Absorption of gas-phase atomic hydrogen by Si(100): Effect of surface atomic structures

Full text of DOI:10.1063/1.1379989

Absorption of gas-phase atomic hydrogen by Si(100): Effect of surface atomic
structures
Jae Yeol Maeng, Sehun Kim, S. K. Jo, W. P. Fitts, and J. M. White
Citation: Applied Physics Letters 79, 36 (2001); doi: 10.1063/1.1379989
View online: http://dx.doi.org/10.1063/1.1379989
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/79/1?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Thermal stability and hydrogen atom induced etching of nanometer-thick a-Si:H films grown by ion-beam
deposition on Si(100) surfaces
J. Vac. Sci. Technol. A 21, 831 (2003); 10.1116/1.1575213
Development of procedures for obtaining clean, low-defect-density Ge(100) surfaces
J. Vac. Sci. Technol. A 19, 976 (2001); 10.1116/1.1367264
In situ scanning tunneling microscopy characterization of step bunching on miscut Si(111) surfaces in fluoride
solutions
J. Appl. Phys. 85, 1545 (1999); 10.1063/1.369285
Atomic structures of hydrogen-terminated Si(001) surfaces after wet cleaning by scanning tunneling microscopy
Appl. Phys. Lett. 73, 1853 (1998); 10.1063/1.122304
In situ Si flux cleaning technique for producing atomically flat Si(100) surfaces at low temperature
Appl. Phys. Lett. 70, 2288 (1997); 10.1063/1.119083
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
155.33.120.209 On: Wed, 03 Dec 2014 18:32:59

APPLIED PHYSICS LETTERS
VOLUME 79, NUMBER 1
2 JULY 2001
Absorption of gas-phase atomic hydrogen by Si„100…:
Effect of surface atomic structures
Jae Yeol Maeng and Sehun Kima)
Department of Chemistry and School of Molecular Science (BK 21), Korea Advanced Institute of Science
and Technology, Taejon 305-701, Korea
S. K. Jo
Department of Chemistry, Kyung Won University, SungNam, KyungKi 461-701, Korea
W. P. Fitts and J. M. White
Science and Technology Center, University of Texas, Austin, Texas 78712
͑Received 16 February 2001; accepted for publication 27 april 2001͒
The atomic-scale surface structural evolution of Si͑100͒ exposed to gas-phase thermal hydrogen
atoms, H(g), has been investigated by scanning tunneling microscopy and temperature-programed
desorption mass spectrometry. For the substrate temperature (Ts) between 420 and 530 K, dihydride
species in 3ϫ1:H domains were selectively etched upon extensive exposures to H(g). As a result,
etch pits grew laterally along Si surface dimer rows. The presence of these pits correlates with the
absorption of H(g) into the bulk of Si͑100͒, confirming our earlier suggestion that atomic-scale
surface roughening caused by etching is a prerequisite for H(g) absorption. © 2001 American
Institute of Physics. ͓DOI: 10.1063/1.1379989͔
Hydrogen, prevalent in semiconductor device process-
ing, deactivates dopants, and creates and passivates defects
on and in crystalline silicon (c-Si)1 and at Si/SiO2
interfaces.2 Passivation of defects is beneficial, but incorpo-
ration of hydrogen into c-Si epifilms or substrates can be
detrimental to device performance. Although the interaction
between Si͑100͒ and gas-phase thermal hydrogen atoms,
H(g), has been investigated intensively over the past several
decades,3 direct H(g) incorporation into c-Si is a recent
observation.4
Surface uptake of H(g), on Si͑100͒-2ϫ1 leads to three
different phases depending on the substrate temperature (Ts)
and H coverage (␪H): ͑1͒ a 2ϫ1:H monohydride phase
(␪Hϭ1 ML) at ϳ600 K; ͑2͒ a 3ϫ1:H phase with the alter-
nating dihydride and monohydride units (␪Hϭ1.33 ML) at
ϳ400 K; and ͑3͒ a 1ϫ1:H dihydride phase (␪Hϭϳ2 ML)
of at ϳ300 K.3–5
The latter two phases give two H2 desorption peaks at
670 K (␤2) and 780 K (␤1), the recombinative desorption
from surface di- and monohydrides, respectively.3,4 Recently,
Kang et al.4 reported an additional and unsaturable H2
temperature-programed desorption ͑TPD͒ peak ͑␣͒ at 860–
900 K from Si͑100͒ surfaces held between 420 and 530 K
during H(g) exposure, which was attributed to hydrogen
evolution from the c-Si bulk. They also showed TPD evi-
dence that surface etching and microscopic surface roughen-
ing were associated with this bulk uptake.4
In this letter, we present atomic-scale scanning tunneling
microscopy ͑STM͒ images of the etch features produced by
H(g) and show a direct correlation between microscopic sur-
face structure and H(g) incorporation into the subsurface of
Si͑100͒. This work provides interesting fundamental and
technological insights.
Experiments were performed in an UHV chamber ͑base
pressure of Ͻ2ϫ10Ϫ10 Torr͒ equipped with an OMICRON
VT-STM. The Si͑100͒ sample used was N-type ͑P-doped;
р1 ⍀cm͒ and cut into a 2ϫ10 mm2 size for STM measure-
ments. The clean surface was prepared by outgassing for 12
h at ϳ900 K and flash annealing to ϳ1350 K in vacuum
without ex situ sample pretreatment. The clean and ordered
Si͑100͒-2ϫ1 surface was confirmed by STM. H2 gas was
introduced into the UHV chamber through a tubular doser
controlled by a variable leak valve. To remove impurities
such as hydrocarbons and water in the gas delivery line, a
liquid-nitrogen cold trap was used. H(g) was produced by
cracking H2 molecules with a hot (ϳ1800 K) spiral W fila-
ment positioned ϳ5 cm away from the Si surface. Exposures
are reported in langmuirs (1 Lϭ1ϫ10Ϫ6 Torr •s) of H2, not
H, because more than 98% of the measured chamber pres-
sure increment is due to H2. All H(g) exposures were done
at a fixed Ts of 460 K lying at the center of the interval
where H(g) absorption occurs. All STM images of 25 nm on
a side were taken at VsampleϭϪ1.6 V and Itϭ0.1 nA at room
temperature using electrochemically etched W tips.
Figure 1 shows STM images of Si͑100͒ surfaces exposed
to H(g) of: ͑a͒ 3000 L; and ͑b͒ 10 000 L H2 at Ts of 460 K.
The surface in Fig. 1͑a͒ exhibits a 2ϫ1:H phase with few
imperfections. For the surface in Fig. 1͑b͒, a 3ϫ1:H phase,
with its alternating parallel rows of bright monohydride
dimer and dark dihydride units, are formed on the surface. A
few atom vacancies due to slight surface etching and several
anti-phase boundaries ͑see the boxed area͒ within 3ϫ1:H
areas are also observed.6 Domains of 2ϫ1:H ͑see the circled
area͒ also persist, reflecting an intermediate stage of the 2
ϫ1:H to 3ϫ1:H phase transition,6 which is in part due to
the relatively high H(g) dosing Ts of 460 K employed here.
Increasing the exposure by factors of 5 and 10 ͓Figs.
2͑a͒ and 2͑b͔͒ indicates significant etching. Images of
a͒
Electronic mail: sehkim@mail.kaist.ac.kr
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
0003-6951/2001/79(1)/36/3/$18.00 36 © 2001 American Institute of Physics
155.33.120.209 On: Wed, 03 Dec 2014 18:32:59

Appl. Phys. Lett., Vol. 79, No. 1, 2 July 2001
Maeng et al.
37
FIG. 2. Filled-state STM images of Si͑100͒ exposed to H(g) of ͑a͒ 50 000
L H2; and ͑b͒ 100 000 L H2 at Ts of 460 K. Etch pit lines ͑circled͒ occurring
in local ͑1ϫ1͒ regions are indicated.
FIG. 1. Filled-state STM images of Si͑100͒ exposed to H(g) of ͑a͒ 3000 L
H2; and ͑b͒ 10 000 L H2 at Ts of 460 K. An antiphase boundary ͑boxed͒ and
2ϫ1:H domain ͑circled͒ are indicated.
mechanism for initiating the surface etching reaction.7 Most
of surface vacancies are located on dihydride rows, and etch
pit lines are mostly observed in local 1ϫ1:H regions ͑see
the circled areas in Fig. 2͒. Figure 2͑a͒ shows that most of
etch pit lines occur between two dihydride rows reflecting
this repulsion-enhanced etching within 1ϫ1:H domains.
Compared with adjacent dihydride units in 3ϫ1:H domains,
those in 1ϫ1:H domains experience increased steric
repulsion,5 and are more vulnerable to H(g) attack.
300 000 and 500 000 L H2 ͑not shown͒ do not differ much
from the 100 000 L H2 image of Fig. 2͑b͒. Figure 2͑a͒ shows
that atom vacancies and vacancy lines are clearly evident and
located mainly along dihydride rows, in contrast to previous
reports that the 3ϫ1:H phase is stable to extended H(g)
exposures.5 Unlike a previous suggestion,6 our results show
no preferential etching of dihydrides at antiphase boundaries.
3ϫ1:H domain sizes in Fig. 2͑a͒ are much smaller than
those in Fig. 1͑b͒, a result attributed to complicated local
surface reconstructions forced by atom vacancies and strain
introduced by etching.6 The terraces formed within larger
etch pits retain the 3ϫ1:H domain structure and contain etch
vacancies, Fig. 2͑b͒.
The linear and rectangular pits are a result of initiation
by selective dihydride etching, subsequent branching and,
finally, coalescence of adjacent linear pits.8 The total etch pit
area after 100 000 L H2, measured by summing missing at-
oms in single atom vacancies and etch pits from the STM
images, is 30% of the total surface area ͓Fig. 2͑b͔͒. For a
five-fold larger exposure ͑not shown͒, the pit area is 35% of
Surface trihydride species, SiH3(a), produced by H in-
sertion into SiH2(a), can be hydrogenated by H(g) and re-
moved as SiH4(g). The random distribution of atom vacan-
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
cies observed in Fig. 2 is in accord with the Eley–Rideal
the total, indicating that 100 000 L H2 is close to steady-state
155.33.120.209 On: Wed, 03 Dec 2014 18:32:59

38
Appl. Phys. Lett., Vol. 79, No. 1, 2 July 2001
Maeng et al.
analysis. Confirming our earlier suggestion,4 the TPD and
STM results provide solid evidence for the correlation be-
tween H(g) absorption into bulk Si͑100͒ and surfaces rough-
ened with atomic-scale etch vacancies and lines. Further
support occurs from preliminary results showing that rough-
ening induced by Arϩ bombardment also facilitates H(g)
absorption.
In summary, our STM results show that H(g) dosing of
Si͑100͒ at Ts of 460 K etches the 3ϫ1:H domain structure
creating atom vacancies and atomic-scale etch pits. These
atomic-scale etch features provide a low activation-energy
path for H(g) to incorporate into bulk Si͑100͒.
This work was supported in part by the US National
Science Foundation Science and Technology Centers pro-
gram, Grant No. CHE8920120, Korea Science and Engineer-
ing Foundation ͑KOSEF Grant No. 98-0501-01-01-03͒, and
Brain Korea 21 project. S.K.J. acknowledges funding from
Korea Science and Engineering Foundation ͑KOSEF Grant
No. 97-0501-01-01-3͒ and Kyung Won University ͑1999͒.
FIG. 3. H2 ͑solid͒ and SiH4 ͑dashed͒ TPD spectra from Si͑100͒ exposed to
H(g) of 500 000 L H2 at 460 K. Also displayed as a reference is a H2 TPD
spectrum ͑dotted͒ from Si(100)-3ϫ1:H prepared by a H(g) dose 1500 L
H2 at Ts of 430 K. The ␣ peak is unsaturable and attributed to H absorbed
into bulk Si͑100͒.
1 Hydrogen in Semiconductors: Bulk and Surface Properties, edited by M.
Stutzmann and J. Chevallier ͑North Holland, Amsterdam, 1991͒; S. J.
Pearton, J. W. Corbett, and M. Stavola, Hydrogen in Crystalline Semicon-
ductors ͑Springer, Berlin, 1992͒.
concentration of missing atoms and, by inference, steady-
state etching rate. Kang et al.4 showed that etching rates at
TsϽ400 K overwhelm H(g) absorption rates while for Ts
Ͼ600 K, removal of surface hydrogen, H(a) via abstraction
by H(g) dominates etching and surface roughening, and
therefore H(g) is not absorbed.
In Fig. 3, we show H2 and SiH4 TPD spectra from a
Si͑100͒ surface exposed to H(g) of 500 000 L H2 at Ts of
460 K. Consistent with previous work, there is a strong ␣
peak from bulk H.4 There is also measurable SiH4 desorption
reflecting surface SiH3(a), which is also confirmed by bright
protrusions in STM images after large H(g) doses at Ts be-
tween 420 and 530 K. From comparison with the reference
3ϫ1:H ␤1 plus ␤2 intensity of H2, we estimate that the
surface area has increased by ϳ30% as result of the etching-
induced roughening, in agreement with the STM image
2 H. J. Queisser and E. E. Haller, Science 281, 945 ͑1998͒; J. W. Lyding, K.
Hess, and I. C. Kizilyalli, Appl. Phys. Lett. 68, 2526 ͑1996͒.
3 H. N. Waltenburg and J. T. Yates, Jr., Chem. Rev. 95, 1589 ͑1995͒; K.
Oura, V. G. Lifshits, A. A. Saranin, A. V. Zotov, and M. Katayama, Surf.
Sci. Rep. 35, 1 ͑1999͒.
4 J. H. Kang, S. K. Jo, J. Lee, B. Gong, D. Lim, J. M. White, and J. G.
Ekerdt, Phys. Rev. B 59, 13170 ͑1999͒; S. K. Jo, J. H. Kang, X. Yan, J. M.
White, J. G. Ekerdt, J. W. Keto, and J. Lee, Phys. Rev. Lett. 85, 2144
͑2000͒.
5 J. J. Boland, Phys. Rev. Lett. 65, 3325 ͑1990͒; Surf. Sci. 261, 17 ͑1992͒.
6 X. R. Qin and P. R. Norton, Phys. Rev. B 53, 11100 ͑1996͒; T. Komeda
and Y. Kumagai, ibid. 58, 1385 ͑1998͒; C. Thirstrup, M. Sakurai, T. Na-
kayama, and M. Aono, Surf. Sci. 411, 203 ͑1998͒.
7
¨
A. Dinger, C. Lutterloh, and J. Kuppers, Chem. Phys. Lett. 320, 405
͑2000͒.
8 M. Chander, D. A. Goetch, C. M. Aldao, and J. H. Weaver, Phys. Rev.
Lett. 74, 2014 ͑1995͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
155.33.120.209 On: Wed, 03 Dec 2014 18:32:59