Metal Cluster's Effect on the Optical Properties of Cesium Bromide Thin Films

Abstract

Cesium Bromide films grown of glass substrates by thermal evaporation showed
interesting optical properties. The UV-visible absorption spectra showed peaks
which showed red shift with time. Structural and morphological studies
suggested decrease in grain size with time which was unusual. Theoretical
simulation shows the optical behaviour to be due to surface plasmon resonance
resulting from Cesium clyindrical rods embedded in the films.

Full text

Page 1

arXiv:1112.3709v1 [cond-mat.mtrl-sci] 16 Dec 2011
Metal Cluster’s Effect on the Optical Properties of
Cesium Bromide Thin Films
Kuldeep Kumara, P.Aruna∗, Chhaya Ravi Kantb, Bala Krishna Juluric
aMaterial Science Research Lab,
S.G.T.B. Khalsa College,
University of Delhi, Delhi-110 007, India
bDepartment of Applied Sciences, Indira Gandhi Institute of Technology,
Guru Gobind Singh Indraprastha University, Delhi 110 006, India
cDepartment of Engineering Science and Mechanics,
The Pennsylvania State University,
Pennsylvania 16802, USA
December 19, 2011
∗e-mail: arunp92@physics.du.ac.in, (T) +91 011 29258401, (F) +91 011 27666220
1

Page 2

Abstract
Cesium Bromide films grown of glass substrates by thermal evaporation showed inter-
esting optical properties. The UV-visible absorption spectra showed peaks which showed
red shift with time. Structural and morphological studies suggested decrease in grain
size with time which was unusual. Theoretical simulation shows the optical behaviour to
be due to surface plasmon resonance resulting from Cesium clyindrical rods embedded
in the films.
PAC No: 68.55.-a, 71.20.Ps, 78.67.Bf
Keywords: Alkali halides, Thin Films, X-Ray Diffraction, UV-visible spectroscopy
1 Introduction
Alkali Halides (AH) have been extensively studied in the early years to understand the variation
in energy band structure with changing lattice constant [1]. The simple cubic crystal structures,
experimentally determinable phase transitions and variation in properties when these samples
are subjected to presure made them popular samples [2]. Also, AH are highly photon-absorbing
materials in X-ray and Ultraviolet region, giving rise to color-centers [3]-[6]. Cesium halides
hence became popular in detectors and optoelectronics devices [7]. Cesium halides also showed
excellent quantum efficiency. Infact, thin CsBr films deposited on thin metal layers have useful
application as electron source and as protective layer of photocathode [8]. However, AH films
suffered due to their high reactivity with atmospheric water vapor [9] making it difficult to
maintain. In this manuscript we focus on Cesium Bromide, since of all the alkali halides, it
exhibits a relatively higher stability when exposure in air [10].
Recent interest in Alkali Halides is due to the appearance of Surface Plasmon Resonance
(SPR) peaks in the UV-visible region of its absorption spectrum, when noble and other metal
nano-particle are embedded in it [11]. However, we have noticed SPR peaks developing in
cesium iodide [12] and cesium bromide (present work) thin films even with aging. We believe
this to be a direct proof of the idea put forward by Emel’yanov and others [13, 14]. They
claimed that the defect (such as F-centers) in alkali halides when provided the right conditions
2

Page 3

would lead to metal cluster formations. We believe it is these alkali clusters that form the
metal nanoparticles responsible for SPR formation. In the present work we report formation
of Cesium cluster in CsBr thin films due to ageing effect and their effect on optical properties.
2 Experimental
Cesium Bromide thin films were deposited by thermal evaporation on microscopy glass sub-
strates. The starting material (powder) was obtained from HiMedia (Mumbai) and was of
99.98% purity. Films depositons were carried out at room temperature with vacuum better
than 1.4 × 10−5Torr. The thickness of the films were monitered with the help of a Digital
Thickness Moniter (DTM-106). One part of the films were kept in desiccators and taken out
only for characterization while the second part were kept outside, exposed to open air. This
was done to compare the ageing effect of these films maintained in different conditions.
The structural and optical characterization the samples were done with Philips PW 3020
diffractometer and Systronics double beam UV-visible spectrophotometer (2202).Surface
morphology and elemental composition of films were examined with Field-Emission Scaning
Electron Microscopy (FE-SEM) FEI-Quanta 200F and the Energy Dispersive Analysis of X-
rays (EDX) system attached with it.
3 Results and Discussion
3.1 Optical Studies
The absorption spectra of Cesium Bromide films showed thickness dependence with films
thicker than 550 nm showing intense absorbance in the visible region. We believe that these
peaks are due to surface plasmon resonance (SPR). SPR peaks results from the interaction
between metal conduction electron and the electric field component of incident electromagnetic
radiation [15]. This interaction leads to collective oscillations of free electrons on the metal
surface if the metal particle’s dimensions are smaller than or comparable to the incident wave-
3

Page 4

length of light. These peaks in CsBr’s absorbance spectra are a result of metal nano-clusters
of cesium existing within the films. Studies on Cesium Iodide too had revealed similar peaks
in their UV-visible absorption spectra in our earlier studies [12].
However, what is intrigging in Cesium Bromide is the systematic variation in film property
with time. UV-visible spectra recorded immediately after film fabrication showed a dominant
peak around 500 nm. Not only did this peak red-shift with time but also its absorption
intensity was also found to decrease. The UV-visible absorption spectra (Fig 1) of a 840 nm
thick sample taken at various intervals after film fabrication exhibits this trend. This ageing
effect of the sample is believed to depend on the ambient atmosphere in which the sample was
maintained, hence as explained above, one set of the samples were maintained in a dessicator
while the second set were maintained in a normal miscroscope slide box. As expected, both
samples that started with identical UV-visible spectra showed different rate of variation. Those
kept in dessicator aged slower. The shift in peak position shows a linear trend with time and
the rate of change was found to be film thickness dependent (fig 2a). Thicker films showed
higher rate of shift as compared to the thinner films (fig 2b). Even the decrease in absorption
showed systematic trend of variation with time (fig 2c).
Both the SPR peak position and its intensity are strongly dependent on the metal cluster’s
size, it’s and that of the surrounding media dielectric constant. These results, hence, indicate
some systematic and continuous changes in the film that require investigation. Below we focus
our attention on how metal clusters arise in our samples and the structural and morphological
changes which would explain the variations observed.
3.2 Morphological, Structural & Compositional Studies
As can be seen from fig 3, sharp edged grains can be found uniformally scattered on the
film surface. The micrographs of this figure compares two samples, (a) sample maintained in
dessicator and (b) that kept outside. The grain density and grain size were determined from
these micrographs using the software ImageJ. Below each micrograph histograms depicting
the grain size distribution in the micrograph are given. As can be seen, samples kept in
dessicator showed slightly larger grains in the sample maintained in the dessicator as compared
4

Page 5

to that maintained outside (1.00µm as compared to 0.95µm). The grain density of samples
kept in air (0.91µm−2) was found to be nearly twice that of the counterpart kept in the
dessicator (0.52µm−2). The full width at half maxima (FWHM) of the Guassian fit to the
histograms also reflect a narrower distribution for the samples maintained in the dessicator.
Viewed in conjucture these results suggest grains of the film split giving rise to smaller grains.
Also, this process of grain breaking is encouraged in samples kept in air. Figure 4 gives a
suggestive sequence of “grain division”. Fig 4a shows two grains in close proximity that are
inter-connected by vesticles. These structures increase in length as the grains move apart. As
the grains moves apart, the vesticle like structure is retained by one of the grains (see fig 4b).
These vesticles finally break off and fall into the background (circles marked in fig 4c highlight
this) leaving behind a smooth spherical grain. EDX (Table 1) on vesticle and grains show
them to be made up of cesium and cesium bromide respectively. These results would indicate
removal of bromide from the surfaces of the grains. The process is accelerated in the samples
kept in air. The sublimation leaves behind cesium metal layer on the surface of the grain,
resulting in a “insulator-metal” “core-shell” structure. Figure 5 is a sample TEM micrograph
which clearly shows distinct regions of core and shell. As expected there is a large variation in
the grain size. Also, seen distributed among the spherical shells are “rod” like structures we
have explained above.
Fig 6(a) shows the Selected Area Electron Diffraction (SAED) taken on one of the core of
the “core-shell” grains. The spots indicate the crystalline nature of the “core”. The major
spots of the SAED are arranged on thwo distinct rings corresponding to the (110) and (211)
peaks of CsBr as indexed in ASTM Card No 73-0391. Similar analysis of the rod region (fig 6b)
shows it to be crystalline with three distinct ring corresponding to the (200), (331) and (220)
planes of Cesium as given in the ASTM Card No 18-0325.
X-ray Diffraction of the samples shows two major peaks of CsBr. The existence of both
CsBr and Cs in our samples are confimred by the broad peaks at 2θ ≈ 29oand ≈ 52o(fig 7).
These peak positions match those given in ASTM Card No 73-0391 and 18-0325. The grain
sizes (Table 2) of CsBr core were also calculated from the X-ray diffraction (fig 7) peak’s Full
Width at Half Maxima (FWHM) using Scherrer formula [16]. The smaller grain sizes indicate
5

Page 6

that the grain boundaries of the core, as viewed in the micrographs, enclose crystalline region
along with amorphous CsBr. While the amount of free Cesium in the samples can be thought
to be low, X-Ray diffraction pattern does show a broad peak at ≈ 52o, formed by merging of
the (211) and (220) peaks of CsBr and Cs. On deconvoluting these peaks, the average grain
size of free Cesium were also calculated. Grain size of CsBr and Cs were found to decrease with
increasing time. This trend is in agreement with those from morphological studies. However,
more importantly, the CsBr peaks were found to have shifted to the right as compared to the
peak positions given in the ASTM Card. This would indicate that the CsBr lattices are in a
state of stress with compressive forces acting on it. The stress in the film were calculated after
evaluating the strain using the relation [17]
∆d
d
=dobs− dASTM
dobs
(1)
where dobsis the experimentally observed d-spacing and dASTMis the corresponding peak’s
d-spacing as reported in the ASTM card. The stress then is determined by multiplying the
strain by the elastic constant of the material. The calculated strain on the (110) plane of CsBr
changed from -0.0023 to -0.0028 with an elapse of 900 hours. We believe it is this stress that
contributes to the required energy for bromide’s disassociation.
3.3Theoretical Modeling
Metal-insulator core-shell structures and nano-rods are known to give rise to SPR peaks in
absorbance spectra in the visible and near IR region [18, 19]. As stated earlier, the SPR peak
position and intensity strongly depend on the metal cluster size, shape, its dielectric constant
and that of the surrounding. Since, the surrounding of the metal clusters in this study is
invariant (CsBr), we may use the results here to investigate the contribution of size, shape and
cesium’s dielectric constant in SPR’s peak position. The systematic variation in SPR peaks
(fig 1 and fig 2) indicate a systematic variation in one of the mentioned properties.
The TEM micrograph (fig 5) shows existence of both cylinderical or nano-rod structures
along with spherical core-shell structures in out sample. Hence, it becomes important to isolate
which of these structures contribute to the SPR peaks in the wavelength region of 500-600nm.
6

Page 7

This can be done using the theoretical framework given by Mie [20] to explain the scattering and
absorption caused by metal clusters. The Mie theory essential is an application of Maxwell’s
equation for electromagnetic plane waves incident on metal particles. Recent works by Balaji
et al [21] extends Mie theory to explain scattering and absorption by core-shell structures. We
have extended those calculation schemes on our Cesium Bromide-Cesium core-shell structures
and find theoretically for the grain diamensions that exist in our samples, there are no or
very shallow peaks in the visible region (fig 8). For the same aspect ratio (rcore/rmantel) but
decreasing grain size, we see a blue shift in the shallow peak observed. This is not in agreement
with our experimental observations of fig 1.
As for theoretically calculating the SPR extinction cross-section of nano-spherical clusters,
the Gans theory [22] is both simple and accurate and is given as [23]
σext(ω) = V
?2π
3λ
?
ǫ3/2
m
?
i
ǫ2(ω)1/P2
i
ǫ2(ω)2+
?
ǫ1(ω) + ǫm1−Pi
Pi
?2
(2)
where ǫ1and ǫ2are the real and imaginary part of the metal nanocluster’s dielectric constant
and ǫmthe dielectric constant of the media in which the metal nano-clusters are embedded
in. The depolarising factors [24] can be easily modified for cylinderically shaped clusters by
taking the aspect ratio (c/a) to be far less than unity [25]. Figure 9 shows the extinction
cross-sections calculated for decreasing average grain size but increasing aspect ratio. For the
calculations here, we have used the frequency dependent dielectric coefficients reported by [26].
The curves and the trend are in agreement with the the experimental trends shown in figure 1.
For completeness, we used eqn (2) to see the shift in SPR peak position with varying
ǫm.This gives the SPR sensitivity to the medium’s refractive index [27].Based on our
calculations we find the slope between SPR peak position and medium’s refractive index to be
145.1 nm/RIU (fig 10). This small slope would in turn demand Cesium Bromide’s refractive
index to change from 1.66 to values greater than 4.10 to explain the results of fig 2(a). Thus,
we conclude that the variation of optical properties seen with ageing in the Cesium Bromide
films can be explained based on basis of formation of Cesium nano-rods in it.
7

Page 8

4 Conclusions
Experimental data suggests that the stress in the as grown films lead to formation of de-
fects caused by Bromide atom’s displacement. These defects start collecting together to form
clusters. With the curvature of the grain’s surfaces contributing the maximum stress, the
formation of Cesium at the grain surfaces can be understood. This surface Cesium around
Cesium Bromide not only contributes to the core-shell structures present in the film but also
the a site for formation of Cesium nano-rods. With time the shell-core grain size decreases
along with decrease in the average grain size of the cylinderical rods. Comparing the results
with theoretical simulations suggests cylinderical grains contribute to SPR peaks in the visi-
ble wavelengths with decreasing grain size accompanied with increasing aspect ratio (ratio of
diameter to length) leading to a red-shift in the peak position.
Acknowledgment
The authors would like to express their sincere gratitude to Department of Science and Tech-
nology (DST) India for the financial assitance (SR/NM.NS-28/2010) given for carrying out
this work.
8

Page 9

Tables
Table 1: Compares the presence of various elements (given in weight percent) in the grain and
vesticles in in the electron micrographs. The chemical composition as measured using EDX
attachment with the SEM.
Element CsBr (at grain) CsBr (at vesticle)
CK 18.04 49.58
OK15.72 22.29
MgK 02.0801.45
SiK20.8519.70
CaK03.35 03.49
CsL17.66 03.50
BrK22.30 00.00
Table 2: Grain size (gs) of CsBr and Cs as calculated from the X-Ray Diffraction Pattern.
Time (Hrs)gs of CsBr (nm) gs of Cs (nm)
22 6588
1125576
7945810
103434–
9

Page 10

References
[1] H. Fujita, K. Yamauchi, A. Akasaka, H. Irie and S. Masunaga, J. Phys. Soc. Japan, 68
(1999) 1994.
[2] M.B.Nardelli, S. Baroni and P. Giannozzi, Phys. Rev. B, 51 (1995) 8060.
[3] P.V. Mitchell, D. A. Wiegand and R. Simoluchowski, Phys. Rev., 121 (1961) 484.
[4] F.T. Goldstein, Phys. Stat. Solidi (B), 20 (1967) 379.
[5] M. Elango, Christian Gahwiller, F.C. Brown, Soid State Comm., 8 (1970) 893.
[6] B.R. Sever, N. Kristianpollar and F.C. Brown, Phys. Rev. B, 34 (1986) 1257.
[7] D.B. Sirdeshmukh, L. Sirdeshmukh and K.G. Subhadra, “Alkali Halides, A Hnadbook of
Physical Properties”, (Springer, Berlin 2001).
[8] A. Brakin, Nucl. Instrum. Methods Phys. Res. A, 367, 325 (1995).
[9] G. Yoshikawa, M. Kiguchi, K. Ueno, A. Saiki, Surf. Sci., 554 (2003) 220.
[10] Ed. W.M. Haynes, “CRC Handbook of Chemistry and Physics”, (CRC Press, USA 2011).
[11] S. Nie and S.R. Emory, Science, 275 (1997) 1102.
[12] K. Kumar, P. Arun, C. R. Kant, N.C. Mehra and V. Methew, Appl. Phys. A, 99 (2010)
305.
[13] V.I. Emel’yanov, Laser Phys., 2 (1992) 389.
[14] S. Seaglione, R.M. Montereali, V. Mussi and E. Nichelatti, J. Optoelect. Adv. ¡ater. 7
(2005) 207.
[15] U. Kreibig and M. Vollmer, “Optical Properties of Metal Clusters”, (Springer, Berlin,
1995).
10

Page 11

[16] B.D. Cullity, “Elements of X-ray diffraction” (2ndEd, Addisson-Wesley, NY).
[17] A.L. Patterson, Phys. Rev., 56 (1939) 978.
[18] J. Perez-Juste, I. Pastoriza-Santos, L.M. Liz-Marzan and P. Mulvaney, Coordination
Chem Rev., 249 (2005) 1870.
[19] C.F. Bohren, D.R. Huffman, “Absorption and Scattering of Light by Small Particles”
(John-Wiley & Sons, NY, 1983).
[20] G. Mie, Ann. Phys, 25 (1908) 329.
[21] B. K. Juluri, Y. B. Zheng, D. Ahmed, L. Jenson and T.J. Hung, J. Phys. Chem C, 112
(2008) 7309.
[22] R. Gans, Ann. Phys., 47 (1915) 270.
[23] S. Link, M. Mohamed, M. El-Sayed, J. Phys. Chem. B, 103 (1999) 3073.
[24] C. Noguez, J. Phys. Chem. C, 111 (2007) 3086.
[25] A. Burchanti, A. Bogi, C. Marinelli, C. Maibohm, E. Mariotti, S. Sanguinetti, L. Moi,
Eur. Phys. J. D, 49 (2008) 201.
[26] N.V. Smith, Phys. Rev. B, 2 (1970) 2840.
[27] H. Chen, L. Shao, K. C. Woo, T. Ming, H. Lin and J. Wang, J. Phys. Chem. C, 113
(2009) 17691.
11

Page 12

Figure Captions
Fig 1 The Absorbance spectra of a representative sample (not kept in dessicator) taken at
various intervals. Interestingly, as the SPR’s peak position shifted to higher wavelengths,
the intensity showed presistent decrease.The SPR’s peak position’s shift to higher
wavelengths with time show a linear trend.
Fig 2 The Absorbance spectra of three different CsBr films varying by thickness show system-
atic variation with time. (A) The peak position varies linearly with time, where the
(B) slope (rate at which peak position varies with time) is directly dependent on film
thickness. (C) shows the absorption intensity also varies with time.
Fig 3 Field-Emission Scanning Electron Microscope (SEM) micrographs of (a) sample main-
tained in dessicator and (b) kept outside. Histograms (c) and (d) shows the grain size
(gs) distribution of sample maintained in dessicator and those kept outside, respectively.
Fig 4 Field-Emission Scanning Electron Microscope images show the sequence of events as CsBr
grains break away. Micrograph (a) shows “vesticles” of Cesium inter-connecting grains
which (b) break away with “vesticle” going with one of the grains. These “vesticles” fall
off (c) giving cesium rods in the film.
Fig 5 Transmission Electron Microscope (TEM) images confirm nature of film throughout the
thickness of the film is same as that seen on the surface using SEM.
Fig 6 Surface Analysis using Electron Diffraction (SAED) of (a) the core and (b) the rod.
Fig 7 X-Ray diffraction of CsBr films after aging. Both broad peaks are deconvoluted to show
the Cs (?) and CsBr (△).
Fig 8 Extinction cross-section of Cesium Bromide-Cesium core-shell structure calculated (see
text) for two different grain sizes, namely (a) 1500 and (b) 1200nm, but same aspect
ratio (rcore/rmantel).
12

Page 13

Fig 8 Theoretically projected variation of SPR peaks caused by Cesium nano-rods. Family of
curves show a redshift with average grain size decreasing abid increasing aspect ratio
(c/a). The calculations where made using Gans Model (see text). The simulation follow
the same trands of fig 1.
Fig 10 Graph shows the relationship between SPR peak position with surrounding media’s re-
fractive index.
13

Page 14

Figures
0.4
0.6
0.8
1.0
1.2
400 600 800
Intensity
λ (nm)
384 hrs
434 hrs
505 hrs
573 hrs
623 hrs
889 hrs
Figure 1: The Absorbance spectra of a representative sample (not kept in dessicator) taken at
various intervals. Interestingly, as the SPR’s peak position shifted to higher wavelengths, the
intensity showed presistent decrease. The SPR’s peak position’s shift to higher wavelengths
with time show a linear trend.
14

Page 15

400
600
800
1000
0 400
time (hours)
800 1200
Peak position (nm)
950 nm
840 nm
540 nm
(a)
0.15
0.25
0.35
500 600 700 800 900
d(Peak Position)/dt
Thickness (nm)
(b)
0.4
0.8
1.2
1.6
0
400
time (min)
800 1200
Intensity
540 nm
840 nm
950 nm
(c)
Figure 2: The Absorbance spectra of three different CsBr films varying by thickness show
systematic variation with time. (A) The peak position varies linearly with time, where the (B)
slope (rate at which peak position varies with time) is directly dependent on film thickness.
15
(C) shows the absorption intensity also varies with time.