Manganese doped Germanium: from clustering to carrier-mediated ferromagnetism

Manganese doped Germanium (Ge:Mn) is a promising candidate for a ferromagnetic semiconductor compatible with silicon technology, since Mn acts as a magnetic ion as well as a double acceptor. In recent years, Ge:Mn thin layers as well as nanostructures have been fabricated, mostly by LT-MBE, and analyzed [1-4]. Whereas ferromagnetism above room temperature has been evidenced by magnetization measurements, the transport behavior (magnetoresistance: MR, anomalous Hall effect: AHE) is entirely different from the GaAs:Mn system, the prototype of a ferromagnetic semiconductor. For instance, the previously reported AHE in Ge:Mn (i) was observed at temperatures above 10 K, (ii) but exhibited no hysteresis, and (iii) changed the sign of its slope. This behavior has been ascribed to Mn-diluted Ge [2, 4], Mn-rich spinodal decomposed phases [3] and MnGe precipitates [5], respectively. We argue that the origin of these observations lies in the less effective substitution of Mn at Ge sites, which results in too low a hole concentration, making carrier-mediated ferromagnetism impossible. The hole concentrations realized in Ge:Mn grown by LT-MBE are mostly well below 10¹⁹ cm^-3.

We have prepared a series of Ge:Mn layer by Mn ion implantation into near-intrinsic, n-type Ge substrates, at 350 °C (resulting in Mn5Ge3 clusters) and -40 °C (without precipitates) [6-8]. The Mn concentration ranges from 0.004% to 10%. For samples with 10% Mn, several annealing procedures have been applied, namely ms flash lamp annealing and nsec pulsed laser annealing with various optical fluences. By this systematic preparation, we obtained three kinds of samples: (1) very dilute Ge:Mn where no ferromagnetic coupling can be expected [6]; (2) nanocrystalline Mn5Ge3 embedded inside the Ge matrix [7]; and (3) diluted Ge:Mn together with Mn-rich spinodal phases [8]. Indeed all samples show p-type conductivity with a hole concentration ranging from 10¹⁸ to 10²⁰ cm^-3. The highest concentrations above 1020 cm^-3 can only be achieved with the help of pulsed laser annealing. A careful characterization of structure, magnetic and transport properties, leads us to the following conclusions.

(1) From 20 to 10 K the resistance of samples with a hole concentration of >10¹⁸ cm^-3 increases in an activated manner with an activation energy of 4 meV, but below 10 K it saturates, i.e. the sample behaves metallic.
(2) We evoke the consideration of a two-band-like conduction in Ge:Mn. Above 10 K another conducting channel with different mobility is active, resulting in the drop of resistivity and the anomalous Hall resistance. The latter can be well described over a wide of parameters by considering two types of carriers with different mobility and population (see Ref. 6).
(3) In the sample with the largest hole concentration of 2×10²⁰ cm^-3, we observed, for the first time to our knowledge, a one-to-one correspondence between the hysteresis in magnetization, magnetoresistance and Hall resistance below 10 K. This is our key result, shown in Fig. 1, and is a strong evidence for carrier-mediated ferromagnetism. Note however that considering mere magnetization data, ferromagnetism remains present up to >100 K.

In summary, we present the magnetic and magnetotransport properties of a series of Ge:Mn samples with hole concentrations ranging from 10¹⁸ to 2x10²⁰ cm^-3. The hole concentration is the critical parameter to establish carrier mediated ferromagnetism in Ge:Mn, similar as is known for GaAs:Mn. A high-concentration co-doping with a shallow acceptor may allow to increase the hole concentration further, possibly resulting in a dramatically increased Curie temperature. In addition to the compatibility to Si technology, ion implantation followed by pulsed laser annealing is an established scalable chip technology and therefore may have a significant industry impact.

References:

[1] Y. D. Park et al., Science 295, 651 (2002).
[2] F. Tsui et al., Phys. Rev. Lett. 91, 177203 (2003).
[3] M. Jamet et al., Nature Mater. 5, 653 (2006).
[4] C. Zeng, et al., Phys. Rev. Lett. 100, 066101 (2008).
[5] O. Riss, et al., Phys. Rev. B 79, 241202(R) (2009).
[6] S. Zhou et al., Appl. Phys. Lett. 95, 172103 (2009).
[7] S. Zhou et al., Appl. Phys. Lett. 95, 192505 (2009).
[8] S. Zhou et al., Phys. Rev. B (2010), submitted.