Researchers from the Electron Microscopy group resolved the fundamental ambiguity behind the formation of misfit dislocations in strained c-plane wurtzite layers. The new findings, published in the Journal of Applied Physics, will help to optimise, in particular, the growth of AlGaN/GaN heterostructures for the deep-UV emitters and achieve either films with low defect density or, on the contrary, to intentionally promote plastic relaxation in such structures for strain-engineering purposes.
The scientists from the Leibniz Institute for Crystal Growth considered the crucial step of misfit dislocation nucleation and investigated the influence of different growth modes on the strain relaxation process. Their work provides for the first time a reliable quantitative model for the plastic relaxation process of strained c-plane wurtzite films and allows to accurately predict the critical thickness as dependent on the actual surface morphology.
Whether to relax the strain between two layers caused by the lattice mismatch or to grow pseudomorphic defect-free heterojunctions, it is critical to understand under which conditions and how misfit dislocations form in the material. Previous quantitative studies of plastic relaxation in wurtzite films adopted the classical models deduced from cubic materials (InGaAs on GaAs, SiGe on Si). In these works, the authors only considered the energy balance aspect, i.e. the moment when plastic relaxation becomes energetically favourable, while completely neglecting the nucleation process of misfit dislocations itself. However, for wurtzite materials, it is particularly important to consider this process, due to the limitation of how dislocations can move within the wurtzite crystal lattice. As an outcome, a significant discrepancy between theoretical predictions and experimental observations of plastic relaxation in III-nitride thin films has existed for many years.
The current study combines thorough experimental analysis of dislocation formation in AlGaN/GaN heterostructures at each growth step by transmission electron microscopy (TEM), scanning electron microscopy (SEM) and atomic force microscopy (AFM) technics with theoretical calculations of the stress distribution at the dislocation nucleation sites obtained from the finite element method.
“We have found, that for c-plane wurtzite films only the presence of three-dimensional surface structures – islands’ edges and corners, crack fronts, surface macro-steps or V-pits – and the high concentration of the shear stress at these sites enables the nucleation and spread of misfit dislocations in the interfacial plane. This provides an efficient pathway for the plastic relaxation.” – says Dr. Toni Markurt, the author of the work and a researcher from the Leibniz Institute for Crystal Growth.
More important, according to the authors, the critical thickness of the films strongly depends on the surface morphology and the actual geometry of surface structures: for example, two-dimensional layers with cracks can be grown thicker than films with three-dimensional islands, before undergoing plastic relaxation.
The work was done in a close collaboration with OSRAM Opto Semiconductors, who provided dedicated samples for the experimental study.
On the 9th of June, the IKZ again opened the doors for the visitors of the Long Night of Sciences – an annual event that aims to increase the awareness about science and technology.
Despite the tropical heat in Berlin, 1452 visitors came on Saturday evening to the institute to learn about crystal growth, modern crystals and their application in technology.
By combining the laser and crystal growth demonstrations with the tours through the growth halls and scientific lectures, we managed to keep the balance between the “wow”-effect and more serious food for thought for the inquisitive guests.
Researchers of the Center for Laser Materials (ZLM) at the IKZ obtained a direct yellow emitting solid-state laser with record efficiency based on a fluoride crystal doped with terbium ions (Tb3+). Since the emission wavelength of this Tb3+-laser is very close to the sodium (Na) absorption D-lines at 589 nm, it may find applications in fields that require sodium detection or excitation such as astronomical optics (e.g., laser guide stars) or microscopy.
Direct laser emission in the visible spectral range is often not easy to achieve. Green laser pointers, for example, rely on a complicated nonlinear mechanism converting infrared (1064 nm) to green light (532 nm), which limits the total device efficiency and the battery lifetime. Other visible lasers are based on organic dyes, which are highly impractical gain media due to their liquid and often toxic nature.
Therefore one of the ZLM research topics is the investigation of rare-earth doped solid-state crystals for lasers directly emitting within the visible range of the electromagnetic spectrum.
In the past, there has been tremendous progress in red, orange and green emitting solid-state lasers with trivalent praseodymium (Pr3+) as an active ion. However, this ion is not suitable for generating emission in the yellow spectral range. Therefore, recently Tb3+-doped fluoride crystals were suggested as laser active materials.
For a long time, terbium was not considered as a good candidate for laser operation due to its low capability to absorb and emit light and a high risk of detrimental parasitic absorption, intrinsic to this ion. In their recent work, ZLM researchers successfully circumvented these limitations by using fluorides as host crystals and significantly increasing the concentration of Tb3+-ions in these hosts.
This approach led to the world’s most efficient direct yellow emitting solid-state laser, pumped by a blue semiconductor laser. A 28% Tb3+-doped lithium lutetium fluoride crystal (Tb:LLF) enabled an output power of 0.5 W at a laser wavelength of 588 nm with an efficiency as high as 25%. The power of the available blue pump light currently limits the output power of the laser, but the prospects for future improvement are good since LLF is a well-established host crystal known to withstand high powers when doped with other laser ions.
In March 2018, the department "Layers and Nanostructures" of the Leibniz-Institute for Crystal Growth (IKZ) successfully completed the EU project CHEETAH. For more than four years this joint scientific project, created within the 7th European Framework Program, supported the development of new photovoltaic technology at every stage of the value chain. The goal was to develop a new process, which will reduce the costs by saving material and increase energy conversion efficiency, in comparison with the existing multi- and polycrystalline silicon-based technology. The project brought together the expertise of 33 member institutions of the European Energy Research Alliance (EERA), including the IKZ.
Firstly, the IKZ department has shown the feasibility of crystalline Si layers grown on thin films of reorganized porous silicon and glass in application to solar cells (fig. 1), which have potential advantages over the conventional Si wafers. In a second work package, they have developed a method to grow insular Cu(InxGa1-x)Se2 (CIGSe) micro-crystals (about 50 µm in size) at defined locations on a glass substrate. These structures are the basis for the development of cost-effective CIGSe micro-concentrator solar cells.
In addition to research activities, the promotion of young European researchers and the establishment of a network for long-term cooperation in the field played a significant role at CHEETAH. The project financed two postdoctoral positions at the IKZ and provided five young scientists the opportunity to make short research stays and share their results at workshops and conferences in the partner institutions, as well as in the USA, Singapore, and Japan.
The direct research partners of the IKZ for the development of thin-film silicon solar cells were IMEC (Belgium), INES (France), SINTEF (Norway), ECN (the Netherlands) and ISE (Germany). In the research of CIGSe micro-concentrator solar cells, the institute cooperated with ENEA (Italy), University of Estonia, INL (Portugal) as well as with Helmholtz-Zentrum and Federal Institute for Materials Research and Testing (BAM) in Berlin, Germany.
Fig 1.: Growth principle of silicon epitaxial layer on reorganized porous Si
Lateral Photovoltage Scanning (LPS) and Scanning Photoluminescence (SPL) are among the key methods used worldwide and in IKZ to visualise electrically active crystalline defects in silicon (Si) or germanium (Ge) samples. The recent upgrade of these systems allows to simultaneously analyse the distribution of both resistivity gradients and structural defects in the crystals.
The latter is well known to reduce charge carrier lifetime in semiconductors – the decisive characteristic of wafers used later on for photovoltaic solar cells and other micro- and nano-electronic devices.
LPS and SPL methods were created and established at IKZ several years ago, and since then our researchers further develop and routinely use these methods to image various defects in Si and Ge crystals, such as grain boundaries, dislocations or growth-induced doping inhomogeneities (striations). The new detection and evaluation mode extends the functionality of these measurement techniques and allows to map the charge carrier lifetimes while scanning across the sample.
The possibility to visualise defects, resistivity inhomogeneities, and carrier lifetime distribution at the same time on the same sample area is the main advantage of the upgraded combined LPS & SPL system. It allows avoiding uncertainties in data interpretation, which might happen, e.g., if the measurements are carried out separately on different stations. Moreover, the new system is equipped with two solid-state lasers with a variable laser power from 1 µW to 100 mW, and the laser beam can be focused down to 5 µm, providing the flexibility of the experiment and precise localisation of recombination centres on the image.
Structural and deep-level impurity defects are among the most important causes of charge carrier recombination events in crystals. When an excited electron-hole pair is trapped by such a defect, it may eliminate – i.e., recombine – and release energy, either thermal or in the form of photons, instead of contributing to the desired (opto)-electronic performance of the device. The longer the charge carriers can stay around before recombining – the longer is their “lifetime” and the better the resulting device performance. Generally, the recombination centres have a detrimental effect on the quality of the wafers used in photovoltaics or electronics. Thus the possibility to visualize the recombination centres and to understand their origin is a significant part of the crystal growth research and development.
The LPS / SPL research work is carried out at IKZ in close collaboration with the company LPCon: https://www.lpcon.com/
In the case of the perovskite prototype CaTiO3, crystal growth using a melt solution is the method of choice. At the IKZ, a mixture of calcium fluoride (CaF2) and titanium (IV) oxide (TiO2) was identified as an advantageous solvent for these crystals.
Calcium titanate (CaTiO3) occurs in the nature as a mineral. However, the direct growth from the melt is impossible since phase transition at 1625 K leads to strong formation of twin defects and thus to damage of the crystals. This problem can be avoided by adding so-called melt solution. This is the substance with a relatively low melting point, which is dissolved in the crystallising substance in liquid phase. During cooling down, the melt solvent is excreting the crystallising component out again.
However, the search for suitable melt solvents is often an almost alchemical procedure. For CaTiO3, potassium fluoride and lead (II) fluoride have been described in the literature. However, both have the disadvantage that only very small portions of CaTiO3 (about 1:12) can be dissolved in them. Furthermore, undesired chemical reactions between the solvent and CaTiO3 lead to contamination of the solvent.
As a part of a master's thesis at the IKZ, a mixture of the lead-free substances calcium fluoride (CaF2) and titanium (IV)-oxide (TiO2) was identified to be a more advantageous solvent for CaTiO3. From a mixture of these substances with the molar ratio 3:1:1 (green dot on the figure) CaTiO3 crystallises below the critical phase transformation, resulting in still small (approx. 2.5 mm edge lengths), but high quality crystals. The results were obtained on the basis of extensive thermoanalytical measurements and a thermodynamic model of the ternary phase diagram based on them. The proportion of dissolved CaTiO3 has been improved to 1:4, no undesirable chemical reactions have been detected and thus no contamination is present.
Although pure CaTiO3 has little technical relevance, the knowledge and understanding of its properties is of fundamental importance. A number of important ferroelectric materials (e.g. barium titanate, (potassium, sodium) niobate etc.) and other functional materials, such as substrates for oxide electronics (e. g. strontium titanate, rare earth scandates etc.) crystallize in the perovskite or related crystal structures. Therefore, the supply of CaTiO3 high quality single crystals as suitable model systems to answer fundamental questions is essential for basic investigations.
The article is published in the Journal of Crystal Growth.
International researchers have revealed the mechanism that limits the indium (In) content in indium gallium nitride ((In, Ga)N) thin films - the key material for blue light emitting diodes (LED).
Increasing the In content in InGaN quantum wells is the common approach to shift the emission of III-Nitride based LEDs towards the green and, in particular, red part of the optical spectrum, necessary for the modern RGB devices. The new findings answer the long-standing research question: why does this classical approach fail, when we try to obtain efficient InGaN-based green and red LEDs?
Despite the progress in the field of green LEDs and lasers, the researchers could not overcome the limit of 30% of indium content in the films. The reason for that was unclear up to now: is it a problem of finding the right growth conditions or rather a fundamental effect that cannot be overcome? Now, an international team from Germany, Poland and China has shed new light on this question and revealed the mechanism responsible for that limitation.
In their work the scientists tried to push the indium content to the limit by growing single atomic layers of InN on GaN. However, independent on growth conditions, indium concentrations have never exceeded 25% - 30% – a clear sign of a fundamentally limiting mechanism. The researchers used advanced characterization methods, such as atomic resolution transmission electron microscope (TEM) and in-situ reflection high-energy electron diffraction (RHEED), and discovered that, as soon as the indium content reaches around 25 %, the atoms within the (In, Ga)N monolayer arrange in a regular pattern – single atomic column of In alternates with two atomic columns of Ga atoms. Comprehensive theoretical calculations revealed that the atomic ordering is induced by a particular surface reconstruction: indium atoms are bonded with four neighboring atoms, instead of expected three. This creates stronger bonds between indium and nitrogen atoms, which, on one hand, allows to use higher temperatures during the growth and provides material with better quality. On the other hand, the ordering sets the limit of the In content of 25%, which cannot be overcome under realistic growth conditions.
The work is a result of a collaboration between Leibniz-Institut für Kristallzüchtung (Berlin, Germany), Max-Planck-Institut für Eisenforschung (Düsseldorf, Germany), Paul-Drude Institut für Festkörperelektronik (Berlin, Germany), Institute of High-Pressure Physics (Warsaw, Poland), and State Key Laboratory of Artificial Microstructure and Mesoscopic Physics (Beijing, China).
Read the full press release.
The article is published in:
© IKZ 1995 - 2018 Berlin