Three main theories have been proposed to explain the ferromagnetic properties of Zn1-xCoxO. The first rests on the hypothesis that Zn interstitial defects are present in the material, i.e. that a small percentage of Zn atoms are not located in their theoretical crystallographic positions, but are lodged in the interstices of the crystal lattice. This would create the conditions for magnetic behavior. The second hypothesis attributes the magnetic properties of the compound to the presence of nanoclusters of metallic Co that would aggregate at the time of crystal growth. In this second hypothesis, the material would lack intrinsic homogeneity and its ferromagnetic properties could not be coupled effectively to its semiconductor properties. A third theory states that the properties of the material would come from oxygen vacancies, i.e. that some sites on the crystal that should be occupied by oxygen atoms would, in fact, be empty.
Based on XANES and EXAFS analysis, a team of researchers operating on the SOLEIL beamlines SIRIUS and SAMBA and CNR researchers in Italy have shown that the magnetization in Zn1-xCoxO is directly proportional to the number of oxygen vacancies present in the material, a quantity which depends on the Co concentration and its production conditions. In addition, they showed that a strongly magnetic compound to which oxygen atoms are added (by annealing under saturated oxygen concentrations) gradually loses its properties as and when the gaps are filled with this element. Note that it is difficult to show the process the other way round, i.e. remove oxygen atoms in the compound after development and measure the increase in its magnetization, because the oxygen atoms are strongly bound in the crystal. It is possible, however, to modify the material by heating and to add oxygen in known quantities.
This is the method that was used in this study.
| In figure 1 (measuring magnetization as a function of the magnetic field applied), it can be seen that the sample with a Co concentration of 4% (red curve, x = 0.04) is more strongly magnetic than the other samples. In figure 2, the XANES study of the different samples shows a spectrum (black curve) with two peaks in the region just after the absorption threshold (indicated by the vertical lines and the numbers 1 and 2) for the least magnetic sample (x = 0.06). These two peaks are separated by a rather obvious dip. On the other hand, in the case of the most magnetic sample (x = 0.04, red curve) the dip has been filled in and there is no distinction between the two peaks. This difference in structure near the absorption threshold reveals two different electron densities, characteristic of different structures of a crystal. The blue curve was recorded when the most magnetic material was degraded by the addition of oxygen [x = 0.04 (a)]. It can be seen that the shape of the blue curve becomes similar to that of the least magnetic material (the two peaks reappear). In parallel, the magnetic moment of the x = 0.04 sample (a) decreases in relation to the value that the same sample had before annealing. All this supports the hypothesis that the oxygen vacancies do govern the properties of ferromagnetic Zn1-xCoxO. | Figure 1
 |
The three curves, purple, black and red, at the top of the figure are simulations of 0, 10 and 30% ,respectively, of oxygen vacancies per cobalt atom present in the material. As a result, the hypothesis that best fits the experimental data for the most magnetic sample would be a Zn1-xCoxO alloy comprising 30% of oxygen vacancies for each Co atom (present in turn at a level of 4% in the alloy). In parallel, our linear dichroism studies of the material were used to determine the local structure of the "oxygen vacancy" defect and its crystallographic orientation. The sites where oxygen is absent are not randomly distributed: they must be close to a cobalt atom with the cobalt – vacancy direction along the c axis of the hexagonal structure of Zn1-xCoxO (Fig. 3)
| Figure 2
 | Figure 3
 |
Reference :
Ciatto, G., Di Trolio, A., Fonda, E., Alippi, P., Testa, A. M., & Amore Bonapasta, A.
Evidence of Cobalt-Vacancy Complexes in Zn1-xCoxO Dilute Magnetic Semiconductors.
Physical Review Letters, 2011, 107(12): art.n° 127206
