In situ x-ray diffraction study of the thermal expansion of silver nanoparticles in ambient air and vacuum

Article abstract of DOI:10.1063/1.1901803

The thermal expansion behavior of silver nanoparticles in ambient air and vacuum was studied by in situ x-ray diffraction (XRD) measurement of the particles dispersed within mesoporous silica in the temperature range of 25-700 °C. It has been shown that t

Full text of DOI:10.1063/1.1901803

In situ x-ray diffraction study of the thermal expansion of silver nanoparticles in
ambient air and vacuum
Jinlian Hu, Weiping Cai, Cuncheng Li, Yanjie Gan, and Li Chen
Citation: Applied Physics Letters 86, 151915 (2005); doi: 10.1063/1.1901803
View online: http://dx.doi.org/10.1063/1.1901803
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/86/15?ver=pdfcov
Published by the AIP Publishing
This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
152.14.136.96 On: Tue, 05 Nov 2013 11:47:10

APPLIED PHYSICS LETTERS 86, 151915 ͑2005͒
In situ x-ray diffraction study of the thermal expansion of silver
nanoparticles in ambient air and vacuum
a͒
Jinlian Hu, Weiping Cai, Cuncheng Li, Yanjie Gan, and Li Chen
Key Lab of Materials Physics, Institute of Solid State Physics, Chinese Academy of Sciences, Hefei, 230031,
Anhui, People’s Republic of China
͑
Received 27 August 2004; accepted 1 March 2005; published online 7 April 2005͒
The thermal expansion behavior of silver nanoparticles in ambient air and vacuum was studied by
in situ x-ray diffraction ͑XRD͒ measurement of the particles dispersed within mesoporous silica in
the temperature range of 25–700 °C. It has been shown that thermal expansion coefficient of Ag
−
5
nanoparticles in vacuum is about 0.6ϫ10 /°C, only near one fourth that of bulk silver ͑2.2
−
5
−5
ϫ10 /°C͒. However, the coefficient in air is about 1.7ϫ10 /°C, about 3 times as high as that in
vacuum and close to the value of bulk Ag. These were explained in terms of Ag particles’ surface
energy, oxygen surface adsorption, and dissolution into lattice. This study is of importance in
architectonics of future nanodevices. © 2005 American Institute of Physics.
͓
DOI: 10.1063/1.1901803͔
1
3,14
Ag nanoparticles exhibit many unique chemical and
physical properties, and have applications in many fields,
annealing at 700 °C for 1 h, as previously described.
The
mesoporous silica prepared in this way, whose pores are in-
terconnected and open to ambient air, had the porosity of
about 50%, a specific surface area of 560 m g , and pore
diameters mainly distributed in the range below 20 nm, de-
termined by isothermal N adsorption measurement, as pre-
viously described. The host was then soaked into silver
1
,2
3,4
5,6
such as, optical devices, gas sensors, catalysis, and
7
,8
2
−1
surface-enhanced Raman substrate. Extensive study has
been done in optical properties, catalysis, and environmental
sensitivity. The lattice constant of Ag nanoparticles and its
size dependence have also been studied, such as, by extended
x-ray-absorption fine structure ͑EXAFS͒ technology when
2
1
5
nitrate ͑AgNO ͒ aqueous solution with 0.5M for 5 days, and
taken out for heat treatment in H at 600 °C for 1 h in a
3
9
the particles are in solid argon, by high resolution transmis-
2
sion electron microscope ͑TEM͒ when the particles are em-
quartz tube oven. The Ag nanoparticles are thus in situ
formed within pores of silica mesoporous solid. The loading
amount of Ag in the host was estimated to be 2.4% in weight
from the host porosity and concentration of the soaking so-
lution, without considering the loss of Ag during its
1
0
bedded in glass, and by electron diffraction when the par-
11
ticles on carbon substrate. It has been revealed that
variation of lattice parameters are of importance in the size
evolution of the surface plasmon resonance of Ag
1
2
14
nanoparticles.
preparation.
As we know, nanoparticles could be used for the build-
ing blocks of various future nanodevices. Obviously, the cor-
responding thermal expansion behavior would be very im-
portant for the structural design, thermal stability, and
reliability of such devices. However, the lattice thermal ex-
pansion of silver nanoparticles in different environments has
not been reported so far. Recently, putting Ag nanoparticles
into the pores of silica mesoporous solid, which is a weak
In situ XRD measurements for the treated samples
were performed at different temperatures in ambient air or
vacuum condition on X’Pert Pro MPD diffractometer with a
sample holder, which can be heated and evacuated
−
5
−4
͑10 –10 mbar͒. CuK radiation was used, and Si͑220͒ dif-
␣
fraction line was chosen as a reference line for the calibration
of the diffraction position. Samples were ground into pow-
ders, mixed with a small amount of pure Si powders for
calibration, before XRD measurement. The measurement
temperature range in this study is 25–700 °C. The tempera-
ture control is within Ϯ1 °C. Temperatures were kept con-
stant for 30 min before each measurement.
1
interacting medium to Ag, and whose pores are intercon-
nected and open to ambient air, we have investigated the
lattice thermal expansion of the quasi-free silver nanopar-
ticles in air and vacuum by in situ x-ray diffraction ͑XRD͒
measurement, and found that the thermal expansion for the
quasi-free silver nanoparticles is much smaller in vacuum
than that in air, and both cases are smaller than that of bulk
Ag. This study is of importance not only in fundamental
academic interest, but also in architectonics of future nan-
odevices since the nanoparticles could be used as building
blocks. The details are reported in this letter.
The existence of loaded silver particles inside pores of
silica has been confirmed by the fact that the isothermal ni-
trogen sorption for H heat-treated soaked sample is lower
2
1
5
than that for SiO host, as previously illustrated. Ag par-
2
ticles dispersed in silica are approximately spherical in
shape, and of about 5 nm in mean size observed by TEM.
Figure 1 shows the in situ measured XRD spectra of the
Ag-loaded sample at different temperatures in ambient air
and vacuum conditions, respectively. It can be seen that the
diffraction peak of the Ag͑111͒ for both cases shifts to
smaller angle with increase of temperature, indicating the
lattice thermal expansion. The lattice parameter of Ag nano-
particles as a function of temperature for both cases is shown
in Fig. 2, in which the corresponding data of bulk Ag is also
The monolithic mesoporous silica host ͑planar-like,
about 1.5 mm in thickness͒ was first prepared by a sol–gel
technique with precursors: tetraethylorthosilicate, water, al-
cohol ͑catalyzed by HNO ͒, followed by drying, and finally
3
a͒Author to whom correspondence should be addressed; electronic mail:
wpcai@issp.ac.cn
This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
003-6951/2005/86͑15͒/151915/3/$22.50 86, 151915-1 © 2005 American Institute of Physics
52.14.136.96 On: Tue, 05 Nov 2013 11:47:10
0
1

151915-2
Hu et al.
Appl. Phys. Lett. 86, 151915 ͑2005͒
FIG. 3. XRD patterns of silver nanoparticles ͑a͒ before and ͑b͒ after expo-
sure to ambient air for 50 min at 700 °C. Curve ͑c͒: the result after evacu-
ation for 50 min at 700 °C for sample ͑b͒.
which are 1.7ϫ10−5 and 0.6ϫ10 /°C for Ag nanoparticles
in ambient air and vacuum, respectively, in the temperature
range of 25–700 °C. The value in ambient air is 3 times as
high as that in vacuum case. Both cases are lower than that
−5
−
5
of bulk Ag which is about 2.2ϫ10 /°C in the same tem-
1
6
perature range.
In addition, if the environment of sample changes from
vacuum to ambient air at the same temperature, the diffrac-
tion peak position will shift significantly, especially at high
temperature. Figure 3 illustrates the results measured before
and after exposure to ambient air for 50 min at 700 °C. We
can see that the diffraction peak of the Ag͑111͒ obviously
shifts to a smaller angle when the environment of sample
changed from vacuum to ambient air. If subsequently ob-
served under vacuum again ͑20 min or longer after evacua-
tion͒, the peak moves back to a larger angle but cannot reach
its original position, as shown in curve ͑c͒ of Fig. 3.
Now let us make a brief discussion. As we know, with
decrease of particle size, the surface effect will get more and
more significant. Obviously, the surface tension of a particle
will induce variation of lattice constants and thermal expan-
sion coefficient. Furthermore, any factor, leading to change
of the surface energy of particles, should induce variation of
lattice constants and thermal expansion coefficient, such as,
adsorption of molecules on the particles during environmen-
tal exposure.
FIG. 1. The XRD patterns of silver nanoparticles in ͑a͒ ambient air ͑arrows:
denotes the diffraction peaks from sample holder͒ and ͑b͒ in vacuum.
included. In the studied temperature range, the lattice con-
stant of Ag nanoparticles in air is larger than that in vacuum
but smaller than that of bulk at the same temperature, and the
differences between them increase with rising temperature. It
can be seen from Fig. 2 that the evolution of lattice constant
shows roughly linear relation with increasing temperature.
We can thus obtain the lattice thermal expansion coefficients,
As mentioned above, the lattice constant and lattice ther-
mal expansion coefficient of silver nanoparticles in vacuum
are less than those of bulk silver. This could be attributed to
the surface tension of Ag particles. It is well known that the
hydrostatic pressure P within a spherical particle, induced by
1
7
the surface tension, can be expressed as P= 2␥ R, where ␥
ր
and R are the surface tension and radius of the particle, re-
spectively. Because the compressibility of solid is defined by
␬
=−⌬V/PV, for a spherical particle with cubic crystal struc-
1
8
ture, we have the relation
⌬
a
1
2␬␥
3R
=
− ␬P = −
,
͑1͒
a
3
where V is the volume and ⌬a is the change in lattice con-
stant a due to the surface tension. So the surface tension will
induce lattice contraction. Obviously, when R is small
FIG. 2. The plots of silver lattice parameter, calculated from Ag ͑111͒ dif-
This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
fraction in Fig. 2, vs temperature. Solid line: linear regression.
enough, P value will be large sufficient to induce detectable
1
52.14.136.96 On: Tue, 05 Nov 2013 11:47:10

151915-3
Hu et al.
Appl. Phys. Lett. 86, 151915 ͑2005͒
lattice contraction. Furthermore, the pressure dependence of
thermal expansion coefficient can be described by Anderson
relation
oxygen should be relatively negligible at 700 °C͒. In a word,
both oxygen adsorption on the surface and dissolution in the
lattice induce lattice expansion compared with the particles
in vacuum. Obviously, the former should be dominant at
lower temperature, and at higher temperature the latter is
dominant.
1
9–21
␦
T͑T,P͒
␤
͑T,P͒
͑T,0͒
V͑T,P͒
V ͑T,0͒
0
=
ͫ ͬ
,
͑2͒
␤
As for the possible effect from the silica support, it
should be negligible compared to the oxygen atmosphere and
the particles’ surface tension, because the silica is a weak
where ␤͑T,P͒=͑1/V͒͑ץV/ץT͒ is the volume thermal expan-
P
sion coefficient, ␦ ͑T,P͒=−͑1/␤B ͒͑ץB /ץT͒ is the so-
T
T
T
P
called Anderson–Gruneisen parameter and Ͼ0, B is isother-
T
1
interacting medium with Ag. Although the thermal expan-
mal bulk modulus, T is the absolute temperature, and
sion of free-standing Ag nanoparticles cannot be measured
by XRD, we believe that the results would be very similar to
those of Ag nanoparticles dispersed within silica.
V͑T,P͒=V ͑T,0͒+⌬V. Combining Eq. ͑1͒, we can rewrite
0
Eq. ͑2͒ as
␦
T͑T,P͒
2
␥␬
␦
T͑T͒
In summary, we have reported a study of thermal expan-
sion behavior for Ag nanoparticles in air and vacuum in the
temperature range of 25–700 °C, by dispersion of the par-
ticles within pores of mesoporous silica and in situ XRD
measurement. The thermal expansion coefficient of Ag nano-
particles in vacuum is much smaller than that of bulk Ag,
only about one fourth of the latter, due to surface effect of
the particles. However, the coefficient in air is about 3 times
as high as that in vacuum and close to the value of bulk Ag
because of oxygen adsorption on particles’ surface and dis-
solution into the lattice. This study could be of importance
not only in fundamental academic interest, but also in the
structural design, thermal stability, and reliability of future
nanodevices.
␤
͑T,P͒ = ␤͑T,0͓͒1 − ␬P͔
= ␤͑T,0͒
ͫ
1 −
ͬ
.
R
͑
3͒
Here, ␤͑T,0͒ corresponds to the value of bulk silver, while
␤
͑T,P͒ to that of silver nanoparticles with surface effect in
vacuum. Since the data of the parameters ␥, ␬, and ␦ for Ag
nanoparticles are not available, which are also size and tem-
2
1
perature dependent, we cannot quantitatively estimate the
thermal expansion coefficient. Qualitatively, however, from
Eq. ͑3͒, we can know the coefficient of lattice thermal ex-
pansion for silver nanoparticles in vacuum is smaller than
that of bulk silver due to surface tension.
When Ag nanoparticles are exposed to ambient air, gas
molecules will adsorb onto the surface of the particles. For
instance: at room temperature, oxygen O in contact with
silver can undergo the two following reactions:
This study was supported by National Natural Science
Foundation of China ͑Grants No. 50271069 and No.
10474099͒.
2
2
2
+
−
Ag + O → Ag O
,
2
2͑ads͒
1
U. Kreibig and M. Vollmer, Optical Properties of Metal Clusters
͑Springer, New York, 1995͒.
R. D. Averitt, D. Sarkar, and N. J. Halas, Phys. Rev. Lett. 78, 4217
͑1997͒.
H. Bi, W. Cai, and L. Zhang, Appl. Phys. Lett. 81, 5222 ͑2002͒.
Y. Gan, W. Cai, G. Fu, and J. Hu, J. Phys.: Condens. Matter 16, L201
2004͒.
R. L. Davies and S. F. Etris, Catal. Today 36, 107 ͑1997͒.
I. E. Wachs and R. J. Madix, Surf. Sci. 76, 531 ͑1978͒.
M. Fleischmann, P. J. Hendra, and A. J. McQuillian, Chem. Phys. Lett.
26, 163 ͑1974͒.
H. Nakai and H. Nakatsuji, J. Chem. Phys. 103, 2286 ͑1995͒.
P. A. Montano, W. Schulze, B. Tesche, G. K. Shenoy, and T. I. Morrison,
Phys. Rev. B 30, 672 ͑1984͒.
+
2−
ads
4
Ag + O → 4Ag + 2O
,
͑4͒
2
2
i.e., chemisorption of O in air will occur on the surface of
2
3
4
metal Ag. The Ag nanoparticles dispersed within pores of
mesoporous silica are in contact with ambient air and hence
with the oxygen in air. It is expected that the chemisorption
will take place at lower temperature. Obviously, oxygen ad-
sorption will decrease surface energy of Ag nanoparticles
significantly. Such decrease will induce lattice expansion and
rise of thermal expansion coefficient relative to the particles
in vacuum based on Eqs. ͑1͒ and ͑3͒, which is qualitatively
in agreement with our experimental results ͑see Fig. 2͒.
However, the adsorbed oxygen will desorb at higher tem-
perature. On the other hand, when Ag is exposed to ambient
air, oxygen in air will dissolve into the Ag lattice, especially
at high temperature, because the solubility will increase with
5͑
6
7
8
9
10
H. Hofmeister, S. Thiel, M. Dubiel, and E. Schurig, Appl. Phys. Lett. 70,
694 ͑1997͒.
1
11
H. J. Wasserman and J. S. Vermaak, Surf. Sci. 22, 164 ͑1970͒.
W. Cai, H. Hofmeister, and M. Dubiel, Eur. Phys. J. D 13, 245 ͑2001͒.
W. Cai and L. Zhang, J. Phys.: Condens. Matter 8, L591 ͑1996͒.
W. Cai and L. Zhang, J. Phys.: Condens. Matter 9, 7257 ͑1997͒.
W. Cai, L. Zhang, H. Zhong, and G. He, J. Mater. Res. 13, 2888 ͑1998͒.
Smithells Metals Reference Book, 6th ed, edited by E. A. Brandes ͑Butter-
worth & Co. Ltd., London, 1983͒, p. 14-5.
1
1
1
1
2
3
4
5
2
3
24
rising temperature. Outlaw et al. found that atomic oxy-
gen can dissolve preferentially into the octahedral intersti-
16
2
5
tials of the silver lattice. Nagy et al. in situ measured XRD
spectra of bulk silver in pure oxygen and ethanol–argon mix-
ture gas, respectively, and found that exposure to oxygen
induces lattice expansion, which is caused by oxygen disso-
lution in the lattice, while exposure to the ethanol–argon
mixture shows no detectable change of lattice constant. In
our case, oxygen in air will enter into the lattice of Ag nano-
particles at high temperature and induce additional lattice
expansion. The oxygen dissolved in Ag lattice is difficult to
escape in subsequent treatments in vacuum, and hence the
diffraction peak cannot return to the original position ͓see
curves ͑a͒ and ͑c͒ in Fig. 3͔ ͑the chemical adsorption of
1
1
7
8
J. D. Verhoeven, Fundamentals of Physical Metallurgy ͑Wiley, New York,
1
975͒.
C. W. Mays, J. S. Vermaak, and D. Kuhlmann-Wilsdorf, Surf. Sci. 12, 134
19͑
1968͒.
O. L. Anderson, Phys. Rev. 144, 553 ͑1966͒.
2
2
2
2
2
0
1
2
3
4
A. Fernandez Guillermet, J. Phys. Chem. Solids 47, 605 ͑1986͒.
O. L. Anderson, J. Geophys. Res. 72, 3661 ͑1967͒.
P. A. Kilty and W. M. Sachtler, Catal. Rev. - Sci. Eng. 10, 1 ͑1974͒.
Reference 16, p. 12-19.
R. A. Outlaw and M. R. Davidsion, J. Vac. Sci. Technol. A 12, 854
5͑
1994͒.
2
A. J. Nagy, G. Mestl, D. Herein, G. Weinberg, E. Kitzelmann, and R.
Schlogl, J. Catal. 182, 417 ͑1999͒.
This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
152.14.136.96 On: Tue, 05 Nov 2013 11:47:10