Canadian Solar Buildings Conference
`Montreal, August 20-24, 2004
`Refereed Paper
`
`OPTICAL, ELECTRICAL AND STRUCTURAL PROPERTIES OF PECVD QUASI
`EPITAXIAL PHOSPHOROUS DOPED SILICON FILMS ON CRYSTALLINE
`SILICON SUBSTRATE
`
`Mahdi Farrokh-Baroughi, Hassan El-Gohary, and Siva Sivoththaman
`Department of Electrical & Computer Engineering, University of Waterloo
`200 University Avenue West, Waterloo, Ontario N2L 3G1, Canada.
`
`EXPERIMENTAL
`Films with different thicknesses were employed in
`this study on three kinds of substrates: CZ - Si wafers
`with resistivity of 8-12 (cid:525).cm, multicrystalline Si
`substrates (mc-Si) with resistivity of 1-2 (cid:525).cm, and
`glass wafers.
`
` For TEM analysis, we used the mc-Si wafers with
`random crystal orientations
`to
`investigate
`the
`structure of the deposited films and the orientation
`dependence of the deposition rates. The CZ, mc-Si
`and glass substrates were employed in conductivity
`measurements. For optical characterization of the
`developed films some test solar cells were fabricated
`on mc-Si substrates.
`
` Since the deposited films follow the crystalline
`order of the c-Si substrates, the clean and oxide-free
`surface is crucial to achieve a high crystallinity in the
`deposited film. We cleaned the Si substrates using
`RCA cleaning processes and performed an HF dip
`for 15 sec in 1% HF solution right before the loading
`of the wafers into the PECVD chamber. The
`hydrogen bonds on wafer surface formed during the
`HF dip prevent the formation of the native oxide
`during the short period of the wafer handling and
`pumping processes. (n+) qEpi-Si films of different
`thicknesses with the deposition conditions listed in
`table 1 were deposited on the substrates. To prepare
`the samples for conductivity measurements, a two
`mask process was employed to pattern the active (n+)
`region and the electrical contacts. The Al/(n+)qEpi-
`Si/Al structure was used for
`the conductivity
`measurements and the Al/(n+)qEpi-Si/(p)mc-Si/Al
`structure was employed as test photodiode for optical
`characterization of the film. The front side and the
`backside Al
`layers of 0.5μm
`thickness were
`deposited using RF magnetron sputtering.
`
`ABSTRACT
`We have developed highly conductive silicon
`emitters using plasma enhanced chemical vapor
`deposition (PECVD) of highly hydrogen-diluted
`silane and phosphine on crystalline silicon substrates.
`High resolution transmission electron microscope
`images show a perfect long range crystalline silicon
`order at the first few tens of nanometers of the film
`thickness. In higher thicknesses, however, a slowly
`varying shorter range order was observed. The
`comparison of the film conductivities on glass (10
`(cid:525)-1cm-1- 20 (cid:525)-1cm-1) and on c-Si substrates (110 –
`450 (cid:525)-1cm-1) confirms an improved doping efficiency
`and carrier mobility due to the higher crystallinity in
`the films deposited on the c-Si substrates. The optical
`properties of the films were assessed by quantum
`efficiency measurements on test solar cells.
`INTRODUCTION
`Silicon-based thin films such as amprphous and
`microcrydtalline (Poissant et al., 2003) films have
`been extensively studied for solar cell applications. It
`has been shown recently that the direct PECVD of a
`thin intrinsic Si film on a crystalline silicon (c-Si)
`using highly hydrogen diluted silane results in a
`quasi-epitaxial (qEpi) growth of Si film where the
`deposited film follows the crystal order of the
`substrate (Pla et al., 2002). Such films have been
`recently employed as intrinsic buffer layer in a
`heterojunction solar cell structure (Centurioni et al.,
`2004). However, to the best of our knowledge, q-Epi
`growth of highly doped highly conductive emitters
`has not been reported. This paper presents a new
`highly conductive (n+) Si emitter for PV application.
`The obtained conductivity values in the range of 100
`(cid:525)-1cm-1–450 (cid:525)-1cm-1 fall into the category of high
`temperature diffusion emitters (Kerr et al., 2001)
`rather than low temperature (LT) PECVD emitters
`(Saha et al., 1997, Alpuim et al., 2003). These
`emitters provide a new work frame for LT Si solar
`cells. This makes the use of simple solar cell
`structures without transparent conductive oxides
`possible (Farrokh-Baroughi et al., 2006). This paper
`presents
`the
`structural, electrical and optical
`properties of the qEpi-Si films.
`
`HANWHA 1062
`
`1
`
`

`

`Canadian Solar Buildings Conference
`Montreal, August 20-24, 2004
`Refereed Paper
`
`to
`
`the
`
`different crystal orientations are close
`deposition rates on glass substrate.
` The quality of the interface between the (n+) qEpi-
`Si film and the c-Si substrate plays an important role
`in the quality of the PV cells based on these
`junctions. Figure 2(a) shows the HRTEM picture of
`the interface of a (n+) qEpi-Si/(p)mc-Si junction. As
`shown in the figure, the deposited film in the
`interface and close to the interface shows a long
`range order similar to the long range order of the c-Si
`substrate. This supports the idea of quasi-epitaxial
`growth of the deposited film at LTs. Meanwhile, the
`figure shows
`that
`there
`is not a well-defined
`boundary between the deposited qEpi-Si film and the
`c-Si substrate and there is just a very thin transition
`region between
`the
`two materials. The high
`crystallinity of the film in the interface and close to
`the interface suggests that the electronic quality of
`the film in this region is good.
`
`Grain
`boundary
`
`Twin
`boundaries
`
`qEpi-Si film
`
`(a)
`
`SiN capping layer
`
`qEpi-Si film
`
`SiN capping layer
`
`(b)
`
`Figure 1. TEM cross-section of a SiN / (n+) qEpi-Si /
`mc-Si structure (a) at the vicinity of a GB and (b) at
`the vicinity of two twin boundaries.
`
`
`
`Figure 2(b) shows that the qEpi-Si film has followed
`the crystal orientation of the substrate up to at least
`30nm. Although
`these
`figures show
`that
`the
`deposited film follows the substrate orientation to
`some extent, figure 2(c) shows that the long range
`order in the deposited film was lost in higher
`thicknesses. Lattice images at different directions far
`
`
`
`Table 1
`Process conditions for the deposition of qEp-Si films
`on c-Si substrates
`
`
`
`RF POWER
`70 mW/cm2
`
`PRESSURE
`900 mT
`
`PH3/SIH4/H2
`0.04/4/500 sccm
`
`MEASUREMENT RESULTS AND
`DISCUSSION
`Structural properties of the (n+) qEpi-Si
`The important structural aspects of the (n+) qEpi-Si
`films are
`the substrate and crystal orientation
`dependence of the deposition rate, the quality of the
`interface between
`the deposited film and
`the
`underlying c-Si substrate, and the quality of the bulk
`of the deposited film. We have employed HRTEM
`analysis for structural study of the (n+) qEpi-Si films.
` As the first application of the HRTEM images, we
`compared the deposition rate of the qEpi-Si films on
`c-Si and glass substrates. Because of the similar
`physical properties of the qEpi-Si and c-Si materials
`it is not possible to selectively etch the qEpi-Si film
`on the c-Si substrate. TEM pictures, however, are
`able to show the difference between two materials.
`Furthermore, since the growth mechanisms of the
`qEpi-Si film using process parameters of table 1 on
`glass and c-Si substrate are different, epitaxial
`growth versus nc-Si growth, the deposition rates can
`be potentially different on these two substrates.
`Therefore, we designed an experiment to investigate
`this matter.
` The deposition rate of 2.05 nm/min was measured
`on glass under the qEpi-Si process conditions listed
`in table 1. Figure 1(a) shows the TEM picture of a
`(n+) qEpi-Si /mc-Si cross-section covered by a
`silicon nitride capping layer. The (n+) qEpi-Si film in
`this structure was obtained by 45 min PECVD of the
`Si film using process parameters of table 1. The film
`thickness in the TEM picture is about 90 nm which is
`in good agreement with the predicted film thickness
`from the measured deposition rate on a glass
`substrate. Another important aspect of the deposition
`is the dependence of the deposition rate to the crystal
`orientation. Figure 1(a) and 1(b) show the qEpi-Si
`films of different thicknesses deposited on mc-Si
`wafers with random crystal orientations. A grain
`boundary (GB) and two twin boundaries (TBs) are
`visible in figure 1(a) and 1(b). It should be noted that
`the crystal orientation on the two sides of a GB or a
`TB are different. Since the film thickness in both
`figures is the same on different sides of the GB and
`TBs, we conclude that the deposition rate is not
`sensitive to the crystal orientation. As a result, we
`think that the deposition rates on c-Si wafers with
`
`2
`
`

`

`
`from the interface are observable in figure 2(c). This
`suggests that the quality of the film decays at higher
`thicknesses. However, looks like at least in less than
`100 nm thick films the changes is the crystal
`orientations is smooth and there are not sharp grain
`boundaries in the film. This is an important point that
`differentiates the LT qEpi-Si film from the nc-Si
`film.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Figure 2: HRTEM picture of the qEpi-Si and c-Si (a)
`at the interface region in 5 nm scale, (b) at the
`interface region in 10 nm scale, (c) qEpi-Si film of
`90nm thickness on the mc-Si substrate, and (d)bulk
`of the qEpi-Si at the vicinity of a low angle grain
`boundary within the film
`
`
`
`
`
`
`Canadian Solar Buildings Conference
`Montreal, August 20-24, 2004
`Refereed Paper
`
`qEpi-Si
`
`Interface
`
`c-Si
`
`(a)
`
`qEpi-Si
`
`Interface
`
`c-Si
`
`(b)
`
`qEpi-Si
`
`c-Si
`
`Interface
`
`(c)
`
`qEpi-Si
`
`Low
`angle GB
`
`qEpi-Si
`
`(d)
`
`
`
`3
`
`

`

`Canadian Solar Buildings Conference
`Montreal, August 20-24, 2004
`Refereed Paper
`
`major assumption in the electrical characterization
`was that the space charge region in (n+) emitter is
`much thinner than the film thickness. We will show
`that this is a valid assumption.
` Figure 3 shows the conductivities of the films of
`different thicknesses deposited on CZ substrates with
`(100) orientation and glass substrates using the
`deposition conditions of table 1. The experiments
`reveal three major differences between the qEpi-Si
`films on CZ-Si and the nc-Si films on glass: (i) the
`conductivity of the qEpi-Si films are at least one
`order of magnitude larger than the conductivity of
`the nc-Si films, (ii) the conductivity of both films
`depend on the films thickness, and (iii) the thickness
`dependence of the films on the two substrates show
`an opposite behavior. The conductivity of the qEpi-
`Si films decreases by increasing the film thickness
`and the conductivity of the nc-Si films increases by
`increasing the film thickness.
` The huge difference in the conductivity of the films
`originates from higher crystallinity of the qEpi-Si
`films on
`the c-Si substrates. The
`thickness
`dependence of the conductivity of the qEpi-Si films
`suggest that the crystallinity and hence the doping
`efficiency and the carrier mobility of the qEpi-Si film
`decreases far from the interface region. This is
`consistent with the results of the structural studies in
`the previous section. On the other hand, thickness
`dependence of the nc-Si conductivity shows that the
`crystallinity of
`the film gets better at higher
`thickness. This is a well known phenomenon that is
`observed on nc-Si
`films developed on glass
`substrates (Luysberg et al., 1997). The observations
`here pinpoint to the major structural difference of the
`qEpi-Si film with normal nc-Si films. This result is
`consistent with our observations from the TEM
`analysis in the previous section.
`
`Conductivity of nc-Si on glass
`
`12
`
`10
`
`16
`
`14
`
`300
`
`250
`
`200
`
`Conductivity of QE-Si on CZ
`
`30
`
`60
`50
`40
`Film thickness (nm)
`
`70
`
`
`
`Figure 3. Conductivity of the deposited films on CZ-
`Si wafer and glass in (cid:525)-1cm-1.
`To better understand the electrical properties of the
`qEpi-Si, the conductivity of the films are discussed
`here in more details. The conductivity of our (n+)
`
`
`two films
`the
`This major difference between
`originates from the substrate effect. Unlike nc-Si
`films on glass that the crystallites start at a point in
`an a-Si phase and grow over
`thickness,
`the
`crystallites of the qEpi-Si film grow directly on the c-
`Si substrate with the same crystal orientation without
`experiencing any incubation layer. In other words the
`c-Si acts as a crystal seed layer for the growth of the
`qEpi-Si film. The qE-Si film loses its crystallinity
`because of the LT nature of the process. Nonetheless,
`changes in the crystal orientations far from the c-Si
`substrate are expected
`to be smooth because
`crystallites start with the same orientations. Figure
`2(d) shows a smooth change in crystal orientation
`deep inside of the qEpi-Si film.
`Electrical properties of the (n+) qEpi-Si
` Electrical conductivity of the emitter depends on
`two parameters: the carrier mobility and the active
`doping density. Unlike high yemperature (HT) Si
`emitters obtained by diffusion of dopants in single c-
`Si wafers, the LT Si emitters normally have low
`conductivities (Alpuim et al., 2003). A good quality
`(n+) a-Si has a conductivity of 0.01 (cid:525)-1cm-1 (Street,
`1991) and a good quality thin (n+) nc-Si has a much
`higher conductivity of 10-20 (cid:525)-1cm-1 [4]. It is known
`that during the deposition of a-Si at LT, most of the
`dopant (phosphorous in n-type material) atoms prefer
`to form three fold bonds (electrically inactive states)
`rather than 4 fold bonds (electrically active donor
`state). This highly reduces the doping efficiency [7].
`For example, the doping efficiency of (n+) a-Si film
`with 1% phosphine concentration in gas phase is
`about 1%. Although the doping efficiency in a nc-Si
`film is much higher than that of a-Si, the presence of
`a-Si
`tissues
`in
`this material especially at
`the
`incubation layer lowers the doping efficiency for a
`thin film. The only key to obtain highly conductive
`Si emitters using PECVD at LTs is to improve the
`crystallinity of the deposited films because the higher
`the crystallinity the higher the doping efficiency and
`carrier mobility. The thin (less that 100nm thickness)
`(n+) Si films in this work show an outstanding
`crystallinity throughout the film. As the TEM
`pictures suggest, the films have the best crystallinity
`at the interface and close to the interface and have
`less crystallinity far from the interface. Nonetheless,
`the crystallinity far from the interface is still good
`and amorphous tissues are not observed in the films.
`As a result, higher conductivities are expected in
`qEpi-Si emitters.
`In order
`to measure
`the
`conductivity of the films, we deposited the qEpi-Si
`films with different thicknesses on glass and CZ-Si
`wafers and measured the sheet resistance of the films
`using
`simple
`electrical
`test
`structures. The
`conductivities of the films were calculated using the
`sheet resistance value and the thickness values of the
`films that were measured on the glass substrates. The
`
`4
`
`

`

`Canadian Solar Buildings Conference
`Montreal, August 20-24, 2004
`Refereed Paper
`
`mobility values for the films with conductivities in
`the range of 100 – 450 (cid:525)-1cm-1 assuming typical
`values for ND and bandgap narrowing. The estimated
`electron mobility values without considering the
`bandgap narrowing for such highly doped material
`are higher than the mobility of c-Si material.
`However,
`the mobility values considering
`the
`bandgap narrowing are close to the reported mobility
`values for (n+) c-Si material (Arora et al, 1982,
`Caughey et al., 1967). As a result, we believe that the
`electrical quality of the (n+) qEpi-Si films falls in the
`category of c-Si emitters rather than LT Si emitters.
`This is the major point of this research work.
`
`EC=EC0
`EC=EC0- 0.05 eV
`EC=EC0- 0.1 eV
`
`ND
`
`n0
`
`1024
`
`1023
`
`1022
`
`1021
`
`1020
`
`1019
`
`1018
`
`1017
`
`Concentration (cm-3)
`
`1016
`-6
`
`-4
`
`-2
`
`2
`
`4
`
`6
`
`0
`(EF-EC0)/kT
`
`Figure 4. Active doping density (ND) and free
`electron concentration (n0) in the c-Si as a function
`of the Fermi level position calculated for different
`values for the conduction band edges. The bandgap
`narrowing at EC of 0, 0.05eV and 0.1eV was
`assumed in this calculation.
`
`
`Table II
`Electron mobility estimation in the (n+) qEpi-Si films
`obtained on mc-Si and CZ-Si wafers
`
`
`BG
`narrowing
`0
`50meV
`100meV
`
`ND=2.5x1020
`(cid:305) =100
`(cid:305) =450
`
`ND=5x1020
`(cid:305)=100
`(cid:305)=450
`
`27.17
`13.52
`7.83
`
`122.26
`60.84
`35.24
`
`20.16
`10.38
`6.25
`
`90.72
`46.71
`28.12
`
`Optical properties of the (n+) qEpi-Si
`Similar to electrical properties, the optical properties
`of the (n+) qEpi-Si film deposited on Si substrate are
`different than the films deposited on glass substrate.
`We have fabricated some test solar cells with 1cm2
`Al/SiN/(n+)qEpi-Si/(p)mc-Si/Al
`area
`using
`structures. The thickness of the SiN layer was 75 nm
`and
`the emitter
`thicknesses was 70 nm. The
`measured internal quantum efficiencies of the solar
`cells were employed for the characterization of the
`films. Figure 5 (a) shows the IQE of a sample test
`solar cell. The drop in the IQE in the short
`wavelength regime, in 375nm – 500 region, is
`
`
`qEpi-Si films developed at LT using PECVD
`technique falls in the range of 100 (cid:525)-1cm-1 to 450
`(cid:525)-1cm-1. In addition, the conductivity of highly doped
`single c-Si wafers is normally between 100 and 1000
`(cid:525)-1cm-1. Therefore, we assume that a thin qEpi-Si
`film has the same material parameters of a single c-Si
`material and we assess the conductivity.
` Consider an (n+) Si film with a conductivity of
`100 – 450 (cid:525)-1cm-1 with phosphorous concentration
`of NPh and an active donor concentration of ND. The
`material is highly degenerate; therefore the Fermi-
`Dirac integral should be employed for finding the
`free electron concentration. The charge neutrality in
`+ = n0 where ND
`+ and n0 are
`this film reduces to ND
`the ionized active donor density and free electron
`density, respectively.
` The free electron concentration and the density of
`ionized donor states in the film are given by,
`N
`E
`E
`2
`(cid:16)
`C
`kT
`(cid:83)
`(cid:12))
`(cid:11)
`N
`1
`(cid:14)
`D
`D
`D
`where the EC, EF, NC, EC, F1/2, f, k and T are
`conduction band edge, , Fermi energy, effective
`density of states in the conduction band edge, Fermi
`integral, Fermi-Dirac distribution function, Boltzman
`constant, and temperature, respectively.
`+ in the charge neutrality
` Substituting n0 and ND
`and simplifying
`the expression we obtain
`the
`relationship between ND and the position of the
`Fermi level.
`
`
`
`(cid:184)(cid:185)(cid:183)
`
`F
`
`C
`
`(cid:168)(cid:169)(cid:167)
`
`F
`2/1
`
`(cid:16)
`
`
`
` (Ef
`
`n
`
`0
`
`
`
`(cid:32)
`
`N
`
`(cid:32)
`
`
`
`(1)
`
` (2)
`d
`(cid:75)
`
`F
`
`E
`
`C
`
`)
`
`(cid:75)
`E
`
`exp(
`(cid:16)
`(cid:75)
`
`1
`
`(cid:14)
`
`(cid:179)(cid:102)
`
`0
`
`(cid:117)
`
`2
`
`N
`C
`(cid:83)
`
`(cid:184)(cid:185)(cid:183)
`
`(cid:117)(cid:184)
`
`E
`
`(cid:16)
`
`F
`
`E
`(cid:16)
`D
`kT
`
`1
`
`(cid:14)
`
`e
`
`(cid:168)(cid:168)(cid:169)(cid:167)
`
`N
`
`D
`
`(cid:32)
`
`(cid:16)
`kT
`where ED is the energy level of the phosphorous
`dopant atoms in the c-Si bandgap. Figure 4 is the plot
`of ND and n0 obtained versus the Fermi-level energy
`respect to EC0 (the conduction band edge of the c-Si
`material without considering bandgap narrowing)
`normalized to the thermal potential. Since the
`bandgap narrowing affects the charge concentration
`at high doping densities, we have calculated the
`curves with 0, 0.05 eV and 0.1 eV bandgap
`narrowings at the conduction band.
`silicon
`to
` The
`gas
`phase
`phosphorous
`concentration ratio during depositing (n+) qEpi-Si
`film was 0.01. If an equal gas phase and solid phase
`phosphorous
`to silicon
`ratio
`is assumed,
`the
`phosphorous concentration in the film becomes
`5x1020 cm-3. The value of ND becomes equal to
`5x1020 cm-3 and 2.5x1020 cm-3 for a 100% and 50%
`doping efficiencies, respectively. For ND of 5x1020
`cm-3 the n0 in the range of 3x1019 cm-3 to 9x1019 cm-
`3 is obtained from figure 4. Table 2 lists the estimated
`
`5
`
`

`

`Canadian Solar Buildings Conference
`Montreal, August 20-24, 2004
`Refereed Paper
`
`CONCLUSION
`A new kind of (n+) qEpi-Si emitters with very high
`conductivities was demonstrated at LTs using
`PECVD technique. HRTEM pictures from the qEpi-
`Si films showed that the film benefits from a high
`degree of crystallinity at the interface region and far
`from the interface. Using TEM analysis and electrical
`measurements, it was confirmed that the films have
`the best structural and electronic quality at the
`interface and close to the interface region and less
`quality far from the c-Si substrate. The obtained
`conductivity values in the range of 100–450 (cid:525)-1cm-1
`are comparable to the conductivities of the diffused
`c-Si emitters. Optical absorption of the emitter was
`characterized at short wavelength
`region and
`absorption coefficient of 2x105 cm-1 was obtained at
`400 nm.
`ACKNOWLEDGMENT
`The authors would like to thank NSERC for funding
`this project.
`REFERENCES
`Poissant, Y., Chatterjee, P., and Roca I Cabarrocas,
`P. 2003, Journal of Applied Physics, 93 170-
`174.
`Pla, J., Centurioni, E., Summonte, C., Rizzoli, R.
`2002, Thin Solid Films 405 248-255.
`Centurioni, E., Iencinella, D., Rizzoli, R., and
`Zignani, F. 2004, IEEE Transaction on Electron
`Devices, 51 1818-1824.
`Kerr, M.J., Schmidt, J., and Cuevas, A. 2001, J.
`Appl. Phys., 89 3821-3826.
`Saha S.C., Rath, J.K., Kshirsagar, S.T., Ray, S. 1997,
`Journal of Physics D: Applied Physics, 30 2686–
`2692
`Alpuim, P., Chu, V., Conde, J.P. 2003, J. of Vacuum
`Science & Technology A 21 1048-1054.
`Farrokh-Baroughi, M., Sivoththaman, S., 2006, Proc.
`IEEE 4th World Conference on Photovoltaic
`Energy Conversion, Hawaii, USA
`Street, R. A., 1991, Hydrogenated Amorphous
`Silicon, Cambridge University Press
`Luysberg, M., Hapke, P., Carius, R., Finger, F. 1997,
`Philosophical magazine A 75 31.
`Arora N.D., Hauser, J.R., and Roulston, D.J. 1982,
`IEEE Transaction on Electron Devices, 29 292-
`295
`Caughey, D.M., and Thomas, R.E. 1967, Proc.
`IEEE, 55 2192-2193.
`
`
`because of the optical loss in the emitter. If we
`assume that the emitter contribution in photocurrent
`is negligible, which is a good approximation for
`degenerate LT emitters, we can easily extract the
`effective emitter absorption coefficient ((cid:302)e) using
`(cid:302)e=-ln(IQE)/tE, where the tE is the emitter thickness.
`the calculated (cid:302)e
`in short
`Figure 5(b) shows
`wavelength regime for the solar cell of figure 5(a). It
`should be noted that the extracted values are not
`exactly equal to the absorption coefficient of the film
`because we have not considered the effect of the
`emitter contribution in IQE. Nonetheless, the (cid:302)e
`values of figure 7(b) are in good agreement with the
`reported values for thick nc-Si films [4]. However,
`the obtained (cid:302)e values in the 375 – 500 nm region are
`less than the absorption coefficients of (n+) a-Si films
`and thin nc-Si films.
`
`400
`
`500
`
`800
`700
`600
`Wavelength (nm)
`(a)
`
`900 1000 1100
`
`1.0
`
`0.8
`
`0.6
`
`0.4
`
`0.2
`
`0.0
`
`Normalized IQE
`
`105
`
`(cid:68)e(cm-1)
`
`475
`
`500
`
`450
`425
`Wavelength (nm)
`(b)
`
`104
`375
`
`400
`
`
`Figure 5. (a) The measured internal quantum
`efficiency of a fabricated photodiode based on the
`(n+)qEpi-Si/(p) mc-Si junction. (b) Extracted effective
`absorption coefficient of the (n+) qEpi-Si thin film
`from the IQE of figure 5(a).
`This clearly indicates that the higher crystallinity of
`the (n+) qEpi-Si films make then more transparent
`than the (n+) a-Si and the thin (n+) nc-Si film.
`Considering the fact that the obtained values from
`IQE measurements are in good agreement with the
`thick,
`it can be concluded
`that
`the emitter
`contribution in the photocurrent is negligible.
`
`6
`
`