Nature of Native Defects in ZnO F. A. Selim, * M.H. Weber, D. Solodovnikov, and K.G. Lynn Department of Physics and Astronomy and Center for Materials Research, Washington State University, Pullman, Washington 99164-2711, USA (Received 9 April 2007; published 21 August 2007) This study revealed the nature of native defects and their roles in ZnO through positron annihilation and optical transmission measurements. It showed oxygen vacancies are the origin for the shift in the optical absorption band that causes the red or orange coloration. It also revealed experimental evidence that the donor nature of oxygen vacancy is 0:7 eV. In addition, this work showed the Zn interstitial was not the donor in the as-grown ZnO and supported recent calculations that predicted hydrogen in an oxygen vacancy forms multicenter bonds and acts as a shallow donor. DOI: 10.1103/PhysRevLett.99.085502 PACS numbers: 81.05.Dz, 78.70.Bj, 81.40.Tv ZnO and, in general, wide band gap semiconductors have experienced difficulties in both n and p type doping. There is normally one type of doping which is difficult to achieve. These difficulties have been attributed to the compensation by native point defects or to lattice relaxa- tions that cause the dopants to have deep energy level [1]. The location of dopants atoms on interstitial sites has also been suggested to be the origin of the doping difficulties in some cases [2]. ZnO, with its wide band gap, has numerous potential applications [3] that have been hindered by the lack of control over the electrical conductivity and the resistance to the formation of p type. Despite the number of studies on ZnO (e.g., [4 – 9]), many fundamental issues remain unresolved. The reason behind the difficulty of obtaining shallow acceptors remains unresolved. The na- ture of native shallow donors in ZnO and the dependence of luminescence and optical absorption on the environmen- tal conditions are still subjects of controversy [7–11]. Oxygen vacancies (O v ), Zn interstitials (Zn i ), and hydro- gen background impurities have been suggested as candi- dates for the native donors in ZnO. The role of hydrogen as a shallow donor has been supported by theoretical and experimental evidence [7,10]. Electron irradiation study has suggested Zn interstitial is the dominant native donor in ZnO [8]. The role of O v as a shallow donor has been widely discussed [9,11], but no agreement has been reached. First-principles calculations [12] predicted that while isolated O v are relative deep donors in ZnO, they are not the major shallow donor and not responsible for the n-type conductivity. However, no clear experimental evi- dence has supported this prediction. Research at Washington State University has identified the defects associated with heating of ZnO in different environments and investigated their potential as the domi- nant native donor in ZnO. Experimental evidence is pre- sented to demonstrate O v are the source for the shift in the optical absorption band that causes the red coloration in ZnO crystals after heating in Zn vapor [5]. From the present measurements, it is consistent this shift in the absorption band is 0:7 eV which indicated O v are deep donors in ZnO. This was consistent with the first principle calculations [12]. Data also suggested Zn i cannot be ac- countable for the n-type conductivity in ZnO and sup- ported recent calculations [13] that predicted hydrogen may occupy an oxygen sublattice and act as a shallow donor. Research also demonstrated different types of va- cancies can be generated and controlled which may facili- tate the doping process. In this investigation, optical transmission measurements were applied with positron annihilation spectroscopy (PAS) [14] which is well suited to identify neutral or negatively charged vacancy type defects in semiconductors. Because of the lack of positive ion cores, vacancy type defects form an attractive potential that traps positrons and leads to characteristic changes in the measured annihilation parameters. Commercial undoped ZnO single crystals were used. The as-grown samples were clear. Samples were sealed in an evacuated quartz ampoule with Zn, Ti, or Zr metal and heated to temperatures ranging from 900–1100  C for various times. Annealing with Zn at 1100  C caused the ZnO samples to turn orange or red as previously reported [5]. Figure 1 shows the optical transmission spectra for the as-grown and annealed samples measured at room tem- perature. The edge shifted from 390 to 500 nm after heating in Zn vapor and caused the red coloration. Depth resolved measurements of PAS Doppler broadening [14] were performed to investigate the source of the shift. The 511 keV annihilation peak was recorded using a HPGe detector. The S parameter [14], which is sensitive to the annihilation fraction with low momentum valence elec- trons, was used to characterize the annihilation peak. It was obtained by dividing the counts in the central region to the total counts in the peak. Trapped positrons at defects are more likely to annihilate with low momentum valence electrons causing an increase in S. S is displayed in Fig. 2 as a function of incident positron energy (E) and mean implantation depth. S is normalized to the minimum S bulk obtained for defect free ZnO single crystals. The high values of S at low energies (E< 5 keV) were associated to positronium formation at the surface. S in the bulk PRL 99, 085502 (2007) PHYSICAL REVIEW LETTERS week ending 24 AUGUST 2007 0031-9007= 07=99(8)=085502(4) 085502-1 © 2007 The American Physical Society