PROGRESS IN PHOTOVOLTAICS: RESEARCH AND APPLICATIONS Prog. Photovolt: Res. Appl. 2002; 10:527–532 (DOI: 10.1002/pip.451) SHORT COMMUNICATION Concerning the Relationship Between Clear-Sky, Global and Direct-Beam, Solar Spectra David Faiman 1, * ,y , Sareet Jacob 1,2 and Arnon Karnieli 1 1 Department of Solar Energy and Environmental Physics, Jacob Blaustein Institute for Desert Research, Ben-Gurion University of the Negev, Sede Boqer Campus, 84990 Israel 2 Department of Chemistry, Brown University, Providence, RI 02912, USA Clear-sky direct-beam and normal global solar spectra were measured with a spectro- radiometer with and without a collimating tube. The ratio of the two spectra was found to be well approximated by a simple two-parameter formula suggested by Rayleigh scattering. Copyright # 2002 John Wiley & Sons, Ltd. INTRODUCTION B ecause of the different spectral sensitivities of PV materials, the need to establish standard reference spectra has long been appreciated. 1 Much effort was accordingly invested in modeling 2 and measuring 3 solar spectra under various atmospheric conditions, leading eventually to the international adoption of the so-called AM15 G standard global spectrum 4 for PV purposes. Recently, with renewed interest in the potential economic benefits of concentrator PV cells (CPV), there would appear to be a pressing need for an internationally accepted standard spectrum for the direct beam component of solar radiation, i.e., an AM15 D standard. The purpose of this paper is, therefore, to present the first stage of an experimental study of the relationship between natural global and direct-beam solar spectra measured on clear days at various times of the year, at Sede Boqer in the Negev Desert (Lat. 308  N, Lon. 348  E, Alt. 475 m). The study was prompted by the fact that this particular site is known 5 to have natural clear-sky spectra in close conformity with the established, albeit computer-generated, AM15 G standard spectrum. 4 Our motivation was to investigate the possible existence of a simple empirical relationship between global and direct spectra. For, if one were discovered to exist, it would not be necessary to define two independent spectra for PV and CPV measurements: the existing AM15 G stan- dard 4 would suffice, mutatis mutandis, for both purposes. To the best of our knowledge, none of the previously published work, whether measurement-based or model-based, has discussed the ratio of the direct-beam and global spectra. EXPERIMENTAL METHOD For our spectral measurements we employed a Li-Cor 1800 spectroradiometer, having a wavelength range 300– 1100 nm and a passband of 6 nm. The instrument had been recently re-calibrated by the manufacturer. Global Published online 17 September 2002 Received 21 January 2002 Copyright # 2002 John Wiley & Sons, Ltd. Revised 5 April 2002 * Correspondence to: David Faiman, Department of Solar Energy and Environmental Physics, Jacob Blaustein Institute for Desert Research, Ben-Gurion University of the Negev, Sede Boqer Campus 84990, Israel. y E-mail: faiman@bgumail.bgu.ac.il Research