Ionic liquid facilitated melting of the metal-organic framework ZIF-8

Hybrid glasses from melt-quenched metal-organic frameworks (MOFs) have been emerging as a new class of materials, which combine the functional properties of crystalline MOFs with the processability of glasses. However, only a handful of the crystalline MOFs are meltable. Porosity and metal-linker interaction strength have both been identified as crucial parameters in the trade-off between thermal decomposition of the organic linker and, more desirably, melting. For example, the inability of the prototypical zeolitic imidazolate framework (ZIF) ZIF-8 to melt, is ascribed to the instability of the organic linker upon dissociation from the metal center. Here, we demonstrate that the incorporation of an ionic liquid (IL) into the porous interior of ZIF-8 provides a means to reduce its melting temperature to below its thermal decomposition temperature. Our structural studies show that the prevention of decomposition, and successful melting, is due to the IL interactions stabilizing the rapidly dissociating ZIF-8 linkers upon heating. This understanding may act as a general guide for extending the range of meltable MOF materials and, hence, the chemical and structural variety of MOF-derived glasses.


REVIEWER COMMENTS
Reviewer #1 (Remarks to the Author): In this study, the ionic liquid [EMIM][TFSI] is incorporated into the pores of the MOF ZIF-8, Zn(mIm)2, to create a composite that melts prior to decomposition. However, TGA and NMR evidence show that the IL molecules themselves degrade during the melting process, and the final chemical composition of the glassy material is not easy to determine. The lowering of the melting point of the composite compared to pristine ZIF-8 is hypothesized to be due to the ionic liquid stabilizing free linkers through supramolecular interactions after they dissociate from the zinc ions. Although the final material is indeed quite complex, the authors use several material characterization techniques to interrogate the structure and composition of the glass. I recommend this paper for publication after several important points are addressed.
1) The DSC in Figure 2C  2) Throughout the article, terms need to be defined and discussed clearly. The authors use the term 'glass' to describe both their low temperature and high temperature materials; however, by PXRD the LT shows clear crystallinity so the definition of 'glass' is not clear. Additionally, they claim that the IL components interact strongly with the MOF components, effectively decreasing the melting temperature, but it seems more likely that decomposed (or decomposing) IL fragments are what stabilizes the amorphous MOF phase. Throughout the article the authors need to state whether they are referring to the pristine IL or decomposed IL fragments.
In addition to the above, we recommend some small changes to the text as follows: The chemdraws to show supramolecular interactions need work. The angles and distances of many of the supramolecular interactions would not be feasible in reality. This reviewer suggests creating a figure with several panels to show portions of possible interactions rather than an entire ZIF-8 pore, in which the details become messy and unclear.
Many future applications of melted MOFs rely on accessible surface area. The melted MOF herein has extremely low surface area and this should be mentioned in the main text. The method of using ILs to stabilize the melting process is interesting and new, but the drawbacks of the current technique should be discussed as well.
In general, the authors place too much interpretation of their results in the Results section. Interpretation of the data and comparing the data to previous work should be in the Discussion section.
Reviewer #2 (Remarks to the Author): This manuscript describes the effect of guest molecules as ionic liquids(IL) infiltrated into the pores of metal-organic framework, ZIF-8, on the melting behavior of IL@ZIF-8 composites. It is quite amazing to see that the IL incorporation decreased the melting temperature below the decomposition temperature. As the authors thoroughly investigated, the IL supports the dissociation of Zn-N coordination bond and stabilized the zinc ions and 2-methyl imidazole after the dissociation. On one hand, I totally understand the importance of this manuscript and the authors' logic as "the incorporated IL decreased the melting point of ZIF-8" as even highlighted in the title. On the other hand, I felt a bit uncomfortable with this statement. This is most likely because I would recognize this as a new composite of "IL@ZIF-8" rather than the ZIF-8 itself. In this case, we cannot say that the IL decreased the melting point of ZIF-8 but rather say that the IL@ZIF-8 showed the lower melting point than the pristine ZIF-8. In particular, the corresponding glasses, which were formed by the thermal treatment at 387 C (LT) or 390 C (HT) followed by quick-quenching, have the decomposed IL molecules therein and thus, the different chemical compositions. If these decomposed IL molecules were removed from the pores and produced a new phase of pristine ZIF-8 glass, I would definitely agree with this argument as the IL supports the melting of ZIF-8. In my opinion, the reversibility of guest incorporation/removal is necessary to justify the authors' claim. I would potentially suggest the acceptance of this manuscript, but this argument and the following scientific concerns should be addressed before the official acceptance.
1. How did the authors decide the temperatures for the thermal treatment at 387 C and 390 C for the LT and HT samples, respectively? Why does the LT sample maintain a partial crystallinity even though the sample was treated at the higher temperature than the melting point (381 C)? Did the authors first try the variable temperature XRD to find the appropriate temperatures?
2. The LT sample looks interesting because it still maintains the crystallinity. However, the thermal analysis was not reported in Figure 2c and d. Please provide these data and discuss how it is different from the HT sample, which is totally amorphous.

The authors choose this specific IL, [EMIM][TFSI]
, because it is hydrophobic. However, after the careful analysis using the solid-state NMR, this IL molecules effectively interact with the ZIF-8 framework via hydrogen bonds. Did the authors try another IL that is hydrophobic but not strongly interact with the framework? In this manuscript, only one IL was used so that it is more difficult to really understand the effect of IL on the melting behavior. Figure 5b, the data of the HT sample is missing. Was this crystalline peak disappeared in the HT sample?

In the
Reviewer #3 (Remarks to the Author): In this manuscript, Nozari, et al, incorporated organic ionic liquid into the pores of ZIF-8 which is a nonmeltable MOF, expecting the formation of ZIF-8 liquid prior decomposition at elevated temperature and hence the formation of glass after quenching the molten. Attempts are then made to understand the melting and the glass formation processes of ZIF-8 loaded with ionic liquid using DSC, FTIR, NMR, PDF, and XRD measurements. In addition, they examined the porosity of the melt-quenched glass using N2 gas adsorption measurement.
The idea of this manuscript may be good which will help to convert the non-meltable MOF crystals into glasses. But, I have several significant issues with this manuscript. My concerns are as follows: 1-The glass transition (@ 322oC, Fig. 2C) of the quenched glass ag(IL@ZIF-8-HT)is not clear! Cp trace is necessary to determine the accurate Tg and whether the observed endothermic peak is correlated with the amorphous MOF structure or it is due to the experimental error. Otherwise, the manuscript will lose its purpose and the melting of IL@ZIF-8 will be useless. In addition, the XRD and Cp measurements are needed for the heated ionic liquid at 390 oC to ensure the heated ionic liquid doesn't have amorphous or glassy nature. 2-Why did you measure the TGA trace of the ag(IL@ZIF-8-HT) glass only up to 400 0C (Fig. 2D)? consistency of curves is important. 3-The most significant concern is that-throughout the discussion/interpretation-I find consistently that the authors deeply focused on the melting mechanism of the IL@ZIF-8-HT, however, the investigation on the glassy nature of the composite sample is lacked. So, I recommend the authors to examine the influence of ionic liquid concentration on the glass formation process, transparency, and porosity (i.e. ZIF- Response. We appreciate the reviewer's comment. Accurate assignment of the glass transition is a challenge in these hybrid materials and the reviewer is absolutely right in that careful Cp scanning should be performed to avoid misinterpretations. As suggested by the reviewer, we now conducted a range of cycle Cp scans for ag(IL@ZIF-8-HT) samples with altered temperature programs (up-/down-scan rates). Accordingly, we updated Figure 2, adding the cycle Cp scan using 20 °C·min -1 as heating and cooling rate as Figure 2d, as shown below. Details of the Cp measurements are summarized in Methods section and highlighted in yellow (in short, the Cp scan is a referenced DSC scan). Corresponding DSC scans are shown in Figure 2c.
Regarding the Tm, as it is mentioned in the paper, the onset of melting of IL@ZIF-8 was detected at 381 °C, however, complete melting occurs once the material is heated for 40 minutes isothermally at 390 °C. Response. We thank the reviewer for this excellent comment. On one occasion in the text where we refered to the LT sample as the melt-quenched glass, and we corrected from "and in the melt-  According to our DSC and TGA data presented in Figure 2c and Supplementary Figure S2, melting of the IL@ZIF-8 occurs before reaching the decomposition temperature. This means that strong interaction between the IL and ZIF-8 is the major parameter leading to the melting. However, the reviewer is absolutely right in that complete melting and formation of the glass requires an isothermal step where some of the IL molecules partially decomposes. Based on this, we added ''(or partially decomposed IL molecules)'' and ''(or partially decomposed IL fragments)'' where we mention that the IL stabilizes the interactions on pages 16 and 23, respectively.
Comment: In addition to the above, we recommend some small changes to the text as follows: Response. We again thank the reviewer for this comment. It is challenging to represent a complicated system with only chemdraws. We have done as the reviewer suggested -created cutaways highlighting the important and unique interactions in Figure 6 and Figure 7. Additionally, we moved the whole pore diagram to the Supplementary Information. Figure  Response. We thank reviewer for this very valuable comment. To investigate the porosity of the ag(IL@ZIF-8-LT) and ag(IL@ZIF-8-HT) samples, we performed additional gas adsorption measurements using N2 at 77 K and CO2 at 273 K, as well as washing experiments by which residual IL/IL fragments were partially removed from the IL@ZIF-8 glass. This comment is essentially equivalent to a comment of reviewer 2. Please refer further to our response to reviewer 2 and reviewer 3.

section. Interpretation of the data and comparing the data to previous work should be in the Discussion section.
Response. We appreciate reviewer's comment. However, we are unsure about the article guidelines of Nat. Commun., where the "Discussion" section is -in our understanding -"only" a short summary/conclusion and outlook paragraph, whereas the major part of the scientific interpretation and data evaluation is indeed placed in the "Results" section, and sub-sections are not commonly used in the journal. We would gladly adapt these sections, but some advice from the editor would be very helpful. Response. We are absolutely with the reviewer; this is an important comment. As suggested by the reviewer we adapted the manuscript in two regards: when referring to the ag(IL@ZIF-8-HT) glass, we point-out that this is a composite glass, not simply a "ZIF-8 glass" (the LT material is a crystalglass composite in itself). More importantly, we investigated the removal of unreacted/decomposed IL from the ag(IL@ZIF-8-HT) and ag(IL@ZIF-8-LT) samples by washing the samples using acetone or DMSO as solvents (whereby the use of acetone turned out to be preferential in terms of observing the results of washing before/after using IR-ATR spectroscopy on the washed sample and on the washing solutions, respectively). After washing, we first investigated the qualitative optical appearance of the samples using a digital microscope as mentioned in the methods section (whereby washed samples became significantly brighter). Then, N2 and CO2 gas adsorption measurements were conducted on samples before and after washing (whereby a significant, i.e., fourfold enhancement of total porosity was detected, taken as evidence for the removal of IL  In summary, these new data clearly show that the residual IL can at least partially be removed from the quenched IL@ZIF-8 glass by washing in acetone. This washing process does not alter the glassy nature of the material, or dissolve the glass. It leads to very significantly enhanced material porosity, which is perfect starting point for future research, e.g., regarding potential applications in gas separation or adsorption. At the present stage, we believe that it provides strong confirmation for the broad interest in using IL-assisted melting of MOFs to enhance the variety of MOF glasses. Response. We thank the reviewer for this comment. IL@ZIF-8 composite melting starts at 381 °C, as can be seen from the melting peak in DSC, and this was also confirmed by in-situ microscopic observation, shown in the supplementary video 2. However, complete melting of the composite occurs only after heating the sample isothermally for a given time; we observe that the melting reaction is slow (which is common in the melting of MOFs and inorganic zeolites, and is related to the viscosity of the obtained liquid phase, which is -being close to Tg -very high, and, eventually, to the kinetics of IL decomposition and IL-ZIF-8 interaction). Based on different heat-treatments which we conducted on IL@ZIF-8 and subsequent analysis using XRD presented in the Figure   below, LT and HT conditions were selected to obtain partially-amorphized and fully-amorphized samples, respectively. We note here that the partially amorphized material might be particularly interesting for potential applications such as gas separation; however, a detailed analysis of the relation between process parameters and stabilization of a given fraction of crystalline ZIF-8 (by only partially melting) is not in the scope of this present study.

This manuscript describes the effect of guest molecules as ionic liquids (IL) infiltrated into
We included a Figure  remains the primary phase in this case, and neither ZIF-8 nor the IL exhibit a DSC feature in this temperature range (Figure 2c) Normalized intensity / a.u.  is not clear! Cp trace is necessary to determine the accurate Tg and whether the observed endothermic peak is correlated with the amorphous MOF structure or it is due to the experimental error. Otherwise, the manuscript will lose its purpose and the melting of IL@ZIF-8 will be useless.
In addition, the XRD and Cp measurements are needed for the heated ionic liquid at 390 o C to ensure the heated ionic liquid doesn't have amorphous or glassy nature.
Response. We thank the reviewer for this comment. This comment is essentially equivalent to comment 1 of reviewer 1. Please refer to our response to the first comment of reviewer 1. In short, we now added Cp cycling data, which unambiguously show that the observed Tg is not a misinterpretation or due to experimental error. We thank the reviewer for pointing this out.

Furthermore, we performed XRD of [EMIM]
[TFSI] and presented those data in Figure 2.

Comment 3-
The most significant concern is that-throughout the discussion/interpretation-I Nevertheless, we synthesized three more IL@ZIF-8 composites with different loadings of 5, 10, and 20 wt%, IL@ZIF-8-5 wt%, IL@ZIF-8-10 wt%, and IL@ZIF-8-20 wt% according to the synthesis procedure explained in the materials section of the manuscript. The obtained composites were characterized using XRD and DSC. Figure 4 shows XRD patterns of the synthesized composites.  Based on the DSC data of newly synthesized samples, lower IL loadings are not sufficient to induce melting of IL@ZIF-8. A specialized study of the exact effects of the amount of IL relative to ZIF-8 might be beneficial in the future, but we point-out that such a study would need to take into account many parameters aside the pure mixing ratio, e.g., including the ZIF-8 crystallite size and morphology, the exact mixing parameters etc. We feel that this is outside of the scope of our present report on the discovery of IL-facilitated melting of ZIF-8. Response. We appreciate the reviewer's comment. This comment is essentially equivalent to comment 1 of reviewer 1, and also to the first comment of this reviewer. Please refer to our response to the first comment of reviewer 1. By definition, the glassy nature of a material is proven when the occurrence of a glass transition is demonstrated. We hope that our newly added Cp scans, together may convince this reviewer in this regard. Further confirmation is from PDF analysis in Figure 8, showing the loss of any long-range order and retention of short-range order in ag(IL@ZIF-8-HT) glass. applications of these new types of glasses is an emerging field. Nonetheless, we investigated the optical, mechanical, and microstructural properties of the ag(IL@ZIF-8-HT) glass in order to provide incentives towards potential applications. 5,11 The following experimental details are included in Supplementary    Response. We thank the reviewer for this comment. Although the ag(IL@ZIF-8-HT) glass reported in this work has a brownish color, measured optical properties illustrated in the response to the previous comment clearly showed optical transparency. In general, glasses can be colored by adding small concentrations of dopants. 18 Furthermore, in the observed increase in optical transparency when approaching the UV edge (~ 450-350 nm), we find evidence for the absence of crystal or other light-scattering inclusions. In response to the first comment, the glassy nature of ag(IL@ZIF-8-HT) was investigated by performing cycle Cp scans, which unambiguously showed a glass transition. We hope that these new data may convince the reviewer that the reported material is indeed a glass. Response. We respectfully disagree with this comment and have pointed towards initial suggested applications elsewhere in this response letter. However, we would also like to point out that i) the study is a fundamental one and fits within the remits of Nature Communication as an advancement in the field of MOF glasses and ii) when metallic glasses were first produced the reaction from some was very similar. As we have stated, melt-quenched MOF glasses have emerged very recently, and it is a nascent field of research. Even so, several promising applications have been presented in the literature, including gas separation and battery applications. 5,11 We hope that the additional data we present in the revised version of this manuscript might further convince this reviewer of the benefits of our present approach.