Elsevier

Journal of Crystal Growth

Volume 310, Issue 18, 15 August 2008, Pages 4046-4049
Journal of Crystal Growth

Properties of InMnP (0 0 1) grown by MOVPE

https://doi.org/10.1016/j.jcrysgro.2008.06.069Get rights and content

Abstract

We investigated growth and incorporation of Mn into InP (0 0 1) by metal-organic vapour phase epitaxy (MOVPE). Depending on the Mn/In ratio and temperature we found four different incorporation regimes. Flat mirror like layers with high Mn incorporation were either produced around 510C or around 600C. At higher temperatures or higher Mn-fluxes, the surfaces roughened. We achieved a maximum Mn incorporation around 0.6%, estimated by X-ray diffraction. The corresponding hole concentration was 1.7×1017cm-3. The hole activation energy for the Mn acceptor in variable temperature Hall measurements was 220 meV, comparable to the onset of a broad photoluminescence. Due to this high activation energy the layers showed no spin polarization.

Introduction

Manganese and other transition metal alloys with III–V semiconductors have attracted much interest in the last few years due to the new ferromagnetic compound semiconductor Ga1-xMnxAs [1]. The maximum concentration of manganese in the case of most materials is relatively low, typically far below 10%. For GaAs it is 7–8% and results in Curie temperatures up to 170 K. Such high incorporations are usually achieved in MBE (molecular beam epitaxy) using low-temperature growth (less than 250C). Attempts to grow GaMnAs in MOVPE (metal-organic vapour phase epitaxy) resulted in the formation of Mn-rich clusters which were embedded defect free in the surrounding heavily Mn-doped GaAs matrix [2], [3]. These clusters are ferromagnetic at room temperature, but do not induce spin polarization comparable to homogenous GaMnAs [3].

Other materials like InMnP received little attention, despite the very first report on diluted magnetic III–V semiconductors (DMS) about photo-induced magnetization in InP:Mn [4]. In the paper the Mn concentration was given as 8×1018cm-3 without the actual hole concentration. Under illumination a magnetization with TC=12K was reported [4]. A newer work on In0.9Mn0.1P nano-clusters (3 nm diameter) reported TC=25K [5]. However, the authors assumed that size effects strongly influenced the magnetic coupling.

Shon et al. [6] claimed to have produced InMnP:Zn diluted magnetic semiconductor (DMS) by Mn indiffusion into an InP:Zn p-doped epilayer. Using EDX a maximum Mn concentration of 3% in the near surface region was found. However, there is strong evidence that the observed magnetization originates from MnP and InMnP clusters. First, there were two unidentified peaks in the X-ray diffraction (XRD) at 2Θ37 and 50. Second, the magnetization curves show two regimes, one where the magnetization drops rapidly until 50 K and a long tail with TC above 300 K. This indicates the formation of MnP inclusions/clusters, because MnP has a TC of 292 K and a second magnetization regime below 50 K [7]. Furthermore, by using ion implantation into undoped InP, no p-type conductivity could be achieved [6], [8], which is assumed to be a prerequisite for coupling between spins to achieve a DMS.

The electronic and optical properties of the Mn acceptor in InP have been studied mostly by low-temperature photoluminescence (PL) [9], [10], [11]. Mn was also intentionally incorporated as p-dopant, but showed low activation with a maximum hole concentration in the middle 1015cm-3 range in MOVPE [11], while in liquid phase epitaxy (LPE) a maximum hole concentrations of 3×1017cm-3 was reported [12]. A newer study of manganese and other transition metals in InP using bulk growth crystals reported a maximum hole concentration below 3×10-16cm-3 [13]. The low hole concentration was explained by the relatively high activation energy for the manganese acceptor of 220 meV, found by temperature dependent Hall, absorption and PL measurements. It was also noted that the absorption band became much broader upon high manganese incorporation [13].

A systematic ab inito theory study of V, Cr, and Mn in several III–V-semiconductors, including InP, suggested InP:Cr as the material with the highest exchange interaction [14]. But this study did not consider the problem of carrier activation to mediate the exchange interaction. Thus, the questions of whether Mn is a good choice for an InP based DMS is still open.

Therefore, in this work we present our investigation on InMnP and InMnP:Zn grown by MOVPE. Using MOVPE instead of implantation or indiffusion allows us very good control and reproducibility. Furthermore, since we use in situ characterization for growth control and stabilize our samples by PH3, we can exclude surface effects or annealing damage.

Section snippets

Experimental procedure

The samples were grown in a horizontal double wall quartz MOVPE reactor under in situ control using reflectance anisotropy spectroscopy (RAS). The carrier gas was either hydrogen or nitrogen at 100 mbar and 3 l/min. Precursors’ partial pressures were PH3 (100 or 200 Pa) and TMIn (typical 0.5 Pa). The manganese source was bis(cyclopentadienly)manganese (bcpMn) with partial pressures between 1 and 40 mPa. bcpMn is a viscous liquid with a vapour pressure of 5 Pa at 20C [15]. The expected decomposition

Growth of InMnP

In MBE, high concentrations of Mn in GaMnAs are incorporated by low-temperature growth, below 250C. Such low growth temperatures are possible, since the materials are supplied in elemental form as beams of atoms or dimers/tetramers. On the other hand, in MOVPE, the precursors require a certain minimum temperature for decomposition to release their elements. In the case of precursors for InP the temperatures for 50% decomposition are about 375C for TMIn and 500C for PH3, i.e. 250C above the

Conclusion

We have grown and characterized InP:Mn and InP:(Mn,Zn). Manganese showed strong segregation even at 500C. The maximum manganese incorporation before the onset of cluster formation was about 0.6%. The resulting hole concentration did not exceed 1.7×1017cm-3, indicating a relatively high activation energy. Using PL and temperature dependent Hall measurements, we estimated an activation energy of 220–230 meV, comparable to values in literature. Without clusters, the samples were paramagnetic and

Acknowledgements

Part of this work was supported by the Deutsche Forschungsgemeinschaft (Sonderforschungsbereich 296).

References (16)

  • T. Hartmann et al.

    Thin Solid Films

    (2000)
  • M. Lampalzer et al.

    J. Crystal Growth

    (2003)
  • Y. Shon et al.

    J. Crystal Growth

    (2005)
  • S. Wen-Bin et al.

    Mater. Chem. Phys.

    (1997)
  • H. Ohno et al.

    Appl. Phys. Lett.

    (1996)
  • R. Laiho et al.

    Phys. Rev. B

    (1990)
  • Y. Sahoo et al.

    J. Phys. Chem. B

    (2005)
  • E.E. Huber et al.

    Phys. Rev.

    (1964)
There are more references available in the full text version of this article.

Cited by (4)

1

Present address: Qimonda Dresden GmbH & Co. OHG, Königsbrücker Str. 180, 01099 Dresden, Germany.

2

Present address: Dipartimento di Fisica Università di Roma II: Tor Vergata, Via della Ricerca Scientifica 1, I-00133 Roma, Italy.

View full text