- Email: [email protected]
Dimer structure of the Si(001)2×1 surface observed by low-temperature scanning tunneling microscope
Physica B 329–333 (2003) 1644–1646
Dimer structure of the Sið0 0 1Þ2 1 surface observed by low-temperature scanning tunneling microscope Masanori Ono*, A. Kamoshida, N. Matsuura, T. Eguchi, Y. Hasegawa Institute for Solid State Physics, The University of Tokyo 5-1-5, Kashiwa-no-ha, Kashiwa, Chiba 277-8581, Japan
Abstract Using a scanning tunneling microscope (STM) which can be operated in ultrahigh vacuum (UHV), low temperature (> 2:8 K), and magnetic field (o11 T), we studied dimer structure on the Sið0 0 1Þ2 1 surface. Asymmetric (buckled) dimer structure was observed with positive sample voltages, while most of the dimers look symmetric with negative voltages. These observations suggest that the observed symmetric dimer in the filled state images is not intrinsic; neither static symmetric dimer nor flip–flopped asymmetric dimer, but a matter of STM imaging mechanism. Magnetic filed, up to 10 T; applied perpendicularly to the surface does not affect the STM imaging at 10 K: r 2003 Elsevier Science B.V. All rights reserved. PACS: 68.35.Bs; 61.16.Ch; 68.35.Rh Keywords: Silicon surface; Silicon dimer structure; Sið0 0 1Þ2 1; Scanning tunneling microscopy
On Si(0 0 1) surface, two silicon atoms making a bond in the /1 1 0S direction to reduce the number of dangling bonds, a dimer structure is formed [1]. The dimers are arranged in the /1 1% 0S direction to form a ‘‘dimer row’’ structure with a long range ordering of 2 1 periodicity. The unique feature that makes the surface interesting is buckling of the dimer, that is, one of the two dimer-composing atoms protruding while the other lowered. In the dimer-row direction, the asymmetric ‘‘buckled’’ dimer exhibits an ‘‘anti-ferromagnetic’’-like interaction, making a zigzag pattern. Between the neighboring dimer-rows, interaction can be either ‘‘ferromagnetic’’ or ‘‘anti-ferromagnetic’’, producing a pð2 2Þ or cð4 2Þ ordering, respectively. A flip–flop motion of the buckled dimer can be induced by thermal activation or by a proximity effect of the probe tip in STM, but the motion can be pinched off near local defects, such as steps and missing dimers. Lots of theoretical and experimental studies using various techniques, including STM [2–7], have been
*Corresponding author. E-mail address: [email protected] (M. Ono).
performed. Generally accepted scenario says that an order–disorder transition occurs at around 200 K; below the temperature, the buckled dimers are arranged with an ordering of cð4 2Þ; and above the temperature the thermally activated flip–flop motion exhibits the 2 1 ordering [3,4,8]. Recently, Kondo et al. [5] reported STM images taken below the temperature showing nonbuckled dimers, and claimed a new ground state structure. Yokoyama and Takayanagi, then, showed an evidence of cð4 2Þ electronic structure at 5 K by STM, attributing the non-buckled dimer imaging to anomalous flip–flop motion [6]. Before their observations, Shigekawa et al. [7] demonstrated buckled-dimer STM images below the temperature. But there is an argument on whether the reported temperature is correct or not [5]. In order to solve the undetermined situation, we carried out STM experiments at correctly measured temperature (10 K). Our results show non-buckled images with positive sample voltages (filled state images). But, with negative sample voltages (empty state images), buckled dimer formation was clearly observed. These observations suggest asymmetric dimer structure at the low temperature and that apparent
0921-4526/03/$ - see front matter r 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0921-4526(02)02439-0
M. Ono et al. / Physica B 329–333 (2003) 1644–1646
1645
Fig. 2. STM image taken at 47 K with a negative sample bias voltage. The size of the observed area is 7:0 nm 3:5 nm: The sample bias voltage and tunneling current is 2:03 V and 68 pA:
Fig. 1. STM images taken on the same area at 10 K with a negative (left) and positive (right) sample bias voltage. The observed area is 14 nm 7 nm: The sample bias voltage and tunneling current is 1:47 V and 42 pA (left), 1:51 V and 42 pA (right), respectively.
symmetric dimer structure in the filled state images is caused by STM imaging mechanism. Silicon samples (n-type, Sb-doped, 0.014–0:005 O cm) were cleaned by 12001C annealing. The sample temperature was measured with a resistance sensor (Cernox), which was calibrated with the one attached on a dummy sample surface. Fig. 1 shows STM images taken on same area with negative (left) and positive (right) sample bias voltages. Filled and empty states of the sample are imaged, respectively. The filled state image shows many linear features, ‘‘dimer-row’’, running parallel each other. Most area apparently shows non-buckled symmetric dimers. We took images with a bias voltage ranging from 3 to 1:3 V; and non-buckled dimer structure always covers most of the surface. The empty state image of the same area (Fig. 1, right), shows remarkable differences from its counterpart. Almost all area shows a zigzag pattern, that is, ‘‘anti-ferromagnetic’’ dimer arrangement along the dimer-row direction, supporting a presence of buckled dimers on the surface. Applied bias voltage ranging from 3 to 1:6 V does not affect the observation. Theoretical calculation [9] suggests the upper atom in the buckled dimer looks brighter (darker) in the filled (empty) state images. Correlation of the zigzag pattern with a neighboring dimer row has two types: in and out of phases, corresponding to pð2 2Þ and cð4 2Þ structures, respectively. Occasionally, we observed a motion of the phase boundary and transitional fuzzy-looking regions in sequentially taken STM images, as was observed by Shigekawa et al. [7]. In spite of the motion, overall phase domain structure was maintained in empty-state images. These observation exclude a possi-
bility that apparently non-buckled dimer is caused by the flip–flop motion either thermally or tip-induced. We speculated that the apparent symmetric dimer structure is due to STM imaging. But, the mechanism itself is not understood yet. One may think that the non-buckled dimmer in the filled state images are due to a limited spatial resolution. In order to check this point, we took the filled state images at an elevated temperature (47 K), shown in Fig. 2. It clearly shows buckled dimer structure, indicative of a transition between the two temperatures. The observed non-buckled dimer structure shown in Fig. 1 is not caused by a low resolution. Theoretical calculation including spin configuration in the dimer structure was reported by Artacho and Yndur!ain [10]. According to their results, symmetric dimer with an anti-ferromagnetic spin arrangement is most stable. Their result encourages us to observe the dimer structure under a magnetic field. By a charge transfer from the upper atom to the lower atom, the buckled dimer structure is considered to have antiparallel singlet spin configuration. Expecting a transition to symmetric dimer, which may have a magnetically favored triplet spin configuration, we took several STM images under magnetic field up to 10 T at 10 K in the perpendicular direction to the surface, both upper and lower directions, but we could not find any significant changes in a structure and arrangement of dimer on the surface. This work was partly supported by a Grant-in-Aid for Scientific Research (Nos. 12450020, 13NP0201) from the Ministry of Education, Science, Sports, and Culture of Japan.
References [1] R.M. Schlier, H.E. Fransworth, J. Chem. Phys. 30 (1959) 917. . [2] R.J. Hamers, U.K. Kohler, J. Vac. Sci. Technol. A 7 (1989) 2854. [3] R.A. Wolkow, Phys. Rev. Lett. 68 (1992) 2636. [4] K. Hata, et al., Phys. Rev. Lett. 86 (2001) 3084.
1646
M. Ono et al. / Physica B 329–333 (2003) 1644–1646
[5] Y. Kondo, et al., Surf. Sci. 453 (2000) L318. [6] T. Yokoyama, K. Takayanagi, Phys. Rev. B 61 (2000) R5078. [7] H. Shigekawa, et al., Jpn. J. Appl. Phys. 35 (1996) L1081.
[8] T. Tabata, et al., Surf. Sci. Lett. 179 (1987) L63. [9] H. Kageshima, M. Tsukada, Phys. Rev. B 46 (1992) 6928. [10] E. Artacho, F. Yndur!ain, Phys. Rev. Lett. 62 (1989) 2491.
Dimer structure of the Sið0 0 1Þ2 1 surface observed by low-temperature scanning tunneling microscope Masanori Ono*, A. Kamoshida, N. Matsuura, T. Eguchi, Y. Hasegawa Institute for Solid State Physics, The University of Tokyo 5-1-5, Kashiwa-no-ha, Kashiwa, Chiba 277-8581, Japan
Abstract Using a scanning tunneling microscope (STM) which can be operated in ultrahigh vacuum (UHV), low temperature (> 2:8 K), and magnetic field (o11 T), we studied dimer structure on the Sið0 0 1Þ2 1 surface. Asymmetric (buckled) dimer structure was observed with positive sample voltages, while most of the dimers look symmetric with negative voltages. These observations suggest that the observed symmetric dimer in the filled state images is not intrinsic; neither static symmetric dimer nor flip–flopped asymmetric dimer, but a matter of STM imaging mechanism. Magnetic filed, up to 10 T; applied perpendicularly to the surface does not affect the STM imaging at 10 K: r 2003 Elsevier Science B.V. All rights reserved. PACS: 68.35.Bs; 61.16.Ch; 68.35.Rh Keywords: Silicon surface; Silicon dimer structure; Sið0 0 1Þ2 1; Scanning tunneling microscopy
On Si(0 0 1) surface, two silicon atoms making a bond in the /1 1 0S direction to reduce the number of dangling bonds, a dimer structure is formed [1]. The dimers are arranged in the /1 1% 0S direction to form a ‘‘dimer row’’ structure with a long range ordering of 2 1 periodicity. The unique feature that makes the surface interesting is buckling of the dimer, that is, one of the two dimer-composing atoms protruding while the other lowered. In the dimer-row direction, the asymmetric ‘‘buckled’’ dimer exhibits an ‘‘anti-ferromagnetic’’-like interaction, making a zigzag pattern. Between the neighboring dimer-rows, interaction can be either ‘‘ferromagnetic’’ or ‘‘anti-ferromagnetic’’, producing a pð2 2Þ or cð4 2Þ ordering, respectively. A flip–flop motion of the buckled dimer can be induced by thermal activation or by a proximity effect of the probe tip in STM, but the motion can be pinched off near local defects, such as steps and missing dimers. Lots of theoretical and experimental studies using various techniques, including STM [2–7], have been
*Corresponding author. E-mail address: [email protected] (M. Ono).
performed. Generally accepted scenario says that an order–disorder transition occurs at around 200 K; below the temperature, the buckled dimers are arranged with an ordering of cð4 2Þ; and above the temperature the thermally activated flip–flop motion exhibits the 2 1 ordering [3,4,8]. Recently, Kondo et al. [5] reported STM images taken below the temperature showing nonbuckled dimers, and claimed a new ground state structure. Yokoyama and Takayanagi, then, showed an evidence of cð4 2Þ electronic structure at 5 K by STM, attributing the non-buckled dimer imaging to anomalous flip–flop motion [6]. Before their observations, Shigekawa et al. [7] demonstrated buckled-dimer STM images below the temperature. But there is an argument on whether the reported temperature is correct or not [5]. In order to solve the undetermined situation, we carried out STM experiments at correctly measured temperature (10 K). Our results show non-buckled images with positive sample voltages (filled state images). But, with negative sample voltages (empty state images), buckled dimer formation was clearly observed. These observations suggest asymmetric dimer structure at the low temperature and that apparent
0921-4526/03/$ - see front matter r 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0921-4526(02)02439-0
M. Ono et al. / Physica B 329–333 (2003) 1644–1646
1645
Fig. 2. STM image taken at 47 K with a negative sample bias voltage. The size of the observed area is 7:0 nm 3:5 nm: The sample bias voltage and tunneling current is 2:03 V and 68 pA:
Fig. 1. STM images taken on the same area at 10 K with a negative (left) and positive (right) sample bias voltage. The observed area is 14 nm 7 nm: The sample bias voltage and tunneling current is 1:47 V and 42 pA (left), 1:51 V and 42 pA (right), respectively.
symmetric dimer structure in the filled state images is caused by STM imaging mechanism. Silicon samples (n-type, Sb-doped, 0.014–0:005 O cm) were cleaned by 12001C annealing. The sample temperature was measured with a resistance sensor (Cernox), which was calibrated with the one attached on a dummy sample surface. Fig. 1 shows STM images taken on same area with negative (left) and positive (right) sample bias voltages. Filled and empty states of the sample are imaged, respectively. The filled state image shows many linear features, ‘‘dimer-row’’, running parallel each other. Most area apparently shows non-buckled symmetric dimers. We took images with a bias voltage ranging from 3 to 1:3 V; and non-buckled dimer structure always covers most of the surface. The empty state image of the same area (Fig. 1, right), shows remarkable differences from its counterpart. Almost all area shows a zigzag pattern, that is, ‘‘anti-ferromagnetic’’ dimer arrangement along the dimer-row direction, supporting a presence of buckled dimers on the surface. Applied bias voltage ranging from 3 to 1:6 V does not affect the observation. Theoretical calculation [9] suggests the upper atom in the buckled dimer looks brighter (darker) in the filled (empty) state images. Correlation of the zigzag pattern with a neighboring dimer row has two types: in and out of phases, corresponding to pð2 2Þ and cð4 2Þ structures, respectively. Occasionally, we observed a motion of the phase boundary and transitional fuzzy-looking regions in sequentially taken STM images, as was observed by Shigekawa et al. [7]. In spite of the motion, overall phase domain structure was maintained in empty-state images. These observation exclude a possi-
bility that apparently non-buckled dimer is caused by the flip–flop motion either thermally or tip-induced. We speculated that the apparent symmetric dimer structure is due to STM imaging. But, the mechanism itself is not understood yet. One may think that the non-buckled dimmer in the filled state images are due to a limited spatial resolution. In order to check this point, we took the filled state images at an elevated temperature (47 K), shown in Fig. 2. It clearly shows buckled dimer structure, indicative of a transition between the two temperatures. The observed non-buckled dimer structure shown in Fig. 1 is not caused by a low resolution. Theoretical calculation including spin configuration in the dimer structure was reported by Artacho and Yndur!ain [10]. According to their results, symmetric dimer with an anti-ferromagnetic spin arrangement is most stable. Their result encourages us to observe the dimer structure under a magnetic field. By a charge transfer from the upper atom to the lower atom, the buckled dimer structure is considered to have antiparallel singlet spin configuration. Expecting a transition to symmetric dimer, which may have a magnetically favored triplet spin configuration, we took several STM images under magnetic field up to 10 T at 10 K in the perpendicular direction to the surface, both upper and lower directions, but we could not find any significant changes in a structure and arrangement of dimer on the surface. This work was partly supported by a Grant-in-Aid for Scientific Research (Nos. 12450020, 13NP0201) from the Ministry of Education, Science, Sports, and Culture of Japan.
References [1] R.M. Schlier, H.E. Fransworth, J. Chem. Phys. 30 (1959) 917. . [2] R.J. Hamers, U.K. Kohler, J. Vac. Sci. Technol. A 7 (1989) 2854. [3] R.A. Wolkow, Phys. Rev. Lett. 68 (1992) 2636. [4] K. Hata, et al., Phys. Rev. Lett. 86 (2001) 3084.
1646
M. Ono et al. / Physica B 329–333 (2003) 1644–1646
[5] Y. Kondo, et al., Surf. Sci. 453 (2000) L318. [6] T. Yokoyama, K. Takayanagi, Phys. Rev. B 61 (2000) R5078. [7] H. Shigekawa, et al., Jpn. J. Appl. Phys. 35 (1996) L1081.
[8] T. Tabata, et al., Surf. Sci. Lett. 179 (1987) L63. [9] H. Kageshima, M. Tsukada, Phys. Rev. B 46 (1992) 6928. [10] E. Artacho, F. Yndur!ain, Phys. Rev. Lett. 62 (1989) 2491.