JOURNAL OF APPLIED PHYSICS 107, 033514 共2010兲
Hydrogen-doped In2O3 transparent conducting oxide films prepared by solid-phase crystallization method Takashi Koida,1,a兲 Michio Kondo,1 Koichi Tsutsumi,2 Akio Sakaguchi,2 Michio Suzuki,2 and Hiroyuki Fujiwara3 1
Research Center for Photovoltaics, National Institute of Advanced Industrial Science and Technology (AIST), Central 2, Umezono 1-1-1, Tsukuba, Ibaraki 305-8568, Japan 2 J. A. Woolam Japan Corporation, Fuji 2F, 5-22-9, Ogikubo, Suginami, Tokyo 167-0051, Japan 3 Center for Innovative Photovoltaic Center, Gifu University, 1-1, Yanagito, Gifu 501-1193, Japan
共Received 14 October 2009; accepted 6 December 2009; published online 8 February 2010兲 We have characterized amorphous to crystalline transformation of hydrogen 共H兲-doped In2O3 共In2O3 : H兲 films by transmission electron microscopy, thermal desorption spectroscopy, spectroscopic ellipsometry, and Hall measurements. The In2O3 : H films that show a mixed-phase structure embedded with small density of crystalline grains in a large volume fraction of amorphous phase have been fabricated at room temperature by the sputtering of an In2O3 ceramic target with introduction of H2O vapor, and the films have been postannealed in vacuum to crystallize the amorphous phase. With increasing annealing temperature up to 200 ° C, the film shows a large increase in Hall mobility 共Hall兲 from 42 to 110 cm2 / V s and a decrease in carrier density 共NHall兲 from 4.6⫻ 1020 to 2.1⫻ 1020 cm−3 with slight decrease in resistivity. The change in Hall and NHall with annealing temperature is strongly correlated with the volume fractions of the amorphous and crystalline phases in the films. Analyses of dielectric functions of the films using the Drude model revealed that the high electron mobility in the crystallized films is attributed mainly to longer relaxation time rather than smaller effective mass, as compared with as-deposited films. Temperature-dependent Hall analysis, relationship between NHall and Hall, and comparison between Hall and optical mobility showed that 共i兲 scattering processes inside amorphous and/or crystalline matrices limit the mobility, 共ii兲 doubly charged ionized impurity scattering is reduced by crystallization, and 共iii兲 phonon scattering becomes dominant after crystallization in the In2O3 : H films. The above results suggest that H-doping reduces carrier scattering in the crystallized In2O3 : H and structural rearrangements during crystallization eliminate oxygen deficiency and generate H+ that acts as a singly charged donor. In this article, we discuss the transport properties with the variation in microscopic and chemical structures in the In2O3 : H films. © 2010 American Institute of Physics. 关doi:10.1063/1.3284960兴 I. INTRODUCTION
Transparent conducting oxide 共TCO兲 films have been widely used as window electrodes in optoelectronic applications such as flat panel displays and solar cells. Continuous development of the high-performance optoelectronic devices stimulates the research on TCO films. So far, low resistivity 共ⱕ1.5⫻ 10−4 ⍀ cm兲 Sn-doped In2O3 共ITO兲 films with high transparency in the visible wavelengths have been manufactured for the display applications. On the other hand, for solar cell applications, TCO films with high transparency up to the near-infrared 共NIR兲 wavelengths are required in order to increase conversion efficiency of solar cells by improving spectral sensitivity in the visible to NIR wavelengths. The optical properties of TCO films in the NIR wavelengths are well understood in terms of the Drude theory,1 and high transparency in the NIR wavelengths can be achieved by reduction in free carrier absorption; i.e., increase in mobility and decrease in carrier density, provided that the same resistivity is required. Various cation-doped In2O3 have been investigated so a兲
Electronic mail:
[email protected].
0021-8979/2010/107共3兲/033514/11/$30.00
far. Among them, crystalline ITO films have been widely used as window electrodes, since addition of Sn to In2O3 effectively increases carrier density, and high electrical conductivity can be attained with high optical transparency in the visible wavelengths.2 On the other hand, Ti-doped, Zrdoped, Mo-doped, and W-doped In2O3 films show high mobility 共⬎80 cm2 / V s兲 at carrier density over 1020 cm−3.3–14 Application of such films to actual devices has also been reported including superstrate-type hydrogenated nanocrystalline Si,15 CuIn1-xGaxSe2,16,17 CdTe,18 and dye sensitized19 solar cells, although the configuration of the solar cells and materials underlying the TCO films limits the growth window of TCO since high growth temperature 共ⱖ300 ° C兲 required to obtain high mobility often leads to defect generation in solar cell materials. Meanwhile, it is possible to produce In2O3 : H films with higher mobility at lower fabrication temperature 共⬃170 ° C兲 by the use of solid-phase crystallization of amorphous In2O3 : H films, which are sputtered at room temperature with H2O addition in sputtering gas.20 The films show enormous potential for window electrodes for heat-sensitive devices and/or substrates. The feasibility of the films has already been demonstrated for hydrogenated amorphous Si/crystalline Si heterojunction solar
107, 033514-1
© 2010 American Institute of Physics
033514-2
J. Appl. Phys. 107, 033514 共2010兲
Koida et al.
cells21 and substrate-type thin-film microcrystalline Si solar cells.22 However, the mechanisms of high electron mobility in the solid-phase crystallized In2O3 : H films are not well understood at present, although H has been proposed to form a shallow donor state in metal oxides,23–26 and the prediction is supported by a number of experimental studies27–30 in ZnO. The film fabrication method itself is sometimes used for growth of ITO films. Amorphous ITO films with low resistivity can be obtained by a dc magnetron sputtering method at room temperature with H2O addition in sputtering gas.31 The influence of water vapor pressure and/or residual H2O during sputtering on the structure of the films has been widely investigated, since amorphous ITO films have a practical advantage compared with polycrystalline ITO films from the view point of patterning due to their superior wet etching properties.31 Also, crystallization process of the films during postannealing procedures and change in electrical properties caused by the crystallization have been investigated.32–39 In these ITO films, however, large increase in mobility observed in In2O3 : H has not been observed. On the other hand, Shigesato et al.40 and Nakazawa et al.41 have reported similar behaviors caused by crystallization of unintentionally doped amorphous In2O3 films that were deposited by high density plasma assisted electron beam evaporation using an arc plasma generator at 120 ° C and by dc/rf sputtering method at room temperature, respectively. The high mobility is also reported in polycrystalline In2O3 films grown directly on heated substrates by thermally evaporating pure In2O3.42–44 In this paper, we have investigated structural, electrical, and optical properties of amorphous and solid-phase crystallized In2O3 : H films in order to study the origin of high electron mobility in the crystallized In2O3 : H films. First, microscopic and macroscopic structure of the films using transmission electron microscopy 共TEM兲 and thermal desorption spectroscopy 共TDS兲 are described, since the structural properties are strongly influenced by water vapor pressure during the growth20,33,39 as well as postannealing procedures. Then, changes in Hall carrier density 共NHall兲 and Hall mobility 共Hall兲 during crystallization are described. Finally, we have investigated the effective mass of carriers in the In2O3 : H films using spectroscopic ellipsometry 共SE兲, and discussed the origin of high Hall in the solid-phase crystallized In2O3 : H films with comparison between Hall, optical mobility 共opt兲, and calculated mobility dominated by ionized impurities. II. EXPERIMENTAL A. Sample preparation and characterization
All the In2O3 : H films were prepared under an identical deposition condition except postannealing temperatures. Approximately 70-nm-thick In2O3 : H thin films were grown on 50-nm-thick SiO2-coated Si without any intentional substrate heating by rf magnetron sputtering method.13 To incorporate H into In2O3 matrix, H2O vapor has been introduced into the chamber during sputtering of an In2O3 ceramic target.20 In this sputtering system, the target was aiming at the substrate
with an angle of 60° to the substrate plane, and an average distance between the substrate and the target was 18 cm. The substrate was rotated for uniform film deposition. The base pressure of the chamber was less than 5 ⫻ 10−6 Pa. The total gas pressure of Ar, O2, and H2O during deposition was 0.5 Pa with a O2 / 共Ar+ O2兲 ratio of 0.0038 and H2O partial pressure of ⬃2 ⫻ 10−4 Pa. The rf power density was 4.4 W / cm2. These deposition conditions yield the relatively slow In2O3 : H deposition rate of ⬃6 nm/ min. After deposition, the films were postannealed in vacuum at 160, 200, 250, 300, 400, and 630 ° C for 2 h. Structural changes in the films were evaluated by planview TEM and electron diffraction. An average grain area in the films was calculated by counting a number of grains in a given area and dividing the total area by the number of grains. The average grain radius, assuming grains to be twodimensional square, was also calculated. Changes in surface morphology were evaluated by atomic force microscopy 共AFM兲. TDS was applied to investigate desorption gases during the postannealing process. Films on SiO2-coated Si were heated at a constant rate of ⬃16 ° C / min by an IR ramp. The temperature was calibrated using a thermocouple in direct contact with the film surface. Gases desorbed from the specimens were identified by a quadrupole mass spectrometry. Electrical resistivity 共兲, NHall, and Hall of the films were obtained from Hall measurements in the van der Pauw configurations with an ac modulation of a magnetic field. Temperature dependence of the electrical properties was also measured from 80 to 523 K 共250 ° C兲. The properties of the films were measured in nitrogen atmosphere from 80 to 400 K and in vacuum from 40 to 250 ° C. In the analyses, film thicknesses determined by SE were used. The SE spectra 共⌿ , ⌬兲 were measured from ultraviolet to IR wavelengths. We used two rotating-compensator ellipsometry systems in the wavelengths from 200 to 1700 nm and from 1.7 to 30 m. B. Data analysis for SE
The theoretical expressions in this section are based of a paper reported by Fujiwara and Kondo,45 and details of the data analysis procedures have been reported in the same reference. The analysis of the SE data allows us to obtain surface roughness, thickness, and optical constants of the In2O3 : H films. From the analyses of the optical constants, effective mass 共mⴱ兲, opt, and optical carrier density 共Nopt兲 of the films can be obtained. The SE analyses of the films were performed using an optical model consisting of air/surface roughness layer/TCO/SiO2/crystalline Si. To simplify the SE analysis, thickness and optical constants of SiO2 film were taken from the measurement of the SiO2/Si substrate prior to deposition. The dielectric function of the In2O3 : H layer was considered to be homogeneous in depth and modeled by combining the Drude model with the Tauc–Lorentz 共TL兲 model.46 Although there have been also studies treating inhomogeneity in depth,47 we characterized thin TCO films 共⬃70 nm兲 to suppress analysis errors arising from the thickness variation in the optical constants. The TL model has been confirmed to provide good fitting to experimental data
033514-3
J. Appl. Phys. 107, 033514 共2010兲
Koida et al.
in various TCO films.45,46 In the TL model, the dielectric function is obtained from five parameters 关ATL, C, ET, E0, 1共⬁兲兴 that represent the amplitude, broadening parameter, Tauc optical gap, peak transition energy, and energyindependent contribution to real part of the dielectric function, respectively. On the other hand, in the Drude model, the expression for the dielectric function is obtained from two parameters 共AD and ⌫D兲 that represent the oscillator amplitude and broadening parameter, respectively. The dielectric function of the Drude model is expressed by D共E兲 = −
A D⌫ D AD AD =− 2 2 , 2 −i 3 E − i⌫DE E E + ⌫D E + ⌫D 2
共1兲
where AD and ⌫D are expressed by AD =
ប2e2Nopt . m ⴱ 0
共2兲
⌫D =
ប បe = ⴱ . opt m opt
共3兲
Here ប, e, and 0 in Eq. 共2兲 are Dirack’s constant, electron charge, and free space permittivity, respectively. opt in Eq. 共3兲 is optical relaxation time. From the two parameters 共AD and ⌫D兲, Nopt and opt can be obtained when mⴱ is known. At sufficiently low energies, the real part of the dielectric function of the TL model shows the constant value of ⬁. At this condition, the real part of the dielectric function of the film can be written as 1共E兲 = ⬁ −
AD 2 . E + ⌫D
FIG. 1. 共Color online兲 SAED patterns of 共a兲 the as-deposited In2O3 : H film and postannealed films at 共b兲 160 ° C, 共c兲 250 ° C, and 共d兲 630 ° C.
amorphous and includes some numbers of crystalline grains in it. Many of the NBD patterns of grains gave a single crystal pattern in each grain. There are no well-defined boundaries in a grain and the order does not decrease at the edge of the crystal grains, as shown in Fig. 2共b兲. The volume fraction of crystalline phase and the number density of the grains estimated from low magnified TEM images 共not
共4兲
2
2 兲, ⬁ can be deterThus, by plotting 1 versus 1 / 共E2 + ⌫D mined from an intercept. The plasma energy 共E p兲 is defined as the maximum energy at which free electrons can follow a disturbing electromagnetic field. 1 is almost zero at this en2 can be neglected in comparison with E2 ergy. In general, ⌫D in TCO, therefore E p can be written as
Ep =
冉 冊 冉 AD ⬁
共1/2兲
=
ប2e2Nopt m ⴱ ⬁ 0
冊
共1/2兲
.
共5兲
E p can be obtained experimentally and ⬁ can be determined from Eq. 共4兲, therefore, mⴱ can be deduced if Nopt is known. Provided that optically active carriers 共Nopt兲 is the same as that of the electrically active carriers 共NHall兲, we can obtain mⴱ using NHall instead of Nopt. III. RESULTS A. Structural properties
Figures 1 and 2 show selected area electron diffraction 共SAED兲 patterns and plan-view TEM images of the asdeposited and postannealed films at 160, 250, and 630 ° C. In Fig. 1共a兲, the as-deposited film shows a broad and diffuse diffracted ring with small numbers of diffracted spots. Figure 2共a兲 shows the plan-view image of the film. A nanobeam electron diffraction 共NBD兲 pattern of a grain depicted by an arrow and a SAED pattern of a dotted region are also shown. The results clearly show that the as-deposited film is mostly
FIG. 2. 共Color online兲 Plan-view TEM images and HRTEM images of 关共a兲 and 共b兲兴 the as-deposited In2O3 : H film and postannealed films at 关共c兲 and 共d兲兴 160 ° C, 关共e兲 and 共f兲兴 250 ° C, and 关共g兲 and 共h兲兴 630 ° C. In 共a兲, a NBD pattern of a grain depicted by an arrow and a SAED pattern of a dotted region are also shown.
033514-4
Koida et al.
shown兲 were 1.5% and 3.5 m−2, respectively. The average lateral grain size of each grain is about 66 nm, and the value is comparable to the film thickness 共⬃74 nm兲. Therefore, the grain might grow from a nuclei having smaller size and the size increased by a progressive advance of the amorphous/crystalline interface during the deposition; initially crystallize three dimensionally until a grain intersects the surface or substrate, and grow laterally. It should be mentioned that no new crystalline nuclei appeared during the TEM or NBD observations at the amorphous region, although the grain growth proceeded from an edge of a grain with keeping the same crystal orientation during the observation of high resolution 共HR兲 TEM images around amorphous/crystalline interfaces. With increasing annealing temperature 共Ta兲, grain growth proceeded. As shown in Fig. 2共c兲, the film postannealed at 160 ° C shows a structure embedded larger crystalline grains 共⬃370 nm兲 in the amorphous matrix. The volume fraction of crystalline phase and the number density of the grains were 48% and 3.6 m−2, respectively. The number density is almost the same as that of the as-deposited film 共3.5 m−2兲. This clearly indicates that the grains grew from the smaller grains observed in the as-deposited film and no new crystal nuclei were generated at Ta lower than 160 ° C. In a grain, thickness-interference fringes can be seen due to the near-perfect planar growth. A HRTEM image at the boundary of three grains shown in Fig. 2共d兲 reveals that no amorphous phase exists at the boundary. With increasing Ta up to 250 ° C, the amorphous phase disappeared. In Fig. 1共c兲, the broad and diffuse diffracted ring observed in Figs. 1共a兲 and 1共b兲 completely disappeared, and very sharp spotty ring pattern appears. The crystalline ring pattern was indexed as bixbyite In2O3. As confirmed previously,48 the lattice constants of the solid-phase crystallized In2O3 : H films were almost the same as that of In2O3 powder, indicating that the films are free from strains. As shown in Fig. 2共e兲, the film shows a grain structure and the averaged grain size is about 440 nm. The boundary between the grains can be seen clearly, therefore, the grain boundaries are almost perpendicular to the surface. In a grain, no well-defined boundary or substructures are observed as shown in Fig. 2共f兲. The number density of the grains 共5.1 m−2兲 is larger than that in the as-deposited film and postannealed film at 160 ° C, indicating that new crystalline nuclei were generated from amorphous regions for the duration of heating up to 250 ° C. Further increase in Ta deteriorates the structures. Figures 1共d兲 and 2共g兲 show a SAED pattern and plan-view TEM image of the film postannealed at 630 ° C, respectively. The film also shows a grain structure and consists of large block grains whose averaged size is almost the same as that of the film postannealed at 250 ° C. However, at the corner of the grain boundaries, there are voids as clearly seen in Fig. 2共h兲 that show a corner of grain boundaries. Also, the film thickness 共⬃57 nm兲 determined by SE was much smaller than that of the other films 共73–75 nm兲 postannealed at Ta lower than 400 ° C. Furthermore, substructures are clearly seen in a grain as shown in Fig. 2共g兲. These phenomena are strongly correlated with desorption of gasses from the film during the postannealing procedure, as described below.
J. Appl. Phys. 107, 033514 共2010兲
FIG. 3. 共Color online兲 AFM images of 共a兲 the as-deposited In2O3 : H film and postannealed films at 共b兲 160 ° C, 共c兲 250 ° C, 共d兲 400 ° C, and 共e兲 630 ° C.
The structural change caused by the postannealing procedure is also observed in surface morphology. Figures 3共a兲–3共e兲 shows AFM images of a 2 ⫻ 2 m region of the as-deposited and postannealed films at 160, 250, 400, and 630 ° C, respectively. The surfaces of the films are very smooth except the film postannealed at 630 ° C, and the values of root-mean-square roughness are 0.14, 0.20, 0.26, 0.25, and 1.8 nm for the as-deposited and postannealed films at 160, 250, 400, and 630 ° C, respectively. From AFM images, no clear difference in surface morphology between the amorphous and crystalline regions is observed in the as-deposited film and postannealed film at 160 ° C, although clear differences are observed in plan-view TEM images as shown in Figs. 2共a兲 and 2共c兲. For films postannealed at 250 ° C, grain boundaries observed in the TEM image 关Fig. 2共e兲兴 are also clearly seen in the AFM image 关Fig. 3共c兲兴. The roughness inside a grain is less than a unit cell 共⬃1 nm兲. With increasing Ta up to 400 ° C, no large difference in surface morphology is observed. However, the surface roughness increases by further annealing up to 630 ° C, and substructures are clearly observed at surface of a grain. During the postannealing process, several gases are desorbed from the films. Figure 4 shows TDS spectra of desorption species from the films prepared at different Ta. We investigated H2 共m / z = 2, where m / z indicates the molecular mass to charge ratio兲, H2O 共m / z = 18兲, O2 共m / z = 32兲, and In 共m / z = 115兲. The intensity of the species reveals that significant amount of H2O-related species are incorporated in the
033514-5
J. Appl. Phys. 107, 033514 共2010兲
Koida et al.
Log Intensity (arb. units)
H2 (m/z=2)
sub. as-deposited Ta=160OC Ta=250OC Ta=400OC Ta=630OC
O2 (m/z=32)
H2O (m/z=18) In (m/z=115)
x 10
0
200
x 10
200 400 400 600 0 Substrate temperature (˚C)
600
FIG. 4. 共Color online兲 TDS spectra of the as-deposited In2O3 : H film and postannealed films at 160, 250, 400, and 630 ° C for H2 共m / z = 2兲, H2O 共18兲, O2 共32兲, and In 共115兲.
films except the film postannealed at 630 ° C. H compositions in the films were estimated from the TDS signals of H2O desorbing from the films during the heat treatment up to 750 ° C, and are summarized in Table I. The atomic concentration of H was obtained by normalizing the TDS signals using H compositions for 240-nm-thick films characterized by hydrogen forward scattering spectrometry and TDS.20 The configurations of H in the as-deposited film are considered to be bonded H with the form of OH as well as H2O. Since H2O molecules are easily decomposed into H and OH during sputtering,32 H and OH have a chance to form In–OH bonds at the growing surface with an oxygen dangling bond and an In dangling bond, respectively. As reported previously,20 introduction of H2O prevents the crystal growth. This is probably because the formation of In–OH bonds prevents construction of In–O–In bond networks. Also, similar H compositions between the as-deposited film and postannealed film at 300 ° C strongly support the existence of chemically bonded H in the films.
In Fig. 4, the films show a complex variation in the TDS signals, which can be categorized into four temperature regions; 共I兲 T ⱕ 170 ° C, 共II兲 170 ° C ⱕ T ⱕ 380 ° C, 共III兲 380 ° C ⱕ T ⱕ 600 ° C, and 共IV兲 600 ° C ⱕ T ⱕ 700 ° C. These desorption states are considered to be closely related to microstructures and chemical structures of the In2O3 : H films. In the region 共I兲, narrow peaks at ⬃70, ⬃135 ° C are observed in the H2O signal for the as-deposited film. However, the two peaks gradually decrease with increase in Ta and disappear for the films annealed at Ta over 200 ° C that show fully crystalline phase. Further, the second peak of H2O desorption occurs with crystallization of the film since In desorption occurs simultaneously at the same temperature. Therefore, the results indicate that the first and second peaks are attributed to physically adsorbed and hydrogen bonded water molecules that exist inside the films, for example, at surface of micropores or macropores in amorphous phase. In the region 共II兲, H2O and H2 signals increase at Ta over 170 ° C and show a broad plateau at 220– 380 ° C for all the films except the films postannealed over 400 ° C. The broad desorption characteristics suggest a necessary prediffusion of internal H2O related species to the grain boundary and/or surface before desorption; during heating, H or OH inside grains might segregate at grain surfaces 共grain boundaries, film surface, and interface between the film and SiO2兲, and react with a neighboring H at the surfaces, and finally generate H2 and H2O molecules. Indeed, segregation of H at an ITO/SiO2-coated glass interface has already been observed in a x-ray photoemission spectroscopy study of crystallized ITO films grown under water vapor on the heated glass substrates.49 In the region 共III兲, a broad and large peak at about 530 ° C is observed in H2O, In, and H2 signals for films except the film annealed at 630 ° C. From AFM and TEM observations, the film postannealed at 630 ° C has substructures in the grains and voids at the grain boundaries that are not observed in the film postannealed at Ta lower than 400 ° C. Therefore, the rapid increase in the TDS signals implies a fast decomposition of indium hydroxide formed on the grain surfaces, and thus, a fast release of H2O, In, and H2 takes place. The decrease in film thickness for the film also supports the validity of the idea. In the region 共IV兲, oxygen desorbs just after finishing desorption of H2, H2O, and In. Due to the voids at the grain boundaries and the structural
TABLE I. H composition and best-fit parameters extracted from the dielectric function modeling using the Drude model and Tauc–Lorentz model for the In2O3 : H films postannealed at different annealing temperature 共Ta兲. Thickness Ta 共°C兲 As-deposited 160 200 250 300 400 630
TL model
Drude model
H composition 共at. %兲
ds 共nm兲
db 共nm兲
ATL 共eV兲
ET 共eV兲
E0 共eV兲
AD 共eV兲
⌫D 共eV兲
3.23 3.28 2.73 2.83 2.74 2.30 0
1.73 3.90 1.59 2.80 2.53 2.22 9.78
73.61 72.11 74.14 72.49 73.16 72.03 52.43
171.4 199.9 254.7 172.8 174.4 191.8 122.3
2.86 3.00 3.28 3.07 3.06 3.03 2.83
5.35 4.69 4.31 6.00 6.00 5.17 ¯
1.591 1.189 0.759 0.703 0.679 0.169 ¯
0.0777 0.0641 0.0331 0.0327 0.0340 0.0504 ¯
J. Appl. Phys. 107, 033514 共2010兲
10
150
10
20 19
100 50 0 0
200 400 600 Annealing temperature (˚C)
FIG. 5. Resistivity, carrier density, and mobility of In2O3 : H films as a function of postannealing temperature.
change inside grains, oxygen might easily desorbs from the films. Finally, the increase in H2, H2O, and O2 signals above 700 ° C is considered to be outgassing from the sample stage due to readsorption of previously desorbed gasses. B. Electrical properties
10
-3
(1) 80K ~ 400K (2) 40˚C ~ 250˚C ~ 40˚C (3) 80K ~ 400K
-4
Carrier density (cm )
10
-3
2
Mobility (cm /Vs)
Resistivity (Ωcm)
Figure 5 shows , NHall, and Hall of the films measured at room temperature as a function of Ta. The electrical properties also show a complex variation. With increasing Ta up to 250 ° C, the film shows a large increase in Hall with a decrease in NHall. The change in Hall and NHall is strongly correlated with the structural change from amorphous to polycrystalline phases. Indeed, the change is clearly observed during the crystallization process. Figure 6 shows temperature dependence of , NHall, and Hall of the asdeposited film. First, the as-deposited film was cooled down to 80 K in nitrogen atmosphere and the electrical properties were measured during heating up to 400 K. Second, the properties of the film were measured during heating from 40 to 250 ° C and also during cooling from 250 to 40 ° C in 10
10
-4
10
10
2
21
20
10 10
2
2
Mobility (cm /Vs)
10
21
-3
-3
-4
10
Carrier density (cm )
10
Mobility (cm /Vs)
as-deposited
-3
-3
10
-2
Carrier density (cm )
10
Resistivity (Ωcm)
Koida et al. Resistivity (Ωcm)
033514-6
10
10
10
2
as-deposited Ta=160˚C Ta=200˚C
1 6
7
8 9
21
20
19
Ta=250˚C Ta=400˚C
2
100 Temperature (K)
3
FIG. 7. 共Color online兲 Resistivity, carrier density, and mobility of the asdeposited In2O3 : H film and postannealed films at 160, 200, 250, and 400 ° C as a function of temperature.
vacuum. During the heating, solid-phase crystallization occurred. Finally, the crystallized film was cooled down to 80 K and the properties were measured during heating up to 400 K. In Fig. 6, the as-deposited film shows the constant NHall at temperatures less than 160 ° C, indicating that the film is a degenerate semiconductor. However, a large decrease in NHall and an increase in Hall are clearly observed at Ta of 160 ° C. TEM analyses showed that the crystallization proceeded at the temperature. Therefore, the change originates from the increase in the volume fraction of crystalline phase in the films. Upon full crystallization, the film also shows the constant NHall at temperatures over the whole range and increase in Hall with decrease in temperature. It should be noted that the temperature dependence of Hall changes after full crystallization. In degenerated semiconductors, mobility is determined mainly by impurity scattering and phonon scattering. The result indicates that the phonon scattering is predominant at higher temperatures in the crystallized film, and the effect of impurity scattering is weakened by the crystallization procedure, since the mobility dominated by ionized impurity scattering in degenerated semiconductors is independent of temperature whereas the mobility dominated by phonon scattering is inversely proportional to temperature.50 This tendency is also observed for the postannealed films at various temperatures as shown in Fig. 7; with increasing Ta up to 250 ° C, the temperature variation in Hall increases with increase in Hall, although it decreases when Ta increases over 400 ° C.
1/T
10
1 6 7 8 9
2
3
4
5
6
100 Temperature (K)
FIG. 6. Resistivity, carrier density, and mobility of the as-deposited In2O3 : H film as a function of temperature. First, the as-deposited film was cooled down to 80 K in nitrogen atmosphere and the electrical properties were measured during heating up to 400 K. Second, the properties were measured during heating from 40 to 250 ° C and during cooling from 250 to 40 ° C in vacuum. During the heating, solid-phase crystallization occurred. Finally, the crystallized film was cooled down to 80 K and the properties were measured during heating up to 400 K.
C. Optical properties
Figures 8共a兲–8共d兲 show the SE spectra for the asdeposited film and postannealed films at 160, 200, and 300 ° C, respectively. The SE spectra were measured at incident angles of 65° and 70°. For clarity, the spectra measured at 70° are shown. The spectral features observed at E ⱕ 1 eV arise from free carrier absorption whereas the spectral features at E ⱖ 3 eV arise from the absorption associated with excitations across the In2O3 band gap. Dotted lines at
J. Appl. Phys. 107, 033514 共2010兲
30 90
Ψ (deg)
(c)
(d)
60 30 0 0.1 1 Photon energy (eV)
10 10 10 10
3 2
(a)
(b)
1
10 10
10
0.1 1 Photon energy(eV)
FIG. 8. 共Color online兲 SE spectra measured for 共a兲 the as-deposited In2O3 : H film and postannealed films at 共b兲 160 ° C, 共c兲 200 ° C, and 共d兲 300 ° C. Dotted lines show fitting results.
1 ⱕ E ⱕ 3.5 eV and 0.055ⱕ E ⱕ 0.28 eV show the calculation results obtained from linear regression analyses using the TL model and the Drude model, respectively. First, we performed the fitting in the higher energy region 共1 ⱕ E ⱕ 3.5 eV兲. We used this energy range to avoid complicated structures observed in dielectric functions at higher energies 共E ⬎ 3.5 eV兲 and the effects of free carrier absorption 共E ⬍ 1 eV兲. In the analysis, fixed values of 1共⬁兲 = 1, and C = 12 were used. Thus, we performed the fitting using ds, db, ATL, ET, and E0 as free parameters, where ds and db represent the thickness of surface roughness and TCO bulk layers, respectively. As shown in Fig. 8, the model provides good fitting to the experimental spectra. The analysis parameters obtained from the fitting are summarized in Table I. Second, we performed the fitting in the lower energy region using the Drude model. We used this energy range to avoid the effect of phonon absorptions caused by In2O3 共E ⱕ 0.055 eV兲 and anomalous absorptions 共E ⱖ 0.28 eV兲 described below. In the analysis, film thicknesses extracted from the above fitting were used. As shown in Fig. 8, the model provides good fitting to the experimental spectra, and the analysis parameters obtained from the fitting are summarized in Table I. When thicknesses of each layer in the optical model are known, the dielectric function can be extracted directly from the measured SE spectra by mathematical inversion using Fresnel equations. Figure 9 shows the dielectric functions of the films obtained from the mathematical inversion in a logarithmic scale versus photon energy. Here, 1 and 2 represent real part and imaginary part of the dielectric functions. Dotted lines in Fig. 9 show the fitting curves at 0.055ⱕ E ⱕ 3.5 eV obtained by the dielectric functions using the TL and the Drude models fitted at 1 ⱕ E ⱕ 3.5 eV and 0.055 ⱕ E ⱕ 0.28 eV, respectively. In Fig. 9共a兲, the as-deposited film shows large absorption at E ⱖ 3 eV. The absorption is associated with excitations across the In2O3 semiconductor band gap. At 1.5ⱕ E ⱕ 3 eV, 2 is very low. The film is transparent in this region. At E ⱕ 1.5 eV, 2 rises monotonically, where 1 goes down and crosses zero at an energy E p. Thus, 1 ⬍ 0 at E ⬍ E p. At E Ⰶ E p, 1 becomes strongly negative and the film reflects light. The E p value is shown in Table II. The fitting curves are in good agreement with the mathematically inverted spectra even at unfitted photon en-
10 10 10
3 2
3 2 1 0 -1
10
(c)
(d)
1
10 10
0
10
-1
10
10
-2
10
10 10
0
-1
10
10
0.1 1 Photon energy (eV)
0.1 1 Photon energy (eV)
|ε2|
60
180 120 60 0 -60 -120 180 120 60 0 -60 -120 -180
|ε1|
(b)
|ε1|
Ψ (deg)
(a)
∆ (deg)
90
10
3 2 1 0
|ε2|
Koida et al.
∆ (deg)
033514-7
-1 -2
FIG. 9. 共Color online兲 Dielectric functions of 共a兲 as-deposited In2O3 : H film and postannealed films at 共b兲 160 ° C, 共c兲 200 ° C, and 共d兲 300 ° C. Dotted lines show fitting results.
ergy region of 0.28⬍ E ⬍ 1 eV and hence the fitting models and parameters summarized in Tables I are consistent. For postannealed films at 200 and 300 ° C, the fitting curves are also in relatively good agreement with the mathematically inverted spectra as shown in Figs. 9共c兲 and 9共d兲. The E p of the films moves to lower energy due to the smaller carrier density as compared with that of the as-deposited film. Also, phonon absorptions caused by In2O3 lattices are observed at E ⱕ 0.055 eV, which cannot be observed in the as-deposited film. This can be explained by lower free carrier absorption. Furthermore, in the mathematically inverted spectra, there is a broad 2 peak at E ⬃ 0.4 eV 共3100 cm−1兲 that is not taken into account in the optical model of the films, and the intensity decreases with Ta. We attributed the peak from OH stretching modes, since the films contain chemically bonded H 共OH兲 and the total amount of H decreases with increasing Ta, as described above. Indeed, a broad absorption caused by OH stretching mode is observed in the regions from 3700 cm−1 共0.335 eV兲 to 2800 cm−1 共0.443 eV兲 for In2O3 films having hydroxylated 共InOH兲 In2O3 surface.51 To confirm this, however, more detailed study on H configuration in In2O3 local structure is necessary. Also, there is a small peak at photon energy of about 1.2 eV in 2 spectra for all the films. This peak is probably due to an artifact, since the peak intensity changes when the film thickness changes from 71 to 75 nm. For the film postannealed at 160 ° C, anomalous behaviors in 1 and 2 spectra are observed at photon energy from 0.3 to 1.0 eV, and the fitting curves does not agree well with the mathematically inverted data at this energy region, as shown in Fig. 9共b兲. The result indicates that the model for the TCO layer that assumes a homogeneous layer is not suitable. The result for the film is discussed later. According to the Eq. 共4兲, ⬁ can be obtained from the 2 兲 = 0. Figure 10 shows 1 of the films intercept at 1 / 共E2 + ⌫D 2 兲. Here, we used ⌫D explotted as a function of 1 / 共E2 + ⌫D tracted from the SE analysis shown in Table I. Solid lines fit to the experimental results. The increase in 1 at 1 / 共E2 2 + ⌫D 兲 ⬃ 0.1 is caused by the effect of band gap and is not accounted for the Drude theory. In the lower energy region, 2 兲. The extracted 1 exhibits a linear dependence on 1 / 共E2 + ⌫D ⬁ values are listed in Table II. As shown in Table II, ⬁ is a constant value of ⬃4.2 irrespective of the carrier concentra-
033514-8
J. Appl. Phys. 107, 033514 共2010兲
Koida et al.
TABLE II. Plasma energy 共E p兲, high frequency dielectric constant 共⬁兲, effective mass 共mⴱ / m0兲, optical mobility 共opt兲, optical carrier density 共Nopt兲, and relaxation time 共opt兲 extracted from the SE analyses for the In2O3 : H films postannealed at different annealing temperature 共Ta兲. Hall mobility 共Hall兲, Hall carrier density 共NHall兲, and relaxation time 共Hall兲 extracted from Hall measurements are also listed. SE
Hall
Ta 共°C兲
Ep 共eV兲
⬁
m ⴱ / m0
opt 共cm2 / V s兲
Nopt 共cm−3兲
opt 共s兲
Hall 共cm2 / V s兲
NHall 共cm−3兲
Hall 共s兲
As-deposited 160 200 250 300 400
0.635 0.625 0.453 0.436 0.420 0.197
4.26 ¯ 4.23 4.18 4.19 4.20
0.373 ¯ 0.336 0.301 0.309 0.288
40.0 ¯ 104 118 110 59.4
4.30⫻ 1020 ¯ 1.85⫻ 1020 1.53⫻ 1020 1.52⫻ 1020 3.71⫻ 1019
8.47⫻ 10−15 ¯ 1.99⫻ 10−14 2.01⫻ 10−14 1.94⫻ 10−14 9.74⫻ 10−15
41.5 64.3 108 122 112 81.5
4.64⫻ 1020 3.11⫻ 1020 2.12⫻ 1020 1.74⫻ 1020 1.66⫻ 1020 3.42⫻ 1019
8.79⫻ 10−15 ¯ 2.05⫻ 10−14 2.08⫻ 10−14 1.97⫻ 10−14 1.34⫻ 10−14
tion in a range from 3.4⫻ 1019 to 4.6⫻ 1020 cm−3. By assuming Nopt = NHall, we determined effective mass 共mⴱ / m0兲 of the films using the ⬁ and E p. Here, m0 is free electron mass. Figure 11 shows mⴱ / m0 plotted as a function of NHall. Reported values of amorphous In2O3 and ITO films,52 and polycrystalline ITO films45,53,54 are also plotted in the figure. mⴱ / m0 of In2O3 : H films gradually increases with NHall, and this tendency is also observed in the reported films. The increase in mⴱ with carrier density in the heavily doped TCO films has been explained by the influence of the degeneracy and the nonparabolicity of the conduction band.45,53–62 In Fig. 11, mⴱ / m0 values of the as-deposited In2O3 : H film are almost equivalent to those of other reported polycrystalline ITO films.45,53,54 Therefore, the difference in mⴱ / m0 between the as-deposited film and the postannealed films can be explained mainly by the difference in carrier density rather than crystal structures, although there is a possibility that mⴱ / m0 is slightly different between amorphous and crystalline In2O3 : H films. The weak effect of the structural disorder on electrical properties has also been reported by Bellingham et al.52 opt and Nopt can be determined by using the ⌫D, mⴱ, and AD by applying Eqs. 共3兲 and 共2兲, respectively. Table II summarizes the opt and Nopt values. Hall and NHall are also listed in the same Table. As shown in Table II, opt and Nopt are in good agreement with Hall and NHall, respectively. The result indicates that the mechanism limiting the electron mobility is attributed to 共i兲 scattering inside the amorphous ma-
4 ε1
Ta=400˚C
2 Ta=250˚C
0 as-deposited
-2 0
1
2 3 2 1/(E +ΓD ) 2
IV. DISCUSSION
The difference in mⴱ for amorphous and solid-phase crystallized In2O3 : H films can be explained mainly by their carrier density as shown in Fig. 11. Thus, the very high electron mobility for the crystallized In2O3 : H films 共Ta = 200– 300 ° C兲 originates mainly from longer relaxation time rather than smaller mⴱ as compared with the asdeposited film, since mobility is functions of and mⴱ as shown in Eq. 共3兲. opt and Hall values for the films are summarized in Table II. As shown in Table II, opt and Hall for the crystallized films are about 2 ⫻ 10−14 s, much longer than those for the as-deposited film that shows a large volume fraction of amorphous phase. However, the longer cannot be simply explained by the decrease in carrier density. In general, Hall of TCO films depends on mechanisms by which the carriers are scattered by phonon scattering, impurity scattering, grain boundary scattering, and so on. Among them, grain boundary scattering seems not to limit the mobility, since Hall show similar values with opt. Also, the temperature variation in Hall shown in Figs. 6 and 7 indicates that the crystallized films have weaker effects of ionized impurity scattering on the transport than the asdeposited film. In order to discuss it, we calculated mobility
Effective mass (m*/m0 )
6
trix for the as-deposited film, and 共ii兲 in-grain scattering rather than grain boundary scattering for the postannealed films.
4
FIG. 10. 共Color online兲 Real part of dielectric function for as-deposited In2O3 : H film and postannealed films at 250 and 400 ° C, plotted as a function of 1 / 共E2 + ⌫D2 兲. The ⌫D is a broadening parameter expressed by Eq. 共4兲. The high frequency dielectric constant ⬁ can be obtained from the intercept at 1 / 共E2 + ⌫D2 兲 = 0.
0.5 as-deposited
0.4 0.3
This study Clanget et al. (poly-ITO) Ohhata et al. (poly-ITO) (a-IO) Bellingham et al. (a-ITO) Fujiwara et al. (poly-ITO)
0.2 0.1 0
2
4
6
NHall ( 10
8 20
10
12
-3
cm )
FIG. 11. Effective mass 共mⴱ / m0兲 of In2O3 : H films plotted as a function of Hall carrier density. In the figure, experimental results reported previously of amorphous In2O3 and ITO films 共after Ref. 52兲, and polycrystalline ITO films 共after Refs. 45, 53, and 54兲 are also shown.
033514-9
J. Appl. Phys. 107, 033514 共2010兲
Koida et al.
2
Mobility (cm /Vs)
(a) Ta=250ºC Ta=200ºC
100 Ta=400ºC
Ta=300ºC Ta=160ºC as-deposited
µHall µopt
µii (z=1) µii (z=2)
10 20
10 -3 Carrier density (cm )
10
(b)
2
Mobility (cm /Vs)
Ta=200ºC
100
21
Ta=160ºC
Ta=160ºC Ta=150ºC
This study Ref. 21
10 20 10
2
as-deposited
3
4
5
-3
Carrier density (cm ) FIG. 12. 共Color online兲 共a兲 Hall mobility of In2O3 : H films as a function of Hall carrier density. Optical mobility of the films as a function of optical carrier density is also shown. The upper and lower bars are calculated mobility at the Hall carrier density assuming that the conducting carriers originate entirely from singly charged and doubly charged ions, respectively, using the effective mass obtained experimentally. 共b兲 Hall mobility of the as-deposited In2O3 : H film and postannealed films at 160 and 200 ° C as a function of Hall carrier density. Experimental results reported previously 共after Ref. 21兲 of an In2O3 : H film postannealed at 100, 120, 140, 150, 160, 170, 180, and 200 ° C are also shown.
共ii兲 dominated by ionized impurity scattering. In Fig. 12共a兲, data of Hall, opt, and the calculated mobility for the In2O3 : H films are shown as a function of carrier density. The upper and lower bars are the mobility at the carrier density assuming that the conducting carriers originate entirely from singly charged and doubly charged impurities, respectively. The details of the expressions for ii are written elsewhere.56 Here, the value of static dielectric constant was taken as 8.9 for In2O3,2 and the values of effective mass derived by SE analyses listed in Table II were used. In Fig. 12共a兲, the opt and Hall for the as-deposited film are in good agreement with the calculated doubly charged ii. In general, stoichiometry of oxygen ions by controlling the oxygen vapor pressure during growth strongly influence on the electrical properties of TCO films, since oxygen deficiencies are easily formed and generate free carriers. Therefore, the carriers of the amorphous films are influenced by stoichiometry of oxygen ions, and the mobility seems to be dominated by doubly charged ions. It should be mentioned that the as-deposited In2O3 : H film seems to have a nearly stoichiometric composition from the results of resistivity change during the crystallization as shown in Fig. 6; large change in resistivity at around In melting point 共157 ° C兲 that is probably due to a solid-liquid transformation of slightly excess In over the stoichiometric composition63 is not observed for the In2O3 : H film.
On the other hand, for the crystallized In2O3 : H films postannealed at 200– 300 ° C, the opt and Hall are not described by the doubly charged ii that described the asdeposited film well, and rather in good agreement with the calculated singly charged ii as shown in Fig. 12共a兲. During the crystallization, H2O molecules inside amorphous phase are desorbed, and H and/or OH inside grains migrate, as indicated by TDS measurements. At the same time, oxidation also occurs in the In2O3 : H films due to the existence of H2O related molecules that work as oxidizing agents, as indicated by previous Rutherford backscattering spectrometry measurements.20 Based on the results, we propose that structural rearrangements during crystallization eliminate oxygen deficiency and generate H+ that acts as a singly charged donor. Fixed coordination around In and O atoms by crystallization may change the doping mechanism in In2O3 : H. Accordingly, opt and Hall for the crystallized films are completely different with those for the as-deposited film. Indeed, NHall for the fully crystallized film 共Ta = 200 ° C兲 is almost a half of NHall for the as-deposited film, as shown in Fig. 12共a兲. Also, NHall for the film postannealed at 160 ° C that shows the crystalline volume fraction of 48% is almost a half of the sum of NHall for the as-deposited film and fully crystallized film at 200 ° C. The observed change in NHall and Hall can be explained if we assume that one doubly charged impurity that generates two free electrons in the amorphous phase is removed during crystallization with generation of one singly charged impurity that also generates one free carrier. It should be mentioned that such large increase in Hall is realized in pure In2O3 by 共i兲 solid-phase crystallization process and 共ii兲 H-doping. Similar large crystalline growth has also been reported during postannealing procedures of amorphous ITO films,64–68 however, such large increase in Hall has not been observed.42,64,68 In case of ITO, Sn that is not electrically activated in amorphous phase40,69,70 is activated by crystallization, however, the Hall and NHall values are far from the singly charged ii,40,64,68,69 indicating that Sn also may provide extra scattering centers. Based on the results, H seems to be crucial to reduce carrier scattering in the solid-phase crystallized In2O3. In these arguments, however, we consider only the effects of ionized impurity scattering although there is a contribution from phonon scattering at room temperature in the crystallized films as indicated by the temperature variation in Hall. Therefore, more detail study is necessary to carry out quantitative discussions on the carrier generation and scattering in the crystallized In2O3 : H. The change in transport properties of the films postannealed at Ta lower than 200 ° C can be explained by the coexistence of amorphous and crystalline phases, as mentioned above. Figure 12共b兲 shows the change in NHall and Hall of the In2O3 : H films. Reported properties of a In2O3 : H film grown under similar conditions except an oxidation condition of a In2O3 ceramic target surface are also shown. The data are obtained from single In2O3 : H film on a glass postannealed several times from 100 to 200 ° C.21 In both series of the films, increase in Hall and decrease in NHall are observed at Ta over 150– 160 ° C at which crystal growth proceeded, indicating that electrical properties are determined by the volume fractions of amorphous and crystalline phases.
033514-10
J. Appl. Phys. 107, 033514 共2010兲
Koida et al.
The coexistence of two phases having different carrier density in the film is also indicated in the SE spectra. In Fig. 8共b兲, the film postannealed at 160 ° C shows two distinct changes at the unfitted photon energy regions 共0.28ⱕ E ⱕ 1 eV兲. The changes are also clearly seen in the mathematically inverted dielectric constants as shown in Fig. 9共b兲; the 2 spectrum has a kink at ⬃0.41 eV, which approximately corresponds to the E p value for the film postannealed at 200 ° C. From TEM images, the film shows planar crystal grains in the amorphous matrix. Hence, we used a model composed of two patterned amorphous and crystalline phases. Here, the dielectric functions of the as-deposited film and postannealed film at 200 ° C were used for the amorphous and crystalline phases, respectively. Using the model, we could demonstrate the two distinct changes in ⌿ and ⌬ in the similar photon energy region, although there are some discrepancies between the fitted curves and experimentally obtained curves. The results 共not shown兲 indicate that the model is, in general, suitable, however, more detail study is necessary to confirm the optical properties of the films having the mixed structures. Finally, further annealing of the crystallized films over 400 ° C deteriorates the electrical properties; namely, large decrease in Hall and NHall. The temperature approximately corresponds to the temperature at the end of the region 共II兲 in TDS signals. The behaviors of H2O and H2 desorption in the region 共II兲 for the films are completely different with those for the films annealed at Ta lower than 250 ° C, indicating the change in microscopic and chemical structures related to H. Furthermore, the H content in the films drastically decreases with Ta over 400 ° C, as shown in Table I. We also found that Hall shows similar values with opt and substructures are generated inside grains for the films annealed at 630 ° C. These results suggest that H2O and H2 desorption generates additional microscopic defects inside grains and decreases free carriers that are generated by H-doping, leading to shorten the carrier relaxation time. V. CONCLUSIONS
Structural, electrical, and optical properties of amorphous and crystalline In2O3 : H TCO films have been studied by applying TEM, TDS, SE, and Hall measurements. The films have been fabricated at room temperature by the sputtering of a In2O3 ceramic target with introduction of H2O vapor, followed by postannealing to crystallize the amorphous phase. The mobility and carrier density depend on Ta, and categorized into two regions; before and after crystallization. At Ta ⱕ 200 ° C, amorphous to crystalline transformation of the films occurs. Simultaneously, Hall increases from 42 to 110 cm2 / V s and NHall decreases from 4.6⫻ 1020 to 2.1⫻ 1020 cm−3. The change in Hall and NHall with Ta is correlated with the volume fractions of the amorphous and crystalline phases in the films. The effective mass of the as-deposited and solid-phase crystallized films are 0.37m0 and 0.34m0, respectively, determined by analyses of dielectric functions of the films using the Drude model. The difference in the effective mass with carrier density can be explained by the influence of the degeneracy and the
nonparabolicity of the conduction band. Since the difference in the effective mass between the films is not so large, the very high electron mobility in the crystallized films is attributed mainly to longer relaxation time rather than smaller effective mass, as compared with the as-deposited films. Temperature-dependent Hall analysis, relationship between NHall and Hall, and comparison between Hall, opt, and calculated mobility dominated by ionized impurity scattering, showed that 共i兲 scattering processes inside amorphous and crystalline matrices limit the mobility, 共ii兲 doubly charged ionized impurity scattering is reduced by crystallization, and 共iii兲 phonon scattering becomes to be dominant after crystallization in the In2O3 : H films. The above results suggest that 共i兲 H-doping reduces carrier scattering in the crystallized In2O3, and 共ii兲 structural rearrangements during crystallization eliminate oxygen deficiency and generate H+ that acts as a singly charged donor. On the other hand, for the fully crystallized films, NHall gradually decreases with increasing Ta up to 300 ° C, whereas both NHall and Hall decrease drastically for the films annealed at Ta over 400 ° C. Desorption of H2O, H2, and In gasses from the films is confirmed to occur at high Ta. We also found substructures inside grains and similar values between Hall and opt for the films. These results suggest that desorption of the gasses generates additional microscopic defects inside grains and decreases free carriers that are generate by H-doping, leading to shorten the carrier relaxation time. ACKNOWLEDGMENTS
This work was supported by New Energy and Industrial Technology Development Organization 共NEDO兲 under Ministry of Economy, Trade, and Industry 共METI兲, Japan. P. Drude, Z. Phys. 1, 161 共1900兲. I. Hamberg and C. G. Granqvist, J. Appl. Phys. 60, R123 共1986兲. 3 R. Groth, Phys. Status Solidi B 14, 69 共1966兲. 4 Y. Kanai, Jpn. J. Appl. Phys., Part 1 23, 127 共1984兲. 5 Y. Meng, X. Yang, H. Chen, J. Shen, Y. Jiang, Z. Zhang, and Z. Hua, Thin Solid Films 394, 218 共2001兲. 6 Y. Yoshida, D. M. Wood, T. A. Gessert, and T. J. Coutts, Appl. Phys. Lett. 84, 2097 共2004兲. 7 C. Warmsingh, Y. Yoshida, D. W. Readey, C. W. Teplin, J. D. Perkins, P. A. Parilla, L. M. Gedvilas, B. M. Keyes, and D. S. Ginley, J. Appl. Phys. 95, 3831 共2004兲. 8 A. E. Delahoy and S. Y. Guo, J. Vac. Sci. Technol. A 23, 1215 共2005兲. 9 M. F. A. M. van Hest, M. S. Dabney, J. D. Perkins, D. S. Ginley, and M. P. Taylor, Appl. Phys. Lett. 87, 032111 共2005兲. 10 P. F. Newhouse, C.-H. Park, D. A. Keszler, J. Tate, and P. S. Nyholm, Appl. Phys. Lett. 87, 112108 共2005兲. 11 T. Koida and M. Kondo, Appl. Phys. Lett. 89, 082104 共2006兲. 12 X. Li, Q. Zhang, W. Miao, L. Huang, Z. Zhang, and Z. Hua, J. Vac. Sci. Technol. A 24, 1866 共2006兲. 13 T. Koida and M. Kondo, J. Appl. Phys. 101, 063705 共2007兲. 14 T. Koida and M. Kondo, J. Appl. Phys. 101, 063713 共2007兲. 15 J. A. A. Selvan, A. E. Delahoy, S. Guo, and Y.-M. Li, Sol. Energy Mater. Sol. Cells 90, 3371 共2006兲. 16 A. E. Delahoy, L. Chen, M. Akhtar, B. Sang, and S. Guo, Sol. Energy 77, 785 共2004兲. 17 T. Miyano, R. Hashimoto, Y. Kanda, T. Mise, and T. Nakada, Proceedings of the Technical Digest of 17th International Photovoltaic Science and Engineering Conference, Fukuoka, December 2007 共unpublished兲, p. 806. 18 X. Wu, J. Zhou, A. Duda, J. C. Keane, T. A. Gessert, Y. Yan, and R. Noufi, Prog. Photovoltaics 14, 471 共2006兲. 19 J. W. Bowers, H. M. Upadhyaya, S. Calnan, R. Hashimoto, T. Nakada, and A. N. Tiwari, Prog. Photovoltaics 17, 265 共2008兲. 1 2
033514-11 20
J. Appl. Phys. 107, 033514 共2010兲
Koida et al.
T. Koida, H. Fujiwara, and M. Kondo, Jpn. J. Appl. Phys., Part 2 46, L685 共2007兲. 21 T. Koida, H. Fujiwara, and M. Kondo, Appl. Phys. Express 1, 041501 共2008兲. 22 T. Koida, H. Sai, and M. Kondo, Thin Solid Films 共to be published兲. . 23 C. G. Van de Walle, Phys. Rev. Lett. 85, 1012 共2000兲. 24 C. Kilic and A. Zunger, Appl. Phys. Lett. 81, 73 共2002兲. 25 A. Janotti and C. G. Van de Walle, Nature Mater. 6, 44 共2007兲. 26 H. Omura, H. Kumomi, K. Nomura, T. Kamiya, M. Hirano, and H. Hosono, J. Appl. Phys. 105, 093712 共2009兲. 27 S. F. J. Cox, E. A. Davis, S. P. Cottrell, P. J. C. King, J. S. Lord, J. M. Gil, H. V. Alberto, R. C. Vilao, J. Piroto Duarte, N. Ayres de Campos, A. Weidinger, R. L. Lichti, and S. J. C. Irvine, Phys. Rev. Lett. 86, 2601 共2001兲. 28 D. M. Hofmann, A. Hofstaetter, F. Leiter, H. Zhou, F. Henecker, B. K. Meyer, S. B. Orlinskii, J. Schmidt, and P. G. Baranov, Phys. Rev. Lett. 88, 045504 共2002兲. 29 Y. M. Strzhemechny, H. L. Mosbacker, D. C. Look, D. C. Reynolds, C. W. Litton, N. Y. Garces, N. C. Giles, L. E. Halliburton, S. Niki, and L. J. Brillson, Appl. Phys. Lett. 84, 2545 共2004兲. 30 G. A. Shi, M. Stavola, S. J. Pearton, M. Thieme, E. V. Lavrov, and J. Weber, Phys. Rev. B 72, 195211 共2005兲. 31 S. Ishibashi, Y. Higuchi, Y. Ohta, and K. Nakamura, J. Vac. Sci. Technol. A 8, 1399 共1990兲. 32 Y. Shigesato, I. Yasui, Y. Hayashi, S. Takaki, T. Oyama, and M. Kamei, J. Vac. Sci. Technol. A 13, 268 共1995兲. 33 M. Ando, M. Takabatake, E. Nishimura, F. Leblanc, K. Onisawa, and T. Minemura, J. Non-Cryst. Solids 198–200, 28 共1996兲. 34 E. Nishimura, M. Ando, K. Onisawa, M. Takabatake, and T. Minemura, Jpn. J. Appl. Phys., Part 1 35, 2788 共1996兲. 35 B. H. Lee, I. G. Kim, S. W. Cho, and S.-H. Lee, Thin Solid Films 302, 25 共1997兲. 36 I. L. Eisgruber, J. R. Engel, R. E. Hollingsworth, P. K. Bhat, and R. Wendt, J. Vac. Sci. Technol. A 17, 190 共1999兲. 37 E. Nishimura, H. Ohkawa, P. K. Song, and Y. Shigesato, Thin Solid Films 445, 235 共2003兲. 38 K. Noda, H. Sato, H. Itaya, and M. Yamada, Jpn. J. Appl. Phys., Part 1 42, 217 共2003兲. 39 M. Ando, E. Nishimura, K. Onisawa, and T. Minemura, J. Appl. Phys. 93, 1032 共2003兲. 40 Y. Shigesato, Y. Hayashi, and T. Haranoh, Appl. Phys. Lett. 61, 73 共1992兲. 41 H. Nakazawa, Y. Ito, E. Matsumoto, K. Adachi, N. Aoki, and Y. Ochiai, J. Appl. Phys. 100, 093706 共2006兲. 42 M. Mizuhashi, Thin Solid Films 70, 91 共1980兲. 43 C. A. Pan and T. P. Ma, Appl. Phys. Lett. 37, 163 共1980兲. 44 Z. Ovadyahu, B. Ovryn, and H. W. Kraner, J. Electrochem. Soc. 130, 917 共1983兲.
H. Fujiwara and M. Kondo, Phys. Rev. B 71, 075109 共2005兲. P. I. Rovira and R. W. Collins, J. Appl. Phys. 85, 2015 共1999兲. 47 R. A. Synowicki, Thin Solid Films 313–314, 394 共1998兲. 48 T. Koida, H. Fujiwara, and M. Kondo, J. Non-Cryst. Solids 354, 2805 共2008兲. 49 Y. Shigesato, Y. Hayashi, A. Masui, and T. Haranou, Jpn. J. Appl. Phys., Part 1 30, 814 共1991兲. 50 D. H. Zhang and H. L. Ma, Appl. Phys. A: Mater. Sci. Process. 62, 487 共1996兲. 51 A. W. Ott, J. M. Johnson, J. W. Klaus, and S. M. George, Appl. Surf. Sci. 112, 205 共1997兲. 52 J. R. Bellingham, W. A. Phillips, and C. J. Adkins, J. Phys.: Condens. Matter 2, 6207 共1990兲. 53 R. Clanget, Appl. Phys. 2, 247 共1973兲. 54 Y. Ohhata, F. Shinoki, and S. Yoshida, Thin Solid Films 59, 255 共1979兲. 55 B. E. Sernelius, K.-F. Berffren, Z.-C. Jin, I. Hamberg, and C. G. Granqvist, Phys. Rev. B 37, 10244 共1988兲. 56 T. Pisarkiewicz, K. Zakrzewska, and E. Leja, Thin Solid Films 174, 217 共1989兲. 57 T. Nagatomo, Y. Maruta, and O. Omoto, Thin Solid Films 192, 17 共1990兲. 58 T. Minami, H. Sato, K. Ohashi, T. Tomofuji, and S. Takata, J. Cryst. Growth 117, 370 共1992兲. 59 S. Brehme, F. Fenske, W. Fuhs, E. Nebauer, M. Poschenrieder, B. Selle, and I. Sieber, Thin Solid Films 342, 167 共1999兲. 60 D. L. Young, T. J. Coutts, V. I. Kaydanov, A. S. Gilmore, and W. P. Mulligan, J. Vac. Sci. Technol. A 18, 2978 共2000兲. 61 M. Chen, Z. L. Pei, X. Wang, Y. H. Yu, X. H. Liu, C. Sun, and L. S. Wen, J. Phys. D: Appl. Phys. 33, 2538 共2000兲. 62 A. V. Singh, R. M. Mehra, A. Yoshida, and A. Wakahara, J. Appl. Phys. 95, 3640 共2004兲. 63 S. Muranaka, Y. Bando, and T. Takada, Thin Solid Films 151, 355 共1987兲. 64 R. G. Dhere, T. A. Gessert, L. L. Schilling, A. J. Nelson, K. M. Jones, H. Aharoni, and T. J. Coutts, Solar Cells 21, 281 共1987兲. 65 H. Morikawa, H. Sumi, and M. Kohyama, Thin Solid Films 281–282, 202 共1996兲. 66 C. W. Ow-Yang, D. Spinner, Y. Shigesato, and D. C. Paine, J. Appl. Phys. 83, 145 共1998兲. 67 D. C. Paine, T. Whitson, D. Janiac, R. Beresford, C. O. Yang, and B. Lewis, J. Appl. Phys. 85, 8445 共1999兲. 68 A. Rogozin, N. Shevchenko, M. Vinnichenko, F. Prokert, V. Cantelli, A. Kolitsch, and W. Moller, Appl. Phys. Lett. 85, 212 共2004兲. 69 T. A. Gessert, D. L. Williamson, T. J. Coutts, A. J. Nelson, K. M. Jones, R. G. Dhere, H. Aharoni, and P. Zurcher, J. Vac. Sci. Technol. A 5, 1314 共1987兲. 70 J. E. Bellingham, W. A. Phillips, and C. J. Adkins, Thin Solid Films 195, 23 共1991兲. 45 46
