Abstract CO2 ice has a phase transition at 30 K when its structure changes from amorphous to crystalline. Using reflection absorption infrared spectroscopy, Öberg et al. observed that the photodesorption yield of CO2 ice deposited at 60 K and irradiated at 18 K is 40% lower than that of CO2 ice deposited and irradiated at 18 K. In this work, CO2 ices were deposited at 16–60 K and UV-irradiated at 16 K to rule out the temperature effect and figure out the relationship between photodesorption yield and ice structure. IR spectroscopy is a common method used for measurement of the photodesorption yield in ices. We found that undetectable C atoms produced in irradiated CO2 ice can account for 33% of the amount of depleted CO2 molecules in the ice. A quantitative calibration of QMS was therefore performed to convert the measured ion current into photodesorption yield. During various irradiation periods, the dominant photodesorbing species were CO, O2, and CO2, and their photodesorption yields in CO2 ices deposited at different temperature configurations were almost the same, indicating that ice morphology has no effect on the photodesorption yield of CO2 ice. In addition, we found that the lower desorption yield reported by Martín-Doménech et al. is due to a linear relationship between the photodesorption yield and the combination of the energy distribution of the Microwave-Discharge Hydrogen-flow Lamp and UV absorption cross section of ices.