The transition between conformal atomic layer epitaxy and nanowire growth.

Abstract

Conformal atomic layer deposition of thin Sb(2)S(3) layers takes place epitaxially on suitable substrates at 90 degrees C. More elevated deposition temperatures increase the mobility of the solid and result in the diffusion of Sb(2)S(3) along surface energy gradients. On an Sb(2)Se(3) wire that presents the high-energy c facet at its extremity, this results in the axial elongation of the wire with a Sb(2)S(3) segment. When an Sb(2)S(3) wire whose c planes are exposed on the sides is used as the substrate, the homoepitaxy collects material laterally and yields a nano-object with a rectangular cross section.

Full text

Page 1

The Transition between Conformal Atomic Layer Epitaxy and Nanowire Growth
Ren Bin Yang,*,†Nikolai Zakharov,†Oussama Moutanabbir,†Kurt Scheerschmidt,†Li-Ming Wu,§
Ulrich Go ¨sele,†,|Julien Bachmann,*,†,‡and Kornelius Nielsch‡
Max Planck Institute of Microstructure Physics, Weinberg 2, 06120 Halle, Germany, Institute of Applied Physics,
UniVersity of Hamburg, Jungiusstr. 11, 20355 Hamburg, Germany, and State Key Laboratory of Structural
Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Fuzhou,
Fujian 350002, P. R. China
Received March 28, 2010; E-mail: renbin@gmail.com; julien.bachmann@physik.uni-hamburg.de
Atomic layer deposition (ALD) is a thin-film technique based on
complementary surface reactions between two thermally stable but
mutually reactive gaseous molecules introduced in alternating fashion
into the reaction chamber.1It offers the unique possibility of coating
substratesofcomplexgeometriesconformally.Thevapor-liquid-solid
mechanism (VLS), on the other hand, occurs when gas molecules
thermally decompose into their elements at a catalytic droplet and
dissolve into it to form a eutectic mixture from which the desired solid
crystallizes.2Here, the growth is anisotropic and yields nanowires.
Although conformal and anisotropic modes of deposition are opposed
in most aspects, we demonstrate in this work the possibility of a
continuous transition between them, and we apply the insight gleaned
from our investigation to the creation of original wire heterostructures.
Nanowire junctions, the functional building blocks for creating
devices,3are created most simply in either the axial or radial direction
(two segments or core-shell). The former structure is usually grown
in VLS mode at a temperature slightly above the relevant eutectic
temperature but usually below the temperature for uncatalyzed
decomposition of the precursor molecules. In contrast to this, a shell
is usually grown around an existing wire at an elevated temperature
in order to allow for efficient, uncatalyzed, isotropic decomposition
of the precursors. The effect of the catalyst particle may in that case
result in a tapered nanowire.4The self-limiting nature of ALD, the
alternative for growing a shell, necessitates that uniform coatings,
mostly amorphous ones, be obtained. Crystalline coatings have been
obtained under the label “atomic layer epitaxy” (ALE) at growth
temperatures typically g300 °C.5
The testbed chosen for our study was a material system of relevance
to thermoelectric and photovoltaic applications: antimony sulfide/
selenide.6For Sb2S3, we were able to establish both the ALD and
VLS growth modes from a single set of precursors, namely, tris(dim-
ethylamino)antimony, (Me2N)3Sb, and hydrogen sulfide, H2S, at 90
and 350 °C, respectively.7This implies that between those two
temperatures, the system undergoes a transition from ALD to VLS.
We first used as the substrate Sb2Se3nanowires grown by 500 cycles
ofAu-catalyzedpulsedVLS(Figure1a).Thesenanowiresweresingle-
crystalline with the c axis of the orthorhombic structure parallel to
their long axis. After the nanowires were cooled, 500 cycles of Sb2S3
ALD carried out at 90 °C deposited a homogeneous shell, as is evident
fromtheincreaseddiametershowninthescanningelectronmicroscopy
(SEM) image in Figure 1b. The transmission electron microscopy
(TEM) image in Figure 1c shows that the entire nanowire structure,
including the gold catalyst, was covered with a smooth shell having a
thickness of 25-30 nm, consistent with isotropic ALD growth at the
previously determined deposition rate (0.58 Å cycle-1).7aThe energy-
dispersiveX-rayfluorescence(EDX)linescanperformedperpendicular
to the longitudinal direction revealed a sharp interface between the
Sb2Se3core and the Sb2S3shell (Figure 2a). No Se was detected in
theshell.Theatomicplanesvisibleintheshellbyhigh-resolutionTEM
(HRTEM; Figure 2b) reveal that in contrast to our ALD of Sb2S3
performed between 65 and 150 °C on other substrates, the ALD layer
on the Sb2Se3wires was crystalline and consisted of one solid crystal
with its c direction running parallel to the long axis, including on the
Au sphere. This establishes the low-temperature ALE of Sb2S3on a
nonplanar single-crystalline substrate. The epitaxy is made possible
by the common crystal structure of Sb2S3and Sb2Se3. The almost
isotropic lattice mismatch of 3-4%, however, must cause a very
significant elastic strain. The crystallinity of both the selenide and
sulfide components in the core-shell structures was confirmed on the
ensemble scale spectroscopically (Figure S1 in the Supporting Infor-
mation).Inmicro-Ramanspectroscopy,thecore-shellsamplefeatured
†Max Planck Institute of Microstructure Physics.
§Fujian Institute of Research on the Structure of Matter.
‡University of Hamburg.
|Deceased Nov 8, 2009.
Figure 1. Conformal heteroepitaxial deposition of Sb2S3at 90 °C onto Sb2Se3
wires. (a, b) SEM and (c) TEM images of (a) the substrate and (b, c) the
core-shell wires.
Figure 2. Conformal heteroepitaxial deposition of Sb2S3at 90 °C onto Sb2Se3
wires. (a) EDX line scan across a core-shell wire (pictured underneath the
graph); the vertical axis displays the signal intensities recorded for the elements
Se, Sb, and S in purple, black, and green, respectively. (b) HRTEM image of
the extremity of a wire, displaying the dark Au particle and identifying the
crystallographic c axis parallel to the long axis of the wire.
10.1021/ja102590v  XXXX American Chemical Society
J. AM. CHEM. SOC. XXXX, xxx, 000 9 A

Page 2

signals observed from the pure wires (186 and 252 cm-1for Sb2Se3;
147, 278, and 302 cm-1for Sb2S3) without the broadening that would
be expected from the presence of amorphous or intermixed material.8
A change in the growth morphology was observed at the slightly
more elevated temperature of 120 °C. The sample obtained after the
Sb2Se3wires were subjected to the conditions of Sb2S3ALD at 120
°C (500 cycles; Figure 3) showed a shell radially encapsulating the
selenide wires but also a considerable axial overgrowth at the tip. This
axial elongation occurred at the expense of the Sb2Se3/Au interface:
the metal catalyst was pushed sideways. Finally, when the experiment
was repeated at 160 °C, the growth of Sb2Se3became purely uniaxial
in the crystallographic c direction. When 500 growth cycles were
carried out at 160 °C, a new segment with a length of several hundred
nanometers and a diameter similar to that of the selenide segment
elongated the wire from its tip (Figure 4). The Au particle remained
on the side at the height of the interface. EDX line scans attested to
the complete absence of radial growth within the last micrometer of
the structures, the purity of each segment, and the sharpness of the
interface perpendicular to the long axis of the structure. Inspection of
the wires at larger distances from the tips, however, revealed the
presence of Sb2S3remnants randomly distributed as small crystallites
on the sides. In HRTEM images, no grain boundary was observed at
the selenide-sulfide interface, although a small number of wires
displayed a crack at the interface (Figure S2), most likely originating
fromtheepitaxialstraincausedbythelatticemismatchbetweensulfide
and selenide.
A clarification is needed before a model is put forth to account for
these observations. Indeed, the low decomposition temperatures of
some amorphous V-VI compounds, whereby loss of the group VI
element takes place, suggests that Se might be exchanged for S under
the conditions of Sb2S3ALD. However, this scenario can be excluded
on the basis of experiments in which Sb2Se3wires were kept at 160
°Covernighteitherundercontinuouspumpingorwithhydrogensulfide
regularly pulsed into the vacuum. Both procedures left the wires and
the gold tips unscathed structurally and without trace of S in EDX
line scans. Thus, we conclude that the Sb2S3material obtained after
the (Me2N)3Sb + H2S step (whether at 90, 120, or 160 °C) was
exclusively deposited onto the selenide substrate. The crystallinity of
the sulfide deposit observed at 90 °C contrasts with the results obtained
in the case of amorphous substrates and suggests that the deposited
solid has a significant mobility and crystallizes quickly during the
deposition. Increasing the temperature thus enhances surface diffusion
during deposition and thereby emphasizes the consequences of the
lattice driving force over those of the chemical reaction. The larger
thermal budget enables the crystal facet of highest energy to attract
material from the other faces. We found by density functional theory
methods that the surface tensions of both Sb2S3and Sb2Se3are larger
on their c planes (at the wire extremities) than on the a and b planes
(on the sides of the cylinder) by 20-58%,9in agreement with the
experimental results available.10At the higher temperature, the crystal
directs the growth, which becomes completely anisotropic: the tip
collects the material initially deposited along the whole wire, as may
also be the case in some cases of standard VLS growth.11We note
the possible advantages offered by this method of growing segmented
wires in comparison with the traditional consecutive VLS steps. First,
it directly prevents the undesired incorporation of catalyst material into
the second segment and its concomitant tapering.12Second, all of the
materialdissolvedinthecatalystdropcrystallizesoutofituponcooling
before the second growth step. This elegantly circumvents the
“reservoir” effect that otherwise smears out the interface.13The
mechanism driving the lateral motion of the catalyst cannot be
determinedfromourdata.Suchmotionhas,however,beeninvestigated
previously in other materials systems.14It was shown to arise when
gold has a lower interfacial energy with the first segment than with
the second one.
If the considerations above are correct and also apply to the
homoepitaxy of Sb2S3, then we can foresee the appearance of a novel
geometry. Indeed, VLS-grown Sb2S3wires furnish a substrate comple-
mentary to their Sb2Se3counterparts in that they mostly grow in the
b direction (instead of c), as shown by HRTEM (Figure S3). Thus,
homoepitaxial c-selective deposition at 160 °C should result in
anisotropic lateral growth. Indeed, after deposition at 160 °C onto
the c faces exposed on the sides, the wire cross section is a rectangle.
Its shorter side equals the diameter of the gold catalyst, which is left
at the tip (Figure 5c-e). The temperature for the onset of anisotropic
growth (65-90 °C) is lower than in the heteroepitaxial case. This
difference may originate from the diffusion distances, which are
micrometers along the length of a wire but only tens of nanometers
on its sides.
Thus, our observations indicate that on a suitable substrate, the
crystal energy becomes prevalent in the thermodynamics of the thin-
film growth initially thought of as “atomic layer deposition”. Surface
diffusion enables the experimentalist to balance it with the ubiquitous
driving force of the chemical reaction to obtain epitaxial growth at
exceptionally low temperatures. Furthermore, one can take control of
the growth morphology and tune it at will between completely
Figure 3. At 120 °C, conformal and anisotropic heteroepitaxial growths of
Sb2S3onto Sb2Se3wires coexist, as shown by (a) TEM and (b) HRTEM images
of one such structure. The region of (a) chosen for closer investigation in (b)
is marked by a rectangle.
Figure 4. PerfectlyanisotropicheteroepitaxialdepositionofSb2S3ontoSb2Se3
wires at 160 °C. (a) SEM and (b) TEM images. (c) HRTEM images near the
interface, in which no crystal defect is visible. (d) EDX linescans of the
segmented wire. The radial scans show the absence of S (green) in the Sb2Se3
segment and of Se (purple) in Sb2S3; the axial scan, scaled to display the
stoichiometry, shows an interface more abrupt than the EDX resolution.
BJ. AM. CHEM. SOC. 9 VOL. xxx, NO. xx, XXXX
COMMUNICATIONS

Page 3

conformal and perfectly uniaxial. As a bonus, judicious exploitation
of surface energy effects can deliver unusual geometries, such as the
rectangular wires converted from their cylindrical parents. Thus, the
deposition of only two compounds, Sb2S3and Sb2Se3, under various
conditions using different combinations of two distinct growth modes
enables one to create a wide variety of structures, including wires,
tubes,core-shellwires,segmentedwires,andrectangularwires(Figure
6). In all heteroepitaxial cases, the interface appears to be experimen-
tally perfect, that is, it displays no crystal defects. No interdiffusion of
the two materials is observed, in contrast to what has been shown to
happen in many other cases.15
The core-shell structures presented here demonstrate the possibility
of driving ALE at temperatures well below those used to date.16The
growth of the double-segment structure relies on the high crystal
anisotropyofalayeredsolid,whichdrivesperfectlyanisotropicgrowth
in a manner reminiscent of the solvothermal growth of wires and
rods.17The catalytic droplet is thus used to define the nanowire
diameterinitially,andthenthefaceofhighestsurfaceenergyisexposed
for the growth of a second segment. We expect that this method will
beadaptedtoothermaterialsystemsfeaturinghighlyanisotropiccrystal
structures.
Acknowledgment. We thank M. Alexe and D. Hesse for helpful
discussions and insightful comments. R.B.Y. was funded by the
InternationalMaxPlanckResearchSchoolforScienceandTechnology
of Nanostructures with a predoctoral fellowship. This work was
supported by the German Ministry of Education and Research (BMBF
03X5519) and the German Priority Program SPP 1386 on Nanostruc-
tured Thermoelectrics.
Supporting Information Available: Additional figures mentioned
in the text (Figures S1-S3), larger versions of the HRTEM images
presented in Figures 2-4 (Figures S4-S6), and experimental and
theoreticalmethods.ThismaterialisavailablefreeofchargeviatheInternet
at http://pubs.acs.org.
References
(1) George, S. M. Chem. ReV. 2010, 110, 111.
(2) Law, M.; Goldberger, J.; Yang, P. Annu. ReV. Mater. Res. 2004, 34, 83.
(3) (a) Wen, C.; Reuter, M. C.; Bruley, J.; Tersoff, J.; Kodambaka, S.; Stach,
E. A.; Ross, F. M. Science 2009, 326, 1247. (b) Kim, C.; Lee, D.; Lee, H.;
Lee, G.; Kim, G.; Jo, M. Appl. Phys. Lett. 2009, 94, 173105. (c) Zakharov,
N.; Werner, P.; Gerth, G.; Schubert, L.; Sokolov, L.; Gos ¨ele, U. J. Cryst.
Growth 2006, 290, 6. (d) Bjork, M. T.; Ohlsson, B. J.; Sass, T.; Persson,
A. I.; Thelander, C.; Magnusson, M. H.; Deppert, K.; Wallenberg, L. R.;
Samuelson, L. Nano Lett. 2002, 2, 87. (e) Wu, Y.; Fan, R.; Yang, P. Nano
Lett. 2002, 2, 83. (f) Gudiksen, M. S.; Lauhon, L. J.; Wang, J.; Smith,
D. C.; Lieber, C. M. Nature 2002, 415, 617. (g) Xiang, J.; Lu, W.; Hu, Y.;
Wu, Y.; Yan, H.; Lieber, C. M. Nature 2006, 441, 489. (h) Hayden, O.;
Greytak, A.; Bell, D. AdV. Mater. 2005, 17, 701. (i) Cui, L.; Ruffo, R.;
Chan, C. K.; Peng, H.; Cui, Y. Nano Lett. 2009, 9, 491. (j) Lauhon, L. J.;
Gudiksen, M. S.; Wang, D.; Lieber, C. M. Nature 2002, 420, 57. (k) Wang,
K.; Chen, J.; Zhou, W.; Zhang, Y.; Yan, Y.; Pern, J.; Mascarenhas, A.
AdV. Mater. 2008, 20, 3248.
(4) (a) Mohseni, P. K.; Maunders, C.; Botton, G. A.; LaPierre, R. R.
Nanotechnology 2007, 18, 445304. (b) Zhou, H. L.; Hoang, T. B.; Dheeraj,
D. L.; Helvoort, A. T. J. V.; Liu, L.; Harmand, J. C.; Fimland, B. O.;
Weman, H. Nanotechnology 2009, 20, 415701.
(5) (a) Ahonen, M.; Pessa, M.; Suntola, T. Thin Solid Films 1980, 65, 301. (b)
Koukitu, A.; Miyazawa, T.; Ikeda, H.; Seki, H. J. Cryst. Growth 1992,
123, 95. (c) Yao, T.; Takeda, T. Appl. Phys. Lett. 1992, 48, 160. (d) Wu,
Y.; Toyoda, T.; Kawakami, Y.; Fujita, S.; Fujita, S. Jpn. J. Appl. Phys.
1990, 29, L727. (e) Dosho, S.; Takemura, Y.; Konagai, M.; Takahashi, K.
J. Appl. Phys. 1989, 66, 2597.
(6) Melting points: Sb2S3, 550 °C; Sb2Se3, 611 °C.
(7) (a) Yang, R. B.; Bachmann, J.; Pippel, E.; Berger, A.; Woltersdorf, J.;
Gos ¨ele, U.; Nielsch, K. AdV. Mater. 2009, 21, 3170. (b) Yang, R. B.;
Bachmann, J.; Reiche, M.; Gerlach, J. W.; Gos ¨ele, U.; Nielsch, K. Chem.
Mater. 2009, 21, 2586.
(8) (a) Zhang, Y.; Li, G.; Zhang, B.; Zhang, L. Mater. Lett. 2004, 58, 2279.
(b) Lu, J.; Han, Q.; Yang, X.; Lu, L.; Wang, X. Mater. Lett. 2008, 62,
2415. (c) Wang, J.; Deng, Z.; Li, Y. Mater. Res. Bull. 2002, 37, 495. (d)
Mernagh, T. P.; Trudu, A. G. Chem. Geol. 1993, 103, 113.
(9) Computed surface energies of the a, b, and c faces: 0.239, 0.254, and 0.311
J m-2, respectively, for Sb2S3; 0.186, 0.226, and 0.293 J m-2, respectively,
for Sb2Se3.
(10) (a) Givargizov, E. J. Cryst. Growth 1975, 31, 20. (b) Wang, Y.; Chen, J.;
Wang, P.; Chen, L.; Chen, Y.; Wu, L. J. Phys. Chem. C 2009, 113, 16009.
(11) Verheijen, M. A.; Immink, G.; de Smet, T.; Borgstro ¨m, M. T.; Bakkers,
E. P. A. M. J. Am. Chem. Soc. 2006, 128, 1353.
(12) (a) Shchetinin, A. A.; Nebol’sin, V. A.; Kozenkov, O. D.; Tatarenkov, A. F.;
Dunaev, A. I.; Novokreshchenova, E. P. Inorg. Mater. 1991, 27, 1137. (b)
Allen, J. E.; Hemesath, E. R.; Perea, D. E.; Lensch-Falk, J. L.; Li, Z. Y.;
Yin, F.; Gass, M. H.; Wang, P.; Bleloch, A. L.; Palmer, R. E.; Lauhon,
L. J. Nat. Nanotechnol. 2008, 3, 168. (c) Perea, D. E.; Allen, J. E.; May,
S. J.; Wessels, B. W.; Seidman, D. N.; Lauhon, L. J. Nano Lett. 2006, 6,
181. (d) Kodambaka, S.; Tersoff, J.; Reuter, M. C.; Ross, F. M. Phys. ReV.
Lett. 2006, 96, 096105. (e) Kodambaka, S.; Hannon, J. B.; Tromp, R. M.;
Ross, F. M. Nano Lett. 2006, 6, 1292. (f) Hannon, J. B.; Kodambaka, S.;
Ross, F. M.; Tromp, R. M. Nature 2006, 440, 69.
(13) Li, N.; Tan, T. Y.; Gos ¨ele, U. Appl. Phys. A: Mater. Sci. Process. 2008,
90, 591.
(14) (a) Paladugu, M.; Zou, J.; Guo, Y.-N.; Zhang, X.; Joyce, H. J.; Gao, Q.;
Tan, H. H.; Jagadish, C.; Kim, Y. J. Appl. Phys. 2009, 105, 073503. (b)
Paladugu, M.; Zou, J.; Guo, Y. N.; Auchterlonie, G. J.; Joyce, H. J.; Gao,
Q.; Tan, H. H.; Jagadish, C.; Kim, Y. Small 2007, 3, 1873.
(15) Gonza ´lez, J. C.; Malachias, A.; Andrade, R.; de Sousa, J.; Moreira, M.; de
Oliveira, A. G. J. Nanosci. Nanotechnol. 2009, 9, 4673.
(16) All of the “low-temperature” ALE reactions reported to date were performed
at or above 300 °C except for those in the following three reports: (a)
Ohta, S.; Kobayashi, S.; Kaneko, F.; Kashiro, K. J. Cryst. Growth 1990,
106, 166. (b) Fujiwara, H.; Gotoh, J.; Shirai, H.; Shimizu, I. J. Appl. Phys.
1993, 74, 5510. (c) Luo, Y.; Slater, D.; Han, M.; Moryl, J.; Osgood, J.
Appl. Phys. Lett. 1997, 71, 3799.
(17) (a) Peng, X.; Manna, L.; Yang, W.; Wickham, J.; Scher, E.; Kadavanich,
A.; Alivisatos, A. P. Nature 2000, 404, 59. (b) Xi, L.; Tan, W. X. W.;
Boothroyd, C.; Lam, Y. M. Chem. Mater. 2008, 20, 5444. (c) Cao, M.;
Hu, C.; Peng, G.; Qi, Y.; Wang, E. J. Am. Chem. Soc. 2003, 125, 4982.
(d) Kim, B.; Koo, T.; Lee, J.; Kim, D. S.; Jung, Y. C.; Hwang, S. W.;
Choi, B. L.; Lee, E. K.; Kim, J. M.; Whang, D. Nano Lett. 2009, 9, 864.
(e) Paek, J. H.; Nishiwaki, T.; Yamaguchi, M.; Sawaki, N. Phys. Status
Solidi C 2009, 6, 1436. (f) Li, S.; Zhang, X.; Yan, B.; Yu, T. Nanotech-
nology 2009, 20, 495604. (g) Zhang, Y.; Tang, Y.; Lee, K.; Ouyang, M.
Nano Lett. 2009, 9, 437.
JA102590V
Figure 5. Homoepitaxial deposition of Sb2S3is isotropic at 65 °C (a, b) and
results in a cylindrical structure, whereas at 160 °C (c-e) it is selective for the
c plane and yields wires with a rectangular cross section. Scale bars: 200 nm.
(a, c, d) SEM images; (b, e) TEM images.
Figure 6. Structural variety of elongated antimony sulfide and selenide
achieved under various growth conditions and combinations: (a) tubes by ALD
in a porous template;7b(b, c) Sb2S3and Sb2Se3wires by VLS;7a(d) segmented
wires by two consecutive VLS steps;7a(e) core-shell wire by conformal
heteroepitaxy of Sb2S3on Sb2Se3at 90 °C (Figures 1 and 2); (f) segmented
wire with core-shell component by deposition of Sb2Se3at 120 °C (Figure
3); (g) segmented wires by anisotropic heteroepitaxy of Sb2S3on Sb2Se3at
160 °C (Figure 4); (h) rectangular wires by homoepitaxy of Sb2S3at 160 °C
(Figure 5).
J. AM. CHEM. SOC. 9 VOL. xxx, NO. xx, XXXX
C
COMMUNICATIONS