Photoluminescence of trivalent rare-earth ions codoped in ZnO thin films: Competing site occupation by Eu³⁺ and Er³⁺ ions

Highlights

• Photoluminescence from Eu³⁺ and Er³⁺ codoped ZnO thin films was investigated.

• Eu³⁺ emission signal was steeply attenuated when Er³⁺ concentration was increased.

• Eu³⁺ emission requires hydrogenated ZnO but Er³⁺ emission prefers hydrogen-free ZnO.

• Concentration quenching for codoping was greater than single doping of Eu³⁺ or Er³⁺.

• Eu³⁺ or Er³⁺ compete for occupation of emission-active sites.

Abstract

The effect of codoping two kinds of trivalent rare-earth ions in ZnO thin films was investigated in terms of photoluminescence (PL) intensities relative to those of singly doped ones. ZnO films codoped with 1 at.% of Eu³⁺ and a variable concentration of Er³⁺ were deposited on Si(100) and SiO₂ substrates. With preference to securing emission signals from Eu³⁺, the deposition was carried out under flow of H₂O vapor gas. After the samples were post annealed at various temperatures, their PL spectra upon bandgap excitation at 325 nm were monitored. It was found that the Eu³⁺ emission signal at 612 nm became steeply attenuated as the Er³⁺ concentration increased, whereas the Er³⁺ emission signal was absent from the films deposited on Si(100) substrate. The necessary conditions for Er³⁺ emission signal at 665 nm to appear were (i) deposition on SiO₂ substrate at low H₂O pressure, (ii) high Er content close to 3 at.%, and (iii) post annealing above 700 °C. Conditions (i) and (iii) correspond to minimizing the influence of hydrogen-containing (OH and H) species. Thus, the emissions from Eu³⁺ and Er³⁺ do not appear together because Eu³⁺ emission requires a hydrogenated ZnO host crystal, whereas Er³⁺ emission prefers a hydrogen-free environment. The observation of Er³⁺ emissions from films on SiO₂ but not from films on Si reflected the stiffness of the Si crystal and the flexibility of the SiO₂ network. Strain in the ZnO host crystal, which was caused by rare-earth ions substituting at Zn²⁺ sites, would have been relieved by slightly deforming the amorphous SiO₂ network. The concentration quenching for codoping with Eu³⁺ and Er³⁺ was greater than that for single doping with Eu³⁺ or Er³⁺. We propose an occupation-site-competing model wherein only limited numbers of emission-active Zn²⁺ sites are available for substitution by Eu³⁺ and Er³⁺

Introduction

ZnO is a most extensively studied oxide useful for various opto-electronic devices. Because its wide bandgap energy of 3.37 eV is advantageous for minimizing thermal relaxation from electronically excited states, application as host crystal for doping luminescent ions is promising. Up to now, Eu³⁺-doped ZnO (ZnO:Eu) [[1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18]] and Er³⁺-doped ZnO (ZnO:Er) [[19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31]] have been investigated by many research groups. These materials are representative combinations of trivalent rare-earth ions with ZnO. We have also reported visible photoluminescence (PL) from sputter-deposited ZnO:Eu [[32], [33], [34], [35]] and ZnO:Er [[36], [37], [38]] thin films through bandgap excitation. While some of the emission characteristics are common to ZnO:Eu and ZnO:Er, others are different. One common feature is that the crystallinity of the host film strongly affects the emission intensities of the rare-earth-ion dopants. While deposition with H₂O vapor gas is a necessary condition for Eu³⁺ emission to be observed, the presence of hydrogen species hinders Er³⁺ emission.

For rare-earth ions to be emission-active upon bandgap excitation, they must be components of the host crystal; i.e., occupation at Zn²⁺ sites is required for quick transfer of the excitation energy to them. If the rare-earth ions segregate at grain boundaries, the excitation energy is consumed via band-edge emissions and/or deep-level emissions of the ZnO crystal. In addition, the excited state should not decay non-radiatively before the excitation energy is transferred to the rare-earth ions. However, substituting Zn²⁺ with Eu³⁺ or Er³⁺ is an energetically unfavorable event. In the first place, the charge difference between divalent (Zn²⁺) and trivalent (Eu³⁺ and Er³⁺) ions is the problem. The replacement must satisfy Pauling's second rule; i.e., charge neutrality around the cation site must be locally conserved. One way to satisfy this requirement is introducing vacancies near the rare-earth ion to adjust the charge imbalance. Moreover, the ionic radius of Zn²⁺ is smaller than those of Eu³⁺ and Er³⁺. Substituting Zn²⁺ sites with Eu³⁺ or Er³⁺ ions exerts strain on the ZnO crystal lattice. If the rare-earth-ion dopants are few in number, the strain may be relieved by slightly deforming the lattice. However, once the rare-earth-ion dopant concentration exceeds some threshold value, the strain cannot be relieved unless vacancies are created at nearby sites or the rare-earth-ions are released to the grain boundaries. Once the rare-earth ions separate from the ZnO lattice, they are no longer emission-active. Moreover, creating vacancies opens deep-level emission channels, which are other energy consumption paths that compete with the emission channel of the rare-earth ions.

Substituting manner of trivalent ions at Zn²⁺ sites varies depending on the ion species. For instance, the doping characteristics of Al³⁺-doped ZnO (AZO) and Ga³⁺-doped ZnO (GZO) films are much more robust than those of rare-earth-ion-doped ZnO films. The high stability of AZO and GZO films have made them popular materials for transparent conductive electrodes [39]. It is based on the smaller ionic radii of Al³⁺ and Ga³⁺ than that of Zn²⁺, which ensures that Al³⁺ and Ga³⁺ can be substituted at Zn²⁺ sites while not causing stress in the crystal. When the dopant concentration is less than 3 at.%, all dopants occupy the cation sites and are electrically activated; thus they act as donors [40]. When the concentration exceeds 3 at.%, some of them fail to enter cation sites and become segregated at grain boundaries, which lowers the electrical activation probability per number of dopant ions. Nevertheless, the GZO crystal maintains c-axis orientation even at a high dopant level of 10 at.% [40]. This means that disruption of the ZnO crystal by Ga³⁺ is not so serious. These characteristics strikingly contrast with the case of rare-earth ions, where even small numbers of rare-earth ions hamper the ZnO crystal lattice. At high dopant levels, Zn vacancies (V_Zn) are created to maintain charge balance. When the Al³⁺ and Ga³⁺ ions occupying the Zn²⁺ sites are accompanied by V_Zn at a nearby site, the resulting Al_Zn–V_Zn [41,42] and Ga_Zn−V_Zn [43,44] pairs behave as acceptors. The transformation from a donor-predominant state to an acceptor-predominant state means that there is an optimum concentration of dopants at which the carrier concentration takes a maximum value. This situation resembles that of emission characteristics of rare-earth ions; the maximum emission intensity is reached at a certain concentration. Securing high-density doping while maintaining optical activation is a key factor to achieving a high emission yield.

In the work reported in this paper, we investigated how codoping of Er³⁺ and Eu³⁺ ions into ZnO crystalline films changes their emission properties, focusing on the interference between these ion species. Since the main emission peaks of Eu³⁺ (612 nm) and Er³⁺ (665 nm) are very close in terms of wavelength, it is worthwhile to clarify how the bandgap excitation energy is transferred to these ions and which emission channel is actually selected. On the basis of our experimental results, we propose a site-competing model for Er³⁺ and Eu³⁺ ions occupying emission-active sites.