Growth and Characterization of One-Dimensional SnTe Crystals within

ARTICLE pubs.acs.org/JPCC

Growth and Characterization of One-Dimensional SnTe Crystals within the Single-Walled Carbon Nanotube Channels L. V. Yashina,*,†,‡ A. A. Eliseev,§ M. V. Kharlamova,§ A. A. Volykhov,‡,§ A. V. Egorov,† S. V. Savilov,† A. V. Lukashin,§ R. P€uttner,|| and A. I. Belogorokhov‡ †

Department of Chemistry, M.V. Lomonosov Moscow State University, 1-3 Leninskie gory 119991 Moscow, Russian Federation GIREDMET OJSC, 5-1 B. Tolmachevsky 119017 Moscow, Russian Federation § Department of Materials Science, M.V. Lomonosov Moscow State University, 1-73 Leninskie gory 119991 Moscow, Russian Federation Institut f€ur Experimentalphysik, Freie Universit€at Berlin, 14 Arnimallee D-14195 Berlin-Dahlem, Germany

)

‡

ABSTRACT: Quasi-free-standing one-dimensional (1D) SnTe crystals were grown within the channels of single-walled carbon nanotubes (SWCNTs) from the melt using the capillary technique. The crystal structure of the 1D SnTe crystal was determined from transmission electron microscopy measurements and density functional theory calculations. The obtained structure for the 1D crystal—(SnTe)5n—shows in comparison to the bulk structure a significant bond length relaxation (contraction) along the SWCNT axis, which is accompanied by a decrease of the Sn and Te effective charges. In contrast to bulk rocksalt SnTe and 1D (SnTe)4n crystal as well, the (SnTe)5n crystal shows the gapless band structure due to the contribution of orbitals mostly composed by three-coordinated atoms. X-ray photoelectron spectroscopy and Raman spectroscopy data indicate that there are no noticeable interactions between the 1D SnTe crystal and the SWCNTs other than a minor donor influence on the electron structure of metallic SWCNTs; this influence is in line with gapless state of (SnTe)5n.

’ INTRODUCTION One-dimensional (1D) crystals of compound semiconductors with a diameter of 3-5 atoms currently attract attention due to their unique properties, in particular size-dependent quantum effects such as van Hove singularities, etc.1,2 From the fundamental perspective, it is also of importance to understand the interplay between the electronic and atomic structure of the 1Dcrystals and the relationship between the structural parameters of a 1D-crystal and the bulk. To a certain extent, the formation of a 1D-crystal resembles a surface formation; the latter topic has been thoroughly investigated for many compound semiconductors. This surface formation has been described in terms of surface relaxation and reconstruction (i.e., superstructure formation). One may suppose that a similar approach is—at least to a certain extent—also applicable to 1D crystals. Unfortunately, in practice it is difficult to prepare and study 1D-crystals in its free state. Up to now, vicinal surfaces or different templates are routinely used to grow stable 1Dcrystals;2,3 however, in these cases the crystal-substrate/template interaction always has to be taken into account. This interaction may include the formation of local chemical bonds, a substrate-induced distortion of the crystal structure and bond geometry, and nonlocal effects (i.e., charge transfer, etc.). Moreover, an asymmetry in the interaction between the 1D crystal and the substrate may play a crucial role and can, consequently, govern the electronic properties of such system. Therefore, it is r 2011 American Chemical Society

essential for a better understanding of the 1D crystal physics to find inert substrates or templates that allow a minimization of the crystal-template/substrate interaction. Within the past decade it has been proposed to use single-walled carbon nanotubes (SWCNTs) as a template for growing 1D crystals since these nanotubes are known to be chemically inert toward most inorganic substances; under certain conditions, it is also possible to avoid a charge transfer between the crystal and the nanotube wall.4-7 Without a formation of local chemical bonds between the 1D crystal and the nanotube the charge transfer can be predicted within the rigid bands model, which describes the interaction in terms of a “doping” effect with the corresponding increase (n-doping) or decrease (p-doping) of the SWCNT Fermi level.8,9 The doping level can be determined experimentally using optical spectroscopy, X-ray photoelectron spectroscopy (XPS), ultraviolet photoelectron spectroscopy (UPS), Raman, or X-ray absorption near-edge spectroscopy (XANES), as well as theoretically by ab initio quantum chemical calculations.10 However, if the work functions for the SWCNT and the 1D crystal are similar and meet the requirements that have been recently developed for graphene,11 a Fermi-level shift is absent. Tin telluride (SnTe), an A4B6 semiconducting material, complies Received: November 9, 2010 Revised: January 6, 2011 Published: February 15, 2011 3578

dx.doi.org/10.1021/jp1107087 | J. Phys. Chem. C 2011, 115, 3578–3586

The Journal of Physical Chemistry C

Figure 1. Raman spectra measured at the excitation energies (a) 2.41 eV, (b) 1.98 eV, and (c) 1.58 eV. Shown are the RBM band and the G band.

with the aforementioned requirements and is, therefore, a model case for studying the 1D-crystal formation effect. For A4B6 crystals with rocksalt structure (PbS, PbSe, PbTe, SnTe), the formation of the (100) surfaces is accompanied by a differential interlayer relaxation due to the ionic nature of the chemical bonds.12,13 These findings suggest to expect a relaxation effect also for the interatomic distances in 1D SnTe crystals. In this paper, we describe the preparation, the atomic structure and the crystal-template interactions for SnTe@SWCNT nanocomposite based on transmission electron microscopy (TEM), XPS, and Raman spectroscopy measurements.

’ EXPERIMENTAL METHODS SWCNTs were obtained by the catalytic arc-discharge method using graphite rods of 0.8 cm diameter and a Y/Ni powder catalyst using a helium pressure 73.3 kPa and currents of 100110 A. Nanotubes were purified by a multistage procedure

ARTICLE

consisting of oxygenation in air and rinsing with HCl to remove the catalyst. To determine the diameter of pristine carbon nanotubes, a Raman spectrum was acquired at the excitation wavelength of λex = 514 nm (2.41 eV), see Figure.1. The diameters d of the SWCNTs were deduced from the Raman shifts caused by the radial breathing mode (RBM) using the relation ϖRBM = A/dþB, and by assuming that A = 232 nm cm-1 and B = 6.5 cm-1, as proposed in another work14 for bundled SWCNTs. The resulting diameters are d = 1.34 ( 0.05, 1.42 ( 0.05, and 1.51 ( 0.05 nm. Our complementary TEM observations confirm the calculated values, with the majority of nanotubes having the smallest value of d = 1.34 ( 0.05 nm. Each sample represents a mixture of semiconductor (s-SWCNT) and metallic (m-SWCNT) nanotubes approximately in equal proportions; their detailed characterization is given elsewhere.6 The purified samples of SWCNTs were preopened by heat treatment at 500 °C in dry air for half an hour. The SWCNTs (0.025 g) were subsequently ground together with polycrystalline SnTe (6N purity) at a molar ratio of 1:2 using an agate mortar and pestle and then evacuated in quartz ampules at 1 Pa for 1 h and sealed. The ampules were heated at a rate of 1 K/min up to a temperature of 100 K above the melting point of SnTe. This temperature was maintained for 6 h, whereupon the samples were slowly cooled down at a rate of approximately 0.02 K/min. Such a slow cooling provides better conditions for the crystallization the SnTe inside the nanotube channels. Both pristine and filled SWCNTs were studied by HRTEM, Raman spectroscopy, and XPS. Raman spectra were acquired on a Renishaw InVia Raman microscope equipped with a 20-mW 514-nm argon laser, a 17-mW 633-nm HeNe laser, and a 300-mW 785-nm NIR diode laser, variable power ND filters (power range 0.00005-100%), and near-excitation tunable (NeXT) filters. The laser spot size varied from 1 to 300 μm. The energy positions of all resonance lines in the Raman spectra were determined by least-squares fitting of the experimental data with sets of Gaussian/Lorentzian convolution functions using the WiRE 3.0 software. For the electron microscopy studies, the samples were prepared by dispersing the nanotubes onto a holey carbon-coated copper grid. HRTEM was performed on a Cscorrected JEOL 2100 microscope at 200 kV. The X-ray photoelectron spectra were acquired with a Kratos Ultra DLD spectrometer using a monochromatic Al KR X-ray source that possesses an analysis area of 300 μm 700 μm. The spectra were recorded in a constant analyzer pass energy mode of 5 eV resulting in a resolution better than 0.3 eV. The energy scale was calibrated with an accuracy of (0.03 eV using the binding energies of core levels Au 4f5/2, Cu 2p3/2, and Ag 3d5/2 of 83.96, 932.62, and 368.21 eV, respectively. Different parts of the samples were analyzed to check the reproducibility of the results. For the data analysis, the spectra were fitted by the Gaussian/Lorentzian convolution functions with simultaneous optimization of the background parameters. The C 1s line asymmetry was described with a Doniach-Sunjic function using R = 0.1004. To model the structure of the 1D SnTe crystal, we used both the cluster approach and periodic boundary conditions (PBC). The hybrid density functional B3LYP method15-17 was applied using the LanL2 pseudopotential17,18 and an appropriate doubleζ basis set augmented by d-type polarization functions.19 The calculations were performed using the GAUSSIAN03 and GAUSSIAN09 program package.17,20 3579

dx.doi.org/10.1021/jp1107087 |J. Phys. Chem. C 2011, 115, 3578–3586


ARTICLE

Table 1. Experimental and Calculated Sn-Te Bond Lengths for Different Models under Consideration experimental, Å free SnTe

2.52

molecule 1D SnTe crystal

2.8 ( 0.1

calculated, Å

assignment

2.55 3.36 - (SnTe)4n

longitudinal

3.01 - (SnTe)5n 3.3 ( 0.2

2.96 - (SnTe)4n

perpendicular

3.14 - (SnTe)5n SnTe (001)

2.97-3.01

surface bulk SnTe

Figure 2. Upper part: Optimized geometry for the two possible structures of the 1D SnTe crystal. Middle part: cross section of the two structures in a SWCNT with a diameter of 1.34 nm. Lower part: the iso-electron density contour plot for an electron density of 0.005 (a.u.)-3 in order to estimate the perpendicular dimension of the crystal structure.

first-to-second layer distance

3.16

3.12

3. RESULTS AND DISCUSSION

Figure 3. TEM image of the SnTe@SWCNT nanocomposite: (a) general view of a filled SWCNT bundle, (b) bundle with SWCNTs ends (perpendicular view), (c) fragment of 2 SWCNTs filled with (SnTe)5n recorded at a higher magnification value. In this image the derived structural parameters are indicated.

3.1. Atomic Structure of SnTe@SWCNTs. It is a matter of common knowledge that SnTe crystallizes in a cubic rocksalt structure at temperatures above 100 K. On the basis of this information and taking into account the size of the cavity (1.06 nm) of a SWCNT with a diameter of 1.34 nm, two different structures can be assumed for the 1D crystal. They are shown in Figure 2. Both structures can be derived from the SnTe bulk structure by cutting it along the [100] direction. The first structure (structure I) possesses 4 atoms in the cross-section and 8 atoms in a unit cell and is also labeled (SnTe)4n; the second one (structure II) has 5 atoms in cross-section and 10 atoms in a unit cell and is also labeled (SnTe)5n. Periodic boundary conditions and different slab sizes were used for the modeling of both structures. In these calculations all atomic positions were fully optimized. Because of the available computation power, the SnTe@SWCNT nanocomposite was modeled within the cluster approach only. For this reason, (SnTe)m clusters with m = 20 and 40 were considered to check whether the results obtained within the cluster approach are identical to the results of the calculations using periodic boundary conditions. For the clusters the periodicity and the cross-section interatomic distances were optimized. To estimate the 1D crystal-SWCNT interaction, the cluster was inserted into a (10,10) SWCNT, which consisted of 120 carbon atoms. To decrease the influence of the dangling bonds at the ends of the SWCNT, the cluster modeling the nanotubes was terminated with hydrogen atoms.

For comparison, the bulk SnTe crystal was also modeled using a (SnTe)56 cluster as described elsewhere.21 The (SnTe)56 cluster geometry allowed us to reproduce the bulk interatomic distances with an accuracy of 0.04 Å (which is comparable with the accuracy obtained for the SnTe molecule, see Table 1). The SnTe (001) surface was modeled as a cluster and in slab geometry with PBC. A cluster (SnTe)34 having the pyramidal shape (25/25/9/9) was used, which was found to be the smallest cluster size that shows an influence of boundary effects at a reasonably low level.22 The atomic positions in the central part of the cluster were fully optimized, while all other atoms were fixed to the experimentally known bulk positions. The slab was composed by three atomic layers; the atomic positions were optimized in the direction perpendicular to the surface. The (100) surfaces for compounds with a rocksalt structure experience surface relaxation due to a lower coordination number of the surface atoms, which includes a first-to-second interlayer distance decrease, a second-to-third interlayer distance increase, and further periodic oscillation of the interlayer distances that diminishes with the depth.12,13 In the case of the SnTe (100) surface, our calculations performed with PBC result in a significant decrease of the first-to-second interlayer distance of approximately 0.15-0.19 Å, see Table 1; within the cluster approach the degree of relaxation is slightly higher. The results of the calculations for the 1D crystals are also summarized in Table 1. For structure I the interatomic distances 3580



ARTICLE

Figure 5. The plasmon structures in the C 1s spectra of pristine SWCNTs and those filled with SnTe. The intensities are normalized to the peak heights of the C 1s main line.

Table 2. Summary of the C 1s Spectral Features of SWCNT and SnTe@SWCNTa plasmon loss values, eV (relative intensity), BE,eV

Γ, eV

ΔE, eV

SWCNT

284.38

0.313

0

6.14 (8.1%), 10.18 (0.34%)

SnTe@SWCNT

284.41

0.357

0.03

6.10 (5.7%), 9.96(0.28%)

ΔEPlasmon, eV (Irel)

Given are the binding energies, BE, the Gaussian widths (fwhm), Γ, the energy shift ΔE, as well as the plasmon loss energies, ΔEPlasmon, and in brackets the corresponding relative intensities, Irel, normalized to the intensity of the C 1s main line. a

Table 3. Mulliken Effective Charges for the Sn and Te Atoms for Different Models under Consideration Sn

Figure 4. C 1s photoelectron spectra: (a) comparison of the spectra obtained for pristine SWCNTs and those filled with SnTe; (b and c) the curve-fitting results obtained within the concept of different peak positions for s- and m-SWCNTs for pristine and filled SWCNTs, respectively. The solid lines through the data points represent the fit results, and the subspectra reflect the contributions of different spectral features.

along the nanotube axis increase by 0.24 Å as compared to the bulk value. This is accompanied by a decrease of the perpendicular distances along the 1D-crystal by 0.16 Å; the latter distances are, however, still essentially larger than those in a free molecule. The bond angles change slightly, i.e., by 3.4° within a crosssectional plane and by 1.5° in the longitudinal direction. For structure II, an opposite behavior of the bond lengths is predicted by the calculations, namely, a decrease of the interatomic distances along the nanotube axis of 0.11 Å. The perpendicular distances are almost identical to those of the bulk and the bond angles remain unchanged within the accuracy of (0.7°. Thus we

Te

bulk SnTe (cluster)

0.52

-0.49

bulk SnTe (PBC)

0.52

-0.52

SnTe (100) surface (cluster)

0.48

-0.41

1D (SnTe)5n, crystal inner position, CN = 6

0.45

-0.40

1D (SnTe)5n, crystal outer position, CN = 3

0.36

-0.37

1D (SnTe)4n, CN = 4

0.41

-0.41

SnTe molecule

0.25

-0.25

can assume that the relaxation is essentially lower for the (SnTe)5n crystal than for the (SnTe)4n crystal. This can be easily explained by considering the coordination numbers for the corresponding atoms. In the (SnTe)4n structure all atoms have undercoordination with only 4 atoms in the nearest environment in comparison to 6 nearest neighbors in the bulk. In structure II, the central atoms have the same coordination as in the bulk and, therefore, each bond in perpendicular direction is formed by one 6-coordinated and one 3-coordinated atom. It should be noted that for both structures these results do not change with an increase of the slab size, i.e., there is no superstructure formation. 3581



ARTICLE

Table 4. Summary of the Observed Raman Shifts Recorded at Different Excitation Energiesa G band, cm-1 RBM band, cm-1

excitation energy, eV SWCNT

2.41

SnTe@ SWCNT SWCNT

1.96

SnTe@ SWCNT SWCNT SnTe@ SWCNT a

1.58

G- m

G- s

Gþ

158

170

179

1555

1568

1591

157 (-1)

169 (-1)

179

1557 (þ2)

1571 (þ3)

1592 (þ1)

154

169

195

153 (-1)

166 (-3)

1543

1563

1591

1548 (þ5)

1569 (þ6)

1588 (-3)

159

171

1552

1570

1593

157 (-2)

173 (þ2)

1555 (þ3)

1572 (þ2)

1596 (þ3)

The numbers in bracket give the changes of the Raman shifts for SnTe@ SWCNT as compared to those for pristine SWCNTs.

Figure 6. Optical absorption spectra of the pristine SWNTs and SnTe@SWNT composite.

Moreover, within the accuracy of our calculations we did not observe any pronounced influence of the (10,10) SWCNT on the optimized geometry of the 1D crystals. To see how the two proposed SnTe structures fit into a channel of a SWCNT with a diameter of 1.06 nm we calculated iso-electron-density contours for different electron densities. Calculations for the iso-electron-density contours of 0.005 (a.u.)-3 show that the (SnTe)5n crystal diameter is 1.03 nm. For comparison, the electron density in the center of the Sn-Te distance is equal to 0.03-0.04 (a.u.)-3. The radial dimensions of the (SnTe)4n crystal using the same electron density contour of 0.005 (a.u.)-3 are found to be essentially smaller, namely 0.82 nm. For this structure the value of 1.03 nm can be reached for the iso-electron-density contours of 0.0002 (a.u.)-3. From this data it can be concluded that the interaction between (SnTe)4n crystal and (10,10) SWCNT is about 20 times weaker than between (SnTe)5n crystal and (10,10) SWCNT. Figure 3 shows a TEM image of the SnTe@SWCNT nanocomposite. The low-magnification image (Figure 3a) provides a certain evidence for the intercalation of SnTe into the nanotube channels. Moreover, a rectangular patterned arrangement of the contrast elements within the SWCNTs indicates that SnTe crystallizes within the channels. However, such an arrangement can—in principle—correspond to both structures (structure I or II) described above. The experimental structural data derived from the TEM images are indicated in the high-magnification

image shown in Figure 3c and are summarized in Table 1. The TEM images show that there are 3 atoms along the diameter direction, which fits structure II rather well. Another evidence in favor of this structure is the measured 1D crystal period, which is essentially lower than that in the bulk structure. As described above, we predicted such behavior for structure II based on our theoretical calculations. One should take into account that the relatively large differences between the unit cell parameters of 1D SnTe and those of the bulk crystal do not essentially influence the specific volume, i.e., the value for the 1D crystal differs from the value for the bulk by about 2% (which is practically the same value taking the accuracy into account), whereas the Sn-Te bond length relaxation is 10-12% for the experimental and 5-6% for both calculated structures. 3.2. Interaction of 1D SnTe with SWCNTs. The C 1s XPS spectra for SnTe@SWCNT and pristine SWCNTs displayed in Figures 4 and 5 show no significant differences. There is, however, a slight shift of 0.03 eV to higher binding energies and also a slight peak broadening of 0.045 eV for SnTe@SWCNT. The observed4 shift between pristine SWCNTs and SnTe@SWCNT is only slightly above the limits for accuracy and reproducibility of the present measurements, thus it must be interpreted with care. The parameters of the spectral features of the C 1s spectra are summarized in Table 2. A detailed analysis of the C 1s spectra can be carried out by taking into account a difference in binding energies for metallic and semiconductor SWCNTs as it was observed in a recent work by Ayala et al.;23 these group of authors compared the spectra obtained for pure metallic and pure semiconductor tubes and reported for m-SWCNTs an approximately 0.1 eV lower binding energy. In addition, they observed a distinct asymmetry in the C 1s peak shape from m-SWCNTs, whereas a comparable asymmetry in the peak shape of s-SWCNTs is much less pronounced. We assume that the reported difference in the binding energies can actually explain the reproducibly higher widths of the C 1s line for the present mixture consisting of s- and m-SWCNTs as compared to highly ordered pyrolytic graphite (HOPG). On the basis of these considerations we described our high resolution C 1s spectra of SWCNTs obtained at a photon energy of 330 eV (not shown here) with two peaks representing metallic and semiconductor nanotubes; in this description the peak-shape parameter for m-SWCNT were similar to those of HOPG. Our spectra, however, are described best if the energetic difference between the two peaks is slightly larger than in the value of ref 23, namely, about 0.175 eV. By use of the peak shape and splitting derived from these highresolution data and by taking into account the present energy resolution, we have approximated our spectra for pristine 3582



ARTICLE

Figure 7. Calculated band structures for one-dimensional crystals (SnTe)4n (left) and (SnTe)5n (right) along the Γ-X direction of the 1D BZ. Fermi level is marked in red; the LUMO and HOMO are shown in blue.

SWCNTs and for SnTe@SWCNT as presented in parts b and c of Figure 4. Within this concept, the relative intensity of the peak for m-SWCNT is diminished after filling tubes with SnTe. Therefore, we assume that the intercalation of the 1D (SnTe)5n crystal results in a slight modification of the m-SWCNTs electronic structure, i.e., slight electron doping effect (and the corresponding BE upshift), which is in line with metal properties of the (SnTe)5n crystal. It should be, however, emphasized that the observed effect in the photoelectron spectra upon intercalation is minor compared to the changes in the photoemission spectra of wide band gap copper and silver halides, for which a pronounced p-doping effect is observed.6,24 This strongly suggests that there is practically no charge transfer within the nanocomposites under consideration. The plasmon-loss structures in the C 1s spectrum (see Figure 5 and Table 2) at binding energies around 290 and 294 eV also does not change significantly upon intercalation. The observed slight decrease of the π-plasmon intensity at binding energies around 290 eV, which is not accompanied by changes in the energy position, indicates also only a weak influence of the SnTe crystal on the nanocomposite electronic structure. The positions of the Sn 3d and Te 3d lines for SnTe@SWCNT (not shown here) are very close to the bulk values, however, with a slight increase of the energy difference between these both lines of about 0.1 eV. This can be attributed to the fact that the effective

charges for the 1D crystal are slightly lower than for the bulk, as obtained from the present calculations and summarized in Table 3. It should be noted that the effective charges for all structures including SnTe molecules (CN = 1) and bulk tin telluride (CN = 6) shows district correlation (nearly linear behavior) with CN—they decrease for the lower CN. Raman spectroscopy, which is widely used to characterize carbon materials, can give more details on the modification of the SWCNTs electronic structure. The Raman spectra for SnTe@SWCNT recorded in the region of the C-C bond (G mode) vibrations and the RBM are shown in Figure 1 together with the spectra obtained for the pristine SWCNT; the corresponding peak positions are listed in Table 4. Three different wavelengths of 514 nm (2.41 eV), 633 nm (1.96 eV), and 785 nm (1.58 eV) were used for the vibrational excitations. The value of 2.41 eV corresponds to the resonant excitation of the s-SWCNTs, whereas at 1.58 eV the metallic tubes are presumably excited. In all cases, only minor changes in the spectral shape and peak positions were observed. The most pronounced changes were registered for the spectra excited with energy of 1.58 eV. These changes are all related to peaks of the G mode, which show a shift of 3 cm-1 to higher binding energies. Therefore, we can assume that the intercalated 1D crystal has a weak influence on m-SWCNTs, which is in line with the XPS results described 3583



ARTICLE

Figure 8. Total density of states for one-dimensional crystals (SnTe)4n (a) and (SnTe)5n (b) and bulk SnTe26 (c).

above. For the RBM peak shifts of 2 cm-1 are observed toward both higher and lower frequencies. This can be explained by the fact that the intercalation influences the electronic structure of SWCNTs in a different manner for metallic and semiconductor tubes. In addition, the most significant changes are observed for low-diameter nanotubes, which have higher RBM frequencies (see Figure 1 and Table 4). We can attribute this effect to a stronger interaction between the SnTe crystal and the narrower nanotubes. This is in line with the electron-density calculations, which predicted a small but nonvanishing interaction between the structure (SnTe)5n and the nanotubes with a diameter of 1.34 nm, see also above. Finally, optical absorption spectrum obtained for SnTe@SWCNT and shown in Figure 6 in comparison with the spectrum for pristine SWCNTs proves also that the electron structure of carbon nanotube does not change essentially due to SnTe intercalation with exception of a minor feature observed at 0.4 eV, which can probably be assigned to either electron structure of 1D SnTe crystal or a weak crystal-SWCNT interaction. 3.3. Electronic Structure of 1D SnTe. Because of wellknown quantum confinement effect, for 1D crystals and other low-dimensional structures composed by semiconductors one could expect an increase of band gap in comparison with bulk state of 0.19 eV as it was observed for example in the case of

HgTe.25 In addition, the band structure should reveal so-called van Hove singularities. The calculated band structures of free 1D crystals (SnTe)4n and (SnTe)5n are shown in Figure.4. The calculations were performed with PBC for 100 point of k-vector in the first Brillouin zone. For the structure I 20 occupied orbitals and 40 vacant ones were calculated. For the structure II the corresponding values comprise 25 and 50. For total density of states (DOS) calculation the numerical integration of orbital energies over kspace was performed. A Gaussian-type broadening of 0.1 eV was assigned for each energy level to avoid an overstructured DOS line shape. The band structure is presented in Figure 7. The total DOS for both 1D crystals, (SnTe)4n and (SnTe)5n, is shown in Figure 8 in comparison with the calculation data obtained for bulk SnTe.26 It is surprising that the (SnTe)5n shows metallic properties. There is intersection of the two dispersion curves at k = 0.63 2π/a, where a is lattice constant (Figure.7, left panel). In contrast to the (SnTe)5n, the (SnTe)4n crystal possesses a band gap of 2 eV, i.e., much higher than for rocksalt SnTe crystal. Its total DOS line shape is similar to those of bulk SnTe (see Figure 8c). The possible reason for the gapless state of (SnTe)5n is the influence of the edge atoms possessing low coordination number (CN), which is equal to 3. The highest occupied crystal orbital 3584



ARTICLE

Figure 9. Isowave function ((0.02 (a.u.)-3) contours for LUCO (left) and HOCO (right) in the case of (SnTe)5n. Green color corresponds to positive values and red color shows negative values.

Figure 10. The optimized structures of 1D crystals of structure III (a), structure IV (b), and structure V (c).

(HOCO) and lowest unoccupied crystal orbital (LUCO) obtained in calculations are shown in Figure 9 and marked in Figure 4 in blue. The HOCO is mostly composed by 5p orbitals of outer Sn atoms with lower contribution of Te 5p orbitals. The LUCO demonstrates π-bonding of SnTe and Te orbitals along the crystal. To study the influence of the coordination number on the band gap in more details we have modeled a number of 1D crystals of larger size, which do not fit definitely to SWCNT. The optimized structures are shown in Figure 10. For (SnTe)6n two different isomers were considered. The structure shown in Figure 10a (structure III) has no three-coordinated atoms. For the structure presented in Figure 10b (structure IV) two atoms for each layer are of CN = 3. For the (SnTe)7n (structure V) one atom in each layer only has CN = 3. The calculations resulted in the band gap of 1.5 eV for structure III, 2 eV for structure IV, and 1.9 eV for structure V. Thus, for the crystal (SnTe)6n the

structure II with 2n low-coordinated atoms the band gap is definitely lower than for (SnTe)6n, and low coordination is responsible for the band gap decrease due to strong interaction between outer atoms. To some extent, this effect is similar to the reconstruction phenomena at the surface of the covalent crystals. Evidently, for (SnTe)6n the tendency to decrease band gap due to low-coordinated atoms is weaker than for (SnTe)5n but still pronounced. (SnTe)5n itself presents an extreme case of the influence of low-coordinated atoms. It should be noted that all calculated structures do not explain the feature observed in optical absorption spectra at ∼0.4 eV. This feature cannot be assign to van Hove singularities of 1D SnTe crystal—the estimated value of first singularity corresponds to ∼1.5 eV for (SnTe)5n. The nature of this optical transition may be related to a kind of the crystal-template interaction, which is beyond of model applied and, therefore, this question requires further examination. 3585



’ CONCLUSIONS 1D SnTe crystals can be grown in SWCNTs with a high loading factor from the melt using a capillary technique. TEM measurements in combination with DFT calculations clearly show that the atomic structure of the 1D SnTe crystals corresponds to (SnTe)5n with a significant bond lengths relaxation (contraction) along the SWCNT axis. The effective charges of both the Sn and Te atoms are lower than in the bulk of SnTe. Calculated band structure of 1D (SnTe)5n shows gapless state which is due to the contribution of the orbitals belonging to low coordinated atoms with CN = 3. XPS and Raman spectroscopy data show no noticeable interaction between the 1D SnTe crystal and the SWCNTs, except for a minor donor influence of the intercalated crystal on metallic SWCNTs. From this observation we conclude that the reported relaxation effects for the bond lengths are inherent to the 1D SnTe crystal. This also suggests that SnTe@SWCNT is a well-suited model system to study the physics of quasi-freestanding 1D-crystals having only few atoms in its cross-section. ’ AUTHOR INFORMATION Corresponding Author

*Fax: þ74959390998. E-mail: [email protected].

’ ACKNOWLEDGMENT The calculations were performed at the calculation centre ZEDAT of Freie Universit€at Berlin. This work was supported by the Ministry of Education and Science of the Russian Federation (State Contract No. P2307 signed on November 16, 2009; federal program “Scientific and Pedagogical Staff of Innovative Russia in 2009-2013”). ’ REFERENCES (1) Garcia de Abajo, F. J.; Cordon, J.; Corso, M.; Shiller, F.; Ortega, J. E. Nanoscale 2010, 2, 717–721. (2) Mugarza, A.; Schiller, F.; Kuntze, J.; Cordon, J.; Ruiz-Oses, M; Ortega, J. E. J. Phys.: Condens. Matter. 2006, 18, S27–S49. (3) Napolskii, K. S.; Chumakov, A. P.; Grigoriev, S. V.; Grigoryeva, N. A.; Eckerlebe, H.; Eliseev, A. A.; Lukashin, A. V.; Tretyakov, Yu. D. Phys. B 2009, 404, 2568–2571. (4) Sloan, J.; Wright, D. M.; Woo, H.-G.; Bailey, S.; Brown, G.; York, A. P. E.; Coleman, K. S.; Hutchison, J. L.; Green, M. L. Chem. Commun. 1999, 23, 699–700. (5) Monthioux, M.; Flahaut, E.; Cleuziou, J. P. J. Mater. Res. 2006, 21, 2774–2793. (6) Eliseev, A. A.; Yashina, L. V.; Brzhezinskaya, M. M.; Chernysheva, M. V.; Kharlamova, M. V.; Verbitsky, N. I.; Kiselev, N. A.; Zakalyuhin, R. M.; Hutchison, J. L; Freitag, B.; Vinogradov, A. S. Carbon 2010, 48, 2708–2721. (7) Eliseev, A. A.; Kharlamova, M. V.; Chernysheva, M. V.; Lukashin, A. V.; Tretyakov, Yu. D.; Kumskov, A. S.; Kiselev, N. A. Russ. Chem. Rev. 2009, 78, 833–854. (8) Kramberger, C.; Rauf, H.; Knupfer, M.; Shiozawa, H.; Batchelor, D.; Kataura, H.; Pichler, T. Phys. Status Solidi B 2009, 246, 2693–2698. (9) De Blauwe, K.; Kramberger, C.; Plank, W.; Kataura, H.; Pichler, T. Phys. Status Solidi B 2009, 246, 2732–2736. (10) Sceats, E. L.; Green, J. C.; Reich, S. Phys. Rev. B 2006, 73, 125441. (11) Khomyakov, P. A.; Giovanetti, G.; Rusu, P. C.; Brocks, G.; van der Brink, J.; Kelly, P. J. Phys. Rev. B 2009, 79, 195425. (12) Batyrev, I. G.; Kleinman, L.; Leiro, J. Phys. Rev. B 2004, 70, 073310.

ARTICLE

(13) Ma, J.; Jia, Y.; Song, Y.; Liang, E.; Wu, L.; Wang, F.; Wang, X.; Hu, X. Surf. Sci. 2004, 551, 91–98. (14) Alvarez, L.; Righi, A.; Guillard, T.; Rols, S.; Anglaret, E.; Laplaze, D.; Sauvajol, J. L. Chem. Phys. Lett. 2000, 316, 186–190. (15) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652. (16) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1998, 37, 785. (17) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, revision C.02; Gaussian, Inc.: Wallingford, CT, 2004. (18) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270–283. (19) Huzinaga, S. Gaussian basis sets for molecular calculations; Elsevier: Amsterdam, 1984. (20) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, Jr., J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, N. J.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, revision A.02; Gaussian, Inc., Wallingford CT, 2009. (21) Zyubin, A. S.; Dedyulin, S. N.; Shtanov, V. I.; Yashina, L. V. Russ. J. Inorg. Chem. 2007, 52, 83–91. (22) Yashina, L. V.; Zyubina, T. S.; Puettner, R.; Zyubin, A. S.; Shtanov, V. I.; Tikhonov, E. V. J. Phys. Chem. C 2008, 112, 19995–20006. (23) Ayala, P.; Miyata, Y.; De Blauwe, K.; Shiozawa, H.; Feng, Y.; Yanagi, K.; Kramberger, C.; Silva, S. R. P.; Follath, R.; Kataura, H.; Pichler, T. Phys. Rev. B 2009, 80, 205427. (24) Kharlamova, M. V.; Eliseev, A. A.; Yashina, L. V.; Petukhov, D. I.; Liu, C. P.; Wang, C. Y.; Semenenko, D. A.; Belogorokhov, A. I. JETP Lett. 2010, 91, 196–200. (25) Carter, R.; Sloan, J.; Kikland, A. I.; Meyer, R. R.; Lindan, P. J. D.; Lin, G.; Green, M. L. H.; Vlandas, A.; Hutchison, J. L.; Harding, J. Phys. Rev. Lett. 2006, 96, 215501. (26) Littlewood, P. B.; Mihaila, B.; Schulze, R. K.; Safarik, D. J; Gubernatis, J. E; Bostwick, A.; Rotenberg, E.; Opeil, C. P.; Durakiewicz, T.; Smith, J. L.; Lashley, J. C. Phys. Rev. Lett. 2010, 105, 086404.

3586


Growth and Characterization of One-Dimensional SnTe Crystals within

Recommend Documents