ARTICLE IN PRESS. Microelectronics Journal

Size: px

Start display at page:

Download "ARTICLE IN PRESS. Microelectronics Journal"

Marybeth Potter
6 years ago
Views:

Microelectronics Journal 40 (2009) 15 19 Contents lists available at ScienceDirect Microelectronics Journal journal homepage: www.elsevier.

Key Laboratory for Thin Film and Microfabrication of the Ministry of Education, Institute of Micro/Nano Science and Technology, Shanghai Jiao Tong University, Shanghai 200030, PR China b Shandong

1 Microelectronics Journal 40 (2009) Contents lists available at ScienceDirect Microelectronics Journal journal homepage: The influence of the AlN film texture on the wet chemical etching Da Chen a,b, Jingjing Wang b, Dong Xu a, Yafei Zhang a National Key Laboratory of Nano/Micro Fabrication Technology, Key Laboratory for Thin Film and Microfabrication of the Ministry of Education, Institute of Micro/Nano Science and Technology, Shanghai Jiao Tong University, Shanghai , PR China b Shandong University of Science and Technology, Jinan , PR China article info Article history: Received 3 July 2008 Accepted 4 September 2008 Available online 23 October 2008 Keywords: Aluminum nitride Film texture Wet chemical etching Anisotropy abstract The influence of the aluminum nitride (AlN) film texture on the chemical etching in KOH solution was invested. The AlN films with the different texture and crystal quality were prepared by sputtering. It is found that the chemical etching behaviors, including the etch rate, the activation energy, the surface morphology and the anisotropy, are strongly dependent on the film texture. There is a faster etching in the case of mixed (10 0) and (0 0 2) texture and a lower rate in the case of only (0 0 2) texture. The etch rate also decreases with the crystal quality. The sample with the only (0 0 2) texture forms discontinuous column structure after etching and exhibits lower porosity compared to that of the mixed (10 0) and (0 0 2) texture. Due to the strong anisotropy of the AlN wurtzite structure, the morphology of the film deposited at 700 1C shows the homogeneous pyramid shape after etching. The cross-section micrographs of etching patterns indicate that the anisotropy of the chemical etching is improved with the improving of the crystal quality. & 2008 Elsevier Ltd. All rights reserved. 1. Introduction With the tremendous progress of III-nitrides research in terms of both fundamental understanding as well as devices applications, aluminum nitride (AlN) attracts increasing interest due to the band gap of approximately 6.2 ev and some excellent properties. The direct wide band gap is very attractive for the ultraviolet light-emitting diodes [1] and the photo detectors [2]. The high thermal conductivity (340 W m 1 K 1 ) makes AlN suitable for the high power applications [3]. Furthermore, the polycrystalline AlN films have been utilized to fabricate the piezoelectric sensors/ actuators [4] and the high-frequency electro-acoustic devices [5] due to the piezoelectric properties and high acoustic velocity. In the devices fabrication, the patterning of the materials is a key step because a number of device performances may be affected by the etching process. Generally, the reactive ion etching using chlorine can be highly anisotropic, an ideal characteristic for producing vertical profiles [6]. However, due to the strong physical component and the high chemical activity of the chlorine plasma, the dry etching has low etch-selectivity between materials and can cause subsurface damage by ion bombardment. In contrast, the wet etching, produces negligible damage, can be highly selective, is relatively inexpensive, and can be done with simple equipment. In addition, wet chemical etching is a Corresponding author. Tel.: ; fax: address: yafeizhan@sjtu.edu.cn (Y. Zhang). reasonable, reliable, and simple method to analyze the defects and crystal polarity in III-nitrides [3]. Unfortunately, there was relatively little success in developing wet etching solutions for AlN because of their excellent chemical stability. The chemical etching was strongly dependent on the crystal quality and the etch temperature. Only KOH or NaOH containing solution can be etch epitaxial and single crystal AlN at the temperature below 80 1C [3,7]. The hot H 3 PO 4 [8], and AZ400 K photoresist developer [9] were also reported can etch the polycrystalline AlN. However, more detail information about etching behaviors of polycrystalline AlN films is required for the micro-device fabrication. In this paper, the influence of the film texture on the chemical etching in KOH solution was invested. The AlN films with the different texture and crystal quality were prepared by sputtering. It is found that the etch rate, the activation energy and the morphology of etching surface depend on the films texture and the crystal quality. The micrographs of the patterns were observed from the top view and the cross-section view to detect the etching evolution and anisotropy. 2. Experimental details The AlN films were deposited using a RF sputtering system (ANEIVA SPF-210F). The target was a 4-in-diameter aluminum with 99.99% purity. Prior to the AlN deposition, the 300 nm W film was sputtering, on in. Si (111) wafers as the bottom electrodes. The chamber was evacuated until the base pressure decreased to /$ - see front matter & 2008 Elsevier Ltd. All rights reserved. doi: /j.mejo

2 16 D. Chen et al. / Microelectronics Journal 40 (2009) Table 1 Sputtering parameters for AlN films deposition Target-to-substrate spacing RF power Pressure Ar flow rate N 2 flow rate Substrate temperature 6 cm W 0.55 Pa 5 sccm 5 sccm RT-600 1C o Pa. High-purity argon was then introduced and presputtered the target for 15 min before the film deposition. The sputtering parameters were summarized in Table 1. The thicknesses of all the films were controlled to about 1 mm. In order to investigate the effect of the crystal quality on the etching behaviors, the substrate temperature was turned from room temperature (RT) to 600 1C. The AlN etchant used in the experiment was 10 wt% KOH solution. In order to demonstrate the suitability of this wet etching for device applications, the test patterns were formed. The Ti layer with the thickness of 200 nm was deposited on the AlN surface and patterned by the conventional photolithography method as the mask for patterning AlN. The Ti layer is a better candidate mask for AlN-based devices because the high cohesion to AlN and the stability in AlN etchants. The etch depths were obtained by Dektak stylus profilometer at different position after the removal of the mask with an approximate 5% error. The crystal structure and the texture of the films were identified by X-ray diffraction (XRD) and rocking curve (BRUKER-AXS) at wavelength of nm (Cu Ka). The morphological investigations were performed using field emission scanning electron microscopy (FE-SEM, FEI SIRION 200). 3. Results and discussion 3.1. The influence of the film texture Fig. 1 displays the XRD patterns of the AlN films for three RF power levels. The substrate temperature was RT. In the case of 50 and 100 W, the diffraction peaks of AlN (10 0) and (0 0 2) were observed at 2y around 331 and 361, respectively. The peak at corresponds to the W (110) orientation. With the increasing of RF power, the (0 0 2) peak becomes stronger and sharper, while the (10 0) peak becomes weaker. The sample deposited at 150 W exhibits the only (0 0 2) texture with the full-width at halfmaximum (FWHM) of 0.311, which indicates the preferred c-axis orientation. The result can be explained by that the AIN complex particles have sufficient kinetic energy to move to the lowest energy state and advantage the formation of the (0 0 2) texture crystalline structure at high power. Fig. 2 shows the etch rates as a function of temperature in KOH solution for the films with the different texture. During etching, the bubbles were observed on the sample surface; meanwhile the white floccules were formed in the etchant. The etching occurred by the following reaction formula [10]: AlN þ 6KOH! AlðOHÞ 3 k þ NH 3 m þ 3K 2 Ok As expected, the etch rates of all the films increase sharply as the etch temperature increases from 30 to 80 1C. The etch rate of the film with the only (0 0 2) texture is around 55 nm/min at 30 1C and increase to 420 nm min 1 at 80 1C, which is faster than the reported value of the epitaxial films and single crystal [3]. The etch rate of the films with the mixed (10 0) and (0 0 2) texture is obviously higher than the case of (0 0 2) texture. Combining the etching rates and the XRD patterns, it can be concluded that the (10 0) plane are etched preferentially compared to the (0 0 2) Fig. 1. The XRD patterns of AlN films prepared in three RF power levels. (a) 50 W; (b) 100 W and (c) 150 W. The substrate temperature was RT. Etch Rate (nm/min) W 100 W 150 W plane. This selective etching behavior was also observed in AlN (10 1) and (0 0 2) planes by Ababneh et al. [11]. For the (0 0 2) texture, the c-axis is normal to the substrate and the plane parallel to the substrate is the close-packed basal plane, with either all aluminum or nitrogen atoms. Other planes should form the lesserpacked atomic array [12] and will suffer higher etch rate than the (0 0 2) plane in the etchant The effects of crystal quality Etch Temperature ( C) Fig. 2. The etch rate as a function of temperature in 10 wt% KOH solution for the films deposited at different RF power. (a) 50 W; (b) 100 W and (c) 150 W. To get more detailed understanding for the correlation between the etching behaviors and the film texture, the substrate temperature was turned from 300 to 600 1C to get the higher (0 0 2) texture. The crystal quality was identified by X-ray rocking curve. As shown in Fig. 3, the rocking curves strengthen in intensity and become narrower with the increasing of substrate temperature, which means the reducing of defects and the improvement of crystal quality. A parameter of interest for the etching experiments is the activation energy E a. This is generally obtained by fitting an

D. Chen et al. / Microelectronics Journal 40 (2009) 15 19 17 (c) Intensity (a.u.) (b) (a) 5 10 15 20 25 30 θ ( ) Fig. 3. The typical X-ray rocking curves of AlN films deposited at different temperature.

3 D. Chen et al. / Microelectronics Journal 40 (2009) (c) Intensity (a.u.) (b) (a) θ ( ) Fig. 3. The typical X-ray rocking curves of AlN films deposited at different temperature. (a) RT; (b) 400 1C and (c) 600 1C. Etch rate (nm/min) 1000 (a) (g) Arrhenius plot to the form [13]. k ¼ A exp E a RT 50 W 100 W 150 W 300 C 400 C 500 C 600 C where k is the etch rate at temperature T, R is idea gas law constant (8.314 J mol 1 K 1 ), and A is the pre-exponential factor which can be thought of as an attempt frequency for the reaction between (OH) ions and the reaction surface. Fig. 4 shows the Arrhenius plots of the samples deposited at 300, 400, 500 and 600 1C. The etch rate of the samples deposited at 50 and 100 W were again shown in this figure. The AlN films deposited above 300 1C were found very high resistant to the KOH solution at the temperature o40 1C. It is remarkable that the effective etching starts at a temperature above 60 1C for the AlN films deposited at 600 1C. The etch rate drops significantly as the substrate temperature increased, which is a clear indication of the dependence of the wet etching on the material crystal quality. The E a is different between the films with varied texture. The E a of the films with the mixed (10 0) and (0 0 2) texture is kj mol 1, which is in the range of the diffusion-limited etching (d) (b) (c) (e) (f) 1000 / T (K) Fig. 4. The Arrhenius plots of the AlN films deposited at different RF power ( W) and substrate temperature ( C). (a) 50 W and RT; (b) 100 W and RT; (c) 150 W and RT; (d) 300 1C and 150 W; (e) 400 1C and 150 W; (f) 500 1C and 150 W (g) 600 1C and 150 W. The etchant was 10 wt% KOH solution. Fig. 5. The surface morphologies before and after etching for the AlN films with different texture. (a) The as-deposited surface of the mixed (10 0) and (0 0 2) texture (deposited at 50 W); (b) the etching surface of the mixed (10 0) and (0 0 2) texture (deposited at 50 W); (c) the as-deposited surface of the only (0 0 2) texture (deposited at 150 W and RT) and (d) the etching surface of the only (0 0 2) texture (deposited at 150 W and RT). The samples were both exposed in 10% KOH solution for 30 s at 40 1C. mechanism ( kj mol 1 ) [3]. However, the E a of the samples with the only (0 0 2) texture is kj mol 1, which is a typical characteristic of the rate-limited etch mechanism. The disparity of the valves means that both reaction rate-limited and diffusion-limited mechanisms lie in the etching reactions. It has been reported that the wet etching of the epitaxial AlN films and single crystal usually proceeds in rate-limited regime [3,9]. In this experiment, due to the improving of the (0 0 2) texture, the ratelimited reaction dominates the etching process. On the other hand, in the case of poor (0 0 2) texture, the etch rate is very fast. The characteristics of diffusion-limited etching can be explained that the reaction is so rapid that the solution becomes depleted of reactants near the surface [14] The morphology of the etching surface Fig. 5 shows the surface morphologies before and after etching for the AlN films with the different texture. The samples shown in Fig. 5(a) and (b) were deposited at 50 W and has the mixed (10 0) and (0 0 2) texture. The as-deposited grains exhibit the packed polygonal topography (Fig. 5a). After exposing in KOH solution for 30 s at 40 1C, the polygonal crystals were heavily destroyed and were replaced by the snatchy hillocks (Fig. 5(b)). As for the samples deposited at 150 W with the only (0 0 2) texture, the surface of as-deposited films consists of the lenticular-like crystals (Fig. 5(c)). After etching for 30 s the resulting pattern (Fig. 5(d)) forms the discontinuous column structure and exhibits the lower porosity compared to the sample deposited at 50 W.Fig. 6(a c) shows the etching evolution of the sample deposited at 600 1C in KOH solution at 50 1C. Fig. 6(d). gives the high-magnification image of Fig. 6(c). After the short exposure (1 min), the caved vein was clearly visible along the grain boundary, which means that the wet chemical etching in polycrystalline AlN starts along the grain boundary. After 5 min etching these veins have extended downward and the homogeneous pyramid with the triangular shape have developed on the etching surface. The faces of the pyramids are seen to be extremely smooth and symmetric. It is

18 D. Chen et al. / Microelectronics Journal 40 (2009) 15 19 Fig. 6. The surface morphologies of the etching evolution of the AlN films deposited at 600 1C in 10% KOH solution at 40 1C.

also noted that all the pyramids have the same azimuthal orientation and are approximately the same size of about 80 nm, which is close to the grain size before etching.

4 18 D. Chen et al. / Microelectronics Journal 40 (2009) Fig. 6. The surface morphologies of the etching evolution of the AlN films deposited at 600 1C in 10% KOH solution at 40 1C. (a) Before etching; (b) 1 min; (c) 5 min (d) the high-magnification image. also noted that all the pyramids have the same azimuthal orientation and are approximately the same size of about 80 nm, which is close to the grain size before etching. The shape of the pyramid is maintained after prolonged etching time. The formation of the pyramids indicated that the etch rate of different planes is very different due to the strong anisotropy of the AlN wurtzite structure, which agrees with the results of Section The lateral etching Fig. 7 shows the cross-section micrographs of etching patterns for the samples deposited at RT and 600 1C. It is can be seen that the Ti mask and the W electrode show a strong resistance to the etching process. There are no obvious etching leavings on the surface after the AlN films were completely removed. A strong lateral etching was observed under the Ti mask for the film deposited at RT. The lateral etching distance is about 600 nm and a little smaller than the film thickness. For the film deposited at 600 1C, the lateral etching distance is only about 200 nm indicating that the anisotropy of the etching is improved with the improving of the crystal quality. Fig. 7. The cross-section micrographs of etching patterns for the sample deposited at (a) RT and (b) 600 1C. The samples were both exposed in 10 wt% KOH solution at 50 1C. Acknowledgments This work is supported by the National Basic Research Program of China No. 2006CB300406; Shanghai Science and Technology Grant No: 0752nm015; National Natural Science Foundation of China No The authors thank the Instrumental Analysis Center of Shanghai Jiao Tong University for the Materials Characterization. 4. Conclusions In summary, the chemical etching of AlN films is strongly dependent on the film texture. There is a faster etching in the case of the mixed (10 0) and (0 0 2) texture and a lower rate in the case of the only (0 0 2) texture. The etch rate also decreases with the crystal quality. The calculated activation energy is different for the films with varied texture, suggesting different mechanisms dominate the etching process. The etching surface of the sample with only the (0 0 2) texture form discontinuous column structure and exhibits lower porosity compared to that of the mixed (10 0) and (0 0 2) texture. Due to the strong anisotropy of the AlN wurtzite structure, the morphology of the film deposited at 600 1C shows the homogeneous pyramids shape after etching. The crosssection micrographs of etching patterns indicate that the anisotropy of the chemical etching is improved with the improving of the crystal quality. References [1] Y. Taniyasu, M. Kasu, T. Makimoto, An aluminium nitride light-emitting diode with a wavelength of 210 nm, Nature 441 (2006) [2] R. Dahal, T.M. Al Tahtamouni, J.Y. Lin, H.X. Jiang, AlN avalanche photodetectors, Appl. Phys. Lett. 91 (2007) [3] D. Zhuang, J.H. Edgar, Wet etching of GaN, AIN, and SiC: a review, Mater. Sci. Eng. R 48 (2005) [4] K. Tonisch, V. Cimalla, C. Foerster, H. Romanus, O. Ambacher, D. Dontsov, Piezoelectric properties of polycrystalline AlN thin films for MEMS application, Sens. Actuator A 132 (2006) [5] H.C. Lee, J.Y. Park, K.H. Lee, J.U. Bu, Preparation of highly textured Mo and AIN films using a Ti seed layer for integrated high-q film bulk acoustic resonators, J. Vac. Sci. Technol. B 22 (2004) [6] S.J. Pearton, et al., Dry etching of thin-film InN, AlN and GaN, Semicond. Sci. Technol. 8 (1993) [7] M. Bickermann, S. Schmidt, B.M. Epelbaum, P. Heimann, S. Nagata, A. Winnacker, Wet KOH etching of freestanding AlN single crystals, J. Cryst. Growth 300 (2007) [8] T.Y. Sheng, Z.Q. Yu, G.J. Collins, Disk hydrogen plasma assisted chemical vapor deposition of aluminum nitride, Appl. Phys. Lett. 52 (1988)

5 D. Chen et al. / Microelectronics Journal 40 (2009) [9] J.R. Mileham, S.J. Pearton, C.R. Abernathy, J.D. MacKenzie, R.J. Shul, S.P. Kilcoyne, Wet chemical etching of AlN, Appl. Phys. Lett. 67 (1995) [10] D. Zhuang, J.H. Edgar, B. Strojek, J. Chaudhuri, Z. Rek, Defect-selective etching of bulk AlN single crystals in molten KOH/NaOH eutectic alloy, J. Cryst. Growth 262 (2004) [11] A. Ababneh, H. Kreher, U. Schmid, Etching behaviour of sputter-deposited aluminium nitride thin films in H 3 PO 4 and KOH solutions, Microsyst. Technol. (2008) 1 7. [12] W.L. Jung, J.C. Jerome, B. Mohamed, Plasma characteristics in pulsed direct current reactive magnetron sputtering of aluminum nitride thin films, J. Vac. Sci. Technol. A 22 (2004) [13] S.A. Campbell, The Science and Engineering of Microekectronic Fabrication, third ed., Oxford University Press, New York, [14] C.B. Vartuli, et al., Wet chemical etching of AlN and InAlN in KOH solutions, J. Electrochem. Soc. 143 (1996)

Surface Micromachining Process for the Integration of AlN Piezoelectric Microstructures

Surface Micromachining Process for the Integration of AlN Piezoelectric Microstructures Saravanan. S, Erwin Berenschot, Gijs Krijnen and Miko Elwenspoek Transducers Science and Technology Laboratory University