Growth and characterisation of cobalt doped zinc oxide 4

For lightly doped samples, x < 0.2, the XRD spectra showed peaks of ZnO 002 and those associated with the substrate; no other peaks due to secondary phases were observed.. can be seen fr

CHAPTER 4

STRUCTURAL CHARACTERIZATION AND

CHEMICAL VALENCY ANALYSIS

4.1 Introduction

In this chapter, the structural properties of Co-doped ZnO:Al characterized by XRD and TEM are discussed The results of chemical valence analysis using XPS and optical transmission spectroscopy are also presented Throughout this thesis, the Co composition is given as a nominal value which is determined by XPS As this technique only has an accuracy of about 5%, caution need to be taken when comparing the properties of the films obtained in this work with those reported in literature It is noted that the Co composition determined in this work tend to be higher than those reported in literature This is particularly true when the MR data of samples with similar nominal Co compositions are being compared

peaks were detected in the θ−2θ scan of XRD patterns in the range of 30-90o This indicates that the films prepared were well textured, with the c-axis pointing in the normal direction of the substrate The lattice parameter determined for the Al-doped ZnO film is 3.361Å (a) and 5.488Å (c) This is comparable to pure ZnO films with values of 3.250Å (a) and 5.207Å (c) As shown in Fig 4-1 is a typical XRD pattern of Al-doped ZnO The FWHM for Al-doped ZnO is about 0.77o which is comparable to the results reported in literature.1,2

In all samples prepared, ZnO was doped with Al For the co-doped samples, the

Co composition, x, was controlled by varying the sputtering power of the Co target In the specific setup, the minimum controllable power was about 3 W, which gives a Co composition of about 5 at.% as determined by XPS As for δ-doped samples, Co was doped into the Al-doped ZnO films digitally, at a Co sputtering power of 10 W for a specific duration of time In general, the co-doped samples have a better structural

quality than their δ -doped counterparts, as revealed from the XRD results (not shown here)

Generally, the crystalline quality of the films degraded as the Co doping amount was increased As observed from Fig.4-2, the FWHM of ZnO (002) peak decreased with an increase in Co concentration initially It began to increase as the Co composition exceeded 0.2 and decreased again after reaching a maximum at about 0.3 The initial decrease of FWHM was somewhat unexpected It indicated that Co was soluble in the ZnO host matrix with Co composition of less than 0.2 The decrease of FWHM indicated that the ZnO:Co film had a better quality than ZnO when both were grown on the sapphire substrates under the specific conditions used in this study This could be attributed to either the difference in thermodynamic properties of ZnO and ZnO:Co or the slightly smaller diameter of the Co2+ ions The upturn at Co composition of 0.2 was due to the formation of secondary phases, as would be discussed later and also in following chapters As the Co sputtering power was further increased, the film becomes increasingly less textured which results in a peak of FWHM when the Co content was around 0.3 A further increase of Co composition beyond this value lead to precipitation of Co nanoparticles, which results in an improvement of the crystallinity of the ZnO and ZnO:Co phases

Figure 4-2 ZnO (002) XRD FWHM versus cobalt composition for co-sputter films

Figs 4-3 ((a) –(c)) show the XRD patterns of ZnO:Co with different Co compositions in the scan range of 2 = 30 - 50o The data displayed was divided into three different ranges of x values, i.e., (a) x < 0.2, (b) 0.2 < x < 0.3 and (c) x 0 3 For lightly doped samples, x < 0.2, the XRD spectra showed peaks of ZnO (002) and those associated with the substrate; no other peaks due to secondary phases were observed These results, in combination with the TEM results that would be discussed later, suggest that the Co was soluble in the ZnO host matrix and no impurity phases were present at x < 0.2

Figure 4-3 XRD scan of co-doped ZnO films.

As the ionic radius of Co2+ is about 96% of that of Zn2+, the in-plane lattice constant of relaxed ZnO:Co film was expected to decrease when Zn atoms are replaced

by Co atoms, leading to an increase of out-of-plane lattice constant due its large Poisson’s ratio This explained why the (002) peak of ZnO:Co shifts to the lower angle side of the original (002) peak of ZnO, as shown in Fig 4-3 (a) This was also an indication that within this composition range, the Co atoms were soluble in the ZnO host matrix The solubility limit of Co in ZnO was found to be about 0.25 in literature.3,4 Shown in Fig 4-4 is the peak position at ~ 34o plotted as a function of the

Co composition It could be seen that up to 0.2 of Co content, there was not much changes in the peak position, which was expected as the films are composed of single

phase co-doped ZnO As the Co content was further increased towards 0.25, a left shift

in peak position was observed, possibly due to the formation of Co-rich ZnO:Co A further increase of Co beyond 0.3 leads to the recovery of the peak to positions near to those for films with x < 0.2, indicating the onset of phase segregation

33.4 33.6 33.8 34.0 34.2 34.4 34.6

Figure 4-4 XRD peak position (around 34 o ) versus Co composition

For samples with x > 0.2, as shown in Fig 4-3 (b), in addition to the (002) ZnO peak, new peaks appear at 2 31.8o, 35.8-36o, 40.6, 42.4 and 44.5o, respectively The assignment of these peaks was nontrivial because Co might exist in the material in question in at least five different forms: ZnO:Co, CoO, Co3O4, ZnCo2O4 and Co The peak at around 31.8o and 36.253o could be assigned to ZnO (100) and (101), respectively The CoO nanoparticles might exist in both cubic and hexagonal structures in the ZnO:Co host matrix Therefore, the peak around 36o may be assigned

to either one of the following peaks due to secondary phases: CoO (111) at 36.493o for cubic CoO, CoO (101) at 36.3o for hexagonal CoO, ZnCo2O4 (311) at 36.803o and

Co3O4 (311) at 36.853o Also, the peak at 44.5o was near peak positions of CoO (200)

FCC at 40.6o and 42.4o, Co (111) at 44.217o, ZnCo2O4 (400) at 44.74o and Co3O4 (400)

of peak position to lower angles could be a result of Zn incorporation into the CoO matrix As the sputtering power was further increased, in sample F (x = 0.25), the CoO (111) peak disappeared and Co (111) peak appeared It should be noted that the lattice constants of hexagonal CoO are very similar to those of ZnO; thus it was difficult to differentiate between the two using XRD, especially if CoO grew pseudmorphically inside ZnO Pole figure measurements (Fig 4-5) had also been carried out for this sample and it could be seen from the results that the film was epitaxially grown and (002) textured The pole diagram showed that this film was still dominantly single crystalline, suggesting that the films with Co content less than 0.25 were indeed single crystalline films with a good texture

Figure 4-5 Pole figure diagram for sample F (Zn 0.75 Co 0.25 O).

As shown in Fig 4-3 (c), as the Co concentration increased further, peaks at 31.8o and 36.1o start to disappear, with the appearance of peaks around 44.5o The peak

at 44.5o was due to Co clusters, though again the existence of other secondary phases such as ZnCo2O4 and Co3O4 could not be excluded The formation of Co clusters was more probable because the formation of ZnCo2O4 and Co3O4 needed an oxygen-rich environment instead of more Co atoms

Before ending this section, there is a need to make the remark that any attempt to find peaks exactly at the same positions of pure ZnO or CoO is meaningless because ZnO would incorporate Co and, vice versa, CoO would contain Zn

4.3 TEM observations

The samples with different Co compositions had been examined by HRTEM The samples with Co compositions lower than the solution limit were found to be homogeneously grown on the substrate As an example, Fig 4-6 shows the cross-sectional TEM image, EDS mapping and diffraction images of the Co15W sample As

can be seen from the TEM image and diffraction pattern, the film was homogeneous and there were no detectable precipitates of Co

100 nm

Figure 4-6 TEM results of co-doped Co15W (Zn 0.80 Co 0.20 O) sample; (a) Cross-sectional TEM image; (b) electron diffraction pattern of the same region; (c)EDS mapping of Al, Co, O and Zn of films, with direction of film growth as indicated

In Fig 4.6 (c), the film’s diffraction pattern showed a single crystalline phase with no impurity spots From XRD studies above, peaks had been observed in the vicinity where CoO (200) is expected This could be due to the fact that the XRD pattern was from a large area of sample, whereas the TEM results were from a very small spot

100 nm

When increasing the Co content to 0.24, it was observed that, similar to the Co15W sample, the Co20W sample also showed the absence of Co-agglomeration or precipitation (Fig.4-7(a)) The film, in general, was still homogeneous, though some external spots were detected in the diffraction pattern, as shown in Fig 4-7 (b) These spots could be due to CoO (200), FCC Co (111), HCP Co (002) or even ZnO (101), as observed in the XRD patterns The homogeneity of the film was found to degrade significantly after the Co content exceeds 0.25, the onset composition of secondary phase formation As a result, the film quality of Co25W was much poorer as compared

to the above-mentioned two samples Also, the TEM cross-sectional image (Fig 4-8 (a)) showed an inhomogeneous film In Fig 4-8 (b), the diffraction spots of secondary phases were seen to increase in quantity

Fig 4-9 (a) shows the TEM images of the Co 32W sample, which illustrated clearly the inhomogeneous nature of the sample Columnar growth, similar to that

reported by Schaedler et al., 5 was observed, together with nanosized secondary phases, as shown by a dark particle in Fig 4-9(b) For the corresponding δ-doped sample, the film structure was found to be very irregular, as shown in Fig 4-10 (a), and it could be observed that Co clusters started to form throughout the film Through analysis of fast-Fourier transform pattern and also EELS measurements on that particular particle shows that it exists as FCC Co

(b) (a)

Al 2 O 3

ZnO : Co

ZnO (002)

Extra spots

Figure 4-7 TEM results of co-doped Co20W (Zn 0.76 Co 0.24 O) sample; (a) Cross-sectional TEM image; (b) electron diffraction pattern of the same region; (c)EDS mapping of Al, Co, O and Zn of films, with direction of film growth as indicated

Figure 4-8 TEM results of co-doped Co25W (Zn 0.75 Co 0.25 O) sample; (a) Cross-sectional TEM image; (b) electron diffraction pattern of the same region

Figure 4-9 TEM results of co-doped Co32W (Zn0.71Co0.29O) sample; (a) Cross-sectional TEM image; (b) HRTEM image of a selected region; (c) electron diffraction pattern of the same region

Figure 4-10 TEM results of δ-doped Co98s ([(ZnO:Al (2.38 nm)/Co (1.0 nm)]× 60) sample; (a) sectional TEM image; (b) HRTEM image of a selected region; (c) electron diffraction pattern of the same region

The EELS analysis was carried out on co-doped Co32W and δ-doped Co98s samples The valence of Co detected from 4 different locations in Fig 4-9 (a) turned out all to be 2+ (Fig 4-11(a)) The results showed that the film consists of mainly Co-incorporated ZnO and/or Zn-incorporated CoO because the valence state of Co in ZnCo2O4 and Co3O4 was 3+ and 4+, and that in Co clusters was 0, respectively In comparison, the corresponding δ-doped Co98s sample, showed a variation of 0 and 2+ valence state (Fig 4-11 (b)), at different positions of the sample These results along with the electron and x-ray diffraction data confirmed that the δ-doped samples contain both substitutional Co and Co clusters, whereas the co-doped samples had Co in valence +2 state, in the form of either Co2+ ions doped in ZnO or in the form of Co2+ in CoO.

Figure 4-11 EELS spectra of L 3 /L 2 of Co for (a) co-doped Co 32W (Zn 0.71 Co 0.29 O) at four different positions and (b) δ-doped Co 98s samples

Selected area electron diffraction of a particle in the above two heavily doped samples, as shown in Fig 4-9 (c) and 4-10 (c), respectively, confirmed the presence of secondary phases Detailed study of the diffraction pattern and HRTEM images showed the presence of hexagonal closed packed (HCP) Co, face center cubic (FCC) Co, hexagonal CoO and also ZnCo2O4 phase, distributed in the ZnO matrix of the co-doped sample This did not contradict the EELS result which suggested that the sample was dominantly composed of phases containing Co2+ ions, which could either

Co-be ZnO:Co or CoO:Zn Before ending this section, it should Co-be stressed again that the TEM observation could only provide information about the material in a very localized region; the results don’t necessarily reflect the macroscopic properties of the materials detected by other techniques

4.3 XPS, AES and UPS studies

The XPS was used to analyze the chemical environment experienced by Co in lightly doped samples Fig.4-12 shows the XPS spectra of samples with different Co compositions (note that the spectra have not corrected for charge-shift; therefore, the O 1s peak was found at 531 eV) As could be seen from the figure, the Co 2p3/2 peak position was at 781.8 eV, while the splits of Co 2p3/2 and 2p1/2 were about 15.4 eV for the samples understudy As the split for Co metal was about 15.05 eV, one can rule out the existence of metallic Co clusters in the samples measured If Co was surrounded by oxygen, the split would be about 15.5 eV.3,6 The above results suggested again that samples with a Co composition less than 0.25 were dominantly ZnO:Co or CoO:Zn, while those at very high doping levels can possibly contain other secondary phases such as Co

From the AES spectra shown in Fig 4-13, a gradual appearance of a satellite feature at the high-energy shoulder at about 781 eV was observed as the Co content increased This could be due to electron-correlations of some Co ions, possibly attributed to some trivalent Co ion formation or lattice distortion in the outer surface

Photon Energy(eV)

a b c d

Figure 4- 13 XAS spectra recorded by LMM AES signal from photoelectrons for samples (a) Co10W (Zn 0.84 Co 0.16 O), (b) Co 20W (Zn 0.76 Co 0.24 O) , (c) Co 25W(Zn 0.75 Co 0.25 O) and (d) Co30W (Zn 0.73 Co 0.27 O) Satellite features at the shoulder are marked by a line

hν=60eV

Figure 4-14 Valence-band UPS spectra, photon energy h =60 eV for samples (a) Co 10W (Zn 0.84 Co 0.16 O), (b) Co 20W (Zn 0.76 Co 0.24 O), (c) Co 25W (Zn 0.75 Co 0.25 O) and (d) Co30W (Zn 0.73 Co 0.27 O) The component indicated by an arrow below E f about 0.92 eV is developing as Co content increases