Views: 0 | Downloads: 0
Epitaxial integration of multifunctional complex oxides with advanced semiconductor
technology is very important and desirable because it opens the door to a wide range of
novel device applications. A c onvenient way of integrating functional oxides on Si(001)
substrate is through the SrTiO 3 (STO) intermediate layer which can be fabricated on
Si(001) in the epitaxial form and with high crystallinity using mainly the molecular beam
epitaxy (MBE) method. However, slowness of the MBE deposition process and difficult
stoichiometry control makes it an inappropriate tool from i n dustrial point of view.
In this PhD thesis the focus is placed on the possibilities of using the pulsed laser deposition
(PLD) te chnique for the synthesis of an epitaxial STO layer on Si(001) as an alternative
manufacturing route to MBE. Deposition performed by PLD is much faster, stoichiometric
transfer more easily controllable, wherein the quality of PLD thin films can be comparab le
to the layers synthesized by MBE. In addition, large area PLD systems became available
recently on the market which indicate that PLD will become an industrial process over the
next few years.
Initial studies were focused on cleaning, i.e. removal of organic compounds from the
substrate surface and elimination of native amorphous SiO 2 layer which covers the Si
surface since the epitaxial growth of STO can only be achieved on carbon free and
crystalline Si surface. Thermal and Sr induced deoxidation proc esses for elimination of
native SiO 2 from the Si surface were investigated. It was demonstrated that by selection of
proper sample holder and flashing the substrate to T>1100 °C a clean (2×1)+(1×2)
reconstructed Si(001) surface can be prepared. Thermal as well as Sr induced deoxidation
processes performed at lower temperatures did not result in clean Si(001) surface due to
the β SiC precipitates which form in the reaction between Si and carbon in the temperature
range from 730 °C to 1100 °C. Temperatures higher than 1100 °C cause dissociation of β
SiC and diffusion of carbon into the bulk silicon. Nevertheless, it was shown that flash
annealed Si surface becomes recontaminated with carbon after a certain period of time as
disclosed by the reflection high energy electron diffraction (RHEED) either as the c(4×4)
or β SiC transmission pattern. It was concluded that the recontaminati on of the Si( 001 )
surface cannot be completely avoided, however, by improving the vacuum (decreasing the
hydrocarbons partial pressure) it can be significantly delayed.
Epitaxial growth of STO directly on (2×1)+(1×2) reconstructed Si(001) surface is not
possible due to the high reactivity of silicon, which readily reacts with oxygen from the
background gas or deposited STO material and forms an amorphous transition layer at the
interface which hinders the epitaxial growth. An effective way of controlling the reactivity
of silicon surface against the oxidation is to form an appropriate buffer layer on top of the
Si(001) surface. Based on a number of MBE studies from the last two decades, Si surface
passivated with 1/2 ML of metallic strontium (Sr) represent s the most suitable buffer layer
in terms of achieved crystallinity of final STO layers. This dissertation presents an
experimental procedure for Sr buffer layer preparation developed with the PLD technique.
Different surface reconstructions formed due to the ordering of the Sr atoms on the Si(001)
surface at elevated temperatures were studied in detail. In situ RHEED characterization
showed surface structure evolution from the two
showed surface structure evolution from the two--domain (2×1)+(1×2) reconstruction, domain (2×1)+(1×2) reconstruction, exhibited by the bare Si surface, to a twexhibited by the bare Si surface, to a twoo--domain (2×3)+(3×2) structure at 1/6 ML, which domain (2×3)+(3×2) structure at 1/6 ML, which is then replaced by a twois then replaced by a two--domain (1×2)+(2×1) structure again at 1/4 ML Sr coverage. domain (1×2)+(2×1) structure again at 1/4 ML Sr coverage. The final (1×2)+(2×1) pattern was observed to be present up to 1/2 ML Sr coverage. The The final (1×2)+(2×1) pattern was observed to be present up to 1/2 ML Sr coverage. The observation of (2×3)+(3×2) surface stobservation of (2×3)+(3×2) surface structure, characteristic for 1/6ructure, characteristic for 1/6 ML Sr coverage, ML Sr coverage, enabled preparation of exactly 1/2enabled preparation of exactly 1/2 ML Sr coverage on Si, which is a prerequisite for the ML Sr coverage on Si, which is a prerequisite for the subsequent epitaxial growth of SrTiOsubsequent epitaxial growth of SrTiO33 layer. In addition, the stability of Sr buffer layer layer. In addition, the stability of Sr buffer layer against oxidation was testedagainst oxidation was tested and the results revealed much higher resistance of and the results revealed much higher resistance of Sr(Sr(1/21/2 ML)/Si(ML)/Si(001001) compared to the bare reconstructed Si surface.) compared to the bare reconstructed Si surface.
The last part of investigation was devoted to the growth of STO on (1×2)+(2×1) ast part of investigation was devoted to the growth of STO on (1×2)+(2×1) reconstructed Sr(1/2reconstructed Sr(1/2 ML)/Si(001) surface with the aim to prML)/Si(001) surface with the aim to prepare epare an an epitaxial layer with epitaxial layer with high crystallinity. Initial deposition proved to be the most critical part of the whole high crystallinity. Initial deposition proved to be the most critical part of the whole deposition process since the integrity of the Srdeposition process since the integrity of the Sr--buffer layer can easily be disrupted by high buffer layer can easily be disrupted by high energetic species from the laser plume. Baseenergetic species from the laser plume. Based on the detailed studies of initial deposition d on the detailed studies of initial deposition parameters in terms of background gas pressure, deposition temperature, fluence, and parameters in terms of background gas pressure, deposition temperature, fluence, and oxidation peculiarities it was demonstrated that the growth is fully epitaxial when the STO oxidation peculiarities it was demonstrated that the growth is fully epitaxial when the STO is deposited at room temperaturis deposited at room temperature and 0.05e and 0.05 mbar Ar pressure with intermediate oxidation mbar Ar pressure with intermediate oxidation periods and crystallized in vacuum (1×10periods and crystallized in vacuum (1×10--88 mbar) at the temperature in the range from mbar) at the temperature in the range from 517517 °C to 546°C to 546 °C. Epitaxial layer of any thickness can be grown using this procedure. °C. Epitaxial layer of any thickness can be grown using this procedure. However, the slowness and coHowever, the slowness and complexity of the process confines its use to mplexity of the process confines its use to the the first few first few nanometers of STO only. The subsequent deposition was performed innanometers of STO only. The subsequent deposition was performed in aa much faster regime much faster regime when both, highwhen both, high--temperature and hightemperature and high--oxygen partial pressure (1.3×10oxygen partial pressure (1.3×10--66 mbar), were mbar), were applied simultaneously. XRDapplied simultaneously. XRD examination of the final STO layer revealed only one outexamination of the final STO layer revealed only one out--ofof--plane (STO(001)||Si(001)) and one inplane (STO(001)||Si(001)) and one in--plane (STO[110]||Si[100]) orientation. On the other plane (STO[110]||Si[100]) orientation. On the other hand, few additional well defined spots with low intensity which emerged in the RHEED hand, few additional well defined spots with low intensity which emerged in the RHEED pattern during subsepattern during subsequent deposition at 517quent deposition at 517 °C indicate not completely single°C indicate not completely single--crystalline crystalline growth and require further optimization of subsequent deposition parameters. The STO growth and require further optimization of subsequent deposition parameters. The STO layer exhibits crystallinity of 1.68°, as derived from FWHM of (002) reflection along omega layer exhibits crystallinity of 1.68°, as derived from FWHM of (002) reflection along omega axis whicaxis which is similar to some of MBE grown films. Posth is similar to some of MBE grown films. Post--deposition annealing in oxygen deposition annealing in oxygen decreased the FWHM to 1.55°. Nevertheless, for some most demanding applications of decreased the FWHM to 1.55°. Nevertheless, for some most demanding applications of STO/Si(001) as a pseudoSTO/Si(001) as a pseudo--substrate, crystal quality of the STO layer needs to be further substrate, crystal quality of the STO layer needs to be further improvedimproved..
Overall, a a new PLD route for the epitaxial growth of STO on Si(001) was developed which new PLD route for the epitaxial growth of STO on Si(001) was developed which is based on the use of 1/2is based on the use of 1/2 ML SrML Sr--buffer layer.buffer layer.