Synthesis of ordered ZnO nanorod film on ITO

Synthesis of ordered ZnO nanorod film on ITO substrate using hydrothermal method X. Tang*a, Z. Q. Maa,b, W. G. Zhaoa, D. M. Wanga,b a b

Department of Physics, Shanghai University, 200444, P.R.China

SHU-SOEN’s R&D Lab, Shanghai University, 200444, P.R.China

ABSTRACT By using low cost, low-temperature hydrothermal approach and spin-coating technique, well-aligned ZnO nanorods have been successfully prepared on ITO substrates. The ITO substrate was pre-patterned with ZnO particles as a seed layer for the subsequent nucleic hybridization. The intuitionistic crystallinity of oriented ZnO nanorod array was characterized by scanning electron microscopy (SEM) and X-ray diffraction (XRD), respectively. It is shown that the preferred orientation growth of the wurtzite structure along (002) plane is easily obtained with the approach. The size of the ZnO nanorods has been found to be dramatically dependent on the concentration of zinc content in the solution. Keywords: Hydrothermal deposition, Zinc oxide, nanorod

INTRODUCTION As a II-VI group compound semiconductor with a wide direct band gap (3.37 eV) and a large exciton binding energy (60 meV), ZnO has attracted much interest. ZnO nanomaterials have been widely studied for high-technology applications ranging from light-emitting diodes[1], photodetectors[2], gas sensors[3] and solar energy conversion[4,5]. Among all the nanostructures, ZnO nanorods are of particular interest due to their versatile applications for the fabrication of different types of nanodevices. Various chemical, electrochemical and physical deposition techniques have created structures of oriented ZnO nanorods so far, such as metal-organic chemical vapor deposition[6], pulse laser deposition[7], spray pyrolysis[8] and radiofrequency magnetron sputtering. However many of these methods require strict growth conditions and very expensive equipments. Due to its low cost and capability to coat large surface areas, hydrothermal method is selected for this study. In our work, a two-step process was used to grow ZnO nanorods on ITO glass substrate at 90℃. The resulting ZnO nanorods were characterized using scanning electron microscopy (SEM), X-ray diffraction (XRD) and step measurement.

EXPERIMENT The experiment was carried out mainly in two steps. The first step was to coat the ITO substrate with a ZnO seed layer using spin coating technique. The basic precursor solution was prepared by dissolving zinc acetate

dihydrate

(Zn(CH3COO)2 · 2H2O)

and

diethanolamine

(HN(CH2CH2OH)2,

DEA)

in

2-methoxyethanol. The molar ratio of Zn(CH3COO)2·2H2O and diethanolamine in the precursor was 1:1. -1-

Then the precursor was stirred by a magnetic stirrer at 60℃ for 1h to get a clear solution. The ITO substrate was then spin-coated with the precursor solution and dried on a hot plate at 300℃ for 30 minutes. The second step was to grow ZnO nanorods on the seeded substrate using hydrothermal method. The seeded substrate was suspended upside down in a temperature controlled chemical bath. The solution of the chemical bath consisted of equimolar amounts of analytic grade hexamethylenetetramine(C6H12N4, HMT) and analytic grade zinc nitrate(Zn(NO3)2·6H2O). The concentration range of the chemical bath was between 0.01 mol/L and 0.05 mol/L. The growth time was about 2 hours and the temperature of the chemical bath was 90℃. The solution was also stirred during the growth process in order to get a homogeneous chemical bath. When the growth of ZnO nanorods was finished, the sample was rinsed with deionized water repeatedly and dried in the air.

3.

RESULTS AND DISCUSSION

Fig. 1 shows spectrum for the typical XRD of the grown ZnO nanorods. The strong (002) diffraction peaks with high intensity in all the three curves show that the ZnO nanorods deposited on the substrate are highly crystalline and grow along the c-axis direction perpendicularly to the substrate in preference. The variation of the full width at half maximum (FWHM) of (002) peak with the concentration of the chemical bath is also shown in Fig.1. It is clearly noticed that the FWHM decreases with the increase of the concentration of the chemical bath. Since the crystallite size can be estimated from the FWHM of (002) diffraction peak by the Scherrer’s relation, we can see that the crystallite size increases when the concentration of the chemical bath changes from 0.01mol/L to 0.05 mol/L. This result also confirms good vertical aligned ZnO nanorod arrays as seen from SEM images.

Fig. 1. XRD images of ZnO nanorods deposited in chemical baths with different concentrations

-2-

Morphologies of the samples obtained in chemical baths with different concentrations were analyzed by SEM shown in Fig.2. The ZnO nanorods are uniformly distributed, mainly hexagonally shaped, and increase in diameter with increasing concentration.

Fig.2. SEM images of ZnO nanorods deposited in chemical baths with different concentrations

A step between the ZnO seed layer and the nanorods has been kept before the growth of the deposition of the nanorods. So the average length of the nanorods can be determined roughly by measuring the height of the step for each sample.

Table.1 Dependence of average length of ZnO nanorods on the variation of the concentration of different chemical baths:

concentration of chemical bath (mol/L)

average length of nanorods (nm)

0.01 mol/L

106.7

0.02 mol/L

218.1

0.03 mol/L

229.9

0.04 mol/L

431.4

0.05 mol/L

301.1

Table.1 shows the length of nanorods deposited in different chemical baths. The average length of nanorods increases with the changing concentration of the chemical bath. The growth mechanism of ZnO nanorods involves a complicated reaction process in the chemical bath. HMT, usually used in the fabrication of ZnO nanostructures, introduces the hydroxide ions (OH-) and ammonia molecules (NH3) to the solution. The detailed chemical equations of the process are given below:

(CH 2 )6 N 4 + 4 H 2O → (CH 2 )6 N 4 H 44+ + 4OH −

(CH 2 )6 N 4 + 6 H 2O → 6 HCHO + 4 NH 3 Zn 2+ + 4OH − → Zn(OH ) 24−

(1)

(2)

(3) -3-

Zn 2+ + 4 NH 3 → Zn( NH 3 ) 42+

(4)

Zn( NH 3 ) 24+ + 2OH − → ZnO + 4 NH 3 + H 2O

(5)

Zn(OH ) 24− → ZnO + H 2O + 2OH −

(6)

and ZnO is a polar crystal, whose polar axis is the c-axis. In the ZnO structure, each Zn2+ ion is surrounded by four O2- ions to form a tetrahedral coordination structure. The growth rate of crystal is related to the orientation of the coordination polyhedron at the interface[9]. Namely, the crystal face with the corner of the coordination polyhedron present at the interface has the fastest growth rate; the crystal face with the edge of the coordination polyhedron present at the interface has the second fastest growth rate; the crystal face with the face of the coordination polyhedron present at the interface has the slowest growth rate. And for wurtzite ZnO, [0001] direction has the fastest growth rate[10]. And it has been known that crystal faces with high growth rates easily disappear and faces with low growth rates are likely to remain. In a chemical bath with high concentration, the growth units almost have the same possibility to combine with one another to form grains, so the diameter of the grain is comparable to the length. But when the concentration is low, it is likely to form grains whose diameters are much shorter than the length. This explains the variation of morphologies of ZnO nanorods on the increase of solution concentration. The result is also consistent with the growth habit of ZnO nanorods observed in the hydrothermal experiment.

4. CONCLUSIONS ZnO nanorods have been successfully deposited on ITO substrate using a two step hydrothermal method. The nanorods are highly crystalline with a preferred orientation in the [0001] direction. The growth process proposed in the present work requires no expensive vacuum equipments and is therefore suitable for large scale of fabrication of aligned ZnO nanorods at a low cost.

REFERENCES 1. N. Saito, H. Haneda, T. Sekiguchi, N. Ohashi, I. Sakaguchi, Adv.Mater. 14 (2002) 418. 2. S. Liang, H. Sheng, Y. Liu, Z,Hio, Y. Lu, H. Shen, J. Cryst. Growth 225 (2001) 110. 3. N. Golege, S.A. Studenikin, M. Cocivera, J.Electrochem. Soc. 147 (2000) 1592). 4. M. Paulose, K. Shankar, O.K. Varghese, G.K. Mor, J. Phys.D: Appl. Phys. 39 (2006) 2498 5. M. Law, L.E. Greene, J.C. Johnson, R. Saykally, P.D. Yang, Nat. Mater. 4 (2005) 455 6. T. Gruber, C. Kirchner, K. Thonke, R. Sauer, A. Waag, Phys. Stat. Sol. A 192 (2002) 166 7. A.V. Singh, R.M. Mehra, A. Yoshida, A. Wakhara, J. Appl. Phys. 90 (2001) 5661 8. M.G. Ambia, M.N. Islam, M.O. Hakim, J. Mater. Sci. 29 (1994) 6575 9. W.J. Li,E.W. Shi , W.Z. Zhong, Z.W Yin, J Cryst Growth 203 (1999) 186 10. J.F Conley, L. Stecker, Y. Ono, Nanotechnology 16 (2005) 292

-4-

The first-principles study of Al adsorption on Si(001)2×1 C. B. Feng*a, Z. Q. Maa,b, F. Honga, Y. H. Lia a Department of Physics, Shanghai University, Shanghai 200444, P.R.China b SHU-SOEN’s R&D Lab., Shanghai University, Shanghai 200444, P.R.China

ABSTRACT First-principles calculations have been performed to study the adsorption of Aluminium(Al) on the Si(001)2×1 surface. The optimized geometries and electronic structures of the adsorption system were investigated. The adsorption energy at various adsorption coverages(Θ) from half a monolayer(ML) to one monolayer has been calculated. The most stable adsorption sites were consequently determined to be HH site and T3-T4 site, respectively. There is obvious evidence that the asymmetric aspect of the Si-Si dimer becomes the symmetric one which has been observed at the coverage of 0.5ML or 1ML. In addition, the bond length of Si-Si was found to be considerably elonged upon the Al adsorption. The work function calculations have shown that the aspect of work-function change for the Si(001) surface due to the adsorption of Al is different from that of the alkali metals adsorption reported in some previous works. The surface formation energy was also calculated. The absolute value of the surface formation energy was found to decrease with increasing coverage indicating that 1ML is not a saturation coverage. In order to shed light on the nature of the Al-Si bond and the character of silicon surface, the density of states and band structure of the system were calculated. Keywords: Density functional theory calculation, Silicon, Aluminium, Work function

INTRODUCTION In the recent decades, the metal/Silicon surface systems have gained much attention. Many experiments have been carried out to study these systems with various techniques of low-energy electron diffraction (LEED), high-energy electron diffraction (HEED), Auger electron spectroscopy (AES) and scanning electron microscopy (SEM). The group-Ⅲ metal/Si(100) systems[1,2] have been widely investigated. We expect that the reaction mechanisms of the two systems will help us in understanding the behavior of Ⅲ-Ⅴ compounds growing on the Si surface. Nogami et al[1]. have studied the growth of the first monolayer(ML) of Al on Si(100) with scanning tunneling microscopy(STM). They have pointed out there were actually two configurations in which the Al dimers were positioned either perpendicular or parallel to the underlying Si dimers. But they haven’t clarified the bonding site of the Al dimers on this surface. Brocks et al[2]. have shown that the adsorption of Al onto the Si(100) surface can be described by a reaction mechanism. They pointed out that the clean Si(100) surface was most stable consisting of “buckled” Si dimers. In contrast to this components, Chao et al[3]. have observed the Cs-saturated surface still has asymmetric dimers with method of photoelectron spectroscopy. Baski et al[4]. have studied Sn-induced reconstructions of the Si(100) surface using LEED and STM. In the experiment, they believed that the metal dimer bond could in fact be -5-

oriented parallel to the underlying Si dimer bonds. Umeno and Kitamura[5] examined Al/Si systems by ab initio molecular dynamics (MD) calculations. Their results suggested that the dense layers of Al could be formed on the surface. Up to now periodic density functional theory(DFT) have been widely and successfully applied to electronic structure, reasonable geometries and energies of metal adsorption on semiconductor surface. In our study, the motives are: (a) to determine the saturation coverage and adsorption site of Al on Si(100)2×1 surface, (b) to explore the nature of the Al-Si bond and the properties of silicon surface, and (c) to investigate Al-induced work-function change as a function of coverage and to study the mechanism of Al-induced work-function change.

METHODOLOGY Ab initio total-energy calculations within the DFT frame-work has been taken to study the adsorption of Al on Si(001)2×1. The interaction between ions and electrons is described using the projector augmented wave method[6]. The general gradient approximation functional proposed by Perdew and Wang known as PW91[7,8] is used. The influence of

Fig. 2. Top view of the Si(001)2×1 asymmetric surface. The 2×1 unit cell is indicated by dashed lines. The HH, HB, T3, T4 sites are indicated. different K-points and plane-wave cut off energy was tested by a series of calculations, which leads to our calculations with 8×8×1 K-points sampling and a cut off energy of 350eV. We used a repeated slab model, in which one slab includes five Si atomic layers and a vacuum region of 10Å. The bottom of the slab has been saturated by two H atoms per Si atom. The adsorbates and the atoms in the top three layers were allowed only to relax. For bulk silicon, the lattice constant was determined to be 5.471Å, which is in excellent agreement with the experimental value of 5.431Å[9]. In our work, four different adsorption sites have been considered, as shown in Fig.1. Four high symmetry adsorption sites are illustrated: pedestal(HH), valley bridge(T3, on top of a third layer Si atom), bridge(HB) sites, and cave(T4, on top of a fourth layer Si atom).

DATAS AND RESULTS 1.1 The adsorption energy and the formation energy

The adsorption energy of Al at different adsorption sites have been calculated, which is defined by the formula

-6-

Eads=  ESi (001) + NE Al − E Al / Si (001)  / N

(1)

where ESi (001) and E Al / Si (001) are the total energy of the Si(001) system with or without adsorbed atoms,

E Al is the total energy of an isolated Al atom, and N is the number of the Al atoms in the surface unit cell. As can be seen from Table.1, the most stable adsorption sites are HH at 0.5ML which was found to be different from some previous results[10]. At the coverage of 1ML, the configuration of T3-T4 was found to be energetically favored. In order to determine the saturation coverage of Al adsorption on Si(001)2×1 surface, we have calculated the surface formation energy which is defined as Eform= ESi + Al − mµ Si − nµ Al

(2)

Where µ Si and µ Al are the total energy of bulk Si and Al. The different surface formation energies of 0.5 and 1ML models are compared in Table.1. We note that the formation energies for the 0.5ML coverage are generally lower than those of 1ML coverage. T3 site adsorption as found to have the lowest surface formation energy, suggesting that the saturation coverage should be 0.5ML coverage rather than 1ML. This is in contrast with the experimental and theoretical investigations[11] for Na and K adsorptions which gave supports to the saturation coverage of 1ML. Table. 1. The adsorption energy and the formation energy of the system at Θ=0.5ML and Θ=1ML. Coverag

Θ=0.5ML

e Site

HH

Eads(eV/

-3.24

at.)

1

Eform(ev)

-13.2 55

Θ=1ML

HB

T3

T4

HH-HB

HH-T3

T3-T4

HB-T4

HH-T4

-3.007

-2.773

-3.153

-3.063

-3.132

-3.236

-2.843

-3.055

-13.021

-12.787

-13.167

-16.141

-16.277

-16.486

-15.701

-16.125

HBT3 -3.04 6 -16.1 06

1.2 The structure analysis

The LEED structure analysis of Over et al[12]. has showed that the reconstruction of the clean Si(001)2×1 surface consists of asymmetric and buckled Si dimers. In our work, the structure of the clean Si(001)2×1 surface with asymmetric dimers was optimized. The Si dimer bond length was calculated to be 2.311Å (it was elonged by 0.22Å and 0.14Å at Θ=0.5ML and Θ=1ML), in excellent agreement with the experimental value of 2.24±0.08Å[12]. This value is shown to be smaller than the Si-Si distance in the bulk (2.369Å). Kobayashi et al[11]. have obtained a dimer bond length of 2.26Å from DFT-LDA calculations. The dimers were found to be tilted by about 18.257° and the vertical separation between up and down atoms in the

-7-

dimers is 0.724Å. The corresponding values by Over et al[12]. are 19°±2° and 0.72±0.05Å. This dimerization results in small lateral and vertical displacements (0.066Å) of silicon atoms in the second layer. Subsequent calculations will show that the asymmetry in the Si-Si dimer is considerably suppressed (the vertical separation between up and down atoms in the dimers is 0.016 Å and 0.193 Å at Θ=0.5ML and Θ=1ML) when Al atoms are adsorbed on the Si(001)2×1 surface with higher coverage (Θ≥0.5ML). 1.3 The DOS analysis

The DOS curves for the adsorption system at silicon are shown in Fig.3. For the clean Si(001)2×1 surface the Fermi level is near the valence bands and the band gap is small, resulting in the surface with semiconducting properties. Several experimental and theoretical investigations have showed that the Si(001) surfaces with a symmetric dimer reconstruction are metallic and the surfaces with an asymmetric dimer are semiconducting[13,14]. For Al adsorption at HH site at Θ=0.5ML, it can be seen that the silicon surface is also semiconducting obviously. However, there is a wide band gap. For Al adsorption at T3+T4 sites at Θ=1ML, it is interesting to find that the DOS curves at the Si atoms are considerably different with both the above situations. The Fermi energy of this system is in a regime with a significant contribution from valence

Work function change(ev)

bands. So the adsorption system at Θ=1ML is metallic.

0.3 0.2 0.1 0.0 -0.1 -0.2 -0.3

Fig.2

0.0

0.5 Coverge(ML)

1.0

20 15

(c)1ML

Fig.3

Density of states(Arb.units)

10 5 20 0 15

(b)0.5ML

10 5 20 0 15

(a)Clean

10 5 0 -10

-5

Energy(eV)

0

5

1.4

Fig. 2. the coverage dependence of the work function change due to Al adsorption. Fig. 3. the DOS of the Al/Si(001)2×1 adsorption system at the coverage of 0ML, 0.5ML and 1ML.

1.5 The work function change

The work function change Δφ, due to Al adsorption has been investigated, which is calculated with the formula Δφ= φ Al − substrate − φsubstrate ,

(3)

Where φ Al − substrate and φsubstrate is the work function value of the system with and without Al adsorbate, respectively. The work function φ is obtained from the formula, -8-

φ= Evac − EF ,

(4)

where Evac and EF represents the energy of vacuum level and the Fermi level, respectively. The calculated work function of the clean Si(001) surface is φ=4.604eV, agreeing well with the experimental value of 4.8eV[15]. In our work, the work function change at the coverage of 0.5ML was found to be 0.259eV. As the coverage was increased to 1ML, the work function change was found to be considerably decreased, Δφ=-0.287eV. This finding is different from the system of Alkali mental/Silicon surface [11,15,16]. The coverage dependence of work function change due to Al adsorption has been illustrated in Fig.2. It is shown that the work function was increased sharply at Θ=0.5ML and considerably decreased at Θ=1ML. The considerable increase of the work function at Θ=0.5ML indicates charge regain of Al upon adsorption on Si(001) surface. The sharp decrease of the work function at Θ=1ML indicates that the adsorbed Al atoms, whose dipoles are not oriented perpendicular to the surface, are strongly depolarized. For analyzing the surface dipole moments (μ), we have calculated this quantity using the Helmholtz equation (in Debye units)[17]: μ= A∆φ /12πΘ

(5)

where A is the area in Å2 per (2×1) surface unit cell, andΔ φ is the work-function change in eV. As the coverage changes from 0.5 to 1ML, the dipole moment per adatom changes from 0.411D to -0.228D, respectively. These results indicating that the dipole moment for the higher coverage was significantly reduced. Thus dipole-dipole repulsion is large and the depolarization effect is significant. The depolarization effect leads to some electronic charge flows back from the substrate to the Al atoms, in agreeing well with what has been concluded from the work function analysis. The regain of electronic charge also results in an increase of the adsorption energy and weaker adsorbate-substrate interaction.

CONCLUSIONS In our work, we used first principle calculation to study the adsorption of Al on Si(001)2×1. The results indicate that the dimer was elonged and became symmetric as the coverage increasing. The surface of the system became metallic properties at Θ=1ML. The work function change suggested that the adsorbed Al atoms were strongly depolarized at Θ=1ML. This finding was also supported by analyzing the surface dipole moments calculated by the Helmholtz equation.

ACKNOWLEDGEMENT This work is supported by "SECE-Institute: Shanghai High Institutions Grid" project. The authors thank them f-or computering time.

REFERENCES 1. J. Nogami A. A. Baski and C. F. Quate Phys. Rev. B 44 1415-1418(1991). 2. G. Brocks, P. J. Kelly and R. Car Phys. Rev. Lett. 70, 2786-2789(1993). 3. Y. C. Chao L. S. O. Johansson and R. I. G. Uhrberg Phys. Rev. B 54, 5901-5907(1996). -9-

4. A. A. Baski, C. F. Quate, J. Nogami, Phys. Rev. B 44, 11167-11177(1991). 5. Y. Umeno and T. Kitamura, Modelling Simul. Mater. Sci. Eng. 12, 1147-1157(2004). 6. G. Kresse and J. Joubert, Phys. Rev. B 59, 1758-1775(1999). 7. J. P. Perdew, J. P. Chevary, S. H. Vosko, K. A. Jackson, M. R. Pederson, D. J. Singh, and C. Fiolhais, Phys. Rev. B 46, 6671-6687(1992). 8. J. A. White and D. M. Bird, Phys. Rev. B 50, 4954-4957(1994). 9. W. L. Bond and W. Kaiser, J. Phys. Chem. Solids 16, 44-45(1960). 10. J. E. Northrup Phys. Rev. Lett. 57, 154-157(1986). 11. K. Kobayashi, Y. Morikawa, K. Terakura, and S. Blügel, Phys. Rev. B 45, 3469-3484(1991). 12. H. Over, J. Wasserfall, W. Ranke, C. Ambiatello, R. Sawitzki, D. Wolf, and W. Moritz, Phys. Rev. B 55, 4731-4736(1997). 13. E. Landmark, C. J. Karlsson, Y. C. Chao, and R. I. G. Uhrberg, Phys. Rev. Lett. 69, 1588-1591(1992). 14. P. Krüger and J. Pollmann, Phys. Rev. Lett. 74, 1155-1158(1995). 15. Y. Enta, T. Kinoshita, S. Suzuki, and S. Kono, Phys. Rev. B 36, 9801-9804(1987). 16. Y. C. Chao, L. S. O. Johansson, R. I. G. Uhrberg, Phys. Rev. B 54, 5901-5907(1996). 17. A. Soon, M. Todorova, B. Delley, C. Stampfl, Phys. Rev. B 73, 165424(2006).

- 10 -

The interface recombination current of the CdS/CdTe heterojunction solar cell L.Zhang, W.Wu, M.Li, Z.X.Zhao, Y.H. Zhang, Z.Q.Ma. SHU-SOLARE’s R&D LAB, Dept.of physics, shanghai University Shanghai 200444 P.R. China ABSTRACT

The recombination current and its relation with voltage and density of interface states of the CdS/CdTe solar cell are discussed. Interface state is the result of lattice mismatch of the two materials used. Here, we regard interface states as impurity lever. In order to find out its influence to the efficiency of solar cell, we use S-R-H model to calculate interface recombination current of the cell on illumination conditions. At the same time, interface recombination current density is compared with photo-generated current density and we find that it is the main loss for thin cell. Also, we discuss relation between interface recombination current and capture cross-section and other factors; we find that interface recombination current is significant when the photo-voltage changes from 0.5V to 0.7V. If photo-voltage is less than 0.5V, interface recombination current could be neglected. At last, we analyze the way to reduce interface recombination current.

Keywords: Recombination Current, CdS/CdTe solar cell

1.

INTRODUCTION

At present, single-crystal silicon is the most popular material in producing solar cells. But silicon is an indirect-gap semiconductor; its band gap is 1.12ev. Its co efficiency of light absorption is very low, not an ideal material for solar cell. The theoretic efficiency is just about 25%; the highest efficiency in laboratory has risen to 24.7%. Due to high cost, it still cannot be used domestically. According to investigation, it is uncompetitive to use silicon comparing to general energy sources. So it is necessary to find new material for solar cell in order to produce high efficiency solar call with low cost. In 60s, with the development of microelectronics, people were interested in heterojunction and used it in producing solar cells. Among these materials, CdTe is an important film material, it has some advantages: (1) an ideal band gap for solar terrestrial photo conversion (1.45 eV), (2) direct-gap material with very high co efficiency of light absorption, it only needs a few microns to absorb 90% of the total incident light. For film solar cell, it has some advantages: higher absorbing efficiency, much thinner, but it still has some shortage comparing to silicon——the interface recombination. For solar cells with only a few microns, the bulk recombination is negligible, so interface recombination comes to front. In this work, interface - 11 -

recombination current is calculated; the range of the current and the influence to the efficiency is also discussed.

2. THE I-V RELATION

Fig.1 Schematic of window-absorber p-n heterojunction solar cell

Here, we use CdS/CdTe model [1], supposing CdS/CdTe heterojunction as abrupt junction, and neglect the photon-generated carriers from the window. The doping is ND=1017, NA=5×1015. Use the widely used standard solar radiation AM1.5, the power density W0=100mw/cm2, the number of photon meet the condition can be calculated by numerical approach: Nλ=1.61057×1017/cm2.s Nλ is the number of photons incident to the unit area of solar cell in unit time. The photo-generated current can be decompounded to 3 parts: the current from absorber and two space-charge regions. J=Jn+J1+J2 The width of two space-charge regions:

2 N DVDε 0 ε 1ε 2 qN A ( N Aε 0ε 1 + N Dε 0ε 2 )

(1)

2 N AVD ε 0 ε 1ε 2 W2= qN D ( N Aε 0ε 1 + N D ε 0ε 2 )

(2)

2

W1=

2

Suppose quantum efficiency as 1, it means that each photon can generate one electron and one hole, the current density of two space-charge regions can be given by: W1

(

J1= ∫ qN λ α 1e − d 2α 2 e −α1x dx = qN λ e − d 2α 2 1 − e −W1α1 0

W2

(

)

(3)

)

J2= ∫ qN λ α 2 e −( d 2 −W2 )α 2 e −α 2 x dx = qN λ e − d 2α 2 eW2α 2 − 1 0

(4)

If the absorber can be seen as a material with long lifetime for minority carries. So, there is no drift field, consequently, n=n(x) must satisfy:

n − n p0 d 2n g (x ) − + =0 2 2 Dn1 dx Ln

(5)

The boundary condition should be written as: - 12 -

n(x1)=np0 ;

Dn1

d (n − n p 0 ) x = d1

dx

= S n (n p − n p 0 )

(6)

(Sn is the recombination velocity at the back surface of solar cell, suppose Sn=0) The current density of absorber can be given by:

J n = jL − jdif (e jL =

jdif =

Sech( M= Lnα1 −

qDn n p 0 Ln

− W1 + d1 )+ Ln

qv kt

− 1)

(7)

e −W1α1 − d 2α 2 qLn N λ α 1 M Ln α 1 − 1 2

2

S n Ln d d Coth( 1 ) + Sech( 1 ) D Ln Ln [ n ] S n Ln d1 d1 Sech( ) + Coth( ) Dn Ln Ln

[−e −( −W1 + d1 )α1 + Coth(

− W1 + d1 )]Ln S n Ln

Dn

Coth(

− W1 + d1 )+ Ln

Sech(

+ e −( −W1 + d1 )α1 Lnα1

− W1 + d1 ) Ln S n Ln Dn

The maximal theoretic current density is:

J sc = J 1 + J 2 + J L = 24.159mA

3.

THE CURRENT OF SOLAR CELL CONSIDERING INTERFA CERECOMBINATION

Interface state is the basic character of heterojunction; it is the result of lattice mismatch of the two materials used in solar cell. Here, we calculate interface recombination current with S-R-H recombination theory. There are two kinds of recombination current in the interface states: jr1, electrons from window are recombined with holes from absorber. According to S-R-H, assuming at interface there is a single combination level Et which lies in the middle of band gap, the recombination current could be written as [2]:

j r1 =

jr 2 =

qrn1 rp1 N t1 nn 0 e

−

qvd 1 kt

−

Pp 0 e

qv d 2 kt

rp1 p p (W1 ) + rp1 p11 + rn1 nn (W1 ) + rn1 n11

qrn 2 rp 2 N t 2 n p 0 Pp 0 rp 2 p p (W1 ) + rp 2 p12 + rn 2 n p (W1 ) + rn 2 n12

(8)

(9)

- 13 -

n1 = n0

Pt 0 ， nt 0

nt 0 pt 0

p1 = p 0

p p (W1 ) = N A e n p (W1 ) = n p 0 e nn (W1 ) = nn 0 e

−

−

qVd 2 kt

qVd 2 kt

qVd 1 kt

e

e

−

−

e

qV2 kt

qV2 kt

qV2 kt

qV

e kt qV

e kt

(10)

E Fn − E Fp = qV (Suppose that EFp is flat across the barrier region in the absorber). V2 =

V Vd 2 V V (Suppose that 1 = d 1 ) Vd 1 + Vd 2 V2 Vd 2

(11)

rn, rp indicates capture co efficiency of electron and hole respectively, Pt is the level density at recombination center not captured by electrons, nt is the level density at recombination center captured by electrons, Nt is the density of recombination center. The number 1，2indicates absorber and window respectively. Substituting Eq. (10) (11) into Eq. (8), (9)

e

j r1 =

−

G 2 kt

qυΩ1 N c1 N t1

ep

2+e

qV qV G − + 2 − d2 kt 2 kt 2 kt kt

+e

−

G qV e p qV2 qVd 2 + + − + kt 2 kt kt 2 kt kt

qυΩ 2 nip N t 2

jr 2 = 2+e

E g 2 e p qV2 qVd 2 − + − 2 kt kt kt kt

+e

−

E g 2 qV e p qV2 qVd 2 + + − + 2 kt kt kt kt kt

(12)

(13)

Eq. (13) uses an effective gap G as:

G = EC1 (W ) − EV 2 (W ) = EC1 − EV 2 = E g 2 − (χ1 − χ 2 )

υ is thermal velocity for electron, Ω is capture cross section for electrons and holes, because what determines the value of jr1 is the height of barrier at CdTe side, assuming capture cross section for holes in CdTe Ω1=10-16[3], Ω2=10-18(the main influence to jr2 is the density of electrons in absorber CdTe.)

Fig.2 Recombination current density – photo voltage curves：

- 14 -

From the calculated result and Fig.2, jr1 is much greater than jr2, and jr2 is almost zero in the working area of solar cell. So we can neglect jr2 when we considering interface combination current. The total current density can be given as:

 qVkt    qVkt   J = (J 1 + J 2 + J L ) − jdif  e − 1 − ( jr1 + jr 2 ) e − 1     From calculation, the light induced current is equal to interface recombination current when photo voltage is 0.603V. In Fig.4, interface recombination current changes with capture cross section. The I-V curve moves to right when the capture crosses section decrease. From (12), (13), we can know that the interface recombination current changes with Nt, Nc1 like Fig.4.

Fig.3. Comparison between recombination current density and photocurrent density

Fig.4. Recombination current density as a function of voltage and capture cross section

4.

CONCLUTION

The interface recombination current of the CdS/CdTe heterojunction solar cell is calculated theoretically, and its relation with photo voltage is given too, point out that recombination current comes main from jr1:electrons from window are recombined with holes from absorber; jr2: electrons from absorber are recombined with holes from absorber, getting the conclusion that the main infection to interface recombination current is capture cross section, the density of recombination center and the effective density - 15 -

of states for the conduction band in the window, providing theoretic basis and derection to minimize the influence of interface recombination to efficiency of solar cell. 5. REFERENCES

1. Peter Nollet, Marc Kontges, M.Burgelman, S.Degrave, Rolf Reineke-Koch; Indications for presence and importance of interface states in CdTe/CdS solar cells; Thin Solid Films 431 –432 (2003) 414–420. 2. Stephen J. Fonash; Solar cell device physics;Academic Press,1981 3. M. Burgelman, P. Nollet, S. Degrave; Modelling polycrystalline semiconductor solar cells; Thin Solid Films 361±362 (2000) 527-532

- 16 -

Parameter optimization of Silicon Solar Cell Back Surface Field (BSF) Formation Li Mina, Wu Weia,Z.X.Zhaoa,Zhang Leia, Z.Q.Maa,b,Q.F.Sub, Y.H.Zhangb,a, a) Department of Physics,Shanghai University,200444,P.R.China B) SHU-SOEN’s R&D Lab.,Shanghai University,200444, P.R.China ABSTRACT: The BSF of mono-crystal silicon cell is studied in this paper . It simply states the principle and formation process of BSF, and studies the evenness of BSF. In the experiment, we analyze BSF formed by taking different sintering temperature , Heating rate and setting different time, develop by chemical reagent and observe at high magnification scope, analyze and get the optimal parameters of forming even BSF and further improve conversion efficiency of silicon chips. Key words: silicon solar cell, even BSF 1. INTRODUCTION In manufacturing of silicon cell, enforcing aluminium BSF on it can further improve conversion efficiency of the cell[1,2]. In silicon solar cell BSF of normal n+/p structure, we can add a stratum of p+ to silicon solar cell BSF in order to form BSF, and the structure as follow: - V + E2

E1 n+

0

p+

p

Xj

L-X

L

Fig.1 Structure of back surface Cell In the structure of n+/p/p+, as shown in Fig.1, potential direction of p/p+ junction is opposite to the direction of cell end voltage, which decreases the Positive Bias endured by the Cell. Moreover, the p/p+ junction holds back minority photon-generated diffuse to p+ zone, as a result，improves the collection efficiency of minority carriers and lowers dark current component. Majority carriers can pass through freely and have no impact on output voltage. Given no consideration of the impact of p zone, open circuit voltage of BSF cell can approximately expressed as follows: Voc ≅

(L-thickness of cell chip X-deepness of BSF

J kT ln  SC2 q  qni

∫

L

L− X

N P+ Dn

 dx  

Dn-diffusion coefficient of electrons)

From Fig.1, we can find that BSF has great help to the improvement of open circuit voltage. Since the absorption peak value of silicon solar cell is about 900nm, long waves penetrate deeply and generate lots of - 17 -

photon-generated carriers as far as near by back surface electrode. The interface of back surface electrode is like a big combining center, but since there exists aluminium BSF that can effectivelyreflect long wave light, it can reduce the recombination of aluminium back surface electrode. We can know from the analysis above that for aluminium BSF of a certain junction depth，it connects with the thickness of aluminium depositing on silicon chip and sintering temperature, the higher the temperature and the bigger the thickness, the more of junction depth of aluminium BSF. The formula of junction depth is[3]： WBSF =

t ⋅ ρ Al  F (T ) F (T0 )  −   ρ Si  1 − F (T ) 1 − F (T0 ) 

(t--thickness of aluminium depositing on silicon chip initially ρAl，ρsi --density of aluminium and silicon F（T）--silicon atom quality percentage at peak value temperature when aluminium and silicon are melted F（T0）-- silicon atom quality percentage at co-melt temperature（577℃）, it is about 12.2%). In practical solar cell, besides to promote to collect minority carriers, BSF can also be taken as electrode output end very well. Aluminium-silicon alloy structure can not only play a role of BSF, but also form ohmic contact very well and be output electrode of current. 1. EXPERIMENT The experiment chooses a silicon solar cell wafer. It is 125×125 mm and its (P-Type) crystal direction is 100, resistivity 1 Ωcm, thickn ess 230 μm and texturing made in NaOH. The PN junction is made according to the method of POCl3 liquid source diffusion, with Plasma etching edges. (Aluminum plasm:ferro53102,Angentine plasm:ferro33462) We take different sintering temperature, observe its conversion efficiency and change of minority carrier lifetime and also look into the formed of BSF, change the curves of sintering temperature in BSF and observe BSF and its evenness, at last we get the optimal sintering temperature curve. 2. RESULTS AND DISCUSSION 2.1 Testing of electric property The experiment takes 4 kinds of sintering temperature: 765℃，770℃，790℃ and 830℃(the highest sintering temperature), we take 5 chips in each temperature, with minority carrier lifetime tester testing lifetime of minority carriers, with xenon light solar energy simulator testing its change of conversion efficiency. The results are shown in Fig.2 and Fig.3. In Fig.2, the three pillars represent minimum, average and maximum lifetime of minority carriers of 5 cell chips. At 790℃, the minority carrier lifetime achieves a relatively ideal value. When the temperature arrive to 830℃, the lifetime descends and the back of the cells are born black, there may be some impurity （Fe）separating out because of too higher temperature. Since silicon and aluminium have different segregation coefficients, the surfaces separate out some iron oxide. At the four kinds of sintering temperature,

- 18 -

the circuit voltage of the chips is all ideal and about 0.6, which means that BSF plays an obvious role in improving open circuit voltage of the cells. In Fig.3, the conversion efficiency of the cells arrives at the maximum at 790℃.

6

τ/μs

5

η

4 3 2 1 0

765

770

790

830

T/℃

Fig.2 Relationship ofminority carrier lifetime and sintering temperature

15.5 15 14.5 14 13.5 13 12.5 12 11.5

765

770

790

830

T/℃

Fig.3 Relationship of conversion efficiency and sintering temperature

2.2 Observation of BSF Put the chips into the solution of HF:HNO3:CH3COOH=1:3:6, and then we can depict the structure of BSF. After erosion, observing by upright metallographical microscope XSP-2C, we can see as Fig.4 and Fig.5. Si

Al-Si

Al Fig.4 Demonstration of BSF formation

Al bulk

Al-Si layer

Si bulk Fig.5 Demonstration of BSF

formation at 790℃

at 830℃

From Fig.4 and Fig.5, we can find that the chips are even and level, and its particles are small and compact，contact well with silicon chips, which means 790℃ is proper to form BSF, but at 830℃ they can hardly form BSF. Here we give some different sintering temperature curves to be a technical reference later on.

- 19 -

Fig.6 sintering curves of different temperature In order to study electric property of chips, we take 10 chips as samples and test their electric parameters at the sintering temperature of 790℃ as follows:

Chart 1 Electric property of monocrystal silicon Cell at the sintering temperature of 790℃

No.

Isc(A)

Voc (V)

FF

η

No.

Isc(A)

Voc (V)

FF

η

1

4.91

0.6

0.78

15.47

6

4.84

0.6

0.77

14.94

2

4.96

0.6

0.77

15.31

7

4.96

0.6

0.76

15.02

3

4.94

0.6

0.77

15.28

8

4.96

0.6

0.73

14.58

4

4.9

0.6

0.77

15.26

9

4.95

0.59

0.73

14.62

5

4.86

0.6

0.76

15.02

10

4.96

0.6

0.77

15.36

3. SUMMARY The article studies BSF of singer crystal silicon solar cell. The experiment applies silicon chips of P-TYPE and CZ/100/220µm, takes different sintering temperature, sets different sintering temperature curves, and tests minority carrier lifetime and efficiency of the cells. And we also observe the BSF formed under different sintering temperature curves by relative instruments, and get optimal sintering temperature curve on even BSF, which will improve conversion efficiency of silicon chips further. REFERENCE [1] A.Kaminski,B.Vandelle,A.Fave,J.P.Boyeaux: Aluminium BSF in silicon solar cells: Solar Energy Materials & Solar Cells 72 (2002) 373-379. [2] S.Narasimha and A.Rohatgi: Optimized Aluminum Back Surface Field Techniques for Silicon Solar Cells: IEEE,26th PVSC:Sept.30-Oct.3,1997. [3] Shreesh Narasimha,Ajeet Rohatgi: An Optimized Rapid Aluminum Back Surface Field Technique for Silicon Solar Cell: IEEE TRANSACTIONS ON ELECTRON DEVICES,VOL.46,NO.7,JULY 1999

- 20 -

Study of the Appropriate Time for Silicon Surface Texturing X.J. Meng, Jack Gillett Abstract We studied the relationship between power output and the texturing time. We put silicon wafers through the ordinary industry process, but we varied the texturing time. As the time increases, the power output of the wafers at first falls, then rises again. In order to obtain the best power output, we need to put the wafers in the NaOH(1.2%) texturing solution for about 40 minutes.

1. Introduction For monocrystalline silicon, we often select dilute solution of alkali to form surface texture (the alkali attacks the (111) surface which forms small surface pyramids on monocrystalline (100) surfaces). We aim to create surface features on the length scale of the incoming light, which increase the absorption by the surface. About 30% of the incident light will be reflected by bare silicon. Good texturing can reduce this to nearly 11%. This will dramatically affect the ultimate power output of solar cells. So we studied the proper time for surface texturing.

2. Experiment We selected thirty solar grade silicon wafers purchased from Hebei Huaer Bandaoti Cailiao Co., LTD as starting materials. We made the wafers go though the ordinary manufacturing process and changed only the texturing time. The wafers were divided into 6 groups with 5 each. The thickness of every wafer was measured. Each group of wafers was textured in 10, 20, 30, 40, 50, 60, 70, 75 minutes separately. We used 160 litres of ultrapure water and 20 litres of alcohol to make the NaOH solution with 1.2% concentration. The temperature of texturing is 88℃. The temperature of texturing is 88°. We measured the thickness of each wafer after texturing to study the thickness dependence. Finally, we measured the power output of each wafer (one of the thirty was broken).

The power output and other parameters of every group is calculated in table 1, the average power of these wafers is 2.332 watts. We can see 40 minutes gives the best light trapping effect. The power output of this group is 2.3526 watts. We can see the fill factor (FF) , working current(Iwork),

Thickness(um)

3. Discussion 230 220 210 200 190 20

30

40

50

60

70

80

Time(min) Original Thickness

Thickness After Texturing

efficiency(η) have the similar trend with the power, they grow as the time grows, until about 40 minutes, when they begin to fall.

Fig.1. The relationship between wafer thickness and texturing time - 21 -

Time(min)

Power(Watt)

FF

Iwork(A)

η

30

2.33175

0.733

4.474

15.69

40

2.3526

0.74

4.5116

15.832

50

2.3338

0.7382

4.4716

15.704

60

2.3084

0.7322

4.422

15.534

70

2.3168

0.7318

4.4408

15.592

75

2.3492

0.7378

4.5138

15.81

Table.1

Average parameters of the sample solar cells

We took an SEM image of an ordinary silicon wafer after texturing, shown in fig.2. It is easy to imagine that the reflectance of area A is smaller than area B. So if we want to generate more power, we need to optimize the texturing condition to form more areas like B. From a clearer image in fig.2.2, we can see the static growing process of the pyramids at area C. So, we can say the first generation of pyramids has already grown at 40 minutes, if texturised too long, the first generation of pyramids will be destroyed and they will experience a new growing process. After 75 minutes, the second generation of pyramids was formed. However, this is too long for the industry process given the minor power increase observed.

2.1

A

2.2 C

B

Fig.2. SEM image of normally texturised silicon surfaces

4. Conclusion We produced several solar cells and aimed to maximize the power output. When only considering the proper time of texturing, we recommend about 40 minutes is best. In forty minutes, we already see surface pyramid formation, which leads to a better absorption of incident light.

- 22 -

Multi-layer antireflection coatings for silicon solar cells using a sol-gel technique B. B. Shi*a, Z. Q. Maa,b, X. Tanga and C. B. Fenga a

Department of Physics, Shanghai University, 200444, P.R.China

b

SHU-SOEN”s R&D Lab., Shanghai University, 200444, P.R.China

ABSTRACT In order to improve the opto-electronic conversion efficiency of solar cells, antireflection coatings (ARCs) have been drew a great attention for the application in the terrene. Generally, the coating is obtained by vacuum processing such as thermal evaporation, reactive sputtering and plasma-enhanced chemical vapor deposition (PECVD). In this work, multi-layer antireflection coatings have been performed by a modified sol-gel technique, which is low-cost and simple. The multi-layer films consisted of SiO2 and TiO2. The physical phase and morphology of each layer were characterized by atomic force microscopy (AFM). The TiO2 single-layer and SiO2/TiO2 double-layer antireflection coatings were respectively annealed at 150 ℃,350℃ and 550℃.T he sols of TiO2 and SiO2 were aged for 24 hours and then were spin coated on the Si substrate. It was found that the reflectance of double-layer ARCs was generally lower than that of single-layer ones. The reflectance of films without being aged was lower than that of sols were aged for 24 hours. In all samples, the SiO2/TiO2 double-layer film which was annealed at 150℃ and w hich sols w ere not aged for 24 hours had the low est reflectance. Keywords: antireflection coatings; sol-gel; anneal; aging time; solar cell

1. INTRODUCTION Antireflection coatings (ARCs) are widely used in various applications such as display panels, solar cells and opticals[1-3]. A thin, dielectric film or several such films were applied to an optical surface to reduce its reflectance and thereby increase its transmittance. A sprayable TiO2 AR coating was developed by Tracy et al[4]. A double-layer interference AR coating was developed by Yoldas [5]. A single-layer ARC can reduce reflectance just only at one wavelength, usually at the middle of the visible region (550nm). Multiple layers are more effective over the entire visible spectrum (380nm – 780nm). Several transparent and high refractive-index material films have already been applied to ARC techniques, for example, SiO(n=1.8-1.9), SiO2(n=1.44), Si3N4(n=1.9), TiO2(n=2.3), Al2O3(n=1.86), Ta2O5(n=2.26), SiO2-TiO2(n=1.8-1.96), and ZnS[6-10]. Silicon solar cells have a high refractive index which leads to a solar-averaged reflectance of about 36%. This large reflection loss can be significantly reduced by coating the silicon with an AR coating. So for mass production of mono- and multi-crystalline Si solar cells, antireflection coatings (ARCs) have become one of the key issues. The sol-gel technique offers a simple and low-cost process to prepare the high quality thin films[11,12]. Kern and Tracy have observed an increase of 44% in the cell efficiency after spraying TiO2 - 23 -

single-layer ARC[13]. Green et al. have used MgF2/ZnS double-layer ARCs on Si cells with 19.1% efficiency[14]. In this work, multi-layer antireflection coatings have been performed by a modified sol-gel technique. The TiO2 single-layer and SiO2/TiO2 double-layer films are annealed at different temperatures. The sols of TiO2 and TiO2/SiO2 are aged for 24 hours respectively and then are spin coated on the Si substrates. Details of the fabrication process and efficiency change will be described as follows.

2.

EXPERIMENT

Tetrabuty titanate / tetraethyl orthosilicate , anhydrous ethanol and hydrochloric acid were used as precursor, solvent and catalyst respectively. The chart of the ARC synthesis process was performed using a sol-gel process as shown in Fig.1. H2O+HCl 1mol

C2H5OH 60-80mol

Mixing

Ti(OC4H7)4

1mol

Si(OC2H5)4 1mol

TiO2, SiO2 transparent sols

Spin-coated (TiO2 sol)

Heated at 90℃ for 15 min Spin-coated (SiO2 sol)

Annealed at 150/350/550℃ for 1hr

Heated at 90℃ for 15 min

Annealed at 150/350/550℃ for 1hr Single-layer (TiO2) ARC Double-layer (SiO2/TiO2) ARCs

Fig.1. Flow chart of sol-gel process for TiO2 single-layer SiO2/TiO2 double-layer ARCs The SiO2 and TiO2 coatings were spin coated on the Si substrate from the above-said solution at 3000-5000 r.m.p for 30s. Each film was heated at 90 ℃ for 15m in atfirstand then annealed at150℃,350℃ and 550℃ respectively. T he refractive index and extinction coefficient w ere measured by ellipsometry spectrometer. The reflectance was measured by UV-Vis recording spectrophotometer. The physical phase and morphology of each layer were characterized by atomic force microscopy (AFM).

3.

RESULTS AND DISCUSSION

Fig.2 shows the absorbance spectra and refractive indices of the deposited films as a function of - 24 -

wavelength. The refractive index of the SiO2 and TiO2 film were determined to 1.48 and 1.71. The extinction coefficients (k) of each film after the annealing process almost approached to zero between 300 and 900 nm. These indicate that the sol-gel deposited SiO2 and TiO2 films can be used as ARCs for Si solar cells.

0.10

2.6

0.08 0.06 TiO2 (n)

2.2

Extinction coeffcient (K)

Refractive index (n)

2.4

0.04

TiO2 (k)

0.02 0.00

2.0

-0.02

SiO2 (K)

1.8

-0.04

SiO2 (n)

1.6

-0.06 -0.08

1.4

-0.10 300

400

500

600

700

800

900

wavelength (nm)

Fig.2. Variations of refractive index (n) and extinction coefficients (k) of SiO2 and TiO2 films 30

点线

28

30

150℃

28

虚线

350℃

22 20

实线

18

550℃

16

150℃

24

点线

350℃

虚线

550℃

22 20 18 16 14

14

12

12 10 300

实线

26

24

Reflectance (%)

Reflectance (%)

26

400

500

600

700

800

10 300

900

400

(A) sols without aging

500

600

700

800

900

Wavelength (nm)

Wavelength (nm)

(B) aging time of sols: 24 hours

Fig.3. Experimental reflectance data as a function of wavelength for TiO2 single-layer ARC on Si substrates at the different annealing temperatures Fig.3 shows the measured reflectance spectra of the TiO2 single-layer ARCs on Si substrates. The reflectance at the annealing temperature of 150 ℃ was the lowest in Fig.3.(A). The average reflectance of approximately 13.2% between 300 and 900 nm was obtained by the TiO2 single-layer ARC on Si substrate. Fig.3.(B) shows that the reflectance of the TiO2 single-layer films with aging time of 24 hours was much higher. The average reflectance was 16%. The effect of antireflection would probably become worse if the sols were aged for 24 hours because of the larger particles of sols. 30

实线

150℃

虚线

350℃

点线

550℃

20 15 10 5 300

400

500

600

700

Wavelength (nm)

(A) sols without aging

800

900

Reflectance (%)

Reflectance (%)

25

30 28 26 24 22 20 18 16 14 12 10 8 6 300

400

500

实线

150℃

虚线

350℃

点线

550℃

600

700

800

900

Wavelength (nm)

(B) aging time of sols: 24 hours

Fig.4. Experimental reflectance data as a function of wavelength for SiO2/TiO2 double-layer ARCs on Si substrates at the different annealing temperatures The measured reflectance spectra of the SiO2/TiO2 double-layer ARCs are shown in Fig.4. It was found that the reflectance of the double-layer ARCs was lower than that of single-layer ones. The minimum value of the reflectance of SiO2/TiO2 double-layer ARCs was 8.3% at the temperature of 150℃ in Fig.4 (A ). Fig.4 (B ) - 25 -

also shows that the effect of antireflection was worse after sols were aged for 24 hours.

(A) 150℃

(B) 550℃

Fig.5. The AFM diagram of the films at the different annealing temperatures The values of surface roughness of the A and B films were 0.276μm and 0.345μm respectively ,which indicate that the surface roughness increased as the annealing temperature increased.

4.

CONCLUSIONS

In this work, our aim is to prepare high-quality single- and double-layer ARCs on Si solar cells using a sol-gel process. The reflectance of TiO2 single-layer and SiO2/TiO2 double-layer ARCs were determined to be 13.2% and 9.8%. The sol-gel derived AR coatings posses comparable or even superior properties, compared to the coatings prepared by traditional vacuum-based technologies. In the future, using a combination of spin coating and rapid thermal annealing process, the sol-gel ARCs could have high potential for continuous mass production of Si solar cells at a fraction of the cost.

REFERENCES 1.

D.Bouhafs, A.Moussi, A.Chikouche, J.M.Ruiz, Energy Master. Sol. Cells 52 (1998) 79.

2. A.Yen, H.I.Smith, M.L.Schattenbyrg, G.N.Taylor, J.Electrochem. Soc. 139 (1992) 616. 3. J.Sczybowski, G.Brauer, G.Teschner, A.Zmelty, Surf.Coat.Technol. 98 (1998) 1460. 4. C.E.Tracy, W.Kern, R.D.Vibronek, US Patent No.4241108, 1980. 5. B.E.Yoldas, US Patent No.4361598 and 4346131,1982. 6. R.Kishore, S.N.Siingh, B.K.Das, Sol. Cells 26 (1992) 27. 7. Z.Chen, P.Sana, J.Salami, A.Rohatgi, IEEE Trans.Electron Devices 40 (1993) 1161. 8. K.Ram, S.N.Singh, B.K.Das, Renewable Energy 12（1997）131. 9. C.Battaglin, F.Caccavale, A.Menelle, M.Montecchi, E.Nichelatti, F.Nieoletti，Thin Solid Flims 351 (1998) 176. 10. U.Gangopadhyay, K.Kim, D.Managalaraj, J.Yi, Appl. Surf. Sci. 230 (2004) 364. 11. A.Morales, A.Duran, J.Sol-Gel Sci.Technol. 8 (1997) 451. 12. T.Schuler, M.A.Aegerter, Thin Solid Films 351 (1999) 125. 13. W.Kern, D.Tracy, RCA Rev. 41 (1980) 133. 14. M.A.Green, Appl .Phys .Lett 44 (1984) 1163.

- 26 -

Studies on the surface and interface of SiO2/Si (111) in mono-crystal Si solar cell by means of AES and XPS Feng Li, Zhong Quan Ma, Bo He, and Xia Jie Meng Based on different experimental techniques-Auger electron spectroscopy (AES) and X-ray photoelectron spectroscopy (XPS), we examined the specific surface and interface characteristics of SiO2/Si (111) in mono-crystal Si solar cells by using practical device wafers. The combination of AES and XPS can explore not only the chemical component of the wafer, but also the chemical states of every element. After detailed experiments, both chemical components and the chemical states of every element in the wafers were determined. In order to understand the real surface and interface structures of SiO2/Si (111) in mono-crystal Si solar cells, we used practical device wafers. The starting material was 12.5cm×12.5cm Cz silicon wafers, P-type, B-doped, 1.0-3.0Ωcm in resistivity, 200µm thickness and (100) oriented. The wafers were chemically eroded in diluted NaOH solution to form the pyramidal surface structure, i.e., (111) oriented surface and interface structures. After rinsed by de-ion water, the cleaned wafers were stacked vertically in an atmosphere of N2 and O2 using POCl3 liquid source with temperatures of 845℃, 840℃, and 854℃. Therefore, the N-P junction and phosphorous silicate glass (PSG) formed in the interior and on the surface of the silicon wafers, respectively. Then the PSG was removed by the above mentioned HF solution. A smooth (111) oriented side of the wafer was selected as micro-analytic object. According to the overall AES spectra for the analytic object in the kinetic energy region of 0-1800ev, phosphorus, silicon, carbon, and oxygen were detected. At the surface of the object, carbon and oxygen were supposed to have the highest and secondly high atomic concentration, respectively, and silicon has the third-high concentration, which was definitely higher than that of phosphorus. When the analytic object was sputtered, the sputtering rate was set at 20nm/min.After 0.5 minutes of sputtering, the atomic concentration of several elements in the wafer changed significantly. The atomic amount of carbon and oxygen descended sharply, and the concentration of silicon ascended liked a vertical line, while that of phosphorus just showed a little rise. With sputtering time more than 0.5 minutes, the atomic concentration of oxygen kept changeless, but that of silicon still moved up. At the same time, the concentration of phosphorus and carbon still declined, but much more slowly than before. The moving trends of the atomic concentration are consistent with the experimental results of XPS. The XPS has the ability to explore bonds with good accuracy in bonding energy (BE), stoichiometry, and spatial distribution. According to the overall XPS spectra for the object in the binding energy (Eb) region of 0-1000ev, P-based bond (P2s and P2p peak), Si-based bond (Si2s and), C-based bond (C1s peak), and O-based bond (O1s, O2s and OKLL peak) were detected. The C1s peaks-284.84ev and 285.30ev, at the surface and the interface respectively, was observed in some samples. The peak observed at 99.74ev, due to the photoelectrons excited from the XPS Si2p at the surface of the sample, was 0.28ev lower than that at the interface. As to the least impurity phosphorus, P2p spectra were fitted to peaks 133.74ev and 134.82ev at the - 27 -

surface and interface, respectively. The peak of O1s was indicated at 534.02ev at the interface, 0.38ev higher than that at the surface. The full wave at half maximum (FWHM) for C1s and Si2p remained nearly unchanged from the surface to the interface of the wafer, while that of O1s and P2p declined. Based on the experimental results of both AES and XPS, the high concentration of carbon at the surface is partly due to the introduction of carbon in the course of eroding by diluted HF acid solution. A lot of carbon atoms assembled in the surface and near-surface. Even after being moved by a thin slice of PSG, the carbon concentration at the surface is still very high, for carbon is the intrinsic impurity in the starting material with a concentration of ~1017/cm3. According to binding energy excursion and shape of peaks, we deduce that C1s can mainly be ascribed to C-H bonding and C-O bonding both at the outmost surface and the interface of the wafer. Silicon is the main component in the wafer, so the peaks of Si2s and Si2p are very high at the surface and interface.Si2p at the surface is ascribed to Si atom, while at the interface the analytic results show that some of the silicon atoms had been replaced by phosphorus atoms. Due to phosphorus is the extrinsic impurity, its atomic concentration is very low. At the outmost surface and in the deeper layer, P2p mainly is related to O-P bonding, which may be included in compound such as (PhO)3PO. Accordant with the aforementioned, most of O1s are the possible formation of Si-O2 bonding and O-C bonding at the surface, while in the deeper interior of the wafer, there existed O-P bonding, C-O-C or Si-O-Si bridges, and Si-O-C linkages. In a word, during the thermal diffusing process of N-P junction formation, solid resultants aggraded at the surface and near-surface, such as SiO2 and P2O5. However, most of these oxides were removed by the diluted HF solution, and only a small quantity of oxide left on the wafers acted as passivation layer that reduced the recombination of carriers.

- 28 -

Thickness impacts of aluminum and silicon on bow of silicon solar cells L. Zhang*a, W. Wua, M. Lia, Q.F.Sub, Y. H. Zhangb, Z. Q. Maa, b a Department of Physics, Shanghai University, Shanghai 200444, P.R. China b SHU-SOLARE’s R&D Lab, Shanghai University, Shanghai 200444, P.R. China

ABSTRACT As the thickness of silicon solar cells becomes thinner, the cells are susceptible to bow because of the metallization of metal and semiconductor on the front and rear contact. In this work, the thickness impacts of aluminum and silicon on bow of the solar cell have been investigated with the perspective of deformation and strain. Keywords: solar cell, bow INTRODUCTION At present, crystalline silicon is the most popular material for solar cells. Crystalline silicon solar cell has taken 90% of the market. How to reduce the cost has become the most urgent task. For most of the solar cells in the market, the price of original wafer is about 60% of the solar cell. The cost can be dramatically reduced if we can reduce the thickness of the cell while maintaining its efficiency. Presently, the thickness of nearly all silicon solar cells is about 200µm. Once made thinner, cell will be seriously bowed. It not only breaks the evenness of the cell, but also impacts the test and encapsulation. When the thickness has been reduced to 150μm, the cell will be bowed to about 6~10mm. In this work, we study the thickness impact of aluminum and silicon on bow of the solar cell from the perspectives of deformation and strain. DEFORMATION THEORY OF BOW During the state-of-art of processing, the cells are flatly transferred into the sinter equipment, and the boundary conditions of enforce on the silicon wafer are very complex. For simplicity, we use the cantilever girder model according to literature [1], fixing one end, and measuring the height of the other (no gravity). Then, we can calculate the bow of the cell using the following equation:

δ =

3(α b − α a )(T f − T )(t b + t a )d 2  t t E t E t 4t b2 4 + 6 a + 4( a ) 2 + ( a )( a ) 3 + ( b )( b tb tb Eb t b Ea t a 

(1)  ) 

Where δ is the height of the bow, αa is thermal coefficient of expansion of silicon, αb is thermal coefficient of expansion of aluminium, Ea is modulus of elasticity of silicon, and Eb is modulus of the elasticity of aluminium. ta is silicon thickness, tb is aluminium thickness, Tf is eutectic temperature (577℃), T is the test temperature (room temperature), d is the length of the cell. In Eq. (1), ta and tb are main variable which are associated with the thickness of silicon and aluminium, respectively. The other parameters are taken as constants and they do not change with the temperature.

- 29 -

Furthermore, ta and tb are independent variables, and ta changes from 100 to 200µm, while tb changes from 10 to 55µm. Fig. 1 shows the calculation results. The deformation degree of the cell is significantly correlative to the thickness of the printed Al layer and the bulk silicon. The thicker the Al layer or the thinner the Si is, the higher the cell is bowed. Considering the fact that the thermal expansion coefficient of Al is 23×10-6K-1 and that of Si is just 2.44×10-6K-1 [2], from eutectic temperature 577℃ to room temperature 23℃, the deformation of Al is much greater than that of Si, so the cell is bowed. When bulk Si is 180µm and printed Al layer is 30µm, bow will be about 3～4mm; which is large enough to pay attention. We first calculate the bow from Al and Si thickness, and then analyze the bow with strain theory.

Fig. 3. Bow as a function of the thickness of bulk silicon and aluminium layer STRAIN THEORY OF BOW Because the solar cell is bowed at one direction, for simplicity, we can use multilayered beams, and just consider elastic deformation. Here we use the double films model in the Ref. 3 to consider the relationship between strain and thickness of bulk Si and printed Al layer.

Fig. 2.Double films model Modulus of elasticity and thermal coefficient of expansion of Si and Al are Ea,αa and Eb,αb, respectively. E stands for modulus of elasticity, α stands for thermal coefficient of expansion. This model is suitable for bow caused by temperature change. If the radius of curvature of neutral axis isρ, the strain at t is [3]:

- 30 -

Є=

t + ε0 p

Є0=

1

ρ

A B

=

(2)

(3)

C (4) B

A = ∆T {E a2 t b4α a + Eb2 t a4α b + E a Eb t b t a K } K = 3t b t a (α a + α b ) + t b2 (3α a + α b ) + t a2 (α a + 3α b ) B = E a2 t b4 + Eb2 t a4 + 2 E a Eb t b t a (2t b2 + 3t b t a + 2t a2 ) C = 6∆TE a Eb t b t a (t b + t a )(α a − α b ) Є0 is the strain in centroidal axis; △T is the temperature change. According to Eqs. (2)- (4), we define Є0 as a function, and Si thickness ta, printed Al layer thickness tb as independent variable, t=0, so we get the strain in centroidal axis shown in Fig. 3.

Fig. 3. Strain in the solar cell as a function of thickness of silicon and aluminium film. From Fig. 3, the thinner the printed Al layer or the thicker the bulk Si is, the smaller the strain is, and the less likely the wafer is going to bow. It is in agreement with our earlier analysis. EXPERIMENTS We have employed 10 pieces solar cells (CZ, p-type, (100) direction, 125×125mm2) for the experiments to study the bow at different thicknesses of Si and Al, and compared with the theoretical calculation. We use vernier caliper to measure the bow, because the solar cells are bowed symmetrically along the edge. Suppose the cells are put flatly on a table, we measure the distance between the middle of cell and the desk. The thicknesses of Si and Al are obtained by a XP Series Stylus Profiler of Ambios Technology, Inc.

- 31 -

Table 1.Bow of the solar cells Si thickness ta

Al thickness tb

Bow δ (mm)

Bow δ (mm)

(μm)

(μm)

(theory)

(experiment)

1

175

43

4.330

3.88

0.450

2

175

44

4.440

4.27

0.170

3

181

40

3.700

3.66

0.040

4

182

30

2.640

3.05

0.410

5

190

32

2.590

3.06

0.470

6

138

33

5.611

5.99

0.379

7

139

42

7.324

7.34

0.016

8

140

39

6.608

7.08

0.472

9

140

40

6.807

6.48

0.327

10

138

42

7.440

6.97

0.470

Cells

Error (mm)

From Table 1, we find that the difference between experimental data and theoretical results is less than 0.5mm, demonstrating the good agreement between the theory and the experiments. It should be noted that we should take the unevenness of the solar cell into account when comparing the experiment with theory. CONCLUSIONS This work calculated the bow of the solar cells with the perspective of the deformation and strain. We have found that the main factor to the bow is the firing temperature and thickness of silicon and aluminium. We conclude that after firing, the thicker the Al layer or the thinner the Si is, the higher the cell is bowed.

REFERENCES 1.

R. J. Roark and R. Jefferson, Formulas for stress and strain, McGraw-Hill, 1976.

2.

F. Huster, Aluminum back surface field: Bow investigation and elimination, 20th European Photovoltaic Solar Energy Conference and Exhibition, Barcelona, 6-10 June 2005.

3.

N. H. Zhang and J. J. Xing, An alternative model for elastic bending deformation of multilayered beams, J. Appl. Phys. 100, 103519 (2006).

- 32 -

Effect of reflectance on efficiency during the solar cell process Zhengshan Yu, Xuzhao Shen, Peng Lv; Zhenguo Gan; Qingfeng Su Technical Department, Solar Enertech (Shanghai) Co., Ltd., 201206 Abstract Key words: PECVD, Reflectance, Anti-reflection film, Nitride-silicon

In the experiment, the reflectance of samples is measured by a Pgeneral TU1901 UV-VIS Spectrophotometer. First, the reflectance of the wafers after each process step is measured. In Fig. 1，the light blue curve is the reflectance of origin wafer，the dark blue curve is the textured wafer, and the red one is the wafer with a SiNx coating. Obviously, there is a great impact on the optical characteristic of the wafer after each step, which reduces the reflectance a lot. Plasma Enhanced Chemical Vapor Deposition (PECVD) technique is used to deposited the SiNx coating. And the reflectance comparing results of the samples processed by different parameter settings is shown in Fig. 2. It’s obvious that different results are obtained by samples with different parameters.

Reflectance(%)

Reflectance(%)

According to the testing results, the reflectance is controllable by adjusting the PECVD process parameters.

200

250

300

350

400

450

500

550

600

650

700

750

800

850

16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 -1 200

250

300

350

400

450

500

550

600

650

700

750

800

850

Wavelength(nm)

Wavelength(nm)

Fig. 1 Reflectance curves of sample after each process step

Fig. 2 Reflectance curves of samples by different PECVD process

Atmospheric Pressure Chemical Vapor Deposition (APCVD) technique is also used to deposite the TiOx coating to make a compare with the SiNx coating. In Fig. 3, the blue curve represents the sample processed by APCVD while the red one is made by PECVD. The reflectance of the APCVD sample is higher than the PECVD sample at the wavelengths from 400nm to 850nm. From Fig. 3, it is inferred that the solar cells coated with SiNx should perform better than the ones coated with TiOx. Couples of cells are made with different coating process. The measured efficiency of them is shown in Fig. 4. The efficiency of the cells increases a large step when coated with SiNx. Definitely the reflectance plays an important role in the solar cells, and the reflectance curve can be - 33 -

adjusted and controlled through different process. The sunlight irradiance under AM1.5 is shown in Fig. 5. A different spectrum is obtained under different wavelength. Next step, the reflectance will be adjusted by changing the sheet resistance to make it possible that cell has a minimum reflectance in the wavelength range at which the sunlight gives strongest energy. 80 70 Number of Cells

60

Reflectance(%)

16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 -1 200

50 40 30 20 10 0

250

300

350

400

450

500

550

600

650

700

750

800

1

2

3

4

5

6

7

8

9

10

Efficiency Range Cells with SiNx coating Cells with TiOx coating

850

Wavelength(nm)

Fig. 3 Reflectance curve of samples by PECVD and APCVD process

Fig. 4 The efficiency of the solar cells

Fig. 5 The spectrum of the sunlight

- 34 -

Effects of packaging materials on spectrum response of solar module Yanting Yin, Ming Luo; Lingbo Zhou; Zhengshan Yu, Qingfeng Su Technical Department, Solar Enertech (Shanghai) Co., Ltd., 201206 Abstract Key words: EVA, low-iron toughened glass, Anti-reflection coating, refractive index

The main encapsulation materials which affect the spectrum response of solar module are EVA mucus、 low-iron toughened glass. The Anti-reflection coating of solar cell also has great effect on the spectrum response. There exists a phenomenon that the power will increase a little in the process of module manufacturing. For example, the same power cells from different manufactures are made into modules. Then the modules combined with these cells are tested under the same testing situation. The power of the modules isn’t the same and the difference ranged about 3% maximum. Because of the different material such as EVA, low-iron toughened glass and the different thickness of anti-reflection coating, the combined refractive index and the combined reflectivity should have difference.

Fig.1 The size of the glass

Fig.2 The sample module

The pictures of the samples are shown in Fig. 1 and fFig.2, respectively. The experiment is aimed to measure the reflectivity of the sample module using a UV-VIS spectrophotometer. The encapsulation materials are changed and then the response of the modules is tested again. The difference between the encapsulation materials and the result is recorded. The process of cells is changed to make the thickness and refractive index of the coating different. Then the combined reflection of the module made of the wafers is tested. According to the results, the experiments are processed time after time so as to find the best matching between the materials and the coatings. Small toughened glass is used instead of the big one so as to be put into the spectrophotometer. Under the normal process, a sample module after lamination is get. The spectrum response ranged from 200nm-900nm at the coordinate of wavelength. The spectrum response test coordinated by wavelength shown in Fig.3 is processed.

- 35 -

Fig.3 The spectrum response coordinated by wavelength

- 36 -

Synthesis and characterization of SiO2 capped ZnCdS nanocrystals Yang Guo-Wei*, Li Dong-Mei, Zhu Xue-Feng, Xu Zhen,Wang Yin-Qiang School of Materials Science and Engineering, Shanghai University, No.149, Yanchang Road, Shanghai 200072, P.R.China

ABSTRACT ZnCdS/SiO2 nanocrystals with core-shell structure were fabricated by Micro-emulsions method via the hydrolysis of tetr-aethyl orthosilicate (TEOS). Using XRD and TEM, the size of capped nanocrystals was around 22nm with the uncaped round 12nm. Emitting peaks in PL spectra of ZnCdS shifted from 390nm to 540nm by adjusting the proportion of Zn/Cd. The water solubility and optical property are both obviously enhanced after the capsulation of SiO2. Keywords: nanocrystals, Silica, Core-shell Structure, ZnCdS

INTRODUCTION ZnS and CdS are both wide band gap semiconductors at room temperature[1] with similar crystal structure and nearly the same lattice constant, which is beneficial to forming ZnCdS nanocrystals of high quality. Thereby, ZnCdS has been widely used in the flat-panel display, electroluminescent devices, photocatalysis, single electron devices and so on [1–3]. Nanocrystals with core-shell structure are widely applied in optics, biology, catalysis and some other domains due to its unique configuration and nice optical property. In the past decade, the capsulation of SiO2 to nanocrystals was extensively investigated. It was reported that the chemical stability and fluorescence efficiency[1,4,5] were obviously enhanced with the change of water solubility[5] after the capsulation of SiO2. Porous SiO2 shell can be easily dressed by assorted functional groups which open up new applications in interconnector or nano-electronic devices, and even biochemical capsules[6]. In this work, micro-emulsions method was firstly used to fabricate ZnCdS nanocrystals for its low cost as well as the small particle size[7] which can be changed easily. Through the hydrolysis of TEOS[5], SiO2 was successfully coated onto ZnCdS. The optical property of prepared nanocrystals was well studied by FT-IR and PL which showed a large peak shift by changing Zn/Cd ratio and an enhancement of fluorescence efficiency after capsulation as well.

EXPERIMENTS All chemical reagents were of analytical level without further purification. In the experiment, Nitrate of both Zinc and Cadmium were dispersed pro rata into water phase to form a liquid with the concentration of positive ions equal to 0.1M. Then, the liquid was introduced with toluene, sodium dodecylsulfate (SDS) and N-Pentyl alcohol (C5H12O) in order under stirring at room temperature till it formed the transparent solution bath. Na2S micro-emulsions system was prepared in the same way and mixed into the above-mentioned bath

- 37 -

with well stirring till its pH value above 7 to prepare ZnCdS. Finally, the ZnCdS/SiO2 core-shell spheres were prepared by the hydrolysis of TEOS. Typically, to spur the hydrolysis of TEOS, a small amount of ammonia solution was added and the mixture was kept at room temperature under continuous agitation for 24h. When the reaction completed, the sample was centrifuged, washed by deionized water several times and freeze-dried. Transmission electron microscopy (TEM) (JEOL ， JEM-2010) and XRD were employed to characterize the samples. The photoluminescence (PL) properties were studied by PL spectrometer (Hitachi F-4500). FT-IR absorption spectra were characterized by Avatar 370 (Nicolit, USA) spectrometer. *[email protected]

RESULTS AND DISCUSSION Appling the TEM, ZnCdS and ZnCdS/SiO2 nanocrystals are 12nm and 22nm in particle size respectively, as shown in Fig. 1. nanocrystals of both samples are nearly in ball shape. The augment of particle size is mainly resulted from coating shell of SiO2. Fig. 2. shows XRD charts of ZnCdS, ZnCdS/SiO2 and pure SiO2 nano-particles. In Fig. 2.(a), three peaks of ZnCdS nano-particles located at 27.56°, 45.92° and 54.57°, respectively between the standard peaks of (111), (220), (311) of beta-ZnS and beta-CdS, proving the ZnCdS mischcrystal of good quality and nice cubic wurtzite structure. From Fig. 2.(b), only one peak at 24° can be observed which is corresponding to the peak of pure SiO2 shown in Fig. 2.(c), proving a nice shell of SiO2 was formed after the hydrolysis of TEOS. All the above peaks were broadened due to quantum size-effects where the particle sizes of ZnCdS and ZnCdS/SiO2 were respectively calculated to be 11nm and 20nm from (111) peaks according to Scherrer formula, which is in accordance with the TEM images.

Fig. 1. TEM images of (a) bare and (b) silica-coated ZnCdS nanocrystals.

- 38 -

FT-IR spectra of ZnCdS/SiO2 nanocrystals are shown in Fig. 3. where the absorption peaks at 1635cm-1 and 3450cm-1 represent the flexural oscillation of H-O-H and stretching oscillation of O-H. The absorption peaks at 1380 cm-1 and 1102 cm-1 are probably attributed to the residual SDS and organic solvent. Two peaks near 1210cm-1 and 800cm-1 resulted from the symmetric oscillation and anti-symmetric oscillation of Si-O-Si, while the peak near 960cm-1 is caused by the stretching oscillation of Si-OH[8], which convincingly proves the existence of SiO2 phase. The quality of mischcrystal is greatly influenced by the concentration of S2- as well as its addition manner. PL spectra of ZnCdS nanocrystals prepared under different addition manners of S2- are displayed in Fig. 4. with the ratio of Zn2+/Cd2+ in each sample equal to 1/4 (M is mechanical mixtures of ZnS and CdS nanocrystals by the same ratio). M1 was prepared by adding 2.4ml of 2M S2--contained solution into Zn2+/Cd2+-contained micro-emulsions quickly. M2 was prepared by adding 47ml of 0.4M S2--contained micro-emulsions into the micro-emulsions fast. M3 was prepared by adding the S2--contained micro-emulsions with slow speed around 1ml/min. The main emitting peaks of three samples (M1, M2 and M3) are all sited near 540nm attributed to band-edge emition[2]. Each main peak has a shoulder peak at 560nm which is caused by precipitated mixture of CdS and ZnS. The inset in Fig. 4. shows the PL spectra of mechanical mixtures with different ZnS/CdS ratios, all of which have an emitting peak near 560nm which is in accordance with the characteristic peak of CdS. The intensity of emitting peak of CdS is enhanced for the existence of ZnS on account of radiation transfer, which explains the increase of peak intensity of CdS as the amount of ZnS rises（as shown in inset）. The intensities of main peaks are obviously enhanced twice from M1 to M3, while the intensities of shoulder peaks are gradually reduced, making the whole peaks more symmetrical and full width at half magnitude (FWHM) lesser. This shows that it is much easier to create mischcrystal of high quality with lower concentration and slow dropping speed of S2-.

Fig. 2. XRD spectra of (a) bare, (b) silica-coated

Fig. 3. FT-IR spectra of (a) silica-coated and

ZnCdS with Zn/Cd=2:3 and (c) silica.

(b) bare ZnCdS with Zn/Cd=2:3.

- 39 -

Fig. 4. PL spectra of ZnCdS, Inset: PL emission

Fig. 5. PL spectra of bare (a,b,c,d) and silica-overcoated

spectra of mechanical mixtures.

(A,B,C,D) ZnCdS.

PL spectra of ZnCdS and ZnCdS/SiO2 nanocrystals are presented in Fig. 5., where ZnCdS nanocrystals are of different Zn2+/Cd2+ ratios. The emitting peak moves toward short-wave range from 540nm to 390nm as the ingredient of Zn rises. FWHM equals to 50nm, much lesser than the results reported in other papers[2,3]. Four peaks at 390nm, 470nm, 505nm and 540nm are attributed to band-edge emitting[2]. The shift of peaks demonstrates different Zn/Cd ratios accompanied by the change of band-gap of ZnCdS. The emitting peak of each sample in Fig. 5. is asymmetric caused by the shoulder peak at 560nm which is attributed to the precipitated mixture of CdS and ZnS in the preparation. As shown in Fig. 5., the emitting peaks of PL spectra of capped samples(A,B,C,D) are unchanged in position and shape, with the fluorescence intensities increasing twice compared with the uncoated one. The surface state of nanocrystals is inactivated after coated with SiO2 accompanied by the reduction of the surface recombination and nonradiative transition of excitons, highly enhancing the fluorescence intensities of emitting peaks and proving that the emitting peaks are attributed to the band-edge emitting. The change of Zn/Cd ratios leads to the change of the position of emitting peak as well as the band-gap of mischcrystal.

CONCLUSIONS In this paper, ZnCdS and ZnCdS/SiO2 nanocrystals were successfully prepared via micro-emulsions method. XRD, TEM results proved the particle size of ZnCdS nanocrystals was 12nm, and was 22nm after the capsulation of SiO2. Through technological improvement, the precipitation of ZnS and CdS was greatly inhibited with the quality of mischcrystal highly improved. PL results demonstrated FWHM of the fluorescence peak of ZnCdS was only 50nm. With different Zn2+/Cd2+ ratios, the band-gap of ZnCdS changed with peak shift of around 150nm according to PL spectra. The intensity of emitting peak was enhanced as the amount of ZnS rises, where the detailed mechanism was still unknown. The encapsulation has highly enhanced the fluorescence efficiency and reduced surface defects and suspending bonds, effectively dressing the surface of nanocrystals and reducing the nonradiative recombination of excitons.

- 40 -

REFERENCES 15. H. Song, et al., Mater. Sci. Eng. B (2007).

16. S.K. Kulkarni, U. Winkler and N. Deshmukh et al., "Investigations on chemically capped CdS, ZnS and ZnCdS nanoparticles, " Applied Surface Science, 169-170, 438-446(2001). 17. O. Raymond, H. Villavicencio and V. Petranovskii et al., "Growth and characterization of ZnS and ZnCdS nanoclusters in mordenite zeolite host, " Materials Science and Engineering, A360, 202-206(2003). 18. Y. Hattori, T. Isobe and H. Takahashi, et al., "Luminescent properties of ZnS:Mn2+ nanocrystals/SiO2 hybrid phosphor synthesized by in situ surface modification co-precipitation, " Journal of Luminescence, 113,69-78(2005). 19. Heesun Yang, "Water-soluble silica-overcoated CdS:Mn/ZnS semiconductor quantum dots", JOURNAL OF CHEMICAL PHYSICS, 121(15), 7421-7427(2004). 20. Kun Han, Zhihui Zhao and Zheng Xiang, et al., "The sol–gel preparation of ZnO/silica core-shell composites and hollow silica structure, " Materials Letters, 61, 363-368(2007). 21. Pierre Bauduin, Didier Touraud and Werner Kunz, et al., "The influence of structure and composition of a reverse SDS microemulsion on enzymatic activities and electrical conductivities," Journal of Colloid and Interface Science, 292, 244-254(2005). 22. Hernandez-Padron G, Rojas F and Garcia-Garduno M, et al., "Development of hybrid materials consisting of SiO2 microparticles embedded in phenolic-formaldehydic resin polymer matrices," Materials Science and Engineering A, 355(1), 338-347(2003).

- 41 -

Synthesis of functionalized ZnS:Mn/ZnS nanocrystals Dongmei Li*, Yiqiang Wang, Guowei Yang, Zhen Xu School of Material Science and Engineering, Shanghai University, No.149, Yanchang Road, Shanghai 200072, P.R.China

ABSTRACT ZnS:Mn/ZnS core/shell nanoparticles were pre-synthesized by microemulsions technique. To prepare water-soluble and biocompatible nanoparticles, thioglycolic acid was directly added into the same reverse microemulsion system to modify the surface of ZnS:Mn/ZnS nanoparticles. The functionalized ZnS:Mn/ZnS nanoparticles were coupled with carboxyl and revealed farther luminescence enhancement at 600nm. The experiment results indicated that passivating of the organic molecules and the polynuclear complex of Zinc may be the possible mechanism leading to enhancement of luminescence. Keywords: microemulsions, nanoparticles, ZnS:Mn/ZnS, thioglycolic acid, photoluminescence

INTRODUCTION In comparison with organic dyes and fluorescent proteins, semiconductor nanoparticles have received a great deal of attention in recent years because of their better brightness, more stability against photobleaching, longer excited-state lifetimes, narrower full-width at half maximum (FWHM). These characteristics make semiconductor nanoparticles attractive materials for biological probe, DNA hybridization detection, immunoassays and biosensors[1-3]. It is possible that shell layer grows on core layer because they have very similar crystal structure. The surface defects of core layer are eliminated greatly and fluorescence quantum productivity increase after modified by shell layer. In addition, functional groups of organic materials were also reported as surface modification for the enhancement[4] of photoluminescence[5]. Recently, more attentions have been paid to the biological detection and imaging application of CdSe, CdSe/ZnS or CdS nanoparticles[6,7]. Although, the toxic transition (‘heavy’) metal of cadmium hampers the application of vivo test, doped nanopaticles provide an alternative approach to acquire variety photoluminescence emissions by choosing various dopants with less toxin[8,9]. As a nontoxic Ⅱ-Ⅵsemiconductor material, ZnS can be synthesized without rigorous reaction condition. Thus, it is considered as a promising host material. In the present work, a new approach of reverse micromulsions method was used to prepare thioglycolic acid capped ZnS:Mn/ZnS nanopaticles due to low cost and non-toxin. Such nanoparticles bonded with carboxyl groups have numerous applications by coupling with various biological molecules.

EXPERIMENTAL ZnS:Mn/ZnS core/shell nanoparticles were prepared in microemulsions. All chemical reagents were of analytical grade and used as received without further purification. In the experiment, Mn2+ doped ZnS - 42 -

nanoparticles was prepared by a reverse microemulsions method. 0.8M SDS toluene solution was mixed with 0.1 M Zn(NO3)2 and 4×10-3M MnSO4 with a Mn2+-to-Zn2+ ratio equal to 0.04. N-Pentyl alcohol was added as a co-surfactant until pellucid microemulsions was obtained. Then, 2M Na2S solution was rapidly injected to the microemulsions in order to the formation of Mn2+ doped ZnS nanoparticles.ZnS:Mn/ZnS nanoparticles were formed by the mixing of ZnS:Mn2+ (previous prepared) and Zn2+-contained microemulsions, followed by the addition of S2--contained microemulsions for the growth of a ZnS shell. After well stirred, 1M thioglycolic acid solutions was added into microemulsions to functionalize ZnS:Mn/ZnS nanoparticles. Transmission electron microscopy (TEM) experiments were carried out with JEM-200CX (JEOL, Japan) electron microscope to obtain images of individual QDs. FT-IR absorption spectra were obtained using Avatar 370 (Nicolit, USA) spectrometer. Photoluminescence spectra were collected using Hitachi FL-4500 luminescence spectrometer with an excitation wavelength of 300nm. All tests were done at room temperature.

RESULTS AND DISCUSSION Thioglycolic acid (TGA) can link to ZnS:Mn/ZnS nanoparticles by covalent bond through the combination of mercapto groups and Zn dangling bond on the surface of ZnS shell. And the polar carboxylic acid group renders the nanoparticles water soluble and biocompatible. The functional ZnS:Mn/ZnS nanoparticles could well disperse in a solution for two months without any precipitation. FT-IR spectra of bare and TGA-overcoated nanoparticles are shown in Fig. 1.,the relative position of adsorption peaking are similar in sample A and B except at 1600 cm-1 and 1380 cm-1, where appearance should be due to C=O stretching peak and COO- symmetric stretching (νS) respectively(Sample C). In all samples, the adsorption peak at 3500cm-1 is a typical peak arising from the bonding of O-H, it indicates water was absorbed from air, which leads to ydroxyl group was absorbed onto the surface. Meanwhile, the peak at 2850cm-1 from the vibration adsorption of C-H bonding can presumably be contributed to the residual SDS and toluene. In sample B, two peaks at 1600 and 1380 cm-1 in carboxyl-functionalized ZnS:Mn/ZnS nanoparticles were caused by C=O stretching peak and COO- symmetric stretching peak (νS), respectively. It indicates the presence of carboxyl groups in ZnS:Mn/ZnS nanoparticles. In particular, the presence of S-H stretching peak (2550cm-1) of sample C confirmed that TGA was indeed covalently bonded to the ZnS:Mn/ZnS nanoparticles. Fig.2. is TEM images of ZnS:Mn/ZnS and TGA-ZnS:Mn/ZnS nanocrystals, respectively, the shape of crystal grains is almost spheric, and grain diameter is about 10-12nm.

- 43 -

Fig. 1.

FT-IR spectra of (a) ZnS:Mn/ZnS,

Fig. 2. TEM image of (a) ZnS:Mn/ZnS,

(b) TGA- ZnS:Mn/ZnS nanoparticles, (c) TGA.

(b) TGA- ZnS:Mn/ZnS nanoparticles.

PL emission spectra of unpassivated, ZnS passivated and TGA-overcoated ZnS:Mn quantum dots are compared in Fig. 3. 4

6

In sample A, the stronger peak at 600nm is attributed to the transition from energy 2+

level T1 to A1 of Mn ions and two weaker peaks at 330nm and 460nm are derived from the band edge and the radiative recombination of the defect-related ZnS host, respectively. However, the emission peaks of the defects and dopants were insignificant due to the existence of abundant surface states. In sample B, the emission intensity at 600nm was five-fold higher than that of the sample A. Thus, the existence of a uniform ZnS

shell structure caused the further decrease of dangling bond on the nanoparticles surface, and just due

to this reason, the enhancement of photoluminescence intensity occurred. Comparing with the 600nm emission from uncoated ZnS:Mn/ZnS nanoparticles, TGA-ZnS:Mn/ZnS nanoparticles enhanced nearly 3 times. TGA- ZnS:Mn/ZnS nanoparticles enhanced nearly 3 times, which is in accordance with the prepare of ZnSe using TGA as stabilizing agent[10].

A well coverage of nanoparticle surface by TGA further

reduces the density of dangling bonds on the surface of ZnS resulted from a decrease in the number of trap sites available for nonradiative recombination. Because the carboxylic acid group can link to ZnS:Mn/ZnS nanoparticles by covalent bond through the combination of Zn dangling bond on the surface of ZnS shell, discontinuous shell can

further passivate surface states. Observed from FT-IR spectroscopy and PL

spectroscopy, it can be inferred that the presence of these covalently bonded TGA can passivate the surface of the particle leading to obvious enhancement of luminescence. Fig.4. show PL spectra of the supernation solution obtained

by decompose

microemulsion and

centrifugal separation. The stronger peak at 400nm was derived by the transition of n electrons in –OH group of pentanol molecule in all sample[11], it is obvious that the emission at 600nm could not be observed in the sanple A and B (in inset), which be attributed poor water-soluble of the ZnS:Mn andZnS:Mn/ZnS nanoparticles .while the dramatic emission at 600nm was found in the sample C, proving the TGA modified ZnS:Mn/ZnS nanoparticles of well water solution.

- 44 -

Fig. 3. PL spectra of the nanoparticles:(A)ZnS:Mn,

Fig. 4. PL spectra of the supernatant: (A )ZnS:Mn,

(B) ZnS:Mn/ZnS and (C) TGA- ZnS:Mn/ZnS.

(B) ZnS:Mn/ZnS and (C) TGA-ZnS:Mn/ZnS.

In TGA-ZnS:Mn/ZnS, the significant peaks at 330nm and 460nm were also detected. However, these peaks could not be found in sample A and B, which may be attributed to the PL of TGA. To affirm this hypothesis, the diluted sample of TGA-ZnS:Mn/ZnS with different dilution ratios were tested. As shown in Fig.5., the intensity of PL at 340nm and 600nm changed synchronously with the increasing of diluent (deionized water). The intensity increased when the solution was diluted to 10 times than the original, and then decreased with further dilution. This is because the initial concentration of nanoparticles (sample A) was too high and resulted in concentration quenching. The weakening of concentration quenching plays a dominative role when dilution with 10 multiples DI water, which results in the intensity reaching to the maximum(sample B). On the contrary, when the dilute multiple over 10, the decrease of luminophor become

Fig. 5. PL spectra of the (A) TGA-ZnS:Mn/ZnS nanoparticles, diluted with (B)10,(C)50,(D)100 multiples DI water. Inset:the relationship between the ratio of intensity at 340nm to 600nm and multiple of diluted. - 45 -

the exclusive influencing factor. This is the reason why the intensity decreased dramatically when the nanoparticles were further diluted. But a phenomenon must be noticed that the fraction of PL intensity at340nm dramatically decrease with diluting. The inset in Fig. 5. shows that the ratio of intensity at 340nm to 600nm is not independence on the concentration of ZnS:Mn/ZnS nanoparticles(decreased from 0.79 to 0.23), which corresponds to the fact that dilution can not influence the inherent photoluminescence properties of ZnS:Mn nanoparticles. These phenomena indicate that passivating of the organic molecules and the polynuclear complex of Zinc may be the possible mechanism leading to enhancement of luminescence.

CONCLUSIONS Functionalized

ZnS:Mn/ZnS

nanoparticles

were

synthesized

successfully

via

the

reverse

microemul-sions technique. We present a simple method for coupling TGA to single ZnS:Mn/ZnS nanoparticles. It is also believed that robust bond between the organic molecules and the ZnS:Mn/ZnS nanoparticle has been formed. Passivating mechanism of organic resulting in enhancement of luminescent intensity of TGA-ZnS:Mn/ZnS nanoparticles were proposed. Functionalized ZnS:Mn/ZnS nanoparticles presents not only well water-soluble and biocompatible capability bout also three-fold stronger PL intensity than un-modify one, which is promising in biological application. Modification of TGA can be used as an effective method to compensate feeble passivating from semiconductor shell or further increase the quantum yields .

REFERENCES 23. Bruchez, Marcel, et al., "Semiconductor Nanocrystals as Fluorescent Biological Labels," Science. 281(5385), 2013-2016 (1998). 24. Pathak S, Choi S K, Arnheim, " Hydroxylated Quantum Dots as Luminescent Probes for in Situ Hybridization," Journal of the American Chemical Society. 123(17), 4103-4108 (2001). 25. Goldman Ellen R,Anderson, George P, et al., "Conjugation of luminescent quantum dots with Antibodies using an Engineered Adaptor Protein to Provide New Reagents for Fluoroimmunoassays," Analytical Chemistry. 74(4), 841-847 (2002). 26. Nikesh V V, Mahamuni S, " Highly Photoluminescent ZnSe/ZnS Quantum Dots," Semiconductor Science and Technology.16(8), 687-690 (2001). 27. Kubo T, Isobe T, Senna M, "Enhancement of Photoluminescence of ZnS:Mn Panocrystals Modified by Surfactants with Phosphate or Carboxyl Groups via a Reverse Micelle Method ," Journal of Luminescence. 99(1), 39-45 (2002). 28. Gao, Xiaohu, Cui,Yuanyuan, Levenson, et al., "In vivo cancer targeting and imaging with semiconductor quantum dots," Nature Biotechnology. 21(4), 969-976 (2004). 29. Kim, Sungjee,Lim, et al., "Near-infrared Fluorescent Type II Quantum Dots for Sentinel Lymph Node Mapping," Nature Biotechnology, 22(1), 93-97 (2004). 30. Bhargava ,R .N. Gallagher, D. Hong, "Optical Properties of Manganese-doped Nanocrystals of ZnS," Physical Review Letters. 72(3), 416-419 (1994). - 46 -

31. Ollinger M., Craciun V., Singh R. K., "Nanoencapsulation of ZnS:Ag Particulates with Indium Tin Oxide for Field Emission Displays,"Applied Physics Letters. 80(11), 1927-1932 (2002). 32. Shavel A, Gaponik N,Eychmuller A, "Efficient UV-blue Photoluminescing Thiol-stabilized Water-soluble Alloyed ZnSe(S) Nanocrystals,"

Journal of Physical Chemistry B. 108(19), 5905-5908

(2004). 33. C.H.Chen, C.W.Tang, J.Shi, K.P.Klubek, " Imaproved red dopants for organic electroluminescent devices," Macromol.Symp. 125, 49-58 (1997).

- 47 -

Electrical properties of radiation detector based on polycrystalline mercuric iodide (HgI2) thick film Weimin Shi*, Yaoming Zheng, Yuying Guo, Yu Zhang, Huan Xu, Linjun Wang, Yiben Xia School of Materials Science and Engineering, Shanghai University, Shanghai 200072, China ABSTRACT Potentially low cost and large area polycrystalline mercuric Iodide (HgI2) is one of the preferred materials for the fabrication of room temperature X-ray and gamma-ray detectors. In this paper, the technique of fabricating polycrystalline HgI2 detectors was studied and the energy resolution of 13.1% for 5.5 MeV Am α particles at room temperature was obtained for the first time. The optimal choice of particle

241

injecting from negative interface enhances the collection efficiency. Keywords: Mercuric Iodide (HgI2), polycrystalline film, radiation detector, electrical properties

INTRODUCTION The growth of mercuric iodide bulk crystals by physical vapor deposition (PVD) has attracted much interest due to its use as γ-ray and X-ray detector material having favorable characteristics such as high atomic numbers and large bandgap (2.13 eV) [1]. Because of the low cost and ease of fabrication of large-area detectors, polycrystalline thin/thick films of mercuric iodide also have considerable potential for X-ray imaging detectors and nuclear detectors for γ-rays and β-particles [2,3]. Thick polycrystalline HgI2 film is most promising for use as a direct photo-conduction converter in X-ray digital radiography. It exhibits the highest X-ray sensitivity among the known polycrystalline or amorphous materials. In the past few years, the research of HgI2 detector was focused on single crystal mercuric iodide materials. It was very difficult to obtain large area, high resistivity, low defect concentration and good uniformity of single crystal material. Therefore, in recent years the research emphasis of HgI2 was transferred from single crystal to low cost and large area polycrystalline materials [1,2]. For many researches are focused on the technique of polycrystalline HgI2, polycrystalline materials with good properties are easily obtained, but the research on the fabrication process and properties of polycrystalline HgI2 detector for α particles has received very little attention in the literature. In this work, a thick polycrystalline HgI2 film with a thickness of 250 µm grown by Hot-Wall PVD method was used to fabricate the radiation detector to detect α particle [4].

EXPERIMENTAL DETAILS Polycrystalline HgI2 thick films were fabricated on ITO-coated glass substrates with an area of 4’×4’ by a Hot-Wall PVD technique. Prior to the deposition, the substrates were coated a layer of aluminum or palladium. After typical cleaning processes, the ITO-coated substrate and the whole PVD system were assembled together. In this study, 1-1.5 g of refined α-HgI2 powder of 99.99% purity was loaded into the bottom of the tube as a source. When the gas pressure of the tube was 10-5 Torr, the tube was sealed. Several parameters should be set before deposition. The temperature was divided three regions, viz. Tsource, Twall and - 48 -

Tsubstrate. Mercuric iodide underwent a solid-solid phase transition from a tetragonal red phase (α-HgI2) to an orthorhombic yellow phase (β-HgI2) at 127 °C. Thus, to obtain high-quality polycrystalline α-HgI2 thick films, the temperature of the substrate should be controlled below 127 °C. Meanwhile, α-HgI2 has high-enough vapor pressure psource (about 0.12 Torr at 120 °C) around these low temperatures; hence, α-HgI2 solid phase can be deposited efficiently from physical vapor. Typical parameters of deposition were given below. 122°C≤Tsource≤126°C, 110°C≤Tsubstrate≤115°C. Twall was a little lower than Tsource. The distance between source and substrate is 7~14cm. The time of deposition was 4~6 hours.The quality of crystal has significant effect on the properties of detector. Polycrystalline HgI2 samples with cense structure and low dislocation density (10-4cm-1) and a high theoretic conversion efficiency of 25% 1.0-1.1eV, respectively

[2]

. Its direct and indirect optical band-gap energies were 1.3-1.5eV and

[3]

, which approaches the optimal value of 1.5eV of solar cells. In addition, its

constituent elements Sn and S were abundant and non-toxic in nature compared with GaAs, CdTe and InP [4]. So, SnS was one kind of efficient, inexpensive and environment friendly solar materials. At present, there were many methods for preparing SnS thin films, such as spray pyrolysis electro-deposition

[5]

, chemical deposition

[6]

, electron beam evaporation

[7]

, two-step process

[2]

,

[1]

, thermal

evaporation [4] and so on. Among these methods, SnS thin films prepared by vacuum evaporation had some advantages, such as high purity, smooth surface, uniform, controlled thickness, and strong adhesion. Therefore, in this study, we prepared the SnS films by vacuum evaporation and investigated the properties of SnS thin films at different substrate temperatures.

2. EXPERIMENTAL Before the film deposition, the slide glass substrates were cleaned in ultrasonic equipment with acetone, anhydrous ethanol, and de-ionized water, respectively. In this experiment, the substrate temperature ranged - 71 -

from 20℃ to 200℃, raw m aterials w ere SnS w ith the purity of 99.999% ,m olybdenum boat w as used as the evaporation boat, the distance between the substrate and the source was about 20cm, and the vacuum was 2.0-2.5×10-4Pa. The crystallography properties of the samples deposited at different substrate temperatures were analyzed by a D/max-III X-ray diffraction (XRD) spectroscopy. The surface morphology of the films was characterized by a JSM—6700F scanning electron microscope (SEM), and the atomic ratio of Sn to S was calculated. The transmission spectra were taken using an ultraviolet visible spectrophotometer (UV/VIS), and the conductive type was measured by means of hot probe method. We deposited a thin layer of interdigitated silver electrode on the SnS thin films samples, and recorded their I-V characteristic curves by a KEITHLEY4200 semiconductor characteristic analyzer in order to measure the dark and photo conductivities.

3. RESULTS AND DISCUSSION 3.1 Structure and surface morphology of SnS thin films

Fig. 1.XRD patterns of SnS thin films grown at different substrate temperatures

Fig.1 shows the XRD patterns for the SnS thin films deposited at various substrate temperatures such as 27℃, 100℃, 150℃ and 200℃. T he strongest diffraction peak at about 2θ =31.8° is assigned to (111) plane of orthorhombic SnS in the pattern. The (111) plane of the film deposited at ℃ 27 has a w eak

intensity of

diffraction peak but a broad diffraction peak of diffuse scattering, which indicates that the amorphous composition is in the majority in the thin films. As the substrate temperature increases continuously, the intensity of diffraction peak increases gradually, and that of (111) plane is strongest with new ones such as (120), (101) appearing when the temperature is 150 ℃. T he reason m ay be that the increasing tem perature provides more energy to the ions as the crystalloid film grow, and the planes with higher surface energy are easier to grow in the orthorhombic structure

[8]

. However, the diffraction peaks of other planes increase

gradually with the increasing substrate temperature while (111) plane reverses, illustrating that directional growth of SnS films becomes poor when the temperature is too high. Therefore, the crystallinity of the thin films is better with the substrate temperature of 150℃.

- 72 -

(a)

(b)

Fig. 2.SEM images of SnS thin films grown at different substrate temperatures (a) 27℃ (b) 150℃

The surface morphology and composition of SnS thin films are investigated at four substrate temperatures (27℃, 100℃,150℃ and 200℃).Fig.2 (a) show s the SE M m icrograph of the film deposited at 27℃. A sm ooth and pinhole-free surface is observed, and the grains have a rod-like shape with a size of about 0.1μm. Fig. 2 (b) shows the SEM micrograph of the films deposited at 150℃ and the size of the grains about 0.5μm. Therefore, the size of the grains becomes larger as the substrate temperature increases during the film deposition. This may be caused by the lack of thermal energy which leads to the smaller grains when they nucleate and grow at low substrate temperature. The chemical composition is measured by means of characteristic X-ray energy dispersive spectrometer (EDS). The EDS analyses of the SnS thin films deposited at four substrate temperatures are listed in the table 1. The stoichiometric ratio of Sn/S approaches 1 if the substrate temperature is high enough. The stoichiometric mismatch of Sn/S is minimum when the substrate temperature is at 150 ℃, w hereas it begins to increase while the temperature further rises to 200℃. Table 1 EDS analysis of SnS thin films at different substrate temperatures

substrate temperatures（℃） 27

100

150

200

element

Weight%

Atomic%

Sn

74.18

43.70

S

25.82

56.30

Sn

74.62

44.26

S

25.38

55.74

Sn

75.58

45.54

S

24.42

54.46

Sn

74.75

44.40

S

25.25

55.60

- 73 -

3.2 Optical properties of SnS thin films 3.0x109

( α hν) 2/cm-2eV2

2.5x109

27°C

2.0x109

1.5x109

200°C

1.0x109

5.0x108

0.0 1.10 1.15 1.20 1.25 1.30 1.35 1.40 1.45 1.50 1.55 1.60 1.65 1.70 1.75 1.80

hν /eV

Fig.3. Plots of （αhν）2 vs. hν for two SnS thin films

The bandgap energy of the SnS thin films can be determined by the plots of (αhν) 2 versus hν, where α is the absorption coefficient and hν is the energy of the indent light. Fig.3 shows the plots of (αhν)2-hν of the films deposited at two different substrate temperatures, and the bandgap energies of the films corresponding to1.375eV and 1.416eV can be calculated on the basis of the intersection points between the straight lines and the cross axis. What’s more, the bandgap energies at other temperatures can be obtained by the same method, and the specific results are obtained in the table 2. Table 2 The optical band-gap of four SnS thin films

substrate temperatures(℃)

band-gap(eV)

27

1.372

100

1.375

150

1.402

200

1.416

3.3 Electrical properties of SnS thin films

The SnS thin films show p-type conductivity using the caloric probe method. A stripe layer of silver electrode is deposited on the films in order to measure the conductivity, and then the related parameters are recorded by a KEITHLEY4200 semiconductor characteristic analyzer to calculate the conductivity. The halogen tungsten lamp with a light intensity of 30mW/cm2 is used as the light source to measure the photo conductivity. The dark conductivity σd and photo conductivity σph at various substrate temperatures are obtained (shown in table 3). The dark conductivity of the films increases at first, then decreases with the substrate temperatures increasing, and it should be noticed that the conductivity is 0.01Ω-1·cm-1 at the temperature of 150℃.T his phe nom enon gives rise to tha t the crystallinity of the SnS film is be tter as w ell as - 74 -

the grain size is larger with the substrate temperatures increasing which leads to the higher conductivity, however, the preferential orientation of the film becomes worse as the temperature above 150 ℃ that causes the conductivity decreasing. The variation of the ratio of σd/σph of the films is the same to the conductivity, because the film deposited at the substrate temperature of 150℃has the better crystallinity and the larger grain size compared with the samples deposited at low temperature. Nevertheless, more photon-induced carriers will recombine because of the abundant internal defects in the films deposited at 200℃, w hich consequently reduces the ratio of σd/σph of the films. Table 3 Dark-conductivity and photo-conductivity of four SnS thin films

Substrate temperatures(℃)

σd(Ω-1·cm-1)

σph(Ω-1·cm-1)

σph/σd

27

0.00165

0.0065

3.9

100

0.0073

0.035

4.8

150

0.01

0.08

8.0

200

0.0061

0.03

4.9

4. CONCLUSIONS SnS thin films have been prepared at different substrate temperatures (27 ℃, 100℃, 150℃ and 200℃) by vacuum evaporation. Experimentally, the properties of the SnS films depend on the substrate temperatures significantly. When the substrate temperature is ℃, 150m os t of the SnS thin film s are polycrystalline with orthorhombic structure and the preferential orientation appears along the plane (111). The films have a grain size of 0.5μm, a stoichiometric ratio of Sn/S approaching 1 and a p-type conductivity. The dark conductivity and photo conductivity of the film is 0.01Ω-1·cm-1and 0.08Ω-1·cm-1, respectively. And the bandgap energy of the film is 1.402 eV.

REFERENCES 1.

K.T.Ramakrishna Reddy, P.Purandhara Reddy, P.K.Datta, et al. Formation of polycrystalline SnS layers by a two-step process [J]. Thin Solid Films, 2002, 403 –404: 116–119.

2.

N. Koteeswara Reddy, K.T. Ramakrishna Reddy. Optical behaviour of sprayed tin sulfide thin films [J]. Materials Research Bulletin, 2006, 41: 414–422.

3.

Sekhar C.Ray, Malay K.Karanjai, Dhruba DasGupta. Structure and photoconductive properties of dip-deposited SnS and SnS2 thin films and their conversion to tin dioxide by annealing in air [J]. Thin Solid Films, 1999, 350:72-78.

4.

M. Devika, N. Koteeswara Reddy, K. Ramesh, et al. Influence of substrate temperature on surface structure and electrical resistivity of the evaporated tin sulphide films [J]. Applied Surface Science, 2006.

5.

Shuying Cheng, Guonan Chen, Yanqing Chen et al. Effect of deposition potential and bath temperature on the electrodeposition of SnS film [J]. Optical Materials, 2006.

6.

Tanusevski, A. Optical and photoelectric properties of SnS thin films prepared by chemical bath - 75 -

deposition [J]. Semicond. Sci. Technol., 2003, 18(6): 505-505 7.

Tanusevski, A. Optical and photoconductive properties of SnS thin films prepared by electron beam evaporation [J]. Solar Energy Materials and Solar Cells, 2003, 80(3): 297-303

8.

Kim Shi Yul, Kim Dong Seop, Ahn Byung Tae, et al., Structural, electrical and optical properties of In-doped CdS thin films prepared by vacuum coevaporation, Thin Solid Films, Vol. 229, No. 2, 1993:227-231

- 76 -

Investigations on Sb2O3 doped-SnS thin films Prepared by vacuum evaporation Yuying Guo*a, Weimin Shi a,b, Yu Zhang a, Linjun Wang a , Guangpu Wei a a School of Materials Science and Engineering, Shanghai University, Shanghai 200072, China b SHU-SOEN’s R&D Lab., Shanghai University, Shanghai 200444, China ABSTRACT Tin sulfide (SnS) is one of promising candidate materials for low-cost thin film solar cells because of its high absorption coefficient and suitable band-gap. The aim of this paper is to study the properties of doped-SnS thin films prepared by vacuum evaporation. Sb2O3 was used as the doping source (the weight ratio of Sb2O3 to SnS in the range from 0.1% to 0.8%). And then the Sb2O3-doped SnS thin films were annealed in the hydrogen atmosphere at different temperatures and times. The structure of all the samples was characterized by X-ray diffraction (XRD). The electrical properties of SnS thin films were investigated as well. From the results, the optimum doping content of Sb2O3 was 0.2% in weight, and the resistivity of the doped-SnS film was 42Ω·cm while that of the pure-SnS film was 99Ω·cm. In addition, the film resistivity of Sb2O3-doped SnS film decreased to 24Ω·cm with the best annealing conditions of 400℃ and 3 hours. Keyword: Solar cells, Tin Sulfide, Vacuum Evaporation

1.

INTRODUCTION

Tin sulfide (SnS) is a IV–VI semiconductor with orthorhombic structure

[1]

. It has been attracted much

attention because of its potential application in the fabrication of optoelectronic and photoconductive devices. SnS has an absorption coefficient greater than 104cm-1[2]. Its direct energy band gap is approximate 1.3 1.5eVwhich is very close to the optimum value of 1.5eV. Besides, the constituent elements Sn and S of SnS are nontoxic to environment and human being. The light conversion efficiency predicted from Loferski diagrams is ~25%

[3]

. A variety of deposition techniques have been reported for the preparation of SnS thin

films, such as chemical deposition electro-deposition

[7]

, brush plating

[4]

, RF sputtering

[8]

[5]

, spray pyrolysis

, close-space vapor transport

[9]

[2]

, electron beam evaporation

, vacuum evaporation

[1]

[6]

,

and so on. In

this paper, our aim is to study the electrical properties of Sb2O3-doped SnS thin films deposited by vacuum evaporation techniques. The electric conductivity of the pure SnS films deposited by vacuum evaporation is not high enough for solar cells, which is overcame by using the doping and annealing treatment with the doping source of Sb2O3. The optimum doping content and the best annealing conditions are expected to be researched.

- 77 -

2.

EXPERIMENTAL PROCEDURES

Using SnS (99.999%) and Sb2O3-doped SnS powders as source materials in the molybdenum boats, thin films were deposited on glass substrates by vacuum evaporation. The glass substrates were rinsed sequentially by the ultrasonic in the ethanol, acetone and de-ionized water, while the molybdenum boats were washed by hydrochloric acid. The weight ratio of Sb2O3/SnS ranged from 0.1% to 0.8%. Several important preparation parameters were shown table 1.

Table1: deposition parameters of SnS thin films by deposited vacuum evaporation

Deposition parameter

Numerical value

Substrate temperature (°C)

150

Source temperature (°C)

980

Source-to-substrate distance (cm)

15

-4

Vacuum system (×10 Pa)

2.0-2.5

Then the SnS thin films with the optimum content of Sb2O3 were annealed in the hydrogen atmosphere at different temperatures and times after evaporation. The structure of all the samples was characterized by X-ray diffraction (XRD). In order to test the electrical properties of thin films, several dot Ag electrodes were deposited on the surface of thin films. The resistivities of all the SnS thin films samples were measured by a conventional SZ-82 four-probe instrument.

3.

RESULTS AND DISCUSSION

3.1 Structure of SnS thin films

All the SnS films deposited were opaque black grey with smooth surfaces which had strong adhesion to the glass substrates. 3.1.1 Structure of SnS thin films before annealing

31.88

0.4%

Intensity

Intensity

Intensity

0.4%

0.3%

0.3%

0.2%

0.2%

0.1%

0.1% 20

30

40

50

60

20

30

40

50

60

30

(a)

31

32

33

34

35

2θ

2θ

2θ

(b)

Fig.1 .XRD patterns of SnS thin films as-deposited (a) pure-SnS; (b) Sb2O3-doped SnS

- 78 -

Table2: the characteristics of dominant XRD peaks of experimental data before annealing 2θ(º)

Weight ratios of Sb2O3 Undope

31.88

0.1%

32.02

0.2%

32.06

0.3%

32.12

0.4%

32.14

Fig.1 shows the typical X-ray diffractogram of as-prepared SnS thin films and Table 2 records the dominant XRD peaks of the samples with different weight ratios of Sb2O3. It can be seen that there is a fair agreement in 2θ values corresponding to the (111) plane, which belongs to the orthorhombic SnS (JCPDS 39-354). From the results, the value of the main peak becomes bigger with the increase of the weight ratios of Sb2O3 from 0.1% to 0.4%, because Sb with the bigger ion radius takes the place of Sn. According to 2d·sinθ=λ, the values of 2θ move to the right 2θ when d changes. 3.1.2 Structure of annealed SnS thin films

Fig.2 shows the XRD pattern for the annealed SnS films deposited at various weight ratios of Sb2O3 and Table 3 shows the dominant XRD peaks of the samples. It can be seen that the interferential peaks decrease and intensity of the main peak increases. The value of the main peak of annealed SnS thin films is bigger than the as-prepared because the annealing treatment can increase the orderliness of atoms and the grain as well provide more energy to the Sb ion to enter into the SnS crystal. Therefore, more and more Sb ions take the place of Sn.

20

30

40

50

60

Intensity

Intensity

Intensity

32.22

20

2θ

30

40

0.4%

0.4%

0.3%

0.3%

0.2%

0.2%

0.1%

0.1%

50

60

31

32

33

34

2θ

2θ

(a)

(b)

Fig.2 .XRD patterns of annealed SnS thin films (a) pure-SnS; (b) Sb2O3 doped-SnS Table3: the characteristics of dominant XRD peaks of experimental data after annealing

Weight ratios of Sb2O3

2θ(º)

Undope

32.22

0.1%

32.28

0.2%

32.32

0.3%

32.34

0.4%

32.36

- 79 -

3.2 Electrical properties of SnS thin films

The composition of the sources affects the properties of the films deeply. Fig.3 shows the resistivity of Sb2O3 doped SnS thin films. The resistivity of the thin film decreases rapidly when the percentage of Sb2O3 increases from 0.1wt% to 0.2wt%. With the increase of the percentage of Sb2O3 from 0.2wt% to 0.5wt%, the film resistivity increases dramatically, because the carriers in SnS thin films are produced by Sb instead of Sn. As the concentration of Sb becomes lower, the concentration of carrier increases with the concentration of Sb, which causes the film resistivity to be descended. However, the concentration of carriers does not increase anymore when the concentration of Sb comes to the maximum that maybe create new compound with high resistivity, for example Sb2S3. The resistivity of thin films begins to decline when the percentage of Sb2O3 is more than 0.5wt%, which maybe caused by the appearance of the dissociative Sb [10]. This can be presumed that the optimum percentage of Sb2O3 is about 0.2wt% in weight for getting low-resistivity SnS material. 36

29

34

28

100 90

32

ρ( Ω

˙cm

70 60

27 ρ(Ω·cm)

ρ(Ω·cm)

)

80

30

26

28

25

50

26 40

24

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

24 340

360

380

Sb2O3 wt.%

400

420

440

T

(a) Fig.3 Resistivities of SnS: Sb2O3 thin films as a function of the doping

1.0

1.5

2.0

2.5

3.0

3.5

4.0

t(h)

(b)

Fig.4 The resistivity of Sb2O3-doped SnS thin films as a function of annealing condition (a) temperature (b) time

level of Sb2O3

Some 0.2wt% Sb2O3-doped SnS thin films were annealed in the hydrogen atmosphere at different temperatures and times. Both annealing temperature and annealing time will influence the resistivity of Sb2O3-doped SnS films deeply. Fig.4 (a) shows the resistivity of Sb2O3-doped SnS films as a function of annealing temperature with the annealing time of 2 hours. With the increasing of annealing temperature from 340℃ to 400℃, the resistivity of SnS thin film s decrease rapidly. T he higher resistivity of the films formed at lower annealing temperatures, is due to smaller grain size. At lower annealing temperatures the grains can not grow sufficiently large so that the inter-crystalline regions are wider and hence they offer high resisitivity to the movement of charge carries. With the increasing of annealing temperature, the grains grow larger and the inter-crystalline regions become smaller. So the resisitivity of films decrease. However, the resistivity increases dramatically when the annealing temperatures increase from 400℃ to 440℃ w ith som e yellow ish things forming on the surface of the sample. It maybe more K+, Na+ of glass diffuse into the films, induce to reverse doping effect and the concentration of carries are reduced. Therefore, the optimum annealing temperature should be about 400℃. Fig.4 (b) shows the resistivity of Sb2O3-doped SnS films as a function of annealing time with the - 80 -

annealing temperature of 400℃. W ith the increasing of annealing tim e from 1.0 to 3.0 hours, the resistivity of Sb2O3-doped SnS films decrease rapidly. Because prolong annealing time, crystal structure of films become perfect and electric property of samples get well. However, the resistivity increases dramatically when the annealing time increases from 3.0 to 4.0 hours. It mirth be some other binary phases are presence or remains oxygen and samples produce tin oxidation. So it can be seen clearly that the optimum annealing time should be about 3.0 hours.

4.

CONCLUSIONS

We investigated the properties of Sb2O3-doped SnS thin films by vacuum evaporation techniques. The electrical properties of the films depend on the doping content, annealing temperature and time with the optimum values of about 0.2wt%, 400 ℃, 3.0 hours respectively, and the resistivities of Sb 2O3-doped SnS thin films was 24Ω·cm.

REFERENCES 1.

M.M. El-Nahass, H.M. Zeyada, M.S. Aziz, et al. Optical properties of thermally evaporated SnS thin films [J]. Optical Materials, 2002, 20:159-170.

2.

K.T. Ramakrishna Reddya, N. Koteswara Reddya, R.W. Miles. Photovoltaic properties of SnS based solar cells [J]. Solar Energy Materials and Solar Cells, 2006, 90: 3041–3046.

3.

N. Koteeswara Reddy, K.T. Ramakrishna Reddy. Optical behaviour of sprayed tin sulfide thin films [J]. Materials Research Bulletin, 2006, 41: 414–422.

4.

M. Ristov, G. Sinadinovski, M. Mitreski, et al. Photovoltaic cells based on chemically deposited p-type SnS [J]. Solar Energy Materials & Solar Cells, 2001, 69:17-24.

5.

Guang-pu, Wei. Investigation on SnS film by RF sputtering for photovoltaic application [J]. Conference Record of the IEEE Photovoltaic Specialists Conference, 1st World Conf. on Photovoltaic Energy Conversion, 1994, 1: 365-368.

6.

Tanusevski, A.

Optical and photoconductive properties of SnS thin films prepared by electron beam

evaporation [J]. Solar Energy Materials and Solar Cells, 2003, 80(3): 297-303. 7.

Nair, M.T.S.

Simplified chemical deposition technique for good quality SnS thin films [J].

Semiconductor Science and Technology, 1991, 6(2): 132-134. 8.

Jayachandran, M. Studies on the brush plated SnS thin films [J]. Journal of Materials Science Letters, 2001, 20(4): 381-383.

9.

Yanuar, F. SnS thin films grown by close-spaced vapor transport [J]. Journal of Materials Science Letters, 2000, 19(23): 2135-2137.

10. A.E. Taverner, A. Gulino, R.G. Egdell, et al. A photoemission study of electron states in Sb-ion implanted TiO2 (110) [J]. Applied Surface Science, 1995, 90: 383-387.

- 81 -

Electrical Properties of Doped SnS Thin Films Prepared by Vacuum Evaporation GUO Yu-ying, SHI Wei-min, WEI Guang-pu (School of Materials Science and Engineering, Shanghai University, Shanghai 200072, China) Abstract：Tin sulfide (SnS) has received much attention because of its high absorption coefficient, suitable band-gap and little toxicity. The resistivity of pure-SnS thin film deposited by vacuum evaporation is too high to make solar cell. In order to solve this problem, doped-SnS thin films were fabricated. Sb, Sb2O3, Se, Te, In, In2O3, Se and In2O3 were used as dopant sources. The thickness and conductance of various doped-SnS thin films were measured, and then the resistivities and the ratio of photo-conductivity to dark-conductivity (Gphoto/Gdark) of these films were calculated. From the experimental results, Sb is the best dopant source. The resistivity of Sb doped-SnS thin film is reduced by four orders of magnitude and the value of Gphoto/Gdark is double. In addition, the influence of Sb doping content on the electrical properties of doped-SnS thin films was also investigated, and the optimum doping content of Sb is 1.3%~1.5% in weight. Keywords: Tin Sulfide; Solar cells; Doped; Vacuum Evaporation; Resistivity

1 Introduction Tin sulfide (SnS) is a IV–VI group semiconductor with orthorhombic structure [1, 2, 3].It is one of promising candidate materials for low-cast thin film solar cells because of its high absorption coefficient and suitable band-gap. SnS has an absorption coefficient larger than 104cm-1[4]. Its direct band-gap is approximate 1.3~1.5ev and it is very close to the optimum value of 1.5ev [5]. Besides, the constituent elements of SnS have quite little toxicity to environment and human being [6]. The light conversion efficiency predicted from Loferski diagrams is ~25% [5].A variety of deposition techniques have been reported for the preparation of SnS thin films, such as RF sputtering[7], spray pyrolysis[4, 5, 8], chemical deposition[9, 10], electro-deposition[11,12], electron beam evaporation[13], two-step process[1], close-spaced vapor transport[14], brush plated[15], vacuum evaporation[3, 6] and others. However, vacuum evaporation techniques have not yet been used for doped-SnS thin films. In this paper, our aim is to study the electrical properties of doped-SnS thin films by vacuum evaporation techniques. The pure SnS films deposited by vacuum evaporation do not possess conductivity sufficient for solar cells. In order to overcome this problem, doping is used. The best dopant source and the optimum doping content are expected to be found.

2 Experimental Procedures Using SnS (99.999%) and doped-SnS powders as source material in the molybdenum boats, thin films were deposited on glass substrates by vacuum evaporation. The glass substrates were rinsed sequentially by the ultrasonic in the ethanol, acetone and deionized water, while the molybdenum boats were washed by - 82 -

hydrochloric acid. The different doped element (Sb, Sb2O3, Se, Te, In, In2O3, Se and In2O3) and SnS weight ratio also was 1.25%. Than different Sb/SnS weight ratios （0.1wt%～12.5wt%）for dope application were investigated. Several important preparation parameters were shown table1. Table 1 Deposition parameters of SnS thin films on glass substrate by vacuum evaporation Deposition parameters

numerical value

Substrate temperature (°C)

150

Source temperature (°C)

980

Source-to-substrate distance (cm) Vacuum system (×10-4Pa)

15 2.0~2.5

We measured all SnS thin films, definding the thickness by weight method. In order to test the electrical properties of thin films, several spacing of 1mm, 1.5mm-wide, 1cm-long Ag electrodes were evaporated on two sides of all the samples (Fig.1). The dark current and photocurrent were measured by a KEITHLEY4200I-V recorder. The resistivities of doped-SnS thin films and the value of Gphoto/Gdark ratio were calculated by related formula.

Fig.1 sketch map of sample for electrical test

3 results and discussion 3.1 Resistivity of different doped-SnS thin films Resistance of pure SnS and doped SnS thin films were calculated according to the related formula （ρ =

a×d × R ）. Here R , d , a , l represent the resistance, thickness of the pure-SnS or doped-SnS l

thin films, the length of the pure-SnS or doped-SnS thin films, spacing of the parallel strip Ag electrodes, respectively. In order to compare the resistivity of pure-SnS and doped-SnS, we calculated the resistivity of all SnS thin films, the results as follow table2. From table2, the resistivities of Sb doped-SnS and Sb2O3 doped-SnS thin films were reduced by three and four orders of magnitude. Because atom state Sb or ion state Sb , which were brought from Sb or Sb2O3 by vacuum evaporation, and Sn, S deposited doped- SnS. In the course, some Sb instead of the place of Sn,

- 83 -

while offering a free electron. So the concentration of carriers increased and the resistivities descended

[16]

.

When Se、Te、In、In2O3 doped SnS thin films, the resistivities of SnS thin films raised. Especially, the resisitivity of In doped-SnS rose for three orders of magnitude. It was because Sb instead of the place of Sn and capture more free electron or maybe new compound created. Table 2 the influence on the resistivity of doped-SnS thin films

resistivity（Ω·cm）

dopant

2.4×102

undope

2.6×10-2

Sb Sb2O3

3.4×10-1

Se

3.6×103

Te

1.4×103

In

1.6×105

In2O3

3.9×103

Se and In2O3

3.2×10

3.2 Gphoto/Gdark of Sb doped-SnS and Sb2O3 doped-SnS thin films In the interest of making out the effecter dopant, we compared the photocurrent with the dark current of Sb doped-SnS and Sb2O3doped- SnS thin films samples. The results were shown in Fig.2 to Fig.4. Both photocurrent (30mW/cm2) and dark current were tested by a KEITHLEY4200I-V recorder.

1.5x10-5

Current/A

1.0x10-5

pure tin sulfide

5.0x10-6 0.0 -5.0x10-6

Gph/Gdark=1.28

-5

-1.0x10

-1.5x10-5 -6

photo current dark current

-4

-2

0 2 Voltage/V

4

6

Fig.2 Characterization of photo current and dark current vs. voltage for pure tin sulfide thin films

- 84 -

8.0x10-3

Sb-doped

current/A

4.0x10-3 0.0 Gphoto/Gdark=2.15~2.25

-4.0x10-3

photo current dark current

-3

-8.0x10 -0.6

-0.4

-0.2

0.0 0.2 Voltage/V

0.4

0.6

Fig.3 Characterization of photo current and dark current vs. voltage for Sb-doped tin sulfide thin films

8.0x10-3 6.0x10-3

Sb2O3-doped

-3

Current/A

4.0x10

2.0x10-3 0.0 -2.0x10-3

Gphoto/Gdark=1.02~1.04

-3

-4.0x10

photo current dark current

-3

-6.0x10

-6

-4

-2

0 2 Voltage/V

4

6

Fig.4 Characterization of photo current and dark current vs. voltage for Sb2O3-doped tin sulfide thin films

These figures show that the value of Gphoto/Gdark was double when SnS thin films doped Sb. Otherwise when Sb2O3 as dopant, the value of Gphoto/Gdark dropped. It was conferred that Sb was optimal dopant for fabricating low-resistivity and high-Gphoto/Gdark tin sulfide thin films.

Electrical Properties of Sb doped-SnS thin films The composition of the sources affects the properties of the films deeply. Fig.4 shows the resistivities of Sb doped SnS thin films. The resistivities of thin films decreased rapidly when the contents of Sb were increased from 0.1 wt % to 1.3 wt %. With increase in the contents of Sb from 1.3 wt % to 2.5wt%, the resistivities of thin films increased dramatically. Because the carriers of SnS thin films are produced by Sb instead of the place of Sn. When the concentration of Sb is lower, the concentration of carriers increased with the concentration of Sb. So the resistivities of thin films were descended. But when the concentration of Sb arrive peak, the concentration of carriers does not increase anymore. It maybe create new high resistivity - 85 -

compound, for example Sb2S3[16]. When the content of Sb is more than 2.5%, the resistivity of thin films were declined. Because maybe create dissociative Sb[16]. It can be presumed that the optimum content of Sb is about 1.3% in weight for getting low-resistivity SnS material.

0.4

ρ/Ω•cm

0.3

0.2

0.1

0.0

0

2

4

6

8

10

12

14

Sb wt.%

Fig.5 Resistivities of SnS: Sb thin films as function of Sb content

4.5 4.0

Gphoto/Gdark

3.5 3.0 2.5 2.0 1.5 1.0 0

2

4

6

8

10

12

14

Sb wt.%

Fig.6 The value of Gphoto/Gdark ratio as a function of Sb content

Fig.5 shows the value of Gphoto/Gdark of all the different Sb/SnS weight ratio samples. When the contents of Sb were increased from 0.1 wt% to 1.5wt%, the value of Gphoto/Gdark increased very fast with the initial value of 1.28. The maximum value of Gphoto/Gdark is about 4.25. But, when the contents of Sb were increased from 1.5% to 12.5% in weight sequentially, the value of Gphoto/Gdark fell to 1.02 dramatically. It can be seen clearly the optimum content of Sb is about 1.5wt. %.

4 conclusions Doped tin sulfide thin films were deposited by vacuum evaporation technique. We characterized the electrical properties of all SnS samples, including pure SnS and doped SnS thin films. The resistivities and the Gphoto/Gdark of tin sulfide depended on doped content, including the type of doped elements and - 86 -

different weight ratios. From the results, Sb is the optimal dopant. The weight ratio from 1.3 to 1.5% is the optimum content for low-resisitivities and high- Gphoto/Gdark SnS thin films.

References [1] K.T.Ramakrishna Reddy, P.Purandhara Reddy, P.K.Datta, et al. Formation of polycrystalline SnS layers by a two-step process [J]. Thin Solid Films, 2002, 403 –404: 116–119. [2] L Ehm, K Knorr, P Dera. Pressure-induced structural phase transition in the IV-VI semiconductor SnS [J]. Journal of Physics Condensed Matter, 2004, 16: 3545-3554. [3] M.M. El-Nahass, H.M. Zeyada, M.S. Aziz, et al. Optical properties of thermally evaporated SnS thin films [J]. Optical Materials, 2002, 20:159-170 [4] K.T. Ramakrishna Reddya, N. Koteswara Reddya, R.W. Miles. Photovoltaic properties of SnS based solar cells [J]. Solar Energy Materials and Solar Cells, 2006, 90 : 3041–3046. [5] N. Koteeswara Reddy, K.T. Ramakrishna Reddy. Optical behaviour of sprayed tin sulfide thin films [J]. Materials Research Bulletin, 2006, 41: 414–422. [6] M. Devika, N. Koteeswara Reddy, K. Ramesh, et al. Influence of substrate temperature on surface structure and electrical resistivity of the evaporated tin sulphide films [J]. Applied Surface Science, 2006. [7] Guang-pu, Wei. Investigation on SnS film by RF sputtering for photovoltaic application[J]. Conference Record of the IEEE Photovoltaic Specialists Conference, 1st World Conf. on Photovoltaic Energy Conversion, 1994, 1: 365-368 [8] N. Koteeswara Reddy, K.T. Ramakrishna Reddy. Growth of polycrystalline SnS films by spray pyrolysis [J]. Thin Solid Films, 1998, 325:4-6. [9] Tanusevski, A. Optical and photoelectric properties of SnS thin films prepared by chemical bath deposition[J]. Semicond. Sci. Technol., 2003, 18(6): 505-505 [10] M. Ristov, G. Sinadinovski, M. Mitreski, et al. Photovoltaic cells based on chemically deposited p-type SnS [J]. Solar Energy Materials & Solar Cells,2001, 69:17-24. [11] Shuying Cheng, Guonan Chen, Yanqing Chen et al. Effect of deposition potential and bath temperature on the electrodeposition of SnS film [J]. Optical Materials, 2006. [12] Nair, M.T.S. Simplified chemical deposition technique for good quality SnS thin films[J]. Semiconductor Science and Technology, 1991, 6(2): 132-134 [13] Tanusevski, A. Optical and photoconductive properties of SnS thin films prepared by electron beam evaporation[J]. Solar Energy Materials and Solar Cells, 2003, 80(3): 297-303 [14] Yanuar, F. SnS thin films grown by close-spaced vapor transport[J]. Journal of Materials Science Letters, 2000, 19(23): 2135-2137 [15] Jayachandran, M. Studies on the brush plated SnS thin films[J]. Journal of Materials Science Letters, 2001, 20(4): 381-383 [16]A.E. Taverner, A. Gulino, R.G. Egdell, et al. A photoemission study of electron states in Sb-ion implanted TiO2 (110) [J]. Applied Surface Science, 1995, 90: 383-387.

- 87 -

掺杂 SnS 薄膜的制备及电学性能的研究 郭余英，史伟民，魏光普，邱永华，夏义本 （上海大学材料科学与工程学院电子信息材料系,上海200072） 摘要：硫化锡(SnS)具有很高的光吸收系数和合适的禁带宽度，又无毒性，因此在太阳电池等光电器 件中具有潜在应用价值。本文用真空蒸发法制备掺杂的 SnS 薄膜，掺杂源有 Sb、 Sb2O3、 Se、 Te、 In、 In2O3、Se 和 In2O3 的混合物。对各种掺杂 SnS 薄膜的厚度、电流－电压（I-V）特性等进行了表 征，并计算了其电阻率和光电导与暗电导的比值（Gphoto/Gdark）。结果表明较有效的掺杂源是 Sb，Sb 掺杂的薄膜电阻率比纯薄膜的电阻率降低四个数量级，Gphoto/Gdark 增加约一倍。同时，研究了 Sb 掺杂 量对 SnS 薄膜电学性能的影响，表明 Sb 的最佳掺入量约为 1.3wt％～1.5wt％。 关键词：硫化锡；太阳电池；掺杂；真空蒸发；电阻率 中图分类号: TN304

文献标识码：A

Electrical Properties of Doped SnS Thin Films Prepared by Vacuum Evaporation GUO Yu-ying, SHI Wei-min, WEI Guang-pu, QIU Yong-hua, XIA Yi-ben (School of Materials Science and Engineering, Shanghai University, Shanghai 200072, China) Abstract：Tin sulfide (SnS) has received much attention because of its high absorption coefficient, suitable band-gap and little toxicity. The resistivity of pure-SnS thin film deposited by vacuum evaporation is too high to make solar cell. In order to solve this problem, doped-SnS thin films were fabricated. Sb, Sb2O3, Se, Te, In, In2O3, Se and In2O3 were used as dopant sources. The thickness and conductance of various doped-SnS thin films were measured, and then the resistivities and the ratio of photo-conductivity to dark-conductivity (Gphoto/Gdark) of these films were calculated. From the experimental results, Sb is the best dopant source. The resistivity of Sb doped-SnS thin film is reduced by four orders of magnitude and the value of Gphoto/Gdark is double. In addition, the influence of Sb doping content on the electrical properties of doped-SnS thin films was also investigated, and the optimum doping content of Sb is 1.3%~1.5% in weight. Key word: Tin Sulfide; Solar cells; Doped; Vacuum Evaporation; Resistivity

- 88 -

1 引言 硫化锡（SnS）具有高的光吸收系数和合适的禁带宽度，而且资源丰富又无毒性[1,2]，因此对制备 低成本、大面积的薄膜太阳电池来说，是一种有前途的半导体材料。SnS 是Ⅳ—Ⅵ族斜方晶体结构的 半导体[2,3,4]，其光吸收系数大于 104cm-1[5]，光学直接带隙和间接带隙宽度分别为 1.3～ 1.5eV[6]和 1.0～ 1.1eV[7]，与太阳辐射中的可见光有很好的光谱匹配，非常适合做太阳电池的吸收层。一般用真空蒸发 法[8]制备的纯 SnS 薄膜的电阻率对太阳电池的吸收层来说太大，为了解决这个问题，本文对 SnS 进行 了掺杂试验，以期找出合适的掺杂源和掺杂量。 2 实验 薄膜沉积之前，载玻片衬底分别用丙酮、无水乙醇、去离子水在超声中清洗并烘干。蒸发原料为 99.999%的SnS，蒸发舟为钼舟，各掺杂物质（Sb、Sb2O3、Se、Te、In、In2O3、Se和In2O3的混合物） 和SnS的重量百分比均为1.25％。另外，还研究了不同的Sb掺杂量（0.1wt%～12.5wt%）对SnS薄膜的 电学性能的影响。蒸发过程中主要工艺参数如表1所示。 表1 真空蒸发法制备SnS薄膜的工艺参数 Table 1 Deposition parameters of SnS thin films on glass substrate by vacuum evaporation 工艺参数

数值

衬底温度( C)

15 0

蒸发源温度( C)

98 0

衬底与蒸发源的距离(cm) 真空度( 10 Pa) -4

15 2.0～2.5

采用称重法测量薄膜的厚度（用探针测厚仪进行校对）。为了测量薄膜的电学特性，对要测量的 样品，采用掩膜方法在其衬底表面、薄膜旁边用真空蒸发法制作间距为 1mm、宽度为 1.5mm、长度 为 1cm 的平行条状银电极，作为测量电极，如图 1 所示。用 Keithley4200 半导体特性分析仪，测量无 光照和光照条件下的电流-电压（I-V）特性。

Fig.1 Sketch map of sample for electrical test 图 1 电学测试样品示意图 3 结果与讨论 3．1 各种掺杂剂对 SnS 薄膜电阻率的影响 根据条状银电极的间距（ l ），长度（ a ），薄膜及电极的厚度（均为 d ），可得薄膜的电阻率 （ρ =

a×d × R ） ，如 表 2 所示。 l

- 89 -

表 2 不同掺杂源的 SnS 薄膜的电阻率 Table 2 Resistivity of doped-SnS thin films 掺杂源

电阻率（Ω·cm）

未掺杂

2.4×102

Sb

2.6×10-2

Sb2O3

3.4×10-1

Se

3.6×103

Te

1.4×103

In

1.6×105

In2O3

3.9×103

Se 与 In2O3 混合物

3.2×103

从上表可以看出：在 SnS 中掺入重量比为 1.25%的 Sb 和 Sb2O3 能使薄膜的电阻率降低三到四个 数量级。其原因可能是，不论 Sb 或 Sb2O3 经真空蒸发后都产生了原子态或离子态的 Sb，它们和 Sn、 S 等原子共同沉积形成掺杂的 SnS 薄膜，在这个过程中一些Ⅴ族的 Sb 占据了Ⅳ族的 Sn 的位置，同时 提供一个自由电子，使薄膜中的载流子浓度上升，电阻率下降[9]。而当掺入 Se、Te、In 和 In2O3 时， 都导致 SnS 薄膜的电阻率有不同程度的升高，特别是 In 的掺入使电阻率升高了三个数量级，这很可 能是 In 替代了 Sn 位置，捕获了晶格中更多的传导电子。 3．2 Sb 掺杂对 SnS 薄膜 Gphoto/Gdark 的影响 掺杂 Sb 或 Sb2O3 的 SnS 薄膜样品测得了有光照和无光照时的 I—V 特性，并根据薄膜厚度等参数 计算出光电导和暗电导。图 2 到图 4 分别给出了薄膜在无光照和有光照条件下的 I—V 特性曲线，其 中实线表示无光照时的 I—V 特性曲线，而虚线表示有光照的情况，光照强度为 30mW/cm2。

1.5x10-5

Current/A

1.0x10-5

pure tin sulfide

5.0x10-6 0.0 -5.0x10-6

Gph/Gdark=1.28

-1.0x10-5 -1.5x10-5 -6

photo current dark current

-4

-2

0 2 Voltage/V

4

6

Fig.2 Photo current and dark current vs. voltage for pure tin sulfide thin films 图 2 SnS 薄膜在有光照和无光照条件下的 I-V 特性曲线

- 90 -

8.0x10-3

Sb-doped

current/A

4.0x10-3 0.0 Gphoto/Gdark=2.15~2.25

-4.0x10-3

photo current dark current

-3

-8.0x10 -0.6

-0.4

-0.2

0.0 0.2 Voltage/V

0.4

0.6

Fig.3 Photo current and dark current vs. voltage for Sb-doped tin sulfide thin films 图 3 掺 Sb 的 SnS 薄膜在有光照和无光照条件下的 I-V 特性曲线

8.0x10-3 6.0x10-3

Sb2O3-doped

Current/A

4.0x10-3 2.0x10-3 0.0 -2.0x10-3

Gphoto/Gdark=1.02~1.04

-3

-4.0x10

photo current dark current

-3

-6.0x10

-6

-4

-2

0 2 Voltage/V

4

6

Fig.4 Photo current and dark current vs. voltage for Sb2O3-doped tin sulfide thin films 图 4 掺 Sb2O3 的 SnS 薄膜在有光照和无光照条件下的 I-V 特性曲线 从图中可以看出：掺 Sb 的 SnS 薄膜 Gphoto/Gdark 相对于纯 SnS 薄膜有一定程度的增大，大约为一 倍；而掺 Sb2O3 的 SnS 薄膜则光电导很小。这足以说明如果需要制备低电阻率和高光电导特性的 SnS 薄膜，Sb 是相对有效的理想掺杂剂。 3．3 Sb 掺杂量对 SnS 薄膜电学特性的影响 掺杂剂的掺入量对薄膜性能有着很大的影响，图 5 表示 Sb 的不同掺入量与 SnS 薄膜电阻率的关 系曲线。当 Sb 掺杂的重量百分比低于 1.3%时，电阻率随掺杂量的增加而减小；当掺入量大于 1.3% 小于 2.5%时，电阻率随掺入量的增加而增加。这可能是因为 SnS 中的载流子主要为Ⅴ族的 Sb 代替了 Ⅳ族的 Sn 占据的位置而产生，因而当 Sb 浓度较低时，提高 Sb 的浓度，则载流子浓度增加，电阻率 降低。但当 Sb 浓度达到一定峰值后，再增加 Sb 浓度时，可能产生 Sb2S3 等电阻率比较高的第二相物 质，导致薄膜的电阻率上升[9,10]。当掺入量大于 2.5%后，可能产生游离的金属 Sb，使电阻率又渐渐 变小，直到接近于零。因此，从电阻率考虑，Sb 的最佳掺入量为 1.3 wt%。

- 91 -

0.4

ρ/Ω•cm

0.3

0.2

0.1

0.0

0

2

4

6

8

10

12

14

Sb wt.%

Fig.5 Resistivities of SnS: Sb thin films as a function of Sb content 图 5 不同的掺 Sb 量对薄膜电阻率的影响 4.5 4.0

Gphoto/Gdark

3.5 3.0 2.5 2.0 1.5 1.0 0

2

4

6

8

10

12

14

Sb wt.%

Fig.6 Gphoto/Gdark ratio as a function of Sb content 图 6 不同的掺 Sb 量对 Gphoto/Gdark 的影响 为了比较不同的 Sb 掺入量对薄膜 Gphoto/Gdark 的影响，计算出各个不同掺 Sb 量的薄膜的 Gphoto/Gdark，其结果如图 6 所示。当掺杂量低于 1.5wt%时，Gphoto/Gdark 随掺杂量的增加而增加，这说 明在 SnS 薄膜中进行适量的 Sb 掺杂很容易在光照条件下产生更多的电子空穴对，增加载流子浓度， 更有利于制作太阳电池的吸收层。当掺杂量高于 1.5wt%时，Gphoto/Gdark 随着掺杂量的增加而降低，直 到接近于 1，这也进一步表明当 Sb 掺杂浓度过高时，SnS 中可能出现游离的金属 Sb[10]，其电学性质 接近于金属导体。由此可见，SnS 用作太阳能电池吸收层时，Sb 的最佳掺入量为 1.5 wt%。 4 结论 本文采用真空蒸发技术制备纯 SnS 薄膜和掺杂 SnS 薄膜，测量各种 SnS 薄膜的电学特性。实验 表明 SnS 薄膜的电阻率和 Gphoto/Gdark 与掺杂剂种类和掺杂量有显著的关系，得出较有效的掺杂源是 Sb。Sb 掺杂的 SnS 薄膜的电阻率比纯 SnS 薄膜的电阻率降低四个数量级，Gphoto/Gdark 增加大约一倍。 此外，对低电阻、高 Gphoto/Gdark 的太阳电池的吸收层来说，Sb 的最佳掺入量为 1.3wt％～1.5wt％。 参考文献 [1] Naoya Sato, Masaya Ichimura, Eisuke Arai, et al. Characterization of electrical properties of SnS thin films prepared by the electrochemical deposition method [J]. Solar Energy Material and Solar cells, 2005, - 92 -

85:153-165. [2] K.T.Ramakrishna Reddy, P.Purandhara Reddy, P.K.Datta, et al. Formation of polycrystalline SnS layers by a two-step process [J]. Thin Solid Films, 2002, 403 –404: 116–119. [3]L Ehm, K Knorr, P Dera. Pressure-induced structural phase transition in the IV-VI semiconductor SnS [J]. Journal of Physics Condensed Matter, 2004, 16: 3545-3554. [4]M.M. El-Nahass, H.M. Zeyada, M.S. Aziz, et al. Optical properties of thermally evaporated SnS thin films [J]. Optical Materials, 2002, 20:159-170 [5]K.T. Ramakrishna Reddya, N. Koteswara Reddya, R.W. Miles. Photovoltaic properties of SnS based solar cells [J]. Solar Energy Materials and Solar Cells, 2006, 90 : 3041–3046. [6] N. Koteeswara Reddy, K.T. Ramakrishna Reddy. Optical behaviour of sprayed tin sulfide thin films [J]. Materials Research Bulletin, 2006, 41: 414–422. [7]Sekhar C.Ray, Malay K.Karanjai, Dhruba DasGupta. Structure and photoconductive properties of dip-deposited SnS and SnS2 thin films and their conversion to tin dioxide by annealing in air [J]. Thin Solid Films, 1999, 350:72-78. [8] Hidenori Noguchi, Agus Setiyadi, Hiromasa Tanamura et al. Characterization of vacuum-evaporated tin sulfide film for solar cell materials [J].Solar Energy Materials and Solar Cells, 1994, 35:325- 331 [9]A.E. Taverner, A. Gulino, R.G. Egdell, et al. A photoemission study of electron states in Sb-ion implanted TiO2 (110) [J]. Applied Surface Science, 1995, 90: 383-387. [10]谢莲革，汪建勋，沈鸽等. 基板温度对 SnO2:Sb 薄膜结构和性能的影响[J]. 功能材料，2005，36： 415-418

- 93 -

衬底温度对 SnS 薄膜性能的影响 郭余英，史伟民，魏光普，夏义本 （上海大学材料科学与工程学院电子信息材料系，上海，200072） 摘要：采用真空蒸发法在玻璃衬底上沉积硫化亚锡（SnS）薄膜，并对不同衬底温度沉积的薄膜性能 进行了探讨。对薄膜的结构、表面形貌、成分、电学特性和光学特性进行了表征。实验发现，最佳的 衬底温度为 150℃；制备 的 SnS 薄膜为多晶的斜方晶系，晶粒大小约为 0.5μm，Sn 和 S 元素的化学计 量比接近 1，导电类型为 P 型，暗电导率、光电导率分别为 0.01Ω-1·cm-1 和 0.08Ω-1·cm-1，禁带宽度为 1.402 eV。 关键词：太阳能电池；硫化亚锡（SnS）薄膜；真空蒸发；衬底温度 中图分类号: TN304

文献标识码：A

Influence of Substrate Temperature on Properties of Tin Sulfide Thin Films GUO Yu-ying, SHI Wei-min, WEI Guang-pu, XIA Yi-ben (School of Materials Science and Engineering, Shanghai University, Shanghai 200072, China) Abstract: Thin films of tin sulfide （SnS）have been deposited by vacuum evaporation at different substrate temperatures. The samples have been characterized with XRD and SEM for structural analysis. The electrical and optical properties of SnS thin films have been investigated also. From the results, 150 ℃ w as the best substrate temperature. When the substrate temperature

was 150℃, the SnS thin film s w as

orthorhombic structure with crystallite size of 0.5μm and the ratio of Sn to S is about 1. Hot probe method showed p-type nature for the deposited films. Dark-conductivity and photo-conductivity were 0.01Ω-1·cm-1 和 0.08Ω-1·cm-1, respectively. The optical energy band gap of the films was 1.402 eV. Key word: Solar cells; Tin Sulfide（SnS）thin films; Vacuum Evaporation; substrate temperature 1 引 言 随着人们环保意识的增强，太阳能的开发应用越来越受到人们的关注，廉价、高效的太阳能转换 材料的研制也越来越受到人们的重视。近些年来，硫化亚锡（SnS）薄膜作为太阳能电池的吸收层材 料的研制已引起材料科学界的广泛注意。SnS 是Ⅳ－Ⅵ族斜方晶体结构的半导体[1,2,3]，其光吸收系数 大于 104cm-1[4]；在理论上其能量转换效率比较高，达到 25％[5]；光学直接带隙和间接带隙宽度分别为 1.3～1.5eV 和 1.0～1.1eV[6]，接近太阳电池的最佳禁带宽度 1.5 eV；组成元素 Sn 和 S 在自然界中含量 丰富，与 GaTe、CdTe,、InP 相比又无毒性[7]。因此，SnS 是一种高效、廉价、环保型的太阳能转换材 料。 目前制备 SnS 薄膜的方法有喷雾高温分解法[4,5,8]、化学沉积法[9,10]、电沉积法[11,12]、电子束蒸发 法[13]、射频溅射法[14]、两步法[1]、近空间气相传输法[15]、刷镀法[16]、热蒸发法[3,7]等等。在这些方法 中，真空蒸发法制备的薄膜具有纯度高、表面光滑、厚度均匀可控、附着力强等优点，因此本实验采 - 94 -

用真空蒸发法制备 SnS 薄膜，并对不同衬底温度沉积的 SnS 薄膜进行了比较研究。 2 实 验 薄膜沉积之前，载玻片衬底分别用丙酮、无水乙醇、去离子水在超声中清洗并烘干。衬底温度范 围从 27℃到 200℃，蒸发原料为 99.999%的 SnS，蒸发舟为钼舟，衬底与蒸发源的距离为 20cm, 真空 度 2.0～2.5×10-4Pa。 在不同衬底温度下制备的样品薄膜用 D/max-III 型 X 射线衍射仪对薄膜的结晶学特性进行分析， 用 JSM—6700F 场发射扫描电镜对薄膜表面形貌和薄膜中的硫与锡原子比进行表征，用紫外-可见分光 光度计测量其透射光谱，用热探针法测定薄膜的导电类型。为了测定薄膜光电导和暗电导，我们在 SnS 薄膜样品上蒸一层薄的梳状银电极， 用 KEITHLEY4200 半导体特性分析仪记录其 I—V 特性曲线。 3 结果与讨论 3.1 SnS 薄膜的结构和表面形貌

图 1 不同衬底温度的 SnS 薄膜的 XRD 图 Fig. 1.XRD pictures of SnS thin films grown at different substrate temperatures 图 1 为衬底温度分别为 27℃、100℃、150℃和 200℃情况下沉积得到的 SnS 薄膜的 XRD 图。所 有的 XRD 图中在 2θ＝31.8° 附近处都有一个最强的衍射峰，对应斜方相 SnS 的（111）晶面。从图中 可以看出在衬底温度在 27℃下沉积得到的薄膜， （111）晶面的衍射峰强度很弱，而漫散射的衍射峰占 据很宽的位置，可见薄膜中非晶态成分较多。随着衬底温度的不断升高，晶面的衍射峰强度逐渐增强。 当衬底温度为 150℃时， （111）晶面的衍射峰强度最强，并出现了（120） 、 （101）等晶面的新的衍射 峰。这可能是因为随着衬底温度的升高，薄膜晶体生长时离子具有更多能量，斜方相中表面能更高的 晶面更容易生长[17]。而随着衬底温度的继续升高，其它晶面的衍射峰逐渐增强，而（111）晶面的衍 射峰逐渐减弱。说明衬底温度太高，SnS 薄膜的定向生长变差。所以衬底温度为 150℃时，薄膜的结 晶性较好。

- 95 -

(a)

(b) 图 2 不同衬底温度的 SnS 薄膜的 SEM 图:(a) 27℃、(b) 150℃

Fig. 2.SEM pictures of SnS thin films grown at different substrate temperatures (a) 27℃ (b) 150℃ 研究了四种不同衬底温度（27℃，100℃，150℃和 200℃）下的 SnS 薄膜的表面形貌和成分。图 2（a）为衬底温度 27℃时 SnS 薄膜的 SEM 图，从图中可以看出薄膜表面均匀光滑，无针孔，晶粒呈 针棒状，大小约为 0.1μm。图 2（b）是衬底温度为 150℃薄膜的 SEM 图，晶粒大小约为 0.5μm。因此 随着衬底温度升高，薄膜中的晶粒大小也随之升高。这可能是因为在衬底温度较低时，晶粒成核和生 长时缺少热能，导致生长的晶粒比较小[7]。 利用特征 X 射线能量散射（EDS）对薄膜中的化学成分进行了测量。表 1 列出了四个不同衬底温 度下生长的 SnS 薄膜的 EDS 分析结果。从 EDS 的测试结果可以看出，衬底温度越高，SnS 薄膜中 Sn 和 S 元素的化学计量比越接近 1，当衬底温度为 150℃时，Sn 和 S 元素的化学计量失配最小；而当衬 底温度进一步升高到 200℃时，Sn 和 S 元素的化学计量比失配又开始增大。 表 1 不同衬底温度下真空蒸发生长的 SnS 薄膜中 Sn 和 S 元素的 EDS 测试结果 Table 1 EDS analysis of SnS thin films at different substrate temperatures substrate temperatures

element

Weight%

Atomic%

Sn

74.18

43.70

S

25.82

56.30

Sn

74.62

44.26

S

25.38

55.74

Sn

75.58

45.54

S

24.42

54.46

Sn

74.75

44.40

S

25.25

55.60

（℃） 27

100

150

200

3.2 SnS 薄膜的光学性能

- 96 -

图 3 不同衬底温度的 SnS 薄膜的透射光谱图 Fig.3. transmittance spectra of four SnS thin films 图 3 是不同衬底温度的 SnS 薄膜的透射光谱图。由图可见，图中曲线的吸收边陡直，对于波长大 于 800nm 的光，透过率显著提高。随着衬底温度的升高，在大于 800nm 波长波段范围内，薄膜的透 过率升高，这可能是薄膜的结晶性改善而引起的[11]。 3.0x109

( α hν) 2/cm-2eV2

2.5x109

2.0x109

27°C

1.5x109

200°C

1.0x109

5.0x108

0.0 1.10 1.15 1.20 1.25 1.30 1.35 1.40 1.45 1.50 1.55 1.60 1.65 1.70 1.75 1.80

hν /eV

图 4 不同衬底温度的 SnS 薄膜的（αhν）2－hν 的关系图 Fig.4. Plots of （αhν）2 vs. hν for two SnS thin films 根据禁带宽度的测试方法，作出（αhν）2－hν的关系图[18]，并进行外推，就可以求出SnS薄膜的 直接禁带宽度。图4是两种不同衬底温度下制备的SnS薄膜的（αhν）2－hν的关系图。根据外推直线与 横轴的交点，可以得到这两种SnS薄膜的禁带宽度分别为1.375eV和1.416eV。采用同样方法可以得到 其它衬底温度下的SnS薄膜的禁带宽度，具体结果如表2所示。需要指出，对于制备的大部分样品， （αhν） 2

与hν 没有线性关系, 因而不能确定间接禁带宽度。

- 97 -

表 2 不同衬底温度下 SnS 薄膜的光学禁带宽度 Table 2 The optical band-gap of four SnS thin films substrate temperatures（℃）

band-gap（eV）

27

1.372

100

1.375

150

1.402

200

1.416

3.3 SnS 薄膜电学性能 采用冷热探针法判别 SnS 薄膜的导电类型为 P 型。为了测量 SnS 薄膜的电导率，在膜上蒸镀一 层条状银电极，用 KEITHLEY4200 半导体特性分析仪记录相关参数，计算电导率。测光电导时用卤 钨灯作为光源，光强为 30mW/cm2。测量了不同衬底温度下 SnS 薄膜的暗电导率 σd 和光电导率 σph， 结果如表 3。从表中可以看出，薄膜暗电导率先随衬底温度升高而增大，然后随着衬底温度升高而降 低。当衬底温度为 150℃时薄膜的电导率为 0.01Ω-1·cm-1。电导率之所以这样变化可能是由于随着衬底 温度的升高，SnS 薄膜晶体结晶性更好，晶粒尺寸增大，所以电导率随着衬底温度的升高而增大[7]。 而当衬底温度高于 150℃时，SnS 薄膜的择优取向性随着衬底温度的升高反而变差，导致了薄膜的电 导率降低。薄膜的 σd/σph 之比也是先随着衬底温度的升高而增大，然后随着衬底温度的升高而降低。 这是因为衬底温度为 150℃的 SnS 薄膜相对于低温衬底样品晶体结晶性更好，晶粒尺寸更大。而衬底 温度为 200℃的薄膜样品中由于薄膜内部缺陷较多，导致更多光生载流子的复合，从而降低了薄膜的 光暗电导之比。 表 3 不同衬底温度下 SnS 薄膜的暗电导率 σd 和光电导率 σph Table 3 Dark-conductivity and photo-conductivity of four SnS thin films substrate σd(Ω-1·cm-1)

σph(Ω-1·cm-1)

27

0.00165

0.0065

3.9

100

0.0073

0.035

4.8

150

0.01

0.08

8.0

200

0.0061

0.03

4.9

temperatures

σph/σd

（℃）

4 结 论 用真空蒸发法在不同的衬底温度（27℃、100℃、150℃和 200℃）上制 备 SnS 薄膜，测量各种 SnS 薄膜的晶体结构和光学、电学特性。实验表明 SnS 薄膜的各种特性都与衬底温度有显著的关系，综合 得出最佳的衬底温度为 150℃。当衬底温度为 150℃时，制备的 SnS 薄膜为多晶的斜方晶系，（111） 晶面有择优取向，晶粒大小约为 0.5μm，Sn 和 S 元素的化学计量比接近于 1，导电类型为 P 型，暗电 导率、光电导率分别为 0.01Ω-1·cm-1 和 0.08Ω-1·cm-1，禁带宽度为 1.402 eV。 参考文献 - 98 -

[1] K.T.Ramakrishna Reddy, P.Purandhara Reddy, P.K.Datta, et al. Formation of polycrystalline SnS layers by a two-step process [J]. Thin Solid Films, 2002, 403 –404: 116–119. [2] L Ehm, K Knorr, P Dera. Pressure-induced structural phase transition in the IV-VI semiconductor SnS [J]. Journal of Physics Condensed Matter, 2004, 16: 3545-3554. [3] M.M. El-Nahass, H.M. Zeyada, M.S. Aziz, et al. Optical properties of thermally evaporated SnS thin films [J]. Optical Materials, 2002, 20:159-170 [4] K.T. Ramakrishna Reddya, N. Koteswara Reddya, R.W. Miles. Photovoltaic properties of SnS based solar cells [J]. Solar Energy Materials and Solar Cells, 2006, 90 : 3041–3046. [5] N. Koteeswara Reddy, K.T. Ramakrishna Reddy. Optical behaviour of sprayed tin sulfide thin films [J]. Materials Research Bulletin, 2006, 41: 414–422. [6] Sekhar C.Ray, Malay K.Karanjai, Dhruba DasGupta. Structure and photoconductive properties of dip-deposited SnS and SnS2 thin films and their conversion to tin dioxide by annealing in air [J]. Thin Solid Films, 1999, 350:72-78. [7] M. Devika, N. Koteeswara Reddy, K. Ramesh, et al. Influence of substrate temperature on surface structure and electrical resistivity of the evaporated tin sulphide films [J]. Applied Surface Science, 2006. [8] N. Koteeswara Reddy, K.T. Ramakrishna Reddy. Growth of polycrystalline SnS films by spray pyrolysis [J]. Thin Solid Films, 1998, 325:4-6. [9] Tanusevski, A. Optical and photoelectric properties of SnS thin films prepared by chemical bath deposition[J]. Semicond. Sci. Technol., 2003, 18(6): 505-505 [10] M. Ristov, G. Sinadinovski, M. Mitreski, et al. Photovoltaic cells based on chemically deposited p-type

SnS [J]. Solar Energy Materials & Solar Cells,2001, 69:17-24. [11] Shuying Cheng, Guonan Chen, Yanqing Chen et al. Effect of deposition potential and bath temperature on the electrodeposition of SnS film [J]. Optical Materials, 2006. [12] Nair, M.T.S. Simplified chemical deposition technique for good quality SnS thin films[J]. Semiconductor Science and Technology, 1991, 6(2): 132-134 [13] Tanusevski, A. Optical and photoconductive properties of SnS thin films prepared by electron beam evaporation[J]. Solar Energy Materials and Solar Cells, 2003, 80(3): 297-303 [14] Guang-pu, Wei. Investigation on SnS film by RF sputtering for photovoltaic application[J]. Conference Record of the IEEE Photovoltaic Specialists Conference, 1st World Conf. on Photovoltaic Energy Conversion, 1994, 1: 365-368 [15] Yanuar, F. SnS thin films grown by close-spaced vapor transport[J]. Journal of Materials Science Letters, 2000, 19(23): 2135-2137 [16] Jayachandran, M. Studies on the brush plated SnS thin films[J]. Journal of Materials Science Letters, 2001, 20(4): 381-383 [17] Kim Shi Yul, Kim Dong Seop, Ahn Byung Tae, et al., Structural, electrical and optical properties of In-doped CdS thin films prepared by vacuum coevaporation, Thin Solid Films, Vol. 229, No. 2, 1993:227-231 [18] Ichimura, M. Electrochemical deposition of SnS thin films[J]. Thin Solid Films, 2000,

361: 98-101. - 99 -

用真空硒化法和叠层法制备铜铟硒多晶薄膜 林飞燕，秦娟*，徐 环，史伟民，魏光普 （上海大学材料科学与工程学院，上海 200072） 摘要：本文采用真空顺序蒸发铜铟金属预置层后真空硒化退火的方法（硒化法），以及真空三元叠层 蒸发后氮气气氛退火的方法（叠层法）分别制备了太阳能电池吸收层材料 CuInSe2 薄膜。通过 X 射线 衍射(XRD)、扫描电子显微镜(SEM)、能量色散 X 射线分析技术(EDX)等分析手段检测了薄膜表面形 貌、微结构和组分。结果表明：两种方法制备的薄膜形貌都比较致密均匀,晶粒直径分别约 1.5 μm 和 约 1 μm。组分分析表明所制薄膜均为富铜 CIS。硒化法制备的 CIS 薄膜具有单一的黄铜矿相结构；而 叠层法制备的薄膜含有少量杂相，如 β-In2Se3 等。因此硒化法制备的薄膜更适于作为太阳能吸收层材 料。 关键词：铜铟硒；薄膜太阳电池；真空蒸发法；硒化法；叠层法 中图分类号: TM 615

文献标识码: A

Fabrication of CuInSe2 Polycrystalline Films by Vacuum Selenization and Stacked Elemental Layers LIN Fei-yan, QIN Quan*, XU Huan, SHI Wei-min, WEI Guang-pu (School of Material Science and Engineering, Shanghai University, Shanghai 200072,China ) Abstract: The CuInSe2 (CIS) films were fabricated by selenization of evaporated metallic precursors and vacuum evaporation of stacked elemental layers (SEL) followed by a thermal annealing step. The morphology, microstructure and composition of the films were investigated by scanning electron microscopy (SEM) , X-ray diffraction (XRD) and Energy Dispersive X-ray spectroscopy (EDX). The results show that all the thin films are compact and uniform. The grain sizes of the two kinds of films are 1.5 μm and 1.0 μm, respectively. All the films made by the two methods are Cu-rich and show p-type conduction. However, the CIS thin films fabricated by the former method consist of a single phase of chalcopyrite structure, while those films fabricated by the later method contain impurity phases such as β-In2Se3 besides CIS phase. Thus the former method is better for fabricating the CIS absorber in solar cells. Key words: CuInSe2; thin film solar cells; vacuum evaporation；selenization; stacked elemental layer

- 100 -

1 引 言 具有黄铜矿结构的I－Ⅲ－Ⅵ族三元化合物CuInSe2（CIS）经实验和理论证实在可见光区具有高吸 收系数（6×105 cm-1），较为适合的禁带宽度（Eg＝1.04 eV）以及能不借助外加杂质制成P型（富铜） 或N型（富铟）半导体等优点，成为薄膜太阳能电池中优秀的吸收层材料[1]。CIS中增加适量的Ga制成 的CuIn1- xGaxSe2太阳能电池的转化效率已经达到了21.5%[2]。 国内外研究者寻求了一些不同的薄膜制备工艺，如真空蒸发法[3]、 磁控溅射法[4]、 电沉积法[5]、 分子束外延气相沉积法等。其中真空蒸发法和磁控溅射法在日本、美国，德国无论在实验室和生产线 上都有采用。但作为实验室里制备面积较小的电池样品，真空蒸发法制备的电池效率较高。包括现在 报道的最高转化率的CIGS电池层也是蒸发法制备的。 本论文采用真空顺序蒸发铜铟金属预置层后真 空 硒化退火的方法(硒化法)，以及Cu、In、Se三元素顺序叠层真空蒸发后氮气气氛退火的方法(叠层 法)制备了CIS薄膜，对两种方法制得的CIS薄膜进行了成分、晶相、薄膜形貌等的检测分析和比较。 2 实 验 本实验采用普通玻璃与镀钼碱石灰玻璃为衬底。镀钼玻璃的钼层厚度约为 2 μm。衬底采用去离 子水、丙酮、无水乙醇超声清洗，然后烘干装片。真空蒸发在型号为 DM-450A 镀膜机中进行。 2.1 硒化法 将按一定比例称量好铜粉（5N）和铟粒（5N）放置于处在同一水平面上的两个 Mo 舟内。蒸发 －3

镀膜时，衬底与舟的距离为 20 cm，真空度为 1.6×10

Pa。铜的蒸发温度为 1100 ℃，铟的蒸发温度

为 700 ℃。 硒化退火时，将制备好金属预置层的衬底和硒粉放在一个密闭的石墨盒中；石墨盒放在管式电阻 炉的石英管中部；石英管由机械泵抽真空至 1 Pa。退火步骤如图 1（a）所示。退火完成后，随炉冷却 至室温。 2.2 叠层法 将按一定比例称量的铜粉（5N） 、铟粒（5N）和硒粉（5N）置于处在同一水平面上的三个 Mo 舟 内；硒的蒸发温度为 250 ℃。In 与 Cu 的蒸发温度分别为 700 ℃和 1100 ℃。此处采用的蒸发顺序为 Se/In/Cu。 氮气退火时，将制备好的薄膜放在一个密闭的石墨盒中；石墨盒放在管式电阻炉的石英管中部； 通氮气排除空气后逐步升温至 227 ℃，保温 1 h，如图 1（b）所示。退火完成后在氮气气氛下，随炉 冷却至室温。实验样品的对应编号见表 1。

Temperature / oC

Temperature / oC

250 500 400 300 200 100 0

(a) 30

60

90

120 150 180

200 150 100 50 0

(b) 30

60

90 120 150 180 210 240

Time / min Time / min 图 1 两种制备方法的退火步骤（a）硒化法； （b）叠层法 Fig.1 The annealing steps of（a）selenization, and (b)SEL - 101 -

表 1 试样编号 Table 1 The serial number of samples 方法

普通玻璃衬底

镀钼玻璃衬底

硒化法

S1—P

S1—M

叠层法

S2—P

S2—M

2.3 样品的表征 薄膜的晶体结构用 Rigaku D/MAX-111C 型 X 射线粉末衍射仪测量，测量时采用 Cu 转靶，电压 40 kV，电流 30 mA。X 射线的波长为 0.1541 nm，扫描范围 2θ 为 20°～60°，步长 0.02°；薄膜的表面 形貌用 KYKY-1000B 型扫描电子显微镜(SEM) 观察并利用能量色散 X 射线分析技术(EDX) 测量薄膜 中 Cu、In、Se 的原子比；薄膜的导电类型用热探针法测量。 3 结果与讨论 3.1 CIS 薄膜的表面形貌分析 图 2 所示为两种方法制得的 CIS 薄膜的 SEM 照片。从图中可以看出两种方法制得的 CIS 均是致 密、均匀的多晶薄膜。其中 S1—P 试样形貌略微粗糙，晶粒较大，直径约为 1.5 μm，而 S2—P 较细 密，晶粒直径约为 1.0 μm。据文献报道，高质量的 CIS 薄膜应具有较好的致密性与较大的晶粒尺寸以 尽量减少晶界缺陷[6]。可见硒化法制得的薄膜形貌更优良。

（a）S1—P

（b）S2—P 图 2 CIS 薄膜的 SEM 照片

Fig.2 SEM micrographs of the CIS films 3.2 CIS 薄膜的结构分析 图 3 为硒化法制得的 CIS 薄膜的 XRD 图谱。 从图中可看出（a）和（b）在 2θ 为 26.86 °、35.78 °、 44.44 °、52.58 °的位置上出现了 CIS 特征衍射峰，表明制得的薄膜具有单一黄铜矿相的 CIS 多晶结构。 图 4 为叠层法制得的 CIS 薄膜的 XRD 图谱。其图 4(a)、(b)中除了有 CIS（112） 、(211)、 （220）/ （204）与（312）的衍射峰外，还出现了 β-In2Se3 (015)的衍射峰。其原因可能是由于实验中采用的蒸 发顺序为 Se/In/Cu，在后继的退火重结晶时，处于顶部的 Cu 未能充分扩散与 In、Se 化合，使得薄膜 中最终残留部分 In 的二元硒化物[7]。另外在 44.9 °和 51.1 °附近也有杂峰，表明薄膜中还有二元相

- 102 -

(312)

S (220)/(204)

211) MO（110）

CIS (112) n2Se3 (015)

arb.units

(312)

IS (220)/(204)

(211) MO（110）

CIS (112)

/ arb.units

Cu2-xSe 的存在[8]。

3.3 CIS 的 EDX 成分分析 表2 给出了CIS薄膜的EDX分析结果。从中可以看出，叠层法制备的薄膜（S2—P）中Cu/In原子 比明显高于硒化法制备的薄膜（S1—P）中Cu/In原子比，达到1.3893。由于Cu/In比偏离1较大，薄膜 中最终可能将残留部分Cu或In的二元硒化物。这与XRD结果表征的S2—P与S2—M薄膜中存在β-In2Se3 和Cu2-xSe相吻合。而硒化法制备的Cu/In原子比为1.0504 非常接近1，在XRD结果分析中也表明薄膜由 成分单一的多晶CIS组成。可见硒化法在合适的原料配比和适当的工艺条件下，可以制备出符合化学 计量比的优良CIS薄膜。 表 2 薄膜 EDX 成分分析（原子百分比） Table2 Results of films components (atomic concentrations) by EDX 试样

S1—P

S2—P

n（Cu）％

n（In）％

n（Se）％

采样点 1

23.7545

22.6127

53.6328

采样点 2

24.3052

23.3739

52.3209

采样点 3

23.7514

22.3772

53.8714

平均值

23.9370

22.7879

53.2750

采样点 1

27.5415

20.0817

52.3768

采样点 2

27.2351

20.1357

52.6292

采样点 3

28.1019

19.4358

52.4623

平均值

27.6262

19.8844

52.4894

拟合化学式

Cu1.05InSe2.33

Cu1.39InSe2.64

用热探针法测量了 CIS 薄膜的导电类型，结果表明两种方法制得的薄膜均为 P 型半导体，这与 EDX 测得的薄膜富铜结果相一致。 4 结 论 本文采用硒化法和叠层法两种方法制备了 CIS 多晶薄膜。通过 XRD、SEM、EDX 等方法对制得 的薄膜进行了表征和分析，得出结论如下： （1）两种方法均制得表面结构致密、均质的 CIS 薄膜。但硒化法制得的薄膜，粒径略大，为 1.5 μm。 - 103 -

（2）硒化法制得的薄膜具有优良的单一黄铜矿结构。而叠层法制得的薄膜存在二元硒化物的杂相。 （3）EDX 的成分分析结果表明硒化法制备的 CIS 薄膜更接近化学计量比。 因此，硒化法制得的 CIS 薄膜更适合作为太阳能电池的吸收层材料。 参 考 文 献 [1] A. Ashour, A.A.S. Akl, A.A. Ramadan, et al. Study of polycrystalline CuInSe2 thin film formation[J]. Thin Solid Films, 2004, 467: 300-307. [2] J.S.Ward, K.Ramanathan, F.S.Hasoon, et al. A 21.5% efficient Cu(In,Ga)Se2 thin-film concentrator solar cell[J].Progress in Photovoltaics: Research and Applications, 2002, 10(1): 41-46. [3] R. Caballero, C. Guillen.CuInSe2 Formation by selenization of sequentially evaporated metallic layers[J]. Solar Energy Materials & Solar Cells, 2005, 86: 1-10. [4] J. Mu ller, J. Nowoczin, H. Schmitt. Composition, structure and optical properties of sputtered thin films of CuInSe2[J]. Thin Solid Films, 2006, 496: 364-370. [5] N.B. Chaure, J. Young, A.P. Samantilleke, et al. Electrodeposition of p–i–n type CuInSe2 multilayers for photovoltaic applications[J]. Solar Energy Materials & Solar Cells, 2004, 81: 125-133. [6] 谢大弢，赵 夔，王莉芳等。用磁控溅射和真空硒化退火方法制备高质量的铜铟硒多晶薄膜。物 理学报，2002，51(6)：1377—1382。 [7] 方 玲，张 弓，庄大明等。Cu- In 膜成分偏析对CIS 膜结构的影响。清华大学学报(自然科学 版)，2004，44(5) ：593—596。 [8] C.J. Huang, T.H. Meen, M.Y. Lai, et al. Formation of CuInSe2 thin films on flexible substrates by electrodeposition (ED) technique[J]. Solar Energy Materials & Solar Cells, 2004,82：533—565.

Fabrication and Properties Study of Cu(In1-xGax)Se2 - 104 -

Films by Vacuum Evaporation Ai-min Li, Juan Qin*, Wei-min Shi, Guang-pu Wei School of Materials Science and Engineering, Shanghai University, Shanghai 200072, China ABSTRACT The Cu(In1-xGax)Se2 (CIGS) thin films were prepared by stacked elemental layers (SEL) method via vacuum evaporation. X-ray diffraction (XRD) analysis showed that the films were consisted of chalcopyrite CIGS phase. Scanning electronic microscopy indicated that the film surface was compact and the grain size was about 1μm. It was found that CIGS film with preparation sequence Ga/In/Cu/Se was of best crystalline quality. Besides, adding Ga greatly improved the crystallinity for all sequences compared with CIS films at the same annealing temperature. Keywords: CIGS, stacked elemental layers, vacuum evaporation

INTRODUCTION As the present state of environmental and energy resource becomes worse and worse in recent years, great attention has been given to the I–III–VI2-type ternary compounds to develop renewable energy sources such as photovoltaic electric generators. Because chalcopyrite semiconductor compounds, CuInSe2 and CuIn1-xGaxSe2 (CIGS), especially for CIGS, have high absorption coefficient, high stability and adjustable energy gap, they have drawn much attention for their application cells. CuInSe2 and CuGaSe2 have a chalcopyrite structure and band gaps of 1.04 and 1.7 eV [1], respectively. The mixed compounds of the latter materials are leading candidates as absorbers in solar cells. CIGS (with x＜0.3) as absorbers exhibit cell efficiencies that surpass those of the corresponding single-crystal-based devices. Currently, CIGS based solar cells have the highest performance among all thin film cells, Efficiencies as high as 19.5% have been achieved by the NREL group using only a soda-lime glass substrate in combination with a high growth temperature[2]. Unlike III–V and II–VI binary semiconductors, CIGS compounds are well known for their ability to accommodate deviation from stoichiometry without affecting their photovoltaic properties. This is convenient for solar cell fabrication because Cu deficient or In deficient stoichiometries are inevitable. Several methods have been used to grow CIGS layers such as co-evaporation (MBE)

[4]

and sputtering

[3]

, molecular beam epitaxy

[5]

. All these techniques are considered heavy and expansive. The stacked elemental

layer (SEL) has emerged as a relative low cost and large scale growth technique. In this paper, CIGS films with good crystallinity and morphology were prepared using SEL method at quite low annealing temperature. This technique has the advantage of lowering danger by eliminating the use of toxic H2Se gas. We present electrical and optical measurements of CIGS thin films. The electrical data has interpreted in relation with the polycrystalline nature of these materials. The optical measurements have shown that the materials characteristics namely their direct band gaps are very suitable for solar cells fabrication. - 105 -

EXPERIMENTAL The Mo-coated soda-lime glasses were used as substrates. The powder of Cu, In, Ga and Se (all with 5N quality) were put into four Mo-boats and evaporated sequentially. The thickness of CIGS layer was about 2.0μm. The CIGS polycrystalline thin-film samples were prepared with different elemental sequence: Ga/Se/In/Cu, Ga/Se/Cu/In, Ga/In /Cu/Se, and Se/Cu/In. The evaporation temperature of Cu, In, Ga and Se source was 1100℃, 700 ℃, 30℃ and 250℃respectively. To avoid the Ga losing, Ga was evaporated first and Cu should not be evaporated next due to the low melting point of Ga. CuInSe2 (CIS) films without Ga layer were also prepared for comparison. The working pressure was maintained at 10-3 Pa. All the pre-deposited films were annealed at 230℃ in N2 atmosphere for 1h and then cooled down naturally. The structure properties of the films were characterized by X-ray diffraction (XRD), and the morphology was evaluated by scanning electronic microscope (SEM). To obtain energy gap, optical transmission spectrum was measured with Perkin-Elmer LAMBDA 9 double beam spectrophotometer in the wavelength range 600–1200nm at room temperature.

RESULTS AND DISCUSSION Figure.1 displays XRD patterns of CIGS（CIS）thin film with different evaporation sequences. Curve a, b, c and d responds to samples with evaporation sequence of Ga/Se/In/Cu, Ga/Se/Cu/In, Ga/In /Cu/Se and Se/Cu/In, respectively. The peak at 40.8° belongs to Mo layer. Peaks located at about 27.2°, 45.1°, 52.9° are the characteristic peaks of chalcopyrite CIGS/CIS (112), (204)/(220), (312) facets respectively, which means that the films are consisted of single phase CIGS. In Fig.1, the homogeneous peaks of those samples were not at the same position exactly. The peaks of curve a deviated to the right side, which means this sample contains higher Ga content. The peaks of curve c are sharpest, indicating that sample c is of best crystallinity. The reason may be as follows: Cu layer was evaporated thirdly, which was carried out with very high source temperature of above 1100℃, resulting high energetic Cu atoms which were propitious to the reaction with the front two elements. Thus the sequence of Ga/Se/Cu/In and Ga/In /Cu/Se were better than the sequence of Ga/Se/In/Cu. On the other hand, depositing Se as the last layer is favorable for Se atoms to diffuse into other layers during annealing process. Therefore, the sample with deposition sequence Ga/In/Cu/Se is of best crystalline quality. Samples b and d in Fig.1 are of the same elemental deposition sequence except that d does not contain Ga. In curve d, the peak at 30.9° may be related to the InSe second phase [6]. It is obvious that sample b has far better crystallinity than sample d which containing other phases besides CIGS. The reason may be that the joining of Ga changes the reaction route of the elements during the annealing process, but the details needs further investigation. However it is clear that adding Ga is favorable for obtaining better crystallinity at the same annealing temperature compared with CIS films. Particularly, this process reduced the process temperature of CIGS thin films, which is vital for flexible CIGS solar cells grown on organic substrates [7].

- 106 -

Fig. 6. XRD patterns of CIGS/CIS thin films.

Figure.2 showed a typical SEM image of CIGS thin films. The surface was smooth with averaged grain size of about 1 μm, which is characteristic for CIGS films applied in solar cells. Large grain size is good for decreasing current loss due to grain boundary. Free charge carriers are then trapped in these GB states, creating a depleted space charge region next to the boundaries, and a potential barrier for electronic transport between adjacent grains.

Fig. 2. SEM image of a typical CIGS thin films prepared by SEL

Fig.3 showed the absorption spectrum of sample shown in Fig. 2 in wavelength range of 600-1200. The absorption of the CIGS thin film decreased rapidly when the wavelength was greater than 1000 nm due to the photons with lower energy could not be absorbed by CIGS thin films. The optical energy gap obtained by linear fitting at the absorption edge was about 1.11eV. On the other hand, according to EDS analysis, the

- 107 -

Ga composition of this sample is about 0.17. Using Vegard law [8], one could get the corresponding energy gap of 1.13eV, which was very close to that obtained by absorbance spectrum.

Fig. 3. Absorption spectrum of CIGS thin film.

To measure the conductivity of CIGS thin film, 1μm thickness CIGS was deposited directly on bare glass substrate, then strip Ag electrode was evaporated on the CIGS thin film. The linearity of I-V curve indicated good ohmic contact between CIGS thin film and electrode. The conductivity obtained from Fig.4 was 0.0103 S.

Fig. 4. I-V curve for CIGS thin film.

CONCLUSIONS CIGS thin films with different composition are prepared by SEL method at low annealing temperature of 230℃. XRD results show that all films have single chalcopyrite CIGS phase, and the sample with the sequence of Ga/In/Cu/Se is of best crystallinity. SEM image show that the films are structurally compact - 108 -

with averaged grain size of about 1μm. In addition, adding of Ga into CIS results in better crystallinity at the same annealing temperature.

REFERENCES 1.

S. Isomura, A. Nagamatsu, K. Shinohara and T. Aono. "Preparation and some semiconducting properties of CuInSe2 thin films," Solar Cells. 16, 143-153(1986)

2.

Miguel

A.

Contreras

K.

Ramanathan

et

al.

"Diode

characteristics

in

state-of-the-art

ZnO/CdS/Cu(In1-xGax)Se2 solar cells," Prog.Photovolt: Res. Appl. 13(3), 209-216(2005). 3.

A. Amara, A. Ferdi, et al. "Electrical and optical study of Cu(In, Ga)Se2 co-evaporated thin films," Catalysis Today.113(3-4), 251–256 (2006).

4.

Shioda R, OkadaY, Oyanagi H. "Characterization of molecular beam epitaxy grown CuInSe2 on GaAs(001)," Journal of Crystal Growth . 150, 1196-1200(1995).

5.

A.F.da Cunha, F.Kurdesau et al. "Performance comparison of hybrid sputtering/evaporation CuIn1-xGaxSe2 solar cells with different transparent conducting oxide window layers," Journal of Non-Crystalline Solids. 352, 1976–1980(2006).

6.

L. Fang, G. Zhang, D. M. Zhuang, M. Zhao et al, "Effects of composition segregation in Cu-In precursor films on microstructures of CIS f ilms," J T singhua U niv (Sci & Tech). 44(5), 593-596 (2004).

7.

F. Kessler, D. Herrmann, M. Powalla, "Approaches to flexible CIGS thin-film solar cells," Thin Solid Films. 480-481, 491-498 (2005).

8.

R. J. Heritage, P. Porteous and B. J. Sheppard, "Determining the composition of InP-GaP alloys using Vegard's Law," Journal of Materials Science. 5(8) , 709-710 (1970).

- 109 -

Passivation Layer of CdZnTe Studied by Spectroscopic Ellipsometry Jianyong Teng, Wenbin Sang, Yue Lu, Yanyan Lou, Jiahua Min, Xiaoyan Liang, Kaifeng Qin School of Materials Science & Engineering, Shanghai University, Shanghai 200072 Yongbiao Qian RAE Engineering Center, RAE Systems Inc. Shanghai 201821, China Abstract The oxidized layer on cadmium zinc telluride (CZT) wafer was obtained by using chemical etching first in a KOH-KCl solution and then in an NH4F/H2O2 solution. The oxidized layer on the CZT obtained by this method was analyzed by ex situ spectroscopic ellipsometry (SE) for the first time. In particular, the optical constants and the thickness of the chemical oxidized layer as a function of the oxidizing time were obtained. PASC：29.30.Kv;29.40-n Keywords: cadmium zinc telluride, spectroscopic ellipsometry, chemical passivation

I. INTRODUCTION Cadmium zinc telluride (CZT) has potential as a room-temperature semiconductor detector [1]. During the process of the CZT detector preparation, mechanical polishing and chemical etching always induce surface damage, dangling bonds, which are considered as the primary factors for high surface leakage current. Passivation is a chemical or physical process that renders the surface of the CZT chemically or electrically inert to its environment, which can decrease the surface leakage current, thereby, improving detector performance. The quality and thickness of surface passivation layer of CZT play a dominant role in detector performance, which can decrease the noise of the detector and improve the spectral energy resolution[2-4]. However, the properties of the passivation layer are not well known. In this paper, the physical and chemical properties of the passivation layer on the CZT obtained by so-called two-step passivation processing [3] were investigated using spectroscopic ellipsometry (SE) for the first time. In particular, the optical constants, refractive index (n) and relative permittivity (ε), and the thickness of the passivation layer as a function of the oxidizing time were first obtained. II. EXPERIMENTS DETAILS The In-doped CZT ingots with resistivities of 108-1010Ωcm, grown by the vertical Bridgman method [5], were cut into wafers of 4×4×2 mm3 with the surface orientation of (111). They were mechanically grounded, then polished with alumina suspension and finally etched chemically with BMLB. After that, the CZT samples were firstly dipped in the KOH-KCl solution to consume the Te-rich layer caused by BMLB in order to achieve a near stoichiometric surface and an NH4F/H2O2 solution [2] was then applied to the etched CZT surface to grow oxidized layer on it, which is called the two-step passivation processing. The optical constants and thickness of the passivation layer were obtained by using a spectroscopic phase modulated ellipsometry UVISEL/460-VIS-AGAS from JOBIN YVON S.A.S., in the energy range of 0.75-3.2eV, an

- 110 -

angle of incidence of 70°. The SE measurements can provide the change in the polarization state that an incident linearly polarized light beam suffers when it is reflected by a surface. This change is expressed as the ratio between the complex reflection coefficients for polarization parallel γp and perpendicular γs to the plane of incidence [6]:

ρ = γ p / γ s = tanψ exp(i∆)

(1)

where tanψ gives the relative amplitude attenuation of γp and γs vector reflected wave and Δ gives phase displacement difference between the two vectors caused by reflection. The two parameters, ψ and Δ, are ellipsometric angles which are as a function of the photon energy. Therefore, as the spectra of Ψ and Δ are determined in the SE measurements, the optical constants, refractive index (n) and extinction coefficient (k), and thicknesses of unknown layers can be obtained. Ⅲ.R E SU L T S A N D D ISC U SSIO N Fig.1 shows the typical XRD spectrum of the oxidized layer obtained by using a diffractometer Rigaku D/max-2100 with copper radiation at an incidence angle of 1.5°. As seen from Fig.1, the peaks for TeO2, TeO3 and CdTeO3 can be observed and the assign of CdO or ZnO is vague, indicating that the oxidized layer on the CZT surface produced consists mainly of TeO2, TeO3 and CdTeO3. The Te-related oxide, like TeO2 or TeO3, is similar to the results reported in the literatures[2,4,7]. However, Cd-oxidized product is a CdTeO3 related oxide[7] instead of the CdO [2,4]. The difference for this might be related to the oxidizing conditions. The morphologies as well as the MicroXRF analysis results obtained by using SEM with a FEI Quanta 200 FEG for the passivation time of 120min are shown in Fig.2. It illustrates that the oxidized layer has a good smooth and compact morphology. In addition, it was also found that the oxygen content in the oxidized layer increases with passivation time and its morphology becomes more smooth and compact. This indicates that the elemental species in the oxidized layer are more fully oxidized with time. Fig.3 shows the typical SE data of the CZT samples for different passivation times, where Is( = Sin(2ψ )× Sin∆ ) and Ic( = Sin(2ψ )× Cos∆ ) are fitting parameters. The symbols correspond to experimental data and the lines to fitted spectra with the model created by the DeltaPsi software, based on the chemical composition obtained above. The fitting procedure was carried out with the DeltaPsi software. As the parameters of the dispersion formula obtained by DeltaPsi software, were determined, the optical constants, refractive index (n) and extinction coefficient (k), and thicknesses of the passivation layer can be obtained by using approximate iterative method. The difference between the experimental data and the fitting parameters is in token of mean square root error (MSRE). Based on the results from Fig.3, the value of the MSRE is estimated to be less than 0.7. Therefore, it could be inferred that this is an ideal fitness. The dependence of the thickness of the passivation layer on the passivation time is shown in Fig. 4. It could be found that the thickness of the passivation layer increases with the time, and it levels out with further increase of the time when it is larger than 60 min. The dependence of the optical constants obtained for the CZT passivation film on the photon energy is shown in Fig5: (a) the refractive index (n), (b) the extinction coefficient (k) and (c) the relative permittivity - 111 -

(ε). It was found that there is a strong dependence of the optical constant of the with the passivation time. The reason for this was ascribed to the changes of composition in the oxidized layer. Based on the MicroXRF analysis mentioned above, it might be inferred that more enriched Te on the oxidized surface was oxidized to TeO2 with time, and alternatively, more TeO2 was further oxidized to TeO3, resulting in TeO2 or TeO3 content increase in the oxidized layer. In addition, it might also be related to the morphology of the layer, like grain size and compactness as observed from SEM. However, further work is needed to fully understand the physical and chemical properties of the oxidized layer, in particular, the relationship between the physical and chemical properties, and the optical and electrical properties.

IV. CONCLUSION The chemical oxidized layer on CZT surface obtained with so-called two-step passivation processing was analyzed by using SE measurement for the first time. In particular, the optical constants and the thickness of the layer as a function of the oxidizing time were obtained. Further work is needed to fully understand the physical and chemical properties of the chemical oxidized layer, in particular, the relationship between the physical and chemical properties, and the optical and electrical properties.

ACKNOWLEDGEMENT This work was supported by the National Natural Science Foundation of China (Grant No. 10575069) (No. 10675080) and Shanghai Leading Academic Discipline Project T0101.

REREFENCES [1] J. F. Butler, C. L. Lingren and F. P. Doty, IEEE Trans.Nucl.Sci., 605 (1992). [2] Wright G W and James R B 2000 Proc. SPIE 4141, 324 (2000) [3] Wenbin Sang, Kunshu Wang, Jiahua Min, Jianyong Teng, Qi Zhang and Yongbiao Qian, Semicond. Sci. Technol 20, 343 (2005) [4] Xiaoqin Wang, Wanqi Jie, Qiang Li, Zhi Gu, Materials Science in Semiconductor Processing 8, 615 (2005) [5] Wenbin Sang, Yongbiao Qian, et al., Journal of Crystal Growth 214/215, 30 (2000) [6] G.E. Jellison Jr., Thin Solid Films 234, 416 (1993) [7] K-T Chen, D.T. Shi, H. Chen, B. Grandeson, M.A. George, W.E.Collins, and A.Burger and R.B.James, J. Vac. Sci. Techn. A 15, 850 (1997)

- 112 -

Fig.1. The typical XRD spectrum of the passivation layer

(a)

(b)

(c) Fig.2.The SEM morphology of the passivation layer for the passivation time of 120min: (a) ×2000, (b) ×100,000, and (c)the MicroXRF analysis result

- 113 -

Fig.3. The typical SE data of the CZT passivation layer for different passivation time: (a) 10min, (b) 20min, (c) 60min, (d) 120min

- 114 -

Fig.4. The dependence of the thickness of the passivation layer on the passivation time

(a)

(b)

(c)

Fig.5. The dependence of the optical constants obtained for the CZT passivation film on the photon energy: (a) the refractive index (n), (b) the extinction coefficient (k) and (c) the permittivity (ε).

- 115 -

高阻In掺杂CdZnTe晶体缺陷能级的研究 ∗ 李刚、桑文斌、闵嘉华、施朱斌、戴灵恩、赵岳 （上海大学 材料科学与工程学院, 上海 200072） 钱永彪 （华瑞科学仪器(上海)有限公司 传感器工程技术中心, 上海 201821） 摘

要: 本文利用低压垂直布里奇曼法制备的不同 In 掺杂量的 CdZnTe 晶体样品，采用低温光致

发光谱(PL)、深能级瞬态谱(DLTS)以及霍尔测试等手段研究了 In 掺杂 CdZnTe 晶体中的主要缺陷能级 及其可能存在的补偿机制。PL 测试结果表明，在 In 掺杂样品中，In 原子占据了晶体中原有的 Cd 空 位，形成了能级位于 Ec-18meV 的替代浅施主缺陷[InCd+]，同时 [InCd+]还与[VCd2-]形成了能级位于 Ev+163meV 的复合缺陷[(InCd+-VCd2-)-]。DLTS 分析表明，掺 In 样品中存在导带以下约 0.74eV 的深能 级电子陷阱能级，这个能级很可能是 Te 反位缺陷[TeCd]施主缺陷造成的。由此，In 掺杂 CdZnTe 晶体 的电学性质是 In 掺杂施主缺陷、 Te 反位深能级施主缺陷与本征受主缺陷 Cd 空位和残余受主杂质缺陷 补偿的综合结果。

关

键 词： 碲锌镉；低温 PL；深能级瞬态谱；缺陷能级

中图分类号： TN304

Study on the defect energy levels of high resisitivity In-doped CdZnTe crystals LI Gang, SANG Wen-Bin, MIN Jia-hua, SHI Zhu-Bin, DAI Ling-En, ZHAO Yue (School of Material Science and Technology, Shanghai University, Shanghai, 200072) QIAN Yong-Biao (RAE Engineering Center, RAE Systems Inc. Shanghai 201821, China) Abstract: Low temperature photoluminescence (PL) spectra, Deep Level Transient Spectroscopy (DLTS) and high resistivity Hall test were used to study major defects in high resistivity In-doped CdZnTe crystal and its possible compensating mechanism. Samples with different In dopant concentration were grown by Low Pressure Vertical Bridgman Method. The PL spectra showed that in the In-doped CdZnTe samples of high resistivity, In dopants occupied Cd vacancies, which would exist in undoped CdZnTe crystal, forming shallow donor defect [InCd+], located at Ec-18meV, and the [InCd+] interacted with [VCd2-] to form a complex defect [(InCd+-VCd2-)-] at Ev+163meV. The DLTS results showed that a deep level donor defect was found at 0.74eV below the conduction band, representing probably the energy level of antisite defect [TeCd]. These results indicated that the electrical properties of In-doped CdZnTe crystals were dominated by a comprehensive compensating consequence among In donor defects, deep level donor defect Te antisites, intrinsic acceptor defect Cd vacancies and other impurities acceptor defects. Keywords: CdZnTe; low temperature PL; deep level transient spectroscopy; defect levels ∗

基金项目：国家自然科学基金(60676002)，上海市重点学科建设项目(T0101) 作者简介：李刚(1983-)，男，硕士研究生，Email: [email protected] - 116 -

1.

引言 CdZnTe 具有较高的平均原子序数和较大的禁带宽度，所以这些材料的探测器具有较大的吸收系 ，2]

数、较高的计数率，尤其是不需任何的冷却设备就能在室温下工作[1

。但由 于 CdZnTe 熔体的 Cd 分

压要比 Te 分压高 1－2 个数量级，所以采用传统的布里奇曼法熔体生长 CdZnTe 晶体时，容易产生化 学计量比偏离，即产生 Cd 空位,导致晶体电阻率的下降。近年来研究发现，在低压布里奇曼法生长 CdZnTe 晶体中加入 In 施主杂质可提高晶体电阻率和载流子寿命，并对γ射线有较高的灵敏度[3-5]。但 是 In 是典型的 IIIA 族元素，其价电子构型为 5S25P1，其在 CdZnTe 中替代 Cd 位为施主缺陷，可以获 得高阻材料，也可获得 n-型低阻材料。这可能与晶体生长工艺参数以及热处理条件相关。对此，还尚 缺乏深入了解，尤其是有关杂质与本征缺陷的相互作用与补偿机理。低温光致发光(PL)谱是研究 CdZnTe 晶体中缺陷能级的重要手段[6,9]，而深能级瞬态谱(DLTS)是分析深能级缺陷的有效方法[7]。本 文将采用低温 PL 和 DLTS，结合高阻霍尔测试，研究高阻掺 In CdZnTe 晶体中的缺陷能级，重点分析 讨论 In 施主与 Cd 空位等本征缺陷的相互作用与补偿机理。 2.

实验 CdZnTe 晶体是由低压垂直布里奇曼法制备获得[8]。测试样品分别来自三次生长试验，其中

CZT-15、 CZT-35 为掺 In 样品， 其掺 In 量分别为 5×1017/cm3 和 7×1017/cm3，同时还选取未掺杂 CdZnTe 晶体(CZT-05)以作比较。样品首先使用金刚砂颗粒进行双面研磨，然后用 Al2O3 悬浮液来机械抛光， 直至样品表面呈镜面后用超声清洗晶片，接下来用 BM 腐蚀液(5％Br+甲醇)和 LB 腐蚀液(2％Br+20％ 乳酸+乙二醇)进行化学腐蚀抛光，并用甲醇冲洗腐蚀表面，最后用氮气吹干备用。 低温 PL 谱测试采用低温氦循环光致发光系统。系统的激光器为氦-氖激光器，激光波长为 655nm， 激光衰减 10000 倍，积分时间 2S×2，采用 600(/mm)光栅，观测镜头为 50 倍。系统利用液氦循环制冷， 最低温度可达 4.2K。DLTS 由 DLTS-HRCV 型深能级瞬态谱仪测得。整个测试台浸入到液氮中，冷却 到温度 78K。 测试样品的升温控制速度：在 77K~90K 时， 0.05~0.08K/sec；90K~300K 时， 0.02~0.05K/sec； 300K~360K 时，0.08~0.1K/sec。最高的温度可升至 360K。晶体的电学性能是通过 HL500PC 型高阻霍 尔效应测试系统得到。 3. 3.1.

结果与讨论 CdZnTe 晶体的低温 PL 谱

图 1 CdZnTe 晶体在 5K 时的 PL 谱：(a) In 掺杂 (a:CZT-15, b:CZT-35) (b)未掺杂 (CZT-05) Fig. 1 PL spectra of CdZnTe crystals at 5K: (a) In-doped (a:CZT-15, b:CZT-35); (b)Undoped (CZT-05)

- 117 -

图 1 为 CdZnTe 晶体在 5K 时的 PL 谱，其中(a)为掺 In CdZnTe 晶体，(b)为未掺杂 CdZnTe 晶体 以作比较。由图 1(a)中的曲线 a 可以看到，掺 In CdZnTe 晶体的 PL 谱，主要包括峰位位于 1.653eV 强度较高的施主束缚激子峰(D0, X)、位于 1.641eV 强度较弱的受主束缚激子峰(A0, X)、位于 1.608eV 施主受主对(D, A)复合发射峰、峰值在 1.488eV 附近发射带以及峰位位于 1.658eV 的自由激子峰 X。 由图 1（b）可知，未掺杂 CdZnTe 晶体的 PL 谱，主要包括峰位处于 1.641eV 中性受主束缚激子峰(A0, X)，以及强度相对较弱峰位在 1.601eV 的施主受主对复合发射峰(D, A)。一般认为，(A0, X)是由与 Cd 空位有关的受主缺陷束缚自由激子造成的[9, 22]。 通过比较 In 掺杂晶体和未掺杂晶体的 PL 谱线可以发现，In 掺杂晶体的 PL 谱有以下特征：其一、 掺 In 晶体在 1.653eV 处出现了一个新的中性施主束缚激子峰(D0, X)，而中性受主束缚激子峰(A0,X)的 强度则大大降低，应是 Cd 空位被掺入的 In 原子所占据，形成了起类氢浅施主作用的替代缺陷[InCd+]， 导致相应的 Cd 空位浓度大为降低。其二、掺 In 晶体中的施主受主对(DAP）发射峰大为增强，这可 能是由于在与 Cd 空位补偿后，依然存在一定数量的 In 施主，这些过量的 In 与晶体中的一些残余受 主杂质如 Na、Li、N 等复合，从而使(D, A)峰得到加强。其三、掺 In 晶体中出现了峰值位于 1.488eV 的所谓的 A 中心发射带，这可能是由[InCd+]和[VCd2-]之间的相互作用，形成的复合缺陷[(InCd+-VCd2-)-] 所致[10]。 图 1（a）中曲线 a 的掺 In 浓度为 5×1017/cm3，而曲线 b 为 7×1017/cm3。通过比较图中的曲线 a 和 b，可以发现，在曲线 b 中中性受主束缚激子峰(A0,X)峰消失，而中性施主受主对(D, A)峰以及 A 中 心发射带却有所增强。这可能是由于掺 In 量的增加，In 相关的缺陷即[InCd+]、[(InCd+-VCd2-)-]的浓度增 高，Cd 空位补偿更加完全。这些现象也进一步说明，(A0, X)、(D, A)以及 A 中心的强度是与 In 有关， 而不是晶体中的其它杂质所致。 In 掺杂 CdZnTe 晶体（CZT-15）的 PL 谱与温度的关系如图 2 所示。从图中可以发现，随着温度 从 5K 提高到 80K，各个位置的发光峰均有不同程度的降低，有些峰下降得很快，直至消失。这主要 是热淬灭的缘故，电子和空穴不再以发光的形式跃迁，而是主要通过热激发非辐射跃迁。此外，还可 以看出，随着温度的升高，所有的发光峰位均向低能量方向移动。这是因为随着温度的升高，样品的 禁带宽度逐渐减小，晶格常数略微变大所造成的。这一实验结果与掺 In CdTe 晶体的理论结果符合较 好[11]。

图 2 In 掺杂 CdZnTe 晶体（CZT-15）在不同温度下的 PL 谱 Fig.2 The PL spectra of In-doped CdZnTe at different temperatures

- 118 -

3.2.

CdZnTe 晶体的深能级瞬态谱（DLTS）

图 3 CdZnTe 晶体的深能级瞬态谱（DLTS） ： （a）掺 In 样品 CZT-35；(b)未掺杂样品 CZT-05 Fig.3 Deep Level Transient Spectra of CdZnTe crystals: (a) In-doped CZT-35, (b) undoped CZT-05 图 3 为 CdZnTe 样品在 80K~360K 的温度范围内所测试得的 DLTS，其中(a)为 CZT-35 样品在 en=0.173 ms-1 条件下测得的；(b)为 CZT-05 样品在 ep=0.043 ms-1 条件下测得的。从图 3 中可以看出， 掺 In 样品的 DLTS 曲线只有一个单峰(a)，说明该样品中存在一个深能级电子陷阱；而未掺杂样品的 DLTS 曲线也只有一个单峰(b)，说明该样品中存在一个深能级空穴陷阱。测得的 DLTS 谱峰的高度反 映了缺陷的隙态密度[12]。电子发射几率 en、空穴发射几率 ep 是由取样积分平均器的输出信号为极大 值时(即瞬态谱的峰值)所对应得指数信号的时间常数 τmax 算出，数学表达式为：

e=

1

τ max

=

ln t 2 − ln t1 t 2 − t1

（1）

其中 t1 和 t2 分别为取样积分平均器的取样时间。(1/τmax)为率窗值。改变率窗值，可以得到不同的 深能级瞬态谱。按 ln(en/T2)对 1/T 和 ln(ep/T2)对 1/T 作函数曲线，根据曲线的斜率可求得深能级位置， 如图 4 所示。从图 4 (a)可得到掺 In 样品中深能级电子陷阱能级为导带以下约 0.74eV，这个结果接近 Berding[13]关于反位缺陷[TeCd]的理论计算值 0.75eV。由图 4 (b) 曲可得到未掺杂样品中深能级空穴陷 阱能级为价带以上 0.45eV，这个数据接近文献[14,15]报道的 Cd 空位能级，其数值在价带上 0.47eV 以内。 但是在未掺杂样品的 DLTS 中反位缺陷[TeCd]并没有得到反映，这可能与晶体生长条件有关，但其确 切原因还不甚清楚，需进一步深入研究。

（a）

（b）

图 4 深能级发射率与温度之间的关系：（a）掺杂样品；（b）未掺杂样品 Fig.4 The Temperature dependence of deep level emission rate: (a) In-doped, (b) Undoped

- 119 -

3.3.

CdZnTe 晶体的电学性能 CdZnTe 晶体样品霍尔测试结果如表 1 所示。从表中可以看出，CdZnTe 样品的电学性能与其掺杂

浓度直接相关。掺 In 样品的电阻率要远高于未掺杂样品，且导电类型不同，掺 In 样品的导电类型为 n 型而未掺杂样品为 p 型。因此可以推测，未掺杂样品中 Cd 空位浓度对其电学性能起决定作用；而 掺杂样品由于 In 施主的引入，对 Cd 空位及其他杂质受主的补偿作用，形成了对其电学性能起决定作 用的 In 替代缺陷[InCd+]，因此样品的导电类型转变成 n 型。同时，从表中还可以看到，掺 In 浓度高 的样品电阻率比掺 In 浓度低的样品要低，这可能是由于 In 施主在与 Cd 空位完全补偿后还有剩余的 In 施主，掺 In 量愈多晶体中剩余载流子浓度愈高，因而电阻率会降低。实际上，晶体的导电类型以 及载流子浓度可能是由晶体中 Cd 空位、Te 反位缺陷以及施主 In 共同作用的结果[11]，因此与掺 In 量 以及熔体的化学计量比条件相关。 表 1 不同掺 In 量的 CdZnTe 样品的电学性质 Table 1 The electrical properties of CdZnTe crystal with different In dopant concentration

Sample

In dopant concentration -3

CZT-03 CZT-13 CZT-35 3.4.

Carrier density

0

1.23×1010

7×10

17

(cm / Vs)

(Ωcm)

type

62.1

8.2×106

-3

(cm )

5×10

Conduction

2

(cm ) 17

Resistivity

Mobility

1.77×10

6

1.06×10

8

189

weak n

8

n

1.87×10

153

p

10

3.85×10

In 掺杂 CdZnTe 晶体中的缺陷能级 根据低温 PL 以及 DLTS 的测试与分析结果，我们可以推测高阻 In 掺杂 CdZnTe 晶体中的一些主

要缺陷能级，如图 5 所示。根据研究表明[16]，CdZnTe 自由激子束缚能约为 10.8 meV，即位于导带下 约 10.8meV 处，由此根据自由激子峰的峰位（1.658eV）可计算出晶体的禁带宽度，其值为 1.669eV。 由于与 In 施主杂质能级相关的（D0，h）比（D0,X）低 2.5meV）[16]，而（D0,X）峰位于 1.653eV，因 此[InCd+]的能级位置在导带下 18 meV，这一结果与 M.Rub 等人的结果[17]相似。施主受主对发射峰（D, A）位于 1.601eV，因此对应的受主能级位置在 Ev+50meV

[6]

，很可能是 Na、Li、N 等残余受主杂质

能级[19,20]；A 中心的峰位在 1.488eV，由于其反映的是[InCd+]与[(InCd+-VCd2-)-]施主受主对跃迁，因此 A 中心对应的能级位置应在 Ev+163meV [7]，这一结果接近文献[21,22]的结果。

图 5 高阻 In 掺杂 CdZnTe 晶体中的主要缺陷能级示意图 Fig.5 Dominant defect energy levels in high resisitivity In-doped CdZnTe crystals - 120 -

4.

结论 本文采用低温 PL 以及 DLTS 等手段研究了 In 掺杂 CdZnTe 晶体中的主要缺陷能级及其可能存在

的补偿机制。结果表明，在 CdZnTe 晶体中的掺 In 量对晶体中的缺陷类型、浓度以及缺陷的补偿作用 产生较大影响。在未掺杂 CdZnTe 样品中，本征缺陷 Cd 空位对晶体电学性能起决定作用。在掺杂样 品中，晶体中的 Cd 空位被 In 原子所占据，形成了替代浅施主缺陷[InCd+]，其能级位于 Ec-18meV；而 [InCd+]还会与[VCd2-]作用形成能级位于 Ev+163meV 的复合缺陷[(InCd+-VCd2-)-]；同时导带以下约 0.74eV 的深能级施主缺陷即反位缺陷[TeCd]，一起参与补偿本征受主缺陷 Cd 空位和残余受主杂质缺陷。因此， In 掺杂 CdZnTe 晶体的电学性质是 In 掺杂施主缺陷、Te 反位深能级施主缺陷与本征受主缺陷 Cd 空位 和残余受主杂质缺陷补偿的综合结果。 5.

参考文献

[1] Tümay O. Tümer, Shi Yin, Victoria Cajipe et al. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2003, 497(1) :21–29. [2] Limousin O. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2003, 504(1-3):24-37. [3] Sang Wenbin, Qian Yongbiao, Shi Weiming et al. Journal of Crystal Growth,2000, 214-215 :30-34. [4] Fochuk P., Panchuk O., Feychuk P. et al. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2001, 458 (1-2):104-112. [5] Chu Muren, Terterian Sevag, Ting David et al. Applied Physics Letters, 2001, 79(17): 2728-2730. [6] Li Qiang, Jie Wanqi, Fu Li et al. Journal of Applied Physics, 2006, 100, 013518. [7] Verity D., Shaw D., Bryant F. J. et al. Journal of Physics C: Solid State Physics, 1982, 15 (19):L573-L583. [8] Liu Hongtao, Sang Wenbin, Yuan Zheng et al. Rare Metal Materials and Engineering, 2007, 36(6):1016-1019. [9] Taguchi T., Ray B.. Progress in Crystal Growth and Characterization, 1983, 6(2):103-162. [10] Fiederle M., Fauler A., Konrath J. et al. IEEE Transactions on Nuclear Science, 2004, 51(41):1864-1868. [11] Seto S., Suzuki K., Abastillas V.N. Jr. et al. Journal of Crystal Growth, 2000, 214-215:5-8. [12] Lang D. V.. Journal of Applied Physics, 1974, 45(7):3023-3032. [13] Berding M. A.. Applied Physics Letters, 1999, 74(4):552-554. [14] Meyer B.K., Stadler W.. Journal of Crystal Growth, 1996, 161(1-4):119-127. [15] Lee E.Y., McChesney J.L, James R.B. et al. Compensation and trapping in semi-insulating CdZnTe, Proceedings of the 1999 Hard X-Ray, Gamma-Ray, and Neutron Detector Physics, Denver, CO, USA, 1999, 115-128. [16] Franc J, Hlidek P, Moravec P, et al. Semiconductor Science And Technology, 2000, 15 (6): 561-564. [17] Rüb M., Achtziger N., Meier J. et al. Journal of Crystal Growth, 1994, 138(1-4):285-289. [18] Chibani L., Hage-Ali M., Siffert P.. Journal of Crystal Growth, 1996, 161(1-4):153-158. [19] Molva E., Pautrat J.L., Saminadayar K. et al. Physical Review B, 1984, 30, 003344. [20] Molva E., Saminadayar K., Pautrat J.L. et al. Solid State Communications, 1983, 48 (11):955- 960. - 121 -

[21] Franc J., Babentsov V., Fiederle M. et al. IEEE Transactions on Nuclear Science, 2004, 51(3):1176-1181. [22] Yang Ge, Jie Wangqi, Li Qiang et al. Journal of Crystal Growth, 2005, 283 (3-4):431-437.

- 122 -

- 123 -

- 124 -

- 125 -

- 126 -

- 127 -

- 128 -

Effect of carbon concentration on the optical properties of nano-crystalline diamond films deposited by HFCVD method L. J. Wang, J. M. Liu, L. Ren, Q. F. Su, W. M. Shi, Y. B. Xia,

Abstract With reducing diamond grain size to nano-grade, the increase of grain boundaries and non-diamond phase will result in the change of the optical properties of chemical vapor deposition (CVD) diamond films. In this paper, the structure, morphology and optical properties of nano-crystalline diamond (NCD) films, deposited by hot-filament chemical vapor deposition (HFCVD) method under different carbon concentration, are investigated by SEM, Raman scattering spectroscopy, as well as optical transmission spectra and spectroscopic ellipsometry. With increasing the carbon concentration during the film deposition, the diamond grain size is reduced and thus a smooth diamond film can be obtained. According to the data on the absorption coefficient in the wavelength range from 200 to 1100 nm, the optical gap of the NCD films decreases from 4.3 eV to 3.2 eV with increasing the carbon concentration from 2.0 % to 3.0 %. From the fitting results on the spectroscopic ellipsometric data with a four-layer model in the photon energy range of 0.75-1.5 eV, we can find the diamond film has a lower refractive index (n) and a higher extinction coefficient (k) when the carbon concentration increases.

Jounal of Nanoscience and Nanotechnology, 8(2007) 1-6

- 129 -

Studies on the heterojunction structure of n-Si/p-nanocrystalline diamond film L. J. Wang, J. M. Liu, L. Ren, Q. F. Su, R. Xu, W. M. Shi, Y. B. Xia

Abstract: An un-doped p-type nanocrystalline diamond (NCD) film was grown by an electron assisted hot filament chemical vapor deposition (EA-HFCVD) technology on n-type single-crystalline Si substrate to fabricate p-NCD/n-Si heterojunction. The structure and morphology of the NCD film, which was analyzed by Raman spectroscopy, X-ray diffraction (XRD) and scanning electron microscopy (SEM), showed that the film consisted of 40-60 nm polycrystalline nano-grains. The results showed that with EA-HFCVD method, not only an undoped NCD film with high-conductivity but also a p-n heterojunction diode between the NCD film and n-Si substrate was fabricated successfully. The p-NCD/n-Si heterostructure was also used for ultraviolet (UV) photodetector application. Operating at a bias voltage of 10V, this photodetector showed a significant discrimination between UV and visible light, and the UV/visible-blind ratio was about three orders of magnitude.

Surface Review and Letters, 14(4)(2007) 761 – 764

- 130 -

Optimal energies for ion-assisted growth of IVA thin films Z.Q. Ma SHU-SEON’s R&D Lab, Department of Physics, Shanghai University, Shanghai 200444, P.R.China [email protected]

Abstract: Basing on a modified interaction potential and experimental results, an extensive study of ion-energy-correlative thin film growth is presented, As a result, an analytical model for ion energy dissipation into top surface layers is proposed. The proposed model extends previously published models and includes analytical expressions for lattice damage by atomic displacement in surface and subsurface layers. Theoretical calculation indicated that there were three distinctive ion energy regimes within which the different processes occur to the condensed adatoms. The medium energy region (in tens of eV) is a favorite "energy window" from which the pre-deposited atoms or atomic clusters on the substrate can be stimulated further to be rearranged by absorbing the effective energy from an energetic ion beam, leading to the uniform film growth. However, over the energy region, the ion can penetrate deeply into the film, and more damage will be caused in the subsurface layer or bulk in this case, which results in the growth of poor-quality or amorphous films. When the ion energy is lower than the “energy window”, there is no driving effect for adatoms in both surface and subsurface layers. The molecular dynamic simulations and experimental results have verified the validity of the proposed analytical model and optimal ion energy regimes.

Keywords: Surface structure optimization, ion beam role, film growth

Published review document:

- 131 -

Determination of Surface Compositions on c-Si Solar Cell by AES/ XPS F. Li, Z.Q. Ma, B. He and X. J. Meng SHU-SEON’s PV Laboratory, Department of Physics, Shanghai University, Shanghai 200444, P.R.China e-mail:[email protected] Review paper ____________________________________________________________________ It is well-known that the various defects introduced within the subsurface or on the surface of the silicon wafer during solar cell processing will badly influence minority carrier lifetime or diffusion length and the more composition centers for carriers will be formed at the local regions. The most of the non-stoichiometric compounds and the impurities accumulated on the interface or surface leads to the electronic-like ramification of hanging bonds, vacancy cluster or dislocations, which considerably reduces the inter quantum efficiency (IQE) and the opto-electronic conversion power. Therefore, the presence of the compositions and the electronic states are necessary to be investigated for the improvement of the silicon-based solar cells.

In this paper, the specific surface and interface characteristics of phosphorous-included silicon oxide / Si (111) materials system for mono-crystal Si solar cells have been examined by using Auger electron spectroscopy (AES) and X-ray photoelectron spectroscopy (XPS), respectively. The typical illustrations of XPS and AES for c-Si wafer doped with phosphorous were shown in figures 1 and 2, respectively. The approach suggested by Ma’s method [1] has been applied to our target. The chemical compositions, such as elemental phosphorus, silicon, carbon, and oxygen were detected, and the respective changing trends of atomic concentration with the depth profile were also analyzed. From the spectrogram of the surface of the wafer, it is verified that the elemental C and O have the highest and second-high atomic concentration, respectively, and Si has the third-high concentration. When the p-type silicon samples just finished thermal diffusion of phosphorous doping were sputtered by Ar+ for 0.3 minutes, a new element-P with comparatively low atomic concentration was explored. From the surface to the depth at 6 nm, the atomic concentration of C and O descends sharply, and the concentration of Si ascends significantly in the XSP depth-survey measurement, while that of P just shows a little rise. With the etching time more than 0.3 minutes, the atomic concentration of O keeps changeless, but that of Si still moves up. At the same time, the concentration of P and C declines, but the rate is much smaller than before.

Besides the changing trends of atomic concentration being analyzed, the chemical states of each element as well as relevant compounds in different depth were explored by XPS. At the surface of the wafer, - 132 -

we deduce that C1s can mainly be ascribed to (CH3)n bonding, while a few of C1s peaks are probably bonded to O atoms. The Full Width at Half Maximum (FWHM) of C1s peak is about 2.2 eV, which is larger than 0.90 eV for the known sample of (CH3)n, suggesting that C1s consists of more than one chemical state. Si2p can be ascribed to Si-C bond. But the FWHM of Si2p is as large as 2.4 eV comparing with that of 0.80 eV for the foregone sample of SiC, so the Si-C bond is not the only chemical state for Si2p. The peaks of Si2p overlap the various chemical states of P2p signal, which makes that it is difficult to the quantitative evaluation of P2p. We suppose that P2p is mainly attributed to P-O bonding, which may be included in compound such as (C6H5O)3PO. In the either imaginable case, the P atom is bonded to four O atoms by one double bond and three single bonds. The peak of O1s suggests that O1s is the possible formation of O-C bonding.

While at the depth of 6 nm, the C1s peak at around 285.32 eV should mainly be attributed to C6H5-O bonding. Based on the fact that the bonding energy (BE) of the peak is 0.02 eV higher than that in the C6H5-O bonding as well as the FWHM of C1s is quite different from the known value, we suggest C-Si and C-O bonding also exists in our wafer. The analytic results show that some of the Si atoms had been replaced by P atoms, i.e., there may be P-Si bonding at the depth of 6 nm in the wafer. At the same position, the P2p peak of 134.82 eV is mostly attributed to (C6H5O)3PO. The probably bonding C6H5-P also has been suggested. Most of O1s are the possible formation of O-C bonding at the depth of 6 nm in the wafer.

When the sputtered depth is more than 20 nm, the chemical states of some elements don’t change obviously with the depth. The C1s peak at about 283.10 eV is mainly attributed to C-Si bonding. However, the FWHM is quite different from that of 0.80 eV for C-Si bonding, suggesting that there exists C-O-C bridge or O-C bonding. The peak of Si2p is 98.10 eV; in addition to the Si-C bonding, other chemical states are not clear yet. The P2p peak of 133.7 eV is mainly attributed to P-O bonding, which may be included in the compound such as (C6H5O)3PO. Most of O1s are probably the formation of O-P bonding, C-O-C or Si-O-Si bridges, and Si-O-C linkages.

Consequently, it is observed that the short-circuit current, open-circuit voltage, series and shunt resistance, filled factor, minority carrier lifetime and the conversion efficiency of the c-Si based solar cell were strongly associated with the surface states and the depth profile of the chemical compositions.

- 133 -

O1s

3.5x104

2.0x104

C1s

1.5x104

OKLL

C/S

2.5x104

Si2p

Si2s

3.0x104

P2p

1.0x104

O2s

5.0x103 1000

800

600 400 200 Bonding Energy (ev) XPS specatrum for the oxide/Si (111) surface of the wafer

0

Fig.1 Typical XPS of Si surface after P diffusion. 2x104 1x104 0

c/s

Si 4

-1x10

Si

-2x104 O C

-3x104 -4x104

0

200 400 600 800 1000 1200 1400 1600 1800 2000

kinetic energy (ev) AES spectrum for the oxide/Si (111) surface of the wafer

Fig.2 Typical AES of Si surface after P diffusion.

Reference: 1. Z.Q.Ma, Q.Guo and Tao.Jin, “Energy spectroscopy studies of radiation-induced damaged surfaces and interfaces in SiO2/Si by light charged particles” Nucl. Instr. & Meth. B71, (1992) 278-290.

- 134 -

Improvement of AZO/ SiO2/p-Si Based Violet Solar Cell B. He, Z.Q. Ma, F.Li, and X. J. Meng SHU-SEON’s PV Laboratory, Department of Physics, Shanghai University, Shanghai 200444, P.R.China e-mail:[email protected] Review paper ____________________________________________________________________ As shown in a previous report, the semiconductor-insulator-semiconductor (SIS) diodes have certain features which make them more attractive for solar energy conversion than conventional Shottky, MIS, or heterojunction structures. For example, the efficient SIS solar cells such as indium tin oxide (ITO) on silicon have been reported, where the crystal structures and lattice parameters of Si (diamond, a = 0.5431nm), SnO2 ( tetragonal, a = 0.4737nm, c = 0.3185nm), In2O3 ( cubic, a = 1.0118nm) show that they are partially compatible and thus not likely to form good devices. The SIS structure is potentially more stable and theoretically more efficient than either a Shottky or a MIS structure. The origins of this potential superiority are the suppression of majority-carrier tunneling in the SIS structure, the absence of a thin metal which absorbs light and is subject to environmental degradation, an induced junction and thin interfacial layer which minimizes the amount and impact of interface states, and the wide choice of conductivities and band gaps allowed in the top layer. In addition, the top semiconducting film can serve as an antireflection coating, a low-resistance window, as well as the collector layer of the junction. The SIS structure with a wide gap semiconductor as the top layer eliminates the surface dead layer associated with homojunction devices. This absence of absorption in a surface layer improves the ultraviolet response. In a device such as the SIS diode, the two semiconductors are separated by an interfacial region. The insulator is sufficiently thin (10-30Å) so that current transport through the interface is by tunneling. While the tunneling process is sensitive to a degree, to the sharp of the barrier, it is charge transport in the semiconductors that determines the I-V characteristics. The tunneling serves largely to provide an “ohmic contact”.

The following is a simple description of current transport in the device. The inverted p-type silicon surface provides a supply of minority carriers (elections) which can tunnel into the TCO

(

J CT = A T e *

2

−

qφB kT

e−

qφT d

). Al-doped ZnO (called AZO) layers are very frequently used as

window layers in photovoltaic solar cell.

In this application high electrical conductivity should be

combined with a low value of the optical absorption constant in the visible range. Thus an appropriate measure of the performance is the ratio and the optimization consists in realizing high by high carrier concentration and mobility with a minimum optical loss due to free carrier absorption at long wavelength. Such layers can be prepared by numerous techniques most frequently by magnetron sputtering either from compound targets or reactively from metallic targets. Optimized doped ZnO:Al-layers have resistivities - 135 -

−4

of 1.2 − 2 × 10 Ω ⋅ cm with Hall mobilities of up to 60 cm2/Vs, refractive indices of approximately 2.0 and an average transmittance of 85 % in the visible range. ZnO is required as window layer forming the front contact and has to combine high conductivity (low series resistance) with high transparency and excellent surface texture (milky appearance). The required surface texture can be realized with sputtered ZnO:Al films by a combination of a proper choice of the process parameters during sputtering and post deposition wet chemical etching in diluted hydrochloric acid HCl. With increasing substrate temperature and decreasing pressure there is a transition from a porous structure (tapered crystallites with voids) to a densely packed film (columnar structure). In the same direction the etching rate decreases. Films which are prepared at low substrate temperature and high deposition pressure have a porous rough surface structure and are homogeneously etched without surface texturing. For optimized surface texture resulting in optimized light scattering over a wide wavelength range the starting structure has to be densely packed with a smooth surface. Therefore, ZnO:Al is a good TCO layer. It has low resistivity, high transmittance, optimized surface texture for light trapping, big bang gap Eg ≈ 3.3eV and low cost. The cell structure is as Fig.1.

Fig.1 The solar cell with SIS structure and high efficiency.

For this structure of multi-layers system, the majority carriers are blocked from tunneling by the AZO band gap. Tunneling can also occur via defects states at the two interfaces, we regard JST to be the dominant tunnel transition via defects in AZO-Si.

JCI and JVI are the effective coupling current flows due to the interchange of the charge between the conduction and valence bands of the silicon by recombination-generation. The degree of tunneling is a function of the interface layer thickness d as equation:

= Voc

J qφB nkT + [ln( * L 2 ) + q AT kT

qφT ⋅ d ]

(1)

The Voc of SIS solar cell increase with interface layer thickness d. However, Jsc reduces with interface layer thickness which makes efficiency reduction. It is reported that the best thickness of interface layer may be 20 Å and theoretically the efficiency of AZO/SiO2/P-Si SIS solar cell at AM1 can reach 21%. While experimental value for this system has yield 16.5% in our study.

- 136 -

- 137 -

- 138 -

- 139 -

- 140 -

- 141 -

- 142 -

- 143 -

- 144 -

- 145 -