Kuula et al. show a new setup using the tandem DMA and 3-wavelength Cantilever-enhanced photoacoustic spectrometer to measure size-resolved aerosol light absorption properties. Overall, the manuscript is well-written, and the setup is novel. I have some minor comments to help authors improve their paper. Please see my comments below.

Major comments:

Why do you use NO₂ not cab-o-jet or other standards? How did you calibrate the CEPAS? Your difference between measured and model values at different wavelengths are not consistent, which could be due to the wrong calibration method. (see “Characterization of light-absorbing aerosols from a laboratory combustion source with two different photoacoustic techniques”).

The specifics of your Cantilever PAS are unclear to me. I suggest a table summarizing your cell length, resonance frequency, quality type, laser power, mean diameter, detection limit, response time, etc. Moreover, it is also unclear to me in L94-95, where you said the measurement point is approx. 10-20 S. Does that refer to one data point that will 10-20 s to collect?

I feel the discrepancy between your measured and model-predicted light absorption properties could be attributed to the calibration method and multiple-charge particles. Please comment on this.

Minor comments:

L126-127, “For a full … checks).” It is not clear what the 12 steps mean here. Are you referring to 12 different particle sizes? If so, could you provide these sizes?

L144-145, “The observed … modeling results.” This is not very clear to me. Typically, measurements is more reliable than models. And what artifact are you referring to here? Does that mean your measurements are not reliable at all?

L145-147, “The root …  the CEPAS.” Why did you use this small tubing? I do not quite understand why smaller tubing is the reason for that. Your CEPAS should always be under pressure since it is connected to a pump. Did you measure the pressure? And if you think this is the reason for the artifact, why didn't you fix that?

L182-183, “The reference … 1:10.” What's the RH after the dryer? RH can affect the PAS measurements.

Table 1, what are the references for nigrosine refractive index.

L209-211, “With respect … Mie-modelled values.” It is unclear to me why you observed that inconsistent discrepancy. I would like to see your explanation.

The paper describes the performance of a three-wavelength cantilever-based photoacoustic spectrometer by comparing absorbing particles from the nebulization of nigrosine ink with Mie calculations. The cantilever technique has great potential, and I think it holds great promise for the future, for that reason I loosely followed the development of this new technology over the past years with interest, and I was excited to have to review the paper. However, I found the paper to be misleading on some key aspects (as discussed below) as well as very dismissive with regard to extensive work done in this field in the past and fully available in the literature. It should not be too difficult to correct these issues, but before the paper can be accepted, a serious effort must be made to address these problems.  

General Comments

-    The introduction and motivation of the work focus mostly on the aging, citing, and absorption enhancement of black carbon particles. However, the paper presents nothing on any of these topics. The paper shows measurements on nigrosine (not black carbon), spherical (not aggregates), and not coated. I do not doubt that the technique can be applied to study the aging, coating, and absorption enhancement of black carbon in the future. Still, the motivations and introduction provided in this work do not match what has actually been done here. The abstract, introduction, and conclusion should make this point clear.

-    The paper, including the introduction and especially Section 3.3, almost completely ignores (with just a couple of exceptions) many other previous photoacoustic experiments (some more than a decade old) that have been carried out in the past including with size selection discrimination and at multiple wavelengths. For example, the work done by Lack et al., Cross et al., Schnaiter et al., Arnott et al., Sharma et al., Smith et al., just to mention a few groups (but more are out there). Therefore, statements like those in lines 304-306 are certainly incorrect. A web search will return a few studies done in the past using photoacoustic even with size selection. In addition, the paper (and especially section 3.3) fails to recognize other types of measurements performed in this field, for example, with the extinction minus scattering approach, or using photothermal interferometry, which even if with their own challenges, are an important contribution to the field as well.

-    Finally, I have some concerns with the “correction” scheme developed and presented in the paper, as discussed in the specific comments.

Specific Comments

-    Section 2 should report more details on the specification of the lasers, in particular, the power and the operation mode (CW or pulsed)

-    Also, section 2 should clarify if the cylinder where the sample is analyzed is acoustically resonant or not.

-    What material is the cylinder made of?

-    Provide specifications about the windows as well as the lens. 

-    What’s the rationale for the specific choice of multiplexing frequencies, where they optimized somehow?

-    Some more detail on the NO2 calibration procedure would be useful. For example, was photodissociation being accounted for? 

-    Deposition losses were calculated assuming the particles are spherical. That’s reasonable, but whether that is the case or not it should be made clear.

-    Is the LabView code for the DMA going to be made openly available?

-    It would have been useful to check the size selected distribution with an additional sizer downstream of the DMA system to verify that indeed the effect of multiple charges was minimized.

-    A sensitivity of the Mie calculations to the assumed index of refraction would have also been valuable. Could it be that the particle-phase index of refraction differs enough in the experiments carried out here from that found in the literature to explain at least part of the discrepancies?

-    I might have missed some key details regarding the development of the correction scheme, but it seems to me that the correction was developed from the divergence between measured and Mie estimates, and then used again to show that after the correction the size-resolved absorption matched the Mie calculations. If that was indeed the procedure, the better agreement in figures 8 and 9 is obviously not surprising, but also of relatively low value without further independent validation that the correction is indeed “correct”. For example, if the issue is an incorrect index of refraction for particle phase nigrosine, then the “correction” would actually make the measurements less correct with respect to the real absorption value. 

-    Line 322: “However, this discrepancy was ultimately resolved.” As mentioned earlier, I do not feel like an empirical adjustment of the data to the theory can be considered as a resolution of the discrepancy between the same measurements and the same theoretical values. Further blind validation of the correction scheme would need to be carried out to test if indeed the correction is effective or not.

Technical Comments

-    Lines 38-40: Not all particles emitted by combustion are black carbon and therefore not all are agglomerates.

-    Line 44: These increases are indeed predicted, if anything, experimental studies have shown enhancement lower than predicted by Mie calculations (several published papers are available on this topic). The enhancement is defined as the absorption cross-section of the combination of the absorbing particle and coating to that of the absorbing particle alone.

-    Line 52: The main key factor is probably the heterogeneity of coating thickness on different particles, which is related to the size but not uniquely.

-    Line 62: This assumes there is no phase change (evaporation of coating) during the exposure to the laser.

-    Line 72: remove “a have” 

-    Line 111: “air flow controls were replaced with aftermarket components” what was the reason to do so?