第二章 CuInSe 2 薄膜硫化處理之研究
2 實驗原理、步驟與分析儀器
Polycrystalline CuInSe2 thin- films were grown on soda lime glass substrates with the MBE technique. High-purity elemental sources of 6N Cu, 6N In, and 5N Se were used. Two commercial effusion cells made by ADDON Inc. were used for the evaporation of Cu and In sources. The temperature of these cells could be precisely controlled within 0.2 oC. A MAXTEK MDC-360 thin- film deposition controller monitored individual source flux with a quartz-crystal sensor. CuInSe2 films with different compositions were prepared by varying the beam flux of the In source, while the beam fluxes of the Cu and Se sources were kept constant. The films were deposited on 1- inch square soda lime glass substrates with a rotation speed of 12 rpm.
The substrate temperature was kept at 550 oC and the deposition time was 1 hour.
The film thickness was measured to be about 1.0 µm. A Cu-rich CuInSe2 epitaxial film was specially prepared on a (001) GaAs substrate at 600 oC for studying the effect of film crystallinity on sulfide conversion. The detail of preparation and characterization of CuInSe2 epitaxial film had published in a separate paper [19].
Sulfurization was carried out in another vacuum chamber equipped with a sulfur effusion cell. Elemental sulfur with 5N purity was used. The beam flux of sulfur source was kept constant of 4.5×1016 atoms/cm2-sec. The annealing temperature was from 300 to 450 oC. The temperature profiles for substrate and sulfur source heating are given in Fig. 1. Having the sulfur source ready prior to substrate heating is essential to improving film adhesion and homogeneity.
A Siemens D5000 X-ray diffractometer was used for the identification of crystal structure and the phases that exist in the film. Chemical compositions of CuInSe2
and CuInS2 films were measured by a JEOL JXA-8900R electron microprobe (EPMA) with beam energy of 10 keV. The detail of composition analysis was published elsewhere [20]. TEM samples prepared by SMI 3050 Focus Ion Beam and measurement was performed in a FEI E.O Tecnai F20 G2 MAT S-TWIN Field Emission Gun Transmission Electron Microscope with EDAX, Gatan Image Filter (GIF) and Electron Energy Loss Spectroscopy (EELS). The electron beam diameter used for performing the measurements was 1 nm.
5
normally found on the film surface. Chemical compositions of film before and after sulfurization are listed in Table 1. The Cu/In ratio was reduced from 1.05 to 0.85 after KCN etching. The Se content was less than 50 % in the as-grown film and slightly increased after KCN etching.
Our experiments showed evidence that the original composition of CuInSe2 film may influence the speed of the sulfide conversion process. The results shown in Fig.
2 indicated that an In-rich CuInSe2 film might not be converted completely after the same annealing conditions used for the Cu-rich film. The second phase of Cu2-xSe that existed on the surface of a Cu-rich film may play an important role in the conversion speed. This argument is verified by comparing the conversion results obtained from a Cu-rich film before and after KCN etching. It is clear that the removal of second phase from the original film significantly slows down the sulfide conversion process, see Fig. 3. We had prepared another Cu-rich sample (Cu/In=1.1) with an epitaxial structure on a (001) GaAs substrate. This sample was only partially converted into sulfide using the same annealing conditions as the polycrystalline Cu-rich CuInSe2 films, see Fig. 4. It is evident that grain boundary diffusion is much faster than bulk diffusion.
TEM observation of these samples consists with X-ray analysis. Figure 5 shows a bright field cross sectional image of the partially converted epitaxial CuInSe2
film and the elemental mappings of the surface. The analysis of the image and the EDX mappings of the surface corroborate that the sulfur are mainly concentrated at the surface of the film and gradually decreases with depth for inner points (1 and 2) to form a quaternary compound CuIn(Se,S)2. Moreover, the composition of the bottom of the film is Cu(In,Ga)Se2 and gradually increases with depth for inner points (3 and 4). This is due to gallium diffusion into CuInSe2 from GaAs substrate. The voids show in figure 5 probably formed due to the thermal stress associated with the sulfide conversion process. The bandgap structure of the partially converted epitaxial CuInSe2 film shows a double profiling similar the highest efficiency polycrystalline Cu(In,Ga)Se2 solar cells.
Figure 6(a) shows XRD of a partially converted polycrystalline CuInSe2 film at temperature of substrate of 425 oC. The intensities of orientation (112) of CuInSe2
and sulfide compound were almost equal; it is suitable for the kinetics study on polycrystalline sulfide conversion process. Figure 6(b) and (c) show a bright field cross sectional image and a STEM image; it presented the Cu2(Se,S) not only exist on surface but also in grain boundaries and intragranular microstructure. These data can be interpreted assuming that a two-step sulfide conversion process, figure 7. First, a surface reaction which is the copper selenide and CuInSe2 converted into copper sulfide and CuInS2. Then followed by solid diffusion which is the copper sulfide filled the film via grain boundary. These copper sulfide acts as a diffusion source in sulfide conversion process.
4 結論
Our experimental results showed that complete conversion of a 1.0
µm-thick
polycrystalline Cu-rich CuInSe2 film into CuInS2 was achieved when the film was annealed in a sulfur beam flux of 4.5×1016 atoms/cm2-sec at 450oC for five minutes.The conversion speed depends on the film crystallinity, and original film composition.
We found that a high sulfur vapor flux is essential for rapid conversion. Another important factor was the original film composition. A Cu-rich CuInSe2 film converted into sulfide much faster than an In-rich film. The copper selenide phase played an important role in the conversion speed. This was confirmed by etching away the copper selenide phase from the Cu-rich sample which resulted in a much slower conversion being observed. In addition, the diffusion of sulfur into the selenide to replace the selenium was mainly through grain boundary in a polycrystalline film. The sulfide conversion process was associated with surface reaction and solid diffusion. The copper selenide on the surface and at the grain boundaries converted into copper sulfide and acts a diffusion source to filled the film.
7
References
[1] M.A. Contreras, B. Egaas, K. Ramanathan, J. Hiltner, A. Swartzlander, F.
Hasoon, R. Noufi, Prog. Photovolt: Res. Appl., 7 (1999) 311.
[2] K. Ramanathan, M.A. Contreras, C.L. Perkings, S. Asher, F.S. Hasoon, J. Keane, D. Young, M. Romero, W. Metzger, R. Noufi, J. Ward, A. Duda, Prog.
Photovolt: Res. Appl., 11 (2003) 225.
[3] Y.B. He, W. Kriegseis, B.K. Meyer, A. Polity, Appl. Phys. Lett., 83 (2003) 1743.
[4] K. Siemer, J. Klaer, L. Luck, J. Bruns, R. Klenk, D. Bräunig, Solar Energy Mater. Solar cells, 67 (2001) 159.
[5] J. Klaer, J. Bruns, R. Henninger, K. Siemer, R. Klenk, K. Ellmer, D. Bräunig, Semicond. Sci. Technol., 13 (1998) 1456.
[6] T. Walter, R. Menner, Ch. Köble, H.W. Schock, Proc. 12th ECPVSEC, Amsterdam, Kluwer Academic Publishing, Dordrecht,1994, p. 1755.
[7] M. Gossla, Th. Hahn, H. Metzner, J. Conrad, U. Geyer, Thin Solid Films 268 (1995) 39.
[8] H. Metzner, Th. Hahn, J.-H. Bremar, J. Conrad, Appl. Phys.Lett. 69 (1996) 1900.
[9] Y. Yamamoto, T. Yamaguchi, Y. Denizu, T. Tanaka, A. Yoshida, Thin Solid Films 281-282 (1996) 372.
[10] Y. Yamamoto, Y. Yamaguchi, T. Tanaka, N. Tanahashi, A.Yoshida, Sol.
Energy Mater. Sol. Cells 49 (1997) 399.
[11] R. Klenk, U. Blieske, V. Diesterle, K. Ellmer, S. Fiechter, I. Hengel, A.
Jäger-Waldau, T. Kampschulte, Ch. Kaufmann, J. Klaer, M.Ch. Lux-Steiner, D.
Braunger, D. Hariskos, M. Ruckh, H.W. Schock, Sol. Energy Mater. Sol. Cells 49 (1997) 349.
[12] Y. Ogawa, A. Jäger-Waldau, Y. Hashimoto, K. Ito, Jpn. J. Appl. Phys. 33 (1994) L1775.
[13] J. Klaer, J. Bruns, R. Henninger, K. Siemer, R. Klenk, K. Ellmer, D. Bräunig, Semicond. Sci. Technol. 13 (1998) 1456.
[14] K. Kondo, S. Nakamura, H. Sano, H. Hirasawa, K. Sato, Sol. Energy Mater. Sol.
Cells 49 (1997) 327.
[15] P. Guha, S. Gorai, D. Ganguli, S. Chaudhuri, Materials Letters 57 (2003) 1786.
[16] T. Ohashi, K. Inakoshi, Y. Hashimoto, K. Ito, Sol. Energy Mater. Sol. Cells 50 (1998) 37.
[17] M. Engelmann, B.E. McCandless, and R.W. Birkmire, Thin Solid Films, 387 (2001) 14
[18] J. Titus, H.W. Schock, R.W. Birkmire, W.N. Shafarman and U.P. Singh, Mat.
Res. Soc. Symp., 668 (2001) H1.5.1
[19] B.H. Tseng, S.B. Lin, G.L. Gu and W. Chen, J. Appl. Phys., 79, (1996) 1391.
[20] G.L. Gu, B.H. Tseng, and H.L. Hwang, J. of Physics and Chemistry of Solids, 64 (2003) 1901.
9
Table 1 A list of composition data of CuInSe2 before and after sulfide conversion
Sample
Cu (at.%)
In (at.%)
Se (at.%)
S (at.%)
Cu / In VI / M S / VI Description
SC-01 27.45 24.19 48.36 ? 1.13 0.94 ? A 27.68 23.64 2.31 46.38 1.17 0.95 0.95 SC SC-02 24.44 25.93 49.63 ? 0.94 0.99 ? A
25.40 25.70 13.89 35.01 0.99 0.96 0.72 SC PL-01 26.66 25.39 47.95 ? 1.05 0.92 ? A
29.00 23.63 1.78 45.6 1.23 0.90 0.96 SC KCN-01 26.63 25.26 48.11 ? 1.05 0.93 ? A
23.48 27.74 48.78 ? 0.85 0.95 ? E
A: as- grown CuInSe2 films
SC: CuInSe2 films after annealing in sulfur beam flux of 4.5x1016 atoms/cm2-sec at 450oC for five minutes
E: CuInSe2 film after 5% KCN etching for five minutes
FIGURE CAPTIONS
Figure 1 The temperature profiles of the substrate (a) and sulfur effusion-cell (b) in sulfurization process.
Figure 2 The XRD spectra of (a) Cu/In=1.13 and (b) Cu/In=0.94 CuInSe2 films on soda-lime glass substrate after sulfurization.
Figure 3 A comparison of X-ray diffraction data showing the completeness of sulfide conversion process for a Cu-rich (Cu/In=1.05) CuInSe2 film (a) before and (b) after KCN etching (Cu/In=0.85).
Figure 4 The XRD spectra of a single crystalline Cu-rich CuInSe2 (Cu/In=1.1) film on GaAs (100) wafer (a) before and (b) after sulfurization.
Figure 5 (a) Bright- field TEM cross sectional image of the partially converted epitaxial CuInSe2 film together with the composition table and (b) the elemental EDX mappings of the surface.
Figure 6 (a) XRD of a partially converted polycrystalline CuInSe2 film at temperature of substrate of 425 oC, (b) bright-field TEM cross sectional image and (c) STEM image.
Figure 7 Proposed two-step sulfide conversion process.
11
1
2
3
4
13
(a)
(c) (b)
Cu2(Se,S)
CuIn(Se,S) 2
6
7
Cu2(Se,S) Au
CuIn(Se,S) 2
SL Glass
15
As-grown Annealing
in Se
Sulfurization In-rich CuInSe2
24 25 26 27 28 29 30
CIS(112) CISe(112)
Intensity (a.u.)
2θ
Sulfurization In-rich CuInSe2
Soda-lime Glass
第三章 CuAlSe 2 薄膜製備與分析
17
1 研究目的
Soda-lime glass/Mo/CuInSe2/CdS/ZnO/Al-contact，為典型的 CuInSe2 薄膜太 陽能電池元件結構，如圖 1。其效率將受下列兩種因素所限制:一為其吸收層與透 明導電層能隙的差【1】，亦即由於 CuInSe2 與 ZnO 之間的導帶差(Conduction-band Difference，△EC)太大，造成 Band-Offset，對於少數載子-電子的傳遞與躍遷會 有所影響，進而影響其電性與轉換效率；一為因晶格邊界失調所造成的漏電流 (Leakage currents)【2】，因此需要緩衝層(Buffer layer)來降低 Band-Offset 所造成 的影響，另一方面可防止濺鍍沉積 ZnO 時所造成對 CuInSe2 薄膜的損害，而緩 衝層需要具備下列條件:首先緩衝層之能隙值須介於吸收層(Absorbing layer)與抗 反射層(Anti-reflection layer)之間，以上述元件設計為例，緩衝層能隙值須介於 1.01∼3.2ev；其次緩衝層須為 n-type 以便和 p-type 的吸收層形成接面(junction)
【3】；此外為了避免造成漏電流，緩衝層的電阻率需高於吸收層，亦即電阻率約 進一步利用在元件的設計上。此外由於未摻雜的 CuAlSe2 呈現 p-type，可用來與 n-type 的 CuInSe2 形成 p-n 接面(junction)，若成長多晶薄膜，可進而製造 CuAlSe2 /CuInSe2 串接式太陽電池；若成長單晶薄膜，可製造量子結構 CuAlSe2/CuInSe2 磊晶太陽電池，因此 CuAlSe2 薄膜品質的好壞，以及 CuAlSe2 與 CuInSe2 接面 的性質，都是本實驗所將研究的方向。
2 CuAlSe2 薄膜介紹
CuAlSe2 屬於 I-III-VI2 族化合物，是高能隙材料，於室溫之能隙值為 2.67eV。其結晶相具黃銅礦(Chalcopyrite)結構，晶格常數 a0=5.617(Å)、
c=10.92(Å)，呈現 p-type 的導電形式。並不像其他低能隙 Chalcopyrite 結構的半 導體材料，可由成分比例的調配，來形成 p-type 或 n-type。若要得到 n-type 的 CuAlSe2薄膜，必須藉由摻雜 Au/K、Zn、Cd 的方式【4】。從 Cu2Se-Al2Se3的擬 二元相圖(Pseudo Binary Phase Diagram)來看(圖 2)【5】，在溫度 1100℃以下 CuAlSe2為具有黃銅礦結構的γ相，而且可偏離化學組成 2mole %。
而在 CuAlSe2的應用方面，除了在薄膜太陽電池中擔任緩衝層的角色，或是 製造 CuAlSe2 /CuInSe2串接式太陽電池與量子結構 CuAlSe2/CuInSe2磊晶太陽電 池外，在製造短波長的光電元件方面也是一個很好的材料，如 blue light-emitting diodes 和 blue laser diodes【6】，特別是應用在 light-emitting diodes 時，和 ZnSe 可形成很好的異質結構，其晶格相差只有 1％左右【7】，若成長在 CuGaSe2 更理
想，其晶格常數相差不到 0.2％【8】。
3 實驗原理、步驟與分析儀器
3.1 分子束蒸鍍原理
分子束蒸鍍(Molecular Beam Deposition)是一種在超高真空(Ultra High Vacuum, UHV) 下，利用蒸發的元素態分子束，直接撞擊基板而沈積元素型
（elemental semiconductor, 如 Si）或化合物型（compound semiconductor, 如 GaAs）半導體的技術。之所以稱為分子束，是因為當氣體壓力減低至 10-4 torr
本實驗所使用的分子束蒸鍍系統(Molecular Beam Deposition)，如圖 3，包括 一超高真空成長腔(UHV Growth Chamber)及用來輸送試片基板的預抽腔
(Load-Lock Chamber)，此預抽腔亦具有 DC Sputter 的功用。而在成長腔與預抽腔 之間還有一個緩衝腔(Buffer Chamber)。成長腔使用 SEIKO SEIKI STP-400 渦輪 分子幫浦(Turbo Molecular Pump)抽氣，讓背景真空可達 2.0×10-8torr。緩衝腔則使 用離子幫浦(Ion Pump)，讓此腔體真空度維持在 8.0×10-7torr。而試片是先放至預 抽腔後，先由機械幫浦(Mechanical Pump )將真空度抽至 5×10-2torr 後，再用小型 渦輪分子幫浦將真空度抽到 5×10-5torr，之後即可將試片從預抽腔送至緩衝腔，
19
3.4 分析儀器
3.4.1 熱探針(Hot Probe)
為了測試薄膜的導電型態，利用兩探針，一為室溫，另一為高溫，利用電流 的正負，來判斷半導體材料為 n-type 或 p-type。其原理是因為 hot probe 瞬速加 熱薄膜，導致此熱端的自由載子動能增加，因此載子以比平常更快的速度擴散到 低溫端。若半導體為 n-type，電子將從熱端移動，而留下一個正電子的施體區，
而熱端
相對於冷端而言為正的，而電流從熱端流到冷端；若半導體為 p-type，則結 果剛好相反。【使用機型: HP34401A Multimeter】。
3.4.2 四點探針(4 point probe)
為了測量薄膜的電阻率，利用四平行探針，外面兩根探針通固定電流，裡面 兩根探針量電壓，將所得到的值代入ρ=(V/I)˙t˙CF(t : 厚度；CF:校正因子)，
即可得到薄膜的電阻率。【使用機型: JANDEL probe head】。
薄膜分析
1.熱探針 2.四點探針 3.X-ray 繞射 4.SEM
5.吸收光譜 6. Hall 量測 7. Rocking curve 8. PL 量測
9.背向式 Laue 繞射照相
21
定律(Scherrer’s Law): t = 0.9λ/Bcos? ( t：晶粒大小；λ：入射光波長；B：半高 寬；?：Bragg 角度)計算出晶粒的尺寸。【使用機型: SIEMENS D5000 XRD】。
3.4.4 吸收光譜儀
掃描式電子顯微鏡(SEM; scanning electronic microscopy)是運用高能的加速 電壓(5-10KeV)電子激發薄膜上的二次電子(secondary electronic)，收集其訊號在 (band gap)之能量，和材料及組成有關。【使用機型: JOEL JSM6330TF (CL System)】。
3.4.6 rocking Curve
本儀器為雙晶繞射儀，由兩塊晶體參與繞射;第一塊晶體通常用來作為單色
型: Bede D1 System(DID50373)】
3.4.7 霍爾量測(Hall Measurement)
霍爾量測系統包括溫控系統(K20)、磁場產生器(MPS50)及電源供應器(H50) 等整組系統，可做變溫量測，其溫度範圍由 80K 至 400K。此系統經由 RS232 連 線接到一對多接線盒(splitter)，再透過 PC 來控制系統運作來取得實驗數據。此 儀器主要是利用霍爾效應之原理，可用來量測半導體的載子濃度、遷移率、電阻
率及電導形式等。 (Reciprocal Lattice Point)相交時便可產生繞射。此繞射之 X 光將可使底片感光而 形成繞射圖形(Diffraction Pattern)。單晶所產生的繞射圖形將會是圓點(Spot)，或 是橢圓點。若是有複晶(Polycrystal)或非晶質(Amorphous)結構存在，繞射圖形將 變成弧線(Arc)，或成圈狀(Ring)。
3.4.9 光激光系統(photoluminescence)
光激光簡稱 PL，是以適當強度的雷射光入射待測試片，將薄膜中位於價帶 為 Donor to Acceptor(D-A)時，入射功率增加，PL 發光強度也會隨著增加，但 不是呈線性增加。而 D-A 躍遷還有另一特徵即發光峰能量會隨著激發光源強度的 增加而往高能量處位移，這是因為一相距 r 的施體(donor)和受體(acceptor)兩
23 閉循環式液態氦冷卻系統冷卻(close-cycle He cryostat)並搭配 LAKESHORE 320 型的溫控器，可降溫至 8°K 左右。激光光譜的偵測使用 ACTON RESEARCH SPECTRAPRO-500 型分光計(monochromator)，其光徑長(focal length)為 500mm，
解析度可達 0.05nm，並搭配 Si 與 Ge 偵測器(detector)，其可偵測的波長範圍分 別為 500~1000nm 與 900~1800nm，於本實驗中使用 Si 偵測器。訊號的處理以 STANFORD RESEARCH SR510 型號 LOCK-IN 擴大器並附有一遮光器(chopper) 可有效濾掉雜訊並放大訊號，同時可以使用 neutral density filters 調變雷射強度，
最後得到的訊號則由 RS-232 介面轉存入電腦中。 的訊號相比對繼而以 ZAF(Z: atomic number ; A: absorption ; F: fluorescence)程序
最後得到的訊號則由 RS-232 介面轉存入電腦中。 的訊號相比對繼而以 ZAF(Z: atomic number ; A: absorption ; F: fluorescence)程序