A general patterning approach by manipulating the evolution of two-dimensional liquid foams

The evolution of gas-liquid foams has been an attractive topic for more than half a century. However, it remains a challenge to manipulate the evolution of foams, which restricts the development of porous materials with excellent mechanical, thermal, catalytic, electrical or acoustic properties. Here we report a strategy to manipulate the evolution of two-dimensional (2D) liquid foams with a micropatterned surface. We demonstrate that 2D liquid foams can evolve beyond Ostwald ripening (large bubbles always consuming smaller ones). By varying the arrangement of pillars on the surface, we have prepared various patterns of foams in which the size, shape and position of the bubbles can be precisely controlled. Furthermore, these patterned bubbles can serve as a template for the assembly of functional materials, such as nanoparticles and conductive polymers, into desired 2D networks with nanoscale resolution. This methodology provides new insights in controlling curvature-driven evolution and opens a general route for the assembly of functional materials.

9. The universal applicability of the technique is unclear since it requires the fabrication of patterned micro posts first. Further, there are numerous other techniques like 3D printing, solution patterning etc. What advantages are provided by this technique compared to other techniques? In this work, Ag NPs were used -and the bubbles were formed by the catalytic activity of Ag. What if we want to deposit other materials -like aluminum, polymers etc. How can bubbles be produced in such systems, and how can such materials be deposited?
10. Line 161-162: The authors mention "So in most cases, it outperforms the conventional ITO transparent thin films". First, the authors have given only one example of comparison and not "most cases". Secondly, they are comparing the performance of their samples with ITO deposited films -which is a different material and hence not a fair comparison. It would be more useful to compare the results with similarly developed patterns through other techniques (such as 3d printing, electrohydrodynamic printing etc.).
11. Towards the end, the authors have mentioned possible applications of their technique can be extended to other systems like water harvesting, anti-icing etc. The implications of Kelvin equation are well-known, but it is unclear how authors work can be used in the application areas that the authors have mentioned. Such type of sweeping statement without significant proof should be avoided. Fig.1 caption: First it is written that ethanol solution of urea peroxide was chosen and a few lines later it is written H2O2 is catalyzed by AgNP. There seems to be an inconsistency there.

The Methods and Supplementary section,
13. Some references appear on Page 6 and some on Page 8. It is unclear why they are separate.
14. There are numerous typographical and grammatical errors in the text, example a. Caption of Figure 1, Line 324: Height is misspelt as hight.
b. Page 18 of supplementary section: As is the area of the gas/liquid interface which has been typed as as at places.
c. Page 18 of supplementary section: From is misspelt as form.
g. Caption of Figure 1, Line 331: "The central bubble consumed surrounding larger bubble." Incorrect tense usage.
i. The entire manuscript refers to the pillars endowing surrounding bubbles. Revisiting this sentence construct is strongly recommended.
k. Line 165: "These results can stimulate research in other fields." l. Line 111: "With evolution, bubbles in dodecagonal cells will has a larger surface curvature radius, which causes an unbalanced pressure between dodecagonal cells and square or hexagonal cells." The word will needs to be deleted.
m. Supplementary section: • "Ethanol solution of urea peroxide was choose because it was easier to spread on the patterned substrate and uniformly distribute than that of hydrogen peroxide water solution." The tense usage in the italicized words above is incorrect.
• Page 18: "..surrounding bubbles is difficult because they varies with time and position". Incorrect tense usage in the italicized word.
It is obvious that the authors have done a significant amount of work to address my concerns. I appreciate the efforts. My only remaining concern is that the application part is still relatively week. It would be helpful if the authors can actually demonstrate some real "device applications" with those additional examples of templated assembly.
Reviewer #2 (Remarks to the Author): The authors have addressed some concerns with this revision, but two problems still remain.
1. There are still many errors in grammar and sentence structure that detract from readability. Here are some of them with line numbers.
36: "other researches for modeling" should be "other research modeling" or "other studies modeling" (avoid using "researches" to refer to multiple studies, see also 264) 49, 174: "when total gas is not enough and distribute uniformly" is unclear and sounds awkward.
(Replace with something like "when the volume fraction of gas is large enough and the gas is distributed uniformly") 74: When bubbles >were< produced in the confined space, 100: "although owning a large size" is awkward 160: To explain that, we calculated quantificationally the radius of curvature -quantificationally is not a word and no adverb is needed there 176: rest bubbles? 199: This is not a quantitative explanation. 252: adaption is not a word (you mean adoption) and the section title is confusing 337: Use a period or semicolon before "for example," because a complete sentence follows it. This is just a sampling, and the authors should carefully go over the text again to improve the writing. Articles (like "the", "a", "an") are often missing and in some places plural forms should be used where singular forms are and vice versa. Examples: 99: Then >the< boundary of the bubble 51, 55: "understandings" should almost always be singular 2. The review article you reference in your rebuttal, "Inter-bubble gas diffusion in liquid foam," states in the introduction: Coarsening (also interchangeably known as Ostwald ripening, inter-bubble gas diffusion or disproportionation) of gas-liquid foam is the process by which big bubbles consume adjacent smaller bubbles due to differences in pressure in the two bubbles caused by the Young-Laplace effect.
There is no need for the new term "jungle law" for a phenomenon that already has multiple names. We believe that sounding impressive is not a good reason to introduce a new term which would add more confusion than insight. Ultimately, whether or not to include the term "jungle law" in the title is up to the editor, but we strongly advise against it (both in the title and the text).
Reviewer #3 (Remarks to the Author): The author has provided adequate answers to the reviewers questions put forward earlier. Although the reviewer still has reservations on the broader aspects of the study, particularly in context with use of this method to fabricate patterns on surfaces, the technique proposed by the authors has novelty factor that the readers of this journal may find very useful.
One of the first concerns by the reviewer (and by the other reviewers as well) was the use of the term "Jungle Law" by the authors. The authors have done their best to convince that the phrase adequately fits in context of foams as well. The reviewer however still disagrees with the usage of this term. The reviewer agrees that "coarsening" is more widely used term in context of foams, and sees no reason why it could not be used in this work as well. The reviewer disagrees with the notion that usage of "jungle law" phrase will make the work more appealing to the readers, and it should be used because it is "more impressive". Certain degree of flexibility is inherent in science, but it should not be done at the cost of glamorizing phenomenon un-necessarily. Coarsening is a perfectly good word that is used by the scientific community across disciplines in context of foams, colloids, lithography etc. The authors are welcome to frame the term "jungle law" in this work and correlate it with "coarsening" more explicitly.
The authors have made welcome changes in the main text and the supplementary information. However, there are still numerous grammatical and syntax mistakes in the text. The new sentences that have been added by the authors appear misfits sometimes with the rest of the story. The reviewer believes that the manuscript will be appropriate for publication in this journal after the concerns mentioned if the concerns mentioned above are addressed properly by the authors.

1
Response to the comments of the reviewers.
Reviewer 1 1. Comments: This paper describes the evolution of bubbles within a liquid foam that is confined between a micropillar substrate and a glass plate. While the observations are interesting and potentially useful, I find the paper needs some major revisions to make it to more appeal to the general audience of Nature Communications.
Reply: The authors thank the reviewer for the positive evaluation of our paper. We have revised the manuscript carefully to present our work more clearly.

Comments:
The paper is not well written. There are numerous grammar errors; the language is difficult to understand. The explanations are not clear. For example, the authors used both "curvatures" and "curvature radius" frequently and sometimes in adjacent sentences, making the paper hard to read/understand. I suggest authors use consistent terms. Reply: Thanks for the reviewer's comments. We have revised the manuscript to make it easier to read and understand. The grammar errors have been corrected carefully. "Curvatures" and "curvature radius" have been replaced by the consistent term, namely, "radius of curvature".

Comments:
It is not clear how the "reverse jungle law" and "defect forming" co-existed. For example, what determines "a"? What does the instantaneous mean radius (Rho) actually mean physically and how/why does it matter? Reply: Thanks for the reviewer's comments. We have added a figure (Fig. 2) to clearly show the situation in which "reverse jungle law" and "defect forming" happen. We will reply the above questions as follows. 1) How did the "reverse jungle law" and "defect forming" coexist? When bubbles evolve, large bubbles consume small bubbles due to interbubble gas diffusion, but the micropillars arranged in hexagonal shapes limit the maximum radius of gas bubbles. Once the bubble surrounded by micropillars reaches its maximum circular size, any additional gas flow leads to the boundary of bubble deforming into several meniscuses which have the low radius of curvature and therefore, the high Laplace pressure. So surrounding smaller bubbles can absorb gas from these large deformed bubbles until having equal radius of curvature. This is the "reverse jungle law", which prevents the large bubbles infinitely growing and assists to form arrays of equal-sized bubbles.
However, the "reverse jungle law" happens only when the pillar interval, a, is not too large. The pillar interval limit the minimum radius of curvature (namely, a/2) for the deformed large bubbles. Once the deformed large bubble reaches its minimum radius of curvature, the surrounding smaller bubbles with radius above a/2 will grow, showing the "reverse jungle law" (Fig. 2a), while bubbles with radius below a/2 will shrink (Fig.  2b), forming a defect. So when a is so large (120 μm, for example) that a/2 is larger than the radius of almost all surrounding smaller bubbles, only the defect forms; When a is so small (35 μm) that a/2 is smaller than the radius of almost all the surrounding smaller bubbles, only the "reverse jungle law" happens; When a is medium (60 μm), some surrounding smaller bubbles with radius larger than a/2 show the "reverse jungle law" and some with radius smaller than a/2 form the defects, so the "reverse jungle law" and "defect forming" coexist. 2) What determines "a"?
The "a" is one of our grammar errors, and we have replaced it with "the". 3) What does the instantaneous mean radius (Rho) actually mean physically and how/why does it matter?
The instantaneous mean radius, ρ, was proposed in Robert Lemlich's theory (Please see: Ind. Eng. Chem. Fundam. 1978, 17, 89-93) for how the foam of arbitrary bubble size distribution might evolve. For a foam with bubbles of different size, the large bubbles grow and the small bubbles shrink until vanishing. In the system, a critical radius was denoted. If radii of bubbles above it the bubbles will grow, and below it they will shrink. The instantaneous mean radius is identical to the critical radius, physically.
In Lemlich's theory, bubbles were viewed as 3D spheres, and ρ was derived as Where n is the number of bubbles in the foams, r is the radius of the bubbles. Equation (1) suggests, the instantaneous mean radius was related to all the bubbles in the foam, and it usually increases with time because of the evolution of bubbles. The instantaneous mean radius matters a lot for predicting the bubble size distribution at any time, as Lemlich deduced, Where K can be viewed as a constant, t is time. The equation suggests, at any time bubbles with radius larger than ρ will grow and that with radius smaller than ρ will shrink. Besides, the equation can be converted to a finite difference equation that can be easily solved. Therefore it can predict the bubble size distribution of the foams at any time. In our work, Lemlich's theory was applied into evolution of 2D foams where bubbles were considered as cylinders instead of spheres. Therefore ρ and evolution equation can be derived as Where L, r is the boundary perimeter and radius of curvature of the bubble in 2D, respectively. These two equations are governing equations for the evolution of 2D foams in our simulation. The instantaneous mean radius also guides us to design the pillar interval for preventing the defect forming. As talked in comment 2, the pillar interval limits the minimum radius of curvature (a/2) for the deformed large bubbles. If r = a/2 < ρ, according to equation (4), the deformed larger bubbles can be adsorbed and show the "reverse jungle law". Since ρ gradually increase during the evolution, the minimum ρ is at the initial state of the evolution (denoted as ρ0). So keeping that a/2 < ρ0, the defect can be effectively forbidden. 4. Comments: What is the mechanical stability of the nanoparticle assembly? Is this approach generally applicable? To make this practically useful (such as ITO glass as authors tried to compare with), the authors need to demonstrate the robustness of these patterns, for example, against abrasion, washing etc. Reply: Thanks for the reviewer's comments. It has proven that the mechanical stability of Ag nanoparticle assembly or Ag nanowire networks will be improved significantly after heat treatment for sintering (Please see: J. Mater. Chem. 2011, 21(39): 15378-15382; Nanoscale, 2014, 6(9): 4812-4818.). Similarly, we carried out the immersion test and tape test for the samples after sintering generated from our experiments. The results are shown in Supplementary Fig. 16 as follows.

Supplementary Figure 16 │ The immersion test (a) and tape test (b) of AgNP
patterns after sintering at 200℃ for 1 h. (a) Three samples were immersed in 10% water solution of NaOH at 60℃, 6% water solution of HCl at 25℃ and solvent composed of water, ethanol and acetone at the rate of 1:1:1 at 25℃, respectively. The conductivity was measured in every 30 minutes. The variation of conductivity is within 5% in the immersion test, showing good resistance to washing. (b) The tape test was carried out. The change in conductivity is within 10% after five tape tests, suggesting good performance against abrasion.
The excellent conductivity and high transparency as well as the robustness of the AgNP assembly enable it to be a promising candidate to replace the expensive ITO. Besides, as an efficient, clean and sustainable strategy to assembly nanoparticles with nano-resolution, it can be tailored for fabricating electronic devices with high precision.

Comments
: I am not sure if it is completely correct to call the liquid foams or bubbles two-dimensional. Obviously they are all 3D objects. What did authors mean by 2D? Actually it might be helpful if the authors can take a look at these foams and bubbles in 3D? Reply: Thanks for the reviewer's comments. Actually, the monolayer bubbles constrained between two plates are usually called 2D or quasi-2D in most of the papers within the foam community. The 3D structure of the foams constrained between two plates has been studied and shown in the figure below. It suggests, the bubbles of different size in the foams have the same height which equals the spacing between the two plates. So the volume and interfacial energy for the bubble i, can be deduced as Where Vi, Ei, is volume and interfacial energy of bubble i respectively, h is the spacing between two plates, and Ai, Li are area and boundary perimeter in 2D from top view respectively. So the area and boundary perimeter of the bubbles can be representative of the volume and interfacial energy. This is why researchers usually called the monolayer bubbles between two plates 2D or quasi-2D. In our work, the liquid foams are viewed as 2D foams similar with discussion above, with details in Supplementary Note 2. It simplifies greatly the simulation by applying Lemlich's theory into 2D foams.
Left: Oblique view of a foam between two plates. Right: A ferrofluid foam with slightly higher liquid content, illuminated from below.
Ref: page 57 in Foams: Structure and Dynamics. (Oxford University Press, 2013) 6. Comments: Again, I appreciate the authors were able to demonstrate a number of different patterns and demonstrate a proof-of-concept application; I just feel the story at its current format may not match the quality and impact of Nature Communications papers. More clear and convincing explanation and demonstration of more broad applications may be helpful. Reply: Thanks the reviewer for the positive evaluation of our paper. We have revised the manuscript following Nature Communications guidelines. We also have added more clear explanation in the revised manuscript. In this paper, we have demonstrated that the evolution of wet 2D foams can be well manipulated, instead of obeying the conventional jungle law. The manipulation enabled us to prepare various 2D bubble patterns. With these patterned bubbles as a template, AgNPs were assembled into desired 2D networks.
To broaden the applications of this strategy, we have managed to control evolution of 2D foams which were produced by other methods, such as hydrolysis of sodium borohydride catalyzed by hydrogen ion. In addition, with these patterned bubbles as a template, The conductive polymer, PEDOT (poly(3,4-ethylendioxythiophene)) and polystyrene (PS) microspheres were assembled into hexagonal networks. As shown in Supplementary Fig. 17 and Supplementary Fig. 18, the evolution of 2D foams generated from hydrolysis of NaBH4 and the assembly of PEDOT/PSS or PS microspheres are the same with the evolution of 2D foams obtained from decomposition of urea peroxide and assembly of AgNPs, respectively. In this work, materials from polymer (PEDOT) to nanoparticles with size of 450 nm (PS microspheres) were assembled, therefore we believe this is a general strategy for assembling functional materials into desired 2D networks for device applications.  The experimental and modeling approaches appear sound. However, the manuscript construction does not follow Nature Communications guidelines (e.g. the use of a bold first paragraph in lieu of an abstract, overly terse main text, and relying too heavily on supplemental material). In addition, attention needs to be given to the English grammar and syntax as there are numerous errors that detract from readability. If properly rewritten, the results of this manuscript seem appropriate for publication in a broad-impact journal such as Nature Communications.

Supplementary
Reply: The authors thank the reviewer very much for the positive evaluation of our work. We have revised the manuscript following the guidelines of Nature Communications. The errors in English grammar and syntax have been carefully corrected. We also moved some figures from supplemental material to Fig. 2 in the main text for less dependence on the supplemental material. Some detailed and clear explanation has been added in the main text for presenting our work clearly.

Comments:
The authors describe the foam evolution in terms of a "jungle law". This choice of words is strange as such evolution is typically referred to as coarsening or Ostwald ripening. Are the authors familiar with other papers within the foam community that refer to this phenomenon as "jungle law", or are there aspects of the phenomenon requiring a new term to be coined? Reply: Thanks for the review's comments. We use the words "jungle law" to describe the evolution within various systems that is dominated by surface tension and curvature. In these systems, curvature determines some physical quantities: curvature determines Laplace pressure for gas-liquid foams; curvature determines saturated vapor pressure or solubility according to Kelvin equation for droplets or crystal particles. For ubiquitous cellular structures, such as metals, ceramics, the growth of grain walls have a speed proportional to the mean curvature (Please see: J. phys. Condens. Matter 1992, 4, 1867-1894). The common phenomenon in these systems is that big individuals absorb the adjacent smaller ones. In other words, only the larger can survive within these system, which is very similar with the jungle law which means that only the stronger can survive in the jungle. The definition of Ostwald ripening recommended by IUPAC in 2007 is that "dissolution of small crystals or sol particles and the redeposition of the dissolved species on the surfaces of larger crystals or sol particles" (Please see: Pure Appl. Chem., 2007Chem., , 79, 1801Chem., -1829. Although Ostwald ripening often refers to the coarsening and recrystallization process in a nearly saturated solution, sometimes it can be used in a wider sense, such as the evolution of droplets or gas-liquid foams (Please see: Current Opinion in Colloid & Interface Science, 2010, 15(5), 374-381). But for the evolution of cellular structures, such as grains and froth, usually is called "coarsening" instead of "Ostwald ripening" (Please see: Nature,9 2007, 446(7139), 1053-1055). So we use a new and consistent term, "jungle law", to describe the evolution within these systems. Besides, we believe that "jungle law" is more impressive and appealing to the general audience of Nature Communications. 3. Comments: Given the space available, the manuscript would be stronger if the evolution equation were brought into the main text. How do the simulations quantitatively compare with the experiments? Reply: Thanks for the reviewer's suggest very much. We have brought the evolution equation into the main text in the revised manuscript. The simulations agree well with the experiments qualitatively. We have simulated evolution of 2D foams with different gas volume fractions, 34.6%, 55.8%, 56.9%, 71% in Supplementary Fig. 7, which also show qualitative agreement with the results in Supplementary Fig. 6. However, there are some problems for the quantitative simulation: 1) It is not well to predict the time the total evolution takes. As the evolution equation deduced by Lemlich shows,  [374][375][376][377][378][379][380][381], so the precise calculation of K is difficult, and the time of evolution cannot predict precisely. In our simulation, K was defined as 2 m 3 s -1 , and what we pay attention to is the process of evolution (gathering effect) and the final bubble patterns. 2) 2D foams with high gas volume fractions (more than 71%) is difficult to simulate. For wet 2D foams with low gas volume fractions, the evolution obey Lemlich's model; for dry foams with very high gas volume fractions (more than 95%), the evolution obey von Neumann's law. However, there has never been a quantitative equation describing the evolution of 2D foams with medium gas volume fractions (Please see: J. Phys.: Condens. Matter, 1992, 4, 1867-1894, Current Opinion in Colloid & Interface Science, 2010: 374-381) As discussed in Supplementary Note 1, we applied Lemlich's theory into 2D foams, so the lower the gas volume fraction is, the better the simulation quantitatively agree with the experiments. Although many efforts still need to make for quantitatively simulating the evolution, we think that our simulations show the "reverse jungle law" due to the narrow space between pillars and "gathering effect" because of the special arrangement of pillars, which agrees well with the experiments.

Comments:
Should the reader view these foams as wet or dry? Does this distinction affect the applicability of the model? Reply: Thanks for the review's comments. As we have pointed in the revised main text (Page 2, paragraph 3 ), in this work we mainly focused on evolution of foams which was usually called wet foams or bubbly liquid in 2D.
In Lemlich's theory, the bubbles in the foams were viewed as separate spheres as shown in the figure below, and the evolution was usually called Ostawald ripening. We applied Lemlich's theory into 2D foams evolution, and the bubbles were considered as separate circles. In the foams with high volume fraction, the bubbles squeeze each other and their deformed shapes apart from sphere or circle need to be considered. So the wetter the 2D foams are, the better our model is applicable to the foams. For the dry foams whose gas volume fraction is usually larger than 95% as shown in the figure below, the evolution obey the von Neumann's law or von Neumann's coarsening (Please see: page 100 in Foams: Structure and Dynamics. (Oxford University Press, 2013)), namely,

The wet foam or bubbly liquid that is applicable to
Where Deff is the effective diffusion constant and positive, Ai, ni are the area and number of sides for the bubble i, and t is time. The equation suggests, the rate of increase or decrease of a bubble's area depends only on its number of sides. A bubble with 6 sides will be invariable in the evolution, and the bubble with sides more or less than 6 will grow and shrink, respectively. This law for dry foams is very different from Lemlich's theory for wet foams. Therefore our model is not applicable to dry 2D foams.
Dry foam between two plates, viewed from above (bubbles of several mm). Ref: page 56 in Foams: Structure and Dynamics. (Oxford University Press, 2013)

Comments:
To what extent do the authors believe the results will extend beyond the specific experimental protocol (urea peroxide solution catalyzed by AgNP)? Reply: Thanks for the review's comments very much. In this work, we studied the evolution of 2D liquid foams constrained by micropatterned surfaces. The narrow space between micropillars endows the evolution with the "reverse jungle law" mechanism. The special arrangement of pillars contributes the "gather effect". These fundamental understandings enable to prepare various 2D bubble patterns.
The patterned bubbles can serve as the template for assembling AgNPs into desired 2D networks for electronic applications. We believe the results can extend beyond the urea peroxide solution catalyzed by AgNPs. We do the simulation regardless of the preparation method for the 2D liquid foams. In the simulation, we produced round bubbles with the random radius (in the range between zero and the maximum circular size that the pillars allow for) and at the random position, on condition that bubbles distribute evenly. Therefore such foams produced by any method will show the "reverse jungle law" and "gather effect".
We believe some other chemical reactions which emit gas will show the similar results we presents, for example, oxygen bubbles from ethanol solution of hydrogen peroxide catalyzed by ferric ion, carbon dioxide bubbles from carbonate reacting with acid, hydrogen bubbles from hydrolysis of sodium borohydride catalyzed by hydrogen ion, and so on. To prove the generality of our results, in the revised manuscript we have added experimental results in Supplementary Fig 17 and Supplementary Fig 18. The evolution of 2D foams consisting of hydrogen bubbles from hydrolysis of sodium borohydride catalyzed by hydrogen ion was studied, showing the same results with urea peroxide solution catalyzed by AgNPs. In addition, with these patterned bubbles as a template, The conductive polymer, PEDOT (poly(3,4-ethylendioxythiophene)) and polystyrene (PS) microspheres were assembled into hexagonal networks. In this work, materials from polymer (PEDOT) to nanoparticles with average size of 450 nm (PS microspheres) were assembled, therefore we believe this is a general strategy for assembling functional materials into desired 2D networks for device applications. Fig. 1e shows the phenomenon of "reverse jungle law", but it's not clear in the videos when this process occurs. The submission would be improved if the authors could include a video or longer series of images corresponding to Fig. 1e. Reply: Thanks for the reviewer's suggestion very much. We have added a longer series of images corresponding to Fig. 1e in the revised manuscript. Please see the supplementary Fig. 5 as follows.

Comments:
Supplementary Fig. 5 │ The longer series of images corresponding to Fig. 2a for showing the "reverse jungle law" in detail. The red arrows denote gas transfer direction among the central bubble and surrounding bubbles due to difference in radius of curvature (denoted with blue dashed circle). Scale bars, 50 μm.

Comments:
The manuscript would be improved if the authors would at least briefly describe the technique for pillar microfabrication in addition to providing the citation. What advantages are there in assembling lattices using two-dimensional foams over more traditional microfabrication techniques? Reply: Thanks for the reviewer's suggest very much. We have added the related description of the technique for pillar microfabrication in the revised Method Section. The traditional techniques to generate the networks at nanoscale are standard "top-down" microfabrication techniques such as electron beam lithography (Please see: Nano lett. 2012, 12(6): 3138-3144). They commonly involve a long preparation time, complex setup and high cost. As a result, the "bottom-up" assembly methods have been investigated, such as by inkjet printing (Please see: ACS Nano, 2009, 3(11), 3537-3542), by drying liquid bridges (Please see: Phys. Rev. Lett. 2009, 102(5), 058303) and using a bubble template (please see: Langmuir, 2012, 28(25), [9298][9299][9300][9301][9302]. Nevertheless, the precision of these assembly methods usually are very low (tens of micrometers), which cannot afford to fabricate electronic devices with high precision. Alternatively, we present a method combining the top-down and bottom-up options. The micropatterned surface was fabricated by the bottom-up method called deep reactive-ion etching, which assists the preparation of patterned bubbles. These bubbles serve as the template for assembling AgNPs with a top-down strategy. This strategy can assemble AgNPs into desired 2D networks with nanosacle resolution. The fabrication is simple, time-saving and low-cost. In addition, the micropatterned substrate can be recycled by a simple washing process. Therefore it is an efficient, clean and sustainable strategy and can be tailored to fabricate high-precision electronic devices.

Comments:
"Surface curvature radius" should be replaced by "radius of t t+10s t+20s t+30s t+40s t+50s t+70s t+60s curvature". Reply: Thanks for the reviewer's suggest very much. We have replaced the "surface curvature radius" with "radius of curvature" in the revised manuscript.
14 Reviewer 3 1. Comments: the main novelty of this work is that the authors have shown that the bubbles can be guided to evolve in particular shapes by designing the 2D porous media. Furthermore, the authors have shown that their technique can be used to fabricate patterns on glass substrates using Ag nanoparticles. While the technique to create patterns on glass using 2D controlled foam is a neat application, the reviewer is concerned with the numerous technical inconsistencies in the manuscript.
Reply: Thank the reviewer for the positive evaluation of our paper. We have revised the manuscript carefully to clarify our work more clearly.

Comments:
The reviewer disagrees with the usage of the phrases "jungle law" and "reverse jungle law" by the authors in context of these studies. There are plenty of examples in biosphere where small animals may be able to consume larger animals. To make a general sweeping statement with respect to evolution in biological systems, and then to extrapolate it to an engineered system is unacceptable, and appears more like an attempt by authors to make their work look more "attractive" rather than good science. What the authors mean by "jungle law" is essentially the phenomenon of Ostwald ripening where smaller bubbles are generally consumed by larger bubbles. Referring to the phenomenon of Ostwald ripening should have itself been sufficient without resorting to phrases like "jungle law" in this work. Further, the reviewer could not find any precedent in literature where the term "jungle law" has been used even in context of biology. Reply: Thanks for the reviewer's comments. The jungle law usually means that "survival of the strongest", "survival of the fittest", "kill or be killed", "eat or be eaten" and "every man for himself". The definition in Oxford English Dictionary is "the code of survival in jungle life, now usually with reference to the superiority of brute force or self-interest in the struggle for survival." The extended meaning of jungle law has been used in many fields The term "jungle law" in the biological system means that "survival of the strongest" or "only the stronger can survive". It doesn't mean that big animals eat smaller ones. Similarly, the code of survival for bubbles in the foam systems is the larger size, and in other words, only the bigger can survive. From this point of view, the jungle law can describe the free evolution of bubbles in the gas-liquid foams. We have revised the manuscript carefully to clarify the similarity of biological system and foams more clearly.
The definition of Ostwald ripening recommended by IUPAC in 2007 is that "dissolution of small crystals or sol particles and the redeposition of the dissolved species on the surfaces of larger crystals or sol particles" (Please see: Pure Appl. paragraph 3, line 3-10.

Comments:
The experiments represented in Figure 2 are very interesting.
However, the logic behind why the authors fabricated these special patterns is unclear. Arguably there could have been many different patterns that the authors could have studiedsquare, triangle, rectangle etc. The reviewers could have chosen the same pattern but varied the pillar spacing. What guided the authors to choose the special patterns for these experiments? Reply: Thank the reviewer for the positive evaluation of our paper. We have revised this part in the manuscript for clearly showing our work. In this paper, we investigated the effect of the pillar interval on the evolution. The same pattern with different pillar spacing has been studied in the preparation of hexagonal bubble arrays of different size, as shown in the revised Fig. 2c-e. The pillar spacing was demonstrated to determine the "reverse jungle law". For studying the effect of arrangement of the pillars. There are various choices, for example Fig. 4d, Fig 4e, Fig 4g and Fig 4h. We choose the pattern of Fig 4f to start our discussion with three reasons as follow.
1) The dodecagonal cell have the much larger area than that of the square (11.2 times) or hexagonal cell (4.3 times), which makes the "gathering effect" very obvious during the evolution. 2) The formed dodecagonal bubble arrays were discrete. And during the evolution, the bubbles in different dodecagonal cells cannot touch each other. This greatly simplifies the simulation because the deformation of bubbles owing to contact (Fig 4d, Fig 4e) need not to be considered. 3) The pattern is regular. The changes in area and radius of curvature for the bubbles can be precisely calculated when the bubbles grow and deform in the square, hexagonal and dodecagonal cells. Therefore quantitative explanations of the "gathering effect" are allowed.

Comments:
The authors state "When the total gas is enough, bubbles in hexagonal cells can survive because consumption of bubbles in square cells has filled the dodecagonal cells." What criterion can be used to know if the "gas is enough"? And how can we make sure that these conditions are met? Reply: Thanks for the reviewer's comments. The simulations of evolution for 2D foams with various gas volume fractions were shown in Supplementary Fig 7. When the gas volume fraction is less than 55.8% (such as 34.6%), the bubbles only fill part of dodecagonal cells. if the volume fraction is more than 56.9% (such as 71% in the simulation), the bubbles in hexagonal cells can survive. So the "enough gas" in the main text means that the gas volume fraction was about more than 57%. We have revised the manuscript to clarity it clearly. The traditional techniques to generate the networks at nanoscale are standard "top-down" microfabrication techniques such as electron beam lithography (Please see: Nano Lett. 2012, 12(6), 3138-3144). They commonly involve a long preparation time, complex setup and high cost. As a result, the "bottom-up" assembly methods have been investigated, such as by inkjet printing (Please see: ACS Nano, 2009, 3(11), 3537-3542), by drying liquid bridges (Please see: Physical review letters, 2009, 102(5), 058303) and using a bubble template (please see: Langmuir, 2012, 28(25), 9298-9302). Nevertheless, the precision of these assembly methods usually are very low (tens of micrometers), which cannot afford to fabricate electronic devices with high precision. Alternatively, we present a method combining the top-down and bottom-up options. The micropatterned surface was fabricated by the bottom-up method called deep reactive-ion etching, which assists the preparation of patterned bubbles. These bubbles serve as the template for assembling AgNPs with a top-down strategy. This strategy can assemble AgNPs into desired 2D networks with nanosacle resolution. The fabrication is simple, time-saving and low-cost. In addition, the micropatterned substrate can be recycled by a simple washing process. Therefore it is an efficient, clean and sustainable strategy and can be tailored to fabricate high-precision electronic devices. 2) In this work, Ag NPs were usedand the bubbles were formed by the catalytic activity of Ag. What if we want to deposit other materialslike aluminum, polymers etc. How can bubbles be produced in such systems, and how can such materials be deposited? We believe some other chemical reactions which emit gas will show the similar results we presents, for example, oxygen bubbles from ethanol solution of hydrogen peroxide catalyzed by ferric ion, carbon dioxide bubbles from carbonate reacting with acid, hydrogen bubbles from hydrolysis of sodium borohydride catalyzed by hydrogen ion, and so on. To prove the generality of our results, in the revised manuscript we have added experimental results in Supplementary Fig 17 and Supplementary Fig 18. The evolution of 2D foams consisting of hydrogen bubbles from hydrolysis of sodium borohydride catalyzed by hydrogen ion was studied, showing the same results with urea peroxide solution catalyzed by AgNPs. In addition, with these patterned bubbles as a template, The conductive polymer, PEDOT (poly(3,4-ethylendioxythiophene)) and polystyrene (PS) microspheres were assembled into hexagonal networks. In this work, materials from