Zinc Oxide (ZnO)
Introduction
The substance ZnO occurs in nature as the rare mineral zincite. Its crystals are typically needles with hexagonal cross section. The color varies mainly from yellow to red, which is the result of iron or manganese impurities. Zincite is sometimes formed also in zinc smelters, because liquid zinc (melting point 419.6°C) evaporates easily (boiling point 908°C under atmospheric pressure) and reacts with oxygen if air is introduced through leaks. Such "artificial zincite" needles are sometimes more than 5 cm long.
ZnO is one of the firstly discovered semiconductors that belongs to a class of materials called transparent semiconducting oxides (TSOs). Semiconducting properties of ZnO were revealed early, and zincite crystals were among the active materials for the first "crystal radios" in the early 20th century. Varistors (voltage-dependent resistors) were developed in the 1960s and are still in use. These are ceramic components made from polycrystalline ZnO (intrinsically n-type), which is doped with additives such as Bi203 (intrinsically p-type). This way diodes are formed between the n-type grains and the p-type interfaces, and the varistor becomes a short circuit above some critical threshold voltage, allowing to save sensitive electronic parts from overvoltage.
Transparent Semiconducting Oxides: Bulk Crystal Growth and Fundamental Properties Zbigniew Galazka
Copyright © 2020 Jenny Stanford Publishing Pte. Ltd.
ISBN 978-981-4800-94-5 (Hardcover), 978-1-003-04520-5 (eBook) www.jennystanford.com
Certainly, ZnO is the TCO material with the largest annual world production, of about 10s tones [1]. The largest share of this huge amount is used for the vulcanization of rubber. Other significant quantities are used in medicine and cosmetics. ZnO is even added to animal food because it raises the growth rate of farm animals.
In addition to semiconducting properties (both n-type and p-type) and wide bandgap (3.3 eV), ZnO also shows piezoelectric [2- 4], pyroelectric [5, 6], ferromagnetic [7-9], and scintillation [10,11] properties. Such a variety of properties puts ZnO among other TSOs in the frontline in material science, growth technology, and device fabrication of different functionality. Indeed, following varistors [12], also other sensing, electronic, and optoelectronic devices were demonstrated, such as surface acoustic wave devices [13], pyroelectric sensors [14, 15], scintillators [16-18], diode devices [19-23], thin-film transistors [24, 25], field-effect transistors [26- 29], UV photodetectors [30-32], and light-emitting devices [25-36].
ZnO single crystals have no significant technical relevance so far; nevertheless, they are widely used for basic research and device demonstration. Single crystals are grown with a large variety of techniques, including (i) growth from the gas phase either by direct evaporation and precipitation of ZnO, by evaporation of Zn and subsequent oxidation to ZnO (like mentioned above in zinc smelters], or by chemical transport; (ii) grown from melt solutions consisting of different oxides and/or alkaline hydroxides; (hi) growth from aqueous or ammonia solutions, often under hydrothermal conditions, and (iv) grown from the melt either by indirect heating in an iridium crucible or by direct heating in the cold crucible technique.
In parallel to bulk single crystals, ZnO was also intensively grown as thin films and nanostructures either as pure ZnO or doped with a wide spectrum of elements. For example, thin films were grown by a diversity of techniques, including sol-gel [37-41], spray pyrolysis [42-44], sputtering [45-50], pulsed laser deposition (PLD) [51-56], chemical vapor deposition (CVD) [57-59], metal-organic chemical vapor deposition (MOCVD) [60-63], molecular beam epitaxy (MBE) [64-67], and many others.
An intensive research on ZnO is granted by a number of review papers, book chapters, and books, e.g. by Triboulet [68], Triboulet and Perriere [69], Pearton et al. [70], Ehrentraut et al. [71], Ozgiir et al. [72], C. Klingshirn [1, 73], Klingshrin et al. [74], Janotti and Van de Walle [75], Rodnyi and Khodyuk [76], Klimm et al. [77], Teherani and Litton (editors] [78], Reynolds and Reynolds [79], and Feng (editor] [80].
Crystal Structure
The crystal structure is among all TCO structures most closely related to that of the important semiconductors silicon and gallium arsenide: Like these, it consists crystals of stacked tetrahedra (Fig. 3.1]. For Si (diamond structure] and GaAs (sphalerite structure], however, the tetrahedra have a stacking sequence A-B-C-A- B - C, which results in a cubic symmetry having three equivalent axes. For ZnO, the larger radius difference of its constituents (Zn2+:
74 pm; 124 pm [81]] in conjunction with larger ionicity leads the alternative stacking sequence A - В - A - B, which is called wurtzite structure, results in hexagonal symmetry with space group P63mc.
The energetic difference between both stackings is rather small, and it can be regarded as curiosity that the eponym minerals for both structures, sphalerite (= zincblende] and wurtzite, are both chemically ZnS. As a result of the similar lattice energy, stacking faults are often found for sphalerite- as well as wurtzite-type crystals, and also for ZnO [82].
Figure 3.1 The wurtzite crystal structure of ZnO consists ofA-B-A-B stacked tetrahedra.
From the lattice parameters a = 3.253(1) A, c = 5.213(1) A [83] follows an axis ratio c/a = 1.603, which represents a significant deviation from an ideal wurtzite structure. Such ideal wurtzite structure is built with perfect tetrahedra and has c/a = (8/3)1/2 = 1.633. In wurtzite ZnO, the Zn-0 bonds consequently have no identical length, because the tetrahedra are slightly compressed along the c-axis [0001] [84], and become disphenoids. The 63 screw axis is polar, and consequently ZnO is a piezoelectric and pyroelectric material.
Typical ZnO substrate orientations are “C-plane”: (0001) or (0001); "A-plane”: (112 0); "M-plane”: (1100). Of these, only M-plane is not a polar face. Under high pressure p > 6 GPa ZnO transforms from wurtzite type (coordination number 4) to rocksalt type (coordination number 6) [85]. The stability of rocksalt-type ZnO can be extended to lower pressures by Mg doping, which is possible up to <7% in the bulk [86], or by epitaxial growth of Znx_xMgxO epilayers [87].
Band Structure and Native Point Defects
Band Structure
ZnO is one of the most investigated TSO materials, including study of the electronic structure thereof. Intensive theoretical and experimental studies of the electronic structure are reflected in a number of reports for pure ZnO and doped with a diversity of elements (H, Li, Na, K, Au, Ag, Cu, Mn, Fe, Co, Ni, Mg, B, Al, Ga, Sc, Ti, V, Cr, Sr, Si, Sn, Zr, N, P, As, La, Ce, Pr, Nd, Eu), e.g., by Schroer et al. [88], Mi et al. [89], Ni et al. [90], Lambrecht et al. [91], Chien et al. [92], Hu and Pan [93], Janotti and Van de Walle [75, 94], Lyons et al. [95], Lee et al. [96], Lee and Chang [97, 98] Erhart et al. [99], Duan et al. [100], Berrezoug et al. [101], Tamura and Ozaki [102], Xiong et al. [103], Petit et al. [104], Karazhanov and Ulyashin [105], Guo et al. [106], Raebiger et al. [107], Schleife et al. [108], Dixit et al. [109], Chowdhury et al. [110], Han et al. [Ill], Deng et al. [112], Qu et al. [113, 114], Liu et al. [115], Zhang et al. [116], Sharma et al. [117], Lee et al. [118], Zhang et al. [119], Wang et al. [120], Peng et al. [121], Lorke et al. [122], Chen and Wu [123], Mahmood et al.
[124], Ataide et al. [125], Hammi et al. [126], Bashyal et al. [127].
The dopants aim to affect structural, electrical, optical, and magnetic (ferromagnetism) properties. Experimental valence-band structure with the use of angle-resolved photoemission spectroscopy (ARPES) and X-ray photoemission spectroscopy (XPS) was reported, e.g., by Sawada et al. [128], Ozawa et al. [129,130], Preston etal. [131], and King et al. [132].
The calculated band structure of ZnO is depicted in Fig. 3.2 (Janotti and Van de Walle [75], DFT, hybrid HSE). At the Г point (momentum к = 0) of the Brillouin zone (BZ), there is a global minimum and maximum in the conduction and valence band (CB and VB), respectively, indicating a direct bandgap. The lowest part of the CB is formed by empty 4s states ofZn2+ or the antibonding sp3 hybrid states, while the top part of the VB originates from the occupied 2p orbitals of O2' or from the bonding sp3 orbitals (Klingshrin [73]).
Figure 3.2 Calculated band structure of ZnO using the HSE hybrid functional. The energy of the valence-band maximum (VBM) was set to zero. Reprinted with permission from Ref. [95], Copyright 2009, American Physical Society.
Theoretical values of the bandgap spread in a wide spectrum depending on the calculation method, and often they are underestimated. Exemplary values of a theoretical direct bandgap (at Г - Г) of ZnO, which is close to experimental values, are 3.21 (DFT, hybrid HSE03 + GW) [108], 3.27 (DFT, GGA-PBE) [101], 3.4 (DFT + U, hybrid-HSE [127], and 3.49 (DFT, LDA+A-1/2) [125] eV.
A computed effective electron mass in ZnO was found to be almost isotropic, with a value of around m*/m0 = 0.28 [73], 0.27 [120], and 0.30 [108], where m0 is the rest electron mass. The effective hole mass was also found to be relatively isotropic, with a value of about mh*/m0h=O.S9 [73].
Native Point Defects
Native point defects in ZnO were also a subj ect of intensive theoretical studies, with some of them been experimentally confirmed. Examples of theoretical investigation of the native defects can be found in the works of Erhart et al. [99], Hu and Pan [93], Janotti and Van de Walle [75,133,134], and Sun and Wang [135].
Following Joanotti and Van de Walle [75, 133, 134], the native point defects in ZnO may include oxygen vacancies (V® and К(, +), oxygen interstitials (0° and Of-), oxygen antisites (0Zn), zinc interstitials (Znf+), zinc vacancies (V|~), and zinc antisites (Zn0). At early times of ZnO development, some of the native point defects, such as oxygen vacancies and zinc interstitials, were ascribed to я-type conductivity. However, oxygen vacancies (V0) were found to be deep donors, have high formation energy, and can compensate p-type doping. Oxygen interstitials (OJ can exist as electrically inactive interstitials or as deep acceptors. Oxygen antisites (0Zn] are deep acceptors with the highest formation energy among all native point defects. Zinc interstitials (ZnJ are, indeed, shallow donors but have high formation energy and are unstable as isolated point defects. Also zinc antisites (Zn0) are shallow donors with a high formation energy. Zinc vacancies (VZn) are deep acceptors with low formation energy. They can compensate the n-type conductivity. Migration barriers of the native point defects span between 0.57 and 2.36 eV and can be annealed out at low and moderate temperatures of 219-909 K. As the result, the underlying n-type conductivity of ZnO originates rather from extrinsic defects, such as impurities and/or complexes of impurities and native point defects (for more details, see Section 3.4).
One particular extrinsic point defect that may induce the electrical conductivity in ZnO is hydrogen, as theoretically described, e.g., by Van de Walle [136, 137], Janotti and Van de Walle [75, 94, 133], Karazhanovand Ulyashin [105], Wardle et al. [138,139], and Du and
Biswas [140]. The presence of hydrogen in any preparation process of ZnO is ubiquitous, and its incorporation cannot be avoided. Hydrogen was found to be a shallow donor when in an interstitial position (Hj+) [75, 137, 140], substituting oxygen (H0+] [75, 94], as a molecule H2 in the oxygen vacancy [140], or in complexes with 3d transition metals [139]. In particular, substitutional hydrogen (H0+) forms a multicenter configuration that is sensitive to the oxygen partial pressure [94].
