University of Bath

Dept. of Physics

Contents:

Resonant Raman spectroscopy of dilute magnetic semiconductor quantum well structures
GR/H57356

Resonantly-excited Raman scattering studies of wide bandgap dilute magnetic semiconductor heterostructures
GR/H93774

Resonant Raman scattering and optical spectroscopy of II-VI epitaxial layers and heterostructures
GR/K04859

Raman scattering studies of doping and compensation processes in wide band gap semiconductor epitaxial layers

GR/M55077 & GR/L62283

(list to be updated!)

Resonant Raman spectroscopy of dilute magnetic semiconductor quantum well structures

Investigators: Professor J J Davies and Dr D Wolverson
EPSRC Grant GR/H57356 (£208,141, November 1st 1992 to October 31st 1995)

The project was part of a co-ordinated inter-university programme for the study of the novel physics associated with quantum well structures that contain dilute magnetic semiconductors. These unusual materials are formed when the metal ions of the semiconductor are partly replaced by ions that are magnetic, the most common examples being those produced when the cadmium or zinc ions in a II-VI compound are substituted for by manganese. Their importance is that the spin-exchange interactions with the magnetic ions results in a giant enhancement of the magnetic behaviour of electrons and holes. One consequence of this is that the energies of the band edges can be altered by the application of an external magnetic field. Thus, in quantum well structures that contain dilute magnetic semiconductors, the depths of the wells can be altered by significant amounts during the course of an experiment, even to the extent of removing a quantum well completely. These systems thus provide an opportunity of studying quantum well behaviour in a way that is not possible with conventional semiconductors.

The structures were produced by the molecular beam epitaxy growth facility at the University of Hull, with additional specimens from the CNRS/CEA team at Grenoble. The particular role of the Norwich team was to use resonant Raman scattering, particularly spin-flip scattering, to investigate the behaviour of charge carriers confined in quantum wells of different depths and geometries. Spin-flip Raman scattering has the special advantage of providing information about the behaviour of the wavefunction of particle (electron or hole) that is being studied without the interpretational difficulties that sometimes arise as a result of the electron-hole Coulomb interaction (in, for example, optical emission or absorption experiments). SERC provided funds for only a 5W laser and this automatically restricted the investigation to structures containing the dilute magnetic semiconductor Cd1-xMnxTe (and then only with x up to 40%). It is therefore with this material that the present report is mainly concerned. A later SERC award (GR/H93774) was made to purchase the more powerful laser and the very successful work on the larger bandgap dilute magnetic structures is described separately.

In bulk dilute magnetic semiconductors an electron in the conduction band (or a hole in the valence band) 'sees' a magnetic field made up of two contributions, the first due to the external field itself and the second due to the spin-exchange interaction with the spin system of the magnetic ions (which we shall hereafter assume to be manganese). In these materials the second of these contributions is, under the normal experimental conditions (low temperatures and moderate magnetic fields), much greater than the first, perhaps by a factor of 100, and it is this that makes the materials of interest. In a spin-flip scattering experiment involving, for example, electrons, the energy difference between the elastically and inelastically scattered photons (the 'Raman' shift) corresponds to the Zeeman splitting of the electronic state in the combined magnetic field and thus leads to a determination of the exchange interaction. Such experiments were carried out by a number of groups (using bulk material and epilayers) during the 1980s and formed the background to the present investigation. They are best carried out at low temperatures (typically 1.5K), since it is then easier to achieve a strong polarisation of the manganese spin system with moderate magnetic fields (e.g. 5 Tesla).

In a quantum well structure the situation becomes more complicated. There are two possibilities to consider as starting points. The first of these is encountered in a heterostructure formed from alternating layers of CdTe and Cd1-xMnxTe. This forms a type I multiple quantum well structure in which both the electrons and holes are confined in the smaller bandgap material CdTe. When a magnetic field is applied, the band edge energies of the Cd1-xMnxTe barrier are altered by up to several tenths of an electron-volt, in directions that depend on the spin state of the carrier. Thus electrons in the CdTe well with one spin direction will see an enhanced well depth whilst those with the opposite spin will be in a well that is reduced. In consequence, the confinement energies for electrons of the two spin directions will be different: it is this difference that is measured directly in a spin-flip Raman scattering experiment on the electrons confined in the wells. The importance of the measurement is that the Raman shift is determined by the width, the shape and the depth of the quantum well, and can thus provide useful information about the confinement behaviour of the carrier. The second simple case is one in which it is the barrier that is made up of non-magnetic material and the well of a dilute magnetic semiconductor (an example here is Cd1 xMnxTe/Cd1 yZnyTe with suitable values of x and y). Here again the depths of the wells seen by the carriers of different spin directions are significantly altered by applying an external field. More complicated structures that contain wells and
barriers that are both magnetic have also been studied during the project.

A key feature of the Raman scattering is that it is a highly resonant process, that is, the scattering rate is considerably enhanced (sometimes by several orders of magnitude) when the laser photon energy is adjusted to coincide with the energy of an appropriate excitonic transition in the material. The adjustment has usually to made to within 1 meV and it is for this reason that tuneable lasers are required. We have made extensive use of these resonance effects in the investigation and have constructed special equipment [1] in which the energies required for resonance can be sought automatically by simultaneously scanning the laser, the triple spectrometer (which uses a CCD array detector) and the pre-filtering
monochromator.

Cd1 xMnxTe/CdTe multiple quantum well structures Electrons in the CdTe quantum wells are subject not only to the well potential but also to that due to the Coulomb interaction with any ionised donor impurities that are present. This has important consequences for the spin-flip spectra and resulted in effects not previously observed in MQW structures. The most striking of these is the appearance of two peaks (denoted SF1 and SF2) in the spin-flip scattering spectra from electrons. This novel observation was completely unexpected and its understanding and exploitation became a major theme of the work. The actual spin-flip spectrum observed depends on the spatial distribution of the donors in the structure. If donors are uniformly distributed, the spin-flip energies will extend over a finite range, with maximum intensities at values for which they show turning points as a function of position. These turning points occur for donors at the centre of the well and for those just inside the barrier and lead to the two peaks observed experimentally.

To account quantitatively for the data, we used a variational approach that employs a wavefunction formed by convolving the hydrogenic donor ground state function with that for an electron in a square well. Excellent agreement with experiment was obtained, the only adjustable parameter being the ratio of the conduction band offset to that in the valence band, here found (when corrected for strain) to be 0.57. The theory has been tested against experiment for series of MQW structures in which both the barrier heights and the well widths were progressively altered. The investigation (which is described in references [2] and [3]) has thus led to a good understanding of the behaviour of electrons that are under the simultaneous influence of potentials due to square wells and to ionised dopants (this is a relatively common situation in quantum heterostructures).
The success of the investigation has also stimulated a series of more detailed theoretical studies [Harrison and co-workers at Hull]. Further, by investigating the relationship between the spin-flip energy and the laser energy required to bring that spin-flip transition into resonance, we have been able to establish (reference [5]) the correlation between the donor binding energy and the localisation energy of the corresponding exciton.

Cd1 xMnxTe/CdTe/Cd1 yMnyTe coupled quantum wells Spin-flip Raman scattering provides an unique method of investigating the coupling between two adjacent quantum wells, since the experiment deals with only one type of carrier (in our case, the electrons), compared with standard optical techniques, which involve electron-hole recombination and the complications associated with excitonic binding energies. When the separation between the wells is large, the two quantum wells are independent of each other and spin-flip signals can be observed from each separately. In contrast, when the width of the separating barrier is reduced, the wells are no longer independent and the electron wavefunction is calculated to extend over both of them and over the region of the barrier. Only one spin-flip signal is now observed, intermediate between those of the isolated wells and in good agreement with the prediction of a transfer matrix calculation. The work is described in reference [6].

Cd1-yZnyTe / Cd1- xMnxTe In addition to the work described above for specimens grown at Hull, we have also studied a series of structures produced (also by MBE) by the joint CNRS/CEA team at Grenoble. These specimens differ from the Hull ones in two respects. The first is the use of substrates of Cd1 yZnyTe and the second is the ability to incorporate zinc instead of manganese. The use of this substrate results in greater strain than specimens produced on InSb and increases the energy splitting between the light-hole and heavy-hole states. The inclusion of zinc makes it possible to study structures in which it is the wells, rather than the barriers that are magnetic (for example Cd1-xZnxTe/Cd1 yZnyTe with low x and high z). Of added interest is the situation in which the values of x and y are chosen so that the application of a magnetic field switches on potential wells that did not exist in zero field (a consequence of this is that one can switch on photoluminescence  transitions, as described in reference [7]): in this context, we studied spin-flip Raman scattering by electrons in potential wells that are being created or destroyed by applying a field, or for which there is a well for one spin state but not for the other.
 

Summary

Extensive studies were made using the technique of resonant Raman spin-flip scattering applied especially to multiple quantum well structures based on the CdTe/Cd1-xMnxTe system. The measurements have led to an excellent understanding of the way in which electrons in a quantum well behave when perturbed by the Coulomb field associated with ionised donors, a situation commonly encountered in quantum well structures but not easily studied in conventional (non-magnetic) systems. The behaviour of a range of MQW structures with varying well widths and barrier heights was studied and analysed successfully in terms of a variational model. The measurements were extended to structures containing asymmetric pairs of quantum wells with an interwell spacing that was reduced so as to study the coupling between the two wells in a unique way. The investigation also included studies of Cd1 xMnxTe/Cd1 yZnyTe structures in which the wells, rather than the barriers, were magnetic. We note finally that the investigation would not have been possible without the very high quality of the specimens provided by the growth teams at Hull and Grenoble.

Publications

1. D.Wolverson and S.V.Railson, 'Automated resonance Raman spectroscopy', Measurement Science and Technology 4, 1080-1084,1983.
2. M.P.Halsall, D.Wolverson, J.J.Davies, D.E.Ashenford and B.Lunn, 'Spin-flip Raman scattering from electrons bound to donors in both wells and barriers of CdTe/Cd0.93Mn0.07Te multiple quantum well structures', Solid State Comm. 86, 15-18, 1993.
3. D.Wolverson, J.J.Davies, S.V.Railson, M.P.Halsall, D.E.Ashenford and B.Lunn, 'Spin-flip Raman scattering by electrons bound to donors in CdTe/Cd1-xMnxTe multiple quantum well structures as a function of barrier composition', J. Crystal Growth 138, 656-660, 1994.
4. M.P.Halsall, S.V.Railson, D.Wolverson, J.J.Davies, B.Lunn and D.E.Ashenford, 'Spin-flip Raman scattering in CdTe/Cd1-xMnxTe multiple quantum wells: a model system for the study of electron- donor binding in semiconductor structures', Phys. Rev. B 50, 11755-11763, 1994.
5. D.Wolverson, S.V.Railson, M.P.Halsall, J.J.Davies, D.E.Ashenford and B.Lunn, 'Selective excitation of spin-flip Raman scattering from electrons bound to donors in semiconductor quantum well structures' Semiconductor Science and Technology 10, 1475-1483, 1995.
6. S.V.Railson, D.Wolverson, J.J.Davies, B.Lunn and D.E.Ashenford, 'Electronic wavefunctions in asymmetric double quantum well structures studied via spin-flip Raman spectroscopy', Superlattices and Microstructures 13, 487-491, 1993.
7. J.J.Davies. R.T.Cox and G.Feuillet, 'Optical studies of quantum wells induced by magnetic fields in Cd1-xMnxTe/Cd1-yZnyTe epitaxial structures', J. Crystal Growth 138, 661-666, 1994.
8. P.J.Klar, C.M.Townsley, D.Wolverson, J.J.Davies, D.E.Ashenford and B.Lunn, 'Photomodulated reflectivity of ZnTe/Zn1 xMnxTe multiple quantum wells with below-bandgap excitation', Semiconductor Science and Technology 10, 1568-1578, 1995.

Resonantly-excited Raman scattering studies of wide bandgap dilute magnetic semiconductor heterostructures

Investigators: Professor J J Davies and Dr D Wolverson
EPSRC Grant GR/H93774 (£77,524, July 1st 1993 to March 31st 1996)

The project formed the second part of a co-ordinated inter-university programme for the study of the novel physics associated with quantum well structures that contain dilute magnetic semiconductors. Dilute magnetic semiconductors are unusual materials that are formed when the metal ions of the semiconductor are partly replaced by ions that are magnetic, the most common examples being those produced when the cadmium or zinc ions in a II-VI compound are substituted for by manganese. Their importance is that the spin-exchange interactions with the magnetic ions results in a giant enhancement of the magnetic behaviour of electrons and holes. One consequence of this is that the energies of the band edges can be
altered by the application of an external magnetic field. Thus, in quantum well structures that contain dilute magnetic semiconductors, the depths of the wells can be altered by significant amounts during the course of an experiment, even to the extent of removing a quantum well completely. These systems thus provide an opportunity of studying quantum well behaviour in a way that is not possible with conventional semiconductors. The research concerned studies primarily of heterostructures based on ZnTe/Zn1 xMnxTe and ZnSe/Zn1 xMnxSe produced by molecular beam epitaxy growth facility at the University of Hull and at the Humboldt University, Berlin (in the first part of the programme the structures studied
were primarily those with small bandgaps, essentially those based on CdTe and Cd1 xMnxTe and Cd1 xZnxTe from Hull and from CEN-Grenoble).

The particular role of the Norwich team was to use resonant Raman spin-flip scattering to investigate the behaviour of charge carriers confined in quantum wells of different depths and geometries. Spin-flip Raman scattering has the special advantage of providing information about the behaviour of the wavefunction of particle (electron or hole) that is being studied without the interpretational difficulties that sometimes arise as a result of the electron-hole Coulomb interaction (in, for example, optical emission or absorption experiments).

In parallel with the Raman investigations, photoreflectivity techniques (in which the laser-induced changes in reflectivity are monitored) were used to obtain detailed information on excitonic transitions involving the quantum wells. By combining these experimental approaches with theoretical modelling we have been able to determine band offsets, to investigate magnetic field-induced transitions from type I to type II multiple quantum well structures and, in particular, to clarify the confusion that has existed in the literature over effects on the behaviour caused by the inevitable fact that quantum wells are never structurally perfect and by the change from three to two dimensional magnetism at the interfaces. The investigation would not have been possible without the very high quality of the specimens provided by the growth teams at Hull and Berlin and
has led to the publications listed below.
 

1. P.J.Boyce, D.Wolverson, J.J.Davies and W.Heimbrodt, 'Spin-flip Raman studies of a ZnSe/Zn1-xMnxSe multiple quantum well structure', Proceedings of the 1994 International Conference on Semimagnetic Semiconductors, Materials Science Forum 182-184, 455-458, 1995.
2. P.J.Klar, D.Wolverson, D.E Ashenford and B.Lunn, 'Photomodulated reflectivity of ZnTe/Zn1 xMnxTe multiple quantum wells', J. Crystal Growth 159, 528-532, 1996.
3. P.J.Klar, C.M.Townsley, D.Wolverson, J.J.Davies, D.E.Ashenford and B.Lunn, 'Photomodulated reflectivity of ZnTe/Zn1 xMnxTe multiple quantum wells with below-bandgap excitation', Semiconductor Science and Technology 10, 1568-1577, 1995.
4. P.J.Klar, P.J.Boyce, D.Wolverson, J.J.Davies, W.Heimbrodt, N.Hoffman and J.Greische, 'Spin-flip Raman scattering of ZnSe/Zn1-xMnxSe multiple quantum well structures', Journal of Crystal Growth 159, 1061- 1065, 1996.
5. P.J.Klar, D.Wolverson, J.J.Davies, B.Lunn, D.E Ashenford and T.Henning, 'Spin-flip Raman scattering in quantum dots based on a Cd0.91Mn0.09Te multiple quantum well structure', Proceedings of the 23rd International Conference on the Physics of Semiconductors, Berlin, vol .I, 1485-1489, 1996.
6. P.J Klar, D.Wolverson, D.E.Ashenford and B.Lunn, 'Observation of a type-I to type-II transition via photoluminescence and photo-modulated refllectivity of ZnTe/Zn1 xMnxTe multiple quantum wells', submitted to Semiconductor Science and Technology
7. P.J.Klar, D.Wolverson, J.J.Davies, W.Heimbrodt and M. Happ, 'Determination of the chemical valence band offset for ZnSe/Zn1-xMnxSe multiple quantum well structures of high x', in preparation
8. P.J.Klar, Magneto-optical spectroscopy of semiconductor quantum structures, PhD Thesis, University of East Anglia, 1997

Resonant Raman scattering and optical spectroscopy of II-VI epitaxial layers and heterostructures

Investigators: Dr. D Wolverson and Professor J J Davies
EPSRC Grant GR / K04859 (£94988, 20th Feb. 1995 to 19th June 1997)

The objectives of the project were linked to those of a EU HCM network on the MOVPE production of II-VI semiconductor materials and were (a) the improvement of wide bandgap materials for opto-electronic applications; (b) the production of II-VI heterostructures of good quality and (c) the further development of MOVPE doping techniques. The task of the UEA group in this network was to use resonance Raman spectroscopy and associated optical techniques in order to provide detailed characterisation of the materials produced and to make comparative studies of MOVPE and MBE-grown samples.

The principal technique employed in this research has been spin-flip Raman scattering (SFRS), an inelastic light scattering process in which the excitation created or annihilated corresponds to a change of the spin direction of an electron or a hole. The Raman shifts therefore depend on the magnitude and direction of the magnetic field and the experiment thus provides information about the spin Hamiltonian parameters of the carriers. These parameters are very sensitive to the environment (e.g., the degree of confinement, the strain and the state of binding of the particle) and their determination can provide extremely useful information. For instance, the are related to the band structure of the material and, thus, to the effective masses: the relationship is relatively simple for carriers in band states and more complicated for impurity-bound carriers.

Within this project, we have made the first observation of the spin-flip Raman scattering of nitrogen acceptors in ZnSe and have determined the spin Hamiltonian parameters of this centre, which were previously unknown. We have developed experimental techniques (angle-dependent SFRS) and associated theoretical models in order to understand the interplay of strain and magnetic-field effects in the acceptor spectra and have demonstrated an extreme sensitivity of the SFRS spectra to small strains in the sample. This has been achieved through the study of epitaxial layers of a wide range of thicknesses and, in some cases, layers from which the substrates have been removed.

The strain-sensitivity of SFRS has made possible the first optical spectroscopic detection of small deviations from biaxial in-plane strain in some ZnSe epilayers, yielding orthorhombic rather than tetragonal symmetry. This reduction in symmetry can be related to the asymmetric nature of the relaxation process at the ZnSe / GaAs interface, a crucial region in, for example, laser structures. We have also commenced SFRS studies of stripes of ZnSe:N fabricated by a combination of ion-beam and wet chemical etching processes, in order to exploit the strain-sensitivity of SFRS for investigations of the strain in one-dimensional nanostructures.

The relationship between nitrogen acceptors and the compensating deep donors that limit the active acceptor concentration in ZnSe at high nitrogen content has become clearer. We have shown that post-growth annealing treatments (which can dramatically increases the active acceptor concentration of MOVPE-grown material) simultaneously increase the concentration of compensating donor centres; the latter process is believed to arise from the in-diffusion of vacancies, which are the basis of the widely-accepted model for the deep donor centre. We are now extending this work to the study of related centres in ZnSe doped with lithium, another substitutional acceptor.

In addition to these published results, we have continued to supply often unpublished but important information to collaborating growth teams on the quality of their material relative both to material grown by alternative techniques and to some of the best material world-wide. This facility has been much in demand and several new interactions have been established. Our ability to detect the shallow donors, deep donors and acceptors in an epilayer produced under given conditions provides a 'fingerprint' for that layer and we have carried out many experiments giving feedback on, for example, (i) purity of precusor materials with respect to halogen contamination; (ii) optimum conditions for post-growth annealing; (iii)
relative efficacy of different nitrogen-doping precursors both with and without the N-H bonds which have been implicated in the hydrogen passivation of N in ZnSe and (iv) optimum microwave power in RF plasma N doping in MBE.
 

Publications

1. J. J. Davies, D. Wolverson and P. J. Boyce, "Raman scattering by acceptors and donors in p-type ZnSe", physica status solidi (b) 187 (1995) 407-413
2. C. M. Townsley, B. Schlichtherle, D. Wolverson, J. J. Davies, K-I. Ogata and Sg. Fujita, "Spin-flip Raman scattering of holes bound to acceptors in p-type nitrogendoped zinc selenide", Solid State Commun., 96 (1995) 437-440
3. C. M. Townsley, J. J. Davies, D. Wolverson, P. J. Boyce, G. Horsburgh, T. A. Steele, K. A. Prior and B. C. Cavenett, "Spin-flip Raman scattering studies of compensating donor centers in nitrogen-doped zinc selenide grown by molecular beam epitaxy", Phys. Rev. B 53 (1996) 10983-10987
4. D. Wolverson, B. Schlichtherle, C. L. Orange, W. Heimbrodt, J.J. Davies, K. Ogata, Sg. Fujita and T. Ruf, "Spin-flip Raman scattering from acceptors in nitrogen-doped ZnSe", proceedings of the 23rd International Conference on Semiconductors, Berlin (1996), vol. II, 2965-2969
5. D.Wolverson, P.J.Boyce, C.M.Townsley, B.Schlichtherle and J.J.Davies, "Raman scattering studies of doped epitaxial zinc selenide", Journal of Crystal Growth 159 (1996) 229-237
6. J. J. Davies, D. Wolverson, B. Schlichtherle and C.M.Townsley, "Raman scattering studies of p-doped II-VI compounds", Proc. Int. Symp. on Blue Laser and Light-emitting Diodes, Chiba, Japan (Ohmsha, Tokyo) (1996) 188-193
7. C. L. Orange, B. Schlichtherle, D. Wolverson, J. J. Davies, T. Ruf, K-I. Ogata and Sg. Fujita, "Angle-resolved studies of the spin-flip Raman scattering of holes bound to acceptors in p-type nitrogen-doped zinc selenide", Phys. Rev. B55 (1997) 1607-1617
8. D. Wolverson, "Raman scattering studies of doping of epitaxial zinc selenide" Inorganic Materials (Russia) 33 (1997) 139-147
9. W. Heimbrodt, C. L. Orange, D. Wolverson, J. J. Davies, K. Kimura and T. Yao, "Determination of nitrogen acceptor spin Hamiltonian parameters in ZnSe epilayers via spin-flip Raman spectroscopy", submitted to Phys. Rev. B in May 1997

Raman scattering studies of doping and compensation processes in wide band gap semiconductor epitaxial layers

1. Introduction

This 24 month project (£91K) involved the use of spin-flip Raman scattering (SFRS) to study wide bandgap semiconductors, with special emphasis on epitaxial ZnSe and related II-VI materials. SFRS is now a well-established technique by which the magnetic properties of electrons and holes can be investigated and by which the behaviour of dopants and defects can be monitored. SFRS can also be used to study the effects of confinement on electrons and holes in quantum wells and quantum dots. Because the spin-flip scattering is very strongly enhanced when the laser is adjusted to be in resonance with the appropriate excitonic transitions, carriers bound under different conditions can be studied selectively, even in the same specimen. Furthermore, the SFRS signals are not subject to the inhomogeneous broadening that affects the photoluminescence spectra, so that the technique can be applied successfully to ternary and quaternary semiconductor alloys.

The project was focussed initially on a greater understanding of the behaviour of p-type dopants in ZnSe, with special attention to nitrogen and to improved methods of incorporating this element without compensation in epitaxial material produced by both MOVPE and MBE. As part of this study we also studied the n-type centres that are usually incorporated in undoped ZnSe and the steps taken to reduce such incorporation in MOVPE material through the use of different precursors. In later phases we also studied alternative p-type dopants (lithium, phosphorus and oxygen). In this we worked closely not only with our colleagues at NEWI and at Heriot Watt University but also with a wide range of collaborators in Germany, France, Italy, the FSU and Japan (see section 5).

During the course of the project, world-wide interest in II-VI materials shifted increasingly towards the behaviour of self-organised systems, notably those formed from ZnCdSe fractional layers in ZnSe. In response to this changing background we adapted our techniques to study excitonic behaviour, making use of higher magnetic fields through a particularly fruitful collaboration with the Max Planck Institute at Stuttgart, using specimens from the Ioffe Institute and Heriot Watt University. Particularly exciting information was obtained about exciton localisation in these systems which will form the spring board for future studies.

Finally, the experience gained in studying the II-VI systems was applied to the study of wide gap III-V materials, in particular the yellow-red emitting Group III phosphide ternary and quaternary alloys, in new collaborative programmes with Epitaxial Products International and with the Sharp Laboratories of Europe.

The original objectives, briefly stated, were the greater understanding of the processes that limit effective p-type doping in ZnSe and related alloys such as ZnCdSe, ZnMgSe, ZnSSe and ZnMgSSe, together with comparisons between material produced by the different growth techniques and feedback to the growth teams as appropriate; we also investigated fundamental materials parameters that are relevant to the functional role of the materials in the opto-electronic context.

As noted above, we extended these objectives in the later part of the work so as to respond to the new aspects of the opto-electronic behaviour that result from the localisation of charge carriers and excitons in both II-VI and III-V heterostructures.

The work has resulted in 7 published papers, plus 2 in press . At least 9 further papers will result from the research, which forms the basis for three PhD theses. The work has also been presented at several international and national conferences, including invited presentations. It has contributed to the training of three research students and has led to new collaborative ventures with research groups in university, industrial and government laboratories. The work has also suggested new research directions for which EPSRC funding has been obtained.

4.1.1. Nitrogen doped material (Collaborators: HW, NEWI, Kyoto, Bremen, Sony, JRCAT)

The nature of the so-called deep donor that appears at a depth of about 50 meV when ZnSe is doped with nitrogen has been a matter of controversy. The centre is of considerable importance since it appears to act as a compensating centre which is (at least in part) responsible for the difficulties of obtaining effective acceptor concentrations in excess of 10¹⁸ cm^-3. Several models have been suggested, the most favoured of which is that of a selenium vacancy associated with a substitutional nitrogen. However, although very recent total energy calculations supported such a model, no conclusive evidence for this identification existed. Electrons trapped by the centre have a gyromagnetic ratio of 1.38, in contrast to the value of 1.12 for electrons bound at the normal, shallow (25 meV) donors in this material. In order to obtain a better understanding of the centre, we used SFRS experiments to determine how this g-value depends on composition in the nitrogen-doped ternary alloy ZnS_xSe_1-x, for x in the range 0 to 0.1. Surprisingly, we found that the g-value remains constant, in contrast to that for the shallow donors, which increases by 0.12 over the same composition range. The results are consistent with the model outlined above but also present a significant challenge for theoretical calculation, since any microscopic model has to be reconciled with this g-value invariance [3,7]. We also carried out post growth annealing studies on MOVPE material in which the relative concentrations of active acceptors and compensating donors were monitored in order to confirm the optimum conditions for optimisation of p-type doping [1].

In MOVPE specimens, compensation can also occur because of halogen impurities (which form shallow donors) in the precursors used during growth. SFRS is very sensitive to the presence of such donors and was used to characterise specimens produced at NEWI using different precursors. The work formed part of a four-way collaboration between NEWI, the University of Lecce (Italy) and Epichem Ltd which was funded partly by the British Council.

A further aspect was the investigation of the effects of growth technique and layer thickness on the strain, which could be determined through the anisotropy induced in the SFR spectra due to the holes [2]. There is also evidence that for MBE layers of ZnSe on GaAs substrates, additional shallow donors (presumed to be nitrogen-related) are formed in nitrogen-doped material, especially if the layer thickness is small. A recent part of our work has been the study of layers from Bremen in which such centres are prevalent.

4.1.2. Lithium doped material (Collaborators: TU-Berlin)

Here the object was two-fold: first, to determine how the magnetic properties of holes trapped at acceptors depended on the nature of the acceptor and, secondly, to determine whether the deep compensating donor (g = 1.38) could be observed in materials that contained p-type dopants other than nitrogen. In contrast to epitaxial layers, in this case the ZnSe was in bulk crystal form and therefore strain free, so that the SFRS spectra were isotropic with respect to the direction of the magnetic field. The acceptor g-value was found to be close to that observed in epitaxial materials for nitrogen, which has a similar depth, but differs from that obtained for oxygen-related acceptors (see below) and from that for free excitons. There was no evidence of the g = 1.38 centre. Instead, particularly interesting signals were obtained (in the region of donor bound excitons) showing spin-exchange interactions between electrons and holes: we believe this to be the first time that such SFRS signals have been observed in bulk ZnSe [13].

4.1.4. Oxygen-containing material (Collaborators: Bremen, Valbonne).

This is a particularly interesting investigation since it follows a demonstration that oxygen contamination can lead to the formation of acceptor centres in ZnSe that do not contain nitrogen. The presence of acceptors of binding energy of about 90 meV was established in the first MBE specimens grown after the chamber was exposed to atmosphere during maintenance. We have used SFRS to study these specimens: preliminary results confirm the existence of these acceptors and indicate that the g-values differ slightly from those for their deeper counterparts nitrogen and lithium [16].

Experiments were also carried out on ZnSe doped with phosphorus, but no hole related SFRS spectra were observed, the spectra being dominated by electron SFR signals, suggesting that the problems due to compensation have not yet been solved for this particular p-type dopant.

4.2. Study of undoped II-VI ternary and quaternary alloys (Collaborators: Heriot-Watt, NEWI, Bremen, Kyoto, Sony, JRCAT, Ioffe)

The k.p perturbation theories that enable one to calculate the electron g-values for binary semiconductors are now well developed and give good agreement with experiment. However, the corresponding theory and experimental determination had not been carried out for several of the ternary and quaternary alloys of interest in opto-electronic applications. Such measurements are important both because of the insight that they provide into the band structures but also because they are the precursor to the study of carriers confined in quantum wells and in quantum dots. We therefore used SFRS to determine the electron g-values in ZnSSe, ZnMgSe, ZnMgSSe and (cubic) ZnCdSe. We also developed the relevant theory and find that we can obtain very good agreement with experiment [5]. We have also determined for the first time the electronic g-value for cubic CdSe [8, 14].

4.3. II-VI structures with n-type quantum wells (Collaborators: Bremen, Lecce, MPI)

The gyromagnetic ratios (g-factors) of electrons in ZnSe quantum wells with ZnMgSSe barriers were determined by resonance SFRS as functions of well width and of the direction of the applied magnetic field. These ZnSe/ZnMgSSe heterostructures were undoped but are n-type. They provided an excellent test of the applicability of k.p perturbation theories to quantum well systems, since in this semiconductor system there is both a large penetration of the electron wave function into the barrier and a considerable difference in the light-hole and heavy-hole confinement energies in the well. The former leads to exposure of the electron to a region in which the spin-orbit coupling is significantly reduced, whilst the latter introduces a marked anisotropy in the g-tensor. The ways in which the principal components of the g-tensor evolve from their value in bulk ZnSe as the well width is decreased was found to be described very well by simple analytic expressions obtained from k.p theory that involves only the conduction band and the J =3/2 heavy and light hole valence bands. These three-band expressions have the advantage of providing excellent insight into the factors that influence the g-tensor of electrons that are partly quantum-confined. The results also show that the conduction band offset in these structures is relatively small, being much closer to the value of 10% quoted by some authors rather than to the value of 30 to 40% quoted by others [10].

We have also studied SFR scattering from excitons localised by well width fluctuations in ZnCdSe/ZnSSe quantum well structures. These (ongoing) experiments enable us to investigate the electron-hole exchange interactions as a function of localisation (see also section 4.5) [15].

4.4. II-VI structures with p-type quantum wells (Collaborators: Bremen)

These structures became available only at the end of the project. They are specially interesting because they contain holes under different conditions of confinement and also because there is evidence for the existence of charged excitons in the wells. We have accumulated a wealth of SFRS data on specimens with well widths ranging from 10 nm to 100 nm, with strong SFRS spectra from electrons, holes and excitons. The analysis of these data is not yet complete but will certainly lead to published work [17].

4.5. Study of self-organised ZnCdSe/ZnSe structures and of excitons localised in ternary and quarternary alloys (Collaborators: Ioffe, MPI, Heriot-Watt, Bremen)

This was an exciting development that was made possible by the availability of specimens that did not exist at the time of award of the grant. SFR was used to determine the g-factors and electron-hole exchange constants for excitons localised at sub-monolayer insertions in a ZnSe matrix. Because of the highly selective nature of the resonance enhancement, we are able to determine the way in which the exchange constant varies with the exciton localisation energy and, in this case, find a strictly linear dependence. We show that the behaviour is that expected from excitons that are localised by compositional variations (including small island formation) whose scale is less than that of the exciton radius. [4,6].

Exciton localisation in wide gap materials has important implications for light-emitting devices and we have therefore investigated the exciton exchange energies as functions of localisation in ZnCdSe/ZnSSe quantum well structures [15] (and also in the group III phosphides, see section 4.7).

4.6. Study of ZnSe:ZnMnSe quantum wells (Collaborators: Würzburg)

This was a short collaboration with the group of Professor J Geurts and a research student, Mr A Keller, who visited Bath during the summer of 1999. The SFRS techniques were used to investigate the behaviour of electrons in ZnSe quantum wells with barriers of the dilute magnetic semiconductor ZnMnSe. The collaboration was especially valuable in enabling the SFRS techniques of the two groups to be compared directly but also in providing insight into the role played by charged excitons in II-VI quantum well systems. The collaboration continues and will lead to at least one publication [18].

4.7. Study of group III phosphide ternary and quaternary layers. (Collaborators: Epitaxial Products International and the Sharp Laboratories of Europe).

Following the success of the II-VI ternary and quaternary work it became clear that the resonance SFRS techniques were equally applicable to alloys formed from other semiconducting compounds. SFRS were therefore carried out on (Al_xGa_1-x)_0.5In_0.5P epitaxial layers grown by MOVPE at EPI (p-doped) and by MBE at Sharp (undoped). Both the electron and hole g-values were determined as functions of composition over the region in which the material remains a direct gap semiconductor and the former used as a means (though k.p theory) of discriminating between different choices that appear in the literature for the spin-orbit coupling parameter (which is of importance in device simulation programmes) [11]. The experiments have also led to measurement of electron-hole exchange interactions as a function of exciton localisation energy [12].

The roles of the different collaborating groups were as follows.

The work described above has enabled us to make substantial contributions to the field under five of the six detailed objectives listed in the original proposal. Objective (d) (the investigation of the advantages of different substrates and of different substrate orientations) could not be addressed in the II-VI context because funding for the growth of such specimens by our collaborators was not in the event available. However, in this context we were able to substitute an investigation of the effects of different substrate orientations on the exciton localisation in quaternary group III phosphide layers. As already mentioned, we were able to respond to changes in the international interest in the field towards quantum confined and self-organised systems and to make substantial contributions in these areas.

A major part of the award was for the consumable budget (mainly liquid helium) and the special items (laser warranty). We obtained funding from elsewhere (the British Council) for the PDRA (Dr K Ogata) and were able to obtain additional travelling funding from the British Council for promoting the interaction with our colleagues at Lecce. The budget for travel was invaluable in helping maintain our collaborative contacts and in establishing new interactions (section 5) and, especially, in making it possible to carry out the high field experiments at the MPI (thus avoiding the need to purchase a new magnet). In addition to enabling the research described above to be carried out, the funding provided excellent training for three research students (two EPSRC funded). We therefore believe that we have made excellent use of the resources provided by EPSRC, the more so since the investigation has stimulated ideas for further novel studies, some of which have already been funded (see form RGR1).

The investigation, which has led to a substantial number of publications, has provided a much clearer understanding of the behaviour of p-type dopants in wide bandgap II-VI materials and has enabled us to provide much useful information on dopant and compensation behaviour to the growth teams. We have developed a very clear understanding of the behaviour of electron g-values in binary, ternary and quaternary II-VI compounds (including carriers under quantum confinement) and have thus established greater confidence in the band structure theories that are used for device modelling. We have broadened our studies to include wide bandgap materials of the III-V series and in both these and the II-VI series we have been able to extend the work on electrons and holes to include situations where the carriers are bound in excitonic states. In particular, it has become clear that the resonance SFR technique presents several advantages for the investigation of excitons localised in quantum wells and quantum dots, including those that are self organised. Because of the selectivity provided by the resonance enhancement of the signals, we are able to investigate excitons under different states of confinement and localisation without the need for recourse to micro-photoluminescence or to the use of mesa structures and this presents a particularly exciting prospect for future studies.

Publications

1. C.L.Orange, D.Wolverson, B. Schlichtherle, J.J.Davies, K.Ogata and Sg.Fujita, 'Spin-flip Raman scattering studies of post-growth annealed p-type nitrogen-doped zinc selenide', Semiconductor Science and Technology 12, 1609-1614, 1997.

2. C. L. Orange, W. Heimbrodt, D. Wolverson and J. J Davies, 'Spin-flip Raman spectroscopy of nitrogen acceptors in ZnSe layers with different biaxial strains', Journal of Crystal Growth,184/185, 510-514, 1998.

3. K Ogata, D. Wolverson, J J Davies, Sz Fujita and Sg Fujita, ‘Spin-flip Raman scattering studies of the compensating donor centre in nitrogen-doped ZnS_xSe_1-x’, Proceedings of the Second International Symposium on Blue Laser and Light-emitting Diodes, Chiba, Japan (Ohmsha, Tokyo) 608-611, 1999.

4. O.Z.Karimov, D.Wolverson, J.J.Davies, T Ruf, and L. N. Tenishev, ‘Resonant spin-flip Raman scattering studies of II-VI semiconductor heterostructures’, Phys. Stat. Sol. (b) 215, 373-378, 1999.

5. D.Wolverson, J.J.Davies, C.L Orange, K. Ogata, Sz. Fujita, Sg.Fujita, K. Nakano, B. Jobst and D.Hommel, ‘Spin-flip Raman scattering studies of wide bandgap II-VI compounds’, Phys. Rev. B 60, 13555-13560, 1999.

6. T Ruf, O.Z.Karimov, D.Wolverson, J.J.Davies, A.N.Reznitsky, A.A.Klochikin, S.Yu.Verbin, L.N.Tenishev, S.A.Permogorov and S.V Ivanov, ‘Spin-flip Raman scattering in sub-monolayer CdSe/ZnSe structures’, Physica B 273-274, 911-914, 1999.

7. K.Ogata, J.J.Davies, D.Wolverson, Sz.Fujita, Sg.Fujita, ‘The gyromagnetic ratio of the compensating donor centre in nitrogen-doped ZnS_xSe_1-x’, Semiconductor Science and Technology 15, 209-213, 2000.

8. D.Wolverson, O.Z.Karimov, J.J.Davies, S.J.C. Irvine, S.Telfer, K.A.Prior, K.Ogata, Sz. Fujita, Sg.Fujita, ‘Band structure parameters of Zn_1-xCd_xSe investigated by spin-flip Raman spectroscopy’, J.Crystal Growth, in press, 2000.

9. J.J.Davies, D.Wolverson, O.Z.Karimov and I.J.Griffin, ‘Spin-flip Raman scattering studies of II-VI heterostructures’, J.Crystal Growth, in press, 2000.

10. J J Davies, D.Wolverson, I.J. Griffin, O.Z.Karimov, C L Orange, D.Hommel and M.Behringer, ‘Investigation of the gyromagnetic ratios of electrons confined in quantum wells in ZnSe/ZnMgSSe heterostructures’, submitted to Physical Review B.

     

    

The support of the EPSRC as well as the support of the Royal Society, the British Council and the Deutscher Akademischer Austauschdienst is gratefully acknowledged.