Chapter Conclusions and Future Work CHAPTER CONCLUSIONS AND RECOMMENDATION FOR FUTURE WORK Since the first theoretical work of Dietl et. al. in 2000 and first experimental work of Ueda et. al. in 2001, considerable efforts have been devoted to the studies of ZnO-based magnetic semiconductors. Although the observation of above roomtemperature ferromagnetism has been reported frequently in literature by magnetometry measurements, so far there has been no report on correlated ferromagnetism in structural, magnetic, optical and electrical measurements. In this work, a systematic study of ZnO:Co thin films has been conducted on both co-doped and δ-doped samples. In the former case, the Co composition was varied systematically from x = to 0.33, while in the latter case, Co layers with thicknesses equivalent to the dopant concentration of co-doped samples were “digitally” doped into the host matrix. Through a systematic study of these samples, more insights and a deeper understanding of the origin of magnetic properties of Co-doped ZnO thin films with different cobalt compositions have been attained. The main contributions and results are summarized as follows: 1. Both the XRD and TEM studies revealed that ZnO:Co thin films grown on (0001) sapphire substrates with x < 0.2 were single phase and were textured with the caxis along the substrate normal direction. Phase separation took place gradually 187 Chapter Conclusions and Future Work from x = 0.2 to x = 0.25 beyond which secondary phases of various types as well as Co precipitates were observed. 2. The XPS and EELS studies had shown that Co existed dominantly in the 2+ valence state even when x was as high as 0.29 in the co-doped samples, whereas a mixture of and 2+ valence state had been observed for δ-doped samples with a same equivalent Co composition of about 0.29. These observations in combination with the XRD and TEM data suggested that, although Co clusters had been formed in heavily co-doped samples, the dominant phase of these samples was still in the form of Co-doped ZnO. 3. The formation of substitutional Co in the host matrix had also been “confirmed” by the observation of characteristic optical absorption lines due to tetrahedrally coordinated Co2+ ion in ZnO. 4. The SQUID measurement results showed that all the films were ferromagnetic up to room–temperature. For heavily doped samples, the M-H loops measured by the MCD were strongly dependent on the photon energy, suggesting that the films were inhomogeneous and consisted of ferromagnetic regions with different phases or origin. 5. The chemical and magnetic inhomogeneity of heavily doped samples had been revealed clearly by the electrical transport measurements which included differential conductance as a function of the bias voltage, MR and Hall effect. 6. The differential conductance measurement which was conducted for the first time for ZnO:Co in this study had proven to be a powerful technique in studying inhomogeneity, in particular, phase separation in DMSs. This technique was expected to be even more powerful once the materials were processed into nanowires so that the current could be confined in a narrow path. 188 Chapter Conclusions and Future Work 7. Using the MR measurements strong sp-d interactions in lightly and moderately doped samples had been observed. From the evolution of MR curves with the increase of Co composition, the onset composition for formation of secondary phases and Co clusters could be identified clearly. On the other hand, the shape of the MR curves for heavily doped samples resembled well those of granular materials. 8. Although all the samples appeared to be ferromagnetic in SQUID measurement, the AHE was only observed for highly doped samples. AHE in the films prepared were due to extrinsic ferromagnetism and not intrinsic ferromagnetism of a DMS system. A comparison of the hysteresis curves obtained by SQUID, MCD and AHE suggests that the dominant ferromagnetic phases in both lightly and moderately doped samples is Co-rich ZnO:Co. As this phase was dominantly antiferromagnetic, the ferromagnetic properties were presumably originating from the uncompensated spins. 9. As a preliminary study, electrical transport studies of ZnO:Co-Nb junctions had also been carried out. The results again confirmed the granular nature of the films with high doping concentrations. Based on the above-mentioned findings, the characteristics of Zn1–xCoxO thin films grown on (0001) sapphire substrates can be summarized as following: a) x < 0.2 The Co atoms were soluble in the ZnO host matrix, leading to a single phase film with the c-axis being oriented in the film normal direction. As the Co content increased near to 0.2, the system formed Co-rich ZnO:Co, which 189 Chapter Conclusions and Future Work acted as a seed to secondary phase formation. Weak ferromagnetism was observed in these films, but its origin could not be revealed using the magnetometry measurement alone. Strong sp-d interactions between the magnetic impurities and itinerant carriers and carrier localization were observed through electrical transport measurements; however, this material does not exhibit carrier-mediated ferromagnetism. The origin of ferromagnetism for this material was still not well understood, but could originate from uncompensated spins of antiferromagnetic clusters. b) x = 0.25 At this composition, the onset of secondary phase formation occured which leads to significant difference in properties of the material as compared to films with x < 0.2; these included the observation of secondary phases in structural studies, increase in magnetic moment, coercivity and resistivity, and observation of anomalous Hall effect. The secondary phases observed included both hexagonal and cubic CoO, ZnCo2O4, and Co3O4. The origin of ferromagnetism for this material was extrinsic in nature, due to ferromagnetic secondary phases. c) x > 0.25 With a further increase of Co composition, the secondary phases would expand to form electrically percolated networks, likely with the help of Co clusters. This drastically altered the electrical transport properties, leading to the appearance of AHE and granular-like MR curves. These films were also observed to be inhomogeneous both structurally and magnetically. The main ferromagnetic contributors for the films include Co clusters and uncompensated spins in antiferromagentic clusters like CoO and ZnCoO. The 190 Chapter Conclusions and Future Work granular nature of the films were also confirmed from the electrical transport properties of ZnO:Co-Nb junctions. Although in this work the experiments have been designed carefully and the study was carried out as systematically as possible, there are still many issues which remain unsolved. Among them, as a recommendation for the future work, the following are listed: 1. Carry out systematic studies of samples grown or post annealed in different gaseous environments (vacuum, H2, O2). This study will be able to determine if defects and carrier concentrations are dependent on growth and post-growth treatment, allowing better understanding on the origin of the magnetic properties of the ZnO:Co system. 2. Carry out point contact transport property studies of ZnO:Co-Nb junctions so as to determine if there is any polarization of carriers in lightly doped samples. 3. Grow the films homoepitaxially on ZnO substrates and study their structural, magnetic and electrical transport properties. 4. Fabricate nanowire samples and study their transport properties, particularly for the inhomogeneous samples. 191 . of co -doped samples were “digitally” doped into the host matrix. Through a systematic study of these samples, more insights and a deeper understanding of the origin of magnetic properties of. Chapter 7 Conclusions and Future Work 1 87 CHAPTER 7 CONCLUSIONS AND RECOMMENDATION FOR FUTURE WORK Since the first theoretical work of Dietl et. al. in 2000 and first experimental work of. and TEM data suggested that, although Co clusters had been formed in heavily co -doped samples, the dominant phase of these samples was still in the form of Co -doped ZnO. 3. The formation of