REPORT on Deng et al.,

" Measurement report: Size distributions of urban aerosols down to 1 nm from long-term measurements"

This manuscript reports the characteristics of atmospheric aerosol size distributions from ~1 nm to 10 μm with an emphasis on sub-3nm particles from four-year measurements in urban Beijing. On the basis of cluster analysis, three typical types of number size distributions (i.e., daytime NPF type, daytime non-NPF type, and nighttime type) were identified. Based on the modified Whitby model, the simplification of aerosol size distributions in sub-3 nm sizes, i.e., the power function, was introduced. From the source identification of sub-3nm particles in urban Beijing, except for NPF, vehicle emission is another important source. Besides, the concentrations and diel patterns of H2SO4 monomer and dimer were also reported to better explain the formation mechanisms of NPF. 

This is a unique dataset for which publication is worthwhile. This manuscript is well structured and written. Hence, I would ask one minor revision and recommend the publication of this article in Atmospheric Chemistry & Physics.

Minor comments:

Compared to lognormal distributions of larger particles, please add more discussion on why the power function was suitable to depict aerosol size distributions in sub-3 nm sizes. 

Figure 2, were H₂SO₄ monomer and dimer concentration converted into dN/dlogDp according to their mass diameters? 

Figure 3, the label of Y-axis “dNdlogDp” should be “dN/dlogDp”. Please check other figures in the manuscript and SI.

Figure 4, for panel C, if available, it would be better to add the diurnal patterns of solar radiation and SO₂.  Also, around 5 a.m., the concentration of sulfuric acid started to rise. Did sulfuric acid follow the diel pattern of radiation at around 5 a.m.?  If not, please comment on it.

I recommend the publication of this manuscript, but suggest the authors to address the following minor concerns.

1. line 120 

Your NPF definition looks like the classification of size distribution, not NPF. It's quite different from conventional NPF definition, like Dal Maso et al. Please describe the difference and how this will affect your conclusions.

2. line 129. why was power law function used? not other function? did you try any others? What information or contribution does the function provide on nucleation mechanistic, regional, and global atmospheric models?

3. line 150-165. Is the PNSD for the time slot before, during or after NPF? how did you pick up the PNSD that you named "typical" from a large set of data? The PNSD is always changing even during a single NPF event.

4. figure 4b where is Nsub-2 curve?

5. line 245-255. is there any evidence from nanoparticle/cluster chemical composition (e.g., H2SO4, amines, organics) measurement to support the source of these sub 3 nm particles were from traffic emission? for example, how were H2SO4 monomer and dimer observed in these traffic events?