Investigation of isotope effects in ozone (O3) photolysis and its contribution to the overall ozone isotope composition is difficult since photolysis always leads to secondary O3 formation and O3 decomposition by reactions with O(3P). Here we use a large excess of carbon monoxide (CO) as O(3P) quencher to suppress O(3P) + O3. This allows disentangling the isotope effects in photolysis and chemical removal when the data are evaluated with a kinetic model. The largest systematic uncertainty arises from an unidentified O3 removal reaction, which is responsible for an unaccounted 20% of the total removal rate. Assuming no isotope fractionation in this reaction, we find inline image = (16J/18J − 1) = −16.1 (±1.4)‰ and inline image = −8.05 (±0.7)‰ for O3 photolysis and inline image = (16k/18k − 1) = −11.9 (±1.4)‰ and inline image = −5.95 (±0.7)‰ for chemical removal via O(3P) + O3. Allowing for isotope fractionation in the unidentified reaction results in lower fractionation values for photolysis and higher fractionations for chemical removal. Several fractionation scenarios are examined, which constrain the fractionation in photolysis to inline image > −9.4‰ and inline image > −4.7‰ and in the chemical removal to inline image < −18.6‰ and inline image < −9.3‰. Both fractionations are thus significant and of similar magnitude. Because our measurements are dominated by photolysis in the peak region of the Chappuis band, isotope fractionation of atmospheric O3 by visible photons should also be in the same range. The isotope fractionation factor for O + O3 directly bears on ozone chemistry in the lower thermosphere.