The authors have paid attention to both sets of reviewer comments in revising their manuscript which will make a valuable contribution to the literature.

I have one word of caution that should be recorded and I would urge the authors to consider. In the response to reviewer 2, it is explicitly recognised that "nitrate partitioning heavily depends on the prevailing atmospheric conditions and therefore emission factors obtained in the PAM cannot be directly scaled to the atmosphere". It is implicit in this statement that this should be contrasted with SOA, the emission factors of which can supposedly be directly scaled to the atmosphere from the PAM reactor in some way. I would urge considerable caution in such a scaling, given the very different chemistry and partitioning regimes. Peng and Jimenez 2017 (doi:10.5194/acp-2017-266) described the regimes of chemical safety within the PAM reactor (or OFR in their terminology), but did not aim to do so with respect to formation of particle mass. The partitioning to form this mass will be dependent upon the chemical regime in addition to the partitioning being dependent on same prevailing atmospheric conditions as nitrate where there is any semi-volatile organic component. As editor, I will not insist on an extensive discussion of this being included in the manuscript, but would advise the authors to reflect on why they consider nitrate emission factors from a PAM reactor to be unreliable, but SOA emission factors to be atmospherically scalable. In this reflection they might think about the fact that SOA scaling will depend upon both the appropriateness of extrapolating from both chemical and physical conditions far from ambient.

I'd like to that the authors for their indulgence in making a sufficient attempt to address my concerns in their revision. Whilst I am still unconvinced by our supposed confidence in the nitrate formation kinetics and thermodynamics in the real atmosphere and its dependence on environmental conditions, I am convinced that the authors have appreciated the difficulties associated with the PAM extrapolations of SOA yield.
The paper is acceptable for publication.