Ceramic nanowelding

Ceramics possess high temperature resistance, extreme hardness, high chemical inertness and a lower density compared to metals, but there is currently no technology that can produce satisfactory joints in ceramic parts and preserve the excellent properties of the material. The lack of suitable joining techniques for ceramics is thus a major road block for their wider applications. Herein we report a technology to weld ceramic nanowires, with the mechanical strength of the weld stronger than that of the pristine nanowires. Using an advanced aberration-corrected environmental transmission electron microscope (ETEM) under a CO2 environment, we achieved ceramic nanowelding through the chemical reaction MgO + CO2 → MgCO3 by using porous MgO as the solder. We conducted not only nanowelding on MgO, CuO, and V2O5 nanowires and successfully tested them in tension, but also macroscopic welding on a ceramic material such as SiO2, indicating the application potential of this technology in bottom-up ceramic tools and devices.

The authors report a new technique to weld nanowires with a ceramic joint, as a result of a chemical reaction occurring in an environmental transmission electron microscope. They synthesized a ceramic nanowelding material through carbonation of the MgO nanowire under the electron beam, without using external heat or current, producing nanocrystalline MgO as resulting material joint which, upon tensile testing, appears to be stronger than the MgO nanowire itself. The work is interesting and provides new insights for the welding technology of oxide ceramics. I recommend publication, with a minor revision: I have one comment regarding an acronym: In p.3, paragraph 2, line 4 the authors write scanning transmission microscopy (STM) to refer to the sample holder. I guess they mean scanning tunneling microcopy.
Reviewer #2 (Remarks to the Author): This manuscript by Zhang et al. presents a novel in situ method to weld several oxide nanowires by using porous MgO as solder and under CO2 environment, generating welding spots with strengths comparable to the welded nanowires. The authors suggest that the present method has a wide applicability not only to the bottom-up buliding of ceramic nanodevices, but also to the macroscopic welding in ceramic industry. The results are valuable, the testing and characterization carried out and presented well; however, I think the authors should address the following issue clearly to make their claims convincing, i.e., what is the role (pros and cons) of the porous structure of the nanowires focused here? It seems that the high specific area of the porous nanowires makes them easily wetted (and permeated) by the MgCO3 fluid, but the high porosity is also responsible for the relatively low strength of the "pristine" nanowires, especially compared to typical single-crystalline ceramic nanowires. Therefore, the authors are prompted to provide a clear evidence that the present method can also weld a single-crystalline nanowires (SiC, ZnO, etc., perhaps the CuO nanowire shown here?) with a satisfactory strength of welding spots, i.e., comparable to the welded nanowires, before this manuscript can be considered for publication.
Reviewer #3 (Remarks to the Author): Referee report for manuscript-NCOMMS-17-18596 High-quality joining of ceramics is a bottleneck challenge that limits the applications of highperformance ceramics in the vast of industrial applications. To address this issue, the authors have proposed a rather novel and efficient approach, named as cold welding, in which the energetic beam irradiation assisted chemical reaction in the presence of CO2 gases plays the key role, and the whole process was performed inside a state-of-art environmental TEM. Successful examples on connecting MgO, CuO and V2O5 nanowires were tested in situ, and the re-joined nanowires exhibit remarkable mechanical properties as evidenced by tensile tests. The authors further explored the potential applications of this method on macroscopic welding on other ceramic materials such as SiO2. All these results have clearly collaborate the promising future of this novel method on the welding of ceramics in the future. As such, the present referee would like to recommend its publication in Nature Communications.
While the present form is still far from mature, I would suggest the authors consider the following issues in their revised manuscript. 1. Can the authors give more explanations on the reasons why the re-joined nanowires are stronger in terms of mechanical properties than the pristine one? Is that mainly due to the polycrystalline or highly defective nature of the chosen nanowires for the welding experiments? Does the contact strong enough and remains unchanged during the whole test? An ideal candidate material would be the single crystalline nanowires. Major point.
2. Is there limits in terms of diameter for the cold welding? Minor point 3. Most important issue-beam effect. i) It is well know that metal oxides may undergo a radiolysis process upon the illumination of highenergy electron beam, and sometimes leads to the formation of metal, such as metal-copper. The authors mostly focus on the direct reaction of metal oxides (MgO, CuO) with CO2, while pay less attention to the former one. During the whole process, if metal, say Cu was formed due to the beam-damage, and then the welding could be easily realized through Cu-Cu metalling bonding, rather than via the beam assisted chemical reaction between CuO and CO2. And then, such a metallic junction may be further oxidized in the presence of surround oxides (may serve as the source oxygen during the beam damage process) and the CO2 gases. This possibility of this procedure should be carefully examined and discussed. By the way, the cold welding of metallic nanowires via in-situ TEM was also demonstrated previous by one of the present authors. ii) As the electron beam plays an indispensable role during the whole welding, it is necessary to quantify the contribution of electron dose. Is there any dose dependence there? iii) Interaction of energy electron beam with CO2 gases will induce the ionization process, and lead to the formation of highly reactive products as a result of radiolysis process, particularly close to the sample surface. It should also be addressed in the discussion.

Reviewer #1:
The authors report a new technique to weld nanowires with a ceramic joint, as a result of a chemical reaction occurring in an environmental transmission electron microscope. They The results are valuable, the testing and characterization carried out and presented well; however, I think the authors should address the following issue clearly to make their claims convincing, i.e., what is the role (pros and cons) of the porous structure of the nanowires focused here? It seems that the high specific area of the porous nanowires makes them easily wetted (and permeated) by the MgCO3 fluid, but the high porosity is also responsible for the relatively low strength of the "pristine" nanowires, especially compared to typical single-crystalline ceramic nanowires. Therefore, the authors are prompted to provide a clear evidence that the present method can also weld single-crystalline nanowires (SiC, ZnO, etc., perhaps the CuO nanowire shown here?) with a satisfactory strength of welding spots, i.e., comparable to the welded nanowires, before this manuscript can be considered for publication.
Response: We thank the referee for his/her constructive suggestions. Herein, we report a novel technique to weld ceramic nanowires with mechanical strength of the welding spots even stronger than that of the pristine nanowires. The referee suggests that we should provide clear evidence that our method is also applicable to single crystalline nanowires. This is an excellent point, as conventional wisdom tells us that a porous structure usually possesses lower mechanical strength than a solid structure. As the referee suggested, we indeed performed welding experiments on single crystalline nanowires with no pores, and the conclusion remains the same, namely the welding spots are so strong that the single crystal nanowires all broke from the nanowires rather than from the welding joints. We conducted welding and then tensile test experiments on  Fig. 16, Video 7). In the meantime, the fact that all nanowires broke from locations other than the welding spots (Fig. 4d, Supplementary Figs. 19, 21) proves that the mechanical strength of the welding spots is stronger than that of the single crystalline CuO and V 2 O 5 nanowires. It should be noted that the above results were presented in our original manuscript, however, we did not point out that the tested CuO and V 2 O 5 nanowires are single crystals. We apologize for this missing information, which is now provided in our revised manuscript (line 240, page 8).  In summary, the welding technology presented in our manuscript is not restricted to porous MgO nanowires but applicable to single crystalline nanowires such as CuO and V 2 O 5 nanowires as well, indicating the broad applications of this novel welding technology. Response: We thank the referee for his/her recognition of our work and very positive comments on our manuscript. Referee 3's main suggestion is essentially similar to that of referee 2: that is the strong welding may originate from the porous MgO nanowire, therefore we should conduct welding and tensile test experiments in single crystalline nanowires to demonstrate that the mechanical strength of the welding spot is stronger than that of the pristine single crystalline nanowires. This is an excellent suggestion. As we already addressed in our response to referee 2, we indeed conducted welding and tensile test experiments on single crystal CuO and V 2 O 5 nanowires. The results reinforce our conclusion that the mechanical strength of the welding spots is stronger than that of the pristine single crystalline nanowires, as both the single crystalline CuO and V 2 O 5 nanowires broke from the nanowires rather than from the welding joints (Fig. 4d,  Supplementary Figs. 19, 21).

It is well known that nanocrystalline materials generally demonstrate superior mechanical properties as compared to its bulk counterpart due to the well-known
Hall-Petch strengthening mechanism. The same principle may apply to the ceramic MgO welding joints, which showed remarkable mechanical strength, which explains why the mechanical strength of the welding spots is even stronger than the pristine nanowires. We recently conducted extensive tensile tests on the nanocrystalline MgO formed in the weld junction, and found that the fracture strength of the MgO nanocrystal junctions can reach as high as 2.8 GPa (Fig. R3), indicating the welding joints is indeed very strong.
In response to the question: "Does the contact strong enough and remains unchanged during the whole test?", the answer is yes. We did not see changes or slippage of the contact during all the tensile testing experiments.

Fig. R3
The in situ tensile tests for the MgO nanocrystals on the weld junction. (a, b) A welding junction was formed between a W tip (on the left) and an AFM cantilever (on the right). (c, d) The welding junction was then pulled towards the left direction until it fractured. The tensile strength for this particular weld was 2.8 GPa.

Is there limits in terms of diameter for the cold welding? Minor point
Response: There appears to be no limits in terms of the diameter for welding of nanowires. In our experiments, MgO is only used as the solder and there is no restriction on the size of the target nanowires. However, the bigger the nanowires, the longer it takes to weld the nanowires. As shown in the Fig. R4, both MgO nanowires with small (Figs. R4 a,b) and large diameters (Figs. R4 c,d) can be welded by using this technique.

Moreover, we can even weld macroscopic ceramic materials by using this method (Figs.
R4 e,f).

Most important issue-beam effect.
i) It is well known that metal oxides may undergo a radiolysis process upon the illumination of high-energy e-beam, and sometimes leads to the formation of metal, such as metal-copper.
The authors mostly focus on the direct reaction of metal oxides (MgO, CuO) with CO2, while pay less attention to the former one. During the whole process, if metal, say Cu was formed due to the beam-damage, and then the welding could be easily realized through Cu-Cu metalling bonding, rather than via the beam assisted chemical reaction between CuO and CO2. And then, such a metallic junction may be further oxidized in the presence of surround oxides (may serve as the source oxygen during the beam damage process) and the CO2 gases.
This possibility of this procedure should be carefully examined and discussed. By the way, the cold welding of metallic nanowires via in-situ TEM was also demonstrated previously by one of the present authors.
Response: We agree that e-beam irradiation does cause a certain degree of irradiation damage to materials, but the welding of two CuO nanowires only through e-beam irradiation without CO 2 gas was not possible in our experiments. As shown in Fig. R5, two CuO nanowires were connected in the ETEM. After focusing the e-beam to the connection junction for 30 minutes with a high dose rate (~900 e/nm 2 ·s) without CO 2 gas, welding did not take place between the two nanowires. Therefore, the proposed cold welding mechanism did not apply to our experiments (Fig. R5). Although the target ceramic nanomaterials with different types and sizes can be welded together by using our technique, it is found that only MgO and CaO can be used as the solder for this ceramic welding (Fig. R6 a,b). As shown in Fig. R6 c,   ii) As the e-beam plays an indispensable role during the whole welding, it is necessary to quantify the contribution of electron dose. Is there any dose dependence there?

It has been reported that
Response: E-beam dose rate plays a significant role in the welding process (Fig. R7).
When the e-beam is blank, no reaction occurred between the MgO and CO 2 (Fig. R7a).
By increasing the e-beam dose rate to about 100 e/nm 2 ·s, the reaction proceeds with medium-speed (Fig. R7b). When the e-beam irradiation dose rate reaches about 870 e/nm 2 ·s, the high-speed reaction quickly leads to formation of large amount of highly mobile bubbles in the interior of the irradiated MgO nanowire (Fig. R7c). This is addressed in our revised manuscript (line 88, page 3).

Fig. R7
E-beam dose rate plays a significant role in the welding process. (a) When the e-beam is blank, no reaction occurred between the MgO and CO 2 . (b) By increasing the e-beam dose rate to about 100 e/nm 2 ·s, MgO reacted with CO 2 with medium-speed. (c) When the e-beam irradiation dose rate reaches as high as 877 e/nm 2 ·s, a large amount of highly mobile bubbles emerged in the interior of the irradiated MgO nanowire.
iii) Interaction of energy e-beam with CO2 gases will induce the ionization process, and lead to the formation of highly reactive products as a result of radiolysis process, particularly close to the sample surface. It should also be addressed in the discussion.