|The authors performed more data analysis and provide further information on the experiments. The answers to the reviewers are often general and interpretation of data still remain fairly speculative. There are two main issues.|
1) From the new Figure H1 it is obvious that there is a strong growth of the particle size distribution over time. However, their “new method” to determine size dependent wall loss requires a stable size distribution. Otherwise, the change of mass concentration in a size bin does not only depend on wall loss, but also heavily on growth. Thus, the eddy diffusivity coefficient and size dependent wall loss coefficient depend on both, wall loss and growth. The loss rate of BC is not affected by this. The reason that kBC and Beta(Dp) correlate is only due to the fact that Beta(Dp) is fit to kBC. The size dependent wall loss rate method is not independent of kBC and would also not work without the knowledge of kBC. The huge scatter of ke may be due to the inappropriate application. In conclusion, this method development does not provide what it is claimed for.
The authors also claim that the discrepancy in calculated ke between ammonium nitrate and vehicle emission experiments could be attributed to the nature of particles. What is the physics behind this statement? What factor in equation (3) should lead to such a pronounced effect?
2) The authors attribute the measured growth of PM to a prolonged condensation of IVOCs, as outlined in Appendix G. Indeed, IVOCs can partition into PM at given conditions. However, according to Figure 6, IVOC emissions are 10-20 times lower than PM emissions (Euro 3 and Euro 4). The increase in PM mass is 30-120% (Fig F1). Thus, there is no way enough IVOCs to contribute to this PM increase.
The paper does not provide a sound explanation of what the reason for the measured PM increase is. I believe this is an important issue as such a large effect, if real, would have far-reaching implications. For example, there are many studies on SOA formation of vehicle emissions. If primary emissions would show such a behavior, all/most of these studies could be heavily biased. Even emission factors, as determined in this paper, would be incorrect. One should take PM not after injection but at a much later time. This is not even addressed in the paper. I am not aware of chamber studies reporting such an effect. Thus, there is still the option of an instrumental or procedural artefact. One possibility could be an evolving change of particle density with coagulation of primary particles due to a high BC fraction.
In conclusion, the method of size dependent wall loss determination is inappropriate under these experimental conditions, the interpretation of measurements is not sufficiently supported or measurements might even have a bias. I cannot support a publication of this paper.