© 2000 The Materials Research Society
This annual meeting was held on the Campus of the University of Nottingham, Prof J Orton was the Conference Chairman. About 200 participants were registered, and about 30 exhibitors. More than 70 oral papers were given, and in addition 55 posters.
The scientific program covered both basic materials properties and devices with related technology.
The activities in the field of violet lasers in Europe are clearly behind Japan and USA, but encouraging progress is reported, in particular from German and English groups. The developments in terms of ir lasers for 1.3 um is more advanced, and a German group reported very interesting results on InGaNAs/GaAs quantum well (QW) lasers. The high temperature performance was so far evidenced with a threshold current parameter To of 110 K at 100 C. For VCSELs in the ir region lasing was obtained with optical pumping only, while a US group has recently reported on such a laser with electrical injection.
An interesting report was given from a leading Japanese group on the development of lasers based on selforganised InGaN/GaN quantum dots (QDs), as well as GaN/AlGaN regular pyramid patterns which had QDs naturally formed at the top of each pyramid. These structures demonstrate a clearly enhanced spontaneous emission β factor of about 0.01 in VCSEL structures, promising that electrically injected VCSELs with a much reduced laser threshold current density compared to the present stripe geometry lasers may be realized soon.
Several groups reported a related development on QD structures for LEDs. Interestingly such structures may be grown on silicon, despite a very high dislocation density, since with an optimized dot size and density (of the order 1011 cm-2) most dots will not encounter any dislocations. Efficient LEDs based on such InGaN/GaN QD structures covering the entire range from blue to red were demonstrated. It was claimed that the controlled deposition of two different sizes of QDs in a multilayer structure grown on silicon would be an attractive concept for white LEDs.
An interesting comparison between hexagonal and cubic MBE-grown AlN/GaN dots was given. It is anticipated that the cubic dots would have a stronger PL since the oscillator strength would not be reduced by any polarization fields. This tendency seems to be borne out by the preliminary experimental data shown. Also, a clear advantage of QD structures vs QW structures was demonstrated: the temperature quenching of the PL intensity was very much reduced for the QD structures up to well above room temperature.
A discussion was given on the development of white LEDs. Apart from the obvious need to drastically increase the radiative efficiency above the present levels for UV LEDs, an important development will be the use of a combination of several phosphors (preferably three phosphors) to get the desired color perception in illumination applications.
Photodetectors are another class of devices that attract a strong interest. In particular various AlGaN/GaN photodiode structures appear very promising, due to their high speed combined with a high detectivity. The importance of a low dislocation density for a sharp spectral cut-off was stressed.
Properties of AlGaN/GaN HEMT transistor structures were reviewed. 2DEG mobilities as high as 2100 cm-2 at RT were reported, although such values are not typical. Dislocations are important electron traps, which are also optically sensitive. Reliability tests are becoming important. It was found that severe problems occur with V-doped SiC substrates, a degradation process is found at high voltage stress in devices grown on V-doped SiC.
Several sessions were focussed on material properties. For InGaNAs QWs recent observations from PLE of the position of excited QW states gives an experimental indication that this QW system is of type I for the heavy holes. For InGaN/GaN QWs it was confirmed that threading dislocations are one obvious source of composition fluctuations in the InGaN QW layers, it seems clear that In does segregate preferentially at these dislocations causing carrier localization. It is not yet clear whether the absence of dislocations in InGaN/GaN QW structures will also possibly reduce the tendency to forms nano-scale QD islands with higher In content.
For AlGaN/GaN MQW structures it was reported that co-doping of Si and In would strongly increase the radiative efficiency. The mechanism for this was not yet clear.
Homoepitaxial GaN layers grown on misoriented bulk GaN substrates with low dislocation density showed interesting PL spectra. The two bound excitons (BEs) at about 3.466 eV (2 K) were found to be of different origin, since only the one at higher energy appears for growth at N-face.
Be-doping of bulk GaN grown at high pressure and high temperature was studied optically. Due to the high doping and compensation no near band edge emissions were observed, only deep level bands.
The structure of seeded bulk GaN crystals grown from the liquid phase at about 1000 C and moderate overpressure (< 2 atm) was studied in detail. A typical crystallite size is 100 μm, with a predominant (0001) orientation, accompanied with cubic inclusions, and with a substantial volume occupied by cavities. This material needs much development before it becomes interesting as a GaN substrate.
Properties of AlN grown by HVPE on SiC and Si substrates were reported. Thick layers (up to 300 μm) were produced, and even freestanding AlN layers with a lateral size of 20 mm and thickness of 1 mm were reported. The material quality was so far not well characterized.
The modeling of recombination in AlGaN/GaN QWs was discussed. At low temperature the recombination is described in terms of localized excitons, strongly affected by the polarization fields. At RT nonradiative processes need to be accounted for to explain the decay times in transient PL.
There is still a controversy about the magnitude of the polarization fields in this system. From AlGaN/GaN HEMT structures a good agreement is observed between the theoretical and experimental values for the polarization coefficients, while from optical studies of MQWs in this system the experimental values are at least a factor two lower than predicted by theory. The MQW data may need to be revisited concerning possible strain relaxation in the samples, and also the surface depletion field should be accounted for properly.
An interesting theoretical result on growth of GaN with simultaneous Si doping was reported. It appears that the total energy for the cubic phase is lowered under such conditions, so that a high Si doping (several percent) might be a way to stabilize the cubic phase during growth.
The Mg acceptor in GaN and AlGaN was studied with magnetic resonance. The g-value is slightly anisotropic in strain free material, with a g-value = 2.065 parallel to the c-axis and g = 2.008 in the perpendicular direction, in sharp contrast to the effective mass expectations for a shallow acceptor. In AlGaN the g factor quickly becomes isotropic with rising Al content. This behavior indicates an anisotropic hole wavefunction, with the hole distributed over the basal plane bonds to the nearest N atoms.
Dislocation properties were also discussed. The glide process was studied in bulk GaN by indentation at 370 C. It was found that for the dislocations formed (with b = 1/3 <11-20>) glide occurs exclusively in polar glide planes like (0001) and {10-11}, since these have a lower Peierls energy. Plastic strain relaxation was also studied in AlGaN on GaN grown by MBE. It was found that relaxation predominantly occurred in part of the layer, while another part remained coherently strained. These studies are relevant for Bragg mirrors to be used in VCSEL structures.
There are previous reports on a negative electron affinity for AlN. A study of this property across the entire AlGaN composition range was reported. The result was that the electron affinity remains positive for all x, with a small margin of about 0.3 eV for AlN.
A study of GaMnN prepared with the AMMONO method was reported. The material obtained contained a mix of different phases, presumably with different Mn composition. The main material was paramagnetic, with small contributions from ferromagnetic and antiferromagnetic phases. Mn retains the d5 configuration, which is promising for the realization of ferromagnetism, provided a very high hole concentration can be realized simultaneously.
B Monemar
|
|
|
|
