Effects of rapid thermal annealing on electrical, optical, and structuralproperties of Ni-doped In2O3 anodes for bulk heterojunction organicsolar cells
Jun Ho Kim and Tae-Yeon SeongDepartment of Materials Science and Engineering, Korea University, Seoul 136-713, South Korea
Han-Ki Kima)
Department of Advanced Materials Engineering for Information and Electronics, Kyung Hee University,1 Seocheon-dong, Yongin-si, Gyeonggi-do 446-701, South Korea
(Received 30 August 2012; accepted 14 December 2012; published 7 January 2013)
The authors investigated the effects of rapid thermal annealing (RTA) on the electrical, optical, and
structural properties, and work functions of Ni-doped In2O3 (INO) anodes prepared by a DC/RF
co-sputtering process for use in bulk heterojunction organic solar cells (OSCs). By RTA processing
at 600 �C, the authors obtained the optimized INO anodes with a sheet resistance of 28 X/sq, anoptical transmittance of 82.93%, and a work function of 5.02 eV, which are acceptable in OSC
fabrication. In particular, the 600 �C annealed INO anode showed much higher opticaltransmittance in the near infrared wavelength region than the conventional ITO film, even though it
had a low resistivity of 5.66� 10�4 X cm. The OSC fabricated on the annealed INO anode showeda higher power convention efficiency of 2.65% than the OSC with as-deposited INO anodes
(2.19%) because the fill factors of the OSC are critically dependent on the sheet resistance of the
anode. VC 2013 American Vacuum Society. [http://dx.doi.org/10.1116/1.4774212]
I. INTRODUCTION
Cost-efficient bulk heterojunction organic solar cells
(OSCs) have been widely investigated due to their potential
applications as disposable energy pads, portable energy sour-
ces, and attachable energy sources.1–5 In particular, the merits
of OSCs, such as light weight, simple structure, simple
printing-based processing, and superior flexibility make
OSCs candidates for next-generation photovoltaics. Since
Tang reported the possibility of OSCs in 1986, the power
conversion efficiency (PCE) of OSCs rapidly increased to
8%–9%, due the rapid advance of organic materials, process-
ing, and device structure technologies.6,7 The increase of
PCE in OSC allows the mass production of OSCs in the near
future. Among several key parameters affecting the PCE of
OSCs, the electrical and optical properties of the anode are
very important because exciton formation and current extrac-
tion are critically influenced by the optical transmittance and
sheet resistance of the anodes. Although DC or RF sputtered
Sn-doped In2O3 (ITO) films have been widely used in OSCs
as transparent anodes, further investigations to develop new
transparent anodes are necessary to improve the performance
of OSCs. For this purpose, oxide-based anode materials, such
as In-Zn-O, In-Ge-O, Ti-In-Sn-O, In-Zn-Sn-O, Nb-Ti-O,
Ga-Zn-O, Al-Zn-O, and In-W-O films, were suggested as
promising anode materials.8–15 Ni-doped In2O3 or Ni-doped
ITO films are also known as transparent oxide electrodes
(TCOs) for optoelectronics.16,17 Hsu et al. reported that theeffective doping of Ni into ITO led to a surface work function
of 5.80 eV.18 They explained that the high work function
of the Ni-doped ITO film is closely related to the effective
doping of Ni into the ITO matrix. Adachi et al. also reportedthat the Ni-doped ITO film acts as an effective anode for or-
ganic light emitting diodes (OLEDs).19 OLEDs with Ni-
doped ITO anodes showed comparable performance to
OLEDs with conventional ITO anodes due to the low sheet
resistance and high work function of the Ni-doped ITO an-
ode. In our previous work, a very thin NiO film on an indium
zinc oxide (IZO) film enhanced the performance of OLEDs
due to the high work function of the NiO layer covering the
IZO anode.20 Although the electrical and optical properties
of Ni-doped ITO films used as transparent electrodes were
investigated, detailed studies of Ni-doped In2O3 (INO) films
are lacking. In addition, applications of INO films in OSCs as
anode have not yet been reported. Furthermore, a detailed
investigation of the effects of rapid thermal annealing (RTA)
on the electrical, optical, structural, and surface properties
and work functions of the INO film has not been reported.
In this work, we investigated INO films prepared by a
NiO and In2O3 co-sputtering process for OSC applications
as transparent anodes. By optimizing RTA temperature, we
obtain an INO film with a resistivity of 5.66� 10�4 X cmand an optical transparency of 82.93%, which are acceptable
values for anodes in OSCs. In addition, we compared the
performances of OSCs fabricated on the as-deposited and
annealed INO film to correlate sheet resistance of the anode
and PCE value of OSCs.
II. EXPERIMENT
The 200 nm thick INO films were grown on glass sub-
strates by using DC/RF magnetron sputtering using NiO and
In2O3 ceramic targets at room temperature. Prior to co-
sputtering, NiO and In2O3 targets were presputtered in an Ar
ambient for 20 min in order to remove impurities anda)Electronic mail: [email protected]
021201-1 J. Vac. Sci. Technol. A 31(2), Mar/Apr 2013 0734-2101/2013/31(2)/021201/5/$30.00 VC 2013 American Vacuum Society 021201-1
Author complimentary copy. Redistribution subject to AIP license or copyright, see http://jva.aip.org/jva/copyright.jsp
http://dx.doi.org/10.1116/1.4774212http://dx.doi.org/10.1116/1.4774212http://dx.doi.org/10.1116/1.4774212http://dx.doi.org/10.1116/1.4774212http://dx.doi.org/10.1116/1.4774212http://dx.doi.org/10.1116/1.4774212mailto:[email protected]://crossmark.crossref.org/dialog/?doi=10.1116/1.4774212&domain=pdf&date_stamp=2013-01-07
contamination on the surface of the targets. The INO films
were co-sputtered at an optimized RF power of 10 W applied
to a NiO target and a DC power of 100 W applied to an
In2O3 target. During the co-sputtering of the NiO and In2O3target, the working pressure of 3 mTorr and Ar flow rate of
10 sccm maintained constant. To obtain uniformly sputtered
INO films, 45� tilted multicathode guns were employed.During the co-sputtering process, the glass substrate was
constantly rotated at a speed of 20 rpm to ensure a uniform
thickness of the INO film. The thickness of the INO film was
controlled by manipulating co-sputtering time. After deposi-
tion of the INO film at room temperature, the INO films
were rapidly thermal annealed in a vacuum ambient for
10 min at 400, 500, and 600 �C to improve the electrical andoptical properties of the INO films. The thickness of the INO
films was measured by means of a surface profiler (a-step,HTSK). Electrical properties of the INO films were exam-
ined by Hall measurement (HL5500PC, Accent optical tech-
nology) as a function of RTA temperature. In addition,
optical transmittance of the INO films was measured using a
UV/visible spectrometer (UV 540, Unicam) as a function of
the RTA temperature in the wavelength region from 200 to
1500 nm. Surface morphology of the as-deposited and
annealed INO films was analyzed by means of atomic force
microscopy (AFM, PUCO Station STD). The microstructure
of the INO films was examined by x-ray diffraction (XRD,
D/Max 2500, Rigaku) using Cu Ka radiation (k¼ 1.54 Å). Inaddition, the work function of the INO films was measured
by using a Kelvin probe (KP Technology) as a function of
RTA temperature.
To correlate the sheet resistance of the INO anode and
PCE of the OSCs, we fabricated conventional bulk hetero-
junction OSCs on the as-deposited INO and 600 �C annealedINO films. After wet cleaning and 5 min UV/ozone treatment
of the INO films, poly (3,4-ethylenedioxythiophene):poly
(styrenesulfonate) (PEDOT:PSS, Clevios PH510) was spin-
coated on the as-deposited and annealed INO anodes and
subsequently annealed for 10 min at 120 �C in air. Then theactive layer (50-mg P3HT, 50-mg PCBM/2-ml dichloroben-
zene) was deposited on the PEDOT:PSS layers under a nitro-
gen atmosphere. Organic materials using a standard active
layer blend with poly(3-hexylthiophene) (P3HT) as electron-
donor and a fullerene derivative [6,6]-phenyl-C61-butyric
acid methyl ester (PCBM) as electron-acceptor. Finally, a
Ca/Al (20/100 nm) cathode with an area of 4.66 mm2 was de-
posited on the active layer using thermal evaporation. The
photocurrent density–voltage (J-V) curves of the OSCs fabri-
cated on the as-deposited and 600 �C annealed INO anodeswere measured using a Keithley 1200 measurement unit
under 100 mW/cm2 illumination with AM 1.5 G conditions.
III. RESULTS AND DISCUSSION
Figure 1 shows the optical transmittances of the INO
films as a function of RTA temperature in the wavelength
region from 200 to 1500 nm. The as-deposited INO films
exhibited an optical transmittance of 71.76% at 550 nm and
an average transmittance of 78.92% between 400 and
800 nm. However, after the rapid thermal annealing process,
the INO films showed significantly improved optical trans-
mittance regardless RTA temperature. For example, the
600 �C annealed INO film exhibited an optical transmittanceof 74.61% at 550 nm and an average transmittance of
82.93% between 400 and 800 nm. Furthermore, the annealed
INO film showed fairly high transmittance in the near infra-
red (NIR) wavelength region, due to the low Ni doping con-
centration. In general, the transparency in the NIR region is
closely related to the plasma absorption frequency, which
depends on the carrier density and the effective mass of the
carrier.21 The conventional ITO anode shows fairly low opti-
cal transmittance in the NIR region since it has a high free
carrier concentration. Therefore, the conventional ITO film
is not a desirable anode for low band-gap active materials or
tandem structure OSCs.22 However, the high transmittance
of the annealed INO films in the NIR wavelength region
indicates that the INO film is an appropriate anode for OSCs
using an organic active layer with a low band-gap.
Figure 2(a) exhibits the Hall measurement results of the
INO films as a function of RTA temperature. An as-
deposited INO film with a thickness of 200 nm showed a
resistivity of 9.51� 10�4 X cm and a sheet resistance of48 X/sq. Even though it was prepared at room temperature,as-deposited INO film showed fairly low resistivity. Below a
RTA temperature of 500 �C, the INO films showed similarresistivity and sheet resistivity to the as-deposited INO film.
However, the 600 �C annealed INO film showed a decreasedresistivity of 5.66� 10�4 X cm and a sheet resistance of28 X/sq. The decrease in the resistivity and sheet resistanceof the 600 �C annealed INO film could be attributed to theformation of oxygen vacancies, which act as donors in the
In2O3 matrix. As discussed by Hsu et al., the low resistivityof the INO film could be due to the curing of structural
imperfections after annealing.18 Adachi et al. also reportedthat the low sheet resistance of INO films could be attributed
to crystal growth with increasing temperature.19 Based on
the optical transmittance (T) in Fig. 1 and sheet resistance
(Rsh) in Fig. 2(a), we calculated the figure of merit (FOM)
using the equation below to determine the optimum RTA
temperature23
FIG. 1. (Color online) Optical transmittance of the INO films as a function of
the RTA temperature.
021201-2 Kim, Seong, and Kim: Effects of rapid thermal annealing on electrical, optical, and structural properties 021201-2
J. Vac. Sci. Technol. A, Vol. 31, No. 2, Mar/Apr 2013
Author complimentary copy. Redistribution subject to AIP license or copyright, see http://jva.aip.org/jva/copyright.jsp
FOM ¼ T10=Rsh; (1)
where T is the transmittance at 550 nm and Rsh is the sheet
resistance of the INO films. As reported by Kaleemulla
et al., the FOM is an important value in determining thequality of the TCO films.24 With increasing RTA tempera-
ture, the INO showed increased FOM values, as shown in
Fig. 2(b). Compared to the as-deposited INO film, the
600 �C annealed INO film showed a higher FOM value(1.91� 10�3 X�1). Therefore, we decided that 600 �C is theoptimum RTA temperature.
The surface morphologies of the INO films were analyzed
by means of AFM. Figure 3(a) showed three-dimensional
AFM surface images of the INO films as a function of RTA
temperature. All INO films showed a smooth morphology on
a scale of 10� 10 lm2 without surface defects. Figure 3(b)shows the root mean square (RMS) surface roughness of the
INO films as a function of RTA temperature. Compared to
the as-deposited INO film (2.89 nm), the 600 �C annealedINO film shows a decreased RMS value of 2.57 nm. How-
ever, as expected from the AFM surface images, all INO
film showed similar RMS roughness between 2.89 and
3.19 nm. This indicates that the surface morphology of the
INO film is not affected by RTA temperature. Considering
the spin coat processing of the organic layer, such as
PEDOT:PSS and P3HT:PCBM on the anode electrode, the
smoothness of the anode film is very important. Therefore,
the INO anode with a smooth surface morphology may be
adaptable in the organic layer coating process.
Figure 4 shows the XRD plots of the INO films deposited
on glass substrate as a function of RTA temperature. The dif-
fraction peaks can be indexed to the In2O3 (JCPDS Card No.
06-0416).25 The as-deposited INO film shows crystalline
peaks at 2h values of 21.60� (211), 30.72� (222), and 51.28�
(440). This indicates that the structure of the INO film is
bixbyite like In2O3. However, the intensity of the crystalline
peaks is fairly weak because the INO film was prepared at
FIG. 2. (Color online) (a) Resistivity and sheet resistance of the INO films as
a function of the RTA temperature. (b) Figure of merit values of the INO
films calculated from the sheet resistance (Rsh) and transmittance at 550 nm
(T) as a function of the RTA temperature.
FIG. 3. (Color online) (a) Surface AFM images (3D) of the INO films as a
function of the RTA temperature. (b) RMS roughness of the INO films as a
function of the RTA temperature.
FIG. 4. (Color online) XRD plots of the INO films as a function of the RTA
temperature.
021201-3 Kim, Seong, and Kim: Effects of rapid thermal annealing on electrical, optical, and structural properties 021201-3
JVST A - Vacuum, Surfaces, and Films
Author complimentary copy. Redistribution subject to AIP license or copyright, see http://jva.aip.org/jva/copyright.jsp
room temperature. The increase in RTA temperature led to
the strongly h222i preferred orientation in the INO film. The(222) XRD peak of the INO showed doublet peaks. These
doublet peaks is attributed to double layer in INO films. As
shown below references,26 the double layer, ITO films with
double layer structure showed the doublet (222) peaks. Like
ITO films, the INO films showed doublet (222) peak because
the RTA process could lead to the formation of doublet
structure. Doublet layer in the annealed INO film can be
attributed to the directional irradiation in the RTA system.
Above 400 �C, all INO films showed strong (222) peak in-tensity, which is similar behavior to that of the conventional
ITO film.27 However, there are no peaks corresponding to
the NiO phase, indicating that the Ni dopant was completely
incorporated into the In2O3 lattice. The average grain size
(D) of the INO thin films was calculated using the full-width
at half-maximum (FWHM) of the (222) peak from the
Scherrer equation28
D ¼ 0:9 k=ðb cos hÞ; (2)
where k is the x-ray wavelength, b is the FWHM of the(222) diffraction peak, and h is the Bragg angle of diffractionpeak. The grain sizes of the INO films calculated from the
above equation are 27.38, 37.44, 34.30, and 34.32 nm,
respectively, with increasing RTA temperature. Compared to
the as-deposited INO film, the annealed INO films showed
larger grain size due to grain growth of the INO film during
the RTA process.
Figure 5 shows the work function of the INO films meas-
ured by Kelvin probe as a function of RTA temperature. The
work function of the INO film increased with increasing
RTA temperature. Compared to the as-deposited INO film
(4.85 eV), the 600 �C annealed INO film has a higher workfunction of 5.02 eV, which is higher than the conventional
ITO film (4.5–4.8 eV). As discussed by Hsu et al., the workfunctions of the Ni-doped ITO anode were affected by
Ni-doping concentration.29 In order to improve the perform-
ance photovoltaics, it is necessary to use a TCO anode with
a higher work function greater than 5 eV, because hole
extraction depends on the barrier height formed between the
anode and the organic materials.30 Therefore, we calibrated
the Kelvin probe measurements before measuring the work
function of the INO films. Whenever we measured the work
function of the TCO films, we calibrated the Kelvin probe
using Au reference electrode (5.1 eV). When we measured
the work function of the INO film, both INO and Au refer-
ence electrodes with different work function are brought to-
gether.31 The tip vibrates with a amplitude of 50 to 2 mm at
a frequency of 80 Hz. Inset of Fig. 5 shows the energy band
diagram of the OSC fabricated on the 600 �C annealed INOanode. Due to the high work function of annealed INO film,
we expect effective carrier extraction from the active organic
layer.
To investigate the effects of anode sheet resistance on the
performance of OSCs, we fabricated conventional OSCs
on the as-deposited and 600 �C annealed INO anodes. Bulk-heterojunction OSC had a structure of INO anode/
PEDOT:PSS/P3HT:PCBM/Ca/Al cathode. Figure 6 shows
representative current density–voltage (J-V) characteristics
of OSCs fabricated on the as-deposited and 600 �C annealedINO anodes with an inset of a cross-sectional FE-SEM
image of the OSCs with the INO anode. Detailed performan-
ces of the OSCs are summarized in Table I. The OSC fabri-
cated on the as-deposited INO anode showed an open circuit
voltage (VOC) of 0.56 V, a short circuit current density (JSC)
of 7.53 mA/cm2, a fill factor (FF) of 52.09%, and a PCE of
2.19%. However, the OSC with the 600 �C annealed INO an-ode exhibited better performance, i.e., a VOC of 0.58 V, a JSCof 7.95 mA/cm2, a FF of 57.41%, and a PCE of 2.65%. In
our previous research, we reported that the sheet resistance
of IZTO anodes with different thicknesses critically affected
the performance of OSCs.32 Like the IZTO anode, the sheet
resistance of the INO anode critically affected the perform-
ance of OSCs. The series resistance of OSCs critically
affects the slope of the J-V curve at J¼ 0 mA/cm2. A higherslope of OSC with an annealed INO anode than that of the
OSC with an as-deposited INO anode indicates that the sheet
FIG. 5. (Color online) Work function of the INO films as a function of the
RTA temperature with energy band diagram for the OSC device fabricated
with the INO electrodes.
FIG. 6. (Color online) Current density (J)–voltage (V) characteristics of
OSCs fabricated on the as-deposited INO, annealed INO, and reference ITO
electrodes. The inset shows a cross-sectional FE-SEM image of the OSCs
with INO anode.
021201-4 Kim, Seong, and Kim: Effects of rapid thermal annealing on electrical, optical, and structural properties 021201-4
J. Vac. Sci. Technol. A, Vol. 31, No. 2, Mar/Apr 2013
Author complimentary copy. Redistribution subject to AIP license or copyright, see http://jva.aip.org/jva/copyright.jsp
resistance and contact resistance between organic layer and
INO anode was reduced by means of the RTA process at
600 �C. The FF and JSC of the OSC critically depended onthe series resistance and transmittance of the transparent
electrode. Therefore, the improved of the FF and JSC of the
OSCs with the annealed INO anode could be attributed to
the low sheet resistance and high transparency of the
annealed INO anode.
IV. SUMMARY AND CONCLUSIONS
We report the effect of RTA on the electrical, optical,
structural properties, surface, and work function of INO ano-
des for use in OSCs. We found that the 600 �C annealing ofthe INO anode resulted in a sheet resistance of 28 X/sq, anoptical transmittance of 82.93% and a work function
5.02 eV, which are comparable to conventional ITO anodes.
In particular, the INO film shows higher transmittance in the
NIR region due to a low free carrier concentration. In addi-
tion, the INO film showed a fairly smooth surface regardless
of the RTA temperature, indicating stable surface properties.
The OSCs fabricated on a 600 �C annealed INO anodeshowed better performance that the OSC with as-deposited
INO anode due to the reduced sheet resistance and increased
optical transmittance.
ACKNOWLEDGMENT
This work was supported by the New & Renewable
Energy of Korea Institute of Energy Technology Evaluation
and Planning (KETEP) Grant No. (2011T100200034)
funded by the Korea government Ministry of Knowledge
Economy.
1S. Gunes, H. Neugebauer, and N. S. Sariciftci, Chem. Rev. 107, 1324(2007).
2J. Chen and Y. Cao, Acc. Chem. Res. 42, 1709 (2009).3Y.-J. Cheng, S.-H. Yang, and C.-S. Hsu, Chem. Rev. 109, 5868 (2009).4D. W€ohrle and D. Meissner, Adv. Mater. 3, 129 (1991).5H. Hoppe and N. S. Sariciftci, J. Mater. Res. 19, 1924 (2004).6C. W. Tang, Appl. Phys. Lett. 48, 183 (1986).7G. Li, R. Zhu, and Y. Yang, Nature Photon. 6, 153 (2012).8C.-K. Cho and H.-K. Kim, J. Nanosci. Nanotechnol. 12, 3346 (2012).9S.-B. Kang, J.-W. Lim, S. h. Lee, J.-J. Kim, and H.-K. Kim, J. Phys. D:
Appl. Phys. 45, 325102 (2012).10J.-W. Lim, S. J. Kang, S. h. Lee, J.-J. Kim, and H.-K. Kim, J. Appl. Phys.
112, 023513 (2012).11J. h. Kim, S.-I. Na, and H.-K. Kim, Sol. Energy Mater. Sol. Cells 98, 424
(2012).12J.-H. Park, Y.-Y. Choi, H. H. Lee, and H.-K. Kim, J. Appl. Phys. 108,
083509 (2010).13J.-H. Park, K.-J. Ahn, S.-I. Na, and H.-K. Kim, Sol. Energy Mater. Sol.
Cells 95, 657 (2011).14J.-H. Park, K.-J. Ahn, K.-I. Park, S.-I. Na, and H.-K. Kim, J. Phys. D:
Appl. Phys. 43, 115101 (2010).15J. H. Kim, Y.-H. Shin, T.-Y. Seong, S.-I. Na, and H.-K. Kim, J. Phys. D:
Appl. Phys. 45, 395104 (2012).16C.-M. Hsu, C.-F. Liu, H.-E. Cheng, and W.-T. Wu, J. Electron. Mater. 35,
383 (2006).17A. Nakasa, M. Adachi, H. Usami, E. Suzuki, and Y. Taniguchi, Thin Solid
Films 498, 240 (2006).18C.-M. Hsu, J.-W. Lee, T.-H. Meen, and W.-T. Wu, Thin Solid Films 474,
19 (2005).19M. Adachi, A. Nakasa, H. Usami, E. Suzuki, A. Hiratai, and Y. Taniguchi,
Jpn. J. Appl. Phys., Part 1 45, 584 (2006).20K.-H. Choi et al., J. Electrochem. Soc. 155, J340 (2008).21D. S. Ginly, H. Hosono, and D. C. Paine, Handbook of Transparent Con-
ductors (Springer, New York, 2010).22Y.-H. Shin, S.-I. Na, and H.-K. Kim, J. Vac. Sci. Technol. A. 30, 061507
(2012).23G. Haacke, J. Appl. Phys. 47, 4086 (1976).24S. Kaleemulla, A. S. Reddy, S. Uthanna, and P. S. Reddy, Optoelectron.
Adv. Mater. 2, 782 (2008).25B. D. Cullity, Elements of X-Ray Diffraction, 2nd ed. (Addison-Wesley,
USA, 1978).26D. G. Neerinck and T. J. Vink, Thin Solid Films 278, 12 (1996).27R. N. Joshi, V. P. Singh, and J. C. McClure, Thin Solid Films 257, 32
(1995).28A. L. Patterson, Phys. Rev. 56, 978 (1939).29C.-M. Hsu and W.-T. Wu, Appl. Phys. Lett. 85, 840 (2004).30C.-M. Hsu, W.-T. Wu, and H.-H. Lee, Displays 29, 268 (2008).31R. Schlaf, H. Murata, and Z. H. Kafafi, J. Electron Spectrosc. 120, 149
(2001).32K.-H. Choi, J.-A. Jeong, and H.-K. Kim, Sol. Energy Mater. Sol. Cells 94,
1822 (2010).
TABLE I. Cell performance of OSCs fabricated on the as-deposited INO,
annealed INO, and reference ITO electrodes.
VOC (V) JSC (mA/cm2) FF (%) PCE (%)
As-deposited INO 0.56 7.53 52.09 2.19
Annealed INO 0.58 7.95 57.41 2.65
ITO (reference) 0.60 8.98 60.33 3.26
021201-5 Kim, Seong, and Kim: Effects of rapid thermal annealing on electrical, optical, and structural properties 021201-5
JVST A - Vacuum, Surfaces, and Films
Author complimentary copy. Redistribution subject to AIP license or copyright, see http://jva.aip.org/jva/copyright.jsp
http://dx.doi.org/10.1021/cr050149zhttp://dx.doi.org/10.1021/ar900061zhttp://dx.doi.org/10.1021/cr900182shttp://dx.doi.org/10.1002/adma.19910030303http://dx.doi.org/10.1557/JMR.2004.0252http://dx.doi.org/10.1063/1.96937http://dx.doi.org/10.1038/nphoton.2012.11http://dx.doi.org/10.1166/jnn.2012.5629http://dx.doi.org/10.1088/0022-3727/45/32/325102http://dx.doi.org/10.1088/0022-3727/45/32/325102http://dx.doi.org/10.1063/1.4737602http://dx.doi.org/10.1016/j.solmat.2011.11.050http://dx.doi.org/10.1063/1.3499282http://dx.doi.org/10.1016/j.solmat.2010.09.036http://dx.doi.org/10.1016/j.solmat.2010.09.036http://dx.doi.org/10.1088/0022-3727/43/11/115101http://dx.doi.org/10.1088/0022-3727/43/11/115101http://dx.doi.org/10.1088/0022-3727/45/39/395104http://dx.doi.org/10.1088/0022-3727/45/39/395104http://dx.doi.org/10.1007/BF02690523http://dx.doi.org/10.1016/j.tsf.2005.07.102http://dx.doi.org/10.1016/j.tsf.2005.07.102http://dx.doi.org/10.1016/j.tsf.2004.08.005http://dx.doi.org/10.1143/JJAP.45.584http://dx.doi.org/10.1149/1.2979143http://dx.doi.org/10.1116/1.4758789http://dx.doi.org/10.1063/1.323240http://dx.doi.org/10.1016/0040-6090(95)08117-8http://dx.doi.org/10.1016/0040-6090(94)06331-1http://dx.doi.org/10.1103/PhysRev.56.978http://dx.doi.org/10.1063/1.1777416http://dx.doi.org/10.1016/j.displa.2007.08.009http://dx.doi.org/10.1016/S0368-2048(01)00310-3http://dx.doi.org/10.1016/j.solmat.2010.05.056
s1s2n1s3d1f1f2af2bf2f3af3bf3f4d2f5f6s4c1c2c3c4c5c6c7c8c9c10c11c12c13c14c15c16c17c18c19c20c21c22c23c24c25c26c27c28c29c30c31c32t1