Note that there is an Erratum correcting a typo in the paper. I also listed this and all other mistakes in chemical rates below. All corrections are minor, and do not have a significant impact on the results in the paper. However, if you want to implement the chemical network yourself, it is always better to use the corrected rates. I try to keep this list and the PDR code up to date. I owe my thanks to Simon Glover and Jeong-Gyu Kim, who spotted many of these mistakes.

• Reaction 1 in Table 1: In the factor in front of the summation term, the exponent of T should be −3/2 instead of 3/2. (In the Erratum)

• Reaction 19 in Table 1: The factor leading the rate equation should be 1.76e-9 instead of 2.84e-9.

• Reactions 11 and 12 in Table 2: Minor corrections of rate coefficients 560 -> 520 (reaction 11) and 90 -> 92 (reaction 12). These values are twice of those in Heays et al. (2017), since their coefficients are based on the primary CRIR of H2 instead of H. Note that all CR included photoionisation rates (reactions 11-13) are multiplied by a factor of H2 abundance x_H2 (see Appendix A.1 point 7).

• The collisional de-excitation coefficients for H2 UV pumping heating and H2 formation on grains heating is updated to the more recent rates from Visser et al. (2018).

• For CR heating in Eq. (29), \xi here should be the primary rate per H nucleus, not the total CR ionization rate.