Role of ion energy in determination of the sp3 fraction of ion beam deposited carbon films

Article abstract of DOI:10.1063/1.115973

The role of ion energy over the range 5 eV≤E≤20 keV in the production of the dense diamondlike sp3-bonded phase of carbon films deposited from ion beams has been investigated. Films with a significant sp3 component (?40%), as determined by electron energy loss spectroscopy (EELS), can be formed over the wide energy region 30 eV≤E≤10 keV at room temperature. The sp3 fraction is completely suppressed only for E≤10 eV or E≥20 keV. For both cases, this suppression is associated with a sharp increase of the surface roughness, as determined by atomic force microscopy (AFM). The different nature of the mechanisms responsible for the suppression of sp3 bonding in both the low and high energy regions is discussed.

Full text of DOI:10.1063/1.115973

Role of ion energy in determination of the sp 3 fraction of ion beam deposited carbon
films
E. Grossman, G. D. Lempert, J. Kulik, D. Marton, J. W. Rabalais, and Y. Lifshitz
Citation: Applied Physics Letters 68, 1214 (1996); doi: 10.1063/1.115973
View online: http://dx.doi.org/10.1063/1.115973
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/68/9?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Effects of interface mixing on adhesion of amorphous carbon films synthesized by variable-energy direct carbon
ion beam deposition
J. Vac. Sci. Technol. A 18, 1983 (2000); 10.1116/1.582458
Modeling the ion energy dependence of the sp 3/sp 2 bonding ratio in amorphous diamondlike films produced
with a massseparated ion beam
J. Appl. Phys. 78, 5837 (1995); 10.1063/1.359652
Direct ion beam deposition of carbon films on silicon in the ion energy range of 15–500 eV
J. Appl. Phys. 70, 5623 (1991); 10.1063/1.350177
Modeling of the sp 3/sp 2 ratio in ion beam and plasmadeposited carbon films
Appl. Phys. Lett. 59, 2391 (1991); 10.1063/1.106026
Vapor deposition processes for amorphous carbon films with sp 3 fractions approaching diamond
J. Appl. Phys. 70, 1706 (1991); 10.1063/1.349540
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:
129.120.242.61 On: Tue, 25 Nov 2014 05:52:01

3
Role of ion energy in determination of the sp fraction of ion beam
deposited carbon films
E. Grossman and G. D. Lempert
Soreq NRC, Yavne 81800, Israel
J. Kulik, D. Marton, and J. W. Rabalais
University of Houston, Department of Chemistry, Houston, Texas 77204-5641
Y. Lifshitza)
Soreq NRC, Yavne 81800, Israel
͑
Received 15 May 1995; accepted for publication 21 November 1995͒
The role of ion energy over the range 5 eVрEр20 keV in the production of the dense diamondlike
3
sp -bonded phase of carbon films deposited from ion beams has been investigated. Films with a
3
significant sp component ͑Ͼ40%͒, as determined by electron energy loss spectroscopy ͑EELS͒,
can be formed over the wide energy region 30 eVрEр10 keV at room temperature. The sp3
fraction is completely suppressed only for Eр10 eV or Eу20 keV. For both cases, this suppression
is associated with a sharp increase of the surface roughness, as determined by atomic force
microscopy ͑AFM͒. The different nature of the mechanisms responsible for the suppression of
3
sp bonding in both the low and high energy regions is discussed. © 1996 American Institute of
Physics. ͓S0003-6951͑96͒02804-9͔
Energetic ions have been employed in a variety of film
deposition technologies.1 The kinetic energy of the imping-
ing species plays an important role in determination of the
minimal energy of Eϭ30 eV was necessary for formation of
–7
3
films with a significant amount of sp bonding and ͑ii͒ films
3
with a high sp fraction could be formed even for Eϭ1 keV,
1
–7
evolving film characteristics. Since the work of Aisenberg
which is a higher energy than previously expected. This let-
ter provides data that allow elucidation of the yet unknown
effect of higher energies ͑1 keVрEр20 keV͒ on the evolu-
tion of carbon film properties; specifically, it shows for the
8
and Chabot, carbon-containing energetic species have been
1
–3,7–17
used
for deposition of diamondlike carbon ͑DLC͒
3
films with properties that vary between those of the sp tet-
rahedrally bonded diamond allotrope and the sp trigonally
2
3
first time that complete suppression of carbon sp bonding
bonded graphite allotrope. It has recently been shown that
amorphous carbon films with the density of diamond, as well
occurs as E approaches 20 keV. The new findings are impor-
tant for resolving the mechanisms that govern stabilization
as other diamondlike properties, can be deposited by using
3
ϩ
and suppression of sp bonding in the C ion deposition
ϩ
ions of the appropriate energy range18,19 on room tem-
C
3
process. The data also indicate that the sp suppression ob-
perature substrates. The deposition using hyperthermal spe-
served in many practical systems at much lower
cies is best described by the subplantation model2
0–22
in
14,16,17
energies
results from intrinsic properties of the deposi-
terms of a shallow implantation process. Incorporation of
carbon species in subsurface layers followed by large inter-
nal stresses is believed to be the dominant mechanism con-
tributing to formation of a dense diamondlike phase. Accu-
tion systems employed.
Carbon films ϳ1000 Å thick were deposited onto silicon
͑
100͒ substrates using
a
mass-selected ion beam
18,19
ϩ
instrument.
Direct C ion deposition on clean silicon
ϩ
rate data regarding the variation of film properties with C
was used for the range 5 eVрEр2 keV. For energies of 10
energy is crucial to the understanding of this deposition
ϩ
and 20 keV, the ranges ͑R͒ of the C ions in silicon calcu-
mechanism, to the development of more detailed models,
lated from the TRIM25 program are 309 and 592 Å, respec-
such as proposed by Robertson23 and Davis,24 and to the
tively, compared to Rϭ46 Å for Eϭ1 keV. Therefore, pure
design of deposition processes for specific applications.
The data from different laboratories regarding the effect
1
000 Å thick carbon films on Si substrates cannot be depos-
ϩ
ϩ
ited directly by C ions with Eу10 keV. In order to over-
come this problem, we bombarded a set of ϳ500–1000 Å
thick DLC films ͑previously deposited by C ions with
of C ion energy, which has been used for comparison to
the proposed models,2
3,24
contain some inconsist-
ϩ
1
3,14,16–19
19,22
encies.
Two of our recent works
have estab-
Eϭ10, 120, and 1000 eV, the data of the 120 eV precoated
lished the variation of the carbon film characteristics with
ϩ
films are presented in the present work but similar data were
C energy ͑E͒ in the range 5 eVрEр1 keV. The films were
obtained for the other films͒ with a dose of 1018 C /cm
ϩ
2
deposited using a controlled mass-selected ion beam deposi-
3
tion system and the main features studied were the sp frac-
with Eϭ10 and 20 keV, producing pure carbon layers of
ϳ1000–2000 Å. These latter films represent the combined
effect of the incorporation and collisional damage of 10 and
tion by electron energy loss spectroscopy ͑EELS͒, the sur-
face morphology by atomic force microscopy ͑AFM͒, and
the density by a combination of Rutherford backscattering
spectroscopy ͑RBS͒ and profilometry. It was found that ͑i͒ a
ϩ
20 keV C ions.
Auger electron spectroscopy ͑AES͒ depth profiles for 10
ϩ
keV C ions implanted into both silicon and a 1000 Å thick
DLC film at doses of 1018 cm
Ϫ2
indeed show that implan-
a͒Author to whom correspondence should be addressed. Electronic mail:
wwilberg@weizmann.weizmann.ac.il
tation into clean silicon yields a carbon layer that contains a
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:
214 Appl. Phys. Lett. 68 (9), 26 February 1996 0003-6951/96/68(9)/1214/3/$10.00 © 1996 American Institute of Physics
29.120.242.61 On: Tue, 25 Nov 2014 05:52:01
1
1

sitions while the silicon atoms on the surface are being sput-
tered and diluted.2
0–22
Successive incorporation of C atoms
into subsurface positions of the evolving films leads to a high
internal stress that is conducive to formation of a dense dia-
mondlike phase. The initial smoothness of the silicon sub-
strate is retained due to the internal carbon film growth, even
for films 1000 Å thick. For energies lower than 30 eV, most
of the carbon atoms do not occupy subsurface positions but
are trapped on the surface, leading to rougher, graphitic
films.
ϩ
As the C energy is increased further, the deposition
becomes a deeper implantation process and the system be-
comes more complicated due to several physical phenomena
that can affect the film growth process: ͑i͒ range ͑R͒ and
ϩ
distribution ͑⌬R͒ of the C ions; ͑ii͒ backscattering coeffi-
cient; ͑iii͒ sputtering yield; and ͑iv͒ damage due to atomic
displacements. These quantities can be simply calculated for
a crude assessment of the deposition process using the
25
TRIM code. TRIM, however, does not treat the incorporation
28
of carbon species for which a dynamic version ͑TRIDYN ͒ is
needed. A code that treats the complete deposition process
including diffusion and density gradients does not exist to
ϩ
the best of our knowledge. Our data indicate that for C ions
with EϽ1 keV, processes ͑ii͒, ͑iii͒, and ͑iv͒ appear to have a
3
minimal effect and are not deleterious to sp bond forma-
FIG. 1. Surface roughness ͓from atomic force microscopy ͑AFM͔͒ and
2
sp fractions ͓from electron energy loss spectroscopy ͑EELS͔͒ of 100 nm
tion. The evolution of the carbon layer is affected mainly by
the incorporation of carbon into subsurface sites. As long as
the substrate temperature is kept low enough to freeze the
trapped carbon atoms in their final subsurface positions after
being thermalized, they are incorporated and dense, sp rich,
smooth films evolve. As the energy increases, the C ions
penetrate deeper into the carbon matrix ͑Rϭ26, 44, 180, 350
Å for Eϭ1, 2, 10, 20 keV, respectively͒ and the damage
N ͓number of vacancies per impinging C ion ͑vac/ion͔͒
increases (N ϭ8.3, 15, 52, 83 vac/ion for Eϭ1, 2, 10, 20
ϩ
films deposited using C energies in the range 5 eV to 20 keV on room
temperature silicon substrates. Note that the complete suppression of the
3
sp component is associated with a sudden increase in the surface roughness
at both low and high energy.
3
ϩ
10%–20% concentration of silicon, while implantation into
the DLC film yields a pure carbon layer. All of the films were
analyzed by EELS,26 Raman,
18,27
and AFM;
19,22
in addition,
and the 10
ϩ
v
18,27
the 5 eV–2 keV films were analyzed by RBS
v
and 20 keV films were analyzed by AES depth profiling.
keV, respectively͒. The backscattering and sputtering yields
3
ϩ
The sp fractions of the films were derived from the
for a normal incidence angle of C on carbon remain
26
21
near K edge EELS spectra as previously described and the
small.
surface roughness (R ͒ was derived from the AFM images
Our AFM data show that the mechanism that suppresses
the formation of sp bonding at low energy, i.e., transition
q
2
1/2
3
using R ϭ͓⌺(Z ϪZ ) /N͔ , where Zav is the average of
q
i
av
the Z height values within a given area, Z is the current Z
from subsurface to surface deposition, does not occur at high
energies. Robertson23 and Davis suggested that the stress or
density relief due to dissipation of the excess energy, i.e., a
thermal spike, is the dominant mechanism leading to sup-
i
24
value, and N is the number of points within the given area.
Figure 1 shows plots of the sp2 fraction and Rq for 5
eVрEр20 keV. The plots show that the optimal ion energy
3
2
3
region for a maximal sp fraction ͑minimal sp ͒ is 50
pression of the sp configuration in the high energy region.
2
eVрEр600 eV. The sp fraction increases sharply with de-
Their calculations, however, indicate a sharp decrease of the
2
3
creasing E for Eр30 eV, reaching 90% sp bonding for
sp fraction with energy, in contradiction to our present data.
Eϭ10 eV. In the high energy region, the sp2 fraction in-
creases with increasing E for Eу2 keV, reaching 90% sp2
bonding for Eϭ20 keV. The figure also shows that for both
the low energy and the high energy regions, full suppression
We suggest that the moderate increase of the sp fraction
2
with energy up to Eϭ10 keV is due to the increasingly
broader range and distribution of the subplanted carbon ͑⌬R
ϭ11, 19, 66, 112 Å for Eϭ1, 2, 10, 20 keV, respectively͒,
and the increasing radiation damage by atomic displacements
that is known to graphitize even real diamond surfaces.29
For energies Eр10 keV, the carbon atoms are incorpo-
rated in subsurface positions and the film grown internally.
The film can be roughly considered as consisting of three
distinguished layers with different properties: ͑i͒ a top, de-
fected layer ͑width ϳRϪ⌬R/2) in which the carbon species
are not trapped and only radiation damage occurs; ͑ii͒ an
evolving layer ͑width ⌬R͒ beneath the top layer; and ͑iii͒ a
3
of sp bonding is associated with a sharp increase of the
surface roughness, while the films retain the initial atomi-
cally smooth texture of the silicon substrate over the wide
energy region 50 eVрEр10 keV.
The effect of ion energy on the properties of carbon films
2
0–22
can be addressed in terms of the subplantation model.
ϩ
The C ions bombard a smooth silicon target (Rq
Ͻ0.2 nm) upon which a carbon layer evolves. For energies
greater than 30 eV, the carbon atoms occupy subsurface po-
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:
Appl. Phys. Lett., Vol. 68, No. 9, 26 February 1996 Grossman et al. 1215
29.120.242.61 On: Tue, 25 Nov 2014 05:52:01
1

deeper layer not affected by the impinging carbon species.
As the film incorporates carbon the initial evolving layer
ing are associated with the subsurface growth of atomically
3
smooth films; ͑iv͒ the suppression of sp bonding by high
ϩ
͑layer ii͒ is gradually covered and ͑as a crude approximation͒
energy C ions is due to ballistic effects of the higher energy
becomes part of the nonaffected layer ͑layer iii͒ after the film
thickness has increased by ⌬R. If ion mixing, diffusion, and
sputtering are neglected ͑the sputtering yield of carbon at
normal incidence angles is low͒, the defected layer ͑layer i͒
remains constant. This analysis indicates that for films thick
enough, the defected layer incorporates very high levels of
radiation damage that may lead to graphitization. The dam-
age in the evolving layer ͑layer ii͒ which can be estimated as
implantation and the resulting enhanced diffusion.
This material is based on work that was partially sup-
ported by the Israeli-U. S. BSF Grant No. 92-197, the
Israeli-EC Grant No. 90-2014, the National Science Founda-
tion Grant No. DMR-9224377, the R. A. Welch Foundation
Grant No. E656, and the State of Texas through the Texas
Center for Superconductivity.
(
(
N /R)⌬R thus increases with increasing ion energy
v
1
Thin Films From Free Atoms and Particles, edited by K. J. Klabunde
N /Rϳconts., ⌬R increases with energy͒ and so does the
v
͑
Academic, Orlando, FL, 1985͒.
2
3
total damage in the final film. The defected layer thickness
RϪ⌬R/2) increases with energy but the total damage in
Plasma Deposited Thin Films, edited by J. Mort and F. Jansen ͑CRC,
Boca Raton, FL, 1986͒.
Ion Bombardment Modification of Surfaces: Fundamentals and Applica-
tions, edited by O. Auciello and R. Kelly ͑Elsevier, Amsterdam, 1987͒.
P. J. Martin, J. Mater. Sci. 21, 1 ͑1986͒.
R. A. Zuhr, S. J. Pennycook, T. S. Noggle, N. Herbots, T. E.
Haynes, and B. R. Appelton, Nucl. Instrum. Methods B 37, 16 ͑1989͒.
D. Marton, in Low Energy Ion-Surface Interactions, edited by J. W. Ra-
balais ͑Wiley, London, 1994͒, p. 481.
Properties and Characterization of Amorphous Carbon Films, Mater. Sci.
Forum. 52–53, edited by J. J. Pouch and S. A. Alterovitz ͑Trans. Tech,
Aerdermannsdorf, 1990͒.
(
this layer for a fixed fluence ␾, (N /R)␾, is constant with
v
energy. The moderate sp2 increase with energy is thus
mainly due to the higher damage in the evolving layer and
partially due to the increased thickness of the defected layer,
4
5
6
7
2
which is predominantly sp bonded. The internal growth na-
ture of the carbon films for high E ͑large R͒ maintains the
initial smooth nature of the films unless one of the above
possible mechanisms occurs: ͑i͒ roughening due to sputter-
ing; ͑ii͒ enhanced diffusion of C species to the surface due to
radiation damage; ͑iii͒ deformation of the defected layer due
to radiation damage. The smooth nature of the deposited
films indicates that none of these mechanisms occurs for
Eр10 keV. At Eϭ20 keV, however, a sudden large increase
of the surface roughness is observed. Since the sputtering
yield for Eϭ10 and 20 keV is similar and so is the damage in
the defected layer, it is very likely that a sufficient enhance-
ment of the diffusion due to the increased damage ͓process
8
9
S. Aisenberg and R. Chabot, J. Appl. Phys. 42, 2935 ͑1971͒.
Y. Lifshitz, S. R. Kasi, and J. W. Rabalais, in Properties and Character-
ization of Amorphous Carbon Films, Mater. Sci. Forum. 52–53, edited by
J. J. Pouch and S. A. Alterovitz ͑Trans. Tech., Aerdermannsdorf, 1990͒, p.
237.
10
E. F. Chaikovskii, V. M. Puzikov, and A. V. Semenov, Kristallografiya 26,
19 ͑1981͒ ͑Sov. Phys. Crystallogr. 26, 122 ͑1981͒.
2
11
J. H. Freeman, W. Temple, and G. A. Gard, Vacuum 34, 305 ͑1984͒.
T. Miyazawa, S. Misawa, S. Yoshida, and S. Gouda, J. Appl. Phys. 55, 188
12
͑
1984͒.
13
J. P. Hirvonen, J. Koskinen, R. Lappalainen, and A. Antilla, in Properties
and Characterization of Amorphous Carbon F, Mater. Sci. Forum 52–53,
edited by J. J. Pouch and S. A. Alterovitz ͑Trans. Tech., Aerdermannsdorf,
͑
ii͔͒ leads to the observed surface roughening at Eϭ20 keV.
1990͒, p. 197.
Similar films bombarded with same fluence of 20 keV Ne
ions did not exhibit a significant roughness increase indicat-
ing that the roughening ͑and probably the complete suppres-
sion of the sp fraction͒ is a complex process that involves
diffusion due to both C incorporation and radiation damage.
14
J. Ishikawa, Y. Takeiri, K. Ogawa, and K. Takagi, Nucl. Instrum. Methods
B 21, 205 ͑1987͒.
W. M. Lau, I. Bello, X. Feng, L. J. Huang, Q. Fuguang, Y. Zhenyu, R.
Zhizhang, and S. T. Lee, J. Appl. Phys. 70, 5623 ͑1991͒.
15
3
16
͑
a͒ D. R. McKenzie, D. Muller, and B. Pailthrope, Phys. Rev. Lett. 67, 773
3
͑1991͒. ͑b͒ D. R. McKenzie, Y. Yin, N. A. Marks, C. A. Davis, B. A.
Pailthrope, G. A. J. Amaratunga, and V. S. Veerasamy, Diam. Relat. Mater.
Contradictory data regarding the sp fraction of carbon
ϩ
films deposited at room temperature as a function of C ion
energy have been reported in the literature.
is the sp bonded configuration that is the stable allotrope
and since system perturbations lead to the suppression of the
sp phase, it is evident that a correct ‘‘phase diagram’’ of the
sp fraction versus E should be an envelope of the maximal
sp fractions ͑minimal sp ͒ obtained in different laboratories
for specific energy values. Figure 1 indeed appears to form
3, 353 ͑1994͒.
13,14,16,17
17
Since it
P. J. Fallon, V. S. Veersamy, C. A. Davis, J. Robertson, G. A. J. Amara-
tunga, W. I. Milne, and J. Koskinen, Phys. Rev. B 48, 4777 ͑1993͒.
Y. Lifshitz, G. D. Lempert, S. Rotter, I. Avigal, C. Uzan-Saguy, R. Kalish,
J. Kulik, D. Marton, and J. W. Rabalais, Diam. Relat. Mater. 3, 542
2
18
3
͑
1994͒.
3
19
Y. Lifshitz, G. D. Lempert, I. Avigal, C. Uzan-Saguy, R. Kalish, J. Kulik,
D. Marton, and J. W. Rabalais, Diam. Relat. Mater. 4, 318 ͑1995͒.
Y. Lifshitz, S. R. Kasi, and J. W. Rabalais, Phys. Rev. Lett. 62, 1290
3
2
2
2
2
0
1
2
͑
1989͒.
3
such an envelope. The lower sp fractions obtained by other
Y. Lifshitz, S. R. Kasi, J. W. Rabalais, and W. Eckstein, Phys. Rev. B 41,
10486 ͑1990͒.
Y. Lifshitz, G. D. Lempert, and E. Grossman, Phys. Rev. Lett. 72, 2759
laboratories at specific energy values reflect intrinsic proper-
ties of the deposition systems employed and not necessarily
the effect of kinetic energy alone. Indeed, films with dia-
mondlike properties, i.e., relatively high sp3 fractions for
EϾ1 keV have been reported by at least two other
͑
1994͒.
23
J. Robertson, Diam. Relat. Mater. 2, 984 ͑1993͒.
C. A. Davis, Thin Solid Films 226, 30 ͑1993͒.
J. P. Biersack and L. G. Haggmark, Nucl. Instrum. Methods 174, 257
͑1980͒.
J. Kulik, Y. Lifshitz, G. D. Lempert, D. Marton, and J. W. Rabalais, J.
Appl. Phys. 76, 5063 ͑1994͒.
Y. Lifshitz, G. D. Lempert, S. Rotter, I. Avigal, C. Uzan-Saguy, and R.
Kalish, Diam. Relat. Mater. 2, 285 ͑1993͒.
W. Moller and W. Eckstein, Nucl. Instrum. Methods B 2, 814 ͑1984͒.
R. Kalish, Diam. Relat. Mater. 2, 621 ͑1993͒.
H. Huck, A. Jeck, E. B. Halac, J. Nikolai, M. A. R. de Benyacar, and R.
Righini, Nucl. Instrum. Methods B 84, 62 ͑1994͒.
2
2
4
5
1
3,30
groups.
26
The findings can be summarized as follows: ͑i͒ carbon
3
27
films with significant amounts of sp bonding can be depos-
ited at room temperature over a wide range of 30 eVрEр10
keV C
bonding at Eр10 eV or Eу20 keV is associated with an
increase in film roughness; ͑iii͒ high fractions of sp bond-
2
2
8
9
ϩ
energy; ͑ii͒ the complete suppression of the sp3
30
3
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:
216 Appl. Phys. Lett., Vol. 68, No. 9, 26 February 1996 Grossman et al.
29.120.242.61 On: Tue, 25 Nov 2014 05:52:01
1
1