A flow-through photoreactor containing a KrCl exciplex lamp has been used for the study of 2,4-dichlorophenoxyacetic acid (2,4-D) removal. Taking into account 2,4-D removal it is possible to affirm that the variation of the flow rate does not have a big influence on the percentage of 2,4-D removal. 2,4-D is faster removed when its initial concentration is lower and the presence of hydrogen peroxide improves the removal efficiency until reaching the mass ratio H2O2:2,4-D 2:1. Fixing this mass ratio H2O2:2,4-D at 2:1, and simultaneously varying 2,4-D and hydrogen peroxide concentrations it is obtained that the time needed for degrading 2,4-D increases when the initial concentration of the compound increases. Besides, with the biggest sample volume the conversion of 2,4-D decreases. Photoproducts, in general, follow the same tendency that the initial 2,4-D compound. It can be observed that the maximum of the compounds appears in the same order that each compound appears in the mechanism. Firstly the primary photoproducts: 2,4-dichlorophenol, 4-chlorophenol and 2-chlorophenol, and later hydroquinone and benzoquinone. A possible reaction pathway based on the primary formation of 2,4-dichlorophenol and chlorophenols is proposed. Finally, and taking into account the dependence of the kinetic constant with the experimental conditions a model based on a pseudo-first order kinetic has been formulated. 2,4-Dichlorophenoxyacetic acid and photoproducts theoretical conversion values were plotted against the experimental ones. A good degree of agreement was reached (X-cal = 0.9991 X-exp, R-2 = 0.9988) confirming the validity of the proposed model. (C) 2015 Elsevier B.V. All rights reserved.