M. Zorn

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.

Surface processes before and during growth of GaAs (001)

GaAs(001) surfaces in both metalorganic vapour phase epitaxy (MOVPE) and molecular beam epitaxy (MBE) have been studied under As-stabilization as well as during growth by reflectance anisotropy spectroscopy (RAS). During MBE growth simultaneously reflection high-energy electron diffraction (RHEED) measurements were performed. With increasing growth rate, the surfaces transform from a c(4 × 4) to a (2 × 4) in MBE. In MOVPE, with trimethylgallium (TMGa) and AsH 3 a transition occurs from a «c(4 × 4)-like» behaviour to a surface best described as a mixture of dielectric contributions from Ga-rich «(1 × 6)-like» and the As-rich «c(4 × 4)-like» regions. Time resolved RAS measurements show oscillations with monolayer periodicity in both epitaxial systems, however, with opposite phase. In MOVPE additionally data with time and spectral resolution are obtained which give insight into the surface modifications during a monolayer growth cycle. A model using an effective medium approach is proposed which relates the RAS oscillations during MBE growth to morphological changes associated with the island formation in the layer-by-layer growth modus. The MOVPE oscillation behaviour is more complex due to the final decomposition steps of trimethylgallium at the surface. In addition to the island formation it has to be included in the model that the surface oscillates between a state «more c(4 × 4)-like As dimers» and a state «more (1 × 6)-like As and Ga dimers»

In situ time‐resolved monitoring of PH3 induced exchange reactions on GaAs under metalorganic vapor phase epitaxy conditions

Exposure of GaAs and InGaAs to PH3 is a standard step in gas switching sequences for metalorganic vapor phase epitaxy (MOVPE) growth of heterostructures in the technologically important GaAsP, InGaP, and InGaAsP material systems. The exchange of group‐V atoms was monitored in situ by reflectance anisotropy spectroscopy when GaAs is exposed to PH3. The c(4×4) reconstructed, As‐terminated GaAs surface is then replaced by a P‐terminated structure. At standard MOVPE growth temperatures and pressures the time constant for this reaction is of the order of 100 ms. The temperature and pressure dependence of the As by P exchange is reported, and the activation energy was determined to be 1.64 eV. It is concluded that PH3 enhances the desorption of As.

Growth oscillations with monolayer periodicity monitored by ellipsometry during metalorganic vapor phase epitaxy of GaAs(001)

In this letter we report on the observation of growth oscillations with monolayer periodicity by ellipsometry. An oscillation amplitude of δ〈e1〉=0.05 was measured using an optimized spectroscopic in situ ellipsometer whose wavelength was tuned to the 2.65 eV resonance energy of the arsenic dimers covering the GaAs (001) growth surface. The monolayer periodicity was verified by parallel monitoring of the growth with reflectance anisotropy spectroscopy (RAS).

In-situ reflectance anisotropy studies of ternary III–V surfaces and growth of heterostructures

Abstract In this paper we describe reflectance anisotropy spectroscopy (RAS) studies of III–V ternary semiconductor alloys with the mixing of the group V element (GaAs - x P x ) as well as the group III element (In x Ga 1- x As). The surface anisotropy turns out to be highly sensitive to the composition of the uppermost monolayers of the ternary alloys. The RAS spectral shape changes with stochiometry x of GaAs 1- x P x as well as In x Ga 1- x As.. Consequently, RAS can be utilized for the determination of the stochiometry. However, several other effects such as thickness, interface and strain influence the spectra and have to be taken into consideration. Oscillation in the time resolved RAS signal with monolayer periodicity are observed for the growth of In x Ga 1- x As on GaAs with low indium content ( x ≤ 0.2) for the first time. They can not only be used for in-situ measurements of the thickness and the growth rate, but also for in-situ measurements of the composition. This was verified by using X-ray, ellipsometry and reflectance measurements. Moreover, by counting the growing monolayers in-situ via RAS oscillations the first monolayer oscillation controlled growth of a InGaAs/GaAs superlattice in a metalorganic vapour phase epitaxy (MOVPE) reactor is reported.

Reflectance anisotropy spectra for the transition from the P-rich to the In-rich surface reconstruction of InP(100)

Abstract Reflectance anisotropy (RA) spectra of InP(1xa00xa00) surfaces were taken either in the MOCVD environment or, after contamination-free transfer to ultra-high vacuum (UHV), in UHV down to 20xa0K. They were correlated with photoemission and Auger electron spectroscopy measurements to investigate the transition from the P-rich to the In-rich surface reconstruction. The strongest surface stoichiometry-induced changes were found in the RA-spectra at around 3xa0eV photon energy, i.e. in the range of the surface-modified bulk E 1 transition. At around 1.8xa0eV, the main peak of a pronounced surface transition was recorded with equal magnitude and sign for the P-rich and for the In-rich surface reconstruction. Two different specific RA-spectra measured with the highest peaks are postulated here to indicate the ordered P-rich and ordered In-rich surface reconstruction, respectively.

In situ surface passivation of III–V semiconductors in MOVPE by amorphous As and P layers

Abstract A new process for chemical passivation of III–V semiconductor surfaces in metalorganic vapour phase epitaxy (MOVPE) is developed. A passivation layer is deposited directly after growth in the reactor. It consists of amorphous arsenic or a double-layer package of amorphous phosphorus and arsenic, which are grown by photo-decomposition of the group-V hydrides. These layers (caps) serve to protect the surfaces against contamination in air after removing the samples from the MOVPE growth reactor. Such passivation is applicable e.g. for a two-step epitaxy or for further surface characterizations.

Efficiency of arsenic and phosphorus precursors investigated by reflectance anisotropy spectroscopy

Abstract The efficiency of several alternative group-V precursors was measured throught the As or P related structures in the reflectance anisotropy spectra from (001) surfaces of GaAs and InP. As arsenic precursors were investigated: arsine (AsH 3 ), tertiarybutylarsine (tBAsH 2 ) and triethylarsine (TEAs). The strength of the As-related feature at 2.5 eV in the reflectance anisotropy spectroscopy (RAS) spectrum of the typical metalorganic vapour phase epitaxy (MOVPE) pregrowth surface GaAs(001)-c(4 × 4) was measured as a function of precursor partial pressure and temperature. From these measurements, a relative efficiency for the arsenic supplied to the surface for the different precursors can be given. A higher efficiency of tBAsH 2 as compared to AsH 3 at 723 and 823 K, but equal efficiencies at 923 K, for both compounds are observed. For TEAs at lower temperatures (723 to 823 K), a new RAS spectrum different from the one for the c(4 × 4) is obtained. This reveals a surface different from the As double layer due to TEAs derivatives absorbed on the surface. At higher temperatures (923 K), a c(4 × 4)-like RAS spectrum is obtained indicating that at this temperature predominantly As is supplied from TEAs to the surface. Using both TEAs and AsH 3 simultaneously, the additional adsorbate structure disappears also at lower temperatures. This effect is attributed to the reaction of atomic hydrogen, derived from arsine, with the organic TEAs derivatives. The efficiencies of the alternative P precursors were evaluated through the P-related peak at 2.7 eV in the RAS spectrum of the phosphorus-rich InP(001)-(2 × 4). At a temperature of 873 K, the precursor tertiarybutylphosphine tBPH 2 revealed a much higher P efficiency than PH 3 . In contrast, with tetraethyldiphosphine (TEDP) no P-rich (2 × 4)-like spectrum could be obtained but rather an In-rich (4 × 2) spectrum was indicated. This reveals a much lower P efficiency for TEDP than for two other P precursors. Accordingly, TEDP seems to be less well suited for MOVPE of P-containing compounds than tBPH 2 .

Spectroscopic ellipsometry applied for in-situ control of lattice matched III-V growth in MOVPE

Abstract Stoichiometry control of lattice matched compound semiconductors during growth is still a challenge for in situ optical techniques. This is because the energy shifts of the semiconductor optical gap induced by strain and stoichiometry compensate one another. Thus conventional critical point analysis for composition measurement is practically impossible. The remaining response of the measured ellipsometric parameters is very small but can be resolved when the signal-to-noise ratio of the ellipsometer is sufficiently high. We exploit this for the closed-loop control of In x Ga 1− x As growth on InP where the trimethylindium (TMIn) flux was directly controlled by the measured optical response in order to reach the desired lattice matched composition ( x =0.53). Ex-situ X-ray diffraction analysis verified ellipsometrically controlled lattice matched growth.

In situ monitoring and control of InGaP growth on GaAs in MOVPE

Abstract The growth of InGaP on GaAs was investigated and controlled by in situ reflectance anisotropy spectroscopy (RAS) and spectroscopic ellipsometry (SE). In contrast to other III–V semiconductors, the CuPt B -type ordering of the group III sublattice in the InGaP material system causes a characteristic bulk anisotropic contribution to the optical constants. This contribution was determined by RAS between room and growth temperature. Based on this result the influence of the composition near the lattice matched composition on the optical data was studied by SE and used for closed-loop controlled growth of lattice matched InGaP on GaAs. For this purpose a photon energy of 3.5xa0eV was used where the contribution of the bulk ordering to the optical constants is negligible. Closed-loop controlled growth of lattice matched In 0.48 Ga 0.52 P on GaAs, i.e., the flux of the indium source was controlled directly by the ellipsometry computer, resulted in a lattice mismatch of 2×10 −4 .