Skip to main content

Surface Physics


The growth of MnSb based compounds on semiconductor substrates has received much attention from the viewpoint of applications such as spintronics [1]. Spintronics or spin electronics is the controlled manipulation of electrons by their intrinsic spin as opposed to charge [2]. Far from just being an academic curiosity, spintronics is expected to revolutionise electronics and lead the way towards quantum information technology. It already has impacted on everyday life in the form of tunnelling magneto-resistance read heads in large capacity hard drives post 2000 [3, 4].

Currently much funding is being directed into the development of a new type of spintronics device, a ferromagnet-semiconductor hybrid spin injection system, in which a spin polarised current is produced in a ferromagnet and injected into the semiconductor. Numerous candidates for spin injection are reviewed by Hatfield [5] and Boeck et al [6].

Possible ferromagnetic and substrate materials must satisfy many pre-requisites to become viable candidates;

1) A Curie temperature high enough to be used in electronic applications.

2) Sufficient spin polarization of electrons.

3) Compatibility between the ferromagnet and the semiconductor substrate.

MnSb is a weak ferromagnetic material and as such does not have high spin polarisation. However many MnSb based compounds such as NiMnSb and CrMnSb demonstrate high spin polarisation [7] and are promising spin injection materials. In this project, thin films of MnSb are grown on semiconductor substrates in order to investigate point 3. It is hoped that the results from this prototype will aid in the growth of the Ni and Cr based compounds.

There are numerous reasons in choosing MnSb based compounds over more conventional ferromagnetic solids. These include its close lattice match with respect to many semiconductors and, perhaps most importantly from the point of device applications, the potential to grow heterostructures. However there are many challenges that must be overcome before this is realised. For example, blurring of the interface by the formation of unwanted Mn interfacial compounds [8], may result in the scattering of polarised electrons during injection from the ferromagnet to the semiconductor. These problems can be overcome by capping the substrate with Sb before MnSb growth [9].

Aims of the Project

The aim of this project is to investigate the growth of MnSb on Sb capped semiconductors including GaAs(001), Si(001) and Ge(001). These substrates have been chosen for two main reasons; their capacity for spin polarization transportation [10-12] and their small lattice mismatch with respect to MnSb. Films were grown using Molecular Beam Epitaxy (MBE), a technique whereby material is deposited onto the surface of a hot substrate, at low growth rates, using elemental effusion cells. High quality films are produced by ensuring the rate of arrival of film material is much greater than the rate of arrival of contaminants, a condition that can only be satisfied under ultra high vacuum (UHV) conditions. The stoichiometry of the film is determined by the flux of material from the effusion cells, which is itself dependant on the temperature and contents of cells. Thus there are numerous variables which must be tailored to produce the best growth results.

Use the links on the left hand side bar to navigate to pages on theory, surface structure, experimental techniques and equipment.

Experimental details

In situ experiments were performed in a UHV MBE system, “Magnusson” equipped with a retractable 3 grid LEED rig, water cooled K-cells (Mn, Sb) and an x, y, z, θ sample manipulator. In situ surface preparation was achieved using an argon ion gun for ion bombardment and annealing using a filament mounted behind the sample plate. UHV in the main chamber was achieved using a diffusion-rotary pump system along with a TSP and ion pump. The Fast Entry Chamber (FEC) was supported by a second diffusion-rotary pump system. After baking, a base pressure of the order 10-10 mbar was achieved in the main chamber.

Some MBE experiments were simultaneously run in another UHV system, “MadgeNETO”, built in house and described in detail in reference 5.

In total 4 film growth experiments were undertaken in the UHV chamber Madge (Due to K-cell failure on Magnusson). Experiment 1 (on the Ge Substrate) was used as a trial run in order to determine suitable deposition times and substrate/cell temperatures. However, often decisions on parameters were made in real time in response to RHEED data, which was monitored throughout deposition (for example, Sb was deposited for capping until a change in the RHEED pattern was seen). Details of the other three experiments are given in the table below.




Sb cap

MnSb growth




TSub (°C)

Temp (°C)

Time (mins)

Temp Mn (°C)

Temp Sb (°C)

Time (mins)

BEP flux ratio Sb/Mn




























Film growth parameters for each experiment. Previous deposition of MnSb using this chamber is detailed in [3], but much higher beam fluxes were used. The cell temperatures were chosen in order to give a Sb rich beam flux. BEP = Beam equivalent pressure measured using a beam flux gauge.

RHEED was performed at 12.5 keV and analysed using a program, RHEEDprofile, designed by Stuart Hatfield [5, appendix B] and edited by James Aldous. Integer order spacing was calculated at 20 different horizontal positions for each pattern and averaged, with the standard deviation used as the error.

XPS was performed in a separate UHV system equipped with a CHA analyser. Mg x-rays (Kα = 1254 eV) were used in early experiments with Al x-rays (Kα = 1487 eV) used later. Survey scans (0.5 eV step, 0.25 s step-1 dwell time) were obtained for the binding energy range ‑16.5 ‑ 983.5 eV, with detailed scans (0.1 eV step, 0.25 s step-1 dwell time) taken for regions of interest. Spectra were offset by an amount in order to bring the C(1s) peak to 185.0 eV. XPSPeak [40] was used to fit spectra, using a suitable background (usually Shirley).

Ex situ Microscopy: Atomic Force Microscopy was performed using an Asylum Research MFP3D AFM and Veeco Multimode AFM.


Full results may be seen here. The following section will highlight the main findings.

The aim of this investigation was to use MBE to grow epitaxial MnSb films on Sb terminated semiconductor substrates and this was realised for Ge(001) and GaAs(001) but not for Si(001). The reason for this was that the Si required a higher anneal temperature during the IBA treatment which could not be reached on the Madge chamber. The clean substrate reconstruction for Ge(001) was double domain 2x1. For the GaAs(001) surface, the reconstruction could not be determined from the RHEED data, although numerous observations suggest that the Ga rich (4x6) or c(2x8) may be present.

During growth, RHEED patterns were monitored in order to discern lattice parameter changes, surface reconstructions and surface quality. During Sb termination a change in surface reconstruction was not seen for any of the samples although the integer order spacing increased, corresponding to a decrease in the lattice parameter size. This may indicate the formation of a metallic Sb film due to the deposition of more than one monolayer of Sb. A future investigation could look into the effect of a post anneal treatment after Sb deposition in order to remove all but one monolayer of Sb as reported in 28.

MnSb deposition was undertaken on the terminated substrates. XPS results from the Sb(3d) region confirmed the presence of MnSb in the case of the Ge substrate, but proved inconclusive for the GaAs substrate (although evidence for the existence of metallic Sb on or near the surface was seen in the O(1s)/Sb(3d3/2) peak). Carrier concentrations and film resistivity, determined from Hall Effect measurements, indicated the films to be borderline metallic, with a larger than expected carrier concentration seen in the GaAs-MnSb film, again suggesting the presence of metallic Sb in the film.

During deposition, surface reconstructions were seen for both Ge and GaAs, along with changes in the IOS. For Ge(001)-Sb-MnSb, a p(2x2) reconstruction appeared after 15 minutes deposition and the IOS reduced by approximately 1.2% with respect to the initial Ge IOS. Although the IOS relaxed towards the bulk c axis MnSb value, the fact that it is not quite fully relaxed suggests the film is still under strain. The IOS and RHEED data suggest a double domain MnSb growth plane, with the c axis aligned along the surface. For GaAs-MnSb a double domain (2x1) surface reconstruction was seen along with an apparent doubling of the IOS with respect to the clean GaAs which has been seen in other studies and attributed to a growth plane of (1,-1,0,1).

We believe that this is the first example of epitaxial MnSb growth on the Ge(001) surface. Further projects supervised by Dr Gavin Bell will investigate this further before results are published.


I produced a poster sumarising the project for the final year presentation, which can be downloaded here (very large file).


Main Supervisor:

Dr Gavin Bell 




Project Partner:

Gavin Butler-Lee 

g dot butler-lee at warwick dot ac dot uk