Surface morphology of yttrium silicides epitaxially grown on Si(1 1 1) by STM

Surface morphology of yttrium silicides epitaxially grown on Si(1 1 1) by STM

Surface Science 482±485 (2001) 1337±1342 Surface morphology of yttrium silicides epitaxially grown on Si(1 1 1) by STM C...

265KB Sizes 0 Downloads 10 Views

Surface Science 482±485 (2001) 1337±1342

Surface morphology of yttrium silicides epitaxially grown on Si(1 1 1) by STM C. Polop *, C. Rogero, J.L. Saced on, J.A. Martõn-Gago Instituto Ciencia de Materiales de Madrid-CSIC, Cantoblanco 28049 Madrid, Spain

Abstract We have studied by means of scanning tunneling microscopy and low energy electron di€raction techniques the surface morphology of very thin layers of yttrium silicide epitaxially grown on a Si(1 1 1) substrate. For an Y coverage of 1 ML a two-dimensional p…1  1† silicide layer is formed. The surface morphology of this phase presents charac and oriented as the faulted half unit cell of the 7  7 reteristic triangular holes with a typical lateral size of 20 A p p construction. Further Y coverage leads to the formation of a three-dimensional (3D) … 3  3†R30° YSi1:7 ®lm. In particular, for a coverage of 3 ML, a 3D YSi1:7 ®lm grows, covering the whole surface. The topography of this phase is characterized by defects consisting of round shaped holes. Ó 2001 Elsevier Science B.V. All rights reserved. Keywords: Scanning tunneling microscopy; Epitaxy; Surface structure, morphology, roughness, and topography; Yttrium; Silicides; Surface defects

1. Introduction Rare-earth silicides (RE ˆ Y, Er, Gd, Dy, Lu, Ho) epitaxially grown on Si(1 1 1) substrates present very interesting properties from the technological point of view: a low value of the Schottky barrier height on n type substrate (0.3±0.4 eV), a sharp interface and extremely low lattice mismatch (0% for the YSi1:7 /Si(1 1 1)) [1±4]. The very low lattice mismatch makes possible the re-epitaxy of Si on the silicide surface, suggesting possible applications for device fabrication [5]. In particular, the investigation of the growth of thin silicide

* Corresponding author. Tel.: +34-91-334-9000; fax: +34-91372-0623. E-mail address: [email protected] (C. Polop).

layers and the Si re-epitaxy on its surface is interesting for the microelectronic industry [5]. Thus, the study of the surface morphology and defects at nanometric scale for the low coverage regime is important because they may limit the device performances. Additionally, RE silicides are also attractive due to the two-dimensional (2D) phase formed by a single silicide layer epitaxially grown on the Si(1 1 1) substrate [6±10]. The 2D properties of this phase can be also in¯uenced by the degree of continuity of the ®lms. The bulk atomic structure of the RE silicides epitaxially grown on Si(1 1 1) is well established. They form a defective AlB2 -type structure, consisting of stacked RE hexagonal planes and  with an graphite-like Si planes separated by 2 A, ordered arrangement of Si vacancies [2]. The vacancies are responsible for the RESi1:7 fractional

0039-6028/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 6 0 2 8 ( 0 1 ) 0 0 9 2 3 - 2


C. Polop et al. / Surface Science 482±485 (2001) 1337±1342

stoichiometry of the ®lms. The silicide surface is formed by a buckled Si layer without vacancies [11±14]. The ErSi1:7 surface has been widely studied by di€erent surface characterization techniques (photoemission spectroscopies, scanning tunneling microscopy (STM), surface di€raction techniques. . .) [11,13,15]. However, there are few experimental works dealing with the YSi1:7 surface [16,17]. In the present work, we report the main topographic features of very thin ®lms of yttrium silicide obtained by solid phase epitaxy. Particularly, we have characterized by STM the epitaxial growth of the silicide on Si(1 1 1) for one and three monolayers (ML) coverage, in order to investigate the evolution of the surface morphology from the ®rst formation stages. For 1 ML of Y, a 2D complete structure showing a p…1  1† periodicity is formed at the surface (1 ML corresponds to 7:8  1014 atoms/cm2 , i.e., the atoms of one Si(1 1 1) plane). Albeit this phase has already been found for other RE silicides (Gd, Dy, Er, Lu, Ho) [6±10] this is the ®rst time that this structure is reported for the Y±Si system. For coverages higher than 1 ML, a three-dimensional p p (3D) yttrium silicide ®lm exhibiting a … 3  3†R30° periodicity develops. In these very thin layers several types of defects are observed.

3. Experimental results and discussion 3.1. The bidimensional silicide YSi2 /Si(1 1 1) p(1  1) Evaporation of 1 ML of Y on the Si(1 1 1)7  7 surface and subsequent annealing to 350°C for 10 min, leads to the formation of a phase characterized by a p…1  1† LEED pattern without any trace of other additional reconstruction. We found a narrow coverage range …0:9 < h < 1 ML† and a narrow annealing temperature range …350°C < T < 450°C† for the formation of a pure p…1  1† phase. As mentioned above, this is the ®rst time that this structure is reported for the Y silicide. Fig. 1a shows a STM topographic image of 2 area scanned on this phase. Large 800  800 A and ¯at terraces characterize the topography of this 2D silicide. On the terraces an important number of holes and clusters can be observed. Fig. 1b is a detailed image recorded on a terrace and it shows triangular shaped holes.  and The terraces width is larger than 1000 A they are separated by monoatomic steps. We have

2. Experimental details The experiments were performed in a UHV chamber equipped with a commercial STM and a low energy electron di€raction (LEED) system, in a base pressure of 2  10 10 mbar. First, the samples were outgassed at 600°C for several hours. Clean and well ordered Si(1 1 1)7  7 surfaces were prepared by series of ¯ashes to 1200°C in a pressure lower than 2  10 9 mbar. Then, the sample was slowly cooled down from 900°C to room temperature (RT) at a rate of 15°C/min. Silicide ®lms were prepared by RT deposition and subsequent annealing. Y was evaporated by electron  bombardment, at rates of 0.5 A/min as controlled by a quartz-crystal thickness monitor, and the deposited ®lm was annealed to temperatures between 350°C and 550°C.

Fig. 1. (a) Topographic STM image of the 2D YSi2 /Si(1 1 1) 2 , V ˆ 1 V, I ˆ 1 p…1  1† surface. Scanned area ˆ 800  800 A nA; (b) STM image of 2D silicide surface showing triangular 2 , V ˆ 1 V, holes on the surface. Scanned area ˆ 200  90 A I ˆ 0:7 nA. The orientation of the F and U parts of the 7  7 unit cell previously to the growth has been drawn overimposed.

C. Polop et al. / Surface Science 482±485 (2001) 1337±1342

achieved atomic resolution images on the ¯at areas of the terraces. Fig. 2a shows an empty-states STM image where a regular hexagonal array of protrusions is observed. The distance between neighboring protrusions has been measured on  was several images. An average value of 3.8 A found. This value corresponds to the periodicity of  Hence, the ideal Si(1 1 1)-p…1  1† surface (3.84 A). the surface displays a p…1  1† structure in agreement with the characteristic p…1  1† LEED pat-

Fig. 2. (a) Atomic scale image of the p…1  1† phase recorded in 2 , V ˆ 0:56 constant current mode. Scanned area ˆ 20  20 A V, I ˆ 0:7 nA. The p…1  1† unit cell has been drawn overimposed. (b) Structural model for the RE2 /Si(1 1 1)-p…1  1†, as determined using surface di€raction techniques by Refs. [6,7,10]. White circles refer to the Y atoms; black circles represent the outermost Si atoms of the surface bilayer and grey circles refer to the rest of Si atoms of the structure.


tern. The p…1  1† structure has been also obtained for others RE silicides, and it has been described by an ordered hexagonal plane of RE atoms accommodated underneath a buckled Si surface bilayer without vacancies, as schematically drawn in Fig. 2b [6,7,10]. The white circles refer to the Y atoms and the black circles represent outermost Si atoms in the surface bilayer. Therefore, this 2D phase can be considered a single RE2 Si layer. Based on the assumption of the same structure for the p…1  1† phase of the Y silicide, we can assign the protrusions of the STM images to the outermost atoms in the buckled surface Si bilayer. The image presents a very small atomic corrugation  suggesting the absence of localized of 0.04 A, dangling bonds perpendicular to the surface. The atomic resolution images of the p…1  1† phase do not present changes with the voltage, being possible to obtain STM images at very low voltages. This supports a metallic character of the surface layer. All these ®ndings coincide with previously published results on the 2D ErSi2 surface [8,15]. The main topographic features present on the surface are holes and clusters (see Fig. 1a). Interestingly, averaging from several images we have found that the total extension occupied by the holes (10%) is close to the occupied by the clusters (11%). The clusters have a very irregular topography and their apparent height depends strongly on the applied voltage. They present the same electronic behavior of those previously reported for yttrium silicide clusters grown on the Si(1 1 1)7  7 surface at RT and imaged with STM [18]. The fact that we ®nd the same electronic behavior suggests that the clusters observed on the terraces are related to an excess of Y atoms that are not incorporated into the structure and then form silicide clusters. On the surfaces, we observe two di€erent types of holes. Some of them, clearly seen in Fig. 1a,  have irregular shape and they are around 3 A deep. However, a detailed look on the surface reveals another speci®c type of hole, marked with arrows in the Fig. 1a and shown in detail in Fig. 1b. These holes have a triangular shape. Their  although this numlateral size is about 20  5 A, ber depends on the preparation conditions (annealing temperature and time). The hole edges are


C. Polop et al. / Surface Science 482±485 (2001) 1337±1342

orientated along the ‰1 0 1Š lattice direction. Recording STM images on the clean Si(1 1 1)7  7 substrate at negative sample bias (occupied states), we can determinate the orientation of the faulted (F) and the unfaulted (U) half cells of the 7  7 reconstruction previously to the growth. From this, we have established that most of the triangular holes (95%) are oriented as the F half cell of the initial 7  7 reconstruction, which is schematically shown in Fig. 1b. A tentative explanation for the origin of the triangular holes described above could be related to the growth process of the silicide. For an Y coverage lower than 1 ML …0:25 < h < 1 ML†, we have observed the formation of triangular silicide islands [19]. The atomic structure of these islands depends of the Y coverage. However, independently of the coverage, most of the islands (90%) are oriented as the U half cell of the 7  7 reconstruction [19]. Therefore, the coalescence of these triangular islands during the silicide growth process could lead to the formation of triangular bare substrate areas oriented as the F half cell, as the holes observed in the Fig. 1. p p 3.2. The epitaxial YSi1:7 /Si(1 1 1) ( 3  3)R30° p p The … 3  3†R30° structure develops for Y coverage greater than 1 ML and annealing temperatures above 350°C. The terrace width of the silicide is determinate by the annealing temperature and time. Here we study the topography for an Y coverage of 3 ML and an annealing temperature of 450°C for min. The LEED shows p 15 p the characteristic … 3  3†R30° superstructure of the 3D silicide. A general view of the surface is shown in Fig. 3a. The detailed image of the surface in Fig. 3b shows the terrace edges orientated along the ‰1 0 1Š lattice direction. A typical pro®le of this image is shown in Fig. 3c. In this ®gure two different step heights can be observed: around 1.2  high. The 4.2 A  height corresponds to and 4.2 A the c lattice parameter of the YSi1:7 structure (4.14  and then can be understood as a silicide step. A),  height terrace can be reasoned as a The 1.2 A  and combination of a single substrate step (3.13 A)  Furthermore, at di€erent silicide heights (4.14 A). the boundary between two terraces separated by a  a partial dislocation is height distance of 1.2 A,

p p Fig. 3. STM images of the YSi1:7 /Si(1 1 1) … 3  3†R30° surface obtained for an Y coverage of 3 ML: (a) 1450  1200 2 , V ˆ 1:5 V, I ˆ 1 nA; (b) 460  410 A 2 , V ˆ 0:8 V, I ˆ 0:7 A nA; (c) Pro®le on (b) along the line.

generated due to the mis®t of the atomic planes in the c direction. As a consequence, surrounding this kind of steps, a depression can be observed which is marked with short arrows in Fig. 3b and c. These depressions could be understood as due to the compressive strain induced by the partial dislocation. Fig. 4 displays an atomic resolution image of empty state taken on the silicide terrace with a sample bias of ‡200 mV and a tunneling current of 0.7 nA. This image p shows p a hexagonal p…1  1† structure with a … 3  3†R30° modulation corresponding p p to the 3D YSi1:7 silicide phase. The … 3  3†R30° unit cell has been drawn overimposed. We have not found any trace of other ordered structures on the surface, indicating that

C. Polop et al. / Surface Science 482±485 (2001) 1337±1342

Fig. 4. STM image obtained on a YSi1:7 ®lm showing the p p 2 , … 3 3†R30° surface reconstruction. Scannedareaˆ3030 A p p V ˆ200 mV, I ˆ0:7 nA. The … 3 3†R30° unit cell has been drawn overimposed.

a complete YSi1:7 ®lm covers the whole surface. Based on similar images for the ErSi1:7 surface, two di€erent atomic models have been previously proposed [11,15]. These works assigned the protrusions observed in the images to the atomic positions. the former work propose that p However, p the … 3  3†R30° superstructure is patterned by atoms at di€erent heights, whereas for the latter, is caused by a lateral relaxation. Indeed, our atomic resolution images show features from both models. Furthermore, we have observed that the relative intensity of the protrusions changes with the bias voltage, suggesting an electronic e€ect rather than a pure structural e€ect. A full understanding of the atomic resolution STM images needs a longer contribution, being now under discussion [20]. A wider look around the surface at atomic scale shows a speci®c type of surface defect. The main morphologic features that are observed at the nanometre scale on the silicide surface are the depressed areas shown in the Fig. 5. This type of defects has been also noticed on the surface of the ErSi1:7 /Si(1 1 1) [13]. To check whether the depressed areas correspond to adsorbed particles or holes, the same image were recorded at di€erent voltages and polarities. The defects appear always as depressions in the STM images. This indicates that they correspond to holes in the surface rather than to adsorbed molecules [21]. These holes are randomly distributed on the whole surface and are formed independently of the annealing tempera-


Fig. 5. STM image showing holes on the surface of an YSi1:7 2 , V ˆ 0:5 V, I ˆ 1 nA. ®lm. Scanned area ˆ 85  85 A

ture in a range between 350°C and 550°C. Most of them seem to have a similar rounded shape and p their p size is around one unit cell of the … 3  3†R30° reconstruction. Their depth range  The maximum depth measured from 0.3 to 0.9 A. for the holes corresponds to the distance between  the two Si planes of the surface bilayer …0.8 A†. Then, these holes might be attributed to a few missing Si atoms of the topmost plane. Its formation could be related to a way of relaxing tensions in the silicide outmost layer rather than to kinetic limitations. The 3D silicide relaxes the stress into the bulk by the formation of an ordered lattice of Si vacancies [22]. The absence of vacancies at the surface could be compensated both by the buckling and by the formation of this type of holes. 4. Conclusions For an Y coverage of 1 ML a 2D YSi2 /Si(1 1 1)p…1  1† silicide is formed, presenting characteristic triangular deep defects that are mainly oriented in coincidence with the faulted half unit cell of the 7  7 reconstruction. Further Y coverage p leads to the formation of a 3D YSi /Si(1 1 1) … 3 1:7 p 3†R30° silicide. In particular, for coverage of 3 ML a YSi1:7 ®lm grows covering the whole surface forming terraces. The topography of this phase at atomic scale is characterized by the presence of defects consisting of round shaped holes.


C. Polop et al. / Surface Science 482±485 (2001) 1337±1342

Acknowledgements We are grateful to I. Jimenez for critical revision of the manuscript. This work has been ®nanced by the CAM-07N/0028/1999 and DGCYT-PB980524 Spanish projects.

References [1] R.D. Thompson, B.Y. Tsaur, K.N. Tu, Appl. Phys. Lett. 38 (1981) 535. [2] J.A. Knapp, T.S. Picraux, Appl. Phys. Lett. 48 (1986) 466. [3] J.Y. Duboz, P.A. Badoz, A. Perio, J.C. Oberlin, F. Arnaud dÕAvitaya, Y. Campidelli, J.A. Chroboczek, Appl. Surf. Sci. 38 (1989) 171. [4] M.P. Siegal, W.R. Graham, J.J. Santiago-Aviles, J. Appl. Phys. 68 (1990) 574. [5] J.Y. Veuillen, C. dÕAnterroches, T.A. Nguyen Tan, J. Appl. Phys. 75 (1994) 223. [6] P. Wetzel, C. Pirri, P. Paki, D. Bolmont, G. Gewinner, Phys. Rev. B 47 (1993) 3677. [7] M. Lohmeier, W.J. Huisman, G. ter Horst, P.M. Zagwijn, E. Vlieg, C.L. Nicklin, T.S. Turner, Phys. Rev. B 54 (1996) 2004.

[8] P. Wetzel, S. Saintenoy, C. Pirri, D. Bolmont, G. Gewinner, T.P. Roge, F. Palmino, C. Savall, J.C. Labrune, Surf. Sci. 335 (1996) 13. [9] S. Vandre, T. Kalka, C. Preinesberger, M. Dahne-Prietsch, Phys. Rev. Lett. 82 (1999) 1927. [10] D.J. Spence, S.P. Tear, T.C.Q. Noakes, P. Bailey, Phys. Rev. B 61 (2000) 5707. [11] T.P. Roge, F. Palmino, C. Savall, J.C. Labrune, P. Wetzel, C. Pirri, G. Gewinner, Phys. Rev. B 51 (1995) 10998. [12] L. Stau€er, A. Mharchi, S. Saintenoy, C. Pirri, P. Wetzel, D. Bolmont, G. Gewinner, Phys. Rev. B 52 (1995) 11932. [13] J.A. Martõn-Gago, J.M. G omez-Rodrõguez, J.Y. Veuillen, Surf. Sci. 366 (1996) 491. [14] L. Magaud, A. Pasturel, G. Kresse, J. Hafner, Phys. Rev. B 58 (1998) 10857. [15] J.A. Martõn-Gago, J.M. G omez-Rodrõguez, J.Y. Veuillen, Phys. Rev. B 55 (1997) 5136. [16] R. Baptist, A. Pellissier, G. Chaubet, Sol. Stat. Com. 68 (1988) 555. [17] R. Baptist, S. Ferrer, G. Grenet, H.C. Poon, Phys. Rev. Lett. 64 (1990) 311. [18] C. Polop, J.L. Saced on, J.A. Martõn-Gago, Surf. Sci. 454 (2000) 842. [19] C. Polop, J.L. Saced on, J.A. Martõn-Gago, in preparation. [20] C. Polop, et al., in preparation. [21] R. Wolkow, Ph. Avouris, Phys. Rev. Lett. 60 (1988) 1049. [22] F. Arnaud dÕAvitaya, A. Perio, J.C. Oberlin, Y. Campidelli, J.A. Chroboczek, Appl. Phys. Lett. 54 (1989) 2198.