IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 61, NO. 6, JUNE 2014 1943 Characterization of Laser-Doped Localized p-n Junctions for High Efﬁciency Silicon Solar Cells Andreas Fell, Sachin Surve, Evan Franklin, and Klaus J. Weber Abstract— To further increase the efﬁciency of industrial crystalline silicon solar cells, a point-contact solar cell concept with localized p-n junctions is considered a promising candidate if implemented by a low cost processing technique like laser doping. For efﬁcient development and optimization of such a process- ing technique, we present a dedicated test structure to derive the fundamental diode characteristics speciﬁc to the localized p-n junction, namely the contact resistance to the metal and the recombination properties, i.e., the dark saturation current. Those properties are ﬁtted to measured dark current–voltage curves by 3-D device simulations using Quokka. We show that in particular, the contact resistance can be accurately extracted and that the method is robust against uncertainties of other device properties of the test structure. Simulations of an idealized point-contact solar cell are performed to judge the usefulness of the extractable value range with respect to the efﬁciency potential. Furthermore, we apply the method to laser doping experiments. We successfully characterize the recombination and contact resistance and identify a ∼24% efﬁciency potential of a nonoptimized two-step laser doping process. Other single step processes show a very high recombination ( J 0pn ≫ 1e -10 A/cm 2 ) likely due to imperfections around the perimeter of the laser processed area. Index Terms— Conductive boundary, modeling, quasi- neutrality, Quokka, simulation, solar cell. I. I NTRODUCTION A WELL-KNOWN concept for achieving a very high efﬁ- ciency crystalline solar cells is the use of localized p-n junctions (emitters) instead of the state-of-the-art large area diffusions [1]. To maintain a high collection efﬁciency of minority carriers, and hence a high short circuit current, the localized p-n junctions are required to be closely spaced. This prevents a front side application due to the difﬁculty of avoiding a high metallization fraction and thus a high shading loss. Therefore, localized p-n junctions require a rear junction concept with a high minority carrier diffusion length bulk to beneﬁt from the high efﬁciency potential. One challenge associated with any localized p-n junction concept is to have sufﬁciently isolating dielectric ﬁlms where the p-type metallization, connecting each of the local junctions, covers the n-type bulk [2]. A further challenge is to ensure that the Manuscript received February 11, 2014; revised April 2, 2014; accepted April 10, 2014. Date of publication May 6, 2014; date of current version May 16, 2014. This work was supported by the Australian Renewable Energy Agency under Project 5-F007 and Project 3-GER002. The review of this paper was arranged by Editor A. G. Aberle. The authors are with the School of Engineering, College of Engineer- ing and Computer Science, Australian National University, Canberra, ACT 0200, Australia (e-mail: andreas.fell@anu.edu.au; sachin.surve@anu.edu.au; evan.franklin@anu.edu.au; klaus.weber@anu.edu.au). 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/TED.2014.2318714 localized doping and contact opening process does not lead to direct contact between the metal and the bulk, in particular caused by imperfections at the contact perimeter. One major advantage of this concept is a relaxed require- ment on the recombination properties of the p-n junction. Because a typical optimum contact area fraction is in the order of 1% [3], recombination can be in the order of 100 times higher compared with a high efﬁciency full area emitter. Similarly, the collection of carriers generated within the doped region is less critical compared with a front side application (often referred to as emitter blue response), as at the rear much less current is generated, allowing for deeper and more heavily doped junctions. Those relaxed requirements motivate the use of low-cost laser processing to form the p-n junc- tions, as opposed to otherwise more complex patterning and diffusion/implantation sequences. Such a cell was fabricated in [4], where they used a laser to form localized Al-alloy p-n junctions from an evaporated Al layer through a dielectric passivation layer (laser ﬁred emitter), achieving an efﬁciency of 19.6% on 100-cm n-type material. This paper deals with the development of localized laser- doped p-n junctions on n-type bulk material. To assess the efﬁ- ciency potential of a speciﬁc laser process, the key properties of the localized p-n junction must be characterized. Evaluation of the metal-silicon contact resistivity, the shunt behavior from metal contact to bulk, and the recombination properties is required. Each of these properties is likely to be strongly spatially varying within the laser affected area A pn , and is thus best described by the area integrated quantities dark saturation current I 0pn and contact resistance R cont . However, the more common notations J 0pn (= I 0pn / A pn ) and r cont (= R cont / A pn ), that is the area average dark saturation current density and contact resistivity, are used as a suitable input for modeling and for better comparability with other work. Neglecting the spatial variation within relatively small feature sizes is valid, as discussed, for example, in [5] Note that as in this paper aluminum metallization is used, the mentioned shunt to the n-type bulk cannot be ohmic but rather forms a Schottky junction [6], which is in parallel to the p-n junction and has the same direction as long as n-type bulk material is used. There- fore, no shunt resistance is considered in the analysis. The high recombination of the Schottky diode [7] is effectively lumped together with the p-n junction recombination into J 0pn , which is, however, best applicable to assess the overall quality of the laser processed region with respect to solar cell performance. For the efﬁcient development of suitable laser processes, it is very valuable to perform the characterization by simpliﬁed test structures rather than by complete solar cells. Standard 0018-9383 © 2014 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.