- Email: [email protected]
Initial Ag adsorption on Si(5 5 12)-2 × 1 and its parasitic Si(7 7 17)-2 × 1
Surface Science 693 (2020) 121547
Contents lists available at ScienceDirect
Surface Science journal homepage: www.elsevier.com/locate/susc
Initial Ag adsorption on Si(5 5 12)-2 × 1 and its parasitic Si(7 7 17)-2 × 1 a,⁎
b
Sanghee Cho , Zhang J. , Jae M. Seo a b c
T
c
Korea Research Institute of Standards and Science, Daejeon 305-340, Republic of Korea Department of Physics, Yunnam University, Kunming 650091, PR China Department of Physics and Institute of Photonics and Information Technology, Chonbuk National University, Jeonju, 561-756, Republic of Korea
A R T I C LE I N FO
A B S T R A C T
Keywords: Ag High-miller index Si surfaces Si(5 5 12) Si(7 7 17) STM Initial adsorption step
The initial adsorption of Ag atoms on Si(5 5 12)-2 × 1 was investigated by scanning tunneling microscopy. The Ag atoms initially adsorb on a special site keeping its periodicity. Preferentially they adsorb on the site between the tetramer in D1 section and π chain. With additional Ag, the Ag-induced protrusion appears to have a 3 × periodicity different from the 2 × periodicity of the substrate along [11¯0]. On the other hand, the adsorption of Ag atoms on Si(7 7 17)-2 × 1 existing on Si(5 5 12)-2 × 1 under compressive stress induces double Ag rows. On both cases, the 1 × chain at the boundary between D2 and D1sections is firstly collapsed by the Ag adsorption. The Ag nanostructure formed by 0.2ML on Si(5 5 12) start to show the metallic feature.
1. Introduction Formation and characteristics of self-assembled one-dimensional (1D) nanostructure on Si(111) or Si(001) have intensively been studied by surface techniques. On the reconstructed high-index Si surfaces between (001) and (111), 1-D structural features along [11¯0] have also been investigated [1]. They can be potential templates for fabricating self-assembled nanowires with regular spacing of nanometer scale and with 1-D symmetry if the deposited atoms preferentially react any of those 1-D features. Hence, in the present study, Ag had been chosen as depositing metal and Si(5 5 12) which is known to be reconstructed as a relatively planar surface with 5.3 nm periodicity had been selected as a substrate in order to understand the formation of self-assembled nanowire fabrication through in-situ scanning tunneling microscopy (STM). Formation of 1-D structure of the noble metals on Si(5 5 12) [2–11], the epitaxial growth like Ge/Si(5 5 12) [12] or Si/Si(5 5 12) [13], and the initial adsorption of the surfactant such as Sb, Bi and Ga had been investigated by STM, PES and LEED [14–18]. In the case of noble metals like Ag and Au, adsorbed atoms form the 1-D nanostructure along the row direction on the tetramer row inside D1 section (that is, a (337) subsection with a tetramer row). Due to the different bond lengths, the adsorption of different atoms is often accompanied by structural deformation to the structure with a different periodicity. Also, the investigation on the initial adsorption of organic molecules like furan, benzene, pyridine, and pyrrole were conducted [19–25]. These organic molecules preferentially adsorbed on the adatom rows are located inside D2(that is, a (225) subsection) or D3(that is, a (337) subsection
⁎
with a dimer and adatom row) subsections. Although the fabrication of low-dimensional metallic structure has been intensively studied, the understanding of the initial adsorption is not sufficient and the various results were reported according to the experimental condition [5–9,16]. Baski et al. reported that the 1-D Ag nanowire with sawtooth phase shows the metallic behavior but the double Ag row phase shows a semiconducting behavior [4,9]. On the other hand, the electronic property of the Ag row studied by ARPES shows the semiconducting character [10]. Depending upon analytical instruments employed, reported are different results for the electronic properties of the 1-D nanostructure. Therefore, it is necessary to study on the self-assembled nanostructure formed by a small amount of metal coverage. Up to now, understandings of initial adsorption and corresponding structural/ electronic properties of 1-D Ag wires on Si(5 5 12) are not sufficient. In the present studies, the reconstructed Si(5 5 12) surface employed as a substrate and Ag used as a metal source are chosen in order to understand the atomic structure of nanowire and the corresponding electronic structure. 2. Experimental methods The experiment was carried out in an ultrahigh vacuum (UHV) chamber with a base pressure 2 × 10−10 Torr. The sample was n-type (P-doped) Si(5 5 12) with dimension of 13 × 2 × 0.25 mm3. After degreasing with pure organic solvents in the air, the sample was mounted on the Mo sample holder and quickly transferred to a UHV chamber. The sample was outgassed at 600 °C for 24 h. The native oxide layer was removed by resistive heating through flashing several times at
Corresponding author.
https://doi.org/10.1016/j.susc.2019.121547 Received 18 April 2019; Received in revised form 20 September 2019; Accepted 26 November 2019 Available online 28 November 2019 0039-6028/ © 2019 Elsevier B.V. All rights reserved.
Surface Science 693 (2020) 121547
S. Cho, et al.
prominent and the number of protrusions has increased but a period of the substrate was still maintained. The feature of the Si-addimer, which appeared only in D2 and D3 sections with D/A row shows a little lower contrast than that of the protrusions appeared as adsorbed Ag atoms. After Ag deposition, there are many protrusions that appear brighter than the feature of Si-addimer appeared as bright spots relative to the clean Si (5 5 12) surface image. It is not a feature of the Si-addimer but of the Ag atoms adsorbed on the surface. At the Ag coverage less than 0.1ML, most of 1 × chains survive with a bright rows appearances maintaining the underlying substrate periodicity, 5.35 nm. The Si-addimer, formed between D and A rows (D/A) in D3 section is still appeared as a bright protrusion and also the number is increased. But π chain forming the boundary between D2 and D1 sections is collapsed and the protrusions indicated by “P” appeared as a brightest dot near the collapsed site. In the images obtained from various area after Ag deposition, we observed that the features appearing by Ag atoms adsorbed at some specific sites, unlike the clean surface images. The adsorption sites of the Ag are indicated by 1 to 4 in Fig. 1(d) and are represented in the structural model of reconstructed Si (5 5 12)-2 × 1 as shown in Fig. 1(e). The numbers 1 through 4 represent the following locations: “1” indicates the site between T row within D2 subsection and π chain, “2” represents the site on the π chain, and “3” represents the site between π chain and T row in D1 section. The “4” represents the site between H chain formed the boundary between D1 section and D3 section and the D row within D3 section. Among them, the most “P” appears on the π chain row, i.e., the “2” site and formed with 3a0~6a0 spacing along the row direction, [−1 1 0]. In order to check the preferentially stable adsorption site of Ag atoms, the Ag-deposited sample is needed to annealing. Fig. 1(f) shows the empty-state image acquired after post-annealed at 500 °C after Ag (~0.01ML) deposited on Si(5 5 12). Unlike the results before post-annealing, Ag atoms preferentially form a 1-D row in T row within D1 section after post-annealing at 500 °C. This adsorption site is a similar to the location of the “3” site appeared before annealing. At this coverage, two kinds of 1 × chain structures (i.e., H and π chain rows) are still maintained and also the period of (5 5 12) remains the same. In a very small coverage(<0.05ML), Ag atoms preferentially form a 1-D Ag row on the T row in D1 section, but the adsorption sites of the Ag atom are located in several places as seen from the results before post-annealing, and the Ag atoms predominantly collapsed the π chain as shown in Fig. 1(b)–(d). The formation of these protrusions appears to be induced by Ag (<0.05 ML) deposited at RT. According the reported result of the absorption of Ag on hydrogen-terminated the Si(001) surface, Ag atoms evaporated onto a prepatterned Si(001)-H surface diffuse on the surface and are adsorbed preferentially on Si dangling-bond sites rather than on defects or step edges [33]. In early reported STM results on noble metal such as Cu, Ag, Au/Si(111) system, the metal atoms sit on top of the dangling bonds of the Si adatoms and appeared to single bright spots in STM images. According to the DFT calculation, these noble metal atoms adsorbed on the H3-like and B2-like sites on the corner hexagonal region of faulted half unit cell (FHUC) [34]. It had reported that π chain row has a lot of dangling bonds [35]. The protrusion of the bright spot in Fig. 1(d) is the brighter and larger than that of Si-addimer formed between D row and A row inside D2 and D3 sections. The brightest protrusions are formed near the π chain row between D1 and D2 subsections rather than other kinds of rows (i.e., H chain, D/A row). It can also be deduced that the Ag atom initially absorbs on the location between T row inside D1 section and the π chain. The protrusions appear to be due to the Ag atom adsorbed on/near the π chain row with many dangling bonds. The structural model of one period of Si (5 5 12) is overlapped above the image as shown in Fig. 1(e). Ag atoms appeared as the protrusion adsorbed between tetramer and π chain row match very well with the structural model. The Si-addimer which appears as a single-dot feature in the dark
1150 °C, and the surface was reconstructed through slow cooling from 900 °C to RT with the rate of 2 °C/sec. The reconstructed Si(5 5 12)2 × 1 surface was routinely checked by STM before depositing of Ag atoms. After well-ordered and wide terrace had been confirmed by STM, Ag atoms were thermally evaporated on the reconstructed Si(5 5 12)-2 × 1 surface held at room temperature at a deposition rate of 0.03ML/min., where one monolayer (ML) is defined as 6.8 × 1014 atoms/cm2 (i.e., consists of 28 Si atoms per unit cell)[13] and subsequently annealed at 450 °C–600 °C. All STM images were acquired with the constant current mode at room temperature. 3. Results and discussion Recently, it is revealed that one of the honeycomb chain(H) which forms the boundary of each section on the structural model of Si(5 5 12)-2 × 1 reported by Jeong et al. [26] was appeared to a different feature in empty-state image [27]. They proposed that this H chain which formed the boundary between D2 section and D1 section is changed to π chain. The other boundary consists of the H chain. From these results, the redefined structural model of reconstructed Si(5 5 12) was been proposed [27]. Until now, there were a several proposed model on the reconstructed Si(5 5 12)-2 × 1 [1,26–29]. In the present Ag/Si(5 5 12) system, the redefined structural model will be accepted as a standard model. [See ref. [27] to detailed information] The reconstructed Si(5 5 12) surface has a large unit cell with 5.35 nm and the unit cell of this surface is divided into two (337) subunits with 1.57 nm and one (225) subunit with 2.21 nm. The structural model of Si(5 5 12)-2 × 1 is displayed in Fig. 1(e). The letters such as H, π, D, A, and T imply honeycomb chain, π chain, dimer row, adatom row, and tetramer row, respectively. These rows are formed a 1D structure along the [−1 1 0] direction with a 2 × periodicity except the H chain and π chain with 1 × cycles. One unit cell of Si(5 5 12)2 × 1 consists of D1, D2 and D3 sections which corresponds (337) with tetramer(T) [T(337)], (225), (337) with dimer/adatom(D/A) row [D (337)], respectively. The one unit of Si(5 5 12) is reconstructed to D3D2-D1 subsections along the [6 6 −5] direction perpendicular to the 1D rows. The boundary of two adjacent sections except for that between D2 and D1 sections is formed with H chain, while the boundary between D2 and D1 sections is formed with π chain. Fig. 1(a) shows the atomic resolution image of the reconstructed Si (5 5 12) surface. The white dotted lines shown on Fig. 1 indicate the one period of Si(5 5 12). Atomic resolution image shown in Fig. 1(a) allows us to distinguish that the H and π chains have a 1 × cycle and the remaining structural units such as D, A and T rows are formed in 2 × cycle along the [−1 1 0] direction. A bright dot appeared in D2 and D3 is a feature of Si-addimer adsorbed between dimer (D) and adatom (A) row. A stable reconstructed Si(5 5 12)-2 × 1 has a large terrace with a period of (5 5 12). Occasionally Si(7 7 17) terrace with a period of (7 7 17) which is composed of T(337) section, i.e., D1, and one (225) section, i.e., D2, is observed on the reconstructed Si(5 5 12) surface under local compressive stress [30]. Also, the mixed periods of (5 5 12) and (7 7 17) appeared in the clean reconstructed Si(5 5 12) terrace. The “a” and “b” presented in all figures indicate the periods of (5 5 12) and (7 7 17), respectively. In order to compare the adsorption sites of Ag atoms deposited on Si (5 5 12) surface before and after the heating, the images obtained immediately after the deposition of Ag less than 0.1 ML and the image scanned after post-annealing at ~500 °C are shown in Fig. 1(b)–(d) and Fig. 1(f), respectively. In case of the Ag deposited on Si(111) surface, the thermal desorption temperature of Ag atoms was reported about 560 °C [31,32]. Our samples were post-annealed at lower temperature than the desorption temperature to avoid the desorption of Ag weakly bound on substrate. Fig. 1(b) and (c)–(d) are filled-state images obtained after 0.03 ML and 0.05 ML Ag deposition at room temperature, respectively. The collapse of π chain, which forms the boundary between D1 and D2, is 2
Surface Science 693 (2020) 121547
S. Cho, et al.
Fig. 1. STM images obtained from a clean reconstructed Si(5 5 12)-2 × 1[(a)] and Ag/Si(5 5 12) before[(b): Θ = 0.03 M L, (c) and (d): Θ = 0.05 ML] and after postannealing[(f): Θ = 0.05 ML, post-annealed at 440 °C], respectively. (a): 10.5 nm × 10.5 nm, Vs = −1.9 V, I = 0.5 nA, (b) and (c): 30 nm × 30 nm, Vs = −2.7 V, I = 0.5 nA, (d): 10.5 nm × 10.5 nm, Vs = −2.7 V, I = 0.5 nA). (e) The structure model of reconstructed Si(5 5 12)-2 × 1 surface. (f) The empty-state image taken after post-annealed at 500 °C after Ag(0.01 ML) deposited on Si(5 5 12). (Size; 30 nm × 30 nm, Vs = +2.5 V, I = 0.5 nA). (g) The line profiling results of [A] and [B] lines shown in Fig. 1(f).
form the Ag wires between the π chain and the T row inside D1 section along the [−1 1 0] direction after heat treatment. The preferential adsorption site after post-annealing is most similar to the adsorption site “3” among the various adsorption sites before annealing. Fig. 2 shows a large image of 0.2 ML Ag-adsorbed Si(5 5 12) surface after heat treatment at 440 °C. Fig. 2(a) shows that the upper-left area written by “B” and lower right-hand area written by “A” have different periods. The period of bright row of B area is shorter than that of A area perpendicular to the row direction, [6 6 −5] direction, and can be seen as a shorter aspect ratio along the row direction, [−1 1 0]. More
rows inside D2 and D3 sections still appeared and its number has increased. The line profiles perpendicular to the row direction are shown in Fig. 1(g). Line A is the line profiling across the 1-D Ag row. The structure of substrate can be identified through the line profiling of line B. This means that the formation position of the Ag row is known. From the line-profile of line A, it has been shown that the brightest row is formed inside D1 section, and the height of the π chain between D2 and D1 section is higher than that of other 1 × chains. As shown from the Fig. 2, the π chain is firstly collapsed by Ag atom, and the chain has been transformed to a different kinds of structure. Ultimately, Ag atoms 3
Surface Science 693 (2020) 121547
S. Cho, et al.
Fig. 2. Filled-state large images obtained after post-annealing at 440 °C on 0.2 ML Ag adsorbed Si(5 5 12) surface. (b) and (c) obtained from the terrace of each area written by “A” on the upper-left area and “B” on the lower-right terrace of (a). (a) Size; 0.3 μm × 0.3 μm, Vs = −2.7 V, (b) 40 nm × 40 nm, Vs = −2.7 V. (c) 40 nm × 40 nm, Vs = −1.9 V.
1 × cycles. Based on the 1 × chain structure with relatively dark contrasts that survives in several places, the period (7 7 17) can be estimated. Line profiling was performed to compare the height and relative distance of the structure represented by the Ag adsorption shown in Fig. 4(b). The line [A], marked by a red line in the image, is the result of line profiling in a direction perpendicular to the row shown below the image. One period of (7 7 17) in the image was matched with one period of (7 7 17) in line profiling. By comparing the height difference in the line profiling results perpendicular to the 1-D structural row direction, one can see whether the 1 × chain row is on a clean substrate or on an Ag-deposited surface. The profiling of the line [B] in the same direction was presented in Fig. 4(d) to determine the relative position of the row in the uppermost layer of Ag. One period of Si(7 7 17) is represented by a white dotted line in the image. As a result of the line [A] analysis, the initial absorption of the Ag atom is indicated by a row forming 1 × cycle on the T row in D2 section. The ``α'' and ``γ'' marked in the line profiling analysis shown in Fig. 4(b) are the locations of adsorption site corresponding to “2” or “1” adsorption sites, shown in Fig 1(d) and (e), respectively. Another row with 1 × period appears at the position of the π chain row that forms the boundary between D1 and D2 sections. The two bright rows formed about 0.7 nm apart. The height of γ site in this double Ag row is higher than that of α site. The distance of double row in the same height holds the period of the substrate. The period of double Ag row is indicated by a black dotted line in the image. The dot with 2 × cycle was found around the bright double row. As a result of Ag adsorbed on Si(7 7 17) terrace, it is difficult to identify the shape of the structural unit such as H chain, D/A row, T row and π chain that originally forms Si(7 7 17). There is a new double row feature that looks brightest on top of the first double Ag row formed. The result of the line profiling of [B] is displayed at the bottom. The relative height difference between the first and second double Ag rows is about less than 0.05 nm. One period of double Ag row is marked by a thick red dotted line in the image. The second layer of double Ag row is located between the first layer of double Ag row and is formed the shift with about 0.44 nm toward [6 6 −5] in the perpendicular to the row direction. The distance between these double Ag rows is also 0.66 nm, which is consistent with the result of the first layer of double Ag row. In line [B], the absorption locations of the second layer of double Ag row are marked with α′ and γ′. The height of γ′ site also is higher than that of α′ site. In terms of the periodicity of the second layer along the row direction, four bright spots appeared as the shape of the teeth was paired with each other and has a 3 × period along the row direction presented a pair with a square of yellow dotted lines. The contrast of two bright spots on the left side of the pair is slightly brighter than that of the two bright spots on the
detailed STM images obtained from each terrace are shown in Fig. 2(b) and (c) respectively. Fig. 2(a) shows that each area shown as A and B is terrace with (5 5 12) period and terrace with (7 7 17) period, respectively. In the Agadsorbed on Si(5 5 12) terrace shown in Fig. 2(b), only the H chain structure that forms the boundary between D2 and D3 sections survives and maintains the period (5 5 12) as the boundary. Ag atoms adsorbed on the site between D1 and D2 sections and form the 1D-like row along the [−1 1 0] direction. Meanwhile, the area “B” marked in Fig. 2(a) is the Ag-adsorbed Si(7 7 17) terrace consisting of D1 + +D2 as shown in Fig. 2(c). This is not a transition from (5 5 12) period to (7 7 17) period as a result of Ag adsorption. The absorption results of Ag over this Si(7 7 17) area show the differences from the characteristics of the feature of Ag adsorbed over Si(5 5 12) terrace. We can see the break to a specific cycle along the row direction and the relatively short Ag rows, unlike the case of Ag adsorbed on Si(5 5 12). It is also difficult to identify all of 1 × chain structures in this area. In order to understand the absorption process of Ag in each region, the analysis of each area are shown in Figs. 3 and 4 respectively. The filled-state STM image of Fig. 3(a) is obtained from 0.2ML Ag-adsorbed on Si(5 5 12) surface after post-annealed at 440 °C. The H chain row that is boundary between D2 and D3 sections only survives. That is indicated by H with a white dotted line. Protrusions caused by Si-addimer appear as bright spots on the left and right sides of this H chain. The red bar represents a period of (5 5 12). The boundary located between D1 and D2 sections, i.e., π chain, is no longer identified by the adsorption of the Ag atoms, and 1D-like Ag rows are formed on T row inside D1 section. The result of line profiling perpendicular to a one-dimensional row is shown in Fig. 3(b) below the image. At the bottom, a structural model was placed as shown in Fig. 3(c), consistent with one cycle of (5 5 12) in the STM image. The highest bright row is formed around the boundary of D1 and D2. The “α” and “β” marked in the line profiling analysis are the locations of adsorption site corresponding to the “2” or “3” adsorption sites shown in Fig 1(d) and (e), respectively. That is, the Ag atoms are preferentially adsorbed on the π chain and break the structure of the weak π chain row with 1 × period due to the additional adsorbed Ag atoms, and the top layer of Ag is located between T row inside D1 section and H chain row located between D1 and D3 sections and form the 1-D Ag rows with a high aspect ratio. In the row direction, the period of Ag row shows a relatively regular periodicity of three times. The period between the brightest Ag rows still follows the period of the (5 5 12) substrate in the direction perpendicular to the row. The filled-state image obtained after post-annealed at 440 °C of 0.2ML Ag/Si(7 7 17) system is shown in Fig. 4(a). Unlike the Ag-adsorbed Si (5 5 12) surface, it is difficult to find a chain structure with 4
Surface Science 693 (2020) 121547
S. Cho, et al.
Fig. 3. (a) Image taken from the area which (5 5 12) period is sustained of Fig. 2(b). Size; 17.4 nm × 17.4 nm, Vs = −2.7 V, I = 0.5 nA. (b) The cross-sectional lineprofiling of image is displayed below image of (a). (c) The structure model of reconstructed (5 5 12) period.
the Ag row, which forms the brightest band. Due to the thermal drift and thermal broadening, it is not easy to distinguish the data from the Ag row formed with a small amount of Ag and the row of the exposed substrate. Therefore, the coverage of Ag was increased to widen the adsorbed Ag area. It is difficult to identify the H chain or π chain that can estimate the period of the substrate as shown in Fig. 5(a). However, it can be observed that a bright region appears due to the adsorption of Ag and a dark region where the substrate is exposed. The results of the STS in the bright and dark areas show distinctly different results. This is the result of repeated experiments of STS in the other scan areas of the same sample, that is, in the dark area and the bright areas. The STS results with blue and red colors were measured on the highest Ag row and that with green color was measured on the lowest dark area. The results are shown in Fig. 5(b). On Ag row, it was observed to metallic character with a small band gap but the semiconducting behavior is appeared at the darkest area that is predicted to bare Si substrates. The results of electronic characteristics of Ag nanostructure formed by increased coverage of Ag are still insufficient. It is expected that it will be helpful to understand the electronic properties of 1-D Ag nanostructure through comparison of the reported results.
right. Unlike the second double Ag row, which breaks into a three-fold periodicity in the direction of row, there is also a double Ag row that creates 1 × -cycle along the row direction. Support of the theoretical calculation will be needed for more detailed adsorption analysis in Agadsorbed Si(7 7 17) system. Baski's group had reported the results for the Ag/Si(5 5 12) system using RHEED, STM/S and PES. For low coverages (θ < 0.25 ML), Ag atoms form very long monoatomic chains on top of the T rows and make a very weak 3 × periodicity along the row direction which agrees with our results [3,4,9]. We performed the STS spectrum at the Ag-adsorbed Si(5 5 12) surface at room temperature. Fig. 5(a) shows the filled-state STM image of Ag-deposited on Si(5 5 12) and sequentially annealed at 450 °C. The H chain rows of the substrate are rarely visible, despite a small amount of Ag adsorption of 0.3 ML. It also shows a sawtooth-like feature as the width of the Ag row is widened in the [6 6 −5] direction, perpendicular to the row. As the height of the Ag row increases, the slope opposite to the direction of gentle slope becomes steep. So it appears as dark contrast in the image. The line profiling results are shown in the image wing to see the period between Ag rows of the same height. It can be seen that the period of the H chain of the substrate is the region of the substrate where the periods of (7 7 17) and (5 5 12) denoted by “a” and “b” are mixed, respectively. STS was performed to understand the electrical characteristics of 5
Surface Science 693 (2020) 121547
S. Cho, et al.
Fig. 4. (a) Image taken from the area which (7 7 17) period is sustained of Fig. 2(c). Size; 10 nm × 10 nm, Vs = −1.6 V, I = 0.5 nA. (b) and (d) The cross-sectional line-profiling of lines of [A] and [B] are displayed. (c) Structure model of reconstructed Si(7 7 17) period.
4. Conclusions
with two pairs of different brightness. The cycle between the double Ag rows also keeps the (7 7 17) period of the substrate. These results imply that the adsorption of Ag atoms takes preferentially over the T rows with high dangling bond density inside the D1 and D2 sections among the structural units forming the structure of the substrate and forms a 1-D Ag nanostructure. The formed 1-D Ag row keeps the identical periodicity of the substrate and shows a metallic characteristics.
The results obtained from Ag adsorbed on Si(5 5 12) terrace are quite matching with those reported previously. However, the Ag-adsorbed Si(7 7 17) surface shows a different aspect. Unlike the results of Ag/Si(5 5 12), 1 × chains of Ag/Si(7 7 17) are difficult to identify. Firstly, the 1 × structure is collapsed by Ag adsorption. The Ag atoms initially adsorbed on top of T row in D2 section and on the π chain, which implies that they form a double Ag row with (7 7 17) period. The second Ag row between the first double Ag rows is formed, and another second Ag row is formed on the T row in D1 section. These second Ag rows also form the double feature appear to be the brightest rows. Similar to the results of Ag/Si(5 5 12), the cycle in the direction perpendicular to the row is three times. Four bright spots are clustered
Declaration of Competing Interest None.
6
Surface Science 693 (2020) 121547
S. Cho, et al.
Fig. 5. STS data at RT from 0.3 ML Ag deposited on Si(5 5 12) and sequentially annealed at 450 °C. (a) filled-state STM image (Size; 40 nm × 40 nm, Vs = −2.6 V, I = 0.7 nA). Line profiling is shown under the image. (b) STS data obtained from the dark area (circle filled with green color), bright area (circles filled with red and blue color) in image (a), respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Acknowledgments
[15] M. Kumar, V.K. Paliwal, A.G. Joshi, Govind S.M. Shivaprasad, Surf. Sci. 596 (2005) 206. [16] S. Cho, J.M. Seo, Surf. Sci. 565 (2004) 14. [17] P. Kumar, M. Kumar, B.R. Mehta, S.M. Shivaprasad, Surf. Sci. 605 (2011) 1426. [18] P. Kumar, M. Kumar, S.M. Shivaprasad, Appl. Surf. Sci. 282 (2013) 348. [19] J.R. Hahn, H. Jeong, S. Jeong, J. Chem. Phys. 123 (2005) 244702. [20] J.R. Hahn, H. Jeong, S. Jeong, S.H. Jang, Jpn. J. Appl. Phys. 45 (2006) 2175. [21] H. Jeong, S. Jeong, S.H. Jang, J.M. Seo, J.R. Hahn, J. Phys. Chem. B 110 (2006) 15912. [22] S.H. Jang, S. Jeong, J.R. Hahn, J. Chem. Phys 130 (2009) 234703. [23] J.R. Hahn, S.C. Bharath, G. Kim, S. Jeong, T. Pearl, J. Phys. Chem C115 (2011) 17111. [24] J.R. Hahn, S.C. Bharath, G. Kim, K.W. Kim, S. Jeong, T. Pearl, J. Phys. Chem. C117 (2013) 11197. [25] J.R. Hahn, G. Kim, K.W. Kim, S. Jeong, Surf. Sci. 616 (2013) 166. [26] S. Jeong, H. Jeong, S. Cho, J.M. Seo, Surf. Sci. 557 (2004) 183. [27] H. Kim, H. Li, Y. Zhu, J.R. Hahn, J.M. Seo, Surf. Sci. 601 (2007) 1831. [28] W. Ranke, Y.R. Xing, Surf. Sci. 381 (1997) 1. [29] J. Liu, M. Takeguchi, H. Yasuda, K. Furuya, J. Cryst, Growth 237 (2002) 188. [30] S. Cho, J.M. Seo, J. Korean Phys. Soc. 49 (2006) 181. [31] S. Hasegawa, H. Daimon, S. Ino, Surf. Sci. 186 (1987) 138. [32] K.J. Wan, X.F. Lin, J. Nogami, Phys. Rev. B 47 (1993) 13700. [33] M. Sakurai, C. Thirstrup, M. Aono, Phys. Rev. B 62 (2000) 16167. [34] C. Zhang, G. Chen, K. Wang, H. Yang, T. Su, C.T. Chan, M.M. Loy, X. Xiao, Phys. Rev. Lett 94 (2005) 176104. [35] H. Kim, H. Li, J.M. Seo, J. Vac, Sci. Technol. B25 (2007) 1511.
This work was supported by the basic R&D program of the Korean Research Institute of Standards and Science. References [1] A. Baski, S.C. E.rwin, L.J. Whitman, Science 269 (1995) 1556. [2] H.H. Song, K.M. Jones, A.A. Baski, J. Vac. Sci. Technol. A 17 (1999) 1696. [3] S.R. Blankenship, H.H. Song, A.A. Baski, J.A. Carlisle, J. Vac. Sci. Technol. A 17 (1999) 1615. [4] K.M. Jones, H.H. Song, A.A. Baski, J. Clust. Sci. 10 (1999) 573. [5] A. Baski, K.M. Saoud, J. Clust. Sci. 12 (2001) 527. [6] Y. Peng, H. Minoda, Y. Tanishiro, K. Yagi, Surf. Sci. 493 (2001) 508. [7] A. Baski, K.M. Saoud, K.M. Jones, Appl. Surf. Sci. 182 (2001) 216. [8] J.W. Dickinson, J.C. Moore, A.A. Baski, Surf. Sci. 561 (2004) 193. [9] A. Baski, K.M. Jones, K.M. Saoud, Ultramicroscopy 86 (2001) 23. [10] J.R. Ahn, Y.J. Kim, H.S. Lee, C.C. Hwang, B.S. Kim, H.W. Yeom, Phys. Rev. B66 (2002) 153403. [11] J.R. Ahn, H.W. Yeom, E.S. Cho, C.Y. Park, Phys. Rev. B69 (2004) 233311. [12] H. Kim, Huiting Li, J.M. Seo, Surf. Sci. 602 (2008) 2563. [13] H. Kim, Y. Cho, J.M. Seo, Surf. Sci. 583 (2005) 265. [14] Huiting Li, Y.-N. Xu, Y.-Z. Zhu, Hidong Kim, Jae M. Seo, Phys. Rev. B 75 (2007) 235442.
7
Contents lists available at ScienceDirect
Surface Science journal homepage: www.elsevier.com/locate/susc
Initial Ag adsorption on Si(5 5 12)-2 × 1 and its parasitic Si(7 7 17)-2 × 1 a,⁎
b
Sanghee Cho , Zhang J. , Jae M. Seo a b c
T
c
Korea Research Institute of Standards and Science, Daejeon 305-340, Republic of Korea Department of Physics, Yunnam University, Kunming 650091, PR China Department of Physics and Institute of Photonics and Information Technology, Chonbuk National University, Jeonju, 561-756, Republic of Korea
A R T I C LE I N FO
A B S T R A C T
Keywords: Ag High-miller index Si surfaces Si(5 5 12) Si(7 7 17) STM Initial adsorption step
The initial adsorption of Ag atoms on Si(5 5 12)-2 × 1 was investigated by scanning tunneling microscopy. The Ag atoms initially adsorb on a special site keeping its periodicity. Preferentially they adsorb on the site between the tetramer in D1 section and π chain. With additional Ag, the Ag-induced protrusion appears to have a 3 × periodicity different from the 2 × periodicity of the substrate along [11¯0]. On the other hand, the adsorption of Ag atoms on Si(7 7 17)-2 × 1 existing on Si(5 5 12)-2 × 1 under compressive stress induces double Ag rows. On both cases, the 1 × chain at the boundary between D2 and D1sections is firstly collapsed by the Ag adsorption. The Ag nanostructure formed by 0.2ML on Si(5 5 12) start to show the metallic feature.
1. Introduction Formation and characteristics of self-assembled one-dimensional (1D) nanostructure on Si(111) or Si(001) have intensively been studied by surface techniques. On the reconstructed high-index Si surfaces between (001) and (111), 1-D structural features along [11¯0] have also been investigated [1]. They can be potential templates for fabricating self-assembled nanowires with regular spacing of nanometer scale and with 1-D symmetry if the deposited atoms preferentially react any of those 1-D features. Hence, in the present study, Ag had been chosen as depositing metal and Si(5 5 12) which is known to be reconstructed as a relatively planar surface with 5.3 nm periodicity had been selected as a substrate in order to understand the formation of self-assembled nanowire fabrication through in-situ scanning tunneling microscopy (STM). Formation of 1-D structure of the noble metals on Si(5 5 12) [2–11], the epitaxial growth like Ge/Si(5 5 12) [12] or Si/Si(5 5 12) [13], and the initial adsorption of the surfactant such as Sb, Bi and Ga had been investigated by STM, PES and LEED [14–18]. In the case of noble metals like Ag and Au, adsorbed atoms form the 1-D nanostructure along the row direction on the tetramer row inside D1 section (that is, a (337) subsection with a tetramer row). Due to the different bond lengths, the adsorption of different atoms is often accompanied by structural deformation to the structure with a different periodicity. Also, the investigation on the initial adsorption of organic molecules like furan, benzene, pyridine, and pyrrole were conducted [19–25]. These organic molecules preferentially adsorbed on the adatom rows are located inside D2(that is, a (225) subsection) or D3(that is, a (337) subsection
⁎
with a dimer and adatom row) subsections. Although the fabrication of low-dimensional metallic structure has been intensively studied, the understanding of the initial adsorption is not sufficient and the various results were reported according to the experimental condition [5–9,16]. Baski et al. reported that the 1-D Ag nanowire with sawtooth phase shows the metallic behavior but the double Ag row phase shows a semiconducting behavior [4,9]. On the other hand, the electronic property of the Ag row studied by ARPES shows the semiconducting character [10]. Depending upon analytical instruments employed, reported are different results for the electronic properties of the 1-D nanostructure. Therefore, it is necessary to study on the self-assembled nanostructure formed by a small amount of metal coverage. Up to now, understandings of initial adsorption and corresponding structural/ electronic properties of 1-D Ag wires on Si(5 5 12) are not sufficient. In the present studies, the reconstructed Si(5 5 12) surface employed as a substrate and Ag used as a metal source are chosen in order to understand the atomic structure of nanowire and the corresponding electronic structure. 2. Experimental methods The experiment was carried out in an ultrahigh vacuum (UHV) chamber with a base pressure 2 × 10−10 Torr. The sample was n-type (P-doped) Si(5 5 12) with dimension of 13 × 2 × 0.25 mm3. After degreasing with pure organic solvents in the air, the sample was mounted on the Mo sample holder and quickly transferred to a UHV chamber. The sample was outgassed at 600 °C for 24 h. The native oxide layer was removed by resistive heating through flashing several times at
Corresponding author.
https://doi.org/10.1016/j.susc.2019.121547 Received 18 April 2019; Received in revised form 20 September 2019; Accepted 26 November 2019 Available online 28 November 2019 0039-6028/ © 2019 Elsevier B.V. All rights reserved.
Surface Science 693 (2020) 121547
S. Cho, et al.
prominent and the number of protrusions has increased but a period of the substrate was still maintained. The feature of the Si-addimer, which appeared only in D2 and D3 sections with D/A row shows a little lower contrast than that of the protrusions appeared as adsorbed Ag atoms. After Ag deposition, there are many protrusions that appear brighter than the feature of Si-addimer appeared as bright spots relative to the clean Si (5 5 12) surface image. It is not a feature of the Si-addimer but of the Ag atoms adsorbed on the surface. At the Ag coverage less than 0.1ML, most of 1 × chains survive with a bright rows appearances maintaining the underlying substrate periodicity, 5.35 nm. The Si-addimer, formed between D and A rows (D/A) in D3 section is still appeared as a bright protrusion and also the number is increased. But π chain forming the boundary between D2 and D1 sections is collapsed and the protrusions indicated by “P” appeared as a brightest dot near the collapsed site. In the images obtained from various area after Ag deposition, we observed that the features appearing by Ag atoms adsorbed at some specific sites, unlike the clean surface images. The adsorption sites of the Ag are indicated by 1 to 4 in Fig. 1(d) and are represented in the structural model of reconstructed Si (5 5 12)-2 × 1 as shown in Fig. 1(e). The numbers 1 through 4 represent the following locations: “1” indicates the site between T row within D2 subsection and π chain, “2” represents the site on the π chain, and “3” represents the site between π chain and T row in D1 section. The “4” represents the site between H chain formed the boundary between D1 section and D3 section and the D row within D3 section. Among them, the most “P” appears on the π chain row, i.e., the “2” site and formed with 3a0~6a0 spacing along the row direction, [−1 1 0]. In order to check the preferentially stable adsorption site of Ag atoms, the Ag-deposited sample is needed to annealing. Fig. 1(f) shows the empty-state image acquired after post-annealed at 500 °C after Ag (~0.01ML) deposited on Si(5 5 12). Unlike the results before post-annealing, Ag atoms preferentially form a 1-D row in T row within D1 section after post-annealing at 500 °C. This adsorption site is a similar to the location of the “3” site appeared before annealing. At this coverage, two kinds of 1 × chain structures (i.e., H and π chain rows) are still maintained and also the period of (5 5 12) remains the same. In a very small coverage(<0.05ML), Ag atoms preferentially form a 1-D Ag row on the T row in D1 section, but the adsorption sites of the Ag atom are located in several places as seen from the results before post-annealing, and the Ag atoms predominantly collapsed the π chain as shown in Fig. 1(b)–(d). The formation of these protrusions appears to be induced by Ag (<0.05 ML) deposited at RT. According the reported result of the absorption of Ag on hydrogen-terminated the Si(001) surface, Ag atoms evaporated onto a prepatterned Si(001)-H surface diffuse on the surface and are adsorbed preferentially on Si dangling-bond sites rather than on defects or step edges [33]. In early reported STM results on noble metal such as Cu, Ag, Au/Si(111) system, the metal atoms sit on top of the dangling bonds of the Si adatoms and appeared to single bright spots in STM images. According to the DFT calculation, these noble metal atoms adsorbed on the H3-like and B2-like sites on the corner hexagonal region of faulted half unit cell (FHUC) [34]. It had reported that π chain row has a lot of dangling bonds [35]. The protrusion of the bright spot in Fig. 1(d) is the brighter and larger than that of Si-addimer formed between D row and A row inside D2 and D3 sections. The brightest protrusions are formed near the π chain row between D1 and D2 subsections rather than other kinds of rows (i.e., H chain, D/A row). It can also be deduced that the Ag atom initially absorbs on the location between T row inside D1 section and the π chain. The protrusions appear to be due to the Ag atom adsorbed on/near the π chain row with many dangling bonds. The structural model of one period of Si (5 5 12) is overlapped above the image as shown in Fig. 1(e). Ag atoms appeared as the protrusion adsorbed between tetramer and π chain row match very well with the structural model. The Si-addimer which appears as a single-dot feature in the dark
1150 °C, and the surface was reconstructed through slow cooling from 900 °C to RT with the rate of 2 °C/sec. The reconstructed Si(5 5 12)2 × 1 surface was routinely checked by STM before depositing of Ag atoms. After well-ordered and wide terrace had been confirmed by STM, Ag atoms were thermally evaporated on the reconstructed Si(5 5 12)-2 × 1 surface held at room temperature at a deposition rate of 0.03ML/min., where one monolayer (ML) is defined as 6.8 × 1014 atoms/cm2 (i.e., consists of 28 Si atoms per unit cell)[13] and subsequently annealed at 450 °C–600 °C. All STM images were acquired with the constant current mode at room temperature. 3. Results and discussion Recently, it is revealed that one of the honeycomb chain(H) which forms the boundary of each section on the structural model of Si(5 5 12)-2 × 1 reported by Jeong et al. [26] was appeared to a different feature in empty-state image [27]. They proposed that this H chain which formed the boundary between D2 section and D1 section is changed to π chain. The other boundary consists of the H chain. From these results, the redefined structural model of reconstructed Si(5 5 12) was been proposed [27]. Until now, there were a several proposed model on the reconstructed Si(5 5 12)-2 × 1 [1,26–29]. In the present Ag/Si(5 5 12) system, the redefined structural model will be accepted as a standard model. [See ref. [27] to detailed information] The reconstructed Si(5 5 12) surface has a large unit cell with 5.35 nm and the unit cell of this surface is divided into two (337) subunits with 1.57 nm and one (225) subunit with 2.21 nm. The structural model of Si(5 5 12)-2 × 1 is displayed in Fig. 1(e). The letters such as H, π, D, A, and T imply honeycomb chain, π chain, dimer row, adatom row, and tetramer row, respectively. These rows are formed a 1D structure along the [−1 1 0] direction with a 2 × periodicity except the H chain and π chain with 1 × cycles. One unit cell of Si(5 5 12)2 × 1 consists of D1, D2 and D3 sections which corresponds (337) with tetramer(T) [T(337)], (225), (337) with dimer/adatom(D/A) row [D (337)], respectively. The one unit of Si(5 5 12) is reconstructed to D3D2-D1 subsections along the [6 6 −5] direction perpendicular to the 1D rows. The boundary of two adjacent sections except for that between D2 and D1 sections is formed with H chain, while the boundary between D2 and D1 sections is formed with π chain. Fig. 1(a) shows the atomic resolution image of the reconstructed Si (5 5 12) surface. The white dotted lines shown on Fig. 1 indicate the one period of Si(5 5 12). Atomic resolution image shown in Fig. 1(a) allows us to distinguish that the H and π chains have a 1 × cycle and the remaining structural units such as D, A and T rows are formed in 2 × cycle along the [−1 1 0] direction. A bright dot appeared in D2 and D3 is a feature of Si-addimer adsorbed between dimer (D) and adatom (A) row. A stable reconstructed Si(5 5 12)-2 × 1 has a large terrace with a period of (5 5 12). Occasionally Si(7 7 17) terrace with a period of (7 7 17) which is composed of T(337) section, i.e., D1, and one (225) section, i.e., D2, is observed on the reconstructed Si(5 5 12) surface under local compressive stress [30]. Also, the mixed periods of (5 5 12) and (7 7 17) appeared in the clean reconstructed Si(5 5 12) terrace. The “a” and “b” presented in all figures indicate the periods of (5 5 12) and (7 7 17), respectively. In order to compare the adsorption sites of Ag atoms deposited on Si (5 5 12) surface before and after the heating, the images obtained immediately after the deposition of Ag less than 0.1 ML and the image scanned after post-annealing at ~500 °C are shown in Fig. 1(b)–(d) and Fig. 1(f), respectively. In case of the Ag deposited on Si(111) surface, the thermal desorption temperature of Ag atoms was reported about 560 °C [31,32]. Our samples were post-annealed at lower temperature than the desorption temperature to avoid the desorption of Ag weakly bound on substrate. Fig. 1(b) and (c)–(d) are filled-state images obtained after 0.03 ML and 0.05 ML Ag deposition at room temperature, respectively. The collapse of π chain, which forms the boundary between D1 and D2, is 2
Surface Science 693 (2020) 121547
S. Cho, et al.
Fig. 1. STM images obtained from a clean reconstructed Si(5 5 12)-2 × 1[(a)] and Ag/Si(5 5 12) before[(b): Θ = 0.03 M L, (c) and (d): Θ = 0.05 ML] and after postannealing[(f): Θ = 0.05 ML, post-annealed at 440 °C], respectively. (a): 10.5 nm × 10.5 nm, Vs = −1.9 V, I = 0.5 nA, (b) and (c): 30 nm × 30 nm, Vs = −2.7 V, I = 0.5 nA, (d): 10.5 nm × 10.5 nm, Vs = −2.7 V, I = 0.5 nA). (e) The structure model of reconstructed Si(5 5 12)-2 × 1 surface. (f) The empty-state image taken after post-annealed at 500 °C after Ag(0.01 ML) deposited on Si(5 5 12). (Size; 30 nm × 30 nm, Vs = +2.5 V, I = 0.5 nA). (g) The line profiling results of [A] and [B] lines shown in Fig. 1(f).
form the Ag wires between the π chain and the T row inside D1 section along the [−1 1 0] direction after heat treatment. The preferential adsorption site after post-annealing is most similar to the adsorption site “3” among the various adsorption sites before annealing. Fig. 2 shows a large image of 0.2 ML Ag-adsorbed Si(5 5 12) surface after heat treatment at 440 °C. Fig. 2(a) shows that the upper-left area written by “B” and lower right-hand area written by “A” have different periods. The period of bright row of B area is shorter than that of A area perpendicular to the row direction, [6 6 −5] direction, and can be seen as a shorter aspect ratio along the row direction, [−1 1 0]. More
rows inside D2 and D3 sections still appeared and its number has increased. The line profiles perpendicular to the row direction are shown in Fig. 1(g). Line A is the line profiling across the 1-D Ag row. The structure of substrate can be identified through the line profiling of line B. This means that the formation position of the Ag row is known. From the line-profile of line A, it has been shown that the brightest row is formed inside D1 section, and the height of the π chain between D2 and D1 section is higher than that of other 1 × chains. As shown from the Fig. 2, the π chain is firstly collapsed by Ag atom, and the chain has been transformed to a different kinds of structure. Ultimately, Ag atoms 3
Surface Science 693 (2020) 121547
S. Cho, et al.
Fig. 2. Filled-state large images obtained after post-annealing at 440 °C on 0.2 ML Ag adsorbed Si(5 5 12) surface. (b) and (c) obtained from the terrace of each area written by “A” on the upper-left area and “B” on the lower-right terrace of (a). (a) Size; 0.3 μm × 0.3 μm, Vs = −2.7 V, (b) 40 nm × 40 nm, Vs = −2.7 V. (c) 40 nm × 40 nm, Vs = −1.9 V.
1 × cycles. Based on the 1 × chain structure with relatively dark contrasts that survives in several places, the period (7 7 17) can be estimated. Line profiling was performed to compare the height and relative distance of the structure represented by the Ag adsorption shown in Fig. 4(b). The line [A], marked by a red line in the image, is the result of line profiling in a direction perpendicular to the row shown below the image. One period of (7 7 17) in the image was matched with one period of (7 7 17) in line profiling. By comparing the height difference in the line profiling results perpendicular to the 1-D structural row direction, one can see whether the 1 × chain row is on a clean substrate or on an Ag-deposited surface. The profiling of the line [B] in the same direction was presented in Fig. 4(d) to determine the relative position of the row in the uppermost layer of Ag. One period of Si(7 7 17) is represented by a white dotted line in the image. As a result of the line [A] analysis, the initial absorption of the Ag atom is indicated by a row forming 1 × cycle on the T row in D2 section. The ``α'' and ``γ'' marked in the line profiling analysis shown in Fig. 4(b) are the locations of adsorption site corresponding to “2” or “1” adsorption sites, shown in Fig 1(d) and (e), respectively. Another row with 1 × period appears at the position of the π chain row that forms the boundary between D1 and D2 sections. The two bright rows formed about 0.7 nm apart. The height of γ site in this double Ag row is higher than that of α site. The distance of double row in the same height holds the period of the substrate. The period of double Ag row is indicated by a black dotted line in the image. The dot with 2 × cycle was found around the bright double row. As a result of Ag adsorbed on Si(7 7 17) terrace, it is difficult to identify the shape of the structural unit such as H chain, D/A row, T row and π chain that originally forms Si(7 7 17). There is a new double row feature that looks brightest on top of the first double Ag row formed. The result of the line profiling of [B] is displayed at the bottom. The relative height difference between the first and second double Ag rows is about less than 0.05 nm. One period of double Ag row is marked by a thick red dotted line in the image. The second layer of double Ag row is located between the first layer of double Ag row and is formed the shift with about 0.44 nm toward [6 6 −5] in the perpendicular to the row direction. The distance between these double Ag rows is also 0.66 nm, which is consistent with the result of the first layer of double Ag row. In line [B], the absorption locations of the second layer of double Ag row are marked with α′ and γ′. The height of γ′ site also is higher than that of α′ site. In terms of the periodicity of the second layer along the row direction, four bright spots appeared as the shape of the teeth was paired with each other and has a 3 × period along the row direction presented a pair with a square of yellow dotted lines. The contrast of two bright spots on the left side of the pair is slightly brighter than that of the two bright spots on the
detailed STM images obtained from each terrace are shown in Fig. 2(b) and (c) respectively. Fig. 2(a) shows that each area shown as A and B is terrace with (5 5 12) period and terrace with (7 7 17) period, respectively. In the Agadsorbed on Si(5 5 12) terrace shown in Fig. 2(b), only the H chain structure that forms the boundary between D2 and D3 sections survives and maintains the period (5 5 12) as the boundary. Ag atoms adsorbed on the site between D1 and D2 sections and form the 1D-like row along the [−1 1 0] direction. Meanwhile, the area “B” marked in Fig. 2(a) is the Ag-adsorbed Si(7 7 17) terrace consisting of D1 + +D2 as shown in Fig. 2(c). This is not a transition from (5 5 12) period to (7 7 17) period as a result of Ag adsorption. The absorption results of Ag over this Si(7 7 17) area show the differences from the characteristics of the feature of Ag adsorbed over Si(5 5 12) terrace. We can see the break to a specific cycle along the row direction and the relatively short Ag rows, unlike the case of Ag adsorbed on Si(5 5 12). It is also difficult to identify all of 1 × chain structures in this area. In order to understand the absorption process of Ag in each region, the analysis of each area are shown in Figs. 3 and 4 respectively. The filled-state STM image of Fig. 3(a) is obtained from 0.2ML Ag-adsorbed on Si(5 5 12) surface after post-annealed at 440 °C. The H chain row that is boundary between D2 and D3 sections only survives. That is indicated by H with a white dotted line. Protrusions caused by Si-addimer appear as bright spots on the left and right sides of this H chain. The red bar represents a period of (5 5 12). The boundary located between D1 and D2 sections, i.e., π chain, is no longer identified by the adsorption of the Ag atoms, and 1D-like Ag rows are formed on T row inside D1 section. The result of line profiling perpendicular to a one-dimensional row is shown in Fig. 3(b) below the image. At the bottom, a structural model was placed as shown in Fig. 3(c), consistent with one cycle of (5 5 12) in the STM image. The highest bright row is formed around the boundary of D1 and D2. The “α” and “β” marked in the line profiling analysis are the locations of adsorption site corresponding to the “2” or “3” adsorption sites shown in Fig 1(d) and (e), respectively. That is, the Ag atoms are preferentially adsorbed on the π chain and break the structure of the weak π chain row with 1 × period due to the additional adsorbed Ag atoms, and the top layer of Ag is located between T row inside D1 section and H chain row located between D1 and D3 sections and form the 1-D Ag rows with a high aspect ratio. In the row direction, the period of Ag row shows a relatively regular periodicity of three times. The period between the brightest Ag rows still follows the period of the (5 5 12) substrate in the direction perpendicular to the row. The filled-state image obtained after post-annealed at 440 °C of 0.2ML Ag/Si(7 7 17) system is shown in Fig. 4(a). Unlike the Ag-adsorbed Si (5 5 12) surface, it is difficult to find a chain structure with 4
Surface Science 693 (2020) 121547
S. Cho, et al.
Fig. 3. (a) Image taken from the area which (5 5 12) period is sustained of Fig. 2(b). Size; 17.4 nm × 17.4 nm, Vs = −2.7 V, I = 0.5 nA. (b) The cross-sectional lineprofiling of image is displayed below image of (a). (c) The structure model of reconstructed (5 5 12) period.
the Ag row, which forms the brightest band. Due to the thermal drift and thermal broadening, it is not easy to distinguish the data from the Ag row formed with a small amount of Ag and the row of the exposed substrate. Therefore, the coverage of Ag was increased to widen the adsorbed Ag area. It is difficult to identify the H chain or π chain that can estimate the period of the substrate as shown in Fig. 5(a). However, it can be observed that a bright region appears due to the adsorption of Ag and a dark region where the substrate is exposed. The results of the STS in the bright and dark areas show distinctly different results. This is the result of repeated experiments of STS in the other scan areas of the same sample, that is, in the dark area and the bright areas. The STS results with blue and red colors were measured on the highest Ag row and that with green color was measured on the lowest dark area. The results are shown in Fig. 5(b). On Ag row, it was observed to metallic character with a small band gap but the semiconducting behavior is appeared at the darkest area that is predicted to bare Si substrates. The results of electronic characteristics of Ag nanostructure formed by increased coverage of Ag are still insufficient. It is expected that it will be helpful to understand the electronic properties of 1-D Ag nanostructure through comparison of the reported results.
right. Unlike the second double Ag row, which breaks into a three-fold periodicity in the direction of row, there is also a double Ag row that creates 1 × -cycle along the row direction. Support of the theoretical calculation will be needed for more detailed adsorption analysis in Agadsorbed Si(7 7 17) system. Baski's group had reported the results for the Ag/Si(5 5 12) system using RHEED, STM/S and PES. For low coverages (θ < 0.25 ML), Ag atoms form very long monoatomic chains on top of the T rows and make a very weak 3 × periodicity along the row direction which agrees with our results [3,4,9]. We performed the STS spectrum at the Ag-adsorbed Si(5 5 12) surface at room temperature. Fig. 5(a) shows the filled-state STM image of Ag-deposited on Si(5 5 12) and sequentially annealed at 450 °C. The H chain rows of the substrate are rarely visible, despite a small amount of Ag adsorption of 0.3 ML. It also shows a sawtooth-like feature as the width of the Ag row is widened in the [6 6 −5] direction, perpendicular to the row. As the height of the Ag row increases, the slope opposite to the direction of gentle slope becomes steep. So it appears as dark contrast in the image. The line profiling results are shown in the image wing to see the period between Ag rows of the same height. It can be seen that the period of the H chain of the substrate is the region of the substrate where the periods of (7 7 17) and (5 5 12) denoted by “a” and “b” are mixed, respectively. STS was performed to understand the electrical characteristics of 5
Surface Science 693 (2020) 121547
S. Cho, et al.
Fig. 4. (a) Image taken from the area which (7 7 17) period is sustained of Fig. 2(c). Size; 10 nm × 10 nm, Vs = −1.6 V, I = 0.5 nA. (b) and (d) The cross-sectional line-profiling of lines of [A] and [B] are displayed. (c) Structure model of reconstructed Si(7 7 17) period.
4. Conclusions
with two pairs of different brightness. The cycle between the double Ag rows also keeps the (7 7 17) period of the substrate. These results imply that the adsorption of Ag atoms takes preferentially over the T rows with high dangling bond density inside the D1 and D2 sections among the structural units forming the structure of the substrate and forms a 1-D Ag nanostructure. The formed 1-D Ag row keeps the identical periodicity of the substrate and shows a metallic characteristics.
The results obtained from Ag adsorbed on Si(5 5 12) terrace are quite matching with those reported previously. However, the Ag-adsorbed Si(7 7 17) surface shows a different aspect. Unlike the results of Ag/Si(5 5 12), 1 × chains of Ag/Si(7 7 17) are difficult to identify. Firstly, the 1 × structure is collapsed by Ag adsorption. The Ag atoms initially adsorbed on top of T row in D2 section and on the π chain, which implies that they form a double Ag row with (7 7 17) period. The second Ag row between the first double Ag rows is formed, and another second Ag row is formed on the T row in D1 section. These second Ag rows also form the double feature appear to be the brightest rows. Similar to the results of Ag/Si(5 5 12), the cycle in the direction perpendicular to the row is three times. Four bright spots are clustered
Declaration of Competing Interest None.
6
Surface Science 693 (2020) 121547
S. Cho, et al.
Fig. 5. STS data at RT from 0.3 ML Ag deposited on Si(5 5 12) and sequentially annealed at 450 °C. (a) filled-state STM image (Size; 40 nm × 40 nm, Vs = −2.6 V, I = 0.7 nA). Line profiling is shown under the image. (b) STS data obtained from the dark area (circle filled with green color), bright area (circles filled with red and blue color) in image (a), respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Acknowledgments
[15] M. Kumar, V.K. Paliwal, A.G. Joshi, Govind S.M. Shivaprasad, Surf. Sci. 596 (2005) 206. [16] S. Cho, J.M. Seo, Surf. Sci. 565 (2004) 14. [17] P. Kumar, M. Kumar, B.R. Mehta, S.M. Shivaprasad, Surf. Sci. 605 (2011) 1426. [18] P. Kumar, M. Kumar, S.M. Shivaprasad, Appl. Surf. Sci. 282 (2013) 348. [19] J.R. Hahn, H. Jeong, S. Jeong, J. Chem. Phys. 123 (2005) 244702. [20] J.R. Hahn, H. Jeong, S. Jeong, S.H. Jang, Jpn. J. Appl. Phys. 45 (2006) 2175. [21] H. Jeong, S. Jeong, S.H. Jang, J.M. Seo, J.R. Hahn, J. Phys. Chem. B 110 (2006) 15912. [22] S.H. Jang, S. Jeong, J.R. Hahn, J. Chem. Phys 130 (2009) 234703. [23] J.R. Hahn, S.C. Bharath, G. Kim, S. Jeong, T. Pearl, J. Phys. Chem C115 (2011) 17111. [24] J.R. Hahn, S.C. Bharath, G. Kim, K.W. Kim, S. Jeong, T. Pearl, J. Phys. Chem. C117 (2013) 11197. [25] J.R. Hahn, G. Kim, K.W. Kim, S. Jeong, Surf. Sci. 616 (2013) 166. [26] S. Jeong, H. Jeong, S. Cho, J.M. Seo, Surf. Sci. 557 (2004) 183. [27] H. Kim, H. Li, Y. Zhu, J.R. Hahn, J.M. Seo, Surf. Sci. 601 (2007) 1831. [28] W. Ranke, Y.R. Xing, Surf. Sci. 381 (1997) 1. [29] J. Liu, M. Takeguchi, H. Yasuda, K. Furuya, J. Cryst, Growth 237 (2002) 188. [30] S. Cho, J.M. Seo, J. Korean Phys. Soc. 49 (2006) 181. [31] S. Hasegawa, H. Daimon, S. Ino, Surf. Sci. 186 (1987) 138. [32] K.J. Wan, X.F. Lin, J. Nogami, Phys. Rev. B 47 (1993) 13700. [33] M. Sakurai, C. Thirstrup, M. Aono, Phys. Rev. B 62 (2000) 16167. [34] C. Zhang, G. Chen, K. Wang, H. Yang, T. Su, C.T. Chan, M.M. Loy, X. Xiao, Phys. Rev. Lett 94 (2005) 176104. [35] H. Kim, H. Li, J.M. Seo, J. Vac, Sci. Technol. B25 (2007) 1511.
This work was supported by the basic R&D program of the Korean Research Institute of Standards and Science. References [1] A. Baski, S.C. E.rwin, L.J. Whitman, Science 269 (1995) 1556. [2] H.H. Song, K.M. Jones, A.A. Baski, J. Vac. Sci. Technol. A 17 (1999) 1696. [3] S.R. Blankenship, H.H. Song, A.A. Baski, J.A. Carlisle, J. Vac. Sci. Technol. A 17 (1999) 1615. [4] K.M. Jones, H.H. Song, A.A. Baski, J. Clust. Sci. 10 (1999) 573. [5] A. Baski, K.M. Saoud, J. Clust. Sci. 12 (2001) 527. [6] Y. Peng, H. Minoda, Y. Tanishiro, K. Yagi, Surf. Sci. 493 (2001) 508. [7] A. Baski, K.M. Saoud, K.M. Jones, Appl. Surf. Sci. 182 (2001) 216. [8] J.W. Dickinson, J.C. Moore, A.A. Baski, Surf. Sci. 561 (2004) 193. [9] A. Baski, K.M. Jones, K.M. Saoud, Ultramicroscopy 86 (2001) 23. [10] J.R. Ahn, Y.J. Kim, H.S. Lee, C.C. Hwang, B.S. Kim, H.W. Yeom, Phys. Rev. B66 (2002) 153403. [11] J.R. Ahn, H.W. Yeom, E.S. Cho, C.Y. Park, Phys. Rev. B69 (2004) 233311. [12] H. Kim, Huiting Li, J.M. Seo, Surf. Sci. 602 (2008) 2563. [13] H. Kim, Y. Cho, J.M. Seo, Surf. Sci. 583 (2005) 265. [14] Huiting Li, Y.-N. Xu, Y.-Z. Zhu, Hidong Kim, Jae M. Seo, Phys. Rev. B 75 (2007) 235442.
7