Shape control of QDs studied by cross-sectional scanning tunneling microscopy
J.G. Keizer, M. Bozkurt, J. Bocquel, P.M. Koenraad, T. Mano, T. Noda, K. Sakoda, E.C. Clark, M. Bichler, G. Abstreiter, J.J. Finley, W. Lu, T. Rohel, H. Folliot, N. Bertru
SShape control of QDs studied by cross-sectional scanning tunneling microscopy
J.G. Keizer, M. Bozkurt, J. Bocquel, and P.M. Koenraad
Department of Applied Physics, Eindhoven University of Technology,P.O. Box 513, NL-5600 MB, Eindhoven, The Netherlands
T. Mano, T. Noda, and K. Sakoda
National Institute for Materials Science, 1-2-1 Sengen, Tsukuba, Ibaraki 305-0047, Japan
E.C. Clark, M. Bichler, G. Abstreiter, and J.J. Finley
Walter Schottky Institut, Technische Universit¨at M¨unchen,Am Coulombwall 3, D-85748 Garching, Germany
W. Lu, T. Rohel, H. Folliot, and N. Bertru
INSA, Universit´e Europ´eenne de Bretagne, 20 Avenue des Buttes de Co¨esmes, F-35043 Rennes Cedex, France
In this cross-sectional scanning tunneling microscopy study we investigated various techniquesto control the shape of self-assembled quantum dots (QDs) and wetting layers (WLs). The resultshows that application of an indium flush during the growth of strained InGaAs/GaAs QD layersresults in flattened QDs and a reduced WL. The height of the QDs and WLs could be controlledby varying the thickness of the first capping layer. Concerning the technique of antimony cappingwe show that the surfactant properties of Sb result in the preservation of the shape of strainedInAs/InP QDs during overgrowth. This could be achieved by both a growth interrupt under Sb fluxand capping with a thin GaAsSb layer prior to overgrowth of the uncapped QDs. The techniqueof droplet epitaxy was investigated by a structural analysis of strain free GaAs/AlGaAs QDs. Weshow that the QDs have a Gaussian shape, that the WL is less than 1 bilayer thick, and that minorintermixing of Al with the QDs takes place.
INTRODUCTION
In the last decade the fabrication of self-assembledquantum dots (QDs) has been intensively studied. Theinterest has been, and still is, stimulated by applicationsof QDs in optoelectronic devices. From previous studiesit is well known that the optical and electronic proper-ties of QDs are strongly affected by their size, shape, andmaterial composition. Despite years of intense studies,control over these properties remains difficult. One ma-jor problem is the change in QD morphology during thegrowth of the capping layer. Traditionally control overthe QD height, one aspect of the change in morphology,can be achieved with monolayer precision by the doublecapping method [1] or the so-called indium flush method[2], a variation on the former technique. In the lattertechnique the growth of the capping layer is interrupted,at which point the temperature is raised to remove anysurface resident indium. This effectively locks the heightof the QD and prevents any further In segregation [3].Another approach in shape control of QDs is the useof surfactants. Recently, antimony has received a greatdeal of attention in its role during the capping processdue to its surfactant properties. It has been shown thatSb reduces the surface diffusion of other atoms but with-out getting incorporated itself [4], allowing the achieve-ment of fully pyramidal shaped QDs [5]. Yet anotherapproach to gain control over the erosion of QDs duringovergrowth and thus over the shape of the QD is the re- moval of the driving force: lattice strain. This can beachieved in lattice matched QDs grown by droplet epi-taxy. First reported by Koguchi et al. [6], this techniqueinvolves low temperature growth of unstrained group III-element droplets that are subsequently crystallized intoQDs by incorporation of group V-elements. It has beenshown that this technique can be used to grow nearlypure nanostructures [7] with a typical size distribution of10–20% [8].In this paper the techniques of indium flush, antimonycapping, and droplet epitaxy are studied by means ofcross-sectional scanning tunneling microscopy (X-STM).We first investigate the degree of control that can beachieved over the height of InGaAs/GaAs QDs and thewetting layer (WL) by means of an indium flush. Wethen go on by showing that antimony capping can beemployed to prevent QD erosion during the capping pro-cess of InAs/InP QDs. Finally, the intermixing in, andthe shape of, GaAs/AlGaAs QDs grown by droplet epi-taxy is examined in detail.
EXPERIMENTAL SETUP
All X-STM measurements were performed at roomtemperature under UHV ( p < × − mbar) condi-tions with an Omicron STM-1, TS2 Scanner. The STMwas operated in constant current mode on in situ cleaved(110)-surfaces. Electrochemically etched tungsten tips a r X i v : . [ c ond - m a t . m e s - h a ll ] N ov were used. The QD layers were grown by molecularbeam epitaxy (MBE). The details of the growth proce-dure for the different material systems will be describedseparately in their corresponding sections. INDIUM FLUSH
The material system used to investigate the indiumflush technique consists of InGaAs QD layers grown byMBE on an n -type GaAs (001) orientated substrate.An undoped GaAs buffer layer of 420 nm was grown at690 ◦ C, followed by a growth interruption of approxi-mately 2 min that allowed the temperature to be low-ered to 600 ◦ C, the nominal growth temperature of theQD layers. Following this, three sequences consisting offour QD layers of 1.98 nm (7 ML) In . Ga . As were de-posited. During the whole growth process the As flux waskept constant at a pressure of 1 . × − mbar. Threeout of the four QD layers were grown with the indiumflush method which consists of the following procedure.First the QD layers are partially capped with a GaAslayer of which the thickness was varied. Next, the tem-perature is raised to 650 ◦ C for 30 s and lowered againto the nominal growth temperature after which a secondGaAs capping layer is deposited. In total, the annealingstep takes place over a time window of ≈
180 s. The totalstructure was capped with 200 nm GaAs.We begin by analysing the WL thickness and compo-sition. In figure 1, four typical X-STM images of WLsgrown with different capping layer thicknesses are de-picted. Even without any statistical analysis it is evidentthat the height of the WL can be controlled by varying
FIG. 1. X-STM images of the InGaAs WL as a function ofthe capping layer thickness. (a) 2 nm (b) 3 nm (c) 6 nm firstcapping layer thickness and (d) conventionally grown cappinglayer. FIG. 2. In segregation as a function of first capping layerthickness and bilayer position from the start of the WL. the height of the capping layer. In addition, the In seg-regation appears to terminate abruptly in case of WLsthat underwent an indium flush. This is a clear indi-cation that most of the surface resident In is removedduring the flush step, preventing further segregation. Inorder to make our analysis more quantitative we countedand marked the bilayer position from the start of theWL for approximately 3000 In atoms. An ≈
400 nmcross-sectional region of each WL present in the sam-ple was analysed in this manner. In figure 2, the resultof our statistical analysis is shown. The conventionallygrown WL exhibits the expected exponential decay ofthe In concentration and In segregation length ( ≈
25 nm)[9]. In contrast, the WLs grown with the indium flushprocedure show a stronger decay and shorter segregationlength. This implies that In segregates out of the WLsand leaves the surface during the indium flush step. Thisadditional loss of already buried In is strongest in case ofthe thinnest capping layer. The total amount of In thatremains after flush-off is thus strongly dependent on thecapping layer thickness due to desorption and additionalsegregation. The thickness of the final WL is found to be6, 8, 10, 12 bilayers for 2, 3, 4, 6 nm thick first cappinglayers, respectively. Note, that the statistical analysispresented in figure 2 reveals that the WL extends furtherthan is expected from figure 1. It is reported that thecritical WL thickness for In . Ga . As QD formation is ≈ d crit ) to the thickness of the first capping layer( d cap ) and compare the resulting sum with the experi-mentally found thickness of the final WL we find goodagreement. This is depicted quantitatively by the dot-ted red line and open red boxes in figure 4, that show d crit + d cap and the experimentally determined average FIG. 3. X-STM image of one conventionally grown QD andthree QDs grown with an indium flush step incorporated inthe growth process. The thickness of the first capping layerwas varied.FIG. 4. QD height (black points) as a function of the thick-ness of the first capping layer. The black line is a linear fit.The dotted red line represents the sum of the critical layerthickness (5 ML) and the first capping layer thickness (dashedblue line). The experimentally determined average thicknessof the final WL is given by the open red boxes.
WL thickness, respectively. This result shows that Insegregation beyond the position of the flush is completelysuppressed; In is absent in the final GaAs capping layer.In order to determine the influence of the indium flushstep on the structural properties of the QDs we deter-mined the width and height of a total of 48 cleaved QDs.The width of the QDs ranged up to 100 nm. The height ofthe conventionally grown QDs was found to vary between7 and 10 nm. The QD layers were found to be weaklycoupled, as one would expect with the GaAs spacer layerbeing 30 nm thick and slightly strained InGaAs QDs [11],resulting in occasional stacking of the QDs. Figure 3 shows one of the sequences consisting of four QD lay-ers where the QDs are stacked. The thickness of the firstcapping layer was varied in the first three QD layers from2, 3 to 6 nm. The last layer is a conventionally grown QDlayer, i.e. without the application of an indium flush step.As can be seen, the application of an indium flush stepresults in lowering of the QD height as compared to theconventionally grown QDs. The shape of the convention-ally grown QD is lens like as expected for typical InGaAsQDs [12]. The heights of all the observed QDs as a func-tion of the first capping layer thickness are plotted infigure 4. Since, the lateral width of all the observed QDswas found to be of the order of 60 nm, we can assume thatnone of the QDs is cleaved through their edge and thatfigure 4 represents the spread in the height distributionof the QDs due to the growth process. We found a lin-ear relation between the QD height and the first cappinglayer thickness up to ≈ ANTIMONY CAPPING
In the previous section we have shown that the indiumflush technique can be used to lower the height of InGaAsQDs. We continue with an investigation of antimony cap-ping, a technique that can be employed to prevent QDerosion during capping. Four InAs QD layers separatedby 30 nm of InP were grown on an n -type (311)B ori-ented InP substrate by solid source MBE. The growthtemperature was set at 450 ◦ C. The QDs were formedby the deposition of 2.1 ML (001) equivalent monolayers.After QD formation, a 30 s growth interrupt (GI) un-der As pressure was performed for all layers. Previously,it has been shown that As/P exchange is limited undersuch GI conditions [13]. The first QD layer was over-grown with an InP capping layer. This first QD layerwill be considered as the reference layer. For the secondQD layer a growth interrupt under a Sb beam equivalentpressure of 2.7 × − Torr (GISb) was performed during30 s before the growth of the InP capping layer. For thethird and fourth layers, respectively a 1 nm and 2 nmGaAs . Sb . (lattice matched to InP) thick layer wasdeposited after a 5 s GISb.In figure 5a, an X-STM image of a typical QD in thereference layer is shown. These QDs are found to have FIG. 5. Two 60 nm ×
15 nm X-STM images. (a) InAs QDcapped with InP after a 30 s GI. (b) InAs QD capped withInP after a 30 s GI + 30 s GISb. The bright spots correspondto Sb atoms. a flat top facet. The homogeneity of the contrast withinthe QD indicates that it consist of almost pure InAs. Nodigging in of the WL in the underlying material as in [13]was observed. The intermixing at the corners is minimal,like in the case of InAs QDs in AlAs [14]. The aver-age height and width estimated from 20 individually ob-served QDs in the reference layer are found to be 2.0 nmand 25 nm, respectively. Before capping, the QDs havean asymmetric pyramidal shape, bounded by low-indexfacets { } , { } B, and { } [15]. Height histogramsof the uncapped QDs deduced from AFM analysis, andheight histograms of the capped QDs in the referencelayer as observed by X-STM are shown in figure 6a-b.For the uncapped QDs, a Gaussian distribution centeredaround 3.3 nm is found, whereas after InP capping theheight distribution is truncated at 2.4 nm. As demon-strated previously [13], the truncated distribution andthe flat top facet of InAs/InP QDs are to a large extendthe consequence of QD decomposition. This decomposi-tion is driven by the strain mismatch between the InPcapping layer and the InAs QDs.Figure 5b, shows InAs QDs for which a 30 s GISb hasbeen performed before the InP capping layer was grown.The bright spots correspond to Sb atoms remaining in theInP capping layer and in the InAs QDs after the GISband the succeeding growth of the capping layer. Giventhe total amount of Sb supplied to the surface and theobserved amount of Sb after capping, we conclude thata large part is desorbed during overgrowth. Segregationof the small fraction of Sb that gets incorporated in theInP capping layer is clearly shown. Within the QDs theback diffusion of Sb is negligible and a preferential incor-poration of Sb is observed at the outermost layers of theQDs. Again, the InAs QD corners appear well defined FIG. 6. QD height distribution of a) uncapped InAs QDs, (b)InP capped InAs QDs, (c) InP capped InAs QDs after 30 sGISb, and d) GaAsSb capped InAs QDs after 5 s GISb.FIG. 7. Two 100 nm ×
20 nm X-STM images of InAs QDs andthe WL capped with (a) 1 nm and (b) 2 nm GaAsSb after a30 s GI + 5 s GISb. A single step edge is visible at the leftside in both images. with minimal intermixing and formation of an InAsP al-loy, just as is the case with the QDs in the referencelayer. The presence of Sb on the surface induces changeson the QD shape; the mean height is now 3.5 nm (seefigure 6c) and the mean diameter 21 nm, correspondingto the dimensions of the uncapped QDs. We can explainthe observed shape preservation by the well documentedsurfactant effect of Sb atoms [4, 16]. An Sb surfactantcan limit the in-plane diffusion of atoms on the surface.Accordingly, the InAs diffusion from the QD apex to theperiphery should be reduced due to the presence of Sbatoms on the surface. This freezing of the mass trans-port on the growth front results in the preservation ofthe shape of the uncapped QDs.X-STM images of the third and fourth QD layers areshown in figure 7a-b. These layers were, after 30 s GI+ 5 s GISb, capped with a thin layer of GaAsSb (latticematched to InP). As was the case with the 30 s GISb,the QDs in these layers are taller than those in the ref-erence layer; for both layers, an average height of 3.2nm (see figure 6)d and a base diameter of 21 nm are de-duced, corresponding to the dimensions of the uncappedQDs. Again, the intermixing in the QDs is negligible.Similar shape conservation has been reported when In-GaAs or GaAsSb strained capping layers are grown onInAs/GaAs QDs [13]. In that case a phase separationis observed in the ternary capping layer on top of theQDs. In our case, the GaAsSb layer is lattice matched tothe InP substrate and the observed conformal growth ofGaAsSb on InAs/InP QD might be related as previouslyto a low group III atoms migration when Sb atoms arepresent on surface.
DROPLET EPITAXY
Having shown that the indium flush technique and thesurfactant properties of Sb allow control over the shapeand height of SK-grown QDs, we now turn our atten-tion to QDs in a lattice matched materials system. Morespecifically, a GaAs/AlGaAs QD layer grown on an n -type (001) oriented GaAs substrate by droplet epitaxy.The sample was grown in the following manner. First anAlGaAs buffer layer is grown at 580 ◦ C. Next, the sam-ple is cooled down to 200 ◦ C, the As flux switched off,and the As evacuated from the growth chamber. Theresult is an As-stabilized c (4 ×
4) surface. Subsequently,3.75 ML Ga, of which the first 1.75 ML changes the excessAs into a two-dimensional GaAs layer [17], is depositedat a rate of 0.5 ML/s. The remainder of 2 ML will formliquid Ga droplets on the surface. Next, these dropletsare crystallized into a GaAs QDs by supply of an As flux (2 × − Torr beam equivalent pressure). Still un-der As flux, the sample is then annealed at 350 ◦ C for10 minutes. Subsequently the structures are capped with50 nm AlGaAs deposited at 350 ◦ C, followed by a secondannealing step at 650, ◦ C under As flux for 5 minutes.This last anneal step is inserted into the growth proce-dure to ensure that the next layer is grown on a defect freesurface. Next, another capping layer of 40 nm is grownat 580 ◦ C. The total structure was capped with 600 nmGaAs. A post growth anneal step, which is usually per-formed to improve the optical properties of the QDs wasnot performed on this sample. All the images presentedin this section are recorded with high negative bias ( ≈ -3.2 V) between sample and STM tip. At these tunnelingconditions and with the color scaling used, dark regionsrepresent AlAs rich regions while bright regions representGaAs rich regions.A total of 11 QDs where observed by X-STM. A typ- FIG. 8. 40 nm ×
34 nm topographic image of a typicalGaAs/AlGaAs QD (top) and an average cross-sectional profile(top graph) and separation between bilayers (bottom graph)along the line in the top figure. ical QD is shown in figure 8. As can be seen in this to-pographic image, the QDs are sharply defined by abruptinterfaces. The thickness of the WL was found to be lessthan 1 bilayer [7], as expected. The bow tie feature ismost likely a foreign atom and is of no interest in thecurrent study. Since AlAs and GaAs are lattice matchedmaterials, the QDs are expected to be strain free. Thisis checked by taking a cross-sectional profile of the QD infigure 8. Three distinct regions can be observed. Fromleft to right: an AlAs rich region, the GaAs QD, andthe AlGaAs matrix. The height difference between theseregions is due to electronic contrast. More importantly,all the regions are flat, there is no outward relaxation asobserved in QDs grown with lattice-mismatched systems[9]. To further illustrate that the GaAs QD is strain free,the distance between adjacent bilayers along the cross-sectional profile was measured. For this analysis the STMpiezo elements were calibrated by performing a 2D FFTon the AlGaAs matrix. The result is shown in the bot-tom graph of figure 8. As can be seen, there is littledeviation from the expected value of 0.565 nm (dashedline), indicating that the QD is indeed strain free. Notethat there is an Al rich region on top of the QD. This can
FIG. 9. 30 nm ×
60 nm topographic image (left) of two QDs.An atomic grid is overlain on top of a close up of the QDdot (right). Al and Ga atoms in the QD are indicate byrespectively red and yellow squares. be explained by the difference in mobility of Al and Gaatoms; the Ga atoms are more mobile and will migratealong the side of the QD during capping while the Alatoms, which are less mobile, are more likely to remainon top of the QD. The driving force behind the migra-tion of the incoming adatoms away from the top of theQD is the convex curvature of the growth front at theposition of the QDs [18]. Note that this different fromthe SK-grown QDs of the previous sections were straininduced by the lattice mismatch is the driving process.Whether intermixing of Al is a factor of importance inthe formation of GaAs/AlGaAs QDs grown by dropletepitaxy is a question frequently raised in the literature[19–21]. In all QDs imaged we have observed some de-gree of intermixing. In figure 9 left panel, two typicalQDs are shown. Even without further analysis it is ev-ident that some intermixing of Al has taken place, seedark spots inside the QDs. To make a more quantitativeanalysis we overlaid a grid with atomic dimensions ontop of a close up of the QD that showed the strongestintermixing, see figure 9 right panels. On this grid, thepositions of the Al and Ga atoms are marked with re-spectively red and yellow squares. The concentration ofAl in this particular QD is determined to be 6%. Herewe would like to point out that the observed Al intermix-ing varied strongly from dot to dot, see for example theQD depicted in figure 8 were the degree of intermixingis considerably lower, and that the 6% can be considered
FIG. 10. Profile of three QDs extracted from the X-STM data(open circles). A Gaussian function is fitted to the largestQD (red line). The other two QDs (green and blue line) areassumed to have the same 3D-structure as the largest QD butcleaved off center. The projection of the (111)-direction onthe cleavage plane is given by the dashed black line. as an upper limit of intermixing in these QDs.Concerning the shape of the QDs, we notice that theside facets of the observed QDs are not exactly straight.The maximum side facet angles were found to be in therange 34–55 ◦ per QD, were the upper limit correspondsto a { } facet (54.7 ◦ ). If we assume that (1) all the QDsare approximately of equal height and (2) the observedheight difference is due to the position of the cleavageplane relative to the center of the QD, this result ex-cludes QD shapes with constant facet angles like rectan-gular (truncated) pyramids [22]. Since it has been re-ported that uncapped AlGaAs/GaAs QDs have { } facets [23, 24], we conclude that the shape of the QDs issomewhat changed during capping. Figure 8 shows thehighest QD we found. Since it is the highest, we assumethat this QD is cleaved directly through its center. Con-sequently, we used the profile of this QD to generate a3D-profile by fitting a Gaussian function, see figure 10(red line), and rotating it around the symmetry axisalong the growth direction. Next, we checked whetherthe profile of all other observed QDs (illustrated for twoexemplary QDs by the green and blue lines) correspondto profiles obtained by cleaving the obtained 3D-profileat specific distances from the center. As can be seen infigure 10, this is the case. From this we conclude thatthe observed QDs are Gaussian shaped QDs of approxi-mately the same height but cleaved at different positionfrom their center. CONCLUSION
To summarize, we have investigated three techniquesthat can be used to gain control over the shape of QDs.The indium flush technique allows control over the heightof the WL and the InGaAs QDs. The resulting QDs havea flatted top facet. We have shown that not only surfaceresident In but also buried In that segregates out of theWL is desorbed during the indium flush. Concerningthe technique of antimony capping, we have shown thata growth interrupt under Sb flux prior to capping pre-serves the shape of the uncapped QDs. The same couldbe achieved by the growth of a GaAsSb capping layer.This capping layer was found to conformally cover thegrowth front. In both case the preservation of QD shapeis attributed to the surfactant properties of Sb. In QDlayers grown by droplet epitaxy the WL was found to beless than 1 bilayer thick. As expected in lattice-matchedsystems, we found no strain present in the GaAs/AlGaAsQDs. Without strain there is no driving force for QD ero-sion during overgrowth, resulting in high QD of which theshape was found to be Gaussian. We conclude that in-dium flush, antimony capping, and droplet epitaxy canall be used as a tool to both shape QDs and WL. [1] C. Paranthoen, N. Bertru, O. Dehaese, A. Le Corre,S. Loualiche, B. Lambert, and G. Patriarche, AppliedPhysics Letters , 1751 (2001), ISSN 00036951, http://apl.aip.org/resource/1/applab/v78/i12/p1751_s1 .[2] Z. Wasilewski, Journal of Crystal Growth , 1131(May 1999), ISSN 00220248, http://dx.doi.org/10.1016/S0022-0248(98)01539-5 .[3] J. G. Keizer, E. C. Clark, M. Bichler, G. Abstreiter, J. J.Finley, and P. M. Koenraad, Nanotechnology , 215705(May 2010), ISSN 1361-6528, http://stacks.iop.org/0957-4484/21/i=21/a=215705 .[4] J. C. Harmand, L. H. Li, G. Patriarche, and L. Travers,Applied Physics Letters , 3981 (2004), ISSN 00036951, http://apl.aip.org/resource/1/applab/v84/i20/p3981_s1 .[5] J. M. Ulloa, I. W. D. Drouzas, P. M. Koen-raad, D. J. Mowbray, M. J. Steer, H. Y. Liu, andM. Hopkinson, Applied Physics Letters , 213105(2007), ISSN 00036951, http://apl.aip.org/resource/1/applab/v90/i21/p213105_s1 .[6] N. Koguchi, S. Takahashi, and T. Chikyow, Journal ofCrystal Growth , 688 (May 1991), ISSN 00220248, http://dx.doi.org/10.1016/0022-0248(91)91064-H .[7] J. G. Keizer, J. Bocquel, P. M. Koenraad, T. Mano,T. Noda, and K. Sakoda, Applied Physics Letters ,062101 (2010), ISSN 00036951, http://apl.aip.org/resource/1/applab/v96/i6/p062101_s1 .[8] T. Mano and N. Koguchi, Journal of Crystal Growth ,108 (May 2005), ISSN 00220248, http://dx.doi.org/10.1016/j.jcrysgro.2004.12.119 . [9] P. Offermans, P. M. Koenraad, R. Notzel, J. H. Wolter,and K. Pierz, Applied Physics Letters , 111903(2005), ISSN 00036951, http://apl.aip.org/resource/1/applab/v87/i11/p111903_s1 .[10] C. Snyder, J. Mansfield, and B. Orr, Physical Review B , 9551 (Oct. 1992), ISSN 0163-1829, http://prb.aps.org/abstract/PRB/v46/i15/p9551_1 .[11] D. M. Bruls, P. M. Koenraad, H. W. M. Salemink,J. H. Wolter, M. Hopkinson, and M. S. Skolnick, AppliedPhysics Letters , 3758 (2003), ISSN 00036951, http://apl.aip.org/resource/1/applab/v82/i21/p3758_s1 .[12] X.-D. Wang, N. Liu, C. K. Shih, S. Govindaraju,and A. L. Holmes, Applied Physics Letters , 1356(2004), ISSN 00036951, http://apl.aip.org/resource/1/applab/v85/i8/p1356_s1 .[13] J. M. Ulloa, C. Celebi, P. M. Koenraad, A. Simon,E. Gapihan, A. Letoublon, N. Bertru, I. Drouzas, D. J.Mowbray, M. J. Steer, and M. Hopkinson, Journal ofApplied Physics , 081707 (2007), ISSN 00218979, http://jap.aip.org/resource/1/japiau/v101/i8/p081707_s1 .[14] P. Offermans, P. Koenraad, J. Wolter, K. Pierz,M. Roy, and P. Maksym, Physical Review B (Oct.2005), ISSN 1098-0121, doi:“bibinfo doi 10.1103/Phys-RevB.72.165332, http://prb.aps.org/abstract/PRB/v72/i16/e165332 .[15] D. Lacombe, A. Ponchet, S. Frechengues, V. Drouot,N. Bertru, B. Lambert, and A. Le Corre, Applied PhysicsLetters , 1680 (1999), ISSN 00036951, http://apl.aip.org/resource/1/applab/v74/i12/p1680_s1 .[16] P. Aivaliotis, L. R. Wilson, E. A. Zibik, J. W. Cock-burn, M. J. Steer, and H. Y. Liu, Applied Physics Let-ters , 013503 (2007), ISSN 00036951, http://apl.aip.org/resource/1/applab/v91/i1/p013503_s1 .[17] S. Sanguinetti, Journal of Crystal Growth , 71(Jun. 2003), ISSN 00220248, http://dx.doi.org/10.1016/S0022-0248(03)01016-9 .[18] Q. Xie, P. Chen, and A. Madhukar, Applied Physics Let-ters , 2051 (1994), ISSN 00036951, http://apl.aip.org/resource/1/applab/v65/i16/p2051_s1 .[19] S. Sanguinetti, Journal of Crystal Growth , 321(Jul. 2002), ISSN 00220248, http://dx.doi.org/10.1016/S0022-0248(02)01434-3 .[20] V. Mantovani, S. Sanguinetti, M. Guzzi, E. Grilli, M. Gu-rioli, K. Watanabe, and N. Koguchi, Journal of Ap-plied Physics , 4416 (2004), ISSN 00218979, http://jap.aip.org/resource/1/japiau/v96/i8/p4416_s1 .[21] C. Heyn, A. Stemmann, A. Schramm, H. Welsch,W. Hansen, and A. Nemcsics, Physical Re-view B (Aug. 2007), ISSN 1098-0121, doi:“bibinfo doi 10.1103/PhysRevB.76.075317, http://prb.aps.org/abstract/PRB/v76/i7/e075317 .[22] D. M. Bruls, J. W. A. M. Vugs, P. M. Koenraad,H. W. M. Salemink, J. H. Wolter, M. Hopkinson, M. S.Skolnick, F. Long, and S. P. A. Gill, Applied Physics Let-ters , 1708 (2002), ISSN 00036951, http://apl.aip.org/resource/1/applab/v81/i9/p1708_s1 .[23] C.-D. Lee, C. Park, H. J. Lee, S. K. Noh, K.-S.Lee, and S.-J. Park, Applied Physics Letters , 2615(1998), ISSN 00036951, http://apl.aip.org/resource/1/applab/v73/i18/p2615_s1 .[24] T. Mano, K. Mitsuishi, Y. Nakayama, T. Noda, andK. Sakoda, Applied Surface Science , 7770 (Sep.2008), ISSN 01694332, http://dx.doi.org/10.1016/j.apsusc.2008.02.025http://dx.doi.org/10.1016/j.apsusc.2008.02.025