114 IEEE JOURNAL OF PHOTOVOLTAICS, VOL. 3, NO. 1, JANUARY 2013 Impurity-Related Limitations of Next-Generation Industrial Silicon Solar Cells Jan Schmidt, Bianca Lim, Dominic Walter, Karsten Bothe, Sebastian Gatz, Thorsten Dullweber, and Pietro P. Altermatt Abstract—We apply highly predictive 2-D device simulation to assess the impact of various impurities on the performance of next- generation industrial silicon solar cells. We show that the light- induced boron–oxygen recombination center limits the efficiency to 19.2% on standard Czochralski-grown silicon material. Curing by illumination at elevated temperature is shown to increase the efficiency limit by +1.5% absolute to 20.7%. In the second part of this paper, we examine the impact of the most important metallic impurities on the cell efficiency for p- and n-type cells. It is widely believed that solar cells on n-type silicon are less sensitive to metallic impurities. We show that this statement is not generally valid as it is merely based on the properties of Fe but does not account for the properties of Co, Cr, and Ni. Index Terms—Charge carrier lifetime, impurities, photovoltaic cells, semiconductor device modeling, silicon. I. INTRODUCTION I N the famous Westinghouse paper published in 1980 [1], the impact of metallic impurities had been experimentally studied on solar cells with a baseline efficiency of only 14%. As next-generation industrial solar cells will have a baseline efficiency exceeding 20% by implementing a rear surface passi- vation and/or a selective emitter, an update of the Westinghouse results is overdue. Our approach is different to the one of the Westinghouse group: We use state-of-the-art measured data of the recombination properties and apply highly predictive numer- ical device simulation using the SENTAURUS DEVICE instead of fabricating actual devices, which gives our analysis greater flexibility. Manuscript received May 11, 2012; revised June 15, 2012 and July 17, 2012; accepted July 17, 2012. Date of publication August 13, 2012; date of current version December 19, 2012. This work was supported by the German State of Lower Saxony; the German Federal Ministry for the Environment, Nature Con- servation, and Nuclear Safety; and industry partners within the research cluster “SolarWinS” under Contract 0325270E. J. Schmidt is with the Institute for Solar Energy Research Hamelin, D-31860 Emmerthal, Germany, and also with the Department of Solar Energy, Institute of Solid-State Physics, Leibniz University of Hannover, D-30167 Hannover, Germany (e-mail: j.schmidt@isfh.de). B. Lim, D. Walter, K. Bothe, and T. Dullweber are with the Institute for So- lar Energy Research Hamelin, D-31860 Emmerthal, Germany (b.lim@isfh.de; d.walter@isfh.de; k.bothe@isfh.de; t.dullweber@isfh.de). S. Gatz was with the Institute for Solar Energy Research Hamelin, D-31860 Emmerthal, Germany. He is now with the SolarWorld Innovations, D-09599 Freiberg, Germany (e-mail: sebastian.gatz@sw-innovations.de). P. P. Altermatt is with the Department of Solar Energy, Institute of Solid-State Physics, Leibniz University of Hannover, D-30167 Hannover, Germany (e-mail: altermatt@solar.uni-hannover.de). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/JPHOTOV.2012.2210030 Fig. 1. Schematics of (a) a standard screen-printed industrial solar cell with full-area Al-p + BSF and (b) a next-generation industrial PERC with local Al-BSF and selective emitter. Fig. 1(a) shows the current standard industrial silicon solar cell with contacts made by screen printing of silver paste at the front and aluminum at the rear. In the next generation of industrial cells, the full-area Al-p + back surface field (BSF) will be replaced by a dielectric rear passivation, as shown in Fig. 1 (b), reducing the rear surface recombination velocity to below 100 cm/s. In addition, the front contacts will be “pas- sivated” by implementing a selective n ++ emitter beneath the fingers. These measures increase the efficiency potential signifi- cantly from around 18% to 21%. However, it also makes the cell performance much more sensitive to recombination losses in the silicon bulk due to impurities. This impact will be analyzed in detail in the following. II. BORON–OXYGEN-RELATED IMPURITIES If boron and oxygen are simultaneously present in crystalline silicon, the bulk lifetime and, hence, the cell efficiency degrade during illumination due to the activation of boron–oxygen- related recombination centers [2]. This phenomenon is observed in solar cells made on boron-doped Czochralski-grown silicon (Cz-Si), but also in cells made on oxygen-rich multicrystalline silicon (mc-Si) [3]. The boron–oxygen center can be perma- nently deactivated by illuminating and annealing the solar cell at the same time [4], [5]. Fig. 2 (green symbols) shows an exper- iment where a standard industrial solar cell, as shown in Fig. 1 (a), is annealed at 140 ◦ C and at the same time illuminated at 1 sun. The initial cell efficiency of 18.2% degrades to 17.7% and then recovers to the initial value. The recovered cell efficiency 2156-3381/$31.00 © 2012 IEEE