The manuscript (MS) describes the production and characterization of aerosol nanoparticles (NP) as single material form and as hetero-aggregates (HA) using a spark discharge generator (SDG) combined with a nebulizer. The HA consist of TiO2-NP, on which four different metal NPs are deposited. The pure metallic particles and the HA are characterized online and offline using a variety of methods such as SMPS and TEM-EDX.

The MS contains many valuable and original results, which, however, still need to be bundled and formulated much better. In particular, there is no common thread linking the research question addressed, its answer and further consequences of the results. What are, up to the conclusion, new findings the authors want to report about the coating of larger NPs with metal NPs from spark ablation? The authors are encouraged to thoroughly rewrite the manuscript with a clear story line and targeted motivation. One interesting part to focus on could be the metal-depending layer quality (layer porosity, thickness, contact angles etc.) of spark generated NPs on TiO2 surfaces. In this context, the oxide affinity of certain metals, e.g. Au vs. Cu or Ni, could be systematically investigated, including many of the already presented data in a condensed way.

The individual points are explained in more detail below, with the page (P) and line (L) indicated.

1)

Introduction

The spark ablation section is well described; however, a detailed motivation for coating TiO2 NPs with spark generated particles is missing. Topical literature regarding this aspect is missing, too. While the introduction covers many interesting topics a clear focus of the presented work at the end of the introduction is missing. i.e. which open research questions should be addressed here. This goes beyond the listing of the performed investigations and covers rather the coherence of the results presented here.

2)

P 4, L 104-108: Collecting small particles by diffusional deposition leads to a size-biased representation of the particle, which is specially agravated for small nanoparticles as outlined in the MS. However, also the amount of deposited particles differs substantially from the number of particles in the aerosol. Therefore, the question needs to be discussed how the fraction of nanoparticles sampled on the TEM grid (and on the Teflon filters) in relation to the total aerosol particles is determined. It is surprising that the number of deposited particles and aerosol particles are so close to each other (e.g. Fig. 4). Which approach was taken to quantify the fraction of deposited particles? Please explain this point in more detail.

3)

How does a quantitative NP sampling work on Teflon filters with 2 µm pore size and NPs being smaller than approx. 50 nm?

4)

P 5, L 126: What is the motivation for depositing agglomerate NPs on a solid TiO2 film and the subsequent sputter coating with a 100 nm Au layer? It is not clear why such a thick platinum layer needs to be applied. Does this not lead to a significant loss on resolution?

5)

P 6, L 151ff: The idea of employing the circularity to define the spherical particles is very good. This approach could make an important contribution to the issue about the first nucleating clusters and their further fate. This approach is certainly a highlight of the MS and would deserve some more evaluation. Why was the limit of 90% also applied for sphericities staring well below 1.0 (cf. Fig. A1, e.g. Pt for 1.3 s residence time)? Is the fact that some particles start already with sphericities well below 1.0 an indication that in fact they consist of much smaller units, i.e. atomic clusters? In Fig. 3 for Pt 3 residence times were analyzed with respect to size distribution. Why was the results for 26 s not also included in Fig. A1 in the Pt column?

6)

The authors state that the coagulation time influenced the primary particle size of the respective materials. How is that possible when primary particle formation is completed approx. 100 µs after the spark discharge? And how can the sphericity of the Pt increase with longer residence time (from 1.3 s to 2.2 s) in Fig. A1?

7)

P 7, L 155 and Table I: The authors refer to “primary particle size” and “agglomerate particle size”. Please include a detailed description which size (primary vs. agglomerate) is shown in the size distributions. What equivalent diameter is presented here?

8)

P 7, L 169-171: The authors state that Au NPs might form oxides as a consequence of using N2 5.0 with oxygen impurities. This argument holds certainly for Cu and Ni. Pt NPs from spark ablation can exhibit thin oxide layers on the particle surface; however, for Au, an oxidation is impossible under the mentioned experimental circumstances. This fact supports, in turn, the relatively large primary particle size of <6 nm for Au NPs with a circularity of C=1 that is mentioned by the authors.

9)

Besides the low melting point, especially the absence of oxidation of the surface of Au NP contributes to the strong necking and coalescence growth of Au primary particles/clusters. On the other hand, oxygen-sensitive materials such as Ni and Cu experience a so called "pinning effect" by adsorption/reaction of the surface with oxygen (e.g. Seipenbusch et al. J. Aerosol Sci. 34, 2003). Such a pinning effect can increase the activation energy for coalescence from typically 50 kJ/mol (clean metal surface) to about 80 kJ/mol (partly oxidized metal surface, e.g. Ni). This point should be more elaborated here.

10)

P 8, L 193-195: Regarding the generally poor charging of very small NP it is surprising that losses due to electric fields should accumulate to such an important amount. Are the substantial losses in the SDG chamber not rather related to the high particle diffusivity?

11)

P 8, L 206-208: The authors mention the ejection of micron sized particles during spark erosion and a subsequent deposition of those within the housing of the spark discharge generator. The observation of Tabrizi et al. is very interesting and is also found for other processes where metal surfaces are locally heated up by sparks or pulsed lasers. It is also reasonable to attribute a high mass loss to the deposition of large micron sized particles. However, since the significance of this mass loss channel is expected to depend very much on the material properties, a more direct confirmation of these droplet-based particles would improve the convincing power of the argument. For instance, samples from the SDG chamber could be taking a wipe sample and analyzing it with electron microscopy (SEM/EDX or TEM/EDX) for micronic spherical particles.

12)

P 12, L 281: The authors derived contact angles for Au and Pt on TiO2. Please show a TEM image with a magnification where the contact angle is outlined. Please increase also the size of the TEM micrographs shown in Fig. 5. The inserts are very hard to see.

13)

P 13/14/18, Fig.6, 8, B1: The coloring of the EDX maps is confusing. Please use a defined color for each element to distinguish, e.g., background from sputter layers from NPs from TiO2 support. Fig.6: The authors present dark field (DF) and bright field (BF) TEM(STEM) images. Please refrain from using BF images for Pt@TiO2 since Pt has a high contrast in HAADF, such as Au in Fig.6 top left.

14)

P 15, L 328: The authors mention a continuous layer of Au on the TiO2 substrate after deposition. This observation can be related to the oxygen affinity of the metals used. Due to the absence of oxygen in Au NPs, partial sintering and necking can be observed even at room temperature.

15)

P 19, L 370-371: The calculation of the corresponding aerosol concentration based on the TEM micrograph is still not clear (cf. comment above). How was this done?

16)

P 19, Caption of Fig. C1: The diffusion losses in laminar flow in tubes does not depend on the tube diameter! Therefore, this argument about the influence of the larger tube does not apply. There must be another reason for the similar losses at different residence times since the amount of loss depends critically on the duration.

Minor comments:

• Check whole manuscript for typos
• Please be consistent with units, e.g., P 2, L 39 “20.000 Kelvin”; please write 20.000 K
• Please check all literature references for typos, e.g., P 2, L 47, P 12,L 273
• P 5, L 119: In front of "... m2 of the grid" seems something to be missing. Please correct it.
• P 12, L 273: The reference “(62)” is completely out of the normal referencing system, which relies on “name and year”. Please correct this.
• P 14, L 313: The sentence “This is a lower estimate of the coating” is not clear. Please be more specific.

Overall the work is written in a vague manner without direct means for relevant confirmations. This work also made the impression for reproducing already published works published years ago but not yet fully cited them. This “gave the chance” of this work to pretend to be the initial discoveries. The morphologies of the NPs demonstrate that this study did not yet optimize the system. Only switching on the generator and collecting them for subsequent analysis cannot add scientific values, particularly ignoring the large volume of literature, which reported almost the same line of argumentations. The authors also try to use artificially texted letters to mark the different metals without showing any experimental proofs. Raising such concern is mainly due to all their similar appearances but undistinguishable nature from the TEM results. 

Large body of relevant literature was severely missing.

Primary particle was defined scientifically incorrect.

where the authors showed the 1-nm particles?

Melting point of the 4 metals cannot be directly used to evaluate their different particle sizes, as the impure surface of the particles also makes dramatic influences.  

How the authors prove that the oxidation took place during NP production?

Why the mass was evacuated using micro molar? As the authors also claimed that the SDG can also deliver NPs of high quantity. How the amount of micro molar supports the aforementioned claim?

For the different production rates of the metals, previous study already clearly explained why. Please cite the relevant literature, not only saving the unnecessarily repeated work but also not making an impression of confined literature review.

The authors also argued that the mismatch of size distributions measured between the TEM analysis and SMPS data is mainly due to the low counting efficiency from the latter. How to explain the overestimation of diffusional deposition of smaller NPs for TEM analysis?

A very puzzled presentation for Fig. 5 onwards is to remove the Cu NPs.  Could the authors provide any scientific reasons for such inconsistencies?

How the edges of the NPs were determined when using eq. 1 to calculate circularity?

How the authors separate the ones coated over TiO2 with those of self-coagulation?

Other mistakes also showed the carelessness and this can also give the reviewer for scientific suspect. Wrong spelling in page1, “Department od Environmental Sciences” should be “Department of Environmental Sciences”.

The paper discusses the deposition by coagulation of small particles generated by spark discharge onto a TiO2 nanopowder to make an unspecified catalyst. Spark ablation is known to generate small particles, but a reliable method to synthesize useful catalysts with them is still lacking. The paper goes in more technical detail than a recent publication with a similar approach (Debecker 2024), and provides several useful insights, and novel approaches compared to the state of the art. The paper does not mention catalytic performance, but focuses on the aerosol synthesis route, and as such is a suitable topic for this journal

The introduction seems written as an afterthought: It contains a lot of puzzle pieces, but the reader is left to guess what the motivations are for including these pieces. A clearer focus here helps highlight which parts of the work are novel, and which ones are not. 

Technical issues:

Residence time for the different experimental parts aren’t completely clear. E.g. Fig 1; L100: The residence time of volume C is specified. But what is the residence time before and after volume C? Figure 2 shows particle size 1.3s “after generation”: depending on where you define generation, this can be somewhere before, in, or after volume C…

The description of Fig 1. implies coated particles are not analysed (either/or?). Is the deposition tree for coating experiments the same as direct particle deposition without volume C? Were flows/volume adjusted to maintain correct residence time?

Is the SDG flow on during nebulizer characterization, and is the nebulizer flow on during SDG only characterization?

L88: Is the flow rate specified otherwise somewhere? It seems redundant.

Section 3.1.1: It’s mentioned later, but already here mention that collection by diffusion will overrepresent small particles. Validity of max. primary particle size and Df of small agglomerates is probably unaffected, but the consideration should be part of the authors analysis. Differences between the metals are consistent with prior literature on spark discharge, not necessarily new.

L151 “Fully coalesced particles are defined here as primary particles.” Is this the same as singlet particles (e.g. Feng 2016) ? The circularity criterion for singlet particle identification is a nice addition to previous analyses, and could be more explicit.

Section 3.1.2 (numbering is wrong in manuscript)

The comparison of electrode mass loss and NP mass on filter is very qualitative. It shouldn't be too difficult to get a first order approximation of expected relative mass rates based on the Llewellyn Jones formula (see Tabrizi 2009, Feng 2016), and confirm whether or not the suggested explanations fit. 

L193: Charge related losses are significant, but not necessarily majority of losses. Losses due to turbulence / poor flow conditions typically are also significant. Collection efficiency on membrane filters is loading dependent, which causes an underestimation of mass arriving at the filter. This effect is most important for low loadings.

L228-235, Fig 4. Are the concentrations for the TEM samples calculated by the diffusion correction, or only the relative abundance?

L281: please provide clear TEM images in SI for the contact angle measurements.

Fig 6: Shouldn’t this be compared to the observed TEM size distribution in figure 3? The reason why the TEM grids collect more of the smaller particles is the same reason why the TiO2 collect more of the smaller particles.

Fig 8. EDX is difficult to read, in particular the Au sputtering layer mentioned in L333-335. This type of information must be clearly visible in the graph, e.g. using labels in the graph.

Minor details:

L84: The set voltage for the spark generator used is the mean voltage, not the breakdown voltage.  

There is no forward reference to appendix C. If it’s not relevant to the work itself, best to include this in the work referred to in L380.

References:

Debecker, Damien P., Plaifa Hongmanorom, Tobias V. Pfeiffer, Bernardus Zijlstra, Yingrui Zhao, Sandra Casale, and Capucine Sassoye. ‘Spark Ablation: A Dry, Physical, and Continuous Method to Prepare Powdery Metal Nanoparticle-Based Catalysts’. Chemical Communications 60, no. 79 (2024): 11076–79. https://doi.org/10.1039/D4CC03469D.

Feng, Jicheng, Luyi Huang, Linus Ludvigsson, Maria E. Messing, Anne Maisser, George Biskos, and Andreas Schmidt-Ott. ‘General Approach to the Evolution of Singlet Nanoparticles from a Rapidly Quenched Point Source’. The Journal of Physical Chemistry C 120, no. 1 (2016): 621–30. https://doi.org/10.1021/acs.jpcc.5b06503.