Measurement of the SO3 concentration in flue gas is important to estimate the acid dew point and to control corrosion of downstream equipment. SO3 measurement is a difficult question since SO3 is a highly reactive gas, and its concentration is generally two orders of magnitude lower than the SO2 concentration. The SO3 concentration can be measured online by the isopropanol absorption method; however, the reliability of the test results is relatively low. This work aims to find the error sources and to evaluate the extent of influence of each factor on the measurement results. The test results from a SO3 analyzer showed that the measuring errors are mainly caused by the gas-liquid flow ratio, SO2 oxidation, and the side reactions of SO3. The error in the gas sampling rate is generally less than 13%. The isopropanol solution flow rate decreases 3% to 30% due to the volatilization of isopropanol, and accordingly, this will increase the apparent SO3 concentration. The amount of SO2 oxidation is linearly related to the SO2 concentration. The side reactions of SO3 reduce the selectivity of SO42- to nearly 73%. As sampling temperature increases from 180 to 300 degrees C, the selectivity of SO42- decreases from 73% to 50%. The presence of H2O in the sample gas helps to reduce the measurement error by inhibiting the volatilization of the isopropanol and weakening side reactions. A formula was established to modify the displayed value, and the measurement error was reduced from 25%-54% to less than 15%. (C) 2017 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences. Published by Elsevier B.V.