P. Kurpas

Real-time monitoring of MOVPE device growth by reflectance anisotropy spectroscopy and related optical techniques

Abstract Reflectance anisotropy spectroscopy (RAS/RDS) so far has been mostly used for basic growth studies in both molecular beam epitaxy (MBE) and metal-organic vapor-phase epitaxy (MOVPE). Due to its sensitivity to the uppermost atomic monolayers, RAS became a very versatile tool for investigating surface stoichiometry, surface reconstruction and surface morphology especially under gas-phase conditions. Meanwhile, however, the performance and adaptability of RAS to standard MOVPE systems has been enhanced significantly and RAS sensors now can also be used for MOVPE device growth monitoring and control. Therefore, after a brief introduction to the basic surface physics and surface chemistry causing the optical signatures, this paper concentrates on device related applications. Examples will be given concerning the optical response to both n-type and p-type GaAs doping levels and the real-time measurement of ternary compound composition for reaching lattice matched growth. The optical surface response during the growth of a complete GaAs/InGaP heterojunction bipolar transistor is visualized. The result indicates on a monolayer level either consistency or deviation from the intended growth process.

Carbon doping for the GaAs base layer of Heterojunction Bipolar Transistors in a production scale MOVPE reactor

In this work different approaches for carbon doping of GaAs in MOVPE are compared with respect to their growth-and device-related material properties. Doping levels up to 6 x 10 19 cm -3 and smooth surface morphologies are achieved with either intrinsically (TMG and AsH 3 or TMAs) or extrinsically (CBr 4 ) doped layers. Despite comparable structural and majority carrier properties differences in GaInP/GaAs-HBT device performance depending on base doping conditions are obtained. Devices with an intrinsically doped base layer (TMG + AsH 3 ) show superior transistor performance with a current gain to base sheet resistance ratio (β/R sb ) exceeding 0.5 for base thicknesses as large as 120 nm. The use of either CBr 4 or TMAs as base growth precursors results in reduced current gains (β/R sb ≤ 0.3). It is shown that the achieved HBT current gain is directly related to recombination centers in the heavily doped base layer depending on doping method.

Growth monitoring by reflectance anisotropy spectroscopy in MOVPE reactors for device fabrication

Abstract Reflectance anisotropy spectroscopy (RAS) has proved its capability to study surface processes during metalorganic vapour phase epitaxy (MOVPE) growth of a variety of III–V compounds. However, these investigations up to now have been mostly restricted to specialized research reactors. Therefore, we studied the feasibility of in-situ monitoring by RAS during growth on two production-type MOVPE reactors: horizontal 2 inch single wafer reactor AIX 200 and Planetary Reactor™ AIX 2000 for 5 × 3 inch. The slight modifications of the reactors necessary to gain normal incidence optical access to the sample do not alter the properties of the grown materials. While in the horizontal reactor the strain-free optical window allows one to obtain well-resolved RAS spectra the signals in the multiwafer reactor are affected by the anisotropy of the ceiling plate. Even in this case RAS spectra can be extracted. First measurements on rotating samples in the horizontal reactor demonstrate the possibility to obtain RAS spectra by multitransient spectroscopy. As an application monitoring of the growth of p-type layers for the base of GaInP GaAs hetero-bipolar-transistors (HBTs) is discussed. The linear electro-optic effect (LEO) gives information on doping type and doping level. Time-resolved transients at specific energies are used to study the impact of different switching schemes on the properties of the base-emitter interface.

Hydrogen in carbon-doped GaAs base layer of GaInP/GaAs heterojunction bipolar transistors

Abstract Hydrogen incorporation into heavily carbon-doped GaAs grown by metal-organic vapour phase epitaxy using carbon tetrabromide (CBr 4 ) has been studied. In the base layer of as-grown GaInP/GaAs heterojunction bipolar transistors (HBTs), about 20% of the carbon acceptors are found to be passivated by hydrogen. The outdiffusion of this hydrogen during an ex situ annealing at 450 °C in nitrogen, which is effective for carbon-doped single layers, is blocked by n-type capping layers in HBTs. An in situ annealing step was found to be suitable to reduce the acceptor passivation in HBTs to about 10%.

MOVPE growth of GaInPGaAs hetero-bipolar-transistors using CBr4 as carbon dopant source

Carbon doping of GaAs with carbon tetrabromide (CBr 4 ) in low pressure MOVPE has been investigated and applied to the fabrication of GaInP/GaAs HBTs. Especially the hydrogen incorporation and the associated acceptor passivation has been studied. The hydrogen found in single GaAs:C layers is predominantly incorporated during cooling the sample under AsH 3 after growth. n-Type capping layers can block this H indiffusion and GaAs:C base layers in HBTs show much lower H concentrations than GaAs:C single layers without a cap. A further reduction of acceptor passivation is possible by optimization of the growth procedure. First HBTs processed from layers with a base that was doped using CBr 4 show promising DC and HF performance (β = 45, f T = 26 GHz for 2 X 20 μm 2 devices).

Correlation of Ingap(001) Surface Structure During Growth and Cupt B -Type Bulk Ordering

The mechanism causing the CuPt B -type ordering of InGaP grown lattice matched to GaAs was investigated by in-situ reflectance anisotropy spectroscopy (RAS/RDS). Experiments were performed during InGaP growth in metal-organic vapour phase epitaxy (MOVPE). From the experiments it can be concluded that bulk ordering only occurs when InGaP growth is performed under phosphorus-rich (2×1)-like surface conditions. Bulk ordering completely disappears under growth conditions which cause a less-phosphorus-rich (2×4)-like surface dimer configuration.