M ö ssbauer studies of dilute magnetic semiconductors Marco Fanciulli

1. Mössbauer studies of dilute magnetic semiconductors Marco Fanciulli MDM – CNR-INFM National Laboratory http://www.mdm.infm.it

2. The collaboration (1) Aarhus – (2) Berlin – (3) CERN – (4) Durban – (5) Milano – (6) Reykjavik M. Fanciulli5, K. Bharuth-Ram4, A. Debernardi5, H. P. Gislason6, H.P. Gunnlaugsson1, Ö. Helgason6, K. Johnston3, R. Mantovan5, S. Olafsson6, D. Naidoo4, R. Sielemann2, G. Scarel5, G. Weyer1 Spokesperson: M. Fanciulli Contact person: K. Johnston

3. Main Objective Understand the origin of the magnetic ordering in semiconducting and insulating oxides and control their magnetic properties

4. Main Motivations Fundamental solid state physics: understand the mechanisms leading to magnetism in oxides and wide band gap semiconductors Applied solid state physics: spintronic devices (MTJ, spin valves, spin transistor), novel nanoelectronic devices (logic and memory), magneto-optic components, and QIP

5. Magnetism in ZnO Doping with few percentages of 3d metal impurities gives rise to room temperature magnetism. The magnetism can be several Bohr magnetons pr. Impurity atom. Ivill et al., JAP97 (2003) 053904 JMD Coey

6. No magnetism under n-type growth conditions (O2) while p-type conditions (N2O) are magnetic • Decrease in magnetic moment pr. Cu atom upon amount of Cu, attributed to close Cu pairs • TC > ~390 K observed

7. Role of defects • Points out that the magentism is likely related to defects • Oxygen pressure during growth has major impact on the magnetic properties; attributed to oxygen vacancies.

8. Samples (2% Mn) grown by solid state reaction method • Sintering at ~900 K optimizes the magnetic properties (700 K - 800 K according to Coey) • Sintering in Ar produces more magnetism than samples sintered in O2.

9. Magnetic ordering in HfO2: conflicting results Undoped HfO2 Undoped HfO2 • Growing interest in developing magnetic oxides and semiconductors for spintronics • Requirement: Curie temperature well above RT Experimental techniques that are able to study magnetism at the atomic scale are needed in order to clarify the nature of the observed macroscopic magnetism

10. Mössbauer Spectroscopy

11. Mössbauer Spectroscopy Isomer shift (position of spectral features): d ~ r(0) High spin Fe3+ : 0.35(10) mm/s High spin Fe2+ : 1.1(10) mm/s Quadrupole interaction (splitting of lines): DEQ ~ Vzz High spin Fe3+ : 0.5-1 mm/s High spin Fe2+ : 2-3 mm/s Cubic crystals : 0.0 mm/s Magnetic interactions (six line splitting of lines): Bhf ~ µ of site High spin Fe3+ : µ = 5 µB, Bhf ~ 50 T High spin Fe2+ : µ = 4 µB, Bhf ~ 40 T

12. Truly dilute semiconductors are easily produced as only ~3·1011 implanted probe atoms are needed to obtain a good spectrum (local concentration ~1017 cm-3). Impurity concentrations and depth profiles are controllable. Lattice defects are produced simultaneously and their interaction with the implanted probe atoms, possibly decisive for the occurrence of FM, can be studied, e.g. by varying the implantation temperature. The half-life and decay properties of 57Mn are favourable for such studies. 57Fe Mössbauer spectroscopy has sufficient sensitivity and resolution to obtain atomic scale information over a large temperature range of ~ 10 – 1000 K for studies of high TC magnetism. Why 57Mn implantation

13. Mössbauer Spectroscopy at ISOLDE

14. Fe Lattice location in ZnO Implanted: Measured: 59Mn (T½ = 4.6 s) → 59Fe (T½ = 44 d) 200 eV recoil b- detected with position sensitive detector > 80% of the Fe atoms are located on substitutional Zn sites at T > 300 K

15. 57Mn implantation in ZnO: test results • Sextet: 47 T • - 57Fe-V complexes? • Magnetic field distribution: 31 T (average) • - 57FeI produced by the 40 eV recoil energy • Quadrupole: 57Fe on perturbed Zn sites • Single line at T>600 K • - Origin? Dissolution of 57Fe-V complexes? Velocity (mm/s)

16. Low temperature measurements to see in more detail the effects of interstitial defects, which become mobile in the temperature range 77 – 300 K [22] as well as a possible mobility of Zn vacancies VZn. This should allow for better determination of the nature of the defects involved. (~ 4 shifts) Measurements in external magnetic fields to determine easy magnetisation axes and the sign of the magnetic hyperfine fields. In our current setup, we can reach external magnetic fields ~0.7 T utilizing permanent magnets. (~3 shifts) Angular dependent measurements to determine Vzz axes and/or the angle between Vzz and Bhf. (~3 shifts) Measurements on differently doped ZnO, pre-implanted/incorporated with 3d impurities and n- and p-type doping. (~4 shifts) Measurements on other oxides and/or related systems (HfO2,, Lu2O3, SnO2, TiO2, GaN, AlN and InN). (~3 shifts) Time-delayed measurements: These are performed at a constant, critical temperature and spectra are measured in different time intervals after the implantation to give the annealing behaviour on a minute timescale. (~3 shifts) Our proposal

17. Beam time: in lumps of 6-10 shifts, during 2-3 years 20 shifts in total. The UC2 target and the laser ion source are requested. Our requests

18. In parallel to the experiments planned at ISOLDE, a similar program involving stable 57Fe atoms will be pursued in the home laboratories, however, much larger concentrations are needed for measurable effects. For 2006 57Fe recoil implantation experiments are also planned at HMI in Berlin. These experiments are complementary (57Fe implantation and measurements during 100 ns) to those planned at ISOLDE. Within the planned activity, we also foresee a collaboration with the PAC and PL groups at ISOLDE. Complementary activities

19. CNR-INFM MDM National Laboratory (M. Fanciulli, R. Mantovan): - Conversion electron Mössbauer spectroscopy (MS) at RT and at 120 K - ESR and EDMR (X and Q band, 300 mK – 600 K, up to 12 T) - Growth by Atomic Layer Deposition (ALD) of oxides (G. Scarel) - First principle calculations of magnetic properties of Fe doped ZnO (A. Debernardi) University of Aarhus: Conversion electron Mössbauer spectroscopy (MS) at RT and transmission MS at 14-300 K on samples with stable isotopes. (H. P. Gunnlaugsson) Theoretical calculation of Mössbauer parameters of model defects (A. Svane) University of Iceland: MBE growth of GaN, AlN and InN samples, with or without Mn doping, C-V and Hall effect and PL characterisation, (S. Olafsson and Prof. H. P. Gislason). Transmission MS from 300-1000 K, (Prof Ö. Helgason). University of KwaZulu-Natal, Durban, and iThemba LABS, South Africa ( K. Bharuth-Ram): Transmission MS at 80 – 1000 K; CEMS at RT Vibrating sample magnetometer Relevant resources at home institutions

20. Mössbauer spectroscopy following implantation of radioactive 57Mn into oxide systems is a powerful method to investigate the magnetism in these systems. The experimental procedures can be expanded providing additional relevant information (other systems, orientation, external magnetic field, extended T range, rotation). Complementary experimental techniques will be also involved at ISOLDE and at home Labs. Theoretical work directly connected with the experiments will be also carried out by the collaboration. Conclusions