Investigation On Ion Implantation Models Impact On I-V Curve And Thin Film Solar Cell Efficiency

Size: px

Start display at page:

Download "Investigation On Ion Implantation Models Impact On I-V Curve And Thin Film Solar Cell Efficiency"

Reynard Pope
6 years ago
Views:

1 Proceedings of the 7th WSES International Conference on Wavelet nalysis & Multirate Systems, rcachon, France, October 13-15, Investigation On Ion Imlantation Models Imact On I-V Curve nd Thin Film Solar Cell Efficiency F. JHNSHH, K.SOPIN, H.BDULLH, I.HMD, M. Y. OTHMN Solar Energy Research Institute Universiti Kebangsaan Malaysia Bangi Selangor Malaysia MLYSI S. H. ZIDI Gratings, Incororated 700 B Broadbent Pkwy., NE lbuquerque, NM UNITED STTES OF MERIC huda@vlsi.ehtt:// bstract: - Solar cell simulation could be useful for time saving and cost consumtion. Different models usually used for imlantation rocess that could affect on the final results. Imact of Dual Pearson, Gaussian and Monte Carlo imlantation models are investigated by SILVCO software for a tyical thin film solar cell and it is found although there are differences between the -n junction deths and net doing rofiles but there is no different in the final results and efficiencies. By time saving consideration during the comutation, the Dual Pearson imlantation model is suggested to be use in this case. Key-Words: - imlantation, modeling, thin solar cell, efficiency 1 Introduction good solar cell simulation involves all the best models for each art a manufacturing rocesses. Ion imlantation is one of the first stes in -n junction rocessing that could effect on the final results [1-]. nalytical models are based on the reconstruction of imlant rofiles from the calculated or measured distribution moments. There are four different analytical imlant models that consider for imlantation according the temerature, imurity, time and articles energy [3]. Gaussian imlant model that is using the Gaussian distribution, Pearson imlant model which calculate the asymmetrical ion imlantation rofile and more better Dual Pearson model that extend toward rofiles heavily affected by channeling [4-8].The statistical technique uses the hysically based Monte Carlo calculation of ion trajectories to calculate the final distribution of stoed articles [13,14]. Silvaco software as a wide alication in VLSI design and articular in solar cell was chosen in order to comare different imlantation models [5]. Mathematical roach.1 Gaussian Imlant Model There are several ways to construct 1D rofile. The simlest way is using the Gaussian distribution, which is secified by: C( = φ π ρ ( x R ) ex (1) where φ is the ion dose er square centimeter secified by the dose arameter. R is the rojected range. R is the rojected range straggling or standard deviation.. Pearson Imlant Model Generally, the Gaussian distribution is inadequate because real rofiles are asymmetrical in most cases. The simlest and most widely aroved method for calculation of asymmetrical ion-imlantation rofiles is the Pearson distribution [].The Pearson function refers to a family of distribution curves that result as a consequence of solving the following differential equation: df ( ( x a) f ( = () dx b0 + b1 x + b x in which f( is the frequency function. The constants a, b0, b1 and b are related to the moments of f( by: γ ( β + 3) a = (3)

Proceedings of the 7th WSES International Conference on Wavelet nalysis & Multirate Systems, rcachon, France, October 13-15, 007 134 (4β 3γ ) b0 = (4) b 1 = a (5) β γ 6 b = (6) where = 10β 1γ 18, γ

.3 Dual Pearson Model To extend alicability of the analytical aroach toward rofiles heavily affected by channeling, l Tasch [3] suggests the dual (or Double) Pearson Method.

2 Proceedings of the 7th WSES International Conference on Wavelet nalysis & Multirate Systems, rcachon, France, October 13-15, (4β 3γ ) b0 = (4) b 1 = a (5) β γ 6 b = (6) where = 10β 1γ 18, γ and β are the skew ness and kurtosis resectively..3 Dual Pearson Model To extend alicability of the analytical aroach toward rofiles heavily affected by channeling, l Tasch [3] suggests the dual (or Double) Pearson Method. With this method, the imlant concentration is calculated as a linear combination of two Pearson functions: C( = Φ1 f1( + Φ f ( (7) where the dose is reresented by each Pearson function f1,(. f1( and f( are both normalized, each with its own set of moments. The first Pearson function reresents the random scattering art (around the eak of the rofile) and the second function reresents the channeling tail region. Equation (7) can is restated as: C( = Φ[ Rf1( + (1 R) f ( ] (8) where Φ = Φ1 + Φ is the total imlantation dose andr = Φ1. Φ.4 Monte Carlo Imlant Model The most flexible and universal aroach to simulate ion imlantation in non-standard conditions is the Monte Carlo Technique [15]. This aroach allows calculation of imlantation rofiles in an arbitrary structure with accuracy comarable to the accuracy of analytical models for a single layer structure. This model based on the Binary Collision roximation (BC) and alies different aroximations to the material structure and ion roagation through it [14]. 3 Comuter Simulation SILVCO software s including TLS and THEN redicts the electrical characteristics of hysical structures by simulating the transort of carriers through a two-dimensional grid. To enter the structure and comosition of a solar cell into SILVCO, several arameters must be defined. These include the definition of a fine, two-dimensional grid, called a mesh Fig 1(a). Once the hysical structure of a solar cell is built in SILVCO, the roerties of the materials used in the cell must be defined. minimum set of material roerties data includes: band ga, dielectric constant, electron affinity, densities of conduction and valence states, electron and hole mobilities, otical recombination coefficient, and an otical file containing the wavelength deendent refractive index n and extinction coefficient k for a material. THEN includes a wide selection of models that can be emloyed in device simulations. These models include the imlantation models that caable to be used in roose. The imlantation stage, annealing rocess, electrode definition are introduce in Fig. 1 (b) to (d) resectively. Photo generation and recombination rates are shown in Fig. 1 (e) and 1 (f). Sectral resonse and internal, external and total quantum efficiency are shown in Fig. 1 (g) to 1 (i) resectively for comarison. In order to define the I-V curve, a subroutine is used by changing the oen circle voltage. The short cut current could be found as it will be introduce in next section. (a) (b)

Proceedings of the 7th WSES International Conference on Wavelet nalysis & Multirate Systems, rcachon, France, October 13-15, 007 135

1: Different outut stages in solar cell simulation including (a) mesh definition (b) imlantation (c) annealing (d) electrode

3 Proceedings of the 7th WSES International Conference on Wavelet nalysis & Multirate Systems, rcachon, France, October 13-15, (c) (g) (d) (h) (i) (e) (f) Fig. 1: Different outut stages in solar cell simulation including (a) mesh definition (b) imlantation (c) annealing (d) electrode definition (e) hoto-generation rate (f) recombination rate (g) sectral resonse (h) external and total quantum efficiency (i) internal and total quantum efficiency. 4 Results and Observation In order to achieve the concentration rofiles for solar cell, a test model were used in µm silicon which boron concentration is 0.5 Ω and its orientation is [100] was selected. By imlant of hoshor with.5 x and energy 10eV, the -n junction deth about 0.1 µm was formed under the to surface. The diffuse time is

Proceedings of the 7th WSES International Conference on Wavelet nalysis & Multirate Systems, rcachon, France, October 13-15, 007 136 considered 10 min and the temerature is considered 850 C.

4 Proceedings of the 7th WSES International Conference on Wavelet nalysis & Multirate Systems, rcachon, France, October 13-15, considered 10 min and the temerature is considered 850 C. In order to finding the I-V curve, the dimensions were increased to µm because of accuracy. By using THEN software we try to change any imlantation model sequentially. The illumination is considered as in geometrical otics as ray tracing. We try to trace the ray with only 90 incident angle. Fig. (a), (b) and (c) show the comuter simulations of solar cell concentration rofile of net doing according to Dual Pearson, Gaussian and Monte Carlo Models with µm, µm and µm -n junction deths under the to surface resectively. In Fig. (3) shows the outut of I-V curve regarding to these three models demonstrated. It shows the all the I-V grah have the same size and shae that introduce the same efficiency. (a) (b) (c) Fig.: Concentration rofile of net doing in (a) Dual Pearson Model (b) Gaussian Model and (c) Monte Carlo Model Fig.3: Outut of I-V curve regarding to Dual Pearson, Gaussian and Monte Carlo models. 5 CONCLUSIONS lthough the -n junction deths in three models are different and the net doing rofile introduce different size and shae but simulation in I-V curve grahs are shown the same results for all three models and the same efficiency. These results shows there is not imortant which models should be used in our thin film solar cell simulation but may be because of time consumtion it is better to use Dual Pearson model. References: [1] J. Lindhard, M. Scharff, and H.E. Schiott. Range Concets and Heavy Ion Ranges, Kgl. Dan. Vid. Selsk. Mat.-fys. Medd., v. 33, [] D.G. shworth, R. Oven and B. Mundin, Reresentation of Ion Imlantation Profiles be Pearson Frequency Distribution Curves, J. Phys. D, v. 3,. 870, 1990 [3]. F. Tasch, n Imroved roach to ccurately Model Shallow B and BF Imlants in Silicon, J. Electrochem. Soc., v. 136,. 810, [4].F. Burenkov, F.F.Komarov, and M.M.Temkin. nalytical Calculation of Ion Imlantation through Mask Windows (in Russian), Microelektronika, v. 16,. 15, [5] S. Michael, "Silvaco tlas as a solar cell modeling tool," IEEE,. 719, 005. [6].F. Burenkov,.G. Kurganov, and G.G. Konolyanik, Two-Dimensional Local Ion Imlantation Distribution (in Russian), Povekhnost (Surface Sciences), v.8,.5, [7] J. Lorenz, W. Kruger, and. Barthel Simulation of the Lateral Sread of Imlanted Ions: Theory, NSECODE-VI, Ed. J.J.H.Miller, Boole Press,.513, [8] D.G. shworth, M.D.J. Bowyer, and R. Oven,

5 Proceedings of the 7th WSES International Conference on Wavelet nalysis & Multirate Systems, rcachon, France, October 13-15, Reresentation of Ion Imlantation Distributions in Two and Three Dimensions, J. Phys. D, v. 4,. 110, [9] G. Hobler, E. Langer, and S. Selberherr, Two- Dimensional Modeling of Ion Imlantation with Satial Moments, Solid-State Electronics, v. 30,. 445, [10] M. Temkin and I. Chakarov, Comutationally Effective Model for D Ion Imlantation Simulation, Semiconductor Process and Device Performance Modeling, Eds, S.T. Dunham, J.S. Nelson, MRS,. 7, [11] J. F. Ziegler, J. P. Biersack, U. Littmark, The stoing and range of ions in solids, v. 1, Pergamon Press, [1] I. R. Chakarov and R. P. Webb, CRYSTL -- Binary Collision Simulation of tomic Collision and Damage Buildu in Crystalline Silicon, Radiation Effects, v ,. 447, [13]. Phillis and P. J. Price, Monte Carlo Calculations on Hot Electron Energy Tails, l. Phys. Lett., v. 30, 1977 [14] S.H.Yang, D. Lim, S. Morris, and. F. Tasch, More Efficient roach for Monte Carlo Simulation of Deely-Channeled Imlanted Profiles in Single-Crystal Silicon, Proc. NUPD,. 97, [15]S.Franssila Introduction to Micro Fabrication,Wiley, 004

Analysis and Design of Si Terahertz Transit-Time Diodes

Analysis and Design of Si Terahertz Transit-Time Diodes Xiaochuan Bi, Jack R. East, Umberto Ravaioli and George I. Haddad Abstract This aer resents a numerical simulation of a Si MITATT diode working in