i l m s in H y d r o g e n C o n t a i n i n g P l a s m a
~
Wavelength
100 ~. . . . ,s
g
8o
~
70
~
60
=
5o
.... ,r-4', w''"
ij
1000
(nm)
, .... 7 .... 4
AZO
30 20 300
400
500
600
700
Wavelength
800
900
1000
(nm)
b Fig. 14. Transmittance spectra for the (a) thermally annealed ITO film A and (b) AZO film C. Solid ( ), dash (- - -), and solid-dash ( - - -) lines represent the spectra for the as-deposited TCO film, following the deposition of a 4000A thick of a-SiN:H film on the surface of TCO film, and after the atomic hydrogen treatment of TCO film, respectively.
t h a t there is a small increase in the optical transmittance of AZO film C after the deposition of a-SiN:H films. This increase of t r a n s m i t t a n c e could be due to the anti-reflection effect introduced by the a-SiN:H film on the top of AZO; i.e., a smaller refractive index difference is obtained between AZO and a-SiN:H films t h a n between AZO and air.
1816
Lan, Kanicki, Catalano, Keane, den Boer, and Gu
D e p o s i t i o n o f AZO f i l m s in DC M a g n e t r o n Sputtering System
In the last paragraph, we will introduce the electrical and optical characteristics of AZO films deposited by a reactive D.C. magnetron sputtering method using a composite target. As shown in Fig. 15a, the target is composed of two different discs: a perforated aluminum disc placed on top of a ceramic ZnO disc. The percentage of perforated area on aluminum disc and oxygen flow rate have been optimized to produce a film having both lowest resistivity and highest transmittance. During the sputtering, the substrate temperature was kept at 213~ Figure 15b shows the resistivity of the sputtered AZO films as a function of oxygen flow rate; the lowest resistivity (3.1 • 10-4 gtcm) AZO films were obtained for an oxygen flow rate ranging from 2.75 to 2.88 sccm. By similarity to ITO films, we have found that the resistivity of AZO films also depends on the film thickness. This phenomenon can be attributed to the variation of free carrier concentration and mobility with film thickness. This observed variation is a direct consequence of the variation in film crystallinity with film thickness. The optical transmittance spectra for two films with different thicknesses are shown in Fig. 15c. By properly controlling the oxygen partial pressure and other sputtering conditions, AZO films can have an electrical resistivity and optical transmittance comparable to ITO films; and the device quality AZO films can be obtained by using a sintered target composed of ZnO with certain percentage ofA1203. CONCLUSIONS In this paper, we have shown that the etch rates of TCO films could depend on the film crystallinity, and the residue formation can be attributed to the etch rate difference between the amorphous and crystalline phases of the film. We have also observed that the patterned edge profiles and the extent of edge roughness depend on both the film properties and the etching solutions used; and a smooth edge profile can be achieved when proper etchants are chosen. We have also shown that ITO films with tapered edge profiles after patterning can be obtained using mixture of HC1 and HNO~, and the taper angles can be tuned by adjusting the mixture concentration to a proper level. The optical transparency of ITO films has been shown to degrade during the deposition of aSiN:H films, and become much worse in the atomic hydrogen environment. In contrast, AZO films were observed to be chemically stable under the same experimental condition. Our results have shown that AZO films can have an optical transmittance and electrical resistivity comparable to ITO films. Furthermore, wet etching of AZO films yields sharp and smooth patterns. AZO films have very high etch rates in m a n y etching solutions; and unlike ITO films, AZO films deposited at either room temperature or an elevated temperature, show no residue formation after the completion of the etching. The properties of high etch rate and no residue formation for the AZO
to2 AI-DOPED ZnO AS-DEPOSITED 0.8
, 9 .,,,
.......
,
I0 I
~L_..
F-
~
,~ . ,,
FILM THICKNESS (~1
rd
(z~l If [ n m ) ~
(2771
9
T(~I (26oi (~o: OCWiAGNEI~ONSPUTTER
(a) 25
50
55
OXYGEN FLOYV(seem}
(b) IOO 80
~_ 40 ~ 20 0
"
/
/
I
ZnO
400
6o0 WhYELEN6D-t(nm)
T~:zc3"c 10o0 1200
(c) Fig. 15. (a) Configuration of the composite sputtering target used for the deposition of AZO film, (b) variation of resistivity of the asdeposited AZ.O films as a function of oxygen flow rate in a D.C. magnetron sputtering system, (The inset shows the resistivity variation with the film thickness), and (c) the optical transmittance spectra for two AZO films having two different thicknesses.
films could make them more attractive than ITO films for a-Si:H TFT-LCDs applications, especially from the throughput and yield point of view. Owing to the advantages such as high electrical-optical quality, chemical stability, and etching feasibility, we propose, for the first time, that AZO can be used as an alternative pixel-electrode material to ITO in the aSi:H TFT-LCDs. ACKNOWLEDGMENT The authors would like to thanks Mr. Bernd Heinz, from Leybold Material Corporation, for providing the ITO films, Mr. Pheng-Piao Liao and Chun-Sung Chiang for their assistance in operating the PECVD system, Mrs. Shuangying Yu for HWCVD, and Jeffrey Politis for helping with the UV-visible spectrum analyzer. This work was supported by the University of Michigan, Center for Display Technology and Manufacturing. REFERENCES 1. G. Tao, M. Zeman and J.W. Metselaar, Solar Energy Materials and Solar Cells 34, 359 (1994). 2. E. Fortunato, M. Vieira, L. Ferreira, C.N. Carvalho and G. Lavareda, Conf. Proc. MRS Spring Mtg. 1993, Syposium A (San Francisco, CA: Mater. Res. Soc., 1993), p. 981.
P a t t e r n i n g of T r a n s p a r e n t Conducting Oxide Thin F i l m s by Wet E t c h i n g for a-Si:H TFT-LCDs
3. G. Kawachi, E. Kimura, Y. Wakui, N. Konishi, H. Yamamoto, Y. Matsukawa and A. Sasano, IEEE Trans. on Electron Dev. 41, 1120 (1994). 4. T. Sunata, T. Yukawa, K. Miyake, Y. Matsushita, Y. Murakami, Y. Ugai and J. Tamamura, IEEE Trans. on Electron Dev. 33, 1212 (1986). 5. T. Yukawa, K. Amano, T. Sunata, Y. Ugai and K. Okamoto, Proc. Japan Display '89, p. 506. 6. Z. Calahorra, E. Minami, R.M. White and R.S. Muller, J. Electrochem. Soc. 136, 1839 (1989). 7. C. Barratt, C. Constantine and D. Johnson, SID 95Digest, p. 681. 8. R.J. Saia, R.F. Kwasnick and C.Y. Wei, J. Electrochem. Soc. 138, 493 (1991). 9. L.Y. Tsou, J. Electrochem. Soc. 140, 2965 (1993). 10. M. Takabatake, Y. Wakui and N. Konishi, J. Electrochem. Soc. 142, 2470 (1995). 11. M. Yokoyama, J.W. Li, S.H. Su and Y.K. Su, Jpn. J. Appl. Phys. 33, 7057 (1994). 12. Y. Kuo, Jpn. J. Appl. Phys. 29, 2243 (1990). 13. I. Adesida, D.G. Ballegeer, J.W. Seo, A. Ketterson, H. Chang, K.Y. Cheng, and T. Gessert, J. Vac. Sci. Technol. B 9, 3551 (1991). 14. M. Mohri, H. Kakinuma and M. Sakamoto, Jpn. J. Appl. Phys. 29, L1932 (1990). 15. K. Nakamura, T. Imura, H. Sugai, M. Ohkubo and K. Ichihara, Jpn. J. Appl. Phys. Part I 33, 4438 (1994). 16. T. Minami, T. Miyata, A. Iwamoto, S. Takata and H. Nanto, Jpn. J. Appl. Phys. 27, L1753 (1988). 17. H. Sakaue, M. Koto and Y. Horiike, Jpn. J. Appl, Phys. 31, 2OO6 (1992). 18. T. Ratcheva and M. Nanova, Thin Solid Films 141, L87 (1986).
1817 19. J.E.A.M. van den Meerakker, J. Electrochem. Soc. 140, 471 (1993). 20. M.V. Tesan, S. McGee and U. Mitra, Thin Solid Films 170, 151 (1989). 21. V. Hochholzer, E. Lueder, T. Kallfass and H-U. Lauer, SID 94 Digest, p. 423. 22. M. Inoue, T. Matsuoka and Y. Fujita, Jpn. J. Appl. Phys. 28, 274 (1989). 23. O. Kuboi, Jpn. J. Appl. Phys. 20, L783 (1981). 24. S. Major, M. Kumar, M. Bhatnagar and K.L. Chopra, Appl. Phys. Lett. 49, 394 (1986). 25. E. Kimura, G. Kawachi, N. Konishi, N. Konishi, Y. Matsukawa andA. Sasano, Jpn. J. Appl. Phys. Part i 32, 5072 (1993). 26. T. Minami, H. Sato, H. Nanto, and S. Takana, Thin Solid Films 176, 277 (1989). 27. H.C. Weller, R.H. Mauch and G.H. Bauer, Conf. Rec. 22 IEEE PVSC. (New York: IEEE, 1991), p. 1290. 28. H. Sato, T. Minami, S. Takata, T. Miyata and M. Ishii, Thin Solid Films 236, 14 (1993). 29. K. Ellmer, F. Kudella, R. Mientus and R. Schieck, Thin Solid Films 247, 15 (1994). 30. F.S. Hickernell and T.S. Hickernell, Ultrasonics Symposium (New York: IEEE, 1992), p. 373. 31. G.D. Swanson, T. Tamagawa and D.L. Polla, J. Electrochem. Soc. 137, 2982 (1990). 32. C.V.R. Vasant Kumar and Abhai Mansingh, J. Appl. Phys. 65, 1270 (1989). 33. Y. Qu, T.A. Gessert, K. Ramanathan, R.G. Dhere, R. Noufi, and T.J. Coutts, J. Vac. Sci. Technol. A 11,996 (1993). 34. Y. Ohhata, F. Shinoki, and S. Yoshida, Thin Solid Films 59, 255 (1979). 35. S. Ray, R. Banerjee, N. Basu, A.K. Batabyal, andA.K. Barua, J. Appl. Phys. 54, 3497 (1983).
