ZnO Thin Films Synthesized by Chemical Vapor Deposition

Transcription

1 ZnO Thin Films Synthesized by Chemical Vapor Deposition Zhuo Chen *1, Kai Shum *2, T. Salagaj #3, Wei Zhang #4, and K. Strobl #5 * Physics Department, Brooklyn College of the City University of New York 2900 Bedford Avenue, Brooklyn, NY zhuochen@brooklyn.cuny.edu 2 kshum@brooklyn.cuny.edu # First Nano, a division of CVD Equipment Cooperation 1860 Smithtown Ave Ronkonkoma, NY tom.salagaj@firstnano.com 4 wei.zhang@firstnano.com 5 kstrobl@cvdequipment.com Abstract - This paper describes the experimental results of our recent attempt to synthesize device quality ZnO thin films on silicon and sapphire substrates by means of chemical vapor deposition. The surface features and crystal quality of these films are studied by scanning electron microscope and optical spectroscopy, respectively. Although it was not successful to deposit crystalline thin film on silicon substrate, high quality thin films on sapphire substrates were synthesized. I. INTRODUCTION Semiconductor industry continuously brings new inventions into our daily life. For example, the massproducible and high-quality blue-ultraviolet GaN-based light emitting diodes and lasers are used in blue-ray players, displays, and traffic signals. ZnO is an II VI semiconductor with a stable wurtzite structure at room temperature. It has a wide direct band gap at 300 K and a large exciton binding energy, theoretically ensuring even more efficient excitonic emission at room temperature than that from GaN counterpart. At present, the most widely publicized application for ZnO is as an ITO replacement for displays and photovoltaic panels, where ZnO could lower costs of transparent conductors. But new applications for ZnO are much broader than that.[1, 2] ZnO may also find applications in thin-film batteries and interesting ZnO nanostructures may be engineered for various new applications down the road. ZnO is already being studied for spintronics.[3, 4] The most significant challenge to transform ZnO to ZnOrelated technologies, especially for ZnO-based optoelectronic devices, is how to produce stable and reproducible p-type doping of ZnO.[5] One of prerequisites to overcome this difficulty is to grow high quality ZnO films with low enough residual donor density which compensates p-type doing. In attempt to grow these high quality ZnO films on suitable substrates, various methods such as molecular beam epitaxy (MBE) and metal-organic vapor-phase epitaxy (MOVPE) have been used.[6, 7] Although, these methods have been proved to be effective to gradually improve the crystal quality of GaN-based materials leading to its commercial applications, it has not been successful so far to produce reproducible and controllable doping for ZnO epitaxial-layers by these methods. In this paper, we present our preliminary work on the growth of ZnO epitaxial layers on two different types of substrates: ZnO-seeded silicon and sapphire substrates. The method of growth is the chemical vapor deposition (CVD) using a specially prepared solid source and a bettertemperature-controlled furnace. These ZnO epitaxial layers are then characterized by scanning electron microscopy (SEM) and optical spectroscopic methods such as absorption and photoluminescence (PL). For the ZnO growth on ZnO-seeded Si substrates, attentions were paid on how to prepare crystalline ZnO seedlayers. These seed-layers on Si substrates were prepared by depositing a thin layer of amorphous ZnO using an e-beam evaporator and then followed by a rapid thermal annealing (RTA) treatment. The growth of the thin layer of crystalline ZnO on ZnO-seed layer is desired with a suitable polycrystalline buffer layer. However, only various ZnO nanostructures (nanorods) were grown. Excellent optical characteristics of these nanostructures such as PL line width, linearity of PL intensity as a function of excitation power density were obtained. For the ZnO growth on sapphire substrates with different orientations, its nucleation mechanism was carefully studied as various parameters such as substrate temperature and O 2 flow rate. SEM images revealed that thin layers of ZnO were successfully grown in sapphire substrates. We also confirmed their excellent optical emission characteristics expected for exciton-related process. In the section of II, we will briefly present the basic properties of ZnO and compare them with GaN. The improved CVD growth apparatus is described in section III. The main experimental results will be given in section IV and V. Finally, a summary is given in section VI.

2 II. BASIC MATERIAL PROPERTIES A group of basic properties of ZnO is listed in Table 1 along with GaN s values for a comparison. There are two distinct properties of ZnO that have made this semiconductor very important. First, it has a direct band gap of 3.4 ev at 300 K and is possible to form high quality hetero-structures with the higher band gap alloy of ZnMgO and the lower band gap alloy of ZnCdO, enabling a range of optical devices to be engineered. Second, a large exciton binding energy of 60 mev was reported for ZnO. This large exciton binding energy can simply translate to much better optical quantum efficiency at room temperature, possibly leading much lower laser threshold. In addition, high quality single-crystal ZnO substrates[8] are more readily available than that for GaN. Table 1 Property comparison between ZnO and GaN at 300 K Property ZnO GaN Stable structure Wurtzite Wurtzite Lattice parameters (c 0/a 0 = for ideal hexagonal structure) a 0 = nm c 0 = nm c 0 /a 0 = a 0 = nm c 0 = nm c 0 /a 0 = Density (g cm 3 ) Melting 1975 ~ 2500 temperature ( o C) Thermal conductivity (W cm 1 K 1 ) Linear expansion coefficient(/ o C) a 0 : c 0 : a 0 : c 0 : Static dielectric constant Energy gap (ev) Electron effective mass (m 0) Electron mobility (cm 2 V 1 s 1 ) Hole effective mass (m 0) Hole mobility (cm 2 5 ~ 50 ~ 100 V 1 s 1 ) Optical phonon energy (mev) Exciton binding energy (mev) substrates. It was equipped with a separated solid source heater, a three-zone (load, center, and end zone) furnace, a gas injector, a vacuum pump, and a quartz tube. The reaction tube was controlled by a three-temperature-zone furnace to obtain a uniform temperature profile across the substrate at the collecting area over 3 inches by 2 inches. The source material was mixed ZnO powder (Alfa Aesar, 99.99%) and graphite powder (Alfa Aesar, 99%) with a mass ratio of 1:4. The solid source was placed at the load zone in the quartz tube and heated up to high temperature by an additional solid source heater to generate Zn vapor which was then carried into the center zone by the Ar carrying gas. The reacting gas (O 2 ) was introduced into the system by an independent gas injector at different locations to achieve the best uniformity over a given substrate size. Silicon or sapphire substrates with various planar cuts were used. Fig. 1a shows the schematic diagram for the CVD apparatus. Fig 1b displays typically measured temperature profiles across load to center zone. Two important parameters, growth (substrate) temperature (T g ) and O 2 flow rate (F O2 ) will be discussed in this work for the optimal ZnO deposition. The surface morphology of deposited thin films was studied by scanning electron microscopy (SEM) and atomic force microscopy (AFM). Optical transmission/absorption spectra were measured from PerkinElmer Lambda-950 optical spectrometer. Photoluminescence (PL) spectra were measured by Horiba NanoLog system coupled with an optical cryostat (capable of ~ 8 to 350 K) from Advanced Research System. Fig. 1 (a) Schematic of CVD apparatus. III. EXPERIMENTAL METHODS A chemical vapor deposition (CVD) system (ET2000) was built by First Nano, a division of CVD Equipment Corporation to deposit thin films on silicon and sapphire Fig. 1 (b) Temperature profiles obtainable during growth.

3 IV. EPITAXIAL DEPOSITION ON ZNO SEEDED SILICON SUBSTRATES In the first phase of this project, the objective was to epitaxially grow thin layers of ZnO on the silicon wafers with thin ZnO seed-layers. A thin ZnO seed-layer was obtained by first depositing a thin layer of amorphous ZnO and then a rapid thermal annealing (RTA) process was used to transform this thin amorphous layer to a thin polycrystalline ZnO seedlayer. ZnO seeded silicon wafers were loaded to the CVD furnace to grow ZnO films with various conditions. For the conditions we experimented, we were not able to grow thin films on the ZnO seeded silicon wafers. Instead, high optical quality of ZnO nanorods were grown on these seeded silicon wafers. Fig. 2a, 2b, and 2c display the SEM images of the amorphous ZnO, RTA-ZnO, and epitaxial nanorods produced with our method. For the amorphous film, the granular particles had a size distribution of 20 to 50 nm. After a thermal annealing treatment, ZnO crystallites appeared with a well defined orientation. The crystallites were served as seeds on which nanorods were epitaxially grown on as shown in Fig. 2c. shows that there are two layers of nanostructures. The first layer is directly grown on the seeds with different sizes acting as a buffer layer. The second layer consists of well defined more uniform nanorods with height of 200 to 300 nm. The temperature dependence of PL spectra is displayed in Fig. 4. These spectra were taken from the ZnO nanorods epitaxially grown on a pre-seeded Si (100) substrate. It clearly demonstrates the high crystalline quality of the nanoepitaxially grown ZnO nanorods from the disappearance of strong donor-related exciton emission peaks and the weak green emission around 500 nm. The evidence of growing zinc blende phase of ZnO was recently reported by Zhuo et al. [9] and reviewed by Ashrafi and Jagadish. [10] Based on the photoluminescence peak position of 378 nm for ZnO nanorods/needles reported in this work, it is consistent that these nanostructures have a wurtzite phase, the energetically allowed phase for bulk ZnO at room temperature. Fig. 3 The cross section SEM image of nanorod layer. (a) (b) (c) Fig. 2 (a) SEM of Amorphous ZnO surface on Si (100), (b) SEM of ZnO surface after RTA, and (c) SEM of Epitaxial vertically aligned ZnO nanorods. Fig. 3 shows a cross section SEM of the cleaved sample edge for the vertically aligned nanorods structure grown epitaxially on Si. A careful analysis of these substantially vertical aligned ZnO nanorod samples with SEM images Fig. 4 Temperature dependence of PL spectra of ZnO nanorods. Fig. 5b displays the excitation power dependence of PL spectra obtained from another ZnO nanorods system using our deposition method with one of individual nanorod (needle) SEM image shown in Fig. 5a. PL peak intensity is plotted as a function of excitation power in Fig. 6. It clearly shows the expected linearity (over the four orders of magnitude of

4 excitation power) for high efficiency exciton emission in these high quality ZnO nanorods. In summary, for this phase of work, vertically aligned ZnO nanorods were grown using our deposition method. SEM images revealed that the resulting ZnO nanorods were singlecrystalline and grown along the c axis. Furthermore, we confirmed their excellent optical emission characteristics expected exciton-related process. deposition time. It is clear that at 720 o C ZnO starts to nucleate on the sapphire surface, but the growth rate of ZnO is too slow, leading to the scattered ZnO islands on the sapphire surface during 1 hour time period. At 820 o C, the speed of nucleation has dramatically increased and left a thin layer of ~ 200 nm during 1 hour time period. Since the ZnO coverage on the sapphire surface is not 100%, the thickness of ZnO islands was measured by AFM. At 920 o C, no visible ZnO film was deposited on the substrate and verified by photoluminescence intensity as shown in Fig. 8 (a). Fig. 5 (a) SEM image of a single ZnO nanorod (needle) of ~ 1.4 μm long and 100 nm radius at the bottom. (b) Excitation power dependence of PL spectra from ZnO nanorods network (the individual nanorods are similar to what is shown on the left) at 8.2 K. Fig. 6 The PL peak intensity as a function of excitation power at 8.2 K for the high optical quality of ZnO nanorods. V. ZNO THIN FILMS DEPOSITED ON SAPPHIRE SUBSTRATES In the second phase of this research project, the objective was to grow device quality thin ZnO layers on sapphire substrates. To better understand how to reach this goal we first investigated the ZnO nucleation process on sapphire surfaces and succeeded in growing high quality thin films. However, the growth rate achieved with our method was very limited. More extensive work is required to find the process parameters which result in a higher and commercially interesting growth rate. Fig. 7 (a), (b), and (c) display the SEM surface images for the 3 samples placed in the center zone with different growth temperature of 720, 820, and 920 o C, respectively, for 1 hour Fig. 7 (a) ZnO deposited on sapphire substrates with growth temperatures at 720 o C, (b) at 820 o C, and (c) at 920 o C Photoluminescence signals were acquired from these 3 samples at room temperature with a Xe-lamp as a photoexcitation source having a 5 nm bandwidth at 300 nm. They are shown in Fig. 8 (a) as green, blue, and red curves for the samples with the growth temperature of 720, 820, and 920 o C, respectively. All three spectra show a band edge emission at ~378 nm with negligible green emission at 550 nm, normally attributed to defect-related emission. PL from the film with growth temperature of 820 o C is certainly strongest among the three samples, consistent with the information provided by their SEM images. The origin of three weak PL peaks from 400 to 500 nm for the 720 o C sample is not clear.

5 In Fig. 8 (b), a Tauc plot [11] is displayed for the 820 o C sample. The optical band gap of 3.30 ev is obtained. This energy gap is consistent with other ZnO thin films obtained by other methods [12, 13] and with generally accepted ZnO band gap [1, 2] at room temperature by taking into account the excitonic effect on absorption edge. intensity increases as the O 2 flow rate increases. These films are clearly of high quality with a low optically active defect density as evidenced by the lack of the so-called green emission around 550 nm. Fig. 8 (a) PL spectra from 3 samples with different growth temperatures are shown, and (b) Tauc plot to obtain the optical band gap for the 820 o C sample. At the preliminary optimal growth temperature of 820 o C on a sapphire substrate, three ZnO films were deposited with different O 2 flow rates at 100, 500, and 1,000 standard cubic center meter per minute (with a short notation of sccm). The SEM images for these three samples are shown in Fig. 9 (a), (b), and (c), respectively. For the 100 sccm sample, it is clearly visible that there is a large amount of surface defects (pits). The estimated number of pit density seems to be on the order of 10 7 cm -2, comparable with these thin films deposited by more expensive MOCVD apparatus.[7] Such defect density decreases as the O 2 flow rate increases as demonstrated by the 500 and 1000 sccm samples. PL spectra for the three samples with different O 2 flow rates are displayed in Fig. 10. The PL peak intensity seems to be consistent with the pit density as obtained from SEM images. Although the data points are limited, apparently, the PL intensity (I PL ) has a square root relationship with the O 2 flow rate, i.e., I PL ~ F O2 1/2. In this part of work, we have presented our preliminary data on ZnO thin films on sapphire substrates synthesized by CVD method using a solid source. We have found that the optimal growth temperature is close to 820 o C and the film surface pit density decreases while photoluminescence Fig. 9 ZnO film on sapphire substrate with O2 flow rate of (a) 100 sccm, (b) 500 sccm, and (c) 1,000 sccm. Fig. 10 PL spectra from the three samples with the different O 2 flow rate as indicated in the figure.

6 VI. SUMMARY We have carried out a study on how high quality ZnO films can be synthesized on silicon and sapphire substrates by the CVD method with a solid source. Although we were not successful in depositing uniform high quality ZnO films on silicon substrates, various high crystal quality nanostructures were epitaxially grown on ZnO pre-seeded silicon substrates. We have clearly demonstrated that high quality ZnO films can be grown on sapphire substrates using our method. The results obtained in this preliminary work encourage a more extensive study to enhance the growth rate. [11] J. Tauc, Amorphous and Liquid Semiconductors (Plenum, London, 1974). [12] S. T. Tan et al., Blueshift of optical gap in ZnO thin film grown by metal-organic chemical-vapor deposition, J. Appl. Phys., Vol. 98, pp , [13] V. Cracium et al., Characteristics of high quality ZnO thin films deposited by pulsed laser deposition, Appl. Phys. Lett., Vol. 65, pp , ACKNOWLEDGMENTS The authors would like to acknowledge the partial financial support from a NYSTAR grant for this work through the CUNY center of advanced technology (CAT) on photonic applications and thank L. Rosenbaum for useful discussions. Participation by C. Jensen, Y. Hu and Mim Nakarmi at the early stage of this research project is appreciated. REFERENCES [1] S. J. Pearton, D. P. Nortona, K. Ipa, Y. W. Heoa, and T. Steinerb, Recent progress in processing and properties of ZnO, Superlattices and Microstructures, vol. 34 pp. 3 32, [2] Ü. Özgür, Ya. I. Alivov, C. Liu, A. Teke, M. A. Reshchikov, S. Doğan, V. Avrutin, S.-J. Cho, and H. Morkoç, A comprehensive review of ZnO materials and devices, J. Appl. Phys. Vol. 98, pp , [3] Hyeon-Jun Lee, Se Young Jeong, Chae Ryong Cho, and Chul Hong Park, Study of diluted magnetic semiconductor co-doped ZnO, Appl. Phys. Lett. Vol. 81 pp , [4] A. C. Tuan, Epitaxial growth and properties of cobaltdoped ZnO on alpha-al2o3 single-crystal substrates, Phys. Rev. B, Vol. 70, pp , [5] J. L. Lyons, A. Janotti, and C. G. Van de Walle, Why nitrogen cannot lead to p-type conductivity in ZnO, Appl. Phys. Lett., Vol. 95, pp , [6] Z. K. Tang et al. Room-temperature ultraviolet laser emission from self-assembled ZnO microcrystalline thin films, Appl. Phys. Lett. Vol. 72, pp , [7] N. Oleynik, MOCVD growth and characterization of ZnO properties for optoelectronic application, Ph. D. Thesis, University of Magdeburg, Germany, July [8] V. Avrutin, J. Z. Zhang, J. J. Song, D. Silverman, and H. Morkoc, Bulk ZnO: current status, challenges, and prospect, IEEE Electron Device [9] S. M. Zhuo et al., Nanotechnology Vol. 19, pp , [10] A. Ashrafi and C. Jagadish, Review of zinc-blende ZnO: Stability of meta-stable ZnO phases, J. Appl. Phys., Vol. 102, pp , 2007.