Enhanced spectral response of semiconducting BaSi2 films by oxygen incorporation

Accepted Manuscript Enhanced spectral response of semiconducting BaSi2 films by oxygen incorporation

Weijie Du, Ryota Takabe, Suguru Yachi, Kaoru Toko, Takashi Suemasu PII: DOI: Reference:

S0040-6090(17)30231-6 doi: 10.1016/j.tsf.2017.03.046 TSF 35894

To appear in:

Thin Solid Films

Received date: Revised date: Accepted date:

27 September 2016 4 March 2017 23 March 2017

Please cite this article as: Weijie Du, Ryota Takabe, Suguru Yachi, Kaoru Toko, Takashi Suemasu , Enhanced spectral response of semiconducting BaSi2 films by oxygen incorporation. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Tsf(2017), doi: 10.1016/j.tsf.2017.03.046

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Enhanced spectral response of semiconducting BaSi2 films by oxygen incorporation Weijie Du,a Ryota Takabe,b Suguru Yachi,b Kaoru Toko,b and Takashi Suemasub,* Key Laboratory of Optoelectronic Material and Device, Mathematics & Science College,

T

a

b

CR

IP

Shanghai Normal University, Shanghai 200234, China

Faculty of Pure and Applied Sciences, Institute of Applied Physics, University of Tsukuba,

AN

US

Tsukuba, Ibaraki 305-8573, Japan

M

We investigated the effect of the incorporation of O atoms into BaSi2 films on their

ED

photoresponse properties. BaSi2 films with higher O concentration exhibited higher photoresponsivity. Time-of-flight secondary ion mass spectrometry measurements showed that

PT

the O atoms were uniformly distributed in the BaSi2 films, in contrast to our prediction that they

CE

would be mostly located around grain boundaries. First-principles calculations revealed that the O atoms occupy the interstitial sites known as the 4c sites rather than substitutional sites.

AC

Moreover, they do not create localized states within the forbidden band gap, which indicates that O atoms incorporated into BaSi2 are inactive.

Keywords: solar cells; oxygen; efficiency; semiconducting silicide; first-principles calculation

*Corresponding author at: University of Tsukuba, Faculty of Pure and Applied Sciences, Institute of Applied Physics, 1-1-1 Tennohdai, Tsukuba, Ibaraki 305-8573, Japan. Tel: +81 29 853 5111. E-mail address: [email protected] (T. Suemasu).

1

ACCEPTED MANUSCRIPT 1. Introduction Over the last few years, thin-film solar cell materials such as cadmium telluride and chalcopyrite have attracted more and more attention owing to their high efficiency and low cost [1-4]. Perovskite solar cells are also attracting increasing attention owing to their high conversion

T

efficiency exceeding 20% and simple fabrication process [5,6]. However, these cells contain

IP

expensive and/or toxic elements. As an alternative, we have focused on the semiconductor

CR

barium disilicide (BaSi2), which consists of only earth abundant elements [7,8]. BaSi2 has many attractive features for thin-film solar cell applications [9]. The band gap of BaSi2 is

US

approximately 1.3 eV, and thus matches the solar spectrum much better than crystalline Si

AN

[10,11]. Theoretical studies have revealed that BaSi2 has a relatively flat band structure, and hence has a very large absorption coefficient α of approximately 3 × 104 cm−1 at 1.5 eV, which is

M

comparable to that of other thin-film solar cell materials [12,13]. a-axis-oriented BaSi2 can be

ED

grown epitaxially on Si(111) substrate owing to a small lattice mismatch of approximately 1%. Besides, intrinsically doped n-BaSi2 has a long minority-carrier lifetime (>10 μs) [14], and so a

PT

large minority-carrier (holes) diffusion length (ca. 10 μm) [15]. Classic solar cells use a pn

CE

junction, which means that control of the carrier type and carrier concentration is very important. Intrinsically doped BaSi2 shows n-type conductivity with an electron concentration on the order

AC

of 1016 cm−3 [10,16]. Hereafter, we denote intrinsically doped n-BaSi2 simply as n-BaSi2. By doping with boron (B), we achieved p-type BaSi2 layers [17,18], a vital step for the formation of BaSi2 homojunction solar cells. Very recently, we achieved a conversion efficiency η approaching 10% with a p-BaSi2/n-Si heterojunction solar cell [19,20]. This is the simplest possible solar cell using BaSi2. These results demonstrate the high potential of BaSi2 for thinfilm solar cell applications. In our previous works, we found that surface native oxides of BaSi2

2

ACCEPTED MANUSCRIPT are beneficial for improving the minority-carrier lifetime [21,22]. We also found that the η of pBaSi2(20 nm)/n-Si solar cells increased with air exposure duration up to 150 h prior to the formation of indium tin oxide (ITO) surface electrodes, in contrast to our prediction [23]. Oxygen (O) atoms are likely to become incorporated into the grown layers during such a long

T

period of air exposure. Indeed, oxygen is the greatest impurity in BaSi2 films. The influence of O

IP

related defects in Si has been widely studied [24,25]. However, the influence of O atoms on the

CR

properties of BaSi2 has yet to be studied. The oxidation of solar cell materials is inevitable because they are exposed to air during practical use. Hence, it is of great importance to

US

understand what happens when O is incorporated into the BaSi2 lattice.

AN

Herein, we examined the influence of O concentration on the photoresponse properties of n-BaSi2 thin films grown under different vacuum levels. Theoretical calculations were also

M

conducted to reveal the most probable O site in the orthorhombic unit cell of BaSi 2 and its

ED

influence on the density of states (DOS).

PT

2. Method

CE

2.1 Experiments

An ion-pumped molecular beam epitaxy (MBE) system (R-DEC) equipped with a

AC

Knudsen cell for Ba and an electron beam gun for Si was used for the growth of BaSi 2 films. An attached electron impact emission spectroscopy (EIES; INFICON) feedback system was used to accurately stabilize the deposition rates of Si and Ba. Before growth, substrates were first cleaned according to standard RCA (Radio Corporation of America) procedure, followed by thermal cleaning at 900°C for 30 min in the ultra-high vacuum chamber. To investigate the effect of O concentration on the photoresponsivity of BaSi2 films, a-axis-oriented epitaxial n-BaSi2

3

ACCEPTED MANUSCRIPT films were grown at 600°C (samples A1A5) on low-resistivity (resistivity ρ < 0.02 Ωcm) Czochralski n-Si(111) substrates by MBE. The contribution of photoexcited carriers originating from such low-resistivity Si substrates to the measured photoresponse spectra could be excluded. The detailed growth method of the a-axis-oriented epitaxial n-BaSi2 films has been reported

T

previously [9]. Table 1 summarizes the BaSi2 layer thickness, background pressure, PBG, and

IP

pressure during MBE, PMBE, for samples A1–A5. PMBE was changed by the presence or absence

CR

of liquid nitrogen in the shroud of the MBE chamber. O depth profiles were evaluated by secondary ion mass spectrometry (SIMS) at least a few days after the sample growth. To

US

measure the spectral response of the films, 200-nm-thick ITO electrodes were deposited on the

AN

front side of the samples by radio frequency (RF) magnetron sputtering. Back contacts were formed with Al by RF magnetron sputtering. The photoresponse spectra were measured under a

M

bias voltage of 1 V between the top and the bottom electrodes at room temperature (RT) using a

ED

lock-in technique employing a xenon lamp with a 25 cm focal-length single monochromator (Bunko Keiki, SM-1700 A). Reflectance spectra were evaluated with a reflection measurement

PT

system using a xenon lamp and integrating sphere. To investigate the three-dimensional

CE

distribution of O atoms, an additional sample with a grain size of more than 2 μm (sample B) [26] was grown for time-of-flight SIMS (TOF-SIMS; PHI TRIFV V nanoTOF) measurements.

AC

Plan-view transmission electron microscopy (TEM; Topcon EM-002B) operated at 120 kV was employed to investigate the grain size of the BaSi2.

2.2 Calculation details O-doped BaSi2 formation energies, DOSs, and band structures were calculated using the Vienna Ab initio Simulation Package code [27] based on density-functional theory with the

4

ACCEPTED MANUSCRIPT projector-augmented wave pseudopotential [28] and Perdew-Wang Generalized Gradient Approximation method [29]. DOSs of Sb, Al, In, and B-doped BaSi2 are given in our previous report [30]. Total energy minimization was obtained via optimization of the lattice parameters and relaxation of the atomic positions in a conjugate gradient routine. Using an energy cutoff of

T

600 eV and a 6 × 8 × 4 grid of Monkhorst–Pack points, the convergence in the total energy was

IP

better than 1 meV/atom [31]. The stoichiometric description of the unit cell is Ba8Si16. In each

CR

BaSi2 unit cell, there are two crystallographically inequivalent sites for Ba (Ba(1) and Ba(2)) and three inequivalent sites for Si (Si(3), Si(4), and Si(5)). Therefore the atoms are distributed over

US

4Ba(1), 4Ba(2), 4Si(3), 4Si(4), and 8Si(5). Hereafter, we describe the O-doped BaSi2 as Ba7O(1)Si16,

AN

in which one of the Ba(1) sites is substituted with O, or Ba8O(3)Si15, in which one of the Si(3) sites is substituted with O. There are 16 candidate interstitial sites in the BaSi2 lattice. According to

M

Imai and Watanabe [32], the most probable insertion sites are the 4c sites, where an impurity

ED

atom is surrounded by three Si atoms, one located at a corner of one Si-tetrahedron and the other two composing an edge of the other Si-tetrahedron. Thus, we chose one of the 4c sites, the

PT

fractional coordinate of which is (0.5841, 0.25, 0.2251). This compound is described as

CE

Ba8Si16O. The calculated formation energies of these compounds are summarized in Table 2.

AC

3. Results and discussion

Figure 1(a) shows the SIMS depth profiles of O in samples A1A5. In samples A2, A3, and A5, a steep rise in O concentration was observed, followed by a rapid decrease. The peak positions corresponded to the BaSi2/Si interface, and such a profile may have been caused by insufficient thermal cleaning of the Si substrate in the MBE chamber. An O concentration of over 1 × 1019 cm−3 was detected in all samples. The O atoms were found to be uniformly

5

ACCEPTED MANUSCRIPT distributed in the depth direction of the BaSi2 layers, indicating that the O atoms did not diffuse from the sample surface. As shown in Table 1, the O concentration in each sample was closely related to PMBE. For samples A1A4, PBG was almost the same. Because the liquid nitrogen cold shroud was not used for samples A3, A4, and B during their growth, PMBE increased from 8 ×

T

10−7 to 3 × 10−6 Pa, while the O concentration also increased from approximately 2 × 1019 cm−3

IP

for samples A1 and A2 to approximately 1 × 1020 cm−3 for samples A3, A4, and B. For sample

CR

A5, we intentionally increased PBG by loosening the nuts of one vacuum flange of the MBE chamber. Therefore, in this case both PBG and PMBE were increased by nearly one order of

US

magnitude compared with those for the other samples. As a consequence, the O concentration in

AN

sample A5 increased to more than 5 × 1020 cm−3. Figure 1(b) shows the normalized internal quantum efficiency (IQE) spectra of samples A1A5 measured under a bias voltage of 1 V at

M

RT. Short-wavelength light was absorbed in the region close to the surface. For example, for

ED

BaSi2, α reaches 4 × 105 cm−1 at a wavelength of 500 nm [11], meaning that most of the photons

PT

of this wavelength are absorbed as they pass only 75 nm (1/α × 3 = 75 nm) through the BaSi2 surface. Therefore, IQEs in the short wavelength range could be compared among samples

CE

A1A5 regardless of their BaSi2 layer thickness. Fig. 1(b) shows that samples with higher O concentration tended to show large IQE. The IQEs of samples A3 and A4 were significantly

AC

higher than those of samples A1 and A2. This result reveals that a certain amount of O is beneficial for enhancing the spectral response of BaSi2 films. Moreover, although the O concentration of sample A5 was much higher than those of the other four samples, the IQE still remained relatively high. We next examined whether the O atoms were positioned around grain boundaries (GBs). An a-axis-oriented epitaxial BaSi2 film contains a large amount of GBs because it has three

6

ACCEPTED MANUSCRIPT epitaxial variants rotated by 120° around the surface normal [26]. Dark-field transmission electron microscopy observations were performed on sample B under a two-beam diffraction condition to clarify the grain size of the BaSi2. The results are shown in Fig. 2(a) alongside a selected-area diffraction (SAED) pattern. The diffraction vector g was set to [004]. Under these

T

conditions, the diffraction spot corresponding to the (004) plane becomes bright in SAED

IP

patterns, while other spots denoted by (00n) (n = ±1, ±2, ±3,...) also become evident. The BaSi2

CR

grains that satisfy Bragg’s condition of diffraction become bright; those of one of the three BaSi2 epitaxial variants, which provides information regarding the grain size. As shown in Fig. 2(a),

US

the grain size of BaSi2 was found to be larger than 1 μm in sample Β, which is much larger than

AN

the in-plane spatial resolution of TOF-SIMS, approximately 0.10.2 μm. To investigate how the O atoms were spatially distributed in the grown layer, TOF-SIMS measurements were conducted

M

on sample B. We aimed to confirm whether the O atoms accumulated along the GBs. Figure 2(b)

ED

shows in-plane O distribution images measured at different depths from the surface of sample B. As can be seen, O atoms were quite uniformly distributed in the BaSi2, and to our surprise, no

PT

accumulation phenomenon was observed. Owing to the much higher diffusion coefficients of the

CE

GBs than the lattice diffusion coefficients normal to the BaSi2 epitaxial film [33], we assumed that O atoms might be more likely to exist along GBs if they had diffused from the surface or if

AC

the GBs acted as impurity gettering sites. The results imply that the O atoms in the BaSi2 mostly originated from inside the chamber during the growth rather than diffused from the surface after the sample was exposed to air. To elucidate the position of O atom in the BaSi2 unit cell, we conducted first-principles calculations as shown in Table 2. From a formation energy point of view, one O atom is likely to occupy the Ba(1) site when it replaces one Ba atom. However, it is more likely to be positioned at

7

ACCEPTED MANUSCRIPT the Si(5) site than the Ba(1) site even though the difference in formation energy between Ba8O(3)Si15, Ba8O(4)Si15, and Ba8O(5)Si15 is small. The results in Table 2 suggest that O is most likely to occupy the interstitial site in the Ba8Si16 unit cell from an energetic point of view, which totally differs from other impurities in BaSi2 such as Al, In, and Sb [30], i.e., these impurities are

T

most likely to occupy the Si site of the BaSi2 lattice.

IP

Figure 3(a) shows the crystal structure of BaSi2, while Fig. 3(b) depicts a Si tetrahedron

CR

in the BaSi2 lattice without the O atom. There are four identical Si tetrahedral in the unit cell, in which one Si(4) (red) and two Si(5) atoms (blue) have almost the same a-axis coordinate, and form

US

an isosceles triangle. When one O atom (yellow) occupies the interstitial 4c site as shown in Fig.

AN

3(c), the atomic positons differ slightly from those in Fig. 3(a). The Si-Si bonds in Fig. 3(c) are colored pink or gray depending on whether the Si-Si interatomic distance is decreased or

M

increased, respectively, compared with those in Fig. 3(b). Figure 3(d) shows the projection of the

ED

BaSi2 unit cell shown in Fig. 3(c) on the b-c plane. Figure 4 shows the partial DOSs of (a) Ba8Si16, (b) Ba8O(5)Si15, and (c) Ba8Si16O near the

PT

Fermi level, EF, which is taken as the energy zero in each case. As shown in Fig. 4(a), the

CE

valence-band maximum (VBM) in BaSi2 is mainly composed of the Si p state, while the conduction-band minimum (CBM) consists mostly of the Si p and Ba d states. In Ba8O(5)Si15,

AC

however, the band gap shrinks significantly, due to localized states mostly composed of Ba-d and Si-p states, just below the CBM (Fig. 4(b)). In Ba8Si16O, which is most likely to occur as indicated by Table 2, no localized state is observed in the band gap, as shown in Fig. 4(c). Figure 5(a) and (b) show the energy band structures of Ba8Si16 and Ba8Si16O, respectively. In Ba8Si16, the CBM is located at Τ(0, 1/2, 1/2) and the VBM is located at approximately (0, 1/3, 0) along the -Υ(0, 1/2, 0) direction [12,34,35]. The direct transition occurs at approximately (0, 1/3, 0),

8

ACCEPTED MANUSCRIPT and its gap value is greater than the band gap by approximately 0.1 eV. In Ba8Si16O, however, the VBM is shifted by approximately (1/2,1/2,0) along the S(1/2,1/2,0)–X(1/2, 0, 0) direction, resulting in a broadening of the band gap. From the theoretical calculations we conclude that the O atoms tend to exist at the interstitial 4c sites of the BaSi2 lattice and do not give rise to a defect

T

state within the band gap, which is in good agreement with the experimental results. We should

IP

also point out that we did not observe the expansion of the band gap macroscopically because the

CR

actual incorporation of O atoms in the BaSi2 was much smaller than that in the calculations. On the basis of these results, we can at least state that O incorporation does not generate defective

AN

US

states in BaSi2.

4. Conclusion

M

We have investigated the relationship between O concentration and photoresponse

ED

properties in n-BaSi2 films grown by MBE. An O concentration of over 1 × 1019 cm−3 was detected in all samples. BaSi2 films possessing a higher O concentration tended to exhibit a

PT

higher photoresponsivity. TOF-SIMS indicated that the O atoms were uniformly spatially

CE

distributed in the BaSi2, while first-principles calculations indicated that they likely occupied the interstitial 4c sites in the unit cell. Both the theoretical and experimental results revealed that the

AC

incorporation of O does not generate defective states within the band gap of BaSi2, which is very important information for further research on this material.

Acknowledgments This work was supported in part by the Core Research for Evolutional Science and Technology (CREST) project of the Japan Science and Technology Agency (JST) and by a

9

ACCEPTED MANUSCRIPT Grant-in-Aid for Scientific Research A (No. 15H02237) from the Japan Society for the Promotion of Science (JSPS). The authors thank Dr. N. Yoshizawa and Mr. N. Saito of AIST,

AC

CE

PT

ED

M

AN

US

CR

IP

T

Japan, for their help with TEM observations.

10

ACCEPTED MANUSCRIPT Fig. 1. (a) SIMS depth profiles of O in samples A1–A5, (b) IQE spectra of samples A1–A5 normalized using the IQE spectrum of sample A4. A bias voltage of 1 V was applied between the top and bottom electrodes at RT.

T

Fig. 2. (a) Dark-field plan-view TEM image with diffraction spots obtained under two-beam

IP

diffraction condition using the diffraction vector g = [004]. One of the three epitaxial variants

CR

appears bright. (b) TOF-SIMS images of in-plane O at different depths (indicated by numerical

US

values) from the surface of sample B.

AN

Fig. 3. (a) Crystal structure of BaSi2. There are two crystallographically-inequivalent sites for

M

Ba (Ba(1) and Ba(2)) and three inequivalent sites for Si (Si(3) in orange, Si(4) in red, and Si(5) in

ED

blue) in the unit cell of BaSi2. (b) Si tetrahedron in the BaSi2 lattice. One Si(4) atom (red) and two Si(5) atoms (blue) have similar a-axis coordinates, and form an isosceles triangle. The Si-Si

PT

interatomic distances are shown. (c) Crystal structure of BaSi2 when one O atom (yellow) is

CE

positioned in the interstitial 4c site. Si-Si bonds are labelled in pink or gray depending on whether their interatomic distance is shortened or lengthened, respectively, compared with

AC

without the O atom in (b). (d) Projection of the BaSi2 unit cell in (c) on the b-c plane.

Fig. 4. Partial DOSs of (a) Ba8Si16, (b) Ba8O(5)Si15, and (c) Ba8Si16O near EF, which was taken as the energy zero.

Fig. 5. Energy band structures of Ba8Si16 and Ba8Si16O. Dotted lines are shown to guide the eye. EF was taken as the energy zero in each case.

11

ACCEPTED MANUSCRIPT References [1] M. G. Panthani, V. Akhavan, B. Goodfellow, J. P. Schmidtke, L. Dunn, A. Dodabalapur, P. F. Barbara, B. A. Korgel, Synthesis of CuInS2, CuInSe2, and Cu(InxGa1-x)Se2 (CIGS) nanocrystal "inks" for printable photovoltaics, J. Am. Chem. Soc. 130 (2008) 16770-16777.

T

[2] W. Jaegermann, A. Klein, T. Mayer, Interface engineering of inorganic thin-film solar cells -

IP

materials-science challenges for advanced physical concepts, Adv. Mater. 21 (2009) 4196-4206.

CR

[3] S. Ahmed, K. B. Reuter, O. Gunawan, L. Guo, L. T. Romankiw, H. Deligianni, A high

US

efficiency electrodeposited Cu2ZnSnS4 Solar Cell, Adv. Ener. Mater. 2 (2012) 253-259.

AN

[4] C. Steinhagen, M. G. Panthani, V. Akhavan, B. Goodfellow, B. Koo, B. A. Korgel, Synthesis of Cu2ZnSnS4 nanocrystals for use in low-cost photovoltaics, J. Am. Chem. Soc. 131 (2009)

M

12554.

Photonics 8 (2014) 506-514.

ED

[5] M. A. Green, A. Ho-Baillie, H. J. Snaith, The emergence of perovskite solar cell, Nature

PT

[6] H. S. Kim, S. H. Im, N. G. Park, Organolead halide perovskite: new horizons in solar cell

CE

research, Journal of Physical Chemistry C 118 (2014) 5615-5625.

AC

[7] J. Evers, G. Oehlinger, A. Weiss, Crystal-structure of barium disilicide at high-pressures, Angew. Chem. Int. Ed. Engl., 16 (1977) 659-660. [8] M. Imai, T. Hirano, Electrical resistivity of metastable phases of BaSi2 synthesized under high pressure and high temperature, J. Alloys Compd. 224 (1995) 111-116. [9] T. Suemasu and N. Usami, Eploring the Potential of Semiconducting BaSi2 for Thin-Film Solar Cell Applications, J. Phys. D: Appl. Phys. 50 (2017) 023001

12

ACCEPTED MANUSCRIPT [10] K. Morita, Y. Inomata, T. Suemasu, Optical and electrical properties of semiconducting BaSi2 thin films on Si substrates by molecular beam epitaxy, Thin Solid Films 508 (2006) 363366. [11] K. Toh, T. Saito, T. Suemasu, Optical absorption properties of BaSi2 epitaxial films grown

IP

T

on a transparent silicon-on-insulator substrate using molecular beam epitaxy, Jpn. J. Appl. Phys.

CR

50 (2011) 068001.

[12] D. B. Migas, V. L. Shaposhnikov, V. E. Borisenko, Isostructural BaSi2, BaGe2 and SrGe2:

US

electronic and optical properties, Phys. Status Solidi (b) 244 (2007) 2611-2618.

AN

[13] M. Kumar, N. Umezawa, M. Imai, BaSi2 as a promising low-cost, earth-abundant material with large optical activity for thin-film solar cells: A hybrid density functional study, Appl. Phys.

M

Express 7 (2014) 071203.

ED

[14] K. O. Hara, N. Usami, K. Toh, M. Baba, K. Toko, T. Suemasu, Investigation of the

CE

112 (2012) 083108.

PT

recombination mechanism of excess carriers in undoped BaSi2 films on silicon, J. Appl. Phys.

[15] M. Baba, K. Toh, K. Toko, N. Saito, N. Yoshizawa, K. Jiptner, T. Sekiguchi, K. O. Hara, N.

AC

Usami, T. Suemasu, Investigation of grain boudaries in BaSi2 epitaxial films on Si(111) substrates using transmission electron microscopy and electron-beam-induce current technique, J. Cryst. Growth 348 (2012) 75-79. [16] W. Du, M. Baba, K. Toko, K. O. Hara, K. Watanabe, T. Sekiguchi, N. Usami, T. Suemasu, Analysis of the electrical properties of Cr/n-BaSi2 Schottky junction and n-BaSi2/p-Si heterojunction diodes for solar cell applications, J. Appl. Phys. 115 (2014) 223701.

13

ACCEPTED MANUSCRIPT [17] M. A. Khan, K. Hara, W. Du, M. Baba, K. Nakamura, M. Suzuno, K. Toko, N. Usami, T. Sueamsu, In-situ heavily p-type doping of over 1020 cm-3 in semiconducting BaSi2 thin films for solar cells applications, Appl. Phys. Lett. 102 (2013) 112107. [18] M. A. Khan, K. Nakamura, W. Du, K. Toko, N. Usami, T. Suemasu, Precipitation control

T

and activation enhancement in boron-doped p+-BaSi2 films grown by molecular beam epitaxy,

IP

Appl. Phys. Lett.104 (2014) 252104.

CR

[19] D. Tsukahara, S. Yachi, H. Takeuchi, R. Takabe, W. Du, M. Baba, Y. Li, K. Toko, N.

US

Usami, T. Suemasu, p-BaSi2/n-Si heterojunction solar cells with conversion efficiency reaching

AN

9.0%, Appl. Phys. Lett. 108 (2016) 152101.

[20] S. Yachi, R. Takabe, K. Toko, T. Suemasu, Effect of amorphous Si capping layer on the

M

hole transport properties of BaSi2 and improved conversion efficiency approaching 10% in p-

ED

BaSi2/n-Si solar cells, Appl. Phys. Lett. 109 (2016) 072103. [21] R. Takabe, K. O. Hara, M. Baba, W. Du, N. Shimada, K. Toko, N. Usami, T. Suemasu,

PT

Influence of grain size and surface condition on minority-carrier lifetime in undoped n-BaSi2 on

CE

Si(111), J. Appl. Phys. 115 (2014) 193510.

AC

[22] R. Takabe, W. Du, K. Ito, H. Takeuchi, K. Toko, S. Ueda, A. Kimura, T. Suemasu, Measurement of valence-band offset at native oxide/BaSi2 interfaces by hard x-ray photoelectron spectroscopy, J. Appl. Phys. 119 (2016) 025306. [23] R. Takabe, S. Yachi, W. Du, D. Tsukahara, H. Takeuchi, K. Toko, T. Suemasu, Influence of air exposure duration and a-Si capping layer thickness on the performance of p-BaSi2/n-Si heterojunction solar cells, AIP Advnaces 6 (2016) 085107.

14

ACCEPTED MANUSCRIPT [24] R. Whan, Oxygen-defect complexes in neutron-irradiated silicon, J. Appl. Phys. 37 (1966) 3378. [25] A. Chroneos, E. Sgourou, C. Londos, U. Schwingenschlögl, Oxygen defect processes in

T

silicion and silicon germanium, Appl. Phys. Reviews 2 (2015) 021306.

IP

[26] M. Baba, K. Toh, K. Toko, K. O. Hara, N. Usami, N. Saito, N. Yosihzawa, T. Suemasu,

CR

Formation of large-grain-size BaSi2 epitaxial layers grown on Si(111) by molecular beam

US

epitaxy, J. Cryst. Growth 378 (2013) 193-197.

[27] G. Kresse and D. Joubert, From ultrasoft pseudopotentials to the projector augmented-wave

AN

method, Phys. Rev. B 59 (1999) 1758.

M

[28] P. E. Blöchl, Projector augmented-wave method, Phys. Rev. B 50 (1994) 17953.

ED

[29] J. P. Perdew, J. A. Chevary, S. H. Vosko, K. A. Jackson, M. R. Pederson, D. J. Singh, C. Fiolhais, Atoms, molecules, solids, and surfaces: Applications of the generalized gradient

PT

approximation for exchange and correlation, Phys. Rev. B 46 (1992) 6671.

CE

[30] D. Tsukahara, M. Baba, S. Honda, Y. Imai, K. O. Hara, N. Usami, K. Toko, J. Werner, T.

AC

Suemasu, Potential variations around grain boundaries in impurity-doped BaSi2 epitaxial films evaluated by Kelvin probe force microscopy, J. Appl. Phys. 116 (2014) 123709.

[31] H. J. Monkhorst, J. D. Pack, Special points for Brillouin-zone integrations, Phys. Rev. B 13 (1976) 5188-5192. [32] Y. Imai, A. Watanabe, Energetic evaluation of possible insertion sites of Cu into BaSi2 using first principle calculations, Intermetallics 19 (2011) 1102-1106.

15

ACCEPTED MANUSCRIPT [33] K. Nakamura, M. Baba, M. Ajmal Khan, W. Du, M. Sasase, K. O. Hara, N. Usami, T. Suemasu, Lattice and grain-boundary diffusions of boron atoms in BaSi2 epitaxial films on Si(111), J. Appl. Phys. 113 (2013) 053511. [34] M. Kumar, N. Umezawa, M. Imai, (Sr,Ba)(Si,Ge)2 for thin-film solar-cell applications:

IP

T

First-principles study, J. Appl. Phys. 115 (2014) 203718.

CR

[35] Y. Imai, A. Watanabe, Consideration of the band-gap tenability of BaSi2 by alloying with

AC

CE

PT

ED

M

AN

US

Ca or Sr based on the electronic structure calculations, Thin Solid Films 515 (2007) 8219-8225.

16

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

AC

CE

PT

Fig. 1

17

AC

CE

PT

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

Fig. 2

18

Fig. 3

AC

CE

PT

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

19

AC

Fig. 4

CE

PT

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

20

US

CR

IP

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

Fig. 5

21

ACCEPTED MANUSCRIPT Table 1. BaSi2 layer thickness, PBG, and PMBE for samples A1–A5 and B.

BaSi2 layer

PBG

PMBE

(Pa)

(Pa)

Cooled by liquid N2

(nm)

A1

690

< 1 × 10-7

8 × 10-7

Yes

A2

940

< 1 × 10-7

8 × 10-7

Yes

A3

630

< 1 × 10-7

3 × 10-6

A4

1100

< 1 × 10-7

3 × 10-6

A5

630

ca. 1 × 10-6

ca. 1 × 10-5

Yes

B

250

< 1 × 10-7

3 × 10-6

No

No No

AC

CE

PT

ED

M

AN

US

CR

IP

T

Sample

22

ACCEPTED MANUSCRIPT Table 2. Calculated formation energies of O-doped BaSi2.

(eV)

Ba8Si16 + O  Ba7O(1)Si16 + Ba

134.446 (132.5341.912)

(2)

Ba8Si16 + O  Ba7O Si16 + Ba

133.963 (132.0511.912) 141.310 (135.8935.417)

Ba8Si16 + O  Ba8O(4)Si15 + Si

141.382 (135.9655.417)

Ba8Si16 + O  Ba8O(5)Si15 + Si

141.398 (135.9815.417)

AN

US

CR

Ba8Si16 + O  Ba8O(3)Si15 + Si

IP

Compound

T

Total energy

141.494

AC

CE

PT

ED

M

Ba8Si16 + O  Ba8Si16O

23

ACCEPTED MANUSCRIPT Highlights ∙BaSi2 films were grown by molecular beam epitaxy. ∙BaSi2 films with higher oxygen (O) concentrations showed higher photoresponsivity.

AC

CE

PT

ED

M

AN

US

CR

∙They do not create localized states within the forbidden gap.

IP

T

∙The O atoms are most likely to occupy the interstitial site from first-principles calculations.

24