Growth and characterisation of cobalt doped zinc oxide 1

In fact, most of the spin-related materials and devices to date still rely primarily on the spontaneous ordering of spins, in the form of different types of magnetic materials.. Therefor

CHAPTER 1 INTRODUCTION

1.1 Spintronics

In the last several decades, the number of transistors per unit area of an integrated circuit (IC) has doubled approximately every 18 months, following a trend known as Moore’s law.1 If this trend continues, the Si-based technology will soon face some physical limits which demand for alternative technologies To this end, intensive research work are being carried out, in both academia and industries, to look for alternative technologies Among them, spintronics, also known as spin electronics, has the potential to create devices with superior performances due to the utilization of both the charge and spin degree of freedoms of electrons.2

Electrons possess two important properties: charge and spin The ability to generate, control and detect the motion of charges in either the free space (vacuum tubes) or in solid state (Si-based electronic devices) forms the basis of modern electronics The controlled transition/recombination of electrons between different energy levels leads to the absorption / generation of photons in the different range of the electromagnetic spectrum, which forms the basis of optoelectronics (light-emitting diodes and laser diodes) Compared to the charge and the photon, it is rather difficult to generate, control and detect the spin of an electron In fact, most of the spin-related materials and devices to date still rely primarily on the spontaneous ordering of spins, in the form of different types of magnetic materials In information processing and storage, charge-based devices are dominant In information transmission, storage, and display photonic devices are used In contrast to the wide range of applications of

devices, or spintronics, are still limited to information storage (e.g., hard disk drives).3From a long term perspective, spintronics has the potential of creating devices with less

or no moving charges, which may lead to devices and systems with faster response and low power consumption.4 It also promises a greater integration between the logic and storage devices The excellent temporal and spatial coherence of spins will also make spintronic devices more suitable than their charge-based equivalents for quantum information processing.5 As illustrated in Fig.1-1, spintronics can be roughly divided into the following categories: (i) metal-based, (ii) dilute magnetic semiconductor (DMS)-based, (iii) pure semiconductor-based systems, (iv) hybrid devices and (v) other technologies

Figure 1-1 Schematic illustration of different areas of spintronics

The metal-based spintronics has its origin in the giant magnetoresistance (GMR) phenomenon, which has already been evolved into useful devices being used in hard disk drives Discovered by Fert6 and Grünberg7, a typical GMR structure consists of two ferromagnetic (FM) layers separated by a non-magnetic (NM) spacer (or the repetition

TiO 2 :Co Oxide-Based

Si:Mn

Ge:Mn

IV

PbTe:Mn PbSe:Mn PbTe:Gd Pb

IV-VI

GaAs:Mn GaN:Mn GaP:Mn InAs:Mn

Semiconductors

Hybrid

ZnO:Co ZnO:Mn

of this basic unit to form multilayers), as shown in Fig 1-2 The resistivity of the trilayer is strongly influenced by the relative orientation of the magnetization between the two magnetic layers, due to dominantly spin-dependent interfacial scattering.8 The resistance of the trilayer is low when the magnetizations of the two layers are aligned in parallel and is high when they are in anti-parallel configuration The GMR is then calculated as the ratio of resistance difference between the anti-parallel and parallel configurations to the resistance in parallel configuration Although the GMR is high in a strongly coupled multilayer structure, it could not be used as it was in hard disk drives because the external magnetic field that is required to switch the magnetizations is too high This has prompted IBM to invent a more practical structure for read sensors which is called the spin-valve.9

The state-of-the-art spin-valve consists of a dozen of thin layers; the heart of which is a trilayer structure consisting of two ferromagnetic layers separated by a non-magnetic spacer, which is usually copper The primary difference between the original GMR structure and spin-valve is that, in the latter, the thickness of the spacer is chosen such that the coupling between the two ferromagnetic layers is minimized This makes

it possible to use the spin-valve to detect rather weak magnetic field The signal detection principle is the same as that of the GMR structure, i.e., the resistance is high when the magnetizations of the two layers are in opposite directions and low when they are in the same direction To have a linear response from the sensor, the angle between the two magnetizations is normally set at 90o at zero-field with one of them “pinned” at

a direction perpendicular to the media surface through exchange-coupling with an antiferromagnet and the other free to rotate in response to the fringe field of magnetic transitions recorded on the magnetic media, also known as the free layer.10

Figure 1-2 Schematic illustration of a GMR trilayer structure, with two ferromagnetic layers separated by

a non-magnetic spacer The resistance is low when the magnetizations are parallel and high when they are

There are two different forms of spin-valve sensors, depending on whether the current flows in the plane of the stack of layers or perpendicular to them The former is called a current-in-plane, or CIP, spin-valve sensor and the latter a current-perpendicular-to-plane, or CPP, sensor So far, CIP is dominant, but CPP is expected to play an important role in future terabit recording systems An alternative of the CPP spin-valve is the magnetic tunnel junction,11,12 or MTJ, in which the current also flows

Parallel configuration (Low resistance)

Anti-Parallel configuration (High resistance)

perpendicular to the plane The major difference between the CPP spin-valve and MTJ

is that the latter is composed of two ferromagnetic layers separated by an insulator instead of a metal Therefore, the electrical conduction in MTJ is based on quantum mechanical tunneling The recent rapid improvement of MTJ devices using crystalline MgO barrier has paved the way for commercialization of MTJs in both hard disk drives and magnetic random access memories (MRAMs) (see Fig 1-3).13

Although the metal-based spintronics has achieved unparalleled success in both fundamental research and practical applications, the lack of capability in charge modulation greatly limits its applications to information processing Therefore, one of the hottest issues in spintronics is how to create a stable source of spin-polarized

carriers in semiconductors, which allows not only the modulation of charges but also

manipulation of spins.14-16 So far, several approaches have been proposed to solve this problem,17 - 20 including the injection of spins from a ferromagnetic metal into a semiconductor via either a Schottky21,22 or tunneling barrier,23-26 generation of spin polarized current using either a ferromagnetic semiconductor27 or a non-magnetic semiconductors based on spin-orbit interaction,28 spin-dependent resonant tunneling,29,30Zeeman effect31-33 and optical orientation.34 In addition to the generation of spins, some

of the aforementioned effects such as spin-orbit interaction also provide a convenient way to control the spins using an electrical field.35 In spite of the significant progress made recently in pure semiconductor or hybrid spintronic structures, the true era of spintronics will probably only be materialized after room temperature magnetic semiconductors become available

1.2 Diluted magnetic semiconductors

1.2.1 Different types of DMSs

Magnetism and semiconducting properties are known to coexist in some materials, such as europium chalcogenides36 , 37 and ferrimagnetic or ferromagnetic semiconducting spinels.38 These materials have been extensively studied since 1960’s, because of their peculiar properties resulting from the exchange interaction between itinerant electrons and localized magnetic spins These interactions lead to a rich variety of optical and transport phenomena, which are strongly affected by the magnetic

field and temperature However, the low Curie temperature, T c, and difficulties in material preparation make this family of compounds less attractive from the application point of view

In addition to these “concentrated” magnetic semiconductors, there were also intensive researches on diluted magnetic semiconductors which are obtained by doping them with a few percent of magnetic ions.39 Most initial work had been centered on II-

VI semiconductors (A,X)B where A = Zn, Cd, Hg, X = Fe, Mn, Co, Ni, Cr and B = S,

Se, Te, and in most cases the valence of group II cations is identical to that of most magnetic transition metals Although these materials are relatively easy to prepare, most of them are random antiferromagnets or spin-glasses, which makes the II–VI DMS unattractive for electronic applications Nevertheless, the presence of sp-d exchange interactions between d electrons of magnetic ions and electrons and holes of the host semiconductor does make some of these materials very useful for magneto-optical applications.40 The ease of bandgap engineering as well as the preparation of heterostructures in these materials also makes them excellent candidates for studying spin-charge interaction and the corresponding dynamics in a variety of quantum structures

1.2.2 (Ga, Mn)As: A Model DMS System

Among all types of DMS materials studied so far, (Ga,Mn)As emerges as one of the best understood model ferromagnet which exhibits not only the properties of a metallic ferromagnet but also new functions such as electrical gating and strain modulation of the magnetic properties 41 The early work on III-V magnetic semiconductors faced the difficulties of low solubility of transition metals in III-V semiconductors, which made it difficult to achieve uniform doping of magnetic impurities in these materials The breakthrough came when attempts were made to grow (In, Mn)As42,43 and (Ga, Mn)As44 using non-equilibrium molecular beam epitaxy

at low temperature, which allows the incorporation of Mn up to a few percent, with a substantially suppressed formation of secondary phase This has led to the discovery of hole-induced ferromagnetic ordering in Mn-doped III-V semiconductors In contrast to Mn-doped II-VI DMS, where the Mn only contributes spins, when Mn is substituted for gallium in GaAs or indium in InAs, it acts as both an acceptor, which provides holes and also localized spins associated with d electrons of Mn2+ ions.38, 41 The former mediates a ferromagnetic interaction among the localized spins of the opened d-shells of

the Mn atoms The T c of (Ga, Mn)As is found to be dependent strongly on the hole concentration, and the collective ferromagnetic behavior of the local spins requires a

minimum doping concentration of 2% The T c was found to be almost a linear function

of the Mn composition up to about 5% beyond which, however, a further increase of

Mn concentration will cause a decreases of T c.45 Annealing at low temperature was

found to greatly enhance the hole density and consequently the T c and saturation magnetization (MS).46-51 A T c up to 173 K has been obtained in annealed samples for a

Mn concentration of 6.4%.52 It was also shown recently that the T c can be further increased to about 250 K in heterostructures consisting of Mn δ-doped GaAs and p-type

AlGaAs layers by varying the growth sequence of the structures followed by temperature annealing.53 It is still not certain when and whether the T c can be pushed up

low-to above room temperature in the (Ga, Mn)As system.54 In order to make DMS a real

technology, however, one needs to find DMS materials with a T c higher than room temperature.55-60

Theoretically, the mean-field Zener model predicts that DMSs with a T C above room temperature are obtainable if the combination of host material, carrier concentration, and magnetic impurity (type and density) is right.61 In particular, if one could introduce 5% of Mn and 3.5×1020 cm-3 of holes into wide-gap semiconductors, such as GaN and ZnO, these materials should be ferromagnetic at room temperature In addition, the first-principles calculations also predict a rather stable ferromagnetism for these materials.62 , 63 Stimulated by these predications, intensive research has been

carried out to explore high T C diluted magnetic semiconductors, particularly oxide and nitride based DMSs.56-60

1.2.3 ZnO-based DMS Systems

Among all different types of materials that have been investigated, oxide-based DMS systems have attracted special attention, in particular TiO2 and ZnO based materials because of their attractive properties and a wide range of applications In this proposal, the focus is on ZnO-based DMS As mentioned above, theoretical studies predicted that V, Cr, Fe, Co, or Ni doped ZnO is a half-metallic double-exchange ferromagnet, whereas Ti or Cu doped ZnO remains paramagnetic; Mn doped ZnO is an antiferromagnetic insulator which changes to a ferromagnet by additional doping of holes, and ZnO doped with 5% Mn and 3.5×1020 cm-3 hole concentration has a T c above room temperature.61-63 It was also shown that electron doping stabilizes the ferromagnetic ordering of Fe, Co, or Ni doped ZnO.63 The first-principles spin-density

functional calculations by Lee and Chang predicted that heavy electron doping and high

Co concentration are required for obtaining ferromagnetism in cobalt doped zinc oxide (ZnO:Co).64 On the other hand, Spaldin argued theoretically that only hole doping promotes ferromagnetism in both ZnO:Co and ZnO:Mn.65 Till recently, Sluiter et al predicted that both hole doping and electron doping promote ferromagnetic ordering in ZnO:Co and ZnO:Mn.66 Hydrogen-mediated spin-spin interaction was also predicted to

be able to induce high temperature ferromagnetism in ZnO:Co.72

Similar to theoretical work, experimental investigations to date have also produced widely diverging results, ranging from non-ferromagnetic to ferromagnetic

with extrinsic origins to intrinsic ferromagnetism with various T C.56,57,59Ferromagnetism up to room temperature was first observed on 15 % Co-doped ZnO thin film.67 The average magnetic moment per Co atom was found to be 2 µB It was suggested that there are three possible mechanisms responsible for the ferromagnetism: (1) carrier-mediated ferromagnetic coupling between Co atoms; (2) weak magnetism of CoO phase and (3) Co clusters The mechanisms (2) and (3) were excluded from the strength of the magnetic properties, Co composition dependence of lattice constant and absence of Co cluster peaks in x-ray diffraction (XRD) peaks

Since this first report, there have been over 200 journal papers published on ZnO-based DMSs If the publications on ZnO-based DMS in the last 3 years64,68-165 has been divided into four categories, based on their magnetic character: homogeneous DMS, extrinsic ferromagnet, paramagnet, and others, the percentages of papers in the four categories were 55%, 9%, 12% and 24%, respectively This material still requires much work before a consensus can be reached on its nature of magnetism The large disparity in experimental results is partially caused by the fact that the properties of ZnO:Co are very sensitive to the structure and chemistry at the nanoscale regime, which

strongly depends on processing conditions It is also caused by the lack of systematic studies, particularly on the same sample Most of the characterization techniques only probe a certain portion or aspect of the sample in either the spatial or energy domain For samples with a nanoscale inhomogeneity, the results are meaningful only when the relationships between the structures and properties at the nanometer scale are well understood

Figure 1-4 Number of papers versus characterization techniques for ZnO-based DMS

Figure 1-5 Possible magnetic phases of ZnO-based DMSs

Figure 1-4 summarizes the techniques that have been used to characterize based DMSs, which can be divided into five main groups: structural, chemical, optical, electrical and magnetic characterization techniques Similar statistics has been obtained for TiO2-based DMS system In order to understand how effective these techniques are

ZnO-in characterizZnO-ing the DMS, the different types of possible magnetism phases ZnO-in magnetically doped oxides is schematically shown in Figure 1-5 At very low doping level, there is no or very little interaction among the magnetic dopants; therefore, the system can be considered as either a paramagnet or weak magnet In the latter case, a

small hysteresis might be originated from interaction between the local spin of electrons and the crystal field of the host material When the dopant concentration further increases, other interactions such as Ruderman, Kittel, Kasuya, and Yosida (RKKY), superexchange and double-exchange (in the case of mixed valence) may become possible, and they can be either ferromagnetic or antiferromagnetic, depending

d-on the type and cd-oncentratid-on of dopants and changed defects Unlike (Ga, Mn)As and (In, Mn)As, so far there has been no established report of carrier mediated or RKKY ferromagnetism in ZnO-based DMSs As the valence of transition metals in ZnO is believed to be 2+, in order to introduce carriers into the material, one must rely on either the doping of non-magnetic impurities or oxygen / zinc deficiency These impurities and dopants make the material highly defective which may easily lead to carrier localization For n-type ZnO, which is true in most cases, the localized electrons surrounding the magnetic impurities will form so-called bound magnetic polarons (Fig 1-5c).39,54,166 The polarons increase in size when temperature decreases When they merge, ferromagnetic ordering may occur either locally or globally in the sample Depending on the growth techniques and conditions, most of the frequently reported materials contain clusters of magnetic dopants or secondary phases (Fig 1-5d and 1-5e) The main difficulty in understanding the true mechanism of magnetic ordering in oxide-based DMSs is that, in most cases, all these different phases are presence in a same sample.56,57,59,68-165 Therefore, it is almost impossible to identify the true mechanism by using a combination of limited number of techniques shown in Fig 1-4 For example, x-ray diffraction is unable to differentiate cases (a)-(c) with cases (d) and (e) (Fig 1-5) unless the density and size of the clusters or secondary phases reach certain threshold values It will be difficult to detect the clusters and secondary phases by conventional transmission electron microscopy (TEM) when they are either very small or too dilute

The shape of the magnetization-field (M-H) or magnetization-temperature (M-T) curves may tell some differences between cases (a), (b) and (c)-(e), but most practical cases fall into categories (c)-(e) The observation of d-d transition in optical spectroscopy is often used to prove that the transition metal dopants substitute the host cations However, most of the optical transmission spectroscopy has a low spatial resolution; moreover, the observation of d-d transition does not necessarily mean that the observed magnetic properties originated from carrier-mediated ferromagnetic alignment of substitutional magnetic ions The recently observed anomalous Hall effect (AHE), in principle, can also originate from any one of the four categories (b)-(e) In the last two cases, the AHE may arise from the fringe field of magnetic clusters.167-169 The magnetic circular dichroism (MCD) spectroscopy might be able to differentiate substitutional magnetic ions from those of magnetic clusters; again MCD alone is unable to differientiate or pinpoint any of the single phases due to limited spatial resolution.170

In addition to chemical or structural inhomogeneity, magnetic inhomogeneity also exists in DMSs due to the formation of bound magnetic polarons Although it is

generally believed that the ferromagnetism below T C in (Ga,Mn)As is caused by mediated RKKY interaction, the real mechanism is still not well understood in

hole-particular at the nanoscale regmine The mechanism of magnetic interaction about the T c

is even more complicated The bound magnetic polaron model has been used to describe

how the ferromagnetic ordering occurs when the sample is cooled across the T c.39,54,166

Due to defects and random distribution of magnetic dopants, carrier localization is expected to happen at local potential minimum when the sample temperature is reduced The carrier localization is particularly enhanced in the insulating regime where impurity band forms in the deep gap of the host semiconductor.171 The interaction of magnetic impurity and localized carriers leads to the formation of magnetic polarons The size of

magnetic polaron increases with decreasing temperature and ferromagnetic ordering

takes place when the magnetic polarons merge at the T c When the magnetic dopant density increases, the direct Mn-Mn antiferromagnetic interaction will gradually

become dominant, which eventually reduces the T c to zero The magnetic polaron picture is believed to be also true for other types of DMSs, in particular, those in which carrier mediated ferromagnetism has not been confirmed such as Co-doped ZnO In addition to the aforementioned study of structural and chemical inhomogeneity, the transport technique should also be very useful in probing the formation and interaction

a four-probe technique In the case that the inhomogeneous region is much smaller than the specimen size, no noticeable effect originated from inhomogeneity will be observed

in both the I-V and dI-dV curves unless the density of inhomogeneity exceeds a certain percolation threshold However, the situation will change when the samples are fabricated into one-dimensional nanowires In this case, due to the geometrical confinement of current path, most of the electrons will have to encounter the

inhomogeneous region before they could reach the other end of the sample Depending

on the nature of electrical conduction in both the host matrix and inhomogeneous region, one may observe I-V curves with different characteristics In oxide-based DMS, one may encounter the following different situations:

Table 1-1 Summary of transport properties in oxide-based DMS system

Ferromagnetic

junctions Host matrix

1.3 Motivation and Objectives of this Work

As discussed above, there is strong demand for DMS materials with T C higher than room temperature for spintronic applications Although room temperature oxide-DMSs have been reported, recent studies have produced widely diverging results, in

magnetic properties, likely due to the different processing conditions used In addition

to the high-sensitivity of magnetic properties to preparation techniques and conditions, the rather chaotic situation is also caused by the lack of a commonly agreeable way to determine if a DMS is intrinsic when a magnetic moment versus applied field curve with a very small hysteresis is observed in a simple magnetometry measurement The common approach taken so far by most experimental work which supports the existence

of intrinsic ferromagnetism in ZnO:Co is as follows First, the existence of ferromagnetism is “confirmed” by either direct measurement using a magnetometer or the combination of this with indirect measurements such as MCD and x-ray MCD Second, the presence of precipitates and other secondary phases is “excluded” by high-resolution transmission electron microscopy (HRTEM) and x-ray diffraction measurements Third, substitution of Zn with Co into the host matrix is confirmed by optical absorption, magneto-optical spectroscopy (including the confirmation of coupling between band carriers with d electrons of the Co ions) or other valence analysis techniques The validity of this approach is only relevant if all the measurements are performed on the same physical location of the same sample as well

as if the measured results are well correlated with each other In the latter case, in addition to magnetic and magneto-optical measurements, it is also of great importance

to perform systematic studies of electrical transport properties Given the fact that it is almost impossible to ensure that all the sample preparation techniques and conditions are identical for experiments performed at different groups, researchers should at least make efforts to conduct systematic studies using their own specific experimental setups

In order to draw meaningful conclusions from the experiments, it is crucial that all different types of characterizations should be carried out on same samples instead of samples with similar chemical concentrations

Based on this background, the main objectives of this work are to conduct a systematic study of ZnO:Co thin films prepared by sputtering with Co composition varying in a large range Special attentions will be paid to ensure that a series of characterization experiments can be performed on each of the series of samples studied Emphasis will be on finding out if there is any correlation between the results obtained from different characterization techniques so as to reveal the real mechanism of magnetism in ZnO:Co thin films

1.4 Organization of the Thesis

After having provided a brief background of this research, an overview of the work done so far on Co-doped ZnO is provided in Chapter 2 Mn doped ZnO and Co doped TiO2 are also briefly introduced in this chapter Chapter 3 describes the growth and characterization techniques used in this work, in which the focus is on those which are more specific to this work The structural and chemical analyses of the obtained films are given in Chapter 4, while Chapters 5 and 6 describe the magnetic and electrical transport properties, respectively This work is summarized and some suggestions for future work are also provided in Chapter 7

3 Y H Wu, “Nano-spintronics for data storage”, Encyclopedia of Nanoscience and Nanotechnology, 7, 493 (2004)

4 D D Awschalom, M E Flatté, and N Samarth, “Spintronics”, Scientific American,

286, 66 (2002)

5 R Ramesh and N A Spaldin, “Multiferroics: progress and prospects in thin films”, Nat Mater 6, 21 (2007)

6 M N Baibich, J M Broto, A Fert, F N Van Dau, F Petroff, P Eitenne, G Greuzet,

A Friederich, and J Chazelas, “Giant magnetoresistance of (001)Fe/(001)Cr magnetic superlattices”, Phys Rev Lett 61, 2472 (1988)

7 G Binash, P Grünberg, F Saurenbach, and W Zinn, “Enhanced magnetoresistance in layered magnetic structures with antiferromagnetic interlayer exchange”, Phys Rev B

39, 4828 (1989)

8 S S P Parkin, “Origin of enhanced magnetoresistance of magnetic multilayers: dependent scattering from magnetic interface states”, Phys Rev Lett 71, 1641 (1993)

Spin-9 B Dieny, V S Speriosu, S S P Parkin, B A Gurney, D R Wilhoit, and D Mauri,

“Giant magnetoresistive in soft ferromagnetic multilayers”, Phys Rev B 43, 1297