Zinc Oxide

The purpose of this introduction is – after a few general words on ZnO – to inform the reader about the history of ZnO research, the contents of this book and the intentions of the authors. Zinc oxide (ZnO) is a II^b–VI compound semiconductor. This group comprises the binary compounds of Zn, Cd and Hg with O, S, Se, Te and their ternary and quaternary alloys. The band gaps of these compounds cover the whole band gap range from E _g ≈ 3. 94 eV for hexagonal ZnS down to semimetals (i.e., E _g = 0 eV) for most of the mercury compounds. ZnO itself is also a wide gap semiconductor with E _g ≈ 3. 436 eV at T = 0 K and (3. 37 ± 0. 01) eV at room temperature. For more details on the band structure, see Chaps. 4 and 6 or for a recent collection of data on ZnO, for example, [Rössler et al. (eds) Landolt-Börnstein, New Series, Group III, Vols. 17 B, 22, and 41B, 1999]. Like most of the compounds of groups IV, III–V, II^b–VI and I^b–VII, ZnO shows a tetrahedral coordination. In contrast to several other II^b–VI compounds, which occur both in the hexagonal wurtzite and the cubic zinc blende type structure such as ZnS, which gave the name to these two modifications, ZnO occurs almost exclusively in the wurtzite type structure. It has a relatively strong ionic binding (see Chap. 2). The exciton binding energy in ZnO is 60 meV [Thomas, J. Phys. Chem. Solids 15:86, 1960], the largest among the II^b–VI compounds, but by far not the largest for all semiconductors since, for example, CuCl and CuO have exciton binding energies around 190 and 150 meV, respectively. See, for example, [Rössler et al. (eds) Landolt-Börnstein, New Series, Group III, Vols. 17B, 22, and 41B, 1999; Thomas, J. Phys. Chem. Solids 15:86, 1960; Klingshirn and Haug, Phy. Rep. 70:315, 1981; Hönerlage et al., Phys. Rep. 124:161, 1985] and references therein. More details on excitons will be given in Chap. 6. ZnO has a density of about 5. 6 g ∕ cm³ corresponding to 4. 2 × 10²² ZnO molecules per cm³ [Hallwig and Mollwo, Verhandl. DPG (VI) 10, HL37, 1975]. ZnO occurs naturally under the name zinkit. Owing to the incorporation of impurity atoms such as Mn or Fe, zinkit looks usually yellow to red. Pure, synthetic ZnO is colourless and clear in agreement to the gap in the near UV. The growth of ZnO and ZnO-based nano-structures is treated in Chap. 3. ZnO is used by several 100,000 tons per year, for example, as additive to concrete or to the rubber of tires of cars. In smaller quantities, it is used in pharmaceutical industries, as an additive to human and animal food, as a material for sensors and for varistors or as transparent conducting oxide. For more details and aspects of present and forthcoming applications, see Chap. 13.

This chapter starts with an overview of the ZnO crystal structure and its conjunction to the chemical binding. ZnO commonly occurs in the wurtzite structure. This fact is closely related to its tetrahedral bond symmetry and its prominent bond polarity. The main part of the first section deals with the ZnO wurtzite crystal lattice, its symmetry properties, and its geometrical parameters. Besides wurtzite ZnO, the other polytypes, zinc-blende and rocksalt ZnO are also briefly discussed. Subsequently, lattice constant variations and crystal lattice deformations are treated. This discussion starts with static lattice constant variations, induced by temperature or by pressure, as well as strain-induced static lattice deformation, which reduces the crystal symmetry. The impact of this symmetry reduction on the electrical polarization is the piezo effect, which is very much pronounced in ZnO and is exploited in many applications. See also Chap. 13. Dynamic lattice deformations manifest themselves as phonons and, in case of doping, as phonon–plasmon mixed states. The section devoted to phonons starts with a consideration of the vibration eigenmodes and their dispersion curves. Special attention is paid to the investigation of phonons by optical spectroscopy. The methods applied for this purpose are infrared spectroscopy and, more often, Raman spectroscopy. The latter method is very common for the structural quality assessment of ZnO bulk crystals and layers; it is also frequently used for the study of the incorporation of dopant and alloying atoms in the ZnO crystal lattice. Thus, it plays an important role with regard to possible optoelectronics and spintronics applications of ZnO. The final section of this chapter focuses on phonon–plasmon mixed states. These eigenstates occur in doped ZnO due to the strong coupling between collective free-carrier oscillations and lattice vibrations, which occurs due to the high bond polarity. Owing to the direct correlation of the plasmon–phonon modes to the electronic doping, they are an inherent property of ZnO samples, when applied in (opto-) electronics and spintronics. See also Chap. 12.

This chapter is devoted to the growth of ZnO. It starts with various techniques to grow bulk samples and presents in some detail the growth of epitaxial layers by metal organic chemical vapor deposition (MOCVD), molecular beam epitaxy (MBE), and pulsed laser deposition (PLD). The last section is devoted to the growth of nanorods. Some properties of the resulting samples are also presented. If a comparison between GaN and ZnO is made, very often the huge variety of different growth techniques available to fabricate ZnO is said to be an advantage of this material system. Indeed, growth techniques range from low cost wet chemical growth at almost room temperature to high quality MOCVD growth at temperatures above 1, 000^∘C. In most cases, there is a very strong tendency of c-axis oriented growth, with a much higher growth rate in c-direction as compared to other crystal directions. This often leads to columnar structures, even at relatively low temperatures. However, it is, in general, not straight forward to fabricate smooth ZnO thin films with flat surfaces. Another advantage of a potential ZnO technology is said to be the possibility to grow thin films homoepitaxially on ZnO substrates. ZnO substrates are mostly fabricated by vapor phase transport (VPT) or hydrothermal growth. These techniques are enabling high volume manufacturing at reasonable cost, at least in principle. The availability of homoepitaxial substrates should be beneficial to the development of ZnO technology and devices and is in contrast to the situation of GaN. However, even though a number of companies are developing ZnO substrates, only recently good quality substrates have been demonstrated. However, these substrates are not yet widely available. Still, the situation concerning ZnO substrates seems to be far from low-cost, high-volume production. The fabrication of dense, single crystal thin films is, in general, surprisingly difficult, even when ZnO is grown on a ZnO substrate. However, molecular beam epitaxy (MBE) delivers high quality ZnMgO–ZnO quantum well structures. Other thin film techniques such as PLD or MOCVD are also widely used. The main problem at present is to consistently achieve reliable p-type doping. For this topic, see also Chap. 5. In the past years, there have been numerous publications on p-type doping of ZnO, as well as ZnO p–n junctions and light emitting diodes (LEDs). However, a lot of these reports are in one way or the other inconsistent or at least incomplete. It is quite clear from optical data that once a reliable hole injection can be achieved, high brightness ZnO LEDs should be possible. In contrast to that expectation, none of the LEDs reported so far shows efficient light emission, as would be expected from a reasonable quality ZnO-based LED. See also Chap. 13. As a matter of fact, there seems to be no generally accepted and reliable technique for p-type doping available at present. The reason for this is the unfavorable position of the band structure of ZnO relative to the vacuum level, with a very low lying valence band. See also Fig. 5.1. This makes the incorporation of electrically active acceptors difficult. Another difficulty is the huge defect density in ZnO. There are many indications that defects play a major role in transport and doping. In order to solve the doping problem, it is generally accepted that the quality of the ZnO material grown by the various techniques needs to be improved. Therefore, the optimization of ZnO epitaxy is thought to play a key role in the further development of this material system. Besides being used as an active material in optoelectronic devices, ZnO plays a major role as transparent contact material in thin film solar cells. Polycrystalline, heavily n-type doped ZnO is used for this, combining a high electrical conductivity with a good optical transparency. In this case, ZnO thin films are fabricated by large area growth techniques such as sputtering. For this and other applications, see also Chap. 13.

This chapter deals with the ordering of the valence bands – a topic that has controversially been discussed for more than 40 years. The Γ ₇, Γ ₉, Γ ₇ ordering is discussed in the light of very recent ab initio band structure calculations, and the important role is emphasized, which the Zn 3d-band position plays to the sign of the spin–orbit splitting. This topic is touched again from a different point of view in Chap. 6 on free excitons. Then we summarize the experimental findings on the cationic and anionic substitutions in ZnO and random alloy formation essential for quantum hetero-structures. The chapter closes with the data on the valence and conduction band discontinuities in iso- and hetero-valent hetero-structures.

In this chapter, the electrical properties of ZnO are discussed, which essentially include doping, carrier mobility, contacts, and some other topics listed below. Nominally undoped ZnO is always n-type. This fact could be possibly due to intrinsic defects or due to hydrogen, which is a donor in ZnO and a rather ubiquitous element (also in most of the growth processes). Therefore, hydrogen in ZnO is treated first and then other donors for efficient n-type doping. The next section is devoted to p-type doping and the persistent difficulties to obtain reliable, stable, and high p-type conductivity. The chapter continues with information on the carrier mobility and the observation of the integer quantum Hall effect. The next electric properties concern the use of ZnO in varistors and high-field transport. The final aspect is photoconductivity.

In this chapter, we review the intrinsic linear optical properties of ZnO close to the fundamental absorption edge. This comprises band-to-band transitions and free excitons and polaritons in bulk samples and epitaxial layers; free and localized excitons and polaritons in quantum wells and wires, including nanorods; also localized excitons in alloys and in quantum dots (or nano crystals) and finally cavity polaritons. By the term “free excitons”, we mean the quanta of the intrinsic electronic excitation in semiconductors (and insulators), which can move freely through the sample and which are described by a plane wave factor exp(i Kr) in d dimensions (d = 3, 2 or 1), where K is the wave vector of the centre of mass motion described by r, multiplied by the envelope function of the relative (hydrogen-like) motion of electron and hole around their common centre of gravity. By the terms “bound exciton complexes” or “bound excitons” [(BEC) and (BE), respectively], we understand excitons that are bound to some centres like neutral or ionized donors or neutral acceptors but also to more complex centres. They will be treated in Chap. 7. In contrast, by the term “localized excitons”, we mean electron–hole pairs, which are localized by disorder like intrinsic alloy disorder, for example, in Mg_1−xZn _x O and/or fluctuations of well (or wire) width in quantum structures. These phenomena are inherent to alloys and to structures of reduced dimensionality and are therefore included in this chapter. The influence of external fields on both free and bound excitons is then covered in Chap. 8.

In the preceding chapter, we concentrated on the properties of free excitons. These free excitons may move through the sample and hit a trap, a nonradiative or a radiative recombination center. At low temperatures, the latter case gives rise to either deep center luminescence, mentioned in Sect. 7.1 and discussed in detail in Chap. 9, or to the luminescence of bound exciton complexes (BE or BEC). The chapter continues with the most prominent of these BECs, namely A-excitons bound to neutral donors. The next aspects are the more weakly BEs at ionized donors. The Sect. 7.4 treats the binding or localization energies of BEC from a theoretical point of view, while Sect. 7.5 is dedicated to excited states of BECs, which contain either holes from deeper valence bands or an envelope function with higher quantum numbers. The last section is devoted to donor–acceptor pair transitions. There is no section devoted specifically to excitons bound to neutral acceptors, because this topic is still partly controversially discussed. Instead, information on these A⁰X complexes is scattered over the whole chapter, however, with some special emphasis seen in Sects. 7.1, 7.4, and 7.5.

The application of external fields provides a powerful tool to investigate a large variety of properties of excitons and exciton related processes. Within this chapter, we focus on the fundamental effects of static magnetic and strain fields on the optical properties of excitons in ZnO. The description is complemented by relevant examples. A general review of this topic can be found for constant fields in [Cho, Excitons, Topics in Current Physics, vol. 14 (Springer, Heidelberg, 1979)] and Hönerlage et al. [Phys. Rep. 124:161, 1985] and for modulation techniques in [Cardona, Modulation Spectroscopy (Academic, New York, 1969); Seraphin, Modulation Spectroscopy (North Holland, Amsterdam, 1973); Goldsmith, NATO Science Series II, Frontiers of Optical Spectroscopy, vol. 168 (Springer Netherlands, 2005)]. Not much has been published on the influence of static electric fields on excitons. A few references are given at the end of Sect. 8.2.

In ZnO, deep centres such as the transition metal ions Cu, Fe, Co, etc., deep acceptors such as Li and Na as well as intrinsic defects such as the cation and anion vacancies are the origin of light emissions in various regions of the visible and infrared spectral range. One example of luminescence in the visible spectral range was already presented in Fig. 7.1. In Fig. 9.1 other examples of deep centre related emissions are given, which are often found in bulk and epitaxially grown ZnO or in nano rods. The presentation of optical properties of various deep centres is complemented by data from electron paramagnetic resonance (EPR). It proceeds from the green and yellow emission bands and their interpretation to transition metals. From the intrinsic defects, the oxygen vacancy is discussed since many years as one of the origins of a broad emission band in the green spectral range. Other intrinsic defects are frequently discussed in recent literature, but the discussion is still rather controversial and the experimental findings on which the various assignments are based are frequently much less elaborate compared, for example, to transition metals. Therefore, no individual section is devoted to these complexes but some information is given, for example, at the end of Sect 5.1.

The investigation of the magnetic properties of ZnO doped with “magnetic” ions such as Mn, Co, Fe, or Ni is still very controversially discussed. Therefore, we give here, after a general introduction, a short overview of the present situation only.

Nonlinear optics is a wide field, comprising in principle all phenomena, where one observes a nonlinear response (e.g. of the polarization of a medium) on the stimulus such as the incident electromagnetic field. It starts with the typical effects such as second or third harmonic generation, four wave mixing or the rectification of the electromagnetic field (the so-called dc-effect) and continues over two and multi photon absorption to the phenomena of extreme nonlinear optics. Another branch deals with the phenomena in dense electron–hole pair systems (so-called high exciton, high density or many particle effects), which can be created, for example by optical excitation, but also by biasing a p–n junction in forward direction or by excitation with an electron beam. The third aspect concerns the stimulated emission resulting (e.g.) from the high excitation effects and the laser emission, which is influenced by the shape of the samples and resonators, (e.g.) of bulk materials, nano rods or powders. All three aspects overlap. We present in the following selected results for ZnO following roughly the above ordering.

The purpose of this chapter is to present the results of the dynamics of exciton (polariton)s or more generally of electron–hole pairs. For a recent review of this topic concentrating on quantum wells, see Davies and Jagadish (Laser Photon. Rev. 3(1), 1(2008)). We neither consider the dynamics of carriers, for example, their relaxation time entering in Hall mobility or electrical conductivity, nor the dynamics of phonons or spins, respectively. We give here only a very small selection of references to these topics (Baxter and Schmuttenmaer, J. Phys. Chem. B, 110:25229, 2006; Queiroz et al. Superlattice Microstruct. 42:270, 2007; Niehaus and Schwarz, Superlattice Microstruct. 42:299, 2007; Lee et al., J. Appl. Phys. 93:4939, 2003; A. K Azad, J. Han, W. Zhang, Appl. Phys. Lett. 88:021103, 2006; Janssen et al., QELS 2008 IEEE 2; D. Lagarde et al., Phys. Stat. Sol. C 4:472, 2007; S. Gosh et al., Appl. Phys. Lett. 86:232507, 2005; W. K. Liu et al. Phys. Rev. Lett. 98:186804, 2007). The main characteristic time constants relevant to optical properties close to the fundamental absorption edge are the dephasing time T ₂, (i.e. the time after which the polarization amplitude of the optically excited electron–hole pair loses the coherence with the driving light field), the intra band or inter sub band relaxation times T ₃ (i.e. the time it takes for the electron–hole pairs to relax from their initial state of excitation to a certain other state e.g. to a thermal distribution with a temperature equal to or possibly still above lattice temperature) and finally the lifetime T ₁ (i.e. the time until the electron–hole pairs recombine). The characteristic time constants T ₂ and T ₁ are also known as transverse and longitudinal relaxation times, respectively. Their inverses are the corresponding rate constants. T ₂ is inversely proportional to the homogeneous width Γ, and T ₁ includes both the radiative and the generally dominating non-radiative recombination (Hauser et al., Appl. Phys. Lett. 92:211105, 2008). For this point, recall Figs. 6.16 and 6.33. Since the polarisation amplitude is gone in any case after the recombination process, there is an upper limit for T ₂ given by T ₂ ≤ 2 T₁. The factor of two comes from the fact that T ₂ describes the decay of an amplitude and T ₁ the decay of a population, which is proportional to the amplitude squared. Sometimes T ₂ is subdivided in a term due to recombination described by T ₁ and another called ‘pure dephasing’ called T ₂ ^∗ with the relation 1 ∕ T ₂ = 1 ∕ 2 T ₁ + 1 ∕ T₂ ^∗. The quantity T ₂ ^∗ can considerably exceed 2 T ₁. In the part on relaxation processes that is on processes contributing to T ₃, we give also examples for the capture of excitons into bound, localized, or deep states. For more details on dynamics in semiconductors in general see for example, the (text-) books [Klingshirn, Semiconductor Optics, 3rd edn. (Springer, Berlin, 2006); Haug and Koch, Quantum Theory of the Optical and Electronic Properties of Semiconductors, 4th edn. (World Scientific, Singapore, 2004); Haug and Jauho, Quantum Kinetics in Transport and Optics of Semiconductors, Springer Series in Solid State Sciences vol. 123 (Springer, Berlin, 1996); J. Shah, Ultrafast Spectroscopy of Semiconductors and of Semiconductor Nanostructures, Springer Series in Solid State Sciences vol. 115 (Springer, Berlin, 1996); Schafer and Wegener, Semiconductor Optics and Transport Phenomena (Springer, Berlin, 2002)]. We present selected data for free, bound and localized excitons, biexcitons and electron–hole pairs in an EHP and examples for bulk materials, epilayers, quantum wells, nano rods and nano crystals with the restriction that – to the knowledge of the author – data are not available for all these systems, density ranges and temperatures. Therefore, we subdivide the topic below only according to the three time constants T ₂, T ₃ and T ₁.

While we presented in the preceding chapters to a large extend basic physical properties of ZnO, which are, however, in many cases just the basis of and thus relevant for applications, we concentrate in this chapter on the applications themselves. We go from past applications in Sect. 13.1 to present and emerging ones in Sect. 13.2. Applications, which are according to the point of view of the author presently still highly speculative, will be presented in Sect. 13.3 as visions of future uses of ZnO and ZnO-based structures. A more humorous overview of ZnO based applications can be found in [ZnO and you, Kentucky Fried Movie (1977)].

In this final chapter, we summarize in the conclusions some aspects of ZnO research and some prospects of applications in the outlook.