Lawrence Berkeley National Laboratory Recent Work Title ON THE ENHANCEMENT OF SILICON CHEMICAL VAPOR DEPOSITION RATES AT LOW TEMPERATURES Permalink https://escholarship.org/uc/item/9pp8142m Author Chang, Chin-An. Publication Date 1976-08-01 eScholarship.org Powered by the California Digital Library University of California u u \J ·i,) "'i .j u ;;) 6 6 9 t({!,-6S Published in Journal of Electrochemical LBL-3938 Rev. Society, Vol.123, No. 8, 1245- 1247 Preprint C. I {August 1976) ON THE ENHANCEMENT OF SILICON CHEMICAL VAPOR DEPOSITION RATES AT LOW TEMPERATURES Chin-An Chang· ·._\,::~FL~:?·{ ;'-~"'-;!'~·~ August 1976 ~. t ~; ... · j\/{ ~;:""" ·...J · r ~-.:, .~· r.-.:... c·. - · r\1 Prepared for the U •. S. Energy Research and. Development Administration under Contract W -7405-ENG -48 For Reference Not to be taken from this room DISCLAIMER This document was prepared as an account of work sponsored by the United States Government. While this document is believed to contain conect information, neither the United States Government nor any agency thereof, nor the Regents of the University of California, nor any of their employees, makes any wananty, express or implied, or assumes any legal responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by its trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof, or the Regents of the University of California. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof or the Regents of the University of California. 0 6 0 Reprinted from JOURNAL OF THF ELECTROCHEMICAL SOCIETY Vol. 123, No. 8, August 1976 Printed in U.S.A. Copyright 1976 I.AWRENC£ BERKElEY LABORATORY ft&MNT NUMIICft · 1976 3 &8-I -, UNIVERSITY Of CALIFORNIA ,;, ,. I On the Enhancement of Silicon Chemical Vapor Deposition Rates at Low Temperatures •. ¥ Chin-An Chang 1 Mater.ials arid Molecular Research Division, Lawrence Betkeley Laboratory, Berkeley, California 94720 '' Chemical vapor deposition (CVD) has been a widely charge carriers due to trapping at the grain boundaries used technique for thin film device fabrication. It is (6) ; low deposition rate makes CVD a noneconomic especially applicable to thin film silicon solar cells. technique 'for deposition. To solve the former prob­ A deposition rate of a few microns per minute can be lem, this ·author has developed a technique to increase easily obtained, and a p-n junction can be made by the silicon crystallinity at low substrate temperatures. mixing silicon chemical vapors, such as silane and An enhancement of two to four orders of magnitude is silicon tetrachloride, with dopant gases, like diborane obtained for the silicon films deposited on quartz and and phosphine. In general, CVD of silicon is carried out graphite at a 600oC substrate temperature (2, 7). The at a substrate temperature ca. 1000o-1200oc, and a next goal is to increase the deposition rate of low tem­ single crystal silicon wafer is used for an epitaxial perature CVD. ln doing so we make use of the obser­ growth of thin films. However, to make economic thin vations that certain dopant gases enhance the silicon film solar cells for terrestrial photovoltaic applications, CVD rate and others hinder it (8). An understanding noncrystalline and often nonsilicon substrates are re­ of these facts will obviously be usefu\ to the ~nhance­ quired. Furthermore; interaction between the substrate ment of the silicon CVD rate at low temperatures. thus chosen and the silic'on thin film deposited should However, as pointed out later, these dopant gas effects be kept minimal. For example, at a substrate tem­ cannot yet be consistently accounted for by the existing perature ca. 1200°C, silicon thin films deposited on theories. Therefore, this paper presents a conceptual graphite show a significant diffusion ·of silicon and model which consistently explains the known dopant carbon and the formation of silicon carbide ( 1). Much gas effects on the rate of silicon CVD. We believe that less diffusion is noted, however, when the substrate this work will not only lead to an enhancement of the temperature is below 800°C (2). Similar high tempera­ low temperature silicon CVD rate needed for economic ture interaction between the silicon films deposited and solar cells, but also be useful to guide experimentalists other types of substrate has also been reported (3). in carrying out the CVD process more effectively and to Preferably, one should use the lowest possible substrate stimulating interest among theorists about the prob­ temperatures to minimize such interactions and diffu- lems related to this topic. The proposed model is fur­ sion. ther tested for its predicted dopant gas effects on differ­ At low substrate temperatures, however, other ent CVD systems and other relevant work which needs problems arise. First, silicon film thus deposited is poly­ to be done is also discussed. crystalline with small grain size ( 4). Second, a much It is well known that the silicon deposition rate using lower deposition rate than that at high substrate tem­ silane and silicon tetrachloride is increased by di­ perature is obtained using the CVD technique (5). borane and boron tribromide, and decreased by ph0s­ Small grain size means a shortening of lifetime for the phine and arsine (8). Existing theories considering 1 Present address: IBM Thomas J. Watson Research Center, either active site blocking· (8b) or strong bonding be­ Yorktown Heights, New York 105gB, Key words: silicon, CVD, enhanced rates, dopant effects, sur­ tween the chemical vapor molecules and. both types of face model. dopant gas molecules (8c) have not been advansed 1246 J. EZectrochem. Soc.: SOLID-STATE SCIENCE AND TECHNOLOGY August 1976 · enough to explain the known effects in a consistent way, ever, the ionic character mentioned would make silane nor can they be used to predict the dopant gas effect on a molecule with four negatively charged hydrogen other CVD systems. In this work we make use of the atoms surrounding a positively charged silicon atom in fact that diborane and boron tribromide, both p-type the center. A molecule of such polarity would be at­ dopant gases for silicon, have an opposite effect on the tracted by a surface with positive surface potential and deposition rate of silicon chemical vapor from that of be repelled from one with negative surface potential. phosphine and a~sine. whi:Ch; are· n-,ty,pe dopant gases. In other words, making the surface potential more This implies ~a· pos~ible c{>,rr~~~tion between the elec­ positive would enhance silane adsorption on the sur­ tronic structure o·~ the d.op'imt atoms with the observed face. Tqe same argument applies to silicon tetrachloride effects cited. The mechanism involved in the CVD which has Si + -Cl- ionic bonds, similar to those in process is, therefore, first' analyzed, in order to see silane. whether and how this property can be incorporated First, we define a reference surface to be a silicon into the. depositi()~ prp~~~s; , i. , l ~·.. · . surface with only adsorbed silicon atoms. This is the Chemical vapor depositiOn can be viewed as a two­ case when pure silicon chemical vapor is used. A step process: adsorption of the chemical vapor mole­ boron-adsorbed silicon surface can be seen to be dif­ cule on the substrate surface followed by its thermal ferent from the reference one. Boron, being a p-type decomposition. As an example, deposition of silane fol- dopant and electron deficient relative to silicon, should, lows · relative to the reference surface, lower the local elec­ heat . / tron density on the surface silicon atoms around the SiH4(g) ~ SilLI(adsorbed) ~ Si(s) + 2H2(g) 1 l adsorption site. This would increase the electron affinity of the local silicon surface around the adsorbed boron The thermal decomposition involves a transfer 'of ( 12). According to Allen and Gobeli, an increase in thermal energy from the substrate· to the chemiCal surface electron affinity increases the positive surface vapor molecule needed for its decomposition. At a fixed potential (12). This has been observed on both Si and substrate temperature, the maximal amount of thermal Ge (9, 12). A boron-adsorbed silicon surface would energy which can be acquired by the molecule is con­ thus attract molecules like silane and silicon tetra­ stant. The efficiency of this energy transfer, however, chloride better than the reference surface. This should. depends on the residence time of the molecule on the enhance the adsorption and, therefore, the deposition substrate surface. In other words, for an effective de­ rate of these molecules. On the other hand, phosphorus composition of the molecule to take place on the sur­ and arsenic, which are n-type dopants and electron­ face, a sufficient amount of thermal energy for such excessive relative to .silicon, would make the surface decomposition should be transferred to the molecule potential more negative than the reference one. Ac­ before it desorbs from the surface. Therefore, at a fixed cordingly, the deposition rate of silane and silicon substrate temperature, adsorption of the molecule de­ tetrachloride should be lowered when they are mixed termines its decomposition rate. This argument is in with phosphine and arsine. The observed dopant gas agreement with the work of Farrow who found that effect on the deposition rates of silicon chemical vapor silane adsorption is the rate-limiting step for the de­ is thus satisfactorily and consistently explained. These composition of this molecule (8c). Accordingly, at a results are summarized in Table I.