Characterization of gold nanoclusters synthesized on carbon nanotubes film: Evaluation of the size distributions by means of X-ray photoelectron spectroscopy

Article abstract of DOI:10.1016/j.susc.2009.12.010

Gold nanoclusters were directly synthesized on thiol functionalized carbon nanotubes film and characterized by means of X-ray photoelectron spectroscopy. A carbon nanotube (CNT) supporting network was produced by spreading a concentrate suspension of thiol-functionalized CNT on platinum films. To synthesize gold nanoclusters, a water solution of tetrachloroauric acid was adsorbed on the CNT substrate and reduced by UV reduction in air. Detailed analysis of the Au 4f core line enabled us to follow the chemical modifications occurring on the substrate. In particular, the XPS analysis of gold features shows a progressive reduction of the gold precursor by increasing the irradiation time. Also information on the nanocluster size distribution after each reducing treatments are obtained.

Full text of DOI:10.1016/j.susc.2009.12.010

Surface Science 604 (2010) 508–512
Contents lists available at ScienceDirect
Surface Science
journal homepage: www.elsevier.com/locate/susc
Characterization of gold nanoclusters synthesized on carbon nanotubes film:
Evaluation of the size distributions by means of X-ray photoelectron spectroscopy
L. Minati a, , S. Torrengo a,b, G. Speranza a
*
a FBK-IRST Sommarive str. 18, 38050 Povo (TN), Italy
b Physics Dep. University of Trento, 38050 Povo (TN), Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 31 July 2009
Accepted for publication 14 December 2009
Available online 21 December 2009
Gold nanoclusters were directly synthesized on thiol functionalized carbon nanotubes film and charac-
terized by means of X-ray photoelectron spectroscopy. A carbon nanotube (CNT) supporting network
was produced by spreading a concentrate suspension of thiol-functionalized CNT on platinum films. To
synthesize gold nanoclusters, a water solution of tetrachloroauric acid was adsorbed on the CNT sub-
strate and reduced by UV reduction in air. Detailed analysis of the Au 4f core line enabled us to follow
the chemical modifications occurring on the substrate. In particular, the XPS analysis of gold features
shows a progressive reduction of the gold precursor by increasing the irradiation time. Also information
on the nanocluster size distribution after each reducing treatments are obtained.
Keywords:
Photoelectron spectroscopy
Carbon nanotubes
Gold nanoparticles
Adsorption
Ó 2009 Elsevier B.V. All rights reserved.
1. Introduction
essential to investigate the interactions between these two compo-
nents. This enables an optimal configuration and tailoring of the
Carbon nanotubes (CNT) are gaining even more importance in
the last years thanks to the unique electronic, optical and mechan-
ical properties. The possibility to easily functionalize both the CNT
walls and the tip with specific reaction pathways has further in-
creased the interest for the potential applications in sensor devices
[1], in fuel cell [2] and in electronic devices [3,4]. One of the most
promising applications of CNT network is its use as a support for
metal nanoparticles in catalysis [5,6]. On the other side, gold nano-
particles since a long time are motive of intense research because
of their special optical properties [7] and electrochemical activity
[8]. Interestingly, the kind of functional groups like thiol, aminic,
carboxylic etc. which may be obtained on CNTs, may be exploited
to bond metal nanoparticles. The hybrid nanostructures obtained
combining these two kinds of materials may, in fact, improve the
performances of the single components. Combination of CNTs
and metal nanoparticles has been exploited for applications in bio-
medical devices, biosensors, optical devices [9,10].
functional hybrids for specific applications. In this paper we pro-
pose the use of X-ray photoelectron spectroscopy to fully charac-
terize gold nanoclusters formed on thiol-functionalized CNT films
deposited on a platinum substrate. It is well known from literature
that XPS is a powerful technique to obtain information on the sur-
face chemistry of the samples. On the other hand a careful analysis
on the Au 4f core lines can also give important information about
the dimensions of the nanoclusters [13–15]. This was demon-
strated in a previous work [16] where we showed that the size dis-
tributions can be obtained for gold nanoclusters supported on
amorphous carbon deposited by magnetron sputtering [17,18].
The calculated size distributions obtained were in good agreement
with the X-ray diffraction single line profile analysis [16] as well as
transmission electron microscopy analysis (data not published).
Because of the great similarity between the two systems we use
the same algorithm to estimate the size distributions of gold
nanoclusters grown on carbon nanotubes film.
Several approaches have been explored. For example Zanella
and co-workers [11] used aliphatic bifunctional thiol functional-
ized SWCNTs to bind gold nanoparticles on the CNT surface. Chen
et al. grew metallic nanoclusters on carbon nanotubes by thermal
decomposition of metal salts [12].
Here the same algorithm to estimate the gold nanocluster size
distribution was used because the high similarity between the
two supporting networks.
To fully exploit the potential of CNT-gold nanoparticles hybrids
as building blocks for functional nanoscale architectures, it is
2. Experimental section
Multi-walled carbon nanotubes (p-CNT) (Nanoamor) were cut
by acid in a (1:3) HNO3/H2SO4 mixture, followed by ultrasonication
for 6 h at 40 °C.
* Corresponding author.
E-mail address: luminati@fbk.eu (L. Minati).
0039-6028/$ - see front matter Ó 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.susc.2009.12.010

After the termination of the reaction, the CNT suspension was
NH2(CH2)2SH
L. Minati et al. / Surface Science 604 (2010) 508–512
HNO3/H2SO4
509
p-CNT
o-CNT
CNT-SH
diluted with water and let overnight for precipitation. The super-
natant was removed and the suspension was filtered under vac-
6 h sonicated
dicyclohexylcarbodiimide
Deposition on
Pt substrate
uum with
a 1 lm filter membrane (o-CNT). The carboxyl-
terminated nanotubes were thiol-functionalized by reacting with
2-aminoethanethiol (Aldrich USA) with the aid of DCC (dicyclohex-
ylcarbodiimide, Aldrich, USA) condensation agent. This is accom-
plished suspending 0.3 mg/ml functionalized carbon nanotubes
in ethanol with the aid of ultrasounds. An excess (referred to the
calculated amounts of carboxylic groups present on the CNT) of
DCC ethanolic solution (2.5 mM) was then added and the relative
suspension was stirred for some minutes. Finally thiol functional-
ities were obtained with an addition of a 5 mM 2-aminoethane-
thiol solution and stirring for 24 h at room temperature.
The resulting thiol functionalized carbon nanotubes (CNT-SH)
are not perfectly stable in ethanol and tend to slowly precipitate
after the reaction. The precipitate obtained 24 h after the reaction
termination was separated by filtration, washed to remove the
unreacted molecules and the by-products, then re-dispersed in
ethanol by prolonged sonication (30 min). The efficiency of the
reactions was followed by means of XPS and IR spectroscopy
[19]. Drops of highly concentrated ethanol suspension of function-
alized carbon nanotubes were spreads on a platinum substrate
forming a black uniform film and dried using a pure nitrogen flux.
To synthesize gold NPs on CNT, the sample was immersed in a
1 mM HAuCl4 (Aldrich USA) water solution for 4 h. Excess of HAuCl4
was washed out in deionized water and the sample was analyzed
with XPS (sample CNTAuref). The sample was then irradiated with
UV light in air using a Deuterium-UV lamp, 25 W with kirr of
160 nm for 1 min (CNTAu1) and 5 min (CNTAu5). The resulting
samples were analyzed again by XPS (see Scheme 1) to estimated
the efficiency of the UV radiation to cause Au reduction.
UV reduction
HAuCl4
CNTAu1-5
CNTAuref
Pt/CNT-SH
Scheme 1. Synthesis of CNTAuref, CNTAu1 and CNTAu5 from the pristine carbon
nanotubes.
In Fig. 2 the Au4f core lines of the three samples are reported.
The straight line represents the position of pure bulk gold refer-
ence. The XPS analysis of the three core lines put in evidence the
marked differences between the samples. In particular the posi-
tions of the Au4f7/2 signal assigned to the Au0 peak present a bind-
ing energy sensitively higher respect to the 84 eV value of bulk
gold (87.6 for the 4f5/2 component). This is an indication of the
presence of very small gold nanoclusters in the samples due to
the presence of metallic nuclei that acts as centers for the succes-
sive grow of the gold nanoparticles. Shifts to higher binding energy
of the core lines obtained from metal nanoclusters, with dimen-
sions lower than 5 nm, with respect to the bulk value are widely
reported in the literature. [13,14] For example shift of around
1 eV in the Au 4f7/2 signal, as in the case of the sample CNTAuref,
are associated to the presence of gold nanoclusters with dimen-
sions around 1 nm [15].
The Au 4f core lines of the CNTAuref, CNTAu1 and CNTAu5 sam-
ples are reported in Fig. 3.
The presence of spin orbit interaction leads to a splitting of the
gold core line in a doublet. The correspondent components are
indicated with Au 4f5/2 and Au 4f7/2. All the changes in the chem-
ical state of gold will be reflected by identical changes in both the
spin orbit components of the gold core line. The presence of spin
orbit splitting leads to a doubling of the components in the peak
fitting.
The first two samples show the presence of at least three gold
chemical species. From data obtained from literature the BE of AuIII
Au 4f7/2 is expected to be around 87–88 eV [20] (90–91 for the Au
4f5/2 component). For AuI the Au 4f7/2 is expected around 86 eV
(89.6 for the 4f5/2 component)1. Finally, the component at ꢀ85 eV
(88.6 for the 4f5/2 component) is assigned to Au0. The quantitative
analysis shows that a considerable fraction of the gold precursor is
reduced in metallic form already before starting the UV reducing
treatment. This could be induced by a reduction process initiated
by the thiol groups of the CNT support that could readily explain
the presence of oxidized sulphur in the sample.
The XPS measurements were carried out using a monochromat-
ed Al K
a (1486.6 eV) X-ray source ESCA200 instrument (Scienta-
Gammadata ESCA 200 Uppsala Sweden). Wide scans were acquired
in the binding energy (BE) range 1200–0 eV using a 500 eV pass en-
ergy while high resolution core line spectra were performed set-
ting the analyzer pass energy at 150 eV and the energy step at
0.05 eV. Because the samples are conductive they do not require
charge compensation, leading to an energy resolution of 0.3 eV.
3. Results and discussion
The XPS analysis of the CNTAuref sample is presented in Fig. 1.
Survey spectrum (not reported) shows a little signal from the plat-
inum substrate, indication of a good coverage of the CNT film.
The C1s core line is fitted with five components, assigned to sp2
bonds (284.4 eV), sp3 bonds (285.3 eV), C–O bonds (286.9 eV),
amidic bonds (288.4 eV) and carboxylic bonds (289.5 eV). The
O1s core line is fitted with four components, the peak at
531.0 eV assigned to the carbonyl oxygens of carboxyl and amidic
bond, the component at 532.5 eV assigned to C–O bonds, that at
533.5 eV assigned to the unreacted oxydrilic oxygens in carboxylic
bonds and finally the component at 535.5 eV is assigned to water
molecules adsorbed on the sample.
The S2p core line shows the presence of oxidized sulphur (1),
probably formed by oxidative processes during the HAuCl4 adsorp-
tion. The doublet at around 163.4 and 165 eV (2) is referred to free
thiol groups present on the CNT surface. The N1s core line is cen-
tered at 399.9 eV in agreement with amidic bond position reported
in the literature1. The quantitative analysis of each element is re-
ported in Table 1.
In literature it is described that X, c, and UV radiations were
used to induce a reductions of AuIII ions to Au0 [21,22]. For example
Pucci et al. reported the reduction of HAuCl4 molecules adsorbed
on dried vinyl alcohol-containing polymers [23].
The reduction mechanism proposed was a photo-reduction pro-
cess accelerated by the presence of
a hydrogens containing alco-
holic groups present in the polymer matrix that act as radical
scavengers. Following the exhaustive review of Vogler, [24] the
gold reduction occurs by formations of chlorine radicals that suc-
cessively form Cl2 molecules with the formations of AuI that under-
goes a further disproportionation with the formation of Au0 and
AuIII. It must be noted that the carbon nanotubes film contain a
great amount of –OH, –COOH and water molecules that can accel-
erate these reactions.
The UV treatments applied to the CNTAuref sample leads to a
progressive disappearance of the AuIII and AuI components due to
the reduction of the gold precursors and intermediate compounds.
At the same time a visible shift to lower BE of the Au 4f7/2 maxi-
mum is obtained (see Fig. 2 samples CNTAu1, CNTAu5).
1
See NIST X-ray photoelectron spectroscopy database http://srdata.nist.gov/xps/
for the assignment of the chemical bonds.

510
L. Minati et al. / Surface Science 604 (2010) 508–512
Fig. 1. XPS core lines of the CNTAuref sample; C1s, O1s, S2p, and N1s core lines.
According to previous works [16] this reveals that the reducing
treatment leads to an increase of the nanoclusters dimensions
showing a progressive growth of the gold NP.
distribution centered at around 1 nm. The very low dimensions
and narrow distributions of the gold nanoclusters of this sample
can be explained with an initial formation of metal nuclei com-
posed by a low number of atoms that acts as seeds for the succes-
sive grow of the nanoparticles.
After 1 min of irradiation the size distribution becomes broader
and centered to a higher nanoparticles dimension (see Fig. 3). Cor-
respondingly, the amount of oxidized gold decreases. A further in-
crease of UV irradiation to 5 min leads to a total reduction of gold.
Differently from CNTAuref and CNTAu1, the CNTAu5 sample do
not show spectral components in the BE range between 90 and
93 eV, indications of the total reductions of the metal precursor,
confirmed also by the disappear of the chlorine component (data
not show) in agreement with the mentioned reduction mechanism.
In these three samples the Au 4f7/2 is shifted respect to the pure
bulk Au of about 1 eV, 0.5 eV and 0.3 eV respectively. In order to
correlate the BE change to a specific nanoparticle size distribution
we extrapolate the Au0 components from the core line peak fitting.
For the size distribution calculation we use the same procedure
presented in a previous paper [16]. The results of the core line anal-
ysis are displayed in Fig. 4.
The distributions were estimated by using lognormal functions
commonly used to describe nanoparticles populations [25]. As ap-
pears from Fig. 4, the CNTAuref sample shows a very narrow size
Table 1
Quantitative analysis (atomic abundance) obtained from the XPS analysis on the
CNTAuref sample.
Core lines
Area
58.02
281.74
33.95
69.50
58.23
80.02
CNTAuref (at.%)
Cl2p
C1s
S2p
O1s
N1s
Au4f
4.7
71.3
4.6
7.1
9.9
2.4
Fig. 2. Au 4f core lines of the three samples with different UV exposing times.

L. Minati et al. / Surface Science 604 (2010) 508–512
511
Fig. 3. Au 4f core line fitting of the CNTAuref (A), CNTAu1 (B) and CNTAu5 (C) samples.
In XPS spectra this corresponds to the disappearance of the AuIII
and AuI components and to the formation of a broader NP distribu-
tion (see Fig. 3). For asymmetric distributions, as the lognormal
used in our work, the mean values of the NP size does not corre-
spond to the curve maximum. For a lognormal distribution the
mean value of the NP size may be estimated using the expression
Au0 fraction present in the sample (left axes) as a function of the
irradiation time.
As it appears, the average dimensions of the gold nanoparticles
are very low respected to other works reported in the literature
[26,22]. This may be explained with the softer reduction strength
of the process used. Moreover the high concentration of functional
groups and defects on the CNT can act as ‘‘cages” preventing the
growth of big Au nanoparticles.
l+r
M = e /2. The calculated mean nanoparticles size for the three
samples are reported in Fig. 5 (right axes) as well as the estimated
4. Conclusions
3.5
In this work high resolution XPS analysis was utilized to de-
scribe the formation of Au nanocluster on CNT. Through a careful
core line analysis we were able to fully characterize the chemical
states of the constitutive elements. Core line modeling enabled
us to obtain information on the nanoparticles size distribution
and their modifications as function of the UV irradiation. The pos-
sibility to follow the reductions of the precursor from the first stage
of the reaction is fundamental to describe the kinetics of reaction.
As a consequence, a clearer knowledge about the Au nanoparticles
synthesis allows tailoring specific NP population for specific appli-
cations. Finally this analytical method is very useful for systems
that cannot be characterized by the traditional techniques as trans-
mission electron microscopy. Examples of this are patterned sub-
strates containing small nanoparticles, carbon nanotubes forests
functionalized with small metal nanoparticles or other systems
with well ordered structures grown on a substrate. In all these
cases the XPS technique allows a direct analysis of the samples
without complex or destructive preparation procedures that could
modify the chemistry and structure of the final product.
CNTAuref
CNTAu1
3.0
2.5
CNTAu5
2.0
1.5
1.0
0.5
0.0
0
1
2
3
4
Diameter (nm)
Fig. 4. Calculated size distributions from the Au4f core line analysis for the
CNTAuref, CNTAu1 and CNTAu5 samples.
2.0
Au0 conc.
Mean dimension
References
100
1.8
[1] Y. Lu, J. Li, J. Han, T. Ng, C. Binder, C. Patridge, M. Meyyapan, Chem. Phys. Lett.
391 (2004) 344.
[2] X. Sun, R. Li, D. Villers, J.P. Dodelet, S. Desilets, Chem. Phys. Lett. 379 (2003) 99.
[3] A.G. Rinzler, J.H. Hafner, P. Nikolaev, L. Lou, S.G. Kim, D. Tomanek, P.
Nordlander, D.T. Colbert, R.E. Smalley, Science 269 (1995) 1550.
[4] Y. Saito, K. Hamaguchi, T. Nishino, K. Hata, K. Tohji, A. Kasuya, Y. Nishina, Jpn. J.
Appl. Phys. 36 (1997) L1340.
[5] G.L. Che, B.B. Lakshmi, C.R. Martin, E.R. Fisher, Langmuir 15 (1999) 750.
[6] H. Fang Cui, J.S. Ye, W.D. Zhang, J. Wang, F.S. Sheu, J. Electroanal. Chem. 577
(2005) 295.
1.6
80
1.4
60
1.2
[7] Y. Dirix, C. Bastiaansen, W. Caseri, P. Smith, Adv. Mater. 11 (1999) 223.
[8] M. Haruta, N. Yamada, T. Kobayashi, S. Ijima, J. Catal. 115 (1989) 301.
[9] A.C. Templeton, W.P. Wuelfing, R.W. Murray, Acc. Chem. Res. 33 (2000) 27.
´
´
[10] T. Sainsbury, T. Ikuno, D. Okawa, D. Pacile, J.M.J. Frechet, A. Zettl, J. Phys. Chem.
C 111 (2007) 12992.
1.0
40
0
2
4
6
[11] R. Zanella, E.V. Basiuk, P. Santiago, V.A. Basiuk, E. Mireles, I. Puente-Lee, J.M.
Saniger, J. Phys. Chem. B 109 (2005) 16290.
Irradiation times (minutes)
[12] B. Xue, P. Chen, Q. Hong, J. Lin, K.L. Tan, J. Mater. Chem. 11 (2001) 2378.
[13] G.K. Wertheim, S.B. DiCenzo, Phys. Rev. B 37 (1988) 844.
[14] M.G. Mason, Phys. Rev. B 27 (1983) 748.
Fig. 5. Calculated mean nanoparticles size (right axes) and the total Au0 concen-
tration (left axes) plotted as a function of the UV exposing time.

512
L. Minati et al. / Surface Science 604 (2010) 508–512
[15] H.-G. Boyen, A. Ethirajan, G. Kastle, F. Weigl, P. Ziemann, G. Schmid, M.G.
Garnier, M. Buttner, P. Oelhafen, Phys. Rew Lett. 94 (2005) 016804.
[16] L. Minati, G. Speranza, L. Calliari, V. Micheli, A. Baranov, S. Fanchenko, J. Phys.
Chem. A 112 (2008) 7856.
[20] S. Suzer, Appl. Spectrosc. 54 (2000) 1716.
[21] E. Ozkaraoglu, I. Tunc, S. Suzer, Surf. Coat. Technol. 201 (2007) 8202.
[22] A.M. Showkat, K.-P. Lee, A.I. Gopalan, S.-H. Choi, Y.C. Nho, Diam. Relat. Mater.
16 (2007) 1688.
[17] A. Baranov, S. Fanchenko, L. Calliari, G. Speranza, L. Minati, A.V. Shorokhov, D.
Fedoseenkov, A. Nefedov, Diam. Relat. Mater. 16 (2007) 1365.
[18] L. Calliari, G. Speranza, L. Minati, V. Micheli, A. Baranov, S. Fanchenko, Appl.
Surf. Sci. 255 (2008) 2214.
[19] L. Minati, G. Speranza, S. Torrengo, L. Toniutti, C. Migliaresi, D. Maniglio, M.
Ferrari, A. Chiasera (unpublished results).
[23] A. Pucci, M. Bernabo, P. Elvati, L.I. Meza, F. Galembeck, C.A.P. Leite, N. Tirelli, G.
Ruggeri, J. Mater. Chem. 16 (2006) 1058.
[24] A. Vogler, H. Kunkely, Coord. Chem. Rev. 219–221 (2001) 489.
[25] T. Jason, C. Mangesh, T. Bore, A.K. Datye, R.J. Davis, J. Catal. 238 (2006) 458.
[26] R. Zhang, M. Hummelgård, H. Olin, Mater. Sci. Eng., B 158 (2009) 48.