Chapter Origin of ferromagnetism from transport studies CHAPTER ELECTRICAL TRANSPORT PROPERTIES OF CoDOPED ZnO 6.1 Introduction As discussed in Chapter 3, all the samples had been fabricated into Hall bars so as to carry out systematic studies on the transport properties of Co-doped ZnO. The electrical transport measurements that had been carried out include temperaturedependence of resistivity, carrier density, differential conductance, MR and Hall effect. The differential conductance measurement for DMS is new and is only performed for the first time in this work (to the best of our knowledge). As the differential conductance is very sensitive to local environment felt by the carriers, it is a powerful technique to characterize the uniformity of ZnO:Co. Some preliminary results on the study of ZnO:Co-superconductor junctions were also presented. The results of electrical transport properties will be presented first, followed by the discussion of origin of ferromagnetism in ZnO:Co, in combination with other results discussed in the previous chapters. 6.2 Differential conductance studies Inhomogeneity and disorder exist virtually in all types of materials, in particular in the material system under study, as seen from the structural properties presented and discussed in Chapter 3. These inhomogeneity and disorder would affect the electrical transport properties in a profound way. In most of the work reported so 147 Chapter Origin of ferromagnetism from transport studies far for ZnO:Co, the electrical transport properties only include the resistivity, carrier concentration and Hall effect. As average parameters, however, these quantities could not reveal directly how the structural properties were related to the transport properties. As inhomogeniety and disordering induce potential fluctuations, and in the extreme case form sub-regions of totally different characteristics, the effects should be reflected more apparently in the dynamic conductance as a function of the bias voltages. For instance, in a system in which metallic particles were embedded in a semiconductor host matrix, one should expect the appearance of Schottky junction behaviour when the density of nanoparticles exceeds a certain threshold value or when the current path was confined in a narrow region. Based on these considerations, the dI/dVxx versus Vxx curves for all the samples were measured at different temperatures. The results were shown in Figs. 6-1 (a)-(h), respectively. Here the discussion was focused on the following two aspects: (1) the shape of the conductance curve and (2) the size of the zero-bias anomaly (ZBA). Before proceeding to discuss the data, the effect of current induced heating will be commented on. When the bias current increased, it was unavoidably that the sample would be heated up to a temperature which was higher than the cryostat temperature. As all the samples under study exhibited a dirty metal behaviour, the local heating might lead to a higher conductance at a higher bias. As all the samples under study had almost the same thickness and the Hall bars were of same dimension, the temperature rise should be inversely proportional to the resistivity of the sample when the bias voltage was at the same level. Bearing this in mind, the results from the dI/dV curves will now be discussed. Shown in Fig. 6-1(a) are the dI/dV curves for Al-doped ZnO without Codoping. As can be seen from the figure, the differential conductance increased with the bias voltage at low temperature and becomes almost a constant above 30 K. As the 148 Chapter Origin of ferromagnetism from transport studies ZBA disappeared at the presence of a magnetic field, it should originate from weaklocalization in this sample. As the sample was further doped with Co, both the shape and ZBA of the dI/dV curve changed remarkably. As shown in Figs. 6-1(b) - (d) (x =0.05, 0.14 and 0.2), the differential conductance curves for x < 0.2 had roughly a “V” shape. With a further increase of Co composition, the low-bias curve evolved gradually into a “U” or parabolic shape (Fig.6-1 (e)-(h)). The V-shape could be understood as being caused by electrical-field assisted “de-trapping” of carriers localized in shallow potential wells. On the other hand, the U-shape and parabolic curves could be attributed to the transport across grain boundaries which could be of either a tunnel junction or nanoscale heterojunctions / Schottky junctions,1 depending on the electrical characteristics of the secondary phases. Although current induced heating may also affect the shape of the dI/dV curve, it should not be the dominative mechanism because the resistivity of the lightly doped samples was almost the same as that of heavily doped samples (to be discussed shortly). Now we turn to the temperature-dependence of ZBA for different samples, as shown in Fig. 6-2 ((a)-(h)). In order to have a meaningful comparison, the ZBA was defined as the conductance ratio between V = V and V, i.e., dI/dVxx(2V)/ dI/dVxx(0). The ZBA, in principle, can appear in many different situations. In a 4point probe measurement configuration like the one used in this study, the influence of sample-electrode contact could be neglected and, apart from weak localization and heating effect, the ZBA was mainly caused by electrostatic potential disorder in the sample. The difference between weak localization and electrostatic potential induced carrier localization could be readily differentiated from the dependence of ZBA on an applied magnetic field. The ZBA for ZnO:Al disappeared completely at an applied field of 1T perpendicular to the sample surface, while those for samples with Co 149 Chapter Origin of ferromagnetism from transport studies doping were insensitive to the external field. The size of ZBA served as an indicator of carrier localization strength and the shape of the differential conductance curve helped to identify the carrier transport mechanism in individual samples. Although currently there is no theoretical model available to explain quantitatively the shape of differential conductance curves observed here, clear changes in both the size of ZBA and shape of the conductance curve as the Co composition was varied had been observed. Focussing on the ZBA at 4.2 K, its value initially increased with the Co composition and reached a maximum at x = 0.25 beyond which it decreased again to unity at high Co composition. As discussed in Chapter 4, the Co atoms were uniformly dispersed in the host matrix in samples with x < 0.2. When x increased to 0.25 at the onset of secondary phase formation, there was a significant increase in the degree of disorder inside the samples. A further increase of Co composition would in turn lead to the enhancement of ordering due to formation of percolated secondary phases or Co clusters. Therefore, the ZBA was expected to increase with the Co composition below the onset point of secondary phase formation and decreased again after it reached a maximum value. It was worth noting that the shape of the dI/dV curves was also different at the two sides of the onset composition. These results demonstrated clearly the usefulness of this technique to establish the relationship between the structural properties with the electrical transport properties. As would be shown later, the ZBA was also correlated with the carrier localization in Co-doped samples. 150 Chapter Origin of ferromagnetism from transport studies 1.966 20K 1.964 -3 1.962 10K 8K 1.960 6K 1.958 (a) -2 50K 7.62 30K 7.60 10K 7.58 4.2K -1 (b) A (x=0.05) 7.64 14K G (10 S) -2 G (10 S) 30K 20K 14K -2 -1 2.65 -3 20K 14K 2.7 10K -2 -1 50K 4.2K 2.55 2.50 8K 2.6 (c) B (x = 0.14) 2.45 -3 10K 30K 20K 8K 6K 4.2K 15K -2 -1 Voltage (V) 2.0 G (10 S) -3 -4 G (10 S) 8K 5K (e) F (x=0.25) -2 1.20 15K 10K 1.15 -6 4.2K -4 -2 Voltage (V) G (10 S) 4.4 -3 -2 -1 -3 -3 G (10 S) 10K 8K 6K 5K 4.2K (g) J (x=0.30) 20K 3.70 15K 4.5 Voltage (V) 20K 4.6 6K (f) G (x = 0.27) 4.2K 1.2 -4 20K 6K 1.4 1.25 10K 1.6 Voltage (V) 14K 1.8 (d) D (x=0.20) 2.60 30K G (10 S) -3 G (10 S) 50K 2.8 Voltage (V) Voltage (V) 2.9 8K 6K 4.2K Voltage (V) 10K 8K 3.60 3.55 -2 14K 3.65 6K (h) L (x=0.33) -1 4.2K Voltage (V) Figure 6-1 Differential conductance at various temperatures as a function of applied voltage of (a) Aldoped ZnO and various co-doped samples, with x values indicated. 151 1.01 1.01 0.99 1.00 ZBA ZBA Chapter Origin of ferromagnetism from transport studies (a) 0.97 100 200 0.99 300 Temperature (K) 1.06 ZBA ZBA (c) B (x=0.14) 1.05 0.95 100 200 1.09 (e) F (x=0.25) 100 200 0.99 300 100 200 300 1.04 (g) J (x=0.30) (h) L (x=0.33) 1.03 ZBA ZBA (f) G (x=0.27) Temperature (K) 1.08 200 1.04 Temperature (K) 0.98 100 Temperature (K) ZBA ZBA 300 (d) D (x=0.20) 1.02 300 1.5 0.5 200 1.04 Temperature (K) 2.5 100 Temperature (K) 1.08 1.15 (b) A (x=0.05) 100 200 Temperature (K) 300 1.02 1.00 100 200 300 Temperature (K) Figure 6-2 Zero bias anomaly as a function of temperature for (a) Al-doped ZnO and various codoped samples, with x values indicated. 6.3 Temperature and Co composition-dependence of resistivity From the dI/dV curves at different temperatures for samples with different Co compositions, one could readily obtain the dependence of resistivity on both temperature and the Co composition. The dependence of resistivity on the Co composition is shown in Fig. 6-3 at 4.2 K. The resistivity increased slowly with the 152 Chapter Origin of ferromagnetism from transport studies increase of Co composition for x < 0.25. After reaching a maximum at x = 0.25, it decreased with a further increase of the Co composition. The trend was almost the same as that of the dependence of ZBA on Co composition, as shown in the inset. This again suggested that the degree of disorder was highest at x = 0.25, the onset composition of formation of secondary phases. 10 20 30 40 Co composition (%) Figure 6-3 Resistivity versus Co composition at 300 K; Inset shows zero bias anomaly dependence on Co composition at 4.2 K. Fig. 6-4 shows the zero-bias resistivity as a function of temperature for samples with different Co compositions. As could be seen from the figure, all the samples at low and high Co compositions exhibited a typical “dirty metal” behaviour, while those in-between behaved like an insulator due to strong disorder caused by the onset of phase separation. On the other hand, as shown in the inset, the Al-doped ZnO sample exhibited a weak semiconductor-like behaviour with the resistivity being almost independent of temperature. Comparing Fig.6-4 with Fig.6-1, one immediately realized that the dI/dV curves provided more insights into the electrical conduction mechanism of the samples. 153 Chapter Origin of ferromagnetism from transport studies Figure 6-4 Resistivity versus temperature for various co-doped samples; Inset shows resistivity versus temperature for Al-doped ZnO. 6.4 Hall effect The presence of AHE is considered as one of the strong evidences for intrinsic ferromagnetism in DMSs.2,3 However, considering the fact that AHE had also been reported in ferromagnetic clusters,4 granular materials5-7 and inhomogeneous DMS in the hopping transport regime,8 the observation of AHE alone cannot support the claim that the DMS under study is a ferromagnet of intrinsic origin, unless it was correlated with ferromagnetism observed by other means. Most importantly, secondary phases and precipitates must be shown to be absent in the sample. 154 Chapter Origin of ferromagnetism from transport studies Figure 6- Hall voltage as a function of applied magnetic field for (a) Al-doped ZnO and various codoped samples, with x values indicated, at 4.2K and 300 K, current applied 0.1mA. 155 Chapter Origin of ferromagnetism from transport studies Fig. 6-5 shows the Hall voltage as a function of the applied field for different samples at different temperatures. For non-magnetic samples, it was well known that the offset due to contact misalignment could be corrected by calculating the Hall voltage as [Vxy(B)-Vxy(-B)]/2. However, for magnetic samples, the offset was dependent on the magnetic field due to the strong MR effect. Therefore, all the data shown in Fig. 6-5 had been corrected for the MR effect. However, the correction for the nonmagnetic field dependent offset by calculating the Hall voltage as [Vxy(B)-Vxy(B)]/2 had not been carried out because the sample contains an antiferromagnetic phase; any operation involving [Vxy(B)-Vxy(-B)]/2 may lead to a wrong Vxy – B loop which would make it difficult to compare the AHE loop with the M-H loop measured by SQUID. However, [Vxy(B)-Vxy(-B)]/2 at the maximum field had been used as the Hall voltage to calculate the carrier concentrations (to be presented shortly). As shown in Figs. 6-5 (a)-(c), the Zn1-xCoxO samples with x < 0.2 showed only OHE. As the Co concentration increased, the AHE appears in samples with x > 0.25. The onset composition at which AHE started to appear also coincided with that of the appearance of secondary phases, i.e., at x = 0.25 in sample F. All other samples (G-L) with x > 0.25 exhibited very clear AHE characteristics, as shown in Fig. 6-5 (d)-(f) at 4.2K and Fig. 6-5 (g)-(i) at 300K for samples F, G and J, respectively. Therefore, now it could be concluded that the presence of AHE in films with x > 0.25 was due to ferromagnetic secondary phases and not due to intrinsic ferromagnetism. Although clear hysteresis loops had been observed for all the samples by SQUID, there was a fundamental difference between the samples with x < 0.25 and those with x 0.25. In the former, the magnetic properties came from magnetic regions which were electrically “isolated” from each other. On the other hand, the magnetic regions in 156 Chapter Origin of ferromagnetism from transport studies -2 2.2×10 (a) -2 2.1×10 -2 2.0×10 -2 1.9×10 -2 1.8×10 -2 2.15×10 (b) -2 2.05×10 -2 1.95×10 -2 1.85×10 Resistance (Ohm) 4.96×10 4.96E+01 4.94×10 4.94E+01 4.92E+01 4.92×10 4.90E+01 4.90×10 4.88E+01 4.88×10 4.86E+01 4.86×10 4.84E+01 4.84×10 4.82E+01 4.82×10 -0.04 (c) -0.02 0.02 0.04 Field (T) Figure 6-14 dI/dV curves as a function of bias voltage for sample J, Co40W (Zn0.70Co0.30O) at (a) 1.6 – 10 K, (b) 1.6 K at various in-plane applied magnetic field, (c) MR curve at 1.5 K for applied field -0.04 – 0.04T. 171 Chapter Origin of ferromagnetism from transport studies 2.05E-02-2 2.05×10 dI/dV (S) 2.00E-02-2 2.00×10 1.4K 2.6K 3.4K 4.2K (a) 1.95E-02-2 1.95×10 1.90E-02-2 1.90×10 1.85E-02-2 1.85×10 -0.04 -0.02 0.02 0.04 Voltage (V) 2.04E-02 2.04×10-2 dI/dV (S) 2.00E-02 2.00×10-2 (b) 0T 0.02T 0.022T 1.96E-02 1.96×10-2 0.025T 0.1T -2 1.92×10 1.92E-02 0.5T 1.88×10-2 1.88E-02 1T 2T 1.84×10-2 1.84E-02 -0.04 -0.02 0.02 0.04 Voltage (V) Resistance (Ohm) 5.30E+01 5.30×10 5.25E+01 5.25×10 5.20×10 5.20E+01 5.15E+01 5.15×10 5.10E+01 5.10×10 5.05E+01 5.05×10 -0.04 (c) -0.02 0.02 0.04 Field (T) Figure 6-15 dI/dV curves as a function of bias voltage for sample L, Co50W (Zn0.67Co0.33O) at (a) 1.6 – 10 K, (b) 1.6 K at various in-plane applied magnetic field, (c) MR curve at 1.4 K for applied field -0.04 – 0.04T. 172 Chapter Origin of ferromagnetism from transport studies Although the results obtained here were preliminary, they agreed well with the results obtained by other characterization techniques discussed in previous chapters. Further studies were required to obtain spatially dependent polarization ratio using point-contact measurement geometry. 6.8 Origin of ferromagnetism in ZnO:Co The most important question now is whether this material exhibits carriermediated ferromagnetism. In Fig. 6-16 below, the possible magnetic contributors in a Co-doped ZnO system are summarized. As the Co content increases, the magnetic contributors are likely to go from a paramagnetic phase to eventually magnetic cluster phase. Before clusters form, it is still debatable whether the origin of ferromagnetism is due to RKKY or magnetic polaron or even other mechanism. (a) Paramagnetic (c) Magnetic polaron (b) RKKY (d) Cluster Figure 6-16 Possible magnetic sources that can be present in Co-doped ZnO system. The general trend in studies of this material is of the order below: (1) use XRD or TEM to show that Co is soluble in ZnO host matrix and films are impurity and secondary phases free; (2) use optical transmission or XPS to show that Co substitutes Zn as Co2+ ions; and (3) magnetic measurements to show that films are ferromagnetic in nature. Satisfying the above three points, it is often assumed the material involved exhibits intrinsic ferromagnetism or carrier-mediated ferromagnetism. However, from 173 Chapter Origin of ferromagnetism from transport studies the results obtained, even if the above three characters are observed, it does not imply that the samples show evidence of intrinsic ferromagnetism behaviour. The characteristic of the films prepared will now be pieced together to determine the origin of ferromagnetism. Since it is quite obvious the samples can be divided into three groups, each group will be discussed individually. 6.8.1 < x < 0.25, intrinsic ferromagnetism? From XRD studies, results have shown that co-doped samples within this Co content range has good crystalline quality and are single phase films, with no impurity phases. Also from TEM images and diffraction patterns, the films are homogeneous in nature and free from precipitation. Although EELS analysis is not carried out on these low Co-doped samples, since in heavily doped samples Co exists only as +2 valence state, without any Co in valence state, it can be implied that Co exist as +2 valence state in these lightly co-doped samples. The +2 valence state of Co can be confirmed from XPS studies, where presence of Co clusters is also eliminated. From structural studies, another feature to show that Co atoms have substituted Zn atoms is through optical transmission. The d-d transitions characteristic of tetrahedrally coordinated Co2+ were observed. From structural studies, it can be confirmed that high quality Co-doped ZnO films has been prepared with good structural quality, free of precipitations. It should be emphasised that the origin of ferromagnetism cannot be implied with the above-mentioned results, so more studies are carried out as follows. Turning the attention to its magnetic properties, these samples exhibit weak ferromagnetic properties with small coercivity of less than 200 Oe. For magnetic moment, contributions from the strong diamagnetic Al2O3 substrate have to be 174 Chapter Origin of ferromagnetism from transport studies subtracted from the raw data to observe the clear hysteresis loop for these samples. After processing of the data, a paramagnetic signal which increases with decreasing temperature has been observed. This paramagnetic signal might be from Co atoms distributed in the host matrix or from defects. These samples also exhibit in-plane anisotropy as observed from the work performed by Sati22. However, it is impetuous to assume the in-plane anisotropy is intrinsic ferromagnetism. From the M-H curves studied at different temperatures, there is a possibly that ferromagnetism observed is due to multi-phases present in the material. In most theories of DMS, it is assumed that the transition metal ions will substitute the cation site of the host matrix. Exchange interaction between the nearest neighbour magnetic ions is mediated by carriers which will then result in ferromagnetic properties. To show carrier-mediated ferromagnetism, there is a need to show that Tc and magnetisation of the films increase as carrier concentration increases. However, it does not seem to fit the characteristic of the discussed samples, as when the Co content is increased, carrier concentration decreases, but magnetisation increases. In BMP model, with decrease of temperature, shallow donor electrons leads to the formation of BMPs23. When these polarons overlap, they are able to lead to longrange interaction, leading to ferromagnetism. Tc, given in equation (2.3), is dependent on the concentration of magnetic cations and donors. Shallow donors electrons, in the form of oxygen vacancies, form split donor impurity band which lie at the Fermi level of the 3d impurity bands, leading to high Tc. It is difficult to observe formation of BMPs in the magnetic studies, but a negative MR, in MR studies when a larger magnetic field is applied at low temperatures, is observed. This negative MR might be 175 Chapter Origin of ferromagnetism from transport studies due to alignment of electron spin with Co2+ ions or formation of BMPs. MR measurements for these samples done at smaller field show a positive MR at low applied field, characteristic of s-d exchange interaction between conducting carriers and localized spins of Co ions. Weak localization is observed clearly in the differential conductance studies. Also with temperature-dependence of carrier concentrations measurements from Hall effect studies, results suggests strongly that carriers are localized at low temperature and become de-trapped as temperature increases. These transport studies, however, where not able to draw any correlation between transport and magnetic properties. The widely believed scenario of ferromagnetic ordering in highly disordered DMSs is as follows. Unless a ferromagnetic ordering is already established above room temperature, otherwise when the temperature decreases, carriers will become gradually localized at small potential valleys surrounding, which BMPs may be formed. The size of the BMP increases with decreasing temperature, which may eventually lead to a ferromagnetic ordering when the BMPs merge globally. From the studies done so far, BMP might be able to fit the lightly co-doped samples, however transport and magnetic properties seem to have no direct correlation with each other. If the theoretical predictions are correct, i.e., high carrier concentrations and Co doping are necessary for realizing carrier-mediated DMS24,25, then the fluctuation of Co must be suppressed before one can obtain carrier mediated DMS in ZnO:Co. In samples with x < 0.2, Co is soluble in ZnO, bandgap increases with Co content increase and carrier localization occurs. One of the possible mechanisms responsible for the absence of carrier-mediated ferromagnetism is attributed to the localization of carriers at non-Co-rich regions. 176 Chapter Origin of ferromagnetism from transport studies Structurally, results have shown that Co is soluble in the ZnO host matrix and that it is in the form of Co2+ ions, and also through transport measurements, the presence of strong s,p-d interaction and carrier localization in these lightly co-doped samples has been shown, but eventually there is no correlation between transport and ferromagnetic properties. The ferromagnetic properties of these samples should be intrinsic in behaviour, but they not exhibit carrier-mediated ferromagnetism. Origin of ferromagnetism could be from Co2+ ions or multi-phases present, however the mechanism of ferromagnetism could not be determined, although it does have certain characteristic of BMPs. 6.8.2 x = 0.25, onset of secondary phase formation Onset of secondary phase formation can be observed clearly from structural, magnetic and also transport properties. From structural studies, a drop in ZnO (002) peak intensity in XRD has been observed, indicating a change in film structure. From the 2θ scans, a change in number of peaks was also observed, with shoulder and broad peaks appearing. From XRD, it is difficult to pinpoint the secondary phase as its peaks is very near to other Zn,Co,O compounds. TEM imaging is also not able to determine specifically the structure of the secondary phase due to the complicated diffraction pattern. However, as can be seen clearly from cross-sectional images, the film changes from a homogeneous to an inhomogeneous film. Thus from XRD and TEM studies, a change of film structure from a homogeneous to an inhomogeneous film can be confirmed. From EELS, XPS studies have been used to determine if Co exists as a Co2+ ion. Presence of Co0 or Co3+ is not observed for these films, eliminating the existence of Co clusters or ZnCo2O4. It is possible that can Co exist as Co2+ in Co-doped ZnO 177 Chapter Origin of ferromagnetism from transport studies and/or Zn incorporated CoO. Since the film is still consists Co2+ in Co-doped ZnO, it is not peculiar that the characteristic d-d transitions in optical transmission studies are still observed. However, when the data for the red-shift peak has been analysed, the data fit well if the wavelength of CoO is used. This can be a lead to believe that the film consists of Zn incorporated CoO and not only Co-doped ZnO. Next in magnetic properties of these films, the onset of secondary phase formation can also be observed clearly. A sharp increase in magnetization, coercivity and also anisotropy indicates a different source of ferromagnetism in these films. The source of ferromagnetism could be from CoO as they can be ferromagnetic when they exist as nanoparticles. No doubt Co-doped ZnO may exist in the films, its ferromagnetic signal is very much drowned by the large magnetisation from the secondary phase. From transport studies, in particular MR studies, a hysteresis loop at low field has been observed. The positive MR is persistent up to 50 K, indicating the presence of presence of s-d interactions. However, above 50 K, a dominating negative MR due to extrinsic ferromagnetic origins was observed. In Hall effect studies, a hysteresis was also observed, but this hysteresis is not likely due to intrinsic ferromagnetism. AHE only become significant when Co content increases past the onset of secondary phase formation, indicating possibly percolated CoO networks becoming the dominant carrier transport channel. Carrier localization is also weakened when x increases further (>0.25) due to the evolution of Co-rich regions into CoO phases. Also observed was the increase of resistivity and ZBA at this composition, which could be used as in indication of secondary phase formation in further studies. 178 Chapter Origin of ferromagnetism from transport studies Thus, gathering all properties above, it can deduced that these characters observed are due to the presence of ferromagnetic secondary phases, likely in the form of CoO network embedded in Co-doped ZnO matrix. Ferromagnetic properties are definitely extrinsic in character and not likely to exhibit carrier-mediated ferromagnetism of DMSs. 6.8.3 x > 0.25, secondary phases Again from structural studies, from TEM images and diffraction patterns, it can be seen clearly that secondary phases are present in these films. However, from XRD scans, it is not a very clear picture as the ZnO and Co peaks are very close to that of other Zn,Co,O peaks. From XRD peaks, as Co composition increases, CoO phases start to disappear and give way to formation of Co nanoclusters. Columnar structure is observed from TEM cross-sectional images, likely consisting of Co-doped ZnO and Zn-incorporated CoO network. From HRTEM, nanosized secondary phases has also been observed, likely to be Co, CoO and/or ZnCo2O4 nanoclusters. Although EELS shows that Co is in +2 valence state, this valence state is due to the host Co-doped ZnO matrix and also from the CoO network. Valence states of and +3 for Co and ZnCo2O4 nanoclusters are not detected likely due to their minute amount. From XPS, the possibility of presence of Co clusters is reduced, but the presence of ZnCo2O4 cannot be ruled out as a trivalent Co oxide is detected. From structural properties, it is a tricky task to determine the secondary phases present due to their minute amount and also the sensitivity of equipment used. However, it can be confirmed that these films are inhomogeneous in nature and have a columnar structure, likely made up of a host Co-doped ZnO matrix and a Zn-incorporated CoO network, with presence of nanoclusters. 179 Chapter Origin of ferromagnetism from transport studies Since from structural studies there are very clear proofs that secondary phases are present, it is not surprising that they exhibit very different magnetic properties from the samples with x < 0.2. These films have much higher magnetic saturation and coercivity, making these secondary phases ferromagnetic in nature. A perpendicular anisotropy is also observed for these films, most likely due to the columnar structure determined from TEM cross-sectional images, and not intrinsic properties of Co2+. From MCD studies, these films have secondary phases as origin of ferromagnetism and are also inhomogeneously ferromagnetic. The MCD M-H curves determined depend very much on photon energy, with the curves at 2.92 eV in particular corresponding well to SQUID and AHE results. This can conclude that CoO is the main ferromagnetic contributing source for these films, but there is still other origins of ferromagnetic. Turning the attention now to transport properties, from MR measurements, a very distinctive negative MR has been observed, which is relatively insensitive to temperature for these films. The MR curves with a hysteresis curve obtained in these samples are the typical curve shapes of granular-like ferromagnetic material, another indication of presence of ferromagnetic clusters in these films. As presence of ferromagnetic clusters is confirmed in these films, AHE is predictably observed in these films. Differential conductance curves have given us an indication from the U-shaped curves that transportation is across grain boundaries, in the form of tunnel junction, nanoscale heterojunctions or schottky junctions. As secondary phases become percolated with Co composition increase, resistivity of these films decreases. As Co 180 Chapter Origin of ferromagnetism from transport studies content increases, carrier localization is no longer dominant in these films, thus there is not much change of carrier concentration with temperature. To further the studies of granular nature of these films, spin polarization studies were carried out. As nanoclusters are expected to be present in these films, they can thus affect the transport properties between the ZnO:Co-Nb junction. Results have shown that it is possible that the nanoclusters on the film surface form Nb/Zn1xCoxO/Co point contacts, resulting in observation of superconducting Nb DOS in the dI/dV curves of these samples. The subgap conductance also decreases in a superconductor/ferromagnet junction in the presence of a magnetic field. With the application of an in-plane magnetic field, the magnetisation of the nanoclusters are aligned, leading to a suppression of crossed AR. Crossed AR can also be suppressed as temperature increases. Crossed AR effect as observed in the samples are weak due to the relatively large spacing among the Co clusters. From the results of all the studies done above, these films are inhomogeneous structurally and magnetically. They are made up of Co-doped ZnO host matrix with an electrically percolated columnar structure of Zn-incorporated CoO, with presence of nanoclusters, likely in the form of Co, CoO and/or ZnCo2O4. The main sources of ferromagnetism for these films are predicted to be from CoO, but this is not its only source of ferromagnetism. As these films are inhomogeneous in nature, it is very difficult to pinpoint every source of ferromagnetism. Results, in transport properties, have also shown the granular nature of these films. 6.9 Conclusions Analysing results from transport studies, it is now possible to correlate all structural, magnetic and transport properties. From MR studies, weak localisation for 181 Chapter Origin of ferromagnetism from transport studies samples, with x < 0.25, were picked up. Positive MR from s,p-d interaction were observed in these samples when T < 50 K. As Co content increases, negative MR with hysteresis, caused by extrinsic ferromagnetism due to secondary phases, is observed. AHE often used as an indication of intrinsic ferromagnetism is not applicable in the study. Observation of AHE coincides with the onset of secondary phase formation. We would have expected AHE to be present for the lightly doped samples, but the phenomenon is not observed. Correlating results from SQUID, Hall effect and MCD, a conclusion that co-doped sample Co32W (x = 0.29) consists of Znincorporated CoO can be drawn. Further transport studies show samples with x < 0.2 have a characteristic “V” shape in differential conductance curve, compared to “U” shape when x > 0.2. Sharp peak in resistivity and ZBA versus Co content curves again indicate the onset of secondary phase formation when x = 0.25. Temperature dependence of carrier concentration shows that samples indeed exhibit localization at low temperature when x < 0.2. It is deduced that it might have caused carrier-mediated ferromagnetism to be absent in these samples. For samples with x > 0.25, carrier localization is insignificant and it is very clear that ferromagnetism is extrinsic in nature. Studies of transport properties between a superconductor / ferromagnet junction shows very clearly the granular nature of ferromagnetism in the highly co-doped samples, no observed in previous techniques used. It is thus expected that there is definitely presence of ferromagnet clusters on the film surface. Summarising results from structural, magnetic and transport studies, the nature of the samples studied can be divided into the following: 182 Chapter Origin of ferromagnetism from transport studies 1) x < 0.2, Co is soluble in ZnO, forming single phase films with no impurity phases present, exhibiting s,p-d interaction, carrier localization and noncarrier-mediated ferromagnetism, 2) x = 0.25, onset of secondary phase formation occurs, however Co still exists in the +2 valence state. Structure of the films is made up of columnar structure of Co-doped ZnO film with electrical percolated network of Znincorporated CoO. The ferromagnetic nature of this material is extrinsic in nature, 3) x 0.3, inhomogeneous Co-doped ZnO films with granular nature, in a network of Zn-incorporated CoO, ZnCo2O4 and Co cluster formation. The ferromagnetic nature of this material is extrinsic in nature. References: G. D. J. Smit, S. Rogge, and T. M. Klapwijk, “Scaling of nano-Schottky-diodes”, Appl. Phys. Lett. 81, 3852 (2002). H. Ohno, A. Shen, F. Matsukura, A. Oiwa, A. Endo, S. Katsumoto, and Y. Iye, “(Ga,Mn)As: A new diluted magnetic semiconductor based on GaAs”, Appl. Phys. Lett. 69, 363 (1996). H. Toyosaki, T. Fukumura, Y. Yasuhiro, Y. Yamada, K. Nakajima, T. Chikyow, T. Hasegawa, H. Koinuma, and M. Kawasaki, “Anomalous Hall effect governed by electron doping in a room-temperature transparent ferromagnetic semiconductor”, Nature Material 3, 221 (2004). 183 Chapter Origin of ferromagnetism from transport studies S. R. Shinde, S. B. Ogale, J. S. Higgins, H. Zheng, A. J. Millis, V. N. Kulkarni, R. Ramesh, R. L. Greene, and T. Venkatesan, “Co-occurrence of superparamagnetism and anomalous Hall effect in highly reduced cobalt-doped rutile TiO2–δ films”, Phys. Rev Lett. 92, 166601 (2004). A.V. Vedyaev, A. B. Granovski, A. V. Kalitsov, and F. Brouers, “Anomalous Hall effect in granular alloys”, JETP. 85, 1204 (1997). H. Sato, Y. Kobayashi, Y. Aoki, and Y. Yamamoto, “Transport properties in granular Co-Ag alloys”, J. Phys.: Condens. Matter 7, 7053 (1995). B. A. Aronzon, D. Yu. Kovalev, A. N. Lagar' kov, E. Z. Meilikhov, V. V. Ryl' kov, M. A. Sedova, N. Negre, M. Goiran, and J. Leotin, “Anomalous Hall effect in granular Fe/SiO2 films in the tunneling-conduction regime”, JETP Lett. 70, 90 (1999). A. A. Burkov and L. Balents, “Anomalous Hall Effect in ferromagnetic semiconductors in the hopping transport regime”, Phys. Rev. Lett. 91, 057202 (2003). G. W. Jr. Pratt and R. Coelho, “Optical absorption of CoO and MnO above and below the Néel temperature”, Phys. Rev. 116, 281 (1959). 10 Z. W. Jin, K. Hasegawa, T. Fukumura, Y.Z. Yoo, T. Hasegawa, H. Koinuma, and M. Kawasaki, “Magnetoresistance of 3d transition-metal-doped epitaxial ZnO thin films”, Physica E 10, 256 (2001). 11 Z. W. Jin, T. Fukumura, K. Hasegawa, Y.-Z. Yoo, K. Ando, T. Sekiguchi, P. Ahmet, T. Chikyow, T. Hasegawa, H. Koinuma, and M. Kawasaki, “Optical and electrical properties of Co-doped epitaxial ZnO films”, J. of Crystal Growth 237–239, 548 (2002). 12 J. H. Kim, H. Kim, D. Kim, Y. E. Ihm, and W. K. Choo, “Magnetoresistance in laser-deposited Zn1-xCoxO thin films”, Physica B 327, 304 (2003). 184 Chapter Origin of ferromagnetism from transport studies 13 T. Andrearczyk, J. Jaroszynski, G. Grabecki, T. Dietl, T. Fukumura, and M. Kawasaki, “Spin-related magnetoresistance of n-type ZnO:Al and Zn1–xMnxO:Al thin films”, Phys. Rev. B 72 121309(R) (2005). 14 P. Stamenov, M. Venkatesan, L. S. Dorneles, D. Maude, and J. M. D. Coey, “Magnetoresistance of Co-doped ZnO thin films”, J. Appl. Phys. 99, 08M124 (2006). 15 of T. Dietl, T. Andrearczyk, A. Lipi ska, M. Kiecana, M. Tay, and Y. H. Wu, “Origin ferromagnetism in Zn1-xCoxO from magnetization and spin-dependent magnetoresistance measurements”, Phys. Rev. B 76, 155312 (2007). 16 G. J. Strijkers, Y. Ji, F. Y. Yang, C. L. Chien, and J. M. Byers, “Andreev reflections at metal/superconductor point contacts: Measurement and analysis”, Phys. Rev. B 63, 104510 (2001). 17 R. Mélin and S. Peysson, “Crossed Andreev reflection at ferromagnetic domain walls”, Phys. Rev. B 68, 174515 (2003). 18 N. M. Chtchelkatchev and I. S. Burmistrov, “Andreev conductance of a domain wall”, Phys. Rev. B 68, 140501 (2003). 19 A. Frydman, and R. C. Dynes, “Disorder-induced Andreev reflections in granular metals”, Phys. Rev. B 59, 8432 (1999). 20 G. Deutscher, “Crossed Andreev reflections”, J. Superconductivity 15, 43 (2002). 21 D. Beckmann, H. B. Weber, and H. V. Lohneysen, “Evidence for crossed Andreev reflection in superconductor-ferromagnet hybrid structures”, Phys. Rev. Lett. 93, 197003 (2004). 22 P. Sati, R. Hayn, R. Kuzian, S. Regnier, S. Schafer, A. Stepanov, C. Morhain, C. Deparis, M. Laugt, M. Goiran, and Z. Golacki, “Magnetic anisotropy of Co2+ as signature of intrinsic ferromagnetism in ZnO:Co”, Phys. Rev. Lett. 96, 017203 (2006). 185 Chapter Origin of ferromagnetism from transport studies 23 J. M. D. Coey, M. Venkatesan, and C. B. Fitzgerald, “Donor impurity band exchange in dilute ferromagnetic oxides”, Nat. Mater. 4, 173 (2005). 24 K. Sato and H. Katayama-Yoshida, “Stabilization of ferromagnetic states by electron doping in Fe-, Co- or Ni-doped ZnO”, Jpn. J. Appl. Phys. 40, L334 (2001). 25 E. C. Lee and K. J. Chang, “Ferromagnetic versus antiferromagnetic interaction in Co-doped ZnO”, Phys. Rev. B 69, 085205 (2004). 186 [...]... studies -3 7 .6 10 × (a) -3 7.2×10 × -3 6. 8×10 × -3 6. 4×10 × -3 6. 0×10 × -3 7.2×10 (b) -3 6. 8×10 -3 6. 4×10 -3 6. 0×10 Resistance (Ohm) 1 .66 E+022 1 .66 ×10 1 .64 E+022 1 .64 ×10 (c) 1 .62 E+022 1 .62 ×10 1 .60 E+022 1 .60 ×10 1.58×10 1.58E+022 1. 56 102 1.56E+02 -0.04 -0.02 0 0.02 0.04 Field (T) Figure 6- 13 dI/dV curves as a function of bias voltage for sample G, Co30W (Zn0 .63 Co0.27O) at (a) 1 .6 – 10 K, (b) 1 .6 K at various... clusters were smaller than the coherence length of the Cooper pairs.21 The CAR was absent in Sample G However, as shown in Figs 6- 14 and 6- 15, the CAR had been seen clearly in samples J and L due to the further increase of density of FM clusters and thus decrease of the distance among the clusters The common feature of the dI/dV curves shown in Fig 6- 14 and Fig 6- 15 was that the conductance exhibited a peak... the conduction band into spin-up and spin-down sub-bands The spin splitting of conduction band enhanced electron- 160 Chapter 6 Origin of ferromagnetism from transport studies electron interactions in a disordered system which leads to a positive MR.11,13 Detailed simulation had been carried out to understand the MR behaviour of lightly doped ZnO:Co (sample 3W and 8W) based on the effect of the field-induced... suppressed by the application of a magnetic field of about 25mT This was because the applied field induced an in-plane alignment of the Co clusters, resulting in a suppression of CAR The increase of zero-bias resistance was also reflected in the magnetoresistance curves shown in Figs 6- 14(c) and 6- 15 (c), respectively, obtained with an applied current of 0.1 mA 169 Chapter 6 Origin of ferromagnetism from... dominant and Co clusters started to appear 161 Chapter 6 Origin of ferromagnetism from transport studies Figure 6- 8 MR of (a) Al -doped ZnO and various co -doped samples, with x values indicated, from 4.2 – 50 K as a function of magnetic field, applied perpendicular to sample plane; Inset in (e) shows hysteresis in MR behaviour for sample Co25W (Zn0.75Co0.25O) for applied field of -1 to 1 T at 50 K 162 Chapter... sample (Fig 6- 12(b)) This suggested that the sample was not ferromagnetic globally, though it might contain ferromagnetic clusters 167 Chapter 6 Origin of ferromagnetism from transport studies 3.2×10 -3 2.8×10 -3 2.4×10 -3 2.0×10 -3 1 .6 10 -3 -3 3.1×10 -3 2 .6 10 -3 2.1×10 -3 1 .6 10 Figure 6- 12 dI/dV curves as a function of bias voltage for sample D as determined at (a) 1 .6 – 10 K, (b) 1 .6 K at various... Co2+ ions of which not all of them contributed to ferromagnetism which would be picked up by SQUID and Hall effect For example, those inside the antiferromagentic clusters contributed to d-d transitions, but they did not contributed to SQUID and Hall signals 157 Chapter 6 Origin of ferromagnetism from transport studies Figure 6- 6 SQUID, AHE and MCD (at various energies) M-H curves for co -doped sample... sample Co25W (Zn0.75Co0.25O) for applied field of -1 to 1 T at 50 K 162 Chapter 6 Origin of ferromagnetism from transport studies Figure 6- 9 MR dependence on temperature of (a) Al -doped ZnO and various co -doped samples, with x values indicated, from 1 .6 – 300 K The temperature dependence of MR for various samples is shown in Fig 6- 9 As could be seen from the figure, significant changes in MR occured when... minute amount and also the sensitivity of equipment used However, it can be confirmed that these films are inhomogeneous in nature and have a columnar structure, likely made up of a host Co -doped ZnO matrix and a Zn-incorporated CoO network, with presence of nanoclusters 179 Chapter 6 Origin of ferromagnetism from transport studies Since from structural studies there are very clear proofs that secondary... of all the studies done above, these films are inhomogeneous structurally and magnetically They are made up of Co -doped ZnO host matrix with an electrically percolated columnar structure of Zn-incorporated CoO, with presence of nanoclusters, likely in the form of Co, CoO and/ or ZnCo2O4 The main sources of ferromagnetism for these films are predicted to be from CoO, but this is not its only source of . 4 1.2 1.4 1 .6 1.8 2.0 -3 -2 -1 0 1 2 3 4.4 4.5 4 .6 -2 -1 0 1 2 3.55 3 .60 3 .65 3.70 -2 -1 0 1 2 1.958 1. 960 1. 962 1. 964 1. 966 -6 -4 -2 0 2 4 6 1.15 1.20 1.25 -2 -1 0 1 2 2 .6 2.7 2.8 2.9 30K 20K 50K 30K 20K . localization in Co -doped samples. Chapter 6 Origin of ferromagnetism from transport studies 151 -2 -1 0 1 2 7.58 7 .60 7 .62 7 .64 -3 -2 -1 0 1 2 3 2.45 2.50 2.55 2 .60 2 .65 -4 -2 0 2 4 1.2 1.4 1 .6 1.8 2.0 -3. applied field of -1 to 1 T at 50 K. Chapter 6 Origin of ferromagnetism from transport studies 163 Figure 6- 9 MR dependence on temperature of (a) Al -doped ZnO and various co -doped samples,