Investigation of silver-glass nanocomposites by positron lifetime spectroscopy

Article abstract of DOI:10.1088/0953-8984/8/30/014

Nanocrystalline silver particles were grown in a glass medium by ion-exchange and reduction techniques and studied by positron lifetime spectroscopy. The particle sizes varied from 5 to about 25 nm as observed by transmission electron microscopy. The positron lifetime spectra of all the samples could be decomposed into three components having lifetimes of around 160 ps, 400 ps and 1500 ps. The first is ascribed to positron annihilation at the interfaces of the nanocrystalline silver and the glass matrix, and it decreases and stabilizes as the silver grain size increases. The second component is explained as arising from positrons trapped and annihilated at the free-volume defects in the glass matrix. The third component arises because of the annihilation of orthopositronium at large free-volume defects. The effects of temperature on the interfacial defects and the processes leading to the formation of additional positron trapping centres are discussed.

Full text of DOI:10.1088/0953-8984/8/30/014

Home Search Collections Journals About Contact us My IOPscience
Investigation of silver - glass nanocomposites by positron lifetime spectroscopy
This content has been downloaded from IOPscience. Please scroll down to see the full text.
1996 J. Phys.: Condens. Matter 8 5649
(http://iopscience.iop.org/0953-8984/8/30/014)
View the table of contents for this issue, or go to the journal homepage for more
Download details:
IP Address: 142.132.1.147
This content was downloaded on 05/09/2015 at 08:37
Please note that terms and conditions apply.

J. Phys.: Condens. Matter 8 (1996) 5649–5657. Printed in the UK
Investigation of silver–glass nanocomposites by positron
lifetime spectroscopy
M Mukherjee†, P M G Nambissan‡ and D Chakravorty†§
† Indian Association for the Cultivation of Science, Jadavpur, Calcutta 700 032, India
‡ Saha Institute of Nuclear Physics, 1/AF Bidhannagar, Calcutta 700 064, India
Received 8 February 1996
Abstract. Nanocrystalline silver particles were grown in a glass medium by ion-exchange and
reduction techniques and studied by positron lifetime spectroscopy. The particle sizes varied
from 5 to about 25 nm as observed by transmission electron microscopy. The positron lifetime
spectra of all the samples could be decomposed into three components having lifetimes of around
160 ps, 400 ps and 1500 ps. The first is ascribed to positron annihilation at the interfaces of
the nanocrystalline silver and the glass matrix, and it decreases and stabilizes as the silver
grain size increases. The second component is explained as arising from positrons trapped and
annihilated at the free-volume defects in the glass matrix. The third component arises because of
the annihilation of orthopositronium at large free-volume defects. The effects of temperature on
the interfacial defects and the processes leading to the formation of additional positron trapping
centres are discussed.
1. Introduction
The remarkable improvement in several useful physical properties of crystalline materials
with particles or grains of size about a few tens of nanometres holds great promise for the
materials science and technology of the future (Schmitt-Ott et al 1980, Birringer et al 1984,
Gleiter 1989, Schaefer et al 1989). The improved characteristics of these nanocrystalline
materials are in fact due to the presence of a very large number of grain interfaces. As
the majority of the atoms reside on these grain interfaces, some of the properties of these
materials appear to be controlled by these atoms (Birringer et al 1986).
The most well known method of preparation of nanocrystalline materials is the technique
incorporating metal evaporation and condensation in a high-purity noble-gas atmosphere
(Schaefer et al 1988). After evacuation of the noble gas, the ultrafine powder is compacted
in a high vacuum under mechanical pressure to disc-shaped pellets. In the present work, we
have used an alternative method of growing nanosized silver particles in a glass medium by
ion exchange and a suitable reduction treatment. This method has the advantages that the
growth of nanocrystalline grains and their interfaces are natural, and no artificial compaction
is required. The grain growth can be efficiently controlled by properly choosing the reduction
temperature and time.
It is now generally accepted that the properties of nanocrystalline materials are very sen-
sitive to the defect structures and the dynamical processes associated with them. Rather than
the commonly encountered crystalline defects, the grain interfaces are the most important
§ Also affiliated to Jawaharlal Nehru Centre for Advanced Scientific Research, Bangalore 560064, India
c
0953-8984/96/305649+09$19.50 ꢀ 1996 IOP Publishing Ltd
5649

5650
M Mukherjee et al
defect structures to be investigated in order to understand the exotic features of nanocrys-
talline materials. These grain interfaces are distinctly different from the macroscopic grain
boundaries in coarse crystalline solids. Several experimental methods have been applied
to the investigation of the properties of nanophase materials with varying degrees of suc-
cess. Transmission electron microscopy (TEM) provides probably the most direct way of
characterizing the nanophase structure. Nanocomposite structures have been used in recent
years to study the properties of the nanocrystalline state of matter (Roy and Chakravorty
1990). Estimation of the fractal dimensions and measurements of the electrical resistivity
have also been carried out by various groups to delineate the mechanisms leading to the
anomalous features exhibited by these materials at low temperatures (Mukherjee et al 1993,
1994). One of the highly successful spectroscopic probes for the study of the structural as-
pects of nanocrystalline materials even at subatomic levels incorporates the measurements
of positron lifetimes as functions of certain well chosen controlled experimental parameters
(Tong et al 1992, Qin et al 1994, Kizuka et al 1994). We report here the results of our
investigation by positron lifetime measurements of silver nanoparticles grown in a glass
medium by the reduction of ion-exchanged glass at different temperatures.
2. Experimental details
2.1. Preparation of the samples
The glass with the composition 55 mol% SiO2–12 mol% ZnO–32.2 mol% Li2O–0.8 mol%
P2O5 was prepared by melting the required amounts of reagent-grade chemicals in alumina
crucibles in an electrically heated furnace at around 1400 ◦C. The melt was poured onto a
brass plate. Several glass plates were cut and powdered by gentle grinding. The samples
were subjected to a lithium ↔ silver ion exchange by immersing them in a molten bath of
silver nitrate at a temperature of 310 ◦C for a period of 8 h. The samples were then washed
in distilled water to remove any adhering traces of silver nitrate and dried. These were
subjected to reduction treatment in hydrogen at temperatures ranging from 300 to 800 ◦C.
After the reduction, the samples were quenched to room temperature to ensure a narrow
size distribution of the silver metal particles. More details have been presented elsewhere
(Chatterjee and Chakravorty 1989).
2.2. Transmission electron microscopy
Transmission electron micrographs were taken using a JEM 200 CX microscope operated
at 160 kV at the Regional Sophisticated Instrumentation Centre, Bose Institute, Calcutta.
2.3. Positron lifetime measurements
A weak (about 1.5 µCi) 22NaCl source was deposited on and covered by folding a thin
(about 2 mg cm−2) nickel foil. It was kept immersed in the centre of a volume of the
powdered specimen in a glass tube. The source was covered by the powdered specimen
from all sides at least to a thickness of 5 mm so that all the positrons emitted by the
source were captured and annihilated in the specimen itself. The glass tube containing the
source–specimen assembly was kept such that the source was in between the two detectors
of a slow–fast gamma–gamma coincidence spectrometer. The spectrometer had a prompt
time resolution (FWHM) of 260 ps. During the positron lifetime data acquisition, the glass
tube was continuously evacuated to remove any possible traces of air, which was otherwise

Positron lifetime spectroscopy of silver–glass composites
5651
likely to be trapped in the powdered specimen. With a calibration of 50 ps per channel
in the multichannel analyser, a total of more than 1 million counts was accumulated for
each lifetime spectrum. The spectra were analysed using the programs RESOLUTION
and POSITRONFIT (Kirkegaard et al 1981). In the multicomponent analysis performed,
variances of fit between 0.96 and 1.14 were obtained.
3. Results and discussion
The positron lifetime spectra of two samples are shown in figure 1. The spectra of all
the samples were decomposed into three components, called τ1, τ2 and τ3, respectively,
in ascending order of their magnitudes. (To avoid scattering of the analysed results,
the value of τ3 was fixed to 1500 ps, which was the mean of the values obtained
in an initial unconstrained analysis.) Their respective intensities I1, I2 and I3 were
also obtained from the analysis (I1 + I2 + I3 = 100%). The mean positron lifetime
τm = (τ1I1 +τ2I2 +τ3I3)/(I1 +I2 +I3) was calculated from these parameters. These values
are shown plotted against the reduction temperatures used for the various samples in figure 2.
Figure 1. Positron lifetime spectra for (a) the ion-exchanged sample and (b) the sample reduced
at 600 ◦C. (c) The prompt time resolution spectrum of 60Co gamma rays is also shown.
3.1. The origin of the different lifetimes
To understand the physical meaning of the various positron lifetimes, we note that the
positrons emitted by the 22Na nuclei became rapidly thermalized on entering the sample. A

5652
M Mukherjee et al
Figure 2. Positron lifetime parameters (a) τ1, τ2, τ3, τm and (b) I1, I2, I3, I3/I2 plotted as
functions of the sample reduction temperature.
significant fraction of these positrons will be captured within the crystalline grains in the
sample. If the size of these grains is less than the thermal diffusion wavelength of positrons
(about 50–110 nm in most of the materials), they will diffuse out of the grains and become
trapped at the grain interfaces. Even the largest grains used in this study have diameters of
about 38 nm and are therefore much smaller than the positron thermal diffusion wavelength.
Hence all the positrons entering into the grains would eventually diffuse out and annihilate
at the grain interfaces. During the analysis of the positron lifetime spectra, specific attempts
were made to search for very short-lived components with lifetimes of the order of 50 ps
or slightly more (the delocalized positron lifetime of 133 ps shortened by trapping in the
grains) but no such lifetimes were obtained. This observation, in other words, ruled out
the presence of any kind of positron traps inside the grains. This is further supported by
the observations of Matuoka and Morinaga (1982) who reported that there are no positron
traps in an ultrafine silver particle of about 70 nm diameter. On the other hand, the disorder
in the arrangement of atoms at the grain interfaces, a specific feature of nanocrystalline
materials (Schaefer et al 1988), can result in ‘vacancy-like’ trapping centre for positrons.
Positrons which are thus trapped and subsequently annihilated at the grain interfaces have
been shown to have lifetimes similar to the values of τ1 obtained in the present work (Qin
et al 1994, Kizuka et al 1994). We therefore attribute the origin of τ1 in our experiments
to positron annihilation at the interfaces of the nanocrystalline silver grains and the glass
medium. This argument is further supported by the work by Sui et al (1992) who had
given similar interpretations in their studies of nanocrystalline Ni–P alloys. Figure 3(a)

Positron lifetime spectroscopy of silver–glass composites
5653
Figure 3. (a) Electron micrograph and (b) the diffractogram of the ion-exchanged sample.
shows the transmission electron micrograph for the ion-exchanged sample. Figure 3(b) is
the electron diffraction pattern for figure 3(a). The diffraction rings confirm the presence
of metallic silver. In table 1, the experimental and the standard ASTM dhkl-values are
listed. The microstructure observed in the case of the ion-exchanged sample is typical of
those found for specimens subjected to reduction in hydrogen. The precipitation of colloidal
silver within the glass without any hydrogen treatment arises owing to photoreduction of
silver ions. It is believed that some impurity ions such as Fe2+ present in the glass act as
the sensitizer for such a reaction to take place (McMillan 1964). In figure 4 is shown the
histogram of silver particles as obtained from figure 3(a). The particle size is found to obey
a log-normal distribution function given by
√
1n = (1/ 2π ln σ) exp{−2 [ln(x/x¯)/ ln σ]21(ln x)
(1)
1
where 1n is the fractional number of particles, x is the diameter of the particles, x¯ is the
median diameter and σ is the geometric standard deviation. The solid curve in figure 4 is
obtained by least-squares fitting of the experimental data to the above equation. Table 2
summarizes the values of x¯ and σ obtained for the different samples.
The value of τ1 obtained in the present work (163 ps) is much higher that that for
pure bulk silver (136 ps) (Debowska 1992) but less than that for positrons trapped in
monovacancies (186 ps) (Stuck and Schaefer 1985) or in dislocations (200 ps) (Linderoth
and Hidalgo 1987) in silver. In the high-pressure compaction studies on nanocrystalline
silver, Qin et al (1994) had attributed the values of τ1 ranging from 203 to 164 ps in their
work to positron trapping at the vacancy-like defects at the interfaces. The value of 163 ps

5654
M Mukherjee et al
Figure 4. A typical histogram obtained from figure 3(a). The solid curve is the fit obtained
using the log-normal distribution.
obtained in the present experiment is more in agreement with the value of 164 ps obtained
by Qin et al (1994) after the shrinkage of the vacancy-cluster defects and the elimination
of vacancy-like defects. Also the fact that the mean size of a defect is less than that of a
monovacancy is more characteristic of an interfacial defect rather than a conventional lattice
defect (Schaefer et al 1988, Kizuka et al 1994). In our work, the nanocrystalline grains
were formed by the process of nucleation and growth at the reduction temperatures used.
This is different from the processes involved in high-pressure compaction.
Table 1. ASTM and observed interplanar spacings dhkl for ion-exchanged sample.
dhkl (nm)
ASTM
Observed Silver
0.272
Silver oxide
0.2734
0.2365
0.236
0.205
0.166
0.145
0.123
0.093
0.2359
0.2044
0.1674
0.1427
0.1445
0.1231
0.09375
The second lifetime component τ2 can be identified as originating from positrons trapped
and annihilated at the free-volume defects in the glass matrix. These free-volume defects are
in fact the counterparts of large three-dimensional cavities called voids in crystalline solids
containing defects. Amorphous materials are in general characterized by a free-volume
defect density which is a few orders of magnitude higher than that in pure crystalline solids.
In the absence of the nanocrystalline silver grains discussed in the previous paragraph, the
presence of such a high concentration of free-volume defects would have resulted in the
saturation trapping of positrons, that is all the positrons injected into the sample would

Positron lifetime spectroscopy of silver–glass composites
5655
Table 2. Reduction treatments, median diameters x¯ and geometric standard deviations σ for
different samples.
Reduction treatment
Temperature
(◦C)
Time
(min) (nm)
x¯
Sample
σ
1
2
3
4
0
300
600
800
0
30
30
30
5.0
8.7
10.1
23.7
1.3
1.5
1.6
1.6
have been trapped and annihilated in these free-volume defects (Mishra et al 1981). The
magnitude of 380 ps for this lifetime component is typical of positrons trapped in voids of
˚
radii 2–3 A and consisting of the ten to 12 monovacancies per void in crystalline solids
(Hautojarvi et al 1977). The actual size of the corresponding free-volume defects in the
amorphous matrix cannot be very different. The relatively larger value (about 42%) of the
intensity I2 further lends support to the interpretation of positron trapping in free-volume
defects.
The long-lived lifetime component τ3 = 1500 ps (fixed) obtained in the analysis is due
to the annihilation of orthopositronium (o-Ps), formed at the internal surfaces of the large
free-volume defects. The o-Ps annihilation takes place after the ‘pick-off’ of the positron
from the Ps atom by an electron with opposite spin from the surroundings (Smedskjaer et
al 1983). The efficiency of this process is dependent on the structure and contamination of
the internal surfaces of the defects (Schaefer et al 1988). Even in the initial unconstrained
positron lifetime data analysis, the values of τ3 centred around 1500 ps, which was kept fixed
in the final analysis. This means that either the large free-volume defects in the glass matrix
are contaminated by hydrogen atoms during the reduction treatment or their size is large
enough to provide a favourable condition for positronium formation. This interpretation
seems to be feasible as the intensity I3 of this component falls at higher temperatures, as
discussed later.
3.2. The results of reduction at higher temperatures
The lifetime τ1 due to positron annihilation at the grain–glass interfaces falls drastically to
153 ps and remains more or less the same in the samples heat treated at higher temperatures.
In agreement with the interpretations given by Tong et al (1992), the reduced positron
lifetime can be attributed to the shrunken sizes of the positron traps at the interfaces of
larger grains. In fact the grain growth would result in the ultimate reduction in the total
interfacial area which in turn could lead to shrinkage of the free volume available at the
interfaces. It is not known at the present stage what exactly leads to the stabilization of τ1.
Studies over a wider range of grain sizes may throw some light on this problem.
Turning back to figure 2, the variation in τ1 with the sample reduction temperature
is understood in terms of the interpretations given above. However, the variations in the
other positron lifetime parameters seem to indicate a number of additional features. The
initial rise in τ2 indicates the agglomeration of the free-volume defects in the glass matrix
due to the heat treatment. The sharp decrease in the corresponding intensity I2 in the
temperature range from 300–600 ◦C provides further support for this argument. The initial
rise in I2 is considered to be the reflection of the decrease in I1 at this stage due to a

5656
M Mukherjee et al
reduction in the total interfacial area brought about by the grain growth. This process in
fact continues but its effect on the intensity parameter I2 becomes less evident because of
the dominating effects of the temperature-induced defect agglomeration in the later stages.
There is, however, no evidence to suggest whether the largest value of τ2 = 408 ps observed
here is the saturation lifetime in large free-volume agglomerates. The increase in I3 in the
initial stages is accounted for in terms of hydrogen contamination of the internal surfaces
of free-volume defect clusters during the reduction at higher temperatures. It may be noted
here that an increased ‘pick-off’ rate had been predicted by Mogensen et al (1985) to be
due to hydrogen absorption in microcrystals.
The subsequent fall in the value of τ2, on the other hand, can be understood from the
contribution coming from the microvoids formed at the intersection of the interfaces of more
than two nanocrystalline grains. During the growth of silver nanoparticles, since the interior
of the grain attains crystallinity, the excess free volumes originally present there become
transferred out of the grain. These free volumes can then agglomerate at the intersection of
the grain interfaces, giving rise to small microvoids. As Schaefer et al (1988) had pointed
out, the formation of these microvoids is plausible since the atomic positions near the centre
of a grain boundary intersection are unstable in the case of strongly differing orientations of
the adjoining grains. The positrons trapped in these microvoids may have a characteristic
lifetime which is less than the lifetime of positrons annihilating in the free-volume defect
agglomerates in the glass matrix. In the case when these two lifetimes are resolution
limited, the resultant value of τ2 will be the weighted average of the two lifetimes. From
this viewpoint, the relatively lower values of τ2 observed in samples containing larger silver
grains can be understood in terms of the excess number of microvoids being formed owing
to the intersection of nanocrystalline interfaces. This is probable especially when the grains
actually grow in size. Also, the rise in the value of I2 at the final stage further supports
this argument. As the positron lifetime at the interfacial intersections cannot be separated
from that in the free-volume defects in the glass matrix, the trapping model (Schaefer et
al 1988) cannot be applied here for the extraction of the residence time of positrons in the
delocalized free state.
Do some of the free-volume defect agglomerates in the glass matrix anneal out during
the heat treatment? It is reasonable not to rule out the possibility of annealing of at least
a fraction of the agglomerates at higher temperatures. I3 and in fact even the fraction of
o-Ps formed in the free-volume defect clusters (indicated by the ratio I3/I2 in figure 2) fall
significantly beyond 600 ◦C. The desorption of hydrogen from the defect clusters may begin
at this stage, causing the clusters to become unstable and to anneal out. It is reasonable to
argue that Ps formation is more favourable in very large vacancy-type defect agglomerates.
In smaller vacancy clusters, the overlap between the electron and positron wavefunctions
may not be negligible enough to favour the formation of Ps atoms, even though the trapped
positron is able to enhance its lifetime. Alternatively, it could be argued that the Ps formation
probability is reduced considerably in vacancy clusters of sizes below a certain critical limit.
Thus, in the sample which is heat treated at 800 ◦C, the majority of the vacancy clusters
contributing to the value of τ2 (= 345 ps) are microvoids formed at the intersection of
the interfaces. A significant fraction of the free-volume agglomerates in the glass matrix
may have already been annealed out owing to the relatively higher temperature of the heat
treatment.
The behaviour of the mean lifetime τm in fact summarizes the above features. Initially
the agglomeration of the free-volume defects in the glass matrix results in an increase in
τm whereas the annealing of these agglomerates and the production of smaller microvoids
gradually cause it to decrease at higher reduction temperatures.

Positron lifetime spectroscopy of silver–glass composites
4. Conclusions
5657
We have presented here an alternative way of systematically investigating the growth of
silver nanoparticles in a glass medium by making use of the ability of positrons to probe
sensitively the properties of interfaces of the grains with the glass matrix. The important
findings in this study can be summarized as follows.
(i) The lifetime of positrons annihilating at the interfaces of the nanocrystalline grains
with the glass reduces and saturates as the average grain size increases. This behaviour is
attributed to the shrinkage of the available free volumes at the interfaces during the grain
growth.
(ii) In samples heat treated for reduction in hydrogen at higher temperatures, the free-
volume defects in the glass matrix agglomerate first but a fraction of them anneal out.
(iii) At higher reduction temperatures, additional trapping centres are generated for
positrons at the intersection of nanocrystallite interfaces. These defects are classified as
microvoids.
Acknowledgments
The authors thank Professor Prasanta Sen for useful discussions and encouragement. MM
and DC acknowledge the support from a foreign research grant N00014-93-1-0040 by Office
of Naval Research, Virginia, USA.
References
Birringer R, Gleiter H, Klein H-P and Marquardt P 1984 Phys. Lett. 102A 365
Birringer R, Herr U and Gleiter H 1986 Trans. Japan Inst. Met. Suppl. 27 43
Chatterjee A and Chakravorty D 1989 J. Phys. D: Appl. Phys. 22 1386
Debowska E 1992 Mater. Sci. Forum 105–110 635
Gleiter H 1989 Prog. Mater. Sci. 33 223
Hautojarvi P, Heinio J, Manninen M and Nieminen R 1977 Phil. Mag. 35 973
Kirkegaard P, Eldrup M, Mogensen O E and Pedersen N J 1981 Comput. Phys. Commun. 23 307
Kizuka T, Nakagami Y, Ohata T, Kanazawa I, Ichinose H, Murakami H and Ishida Y 1994 Phil. Mag. A 69 551
Linderoth S and Hidalgo C 1987 Phys. Rev. B 36 4054
Matuoka Y and Morinaga M 1982 J. Phys. C: Solid State Phys. 15 4207
McMillan P W 1964 Glass Ceramics (London: Academic)
Mishra T, Otake S, Fukushima H and Doyama M 1981 J. Phys. F: Met. Phys. 11 727 and some of the references
therein
Mogensen O E, Eldrup M, Morup S, Ornbo J W and Topsoe H 1985 Positron Annihilation ed P C Jain, R M
Singru and K P Gopinathan (Singapore: World Scientific) p 980
Mukherjee M, Datta A and Chakravorty D 1994 Appl. Phys. Lett. 64 1159
Mukherjee M, Saha S K and Chakravorty D 1993 Appl. Phys. Lett. 63 42
Qin X Y, Zhu J S, Zhou X Y and Wu X J 1994 Phys. Lett. 193A 335
Roy B and Chakravorty D 1990 J. Phys.: Condens. Matter 2 9323
Schaefer H-E, Eckert W, Stritzke O, Wurschum R and Templ W 1989 Positron Annihilation ed L Dorikens-
Vanpraet, M Dorikens and D Segers (Singapore: World Scientific) p 79
Schaefer H-E, Wurschum R, Birringer R and Gleiter H 1988 Phys. Rev. B 38 9545
Schmitt-Ott A, Schurtenberger P and Siegman H C 1980 Phys. Rev. Lett. 45 1284
Smedskjaer L C and Fluss M J 1983 Solid State: Nuclear Methods (Methods of Experimental Physics 21) ed J N
Mundy, S J Rothman, M J Fluss and L C Smedskjaer (New York: Academic) p 78
Stuck W and Schaefer H-E 1985 Positron Annihilation ed P C Jain, R M Singru and K P Gopinathan (Singapore:
World Scientific) p 454
Sui M L, Xiong L Y, Deng W, Lu K, Patu S and He Y Z 1992 Mater. Sci. Forum 105–110 1249
Tong H Y, Ding B Z, Wang J T, Lu K, Jiang J and Zhu J 1992 J. Appl. Phys. 72 5124