Microscopic mechanism of room-temperature ferromagnetism in Ga$_{1-x}$Cu$_x$N nanowires: exchange pathway, role of carriers and defects, and mixed-valence effects

Single-crystalline Ga$_{1-x}$Cu$_x$N nanowires (x = 0.01, 0.024) show ferromagnetic hysteresis at room temperature with saturation moments above 0.86 $\mu_B$ per Cu, substitutional Cu on Ga sites, and Cu L$_{2,3}$-edge X-ray absorption and magnetic circular dichroism indicating local moments in a predominantly 3d$^9$ configuration mixed with a small trivalent component. The discoverers left open what produces the long-range ferromagnetic coupling between these dilute Cu moments -- Cu 3d--N 2p $p$--$d$ hybridization, carrier-mediated exchange, or defect-mediated coupling -- while how the mixed valence and the intrinsic carrier density set the moment size, exchange pathway, and Curie temperature remains unresolved in the subsequent literature. This question remains unresolved: subsequent first-principles work disagrees on whether Cu-doped GaN is robustly or only weakly ferromagnetic, later experiments implicate vacancy-like defects and even ferromagnetism in nominally undoped GaN nanostructures, and no element-specific or carrier-control experiment has discriminated among the candidate mechanisms.

Background

Spintronics seeks to transport and manipulate carrier spin in addition to charge, and one route is the dilute magnetic semiconductor (DMS): a semiconductor in which a small fraction of host atoms is replaced by species carrying local magnetic moments, so that ferromagnetic order and semiconducting transport coexist in one material. For wide-band-gap hosts such as GaN, mean-field (Zener) theory of hole-mediated exchange -- coupling of local moments by itinerant valence-band holes -- predicted Curie temperatures above room temperature, which motivated an intense search for GaN-based DMS materials; later reviews emphasized that in many such systems the observed magnetism instead reflects defects or dopant-rich nanoregions rather than uniform carrier-mediated exchange (Dietl, 2008). Most efforts doped GaN with magnetic 3d transition metals (Mn, Cr, Fe), which risk forming magnetic secondary phases that mimic dilute ferromagnetism. Doping with copper avoids this: Cu is a nominally non-ferromagnetic element, so ferromagnetism cannot be dismissed as arising from magnetic precipitates. First-principles calculations by Wu et al. (2006) predicted that GaN doped with 6.25% substitutional Cu is a ferromagnetic semiconductor with fully spin-polarized carriers, attributed to $p$--$d$ hybridization -- the mixing between the dopant 3d states and the neighboring nitrogen 2p valence states, which can both create a local moment on Cu and mediate exchange through the N sublattice -- with an estimated Curie temperature near 350 K. On the experimental side, Lee et al. (2007) reported room-temperature ferromagnetism in Cu-ion-implanted GaN films that vanished when Cu clustered during 900 degree C anneals. The candidate source paper (Seong et al., 2007) then reported magnetism in single-crystalline Ga$_{1-x}$Cu$_x$N nanowires grown by the vapor-liquid-solid mechanism (a catalyzed nanowire growth mode). Anomalous X-ray scattering -- element-specific diffraction that isolates the contribution of Cu atoms -- together with X-ray diffraction showed that Cu atoms substitute Ga sites within the wurtzite (hexagonal) lattice network of GaN; X-ray absorption spectroscopy (XAS) and X-ray magnetic circular dichroism (XMCD, the difference in absorption of opposite circular light polarizations, which detects element-specific local moments and valence) at the Cu L$_{2,3}$ edges showed that doped Cu carries a local magnetic moment with an electronic configuration mainly 3d$^9$ mixed with a small trivalent component; SQUID (superconducting quantum interference device) magnetometry, an ultrasensitive measurement of net sample magnetization, showed hysteresis at 300 K with saturation above 0.86 $\mu_B$/Cu and Curie temperatures far above room temperature. The authors proposed only qualitatively that the ionocovalent bonding of the Cu 3d orbital with the surrounding semiconductor medium produces the mixed configuration and the moments, and did not identify the coupling mechanism; the role of charge carriers (whether intrinsic carrier densities suffice, or additional doping or gating is needed to enable or tune carrier-mediated coupling) and the quantitative impact of the mixed valence on the moment size, exchange pathways, and Curie temperature remain open in the subsequent literature. Three candidate coupling channels must be distinguished: $p$--$d$ hybridization through the N sublattice, carrier-mediated exchange of Zener/RKKY type (indirect exchange between local moments transmitted by itinerant electrons or holes), and defect-mediated coupling through vacancies or surface states. The last is not a straw man: nominally undoped GaN nanoparticles are predicted to be ferromagnetic via nitrogen dangling bonds at Ga vacancies (Jin et al., 2009), undoped GaN nanowires grown by molecular beam epitaxy show room-temperature hysteresis associated with Ga-vacancy-related defects (Jeganathan et al., 2014), and Cu incorporation itself generates vacancy clusters in GaN (Elsayed et al., 2011) -- so isolating the Cu-specific contribution is an intrinsic part of the mechanistic question.

Problem Statement

Establish the microscopic mechanism of the long-range ferromagnetic coupling between substitutional Cu local moments in Ga$_{1-x}$Cu$_x$N -- posed for the single-crystalline nanowire system (x = 0.01, 0.024) of Seong et al. (2007). A complete answer must identify the dominant exchange channel among Cu 3d--N 2p $p$--$d$ hybridization, carrier-mediated exchange, and defect/vacancy-mediated coupling, using evidence that discriminates among these channels (for example, controlled variation of carrier density by doping or gating, controlled variation of defect density by annealing or growth conditions with matched characterization, element-specific spectroscopy such as Cu L$_{2,3}$ XMCD that separates Cu moments from defect moments against undoped GaN nanowire controls, or first-principles calculations whose predicted moments, exchange constants, and Curie temperatures are quantitatively compared with measured values) rather than evidence merely consistent with one of them. Any account must be consistent with the experimental record established by the source: substitutional Cu on Ga sites of the wurtzite lattice, element-specific Cu local moments in the predominantly 3d$^9$ configuration mixed with a small trivalent component, and room-temperature hysteresis with saturation above 0.86 $\mu_B$ per Cu and Curie temperatures far above room temperature. Two closely related open issues frame the context of the question rather than forming additional completion requirements: whether the intrinsic carrier density in such nanowires can sustain the coupling or whether additional doping or gating is required, and how the mixed Cu valence quantitatively affects the local moment size, the exchange pathway, and the Curie temperature.

References

1. H.-K. Seong, J.-Y. Kim, J.-J. Kim, S.-C. Lee, S.-R. Kim, U. Kim, T.-E. Park, H.-J. Choi, Room-Temperature Ferromagnetism in Cu Doped GaN Nanowires, Nano Letters 7, 3366-3371 (2007). DOI: 10.1021/nl0716552, https://doi.org/10.1021/nl0716552
2. R. Q. Wu, G. W. Peng, L. Liu, Y. P. Feng, Z. G. Huang, Q. Y. Wu, Cu-doped GaN: A dilute magnetic semiconductor from first-principles study, Applied Physics Letters 89, 062505 (2006). DOI: 10.1063/1.2335773, https://doi.org/10.1063/1.2335773
3. J.-H. Lee, I.-H. Choi, S. Shin, S. Lee, J. Lee, C. Whang, S.-C. Lee, K.-R. Lee, J.-H. Baek, K. H. Chae, J. Song, Room-temperature ferromagnetism of Cu-implanted GaN, Applied Physics Letters 90, 032504 (2007). DOI: 10.1063/1.2431765, https://doi.org/10.1063/1.2431765
4. A. L. Rosa, R. Ahuja, Weak ferromagnetism in Cu-doped GaN, Applied Physics Letters 91, 232109 (2007). DOI: 10.1063/1.2819602, https://doi.org/10.1063/1.2819602
5. T. Dietl, Origin and control of ferromagnetism in dilute magnetic semiconductors and oxides (invited), Journal of Applied Physics 103, 07D111 (2008). DOI: 10.1063/1.2832613, https://doi.org/10.1063/1.2832613
6. L. Sun, F. Yan, H. Zhang, J. Wang, Y. Zeng, G. Wang, J. Li, Strong room-temperature ferromagnetism in Cu-implanted nonpolar GaN films, Journal of Applied Physics 106, 113921 (2009). DOI: 10.1063/1.3266013, https://doi.org/10.1063/1.3266013
7. X. L. Yang, Z. T. Chen, C. D. Wang, Y. Zhang, X. D. Pei, Z. J. Yang, G. Y. Zhang, Z. B. Ding, K. Wang, S. D. Yao, Structural, optical, and magnetic properties of Cu-implanted GaN films, Journal of Applied Physics 105, 053910 (2009). DOI: 10.1063/1.3079519, https://doi.org/10.1063/1.3079519
8. F.-Y. Ran, M. Subramanian, M. Tanemura, Y. Hayashi, T. Hihara, Ferromagnetism in Cu-doped AlN films, Applied Physics Letters 95, 112111 (2009). DOI: 10.1063/1.3232238, https://doi.org/10.1063/1.3232238
9. H. Jin, Y. Dai, B. Huang, M.-H. Whangbo, Ferromagnetism of undoped GaN mediated by through-bond spin polarization between nitrogen dangling bonds, Applied Physics Letters 94, 162505 (2009). DOI: 10.1063/1.3123169, https://doi.org/10.1063/1.3123169
10. P. R. Ganz, C. Sürgers, G. Fischer, D. M. Schaadt, Cu-doped GaN grown by molecular beam epitaxy, Journal of Physics: Conference Series 200, 062006 (2010). DOI: 10.1088/1742-6596/200/6/062006, https://doi.org/10.1088/1742-6596/200/6/062006
11. P. R. Ganz, G. Fischer, C. Sürgers, D. M. Schaadt, Cu-doped nitrides: Promising candidates for a nitride based spin-aligner, Journal of Crystal Growth 323, 355-358 (2011). DOI: 10.1016/j.jcrysgro.2010.10.115, https://doi.org/10.1016/j.jcrysgro.2010.10.115
12. Q. Yang, F. Sun, Z. Tang, First-principles studies for magnetism in Cu-doped GaN, Wuhan University Journal of Natural Sciences 16, 245-248 (2011). DOI: 10.1007/s11859-011-0744-9, https://doi.org/10.1007/s11859-011-0744-9
13. M. Elsayed, R. Krause-Rehberg, O. Moutanabbir, W. Anwand, S. Richter, C. Hagendorf, Cu diffusion-induced vacancy-like defects in freestanding GaN, New Journal of Physics 13, 013029 (2011). DOI: 10.1088/1367-2630/13/1/013029, https://doi.org/10.1088/1367-2630/13/1/013029
14. T.-H. Huang, P. R. Ganz, L. Chang, D. M. Schaadt, Formation Mechanisms of Islands on Cu-Alloyed GaN Grown by Plasma Assisted Molecular Beam Epitaxy, Journal of The Electrochemical Society 158, H860 (2011). DOI: 10.1149/1.3607425, https://doi.org/10.1149/1.3607425
15. R. González-Hernández, A. González-Garcia, W. López-Pérez, Ferromagnetism in Cu-doped polar and nonpolar GaN surfaces, Computational Materials Science 83, 217-221 (2014). DOI: 10.1016/j.commatsci.2013.11.024, https://doi.org/10.1016/j.commatsci.2013.11.024
16. K. Jeganathan, V. Purushothaman, R. K. Debnath, S. Arumugam, Ferromagnetism in undoped One-dimensional GaN Nanowires, AIP Advances 4, 057116 (2014). DOI: 10.1063/1.4878976, https://doi.org/10.1063/1.4878976
17. Y. H. Park, R. Ha, T.-E. Park, S. W. Kim, D. Seo, H.-J. Choi, Magnetic In$_x$Ga$_{1-x}$N nanowires at room temperature using Cu dopant and annealing, Nanoscale Research Letters 10, 3 (2015). DOI: 10.1186/1556-276X-10-3, https://doi.org/10.1186/1556-276X-10-3
18. M. Shakil, A. Hussain, M. Zafar, S. Ahmad, M. I. Khan, M. K. Masood, A. Majid, Ferromagnetism in GaN doped with transition metals and rare-earth elements: A review, Chinese Journal of Physics 56, 1570-1577 (2018). DOI: 10.1016/j.cjph.2018.05.018, https://doi.org/10.1016/j.cjph.2018.05.018
19. R. Xie, H. Xing, Y. Zeng, Y. Huang, A. Lu, X. Chen, Room temperature ferromagnetism in Cu-Gd co-doped GaN nanowires: A first-principles study, Physics Letters A 383, 54-57 (2019). DOI: 10.1016/j.physleta.2018.09.009, https://doi.org/10.1016/j.physleta.2018.09.009
20. G. Guzmán, D. Maestre, M. Herrera, Hole-mediated ferromagnetism in GaN doped with Cu and Mn, Journal of Materials Science: Materials in Electronics 31, 15070-15078 (2020). DOI: 10.1007/s10854-020-04070-7, https://doi.org/10.1007/s10854-020-04070-7