Conventional photoluminescence (PL) yields at most one emitted photon for each absorption event. Downconversion (or quantum cutting) materials can yield more than one photon by virtue of energy transfer processes between luminescent centers. In this work, we introduce Gd 2 O 2 S:Tm 31 as a multi-photon quantum cutter. It can convert near-infrared, visible, or ultraviolet photons into two, three, or four infrared photons of ,1800 nm, respectively. The cross-relaxation steps between Tm 31 ions that lead to quantum cutting are identified from (time-resolved) PL as a function of the Tm 31 concentration in the crystal. A model is presented that reproduces the way in which the Tm 31 concentration affects both the relative intensities of the various emission lines and the excited state dynamics and providing insight in the quantum cutting efficiency. Finally, we discuss the potential application of Gd 2 O 2 S:Tm 31 for spectral conversion to improve the efficiency of next-generation photovoltaics. Keywords: downconversion; infrared emission; quantum cutting; solar cells; spectral conversions INTRODUCTION Over the last decade, advanced luminescent materials exhibiting downconversion, also known as quantum cutting or quantum splitting, have been developed 1,2 . In this process, a high-energy photon is converted into two or more lower-energy photons, with a quantum efficiency of potentially well over 100% 3 . If the downconverted photons are in the visible range, this concept is of interest for color conversion layers in conventional lighting applications 1,2,4-6 . New exciting possibilities of downconversion to infrared (IR) photons lie in next-generation photovoltaics, aiming at minimizing the spectral mismatch losses in solar cells [7][8][9][10][11] .Spectral mismatch is the result of the broad width of the spectrum emitted by the sun. Semiconductor absorber materials absorb only photons with an energy hn higher than the band gap E g 7 . To absorb many photons and produce a large electrical current, absorber materials must therefore have a small bandgap. However, because excited charge carriers rapidly thermalize to the edges of a semiconductor's conduction and valence bands, a high voltage output requires an absorber with a large band gap. Hence, materials with low transmission losses (leading to a large current) have high thermalization losses (leading to a low voltage) and vice versa 12,13 . These spectral mismatch losses constitute the major factor defining the relatively low ShockleyQueisser limit 14 , the maximum light-to-electricity conversion efficiency of 33% in a single-junction solar cell, obtained for a band gap of approximately 1.1 eV (1100 nm) 13 .Efficiencies higher than 33% can be reached with multi-junction solar cells, where a stack of multiple absorber materials are each optimized to efficiently convert different parts of the solar spectrum to