effect of temperature and micro morphology on the ag raman peak in nanocrystalline cuo thin films

Effect of temperature and micro morphology on the Ag Raman peak in nanocrystalline CuO thin films Shrividya Ravi, Alan B Kaiser, and Chris W Bumby, Citation Journal of Applied Physics 118, 085311 (201[.]

Effect of temperature and micro-morphology on the Ag Raman peak in nanocrystalline CuO thin films

Shrividya Ravi, Alan B Kaiser, and Chris W Bumby,

Citation: Journal of Applied Physics 118, 085311 (2015); doi: 10.1063/1.4929644

View online: http://dx.doi.org/10.1063/1.4929644

View Table of Contents: http://aip.scitation.org/toc/jap/118/8

Published by the American Institute of Physics

Effect of temperature and micro-morphology on the AgRaman peak

in nanocrystalline CuO thin films

ShrividyaRavi,1Alan B.Kaiser,1and Chris W.Bumby1,2,a)

1

The MacDiarmid Institute for Advanced Materials and Nanotechnology, SCPS, Victoria University

of Wellington, Kelburn Parade, Wellington 6140, New Zealand

2

Robinson Research Institute, Victoria University of Wellington, P.O Box 33436, Lower Hutt 5046,

New Zealand

(Received 24 April 2015; accepted 16 August 2015; published online 28 August 2015)

Raman spectra obtained from a nanocrystalline CuO thin film are observed to exhibit significant

variation in the peak position and peak line-shape as a function of spatial position within the film

We attribute this effect to variation in the degree of local heating beneath the focused spot of the

Raman probe laser To understand this, we have undertaken a detailed study of the

temperature-dependence of the CuOAgRaman peak We observe a linear relationship between line-width and

peak position, which persists over a wide temperature range, and is characteristic of a Raman

cess in which the temperature-dependence is dominated by anharmonic 3-phonon decay We

pro-vide an analytical description of the Raman line-shape as a function of temperature and use this

model to interpret the degree of laser heating observed within our sample Using this relationship,

we identify that the local micro-morphology of the CuO sample under study can dramatically affect

the temperature achieved due to laser heating We find that spectra collected from the surface of

“micro-bubbles” within the CuO film studied can reach temperatures of >1000 K beneath the

focused spot of our low power (5 mW) probe laser.V C 2015 Author(s) All article content, except

where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported License

[http://dx.doi.org/10.1063/1.4929644]

I INTRODUCTION

Confocal micro-Raman spectroscopy is a non-contact

and non-destructive technique that is widely used as an

analytical tool in the study of new materials The observed

line-shape of a Raman peak obtained from an inorganic

crystalline material can be influenced by a number of

con-tributing factors including: phonon confinement,1 intrinsic

stress,2electron/hole-phonon coupling,3and local heating.46

In particular, local laser heating can lead to significant

red-shifts and broadening of Raman peaks obtained from

nano-crystalline semiconductor materials Such materials

can exhibit large absorption coefficients at the wavelength of

the probe laser2in conjunction with suppressed heat

conduc-tivities due to interface scattering mechanisms.7 This can

result in substantially elevated temperature distributions

under the probe laser, which may be either homogeneous4or

inhomogeneous.8,9

Cupric Oxide (CuO) is a crystalline semiconducting

ox-ide, which has attracted continued interest due to the wide

range of nano-morphologies,10–12which can be formed using

inexpensive and facile synthesis methods such as solution

synthesis and thermal oxidation CuO has an optical bandgap

in the visible region13(1.35 eV at 300 K) and reported

de-vice applications of nanostructured CuO include infra-red

photodetectors,14gas sensors,15catalyst surfaces,11and field

emission sources.16 Many authors17–26 have reported

red-shifts and broadening of theAgRaman peak obtained from

various nanostructured CuO samples and have cited this as evidence of phonon confinement effects in their materials However, these reports rarely account for local laser heating

of the CuO samples—which can give rise to similar effects upon the observed experimental spectra Given that CuO has

a very strong absorption coefficient in the visible region (102 greater than silicon), laser heating cannot be lightly ignored However, there is very limited published data on the effect of elevated temperature upon the CuO Raman spectra, with earlier studies27,28having largely concentrated

on behaviour below 300 K In this work, we illustrate the potential for laser heating in CuO We first report a study of the temperature-dependence of the CuO AgRaman peak at elevated temperatures up to 673 K, which we describe using the 3-phonon scattering model of Klemens.27We have then used this model to investigate the impact of local laser heat-ing upon this Raman peak and highlight the large variability

in the observed degree of local heating, which occurs as a result of micro-morphological features which are present in our CuO thin film samples

II EXPERIMENTAL

A set of nanocrystalline CuO thin film samples were produced through a two-stage process First, a thin film of copper (thickness: 500–1000 nm) was deposited upon a sapphire (006) substrate by thermal evaporation under vac-uum The Cu films were then placed in a tube oven and annealed at 500C in air for 2 h Initial sample character-isation has been carried out using scanning electron mi-croscopy (SEM) and X-ray diffraction (XRD).29 Raman

a) Author to whom correspondence should be addressed Electronic mail:

chris.bumby@vuw.ac.nz

spectra were collected using a Jobin Yvon LabRam

spec-trometer with a 632.8 nm probe laser An initial power

de-pendence study at a single spot on the CuO film was

acquired using four different levels of incident laser

power: 5 mW, 3 mW, 1 mW, and 500 lW The laser

power incident upon the sample was varied through using

a series of neutral density filters, which were placed in the

optical path between the laser and the microscope

objec-tive A series of spectra were then collected at

tempera-tures between 200 and 700 K at 500 lW laser power and

under a nitrogen gas atmosphere This was achieved using

a sample heating stage located within a N2 gas-cooled

cryostat A 100 (0.6 NA) objective was used for the

cryostat measurements, whilst a 100 (0.9 NA) objective

was used for the micromorphology study Both objectives

have a beam waist at the focal point of the probe laser,

which has been measured30to be0.6 lm

III RESULTS

A Sample characterisation

Figure1shows SEM images of a sample nanocrystalline

CuO film at 100 and 30 000 magnification At high

mag-nification (30 000), we observe that the nano-morphology

of the film comprises tightly packed spherical crystallites

with diameters between 40 and 60 nm.29This morphology is

a result of the thermal oxidation process used to produce our

sample films, which drives a large increase in the film

vol-ume (The volume of the oxidised CuO film is 1.7 times

larger than the volume of the initial copper film.) During this

process, stress relaxation occurs both at grain boundaries and

at the film surface Earlier XRD studies have confirmed that

all the sample films produced for this study were monophase

CuO and Williamson-Hall analysis indicated that the films

were free of intrinsic stress29 at the macro-scale (using an

XRD spot size of 1 cm) At lower magnification (100),

the SEM image in Figure 1(a) reveals the presence of

bubble-like morphological features in the film and we

believe that these micro-bubbles also result from the

expan-sion of the film during growth, in this case leading to stress

relaxation through out-of-plane deformation of the film such

that it becomes detached from the substrate Surface profile

measurements of the sample using a Dektak profilometer

indicated that these micro-bubble features vary in height from approximately 1 lm to 5 lm, with the regions between the micro-bubbles comprising a flat planar film, as is expected for a CuO film formed directly upon the flat under-lying sapphire substrate

B Temperature dependence of Agmode Cupric oxide exhibits three Raman active modes27 (Ag,Bg, Bg), and when probed with an excitation energy of 1.96 eV at 300 K, the dominant signal31 is the Ag mode at

298 cm1 Figure2shows theAgRaman peak obtained from

a nanocrystalline CuO thin film sample measured at room temperature under four different incident laser power levels

We observe a clear power dependence of the peak position and width, whilst retaining a symmetrical peak line-shape This effect is most marked for spectra taken using a probe laser powers of more than 1 mW, where xpeak under-goes a significant red-shift of up to 16 cm1 whilst the line-width broadens from 8 cm1 to 22 cm1 However, for probe laser powers of1 mW, the peak position is observed

to remain approximately constant at 298 cm1, which is consistent with room-temperature values reported in the ear-lier literature.27,28 We shall show that the observed power dependence is caused by laser heating of the CuO thin film

FIG 1 Scanning electron microscope (SEM) images of the nanocrystalline CuO films studied in this work Images acquired at (a) 100 magnification and; (b) 30,000 magnification.

FIG 2 Raman spectra of the CuO A g peak collected at room temperature (293 K) from the nanocrystalline CuO thin film sample under 4 different incident probe laser powers (5 mW, 3 mW, 1 mW, and 0.5 mW).

In order to understand the extent of laser heating, we have

first undertaken a temperature-dependence study in which

laser heating effects were eliminated by using a low laser

power (500 lW) with extended acquisition times, whilst the

CuO sample was uniformly heated using an external high

temperature heating stage Care was taken to ensure that all

spectra collected were taken from sites on the sample located

within an extended region of uniform flat planar

morphol-ogy This was achieved by refocusing the laser beam and

checking the morphology using the attached optical

micro-scope prior to every measurement

Figure 3 shows the evolution of the Ag peak with

increasing sample temperature This shows qualitatively

sim-ilar red-shift and broadening behaviour to that seen in the

power-dependent data The measured line-shape of the Ag

peak comprises a linear combination of Gaussian and

Lorentzian components, and across the range of temperatures

measured we empirically find that the line-shape is

well-described by a pseudo-Voigt function32of the form given in

the following equation:

I x ð Þ ¼I0

2

C=2

ð Þ 2

x  x peak

ð Þ2þ C=2 ð Þ2

þI0

2 exp

4ln 2 ð Þ x  x ð peak Þ2

C 2

!

: (1)

We have used this equation to define the two key

temperature-dependent parameters, which are required to

describe theAgline-shape, namely, the peak position, xpeak

and the line-width, C (C denotes the full-width at

half-maximum of Eq.(1)) We obtained values for each of these

parameters from each of our measured spectra by

iterative-regression fitting of Eq.(1) to the experimental data (using

the Levenberg-Marquardt algorithm) as shown in Figure3

This approach enabled robust and unambiguous values to be

assigned to xpeakand C for the individual spectra measured

at each temperature Using the calculated values, we can

then examine the relationship between the values of xpeakðTÞ and CðTÞ obtained as the temperature of the sample varied, and this is shown in Figure4 There is a clear linear correla-tion between the peak posicorrela-tion and line-width As discussed below, this is characteristic of a Raman peak for which the change in line-shape due to temperature is dominated by the temperature-dependent three-phonon anharmonic decay33 rate

The Raman peak position of a crystalline solid,

xpeakðTÞ, is conventionally expressed as a function of tem-perature in the form of a linear combination of three terms34–38

xpeakðTÞ ¼ x0þ Dx1ðTÞ þ Dx2ðTÞ: (2) The second and third terms on the right of Eq (2)describe the contributions from different effects that lead to a shift in peak position from the pure harmonic frequency of the opti-cal mode, x0 The second term describes the contribution due to lattice softening from thermal expansion, which is expanded in Eq.(3) Here, c is the Gruneisen parameter and aðTÞ is the temperature dependent coefficient of thermal expansion39

Dx1¼ x0 exp 3c

ðT 0 aðT0ÞdT0

 1

We can estimate the contribution from Dx1ðTÞ to the peak-shift in CuO using values from Ref 40for cCuO 0.37 and

aCuOðT > 250 KÞ  6.0  106K1 In this manner, we esti-mate the expected peak shift of the Ag line due to thermal expansion between 273 K and 673 K to be approximately 0.6 cm1 This is far lower than the experimentally observed shift of 18 cm1, which implies that in this case we can neglect the Dx1term in Eq.(2)

The third term in Eq.(2), Dx2ðTÞ, describes the peak-shift observed due to the anharmonic decay of Raman optical phonons into two or more phonons from other branches Following the approach of Klemens,33Dx2ðTÞ can be writ-ten as shown in Eq (4), which describes the contribution

FIG 3 Plot showing Raman spectra of a CuO thin film sample acquired at

temperatures from 198 K to 673 K at approximate temperature intervals of

25 K Note the red-shift of the A g peak position and broadening of the peak

line-width with increasing temperature Open circles show experimental

data and solid red lines show fits to the experimental data using Eq (1) For

clarity, spectra taken at each temperature have been uniformly offset along

the y-axis The spectra marked “*” were taken at 298 K and approximately

correspond to the 0.5 mW data shown in Figure 2

FIG 4 Plot showing linear correlation between the measured line-width, CðTÞ, and the measured peak position, x peak ðTÞ, for the CuO A g Raman peak spectra obtained across the temperature range of 198 K to 673 K.

from the decay of an optical phonon into two identical

pho-nons, each with frequencyx0

2 and of opposite momenta (the so-called “three phonon” decay process)

Dx2ð Þ ¼ AT x 1þ 2

exp hcx0 2kBT

 1

0 B

1

The temperature dependence of the three-phonon process

arises from the thermal occupation distribution of the

daughter phonons, given by the Planck distribution:

exp hcx0

2 B T

 1

The observed broadening of the Raman peak with

increasing temperature arises from damping of the excited

optical phonon, and the line-width is inversely proportional

to the phonon lifetime.41The lifetime of the excited phonon

is determined by its decay rate into lower energy phonons,

which follows the same three-phonon process as described

above Hence, CðTÞ can be described using an equation

anal-ogous to Eq.(4), where only the independent fitting

parame-ter is changed (fromAxtoAC)

C Tð Þ ¼ AC 1þ 2

exp hcx0 2kBT

 1

0 B

1

Both DCðTÞ and Dx2ðTÞ follow the same functional form

with temperature, leading to a direct linear relationship

between xpeakand C of the form:

xpeakð Þ ¼ xT 0þAx

AC

We expect this linear relationship to hold for all Raman

spectra obtained from a uniformly heated sample in the

re-gime where 3 phonon decay dominates the temperature

dependence of the Raman peak Using Eq.(6), we find from

Figure 4 that Ax/AC¼ 1.21 and x0¼ 307.1 cm1 for the

Agpeak in our nanocrystalline CuO films

In the high temperature limit where hcx0

2BT< 1, Eqs (2), (4), and(5)can be approximated to the form:

xpeak x0þ Axþ4kBAxT

hcx0

(7) and

C Tð Þ  ACþ4kBACT

hcx0

¼ ACþAC

Ax

For the CuO Ag peak, Eqs (7) and (8) are valid for

T > 215 K Figure 5 shows the values obtained for C and

xpeak as a function of temperature, along with linear fits to

the data points obtained at temperatures above 225 K From

these fits, we obtain values of Ax¼ 3.75 cm1 and

AC¼ 3.10 cm1 Using these parameters, we are now able to

describe CðTÞ and xpeakðTÞ across the entire temperature

re-gime studied, and substituting these expressions into Eq.(1)

yields an analytical description of theAgline-shape in CuO,

which is valid up to temperatures of at least 673 K

C Impact of morphology on Raman signal The results of SectionIII Bimply that if the substantial red-shift (16 cm1) observed in Figure 2is due solely to laser heating by the 5 mW probe laser, then the CuO sample beneath the laser spot must experience a substantial rise in the local temperature To examine this effect in detail, we acquired datasets of multiple spectra from our CuO films at two different laser powers under ambient conditions (5 mW and 0.5 mW) A series of spectra were acquired at each laser power level by rastering the laser spot across the film sur-face, with spectra being collected at spacings of 10 lm Raman spectra acquired from each independent location were found to be symmetric about the peak position and hence were analysed by fitting to Eq.(1), following the same procedure as used in SectionIII B Figure6shows the values

of xpeakðTÞ versus CðTÞ obtained from the two datasets obtained using this method Also shown is the uniformly heated temperature-dependent data obtained in SectionIII B (Figure4)

FIG 5 Plot showing values of x peak ðTÞ (circles) and CðTÞ (triangles) versus temperature for the CuO A g peak measured in this work Dotted lines show fits to this data using Eqs (7) and (8) , as described in the text.

FIG 6 Plot showing measured values of xpeakðTÞ and CðTÞ obtained from spatially rastered sampling across the CuO film surface using 2 different laser powers (0.5 mW and 5 mW) Also shown are values obtained from the temperature dependence study (shown earlier in Figure 4 ) and the linear relation (dotted line) derived from that data.

Figure 6shows that data collected using the low laser

power laser (0.5 mW) are tightly clustered around a single

value of xpeak¼ 29761 cm1, which is consistent with the

300 K value measured in Section III B However, at the

higher laser power (5 mW), we observe a strong spatial

vari-ation in the measured values of both the peak position and

line-width, with a difference of22 cm1between the

high-est and lowhigh-est observed values of xpeak We must exclude

the possibility that this spatial variation is caused by material

inhomogeneities that affect the local phonon modes, as we

observe no variation in Raman spectra at the lower laser

power In addition, such material inhomogeneities (e.g.,

impurities,42 defects,43 phonon confinement,44,45

surface-modes46) would all be expected to give rise to an asymmetric

Raman peak, which we do not observe Similarly,

homoge-neous stress (beneath the probe laser) should be expected to

lead to shifts in xpeak that are uncorrelated2with C Again,

this is not consistent with our observations

From Figure 6, we also see that all of the data points

obtained at the higher laser power agree closely with the

lin-ear relationship between xpeak and C obtained in the earlier

external heating study (Figure4) We take this as conclusive

proof that the spatial variation must arise from differences in

the local temperature arising from laser heating under the

probe laser It is highly improbable that any alternative

mechanisms could simulate such a reproducible match

across the entire dataset One aspect to note is that this close

agreement would not be observed if laser heating gave rise

to a highly inhomogeneous temperature profile beneath the

laser spot, as this scenario would cause substantial additional

broadening of the measured peak.47 Instead, our results

imply that all of theAgline-shapes obtained at 5 mW can be

described by a single uniform apparent local temperature

under the focused laser spot This situation is unusual, as the

beam profile of the probe laser is generally expected to

gen-erate a thermal gradient beneath the focused laser spot,5,42,48

which drives heat flow via in-plane thermal conduction We

speculatively suggest that a possible explanation in this case

could be that photo-excitation by the probe laser causes a

substantial increase in local carrier density within our CuO

thin film This would then lead to an increase in the local

thermal conductivity beneath the tightly focused laser spot,

thus promoting temperature equalization within the region

from which spectra are obtained We would expect this

effect be most marked when the dark in-plane thermal

con-ductivity beyond the laser spot remains extremely low, as is

the case studied here

Regardless of the underlying heat flow processes, the

apparent uniform temperature beneath the laser spot can be

determined from C (or xpeak), using the parameters derived

in SectionIII B(as shown in Figure5) Figure5shows that

the linear correlation between these parameters persists

even for values that lie well beyond those which were

experimentally accessible using the external sample heating

stage Extrapolating from Eq.(7), we find that the lowest

experimentally-observed values of xpeak(274 cm1)

cor-responds to a calculated maximum apparent local

tempera-ture that is in excess of 1000 K Figure7shows a histogram

of the same data points binned according to the apparent

temperature The corresponding observed line-width is also plotted on the upper x-axis There is little variation in the observed line-width for the dataset obtained at 0.5 mW; however, two distinct clusters of spectra are observed in the high power dataset; a lower temperature cluster in the range

390 K to 500 K, and a higher temperature cluster between

800 K and 1050 K We believe that this bi-modal distribu-tion reflects the underlying cause of the observed spatial variation in observed Raman spectra, which arises from the numerous micro-bubbles located within the flat planar CuO film (Figure 1) We observe that spectra collected from larger micro-bubbles (which are visible through the optical microscope) correlate with the histogram cluster between

800 and 1050 K, whilst spectra from the flat planar region correlate with values, which are consistent with the values clustered between 390 and 500 K

The local temperature attained by laser heating can be affected by two key parameters that may vary with the micro-morphology of the film: (i) local rate of thermal con-duction away from the laser spot and (ii) focal position of the laser relative to the sample surface In order to distin-guish between the relative contributions from these two com-ponents, we conducted focal-length scans at a micro-bubble site and a planar site, in which the con-focal plane of the Raman microscope was scanned through the surface of the CuO film whilst acquiring spectra at a laser power of 5mW Figure8shows the data obtained from this procedure, where

an axial displacement of zero indicates that the laser is focused on the surface of the CuO film Negative axial dis-placement values indicate that the focal spot is located within the film The profiles obtained from the two sites dif-fer markedly, with the maximum apparent temperature achieved at the micro-bubble site being500 K more than at the flat planar site It should be noted that spectra taken at the two sites under the low laser power of 0.5 mW were iden-tical; hence, the differing behaviour cannot be attributed to variations in local stress within the CuO film Rather, it appears that the local heat transport properties between the

FIG 7 Histogram of A g Raman peak line-width showing the bimodal distri-bution of the line-width data obtained at full laser power The bottom axis shows the line-width converted to an effective temperature using the inverse

of Eq (5)

two sites differ, presumably due to the fact that the CuO film

has become physically detached from the underlying

sap-phire substrate at the micro-bubble site The absence of

ther-mal coupling to the substrate will greatly reduce the rate of

heat conduction away from the laser spot By comparison,

the flat planar regions of the CuO film are strongly coupled

to the underlying substrate and this depresses the maximum

temperature, which can be attained by laser heating at these

sites The dramatic reduction in the local thermal

conductiv-ity at a micro-bubble site is such that this effect continues

to be observed even when the focal plane is displaced by

>20 lm from the sample surface

The data shown in Figure 7were acquired using a

sim-ple raster motion-control and as a result we would expect to

see some variation in focal distance across the sample due to

variations in the surface height of the CuO film The

high-lighted region in Figure 8corresponds to variations in

sur-face height of up to 5 lm, which is consistent with sursur-face

profile measurements of our samples (Section III A) The

temperature values which lie within the highlighted regions

closely match the two histogram clusters (at 390–500 K and

800–1050 K) shown in Figure7 As such, we infer that the

observed bi-modal distribution of temperatures is a result of

the different extent of laser heating experienced at

micro-bubble sites and flat planar sites, respectively

IV CONCLUSION

TheAgRaman peak from our nano-crystalline CuO thin

films exhibits a strong power-dependence, which is due to

local heating effects caused by the focused probe laser We

have carried out a detailed study of the

temperature-dependence of this peak and shown that it can be fitted using

a pseudo-Voigt line-shape over a wide range of

tempera-tures We observe a linear relationship between the peak

position and line-width, which holds for the entire range of

temperatures studied here, and we note that this behavior is

characteristic of any material in which the dominant

temperature-dependent contribution to the Raman peak line-shape is due to anharmonic 3-phonon decay via the Klemens process We report material dependent parameters, which provide an analytical description of the temperature depend-ence of the CuO AgRaman peak that is valid for our nano-crystalline CuO film in the temperature range between 300 K and 650 K

We have then shown that local laser heating under a

5 mW probe laser leads to values of xpeakand C, which indi-cate an apparent uniform local temperature beneath the laser spot From this observation, we conclude that the observed power dependence of the Ag peak is entirely due to local laser heating of the sample We extrapolate that the maxi-mum apparent temperature attained by the CuO film due to laser heating by the focused 5 mW, 633 nm, CW, laser exceeded 1000 K

Finally, we observe a strong spatial variation in the shape and position of the Agpeak obtained from across our CuO film This is due to variation in the local heat transport rates within the film—which determine the temperature reached due to laser heating Local heat transport rates are dramatically decreased when the film becomes detached from the underlying substrate, and this occurs at micro-bubble sites within our samples We highlight the need for caution when interpreting the Raman spectra of any nano-structured material, as heat transport within this class of materials should be expected to exhibit significant inhomo-geneity at the microscale In particular, several previous reports of red-shifting and broadening in the Raman spectra

of nanostructured CuO samples may simply be due to local laser heating effects rather than the phonon-confinement effects, which have been suggested by other authors.17–26

1

H Richter, Z Wang, and L Ley, Solid State Commun 39, 625–629 (1981).

2

T Beechem, S Graham, S P Kearney, L M Phinney, and J R Serrano,

Rev Sci Instrum 78, 061301 (2007).

3

D M Rie, Phys Rev B 84, 064308 (2011).

4

G Faraci, S Gibilisco, and A R Pennisi, Phys Rev B 80, 193410 (2009).

5

S Kouteva-Arguirova, Tz Arguirov, D Wolfframm, and J Reif, J Appl Phys 94, 4946–4949 (2003).

6

H Scheel, S Reich, A C Ferrari, M Cantoro, A Colli, and C Thomsen,

Appl Phys Lett 88, 233114 (2006).

7

Z Wang, J E Alaniz, W Jang, J E Garay, and C Dames, Nano Lett 11, 2206–2213 (2011).

8

K Adu, H Gutierrez, U Kim, and P Eklund, Phys Rev B 73, 155333 (2006).

9

R Jalilian, G U Sumanasekera, H Chandrasekharan, and M Sunkara,

Phys Rev B 74, 155421 (2006).

10 X Jiang, T Herricks, and Y Xia, Nano Lett 2, 1333–1338 (2002) 11

F Teng, W Yao, Y Zheng, Y Ma, Y Teng, T Xu, S Liang, and Y Zhu,

Sens Actuators, B 134, 761 (2008).

12

F Bayansal, S Kahraman, G Cankaya, H A Cetinkara, H S Guder, and

H M Cakmak, J Alloys Compd 509, 2094–2098 (2010).

13

F Marabelli, G Parravicini, and F Salghetti-Drioli, Phys Rev B 52, 1433–1436 (1995).

14

S Wang, C Hsiao, S Chang, K Lam, K Wen, S Hung, S Young, and B Huang, Sens Actuators, A 171, 207–211 (2011).

15

D Li, J Hu, R Wu, and J G Lu, Nanotechnology 21, 485502 (2010) 16

J Chen, N Y Huang, S Z Deng, J C She, N S Xu, W Zhang, X Wen, and S Yang, Appl Phys Lett 86, 151107 (2005).

17

M H Chou, S B Liu, C Y Huang, S Y Wu, and C L Cheng, Appl Surf Sci 254, 7539–7543 (2008).

FIG 8 Calculated temperature obtained from the observed Raman spectra

as the confocal height of the Raman microscope was scanned through the

sample surface at two different sites located at (i) a micro-bubble (crosses)

and (ii) a flat planar region (circles) An axial displacement of zero denotes

the laser was focused on the film surface and negative values denote the

focal point was below the surface (i.e., within the film).

18 Y L Huang, M H Chou, S Y Wu, and C L Cheng, Jpn J Appl Phys.,

Part 1 47, 703–705 (2008).

19

J F Xu, W Ji, Z X Shen, W S Li, S H Tang, X R Ye, D Z Jia, and

X Q Xin, J Raman Spectrosc 30, 413–415 (1999).

20 H Fan, B Zou, Y Liu, and S Xie, Nanotechnology 17, 1099 (2006).

21 A Das, B Venkataramana, D Partheephan, A K Prasad, S Dhara, and

A K Tyagi, Physica E 54, 40–44 (2013).

22

F Emanuela, T Marco, S Tiziana, C Daniela, M Cosimo, and G M.

Rachele, Physica E 60, 59–64 (2014).

23 W Wang, Y Zhuang, and L Li, Mater Lett 62, 1724–1726 (2008).

24

W Wang, L Wang, H Shi, and Y Liang, CrystEngComm 14, 5914–5922

(2012).

25 U Kumar Gaur, A Kumar, and G D Varma, CrystEngComm 16,

3005–3014 (2014).

26

C L Chen, Y R Ma, M H Chou, C Y Huang, V Yeh, and S Y Wu,

Nanotechnology 18, 245604 (2007).

27 J Chrzanowski and J C Irwin, Solid State Commun 70, 11–14 (1989).

28

W Reichardt, F Gompf, M Ain, and B M Wanklyn, Z Phys B:

Condens Matter 81, 19–24 (1990).

29

S Ravi, A B Kaiser, and C W Bumby, Mater Sci Forum 700, 71–79 (2012).

30 E Le Ru, E Blackie, M Meyer, and P Etchegoin, J Phys Chem C 111,

13794–13803 (2007).

31

N A Mohemmed Shanid, M Abdul Khadar, and V G Sathe, J Raman

Spectrosc 42, 1769–1773 (2011).

32 G K Wertheim, M A Butler, K W West, and D N E Buchanan, Rev.

Sci Instrum 45, 1369–1371 (1974).

33 P G Klemens, Phys Rev 148, 845–848 (1966).

34

H Tang and I P Herman, Phys Rev B 43, 2299–2304 (1991).

35

M Gu, Y Zichou, L Pan, Z Sun, S Wang, and C Q Sun, J Appl Phys.

102, 083524 (2007).

36 W S Li, Z X Shen, Z C Feng, and S J Chua, J Appl Phys 87, 3332–3337 (2000).

37

X B Chen, J Huso, J L Morrison, L Bergman, and A P Purdy, J Appl Phys 98, 026106 (2005).

38 X D Pu, J Chen, W Z Shen, H Ogawa, and Q X Guo, J Appl Phys.

98, 033527 (2005).

39

J Menendez and M Cardona, Phys Rev B 29, 2051–2059 (1984) 40

M Ohashi, A Tashiro, G Oomi, E Maeda, and X G Zheng, Phys Rev B

73, 134421 (2006).

41 W Demtroder, Laser Spectroscopy: Basic Concepts and Instrumentation (Springer-Verlag, 2003).

42

O Brafman and R Manor, Phys Rev B 51, 6940–6946 (1995).

43 S Nakashima, S Oima, A Mitsuishi, T Nishimura, T Fukumoto, and Y Akasaka, Solid State Commun 40, 765–768 (1981).

44

M N Islam, A Pradhan, and S Kumar, J Appl Phys 98, 024309 (2005).

45 P Mishra and K P Jain, Phys Rev B 62, 14790–14795 (2000).

46 I M Tiginyanu, G Irmer, J Monecke, and H L Hartnagel, Phys Rev B

55, 6739–6742 (1997).

47

H W Lo and A Compaan, J Appl Phys 51, 1565 (1980).

48 G D Pazonis, H Tang, P Irving, and M Herman, IEEE J Quantum Electron 25(5), 976–988 (1989).