General

The manuscript describes the development of a low-cost, 3-wavelength photoacoustic spectrometer (SiMPLE-PAS). The emphasis is on engineering design choices (mechanical, electrical, and software), with some laboratory validation and a limited field deployment. While the realization of the device is technically competent, I do not find significant novelty from a scientific instrumentation perspective: the underlying working principle is that of a standard photoacoustic instrument, and the use of 3D-printed parts and consumer electronics is incremental rather than conceptually new. The authors also do not clearly articulate the specific need or scientific problem that this instrument addresses beyond low cost.

That said, the work could still be of interest to Aerosol Research if framed as a reproducible, open-source, educational, or accessibility-focused contribution. To reach that point, the manuscript requires major revision, both in structure and in content.

Manuscript length and structure

At ~12,000 words (not accounting figures and tables), the manuscript is 4,000–6,000 words too long. Sections should be significantly shortened or moved to the supplement. Moreover, the current organization is confusing: for example, calibration methods are embedded in the Results section rather than Methods. The paper should follow standard structure before detailed discussion is considered.

Reproducibility and open-source availability

If the intention is to provide a community-sensor-type instrument, all essential resources (software, CAD files, PCB designs) must be openly available in a long-term, independent repository. “Available by request” is not sufficient. An assembly guide with photographs would further enhance reproducibility and impact.

Calibration and evaluation

The methods used to calibrate and evaluate the device are not sufficiently thorough or clearly explained. In particular, the lack of a conventional field evaluation with side-by-side reference instruments is unfortunate, as this is typically the best way to obtain a general understanding of the instrument response characteristics (e.g. susceptibility to varying relative humidity and temperature, long-term drift, influence of aerosol composition etc). I encourage the authors to seriously consider whether such an evaluation could be arranged. The comparison of Ångström exponents with denuded and non-denuded samples is not without interest, but the measurement arrangement introduces multiple sources of uncertainty that make explicit conclusions about the device performance difficult to draw. Below are some specific concerns related to the O₃ evaluation:

• It is unclear why the calibration is conducted first with NO₂ and then again with O₃ using the corona discharge and prior CRDS calibration. The manuscript states that O₃ is used for accuracy determination, but no quantitative metric of accuracy is provided. In addition, why should the PAS and CRDS produce a 10× difference? Why is the slope of the linear fit forced through zero? Reporting R² = 0.9999 for a fit with three data points spanning 0–5000 Mm⁻¹ is potentially misleading. Values should be reported with realistic precision (not five significant decimals).
• The current presentation largely omits the 0–10 Mm⁻¹ range, which is highly relevant for outdoor measurements. Figures 4 and 5 would be more informative if e.g. replotted with log-scaled axes to highlight this range.

Recommendation

I recommend major revision. The manuscript is not suitable for publication in its current form, but with substantial shortening, restructuring, open-source dissemination of design files, clearer justification of calibration, and a more thorough evaluation against reference instruments, it could become publishable.

The manuscript describes a low-cost photoacoustic instrument, capable of measuring aerosol absorption at three visible wavelengths. The authors give a highly detailed description of the design of the instrument, its calibration against gas phase absorption, and evaluation of its detection limit. A comparison test with aerosol sample against an existing photoacoustic instrument is also presented. The performance is comparable to many previously reported PAS instruments.

However, the manuscript reads a lot more like a technical tutorial rather than a research article, and the scientific novelty is quite limited. The design seems fairly conventional, outside of demonstrating that good sensitivity can be achieved with the low-cost options and technical effort has gone to designing a custom amplifier for the microphone and ensuring that the 3D-printed parts are suitable for handling gas samples and acoustic noise.

On the other hand, the topic is quite interesting and relevant, particularly because, like the authors justify in the introduction, the option for commercial PAS instruments is currently somewhat limited. Offering a detailed instructions for a low-cost starting point using readily available parts has potential for impact. So, while I would like to support the publication of the manuscript, I question if Aerosol Research is the correct publication in this case, due to limited scientific novelty of the manuscript and since AR offers no options for something like technical notes or tutorial type articles. I agree with comments of RC1 (https://doi.org/10.5194/ar-2025-31-RC1) that there is likely a way to reformat the manuscript towards a more conventional research article, but I feel like substantially shortening it could be detrimental, since main impact seems to come from the potential reproducibility of the instrument. If the authors decide to move to this direction, I suggest preserving the details at least in a supplement.

I have, in any case, added some specific comments below that I think should be considered by the authors:

• Comparing the sensitivity, for example to the MultiPAS-IV, the resonator Q and laser power seem to be on similar levels, but MultiPAS-IV uses multipass configuration, which gives an power enhancement of 30x or more, according to Fischer and Smith (2018a). Are you able to comment on what factors are contributing to the fact that the SiMPLE-PAS reaches close to same level of a detection limit with a single pass instrument? Is this mostly due to a more sensitive microphone and amplifier, or are there other contributing factors that might explain this difference?
• The authors state that SiMPLE-PAS is the lowest cost PAS instrument to date, which, while likely to be true, is difficult to justify properly, as the costs are rarely addressed in scientific publications, and can be out of date relatively quickly, and comparisons of only material costs to commercial instruments is typically unfair.
• One concern I would have when using a low-cost laser module without temperature stabilization and having left out the laser power monitoring, is the long-term drifting. That is, are they large enough to significantly affect the calibration over time. For example, fig. 6 shows that the drifts start to overcome the noise after around 30 min mark, but it is unclear how large the drifts are over the whole 3-day measurement, and if the relative change in the signal level can be considered negligible after the background subtraction.
• Fig. 4 caption: considering mentioning that the points in the figure are 2 min averages. This would clarify the point made in the last sentence of the caption.
• Line 389: I assume this should say “decreased by 100 Hz”?
• Fig. 5: The slope of the blue line in the figure does not seem to be 1.21. At b_crd=100 Mm^-1, it looks to be something around 600 Mm^-1. For clarity, I would also consider plotting, instead, the measured absorption as a function of the absorption calculated from the CRDS and the wavelength dependence, like was done in figure 4.
• Fig. 7: what is the fit here? Or is it the function calculated using eq. (3)-(5)?
• line 506: Should this say “given that AAE does not decrease fully to 1.0.”? This AAE treatment is currently a little unclear to me. It seems that the manuscript only reports AAE_BrC, but nowhere the effective AAE of the total sample. So, in the limiting case that there was only BC left after denuding, would AAE_BrC not become undefined (i.e. the residuals in eq. 4 would be zero)? In my opinion the current treatment leaves unclear, how far from BC the sample actually is after denuding.
• Related to the AAE treatment, I think there may be an error in eq. 5: are you not double counting b_BC,515 here? For example, if you were to calculate b_abs,515 using this, you would end up with b_abs,515+b_BC,515.