IKZ logo logo of TU Berlin Crystal Growth of Advanced Materials
Homepage of Prof. Dr. M. Bickermann
Past research activities

My research activities in Erlangen (1996-2011)

The AlN research team in Erlangen

On his appointment as the Chair of Materials Science 6 at the University of Erlangen in 1992, Prof. Dr. Albrecht Winnacker's research activities focused on sublimation growth of bulk crystals of wide bandgap materials and their characterization. During the nineties, the team developed the preparation of SiC bulk single crystals what eventually led to the formation of the spin-off company SiCrystal AG. I studies Materials Science at this chair and took my share in developing SiC growth technology in my diploma theses 1998 and my PhD research work that was finshed in 2002 (see also below).

Starting in 2001, Boris Epelbaum and me worked to establish bulk single crystal growth of AlN. For this we used the SiC grwoth stations and soon after, we developed and built our own machines. At that time, AlN bulk growth was still in its infancy, and only two teams in the U.S. (and one in Russia) were investigating this topic. I was active in crystal growth and later I focused on building the growth stations and characterizing the AlN samples. Aside from that I advanced my academic career by giving lectures and providing seminars and live exercises at the university. In 2008, I finished my 'habilitation', a German qualification as a lecturer, on growth and characterization of AlN single crystals.

The AlN research team in Erlangen

By the time, Dr. Octavian Filip and Paul Heimann joined our team (see group photo), as well as Dr. Shunro Nagata who worked for a joint development with his Japanese company. Due to active help from technicians and students at the chair of Materials Science, we were able to advance our knowledge and technology to eventually found the spin-off company CrystAl-N GmbH in 2010.

I substantially contributed to the successful application for funding through the German EXIST-Forschungstransfer (research transfer) programs to prepare a spin-off company. As a founding partner my role in the company was the characterization of the crystals and the involvement of students in scientific aspects of the technology. Therefore I continued by Work and employment at the University of Erlangen. Finally, the successful appointment as a professor at the Technische Universität Berlin in 2011 ended my direct work for the company, and after initial success the company went out of business in 2018. At the IKZ my activities are still in the field of applied research, but dealing with pre-industrial research and development.

The company Firma CrystAl-N GmbH

AlN wafer and logo of the company CrystAl-N GmbH

The company CrystAl-N GmbH was a spin-off to prepare the commercial use of our technology for the production of AlN crystals and substrates. The preparation to found the spin-off was funded by the "Exist Research Transfer" program of the Federal Ministry of Economics and Technology. In 2009, we successfully participated in the Northern Bavaria network's business plan competition, winning of the first and second rounds and being finalist of the third round! With our business idea, we also became a finalist in the 2010 Bavarian Founder Prize and winner of the Middle Franconian Founder Prize.

The founding persons and the University of Erlangen were invested right from the start, followed later by other personal investors, HighTech Gründerfonds and BayernKapital SeedFonds Bayern. The company produced AlN crystals on SiC seeds and commercially sold AlN substrates. In 2013, the company moved from the Erlangen-Tennenlohe start-up centre to the "Uferstadt" technology park in Fürth. Unfortunately, no industrial company was ultimately interested to transfer the technology.

AlN - Research Summary

Die Entwicklung von AlN in Erlangen

Growth and characterization of AlN bulk single crystals (my habilitation thesis 2008)

In recent years, single-crystalline aluminum nitride (AlN) became a candidate as substrate material for group-III nitride epilayers. Devices based on such epilayers constitute a rapidly emerging market in the field of optoelectronics and high-frequency communications. Particularly UV optoelectronics would greatly benefit from the use of AlN substrates, if these were available in sufficient amount, quality, and size. That is still not the case.

Bulk AlN crystals are most successfully grown by a sublimation-recondensation process (PVT growth) at temperatures exceeding 2000°C. A set-up based on tungsten parts was found to be most stable against the aggressive vapor and provides very low contamination of the growing crystal. We show that mass transport and unidirectional solidification may be easily established. However, grain selection and enlargement is very slow, as the crystals tend to form a columnar structure. Under nearly isothermal conditions, freestanding crystals of up to 25 mm in size and high structural quality have been fabricated. But further enlargement seems infeasible, and their special habit renders them unsuitable for preparation of substrates.

Large-area AlN single crystals are obtained by seeding on SiC substrates, alas at the cost of severe problems in material compatibility which lead to highly contaminated AlN crystals. As the set-up degrades in the presence of SiC during long-time growth runs, the achievable crystal thickness is limited. Still, such large-area AlN crystals have been successfully used as templates for further growth of AlN single crystals up to 30 mm in diameter and 10 mm in height.

In the main volume of such crystals, the density of dislocations threading the basal plane is lower than 105 cm-2. However, low angle grain boundaries are inherited from the SiC seed and the template, leading to a mosaic structure of elongated subgrains which may deviate by more than 1° from the main orientation. Structural quality improves in crystals that were grown under optimized conditions.

According to chemical analysis, oxygen is found to be the dominating impurity regardless of growth orientation. Additionally, silicon and carbon are clearly detectable even in bulk AlN grown homoepitaxially on the templates. The wurtzite structure of AlN is highly anisotropic. As growth takes place on different facets simultaneously, the resulting crystal shows a zonal structure. Different zones of AlN single crystals showed very similar threading dislocation densities and topographies. On the other hand, impurity incorporation and formation of intrinsic defects clearly depends on the zone or growth orientation. The differences are evidenced by optical absorption and cathodoluminescence.

We show that both intensity and peak position of UV optical absorption are closely linked to the 'violet luminescence' in the 3-4 eV energy range. Comparing our data with experimental results and first principles calculations reported in literature, we establish a model of the defect content in nominally undoped AlN crystals. As a conclusion, oxygen contamination as well as aluminum vacancy formation increases in the order Al-polar (0001) zone, rhombohedral {1012} zones, prismatic {1010} zones, N-polar (0001) zone. In the same sequence, the absorption at 2.8 eV and thus the yellowish coloration of the crystal areas increases.

The model further suggests a DX-center formation of oxygen, whose thermal activation energy of 0.6-0.8 eV is evidenced by thermally stimulated luminescence and high-temperature resistivity measurements. In contrast, the existence of nitrogen vacancies in significant concentrations is disputed at least for samples in which the oxygen content governs the electrical behavior.

In conclusion, a procedure for homoepitaxial growth of AlN bulk crystals is developed. The structural quality of the resulting substrates is very promising for group-III nitride epitaxy. Additionally, important insights into the defect occurrence and distribution in this novel material are obtained, which may aid in further optimization of crystal growth in regard to applicability e.g. in UV optoelectronics.

SiC - Research Summary

Various doped SiC crystal slices from my research in Erlangen

Growth of semi-insulating SiC bulk single crystals (my PhD thesis 2002)

The goal of this work, the production of semi-insulating silicon carbide single crystals, was realized by vanadium doping. By the appropriate addition of vanadium to the SiC source material during the growth of SiC crystals according to the modified-Lely method, the reproducible production of SiC crystal discs with homogeneous electrical properties, i.e. specific resistances at room temperature of more than 1010 Ohm-cm, with a high yield is possible. As a substrate material, these crystal discs can significantly improve the electrical properties of high-performance and high-frequency components.

A detailed investigation of all known impurities in SiC leads to the conclusion that the vanadium donor is best suited to compensate flat dopants by its band-centered impurity and to generate semi-insulating behavior with more than 1015 ohm-cm at room temperature. Since vanadium acts as an amphoteric dopant in SiC, it is necessary to generate a slight excess of flat acceptors in the crystal to compensate them with vanadium. On the other hand, the activation of the vanadium acceptor level at compensation with flat donors also leads to semi-insulating behavior, but at lower specific resistances at room temperature in the order of 1011 Ohm-cm.

The impurities introduced during nominally undoped growth are decisive for the development of the manufacturing process for semi-insulating SiC. Nitrogen, which is molecularly adsorbed on the materials used in the cultivation process, dominates the electrical properties as a flat donor. It desorbs during growth and is built into the crystal in a time-inhomogeneous manner, decreasing exponentially with growth time. Through the development and improvement of a high vacuum heating step prior to cultivation, it was possible to reduce the nitrogen concentration in crystal regions remote from the germ to up to 4 x 1016 cm-3, but at the beginning of cultivation more than 1 x 1018 cm-3 was measured.

With the aim of producing a slight excess of acceptors in the crystal, the doping with boron was investigated by adding boron carbide directly to the SiC powder. This process is efficient - the transfer coefficients are k = 0.1 depending on the configuration ... 0.22 - and enables constant boron incorporation during growth. This was demonstrated by the evaluation of temperature-dependent Hall effect measurements on several crystals with different boron additions. The electrical properties of boron-doped crystals are clearly influenced by the time inhomogeneous nitrogen incorporation, especially with low boron addition. Nevertheless, the production of boron-doped crystal discs with homogeneous electrical properties has been successful.

Doping with vanadium was also realized by adding vanadium carbide to the SiC powder. This process is inefficient because the vanadium species evaporate rapidly and the vanadium concentration in the crystal decreases by up to three orders of magnitude during growth. By the addition in an inner crucible and a slight lowering of the growth temperature the evaporation of the vanadium species can be limited and an almost homogeneous vanadium incorporation during the growth can be achieved. The formation of vanadium carbide precipitates can also be limited, which occur due to the exceeding of the maximum solubility of vanadium of 5 x 1017 cm-3 especially at the beginning of the growing process in the crystals.

In vanadium-doped crystals, the electrical properties are determined by the vanadium acceptor, which compensates the flat nitrogen donors. Absorption and ESR measurements can be used to identify the active vanadium charge states. Measurements of the Hall effect and the specific resistance confirm homogeneous electrical properties as well as semi-isolating behaviour with charge carrier concentrations in the range of 108 cm-3, specific resistance values around 1011 Ohm-cm and an activation energy of EA > 700 meV at room temperature in the crystal disks.

Crystals with co-doping of vanadium and boron were produced in which the electrical properties are determined by the vanadium donor. The usable semi-insulating range has activation energies between 1.2 eV and 1.8 eV and exhibits extremely high specific resistances limited at room temperature by leakage currents and poor contacts. It is limited by the high nitrogen concentration and precipitate formation at the beginning of cultivation and by the decreasing vanadium incorporation at the end of cultivation. By optimizing the co-doping, it has been possible to significantly increase the semi-insulating range in the crystals.

The highly pure semi-insulating SiC without vanadium doping developed by some research groups, in which the compensation is performed by intrinsic defects, shows ways to further optimize the production of semi-insulating SiC single crystals: A further reduction of the impurity concentrations and the investigation of alternative impurities for compensation in volume crystal growth, together with the improvement of the homogeneity of the vanadium incorporation, can be regarded as the most urgent tasks.