Stepwise introduction of three different transition metals in metallo-supramolecular polymer for quad-color electrochromism

Metallo-supramolecular polymers (MSPs) show unique electrochemical and optical properties, that are different to organic polymers, caused by electronic interactions between metals and ligands. For the development of quad-color electrochromic materials, here we report the stepwise introduction of three different transition metal ions into an MSP, utilizing the different complexation abilities of the transition metals. An MSP with Os(II), Ru(II), and Fe(II) (polyOsRuFe) was synthesized via a stepwise synthetic route through the formation of an Os(II) complex first, followed by the introduction of Ru(II) to the Os(II) complex, and finally the attachment of Fe(II) to the Os(II)-Ru(II) complex to produce the polymer. This synthetic procedure was extended to fabricate MSPs that comprised Co(II)/Ru(II)/Os(II) and Zn(II)/Ru(II)/Os(II). The synthesized MSPs showed a broad optical and electrochemical window due to the coupling of three heterometallic segments into the polymer. Introducing acetate anion as the counter anion greatly enhanced the solubility of polyOsRuFe in methanol. A thin film of polyOsRuFe was prepared on ITO/glass by spin-coating the methanol solution, and its reversible quad-color electrochromism was demonstrated.

<15> In Supplementary Figures 12, 17 (and also Figure 5), the differences between the exp. and calcd. mass values are not within error range normally accepted for HR-MS. For LR-MS, these are fine. Is there any explanation for the same?

Answer for Reviewer-1's Comments
Remark: This manuscript  by Higuchi and coworkers demonstrated an excellent synthetic approach to access well-engineered hetero-trimetallic assembly in solution. The synthetic work deserves special credit in this manuscript. The problem and aim of the work were explained nicely, and the importance of ready access of well-engineered redox-active multi-metallic assembly for developing future voltage-tenable multi-electrochromic materials and devices was explained clearly. The claims and conclusions are coherent. I agree fully with the fact that assembly of multiple, different, properly tunable redox-active metal complexes into a processable polymeric/any other composite material is highly important when looking toward development of voltage-tuned electrochromic and similar devices. This work certainly fits for this journal considering the novelty, importance, competence, and standard/quality of the research described herein. However, according to the opinion and assessment of this reviewer, some crucial points (noted below) need to be addressed in the revision cycle before acceptance. These suggestions are intended to not only improve the competence, quality and strength of the current version but also to benefit the other peer researchers in terms of reproducibility and validity of the work.
Reply: We are highly thankful to the reviewer-1 for reviewing and spending your valuable time with our manuscript. We are also grateful for your excellent remark and positive assessment about our manuscript. Thank you for your valuable suggestions to revise our manuscript. We have revised our manuscript accordingly.
Comment 1: Page 3, 2nd paragraph, Line 1: "We noticed that a reaction of a metal complex with an organic compound was useful to." What does it mean? Is it incomplete somehow to convey the intended message? Please check.
Reply: Many thanks for your pointing out. We corrected the sentence as follows.
In Page 3, "We noticed that a reaction of a metal complex with an organic compound was useful to create a metal-containing ligand which can undergo further complexation with heterometals ions to give heterometallic supramolecular complexes or polymers. Therefore, we assumed that similar strategy could be utilized to organize three heterometal ions into MSP." Comment 2: The authors stated that they accomplished "stepwise harness of strong coordination metal ion Os(II) followed by another strong coordination metal ion Ru (II)  Reply: Many thanks for your nice suggestion. According to your suggestion, we tried to find the binding constant value for [Os(tpy)2]2+, [Ru(tpy)2]2+, and [Fe(tpy)2]2+ complexes. Unfortunately, we don't find any reports showing the exact value of the binding constant of three metal complexes (only for [Fe(tpy)2]2+ was found). However, we found two literatures (J. Mass Spectrom. 2003, 38, 510−516 and Eur. J. Inorg. Chem. 2015, 2015 that clearly show the binding strength of [Os(tpy)2]2+, [Ru(tpy)2]2+, and [Fe(tpy)2]2+ complexes follows the order Ru>Os>Fe. That mean Os(II) and Ru (II) form strong coordination compared with Fe(II). Here, we should mention one thing that although the study by Newkome et al. (Eur. J. Inorg. Chem. 2015, 2015 revealed that the binding strength of [Os(tpy)2]2+, [Ru(tpy)2]2+, and [Fe(tpy)2]2+ complexes follows the order Ru>Os>Fe, our synthetic strategy is not exactly based on the consideration of the order of the binding strength of metal ions. We have designed and developed the synthetic route by considering two things; reaction conditions of Os(II), Ru(II), and Fe(II) with tpy and the stability of tpy-M(II)-tpy connectivity. That is why, we first made complexation with Os(II), followed by Ru(II) and then Fe(II).
We have updated the text in revised manuscript by informing the reported result of these two reported literatures for giving a quantitative idea to the readers. We have modified this in introduction section as well as in result section as follows. If yes, please mention in the manuscript/supporting information so that future researchers will not be confused.
Reply: According to your suggestion, we have modified Scheme 1. Actually, the Ru(Br-tpy) can be attached in both side of TOsT. But we can get the desired product with high yield (one side coupling) by controlling the molar ration of the reactants to 1:1. Also, during column chromatography we collected the first fraction as the desired product. Following the first fraction, a second fraction of undesired product was appeared, which was probably the product of both side coupling. We have updated this information in the revised manuscript as well as in revised supporting information (SI) as follows.
In Page 6, "However, during separation of TOsRuBr by column chromatography, the first fraction was collected as the desired product (following first fraction, a second fraction of undesired product was appeared, which was probably the product of both side attachment of compound 4 with TOsT)." In Page 13 of Supporting Information, "to collect first fraction as desired product (a second fraction of undesired product was appeared, which was probably the product of both side attachment of compound 4 with TOsT)" Reply: Many thanks for your concern about this matter. During study of complexation behavior of TOsRuT, we successively added Fe(BF4)2 to the solution of TOsRuT and recorded absorption spectra of the solution. Upon gradual addition of Fe(BF4)2, the solution of TOsRuT got saturated and the polymer was precipitated. We recorded the absorption spectra of the solution mixture until it forms the precipitate. The last point (and the absorption spectrum) shown in Fig. 1b and 1c was recorded before the precipitation of the solution mixture. From the titration and plot of [Fe(BF4)2]/[TOsRuT] ratio, we estimated the molar ration to 1:1. We have updated the information about the precipitation of the polymer in our revised manuscript as follows.
In Page 10, "(when the solution mixture was reached to saturation, the polymer was precipitated)" Comment 5: Why was no mass spectrometric (ESI, or MALDI or TOF-SIMS) experiment attempted for polyOsRuFe?
Reply: Thank you for your suggestion. Actually, we attempted MALDI experiment for polyOsRuFe previously (after following few recent literatures), but no fruitful result was obtained. That is why we measured molecular weight by RALLS. Anyway, still our effort is in progress. If we get success, the result will be explored in a later manuscript.  And the following explanation was added in the revised manuscript.
In Page 15, "The scan-rate-dependent (0.01−0.1 V/s) CV study of polyOsRuFe displayed linear proportionality of the peak current with the scan rate ( Fig. 3d and Supplementary Fig. 21b Reply: Many thanks for your kind advice. Actually, we measured CV of TOsRuT to confirm the presence of Os(II) and Ru(II). The TOsRuT contains two free tpy units along with Os(II) and Ru(II) complex where as polyOsRuFe contains the complex of Os(II, Ru(II) and Fe(II). So, it seems difficult to compare the redox potential data between them. Still, we may think the less E1/2 value of Os(II)/Os (III) and Ru(II)/Ru(III) pairs in TOsRuT compared with polyOsRuFe is due to the presence of free tpy units in TOsRuT.
In case of polyOsRuCo, the CV was measured using 0.1 M tetrabutylammonium perchlorate (TBAP) in acetone as electrolyte, whereas the CV of polyOsRuFe was measured in 0.1 M lithium perchlorate (LiClO4) in CH3CN as electrolyte. So, the redox potential of Os(II)/Os(III) and Ru(II)/Ru(III) pairs will vary in polyOsRuCo and polyOsRuFe as the electrolytes and solvents are different. So, no comparison between them can be made.
We tried to measure the electrochemical property of polyOsRuCo using 0.1 M lithium perchlorate (LiClO4) in CH3CN as electrolyte, but no redox property of Co was detected. This may be due to counter anion effect, which may suppress the oxidation/reduction of Co.  Table 3). And the following explanation was added in the revised manuscript.
In Page 15, "The observed redox potential of three heterometal ions in polyOsRuFe has been found to be comparable with the previously reported heterometallic supramolecular complexes and polymers (see Supplementary Table 3 for the details of comparison)." A brief discussion on this issue would be beneficial to the readers.

Supplementary
Reply: Actually, this kind of peak is quite common for polypyridyl containing metallo-supramolecular polymers with Os and/or Ru and/or Fe. When the electrochromic property of polypyridyl containing supramolecular systems is studied on ITO, this type of peak appears after oxidation of metal center into the materials. This peak is mainly originated from the ligand-oxidized metal part. In our case, gradual oxidation of three different metal ion, the amount of ligand-oxidized metal part increases, which results gradual increase of the intensity of 401 nm peak. According to your suggestion, we have added this discussion in our revised manuscript as follows.
In Page 24-25, "This 401 nm peak may be due to the terpyridine-oxidized metal part which gradually increases upon successive oxidation of the heterometal ions." Reply: Thank you for your concern in this matter. Actually, we studied few more redox cycles of the spin coated polymer film on ITO. But we wanted to show the result in a later manuscript where we want to disclose details electrochromic property of the polymer and its device property. Anyway, according to your suggestion, here we have given the date for cyclic stability for at least 300 cycles. The result is shown below. They were added in SI as Supplementary Figure 33. c & d. If you see the redox stability of the polymer (polyOsRuFe-OAc), it is stable for at least 1000 switching cycles (see Supplementary   Figure 32d). So, we expect, our material will show high durability in terms of redox/electrochromic switching, which will be explored later. And the following explanation was added in the revised manuscript.
In to 300 cycles; monitored at 502 nm upon switching the potential between 0 and 1.2 V during chronoamperometry measurement.

Reply:
We are sorry for this mistake. We have corrected it in the revised SI file (the table 5 have been   changed to Table 6 for revision purpose).
Comment 13: Supplementary File, Section 1.9: Please mention the solvent used in the spectroelectrochemistry experiment.
Reply: Thank you for your suggestion. We have mentioned this in the revised SI (Page 50).    Reply: This is probably because we calculated the mass considering the peak with the highest intensity.
But now we modified it by the comparison with the theoretical peak pattern. Hope these results will be satisfactory. Thank you for your kind suggestion. Please check the revised figures ( Figure A, B & C shown above) in SI.
Finally, thank you very much once again for taking precious time from your work and helping us in improving this manuscript. We have tried our best to address every single concern that you have raised.
We sincerely hope that current fully revised manuscript would meet all your concerns. Reply: We are highly thankful to the reviewer-2 for reviewing and spending your valuable time with our manuscript. We are also grateful for your positive consideration of our manuscript for publication.

Answer for Reviewer-2's Comments
We have considered the concerns raised by you and revised our manuscript according to your valuable suggestions.
Comment 1: Despite the success in making compound TOsRuT, the synthetic strategy towards the final coordination polymer leads to a mixture of species where, within the linear arrays, the sequence of metals is not unique (see Fig. 2 on page 11). This poses a question of whether the &#x201C;precise synthesis of the polymer&#x201D; could be claimed, as stated, and whether the strategy described can be in the future modified in order to fix this problem.
Also, the way the authors describe this problem should be made clearer (writing that &#x201C;the polymer may be formed in two possible structures&#x201D; is quite misleading).

Reply:
The TOsRuT can be considered as a modified ditopic ligand containing two heterometal ions. Now, if TOsRuT undergoes to coordination-driven self-assembly with a third heterometal ions, the resultant polymer definitely will be a heterotrimetallic supramolecular polymer. We think, the synthesis could be claimed to precise synthesis as no other mode of synthesis is possible to introduce (as well as organize) three heterometal ions in a metallo-supramolecular polymer chain in homoleptic environment. Now, we extended our thought in a micro-scale observation of the structure of TOsRuT. We can consider the final ligand as asymmetric structure due to presence of one Os and one Ru center. Therefore, we anticipated that the polymer may be formed in either head-to-tail structure or a mixture of head-totail and head-to-head/tail to tail. However, if there is any head-to-head/tail-to-tail possibility, we may observe additional characteristic in UV/CV..etc. But we don't observe that kind of signature to proof our thought. But theoretically, it should be possible. That is why we mentioned the fact in a paragraph, so that the readers/researchers can get a clear idea about the possibilities (and to find/investigate more).
We think, using bulk synthesis, this defect (if there?) can not be controlled. The only way, if anyone find any experimental evidence to support the above assumption.
We have modified the paragraph for Fig. 2 in our revised manuscript to make it more clear to the readers as follows.
In Page 13, "It should be noted that if we look the chemical structure of the modified ditopic ligand TOsRuT closely, an asymmetric structure can be considered for this ligand due to presence of one Os(II) and one Ru(II) complex into the modified ditopic ligand. Therefore, the reactivity of two tpy units at the two ends of TOsRuT may differ little bit due to presence of two heterometal ions. In other way, the TOsRuT can be imagined as a structure with one side as head and another side as tail. Thus, when TOsRuT reacts with Fe(II) to make the polymer, it could be anticipated that the resultant polymer (polyOsRuFe) may be formed in regular head-to-tail structure (Fig. 2a) or regular structure with some irregular head-to-head/tail-to-tail structure (Fig. 2b). That means the sequence of Fe(II)/Ru(II)/Os (II) in polyOsRuFe may be repeated in alternate fashion or the sequence may be repeated with small irregularity. However, at present we did not find any experimental evidence to support this kind of assumption (theoretical prediction). More investigation in this direction is currently underway." Comment 2: As to Fig. 1

b, I was not expecting to see a plateau reached at ca. 1 equivalent of Fe(II)
ions added, as the compound TOsRuT has two binding sites which are quite far apart and probably independent.
Reply: Thank you for your concern about this matter. The plateau reached to not exactly 1 but it close to 1. Actually, for metallo-supramolecular polymers, when ditopic ligand is used for titration experiment (study of complexation behavior) with divalent metal ion, such king of plateau appears (close to equivalent point). So, we think, in our case the same thing happened. As Fe(II) reacts at room temperature, so probability of binding of Fe(II) at two sides of TOsRuT is more to build a linear polymeric structure. Another thing is that as the both Os and Ru forms nonlabile complexes, the reactivity of terpyridines at two ends of TOsRuT will not differ so much we think.
During study of complexation behavior of TOsRuT, we successively added Fe(BF4)2 to the solution of TOsRuT and recorded absorption spectra of the solution. Upon gradual addition of Fe(BF4)2, the solution of TOsRuT got saturated and the polymer was precipitated. We recorded the absorption spectra of the solution mixture until it forms the precipitate. The last point (and the absorption spectrum) shown in Fig. 1b and 1c was recorded before the precipitation of the solution mixture. From the titration and plot of [Fe(BF4)2]/[TOsRuT] ratio, we estimated the molar ratio of the reactants is close to 1.

Comment 3:
The last concern is related to the solubility of the final coordination polymers.

Is there any other way to confirm the presence in solution of such high MW polymeric species?
On the same note, Fig. 3c shows a peak at 575 nm for the absorption of the polymeric species, which is however in line with what seen in Fig. 1 In summary, I would be able to recommend this manuscript for publication in Communications Chemistry only after the authors have carefully responded to the points raised above.
Reply: Yes, the final polymer is soluble in DMSO and DMF. However, by varying counteranions, we can tune its solubility in low boiling solvents for various application purpose.
Sometime MALDI mass can be employed. We also tried that, but no fruitful result was obtained. That is way, we determined by RALLS.
Actually, the polymer is not soluble in MeOH. During titration (Fig. 1), as we mentioned above, when the solution mixture reached to saturation, the polymer is precipitated (not soluble in MeOH or MeOH/DCM mixture).
Thank you for your suggestion. We have carefully checked the English language in our whole revised manuscript and SI.
Finally, thank you very much once again for taking precious time from your work and helping us in improving this manuscript. We have tried our best to address every single concern that you have raised.
We sincerely hope that current fully revised manuscript would meet all your concerns. Reply: We are very thankful to the reviewer-3 for reviewing and spending your valuable time with our manuscript. We are also grateful for your excellent remark about our manuscript. Thank you again for your recommendation to accept our manuscript after some minor revision and improvement. We have revised our manuscript according to your valuable suggestions. Please check the revised manuscript. Reply: Thank you so much for bringing this point to our notice. We have corrected the tables in our revised manuscript. Please see the revised manuscript. Table 4 has been changed to Table 5 in our revised manuscript for revision purpose.

Comment 2: The cyclic voltammograms included in the SI all obtain different current densities
(sometimes double), where the peak height is related to the amount of electrochemically accessible material. For example, SI Fig. 21, 30, 31, and 32. The authors should briefly discuss the discrepancies with the current densities between the different CVs. The result might be due to these two factors: human error in the drop casting volume, or a result caused by the diffusion of the counter ions to and from the working electrode. Because the CVs in question involve different counter anions, the latter is a strong possibility. Additionally, the individual peak height ratios should be considered between metal to metal.
For example, In SI Fig. 30a and 32a, there is an approximately 1:1 ratio between Os to Fe, but more current at the Ru oxidative potential. In SI Fig. 31  And the following explanation was added in the revised manuscript. In Page 15, "The scanrate-dependent (0.01−0.1 V/s) CV study of polyOsRuFe displayed linear proportionality of the peak current with the scan rate ( Fig. 3d and Supplementary Fig. 21b), suggesting a surface-confined electrochemical redox process which is not restricted by slow electrolyte diffusion. 10 "     TOsRuT can also bind other divalent metal ion, we extended the metal to Co(II) and Zn(II). That was our main aim.
In addition, we think this Zn (II) containing polymer (polyRuOsZn) may be interested in researchers who deal with emission property of Os/Ru/Zn system. Finally, thank you very much once again for taking precious time from your work and helping us in improving this manuscript. We have tried our best to address every single concern that you have raised.
We sincerely hope that current fully revised manuscript would meet all your concerns. My only concern left is still related to the claim of having attained control over the sequence of metals within the polymeric structure. I suggest toning down this claim (see abstract, intro and conclusion) for in my opinion the authors do not present any definite proof of being able to do that. After this only modification is implemented, I would be glad to recommend the manuscript for publication in Communications Chemistry.
The reason for my comment above is briefly the following: To my understanding, in the case that, as it is suggested, "the reactivity of two tpy units at the two ends of TOsRuT may differ little bit due to presence of two heterometal ions" (pag.13), it would be important to define that "little bit", otherwise the most probable situation is that of a complete random sequence, and thus no control over the assembly. If the two sites act very differently instead (say the Ru side binds Fe a lot better), then a controlled sequence could be within reach (first step of the assembly process would be the formation of OsRuFeRuOs "dimers" which then would lead to an ordered (OsRuFeRuOs-Fe-OsRuFeRuOs)n polymeric sequence. Although the control of the sequence might be not so relevant for the final properties of the material (if the communication between neighboring centres is not very effective) it indeed remains an interesting point in terms of understanding the system and future designs. I am looking forward to reading more about developments on that in the authors' future investigations.
Reviewer #3 (Remarks to the Author): The authors carefully addressed all comments raized by me and all other referees. I would suggest publishing this manuscript in its current version.

Abstract:
We changed the words of "sequence control" in the 2 nd and 3 rd sentences of "Utilizing different 2) Instead of deleting the above sentences, we added the explanation of metallo-supramolecular polymers and our motivation to the research as follows.
"Metallo-supramolecular polymers (MSPs), which are synthesized by a 1:1 complexation of metal ion and ditopic ligand, have received attention due to a wide range of the applications including electrochromic displays, memory devices, sensors, energy storage devices, and anticancer therapies. [1][2][3][4][5][6][7] The electronic interaction between the metal and the ligand in the polymer chains causes unique electrochemical, optical, emissive properties unlike the conventional organic polymers. In general, one metal species is included in MSPs, but introduction of two metal ion species in the polymer chain is one of the recent hot topics, because the dual metal species controlled in the polymer are expected to expand the functions of MSPs. 2,[8][9][10][11][12][13][14][15][16][17] "

Main text (Page 13)
We revised the following sentences, "It should be noted that if we look the chemical structure of the modified ditopic ligand TOsRuT closely, an asymmetric structure can be considered for this ligand due to presence of one Os (II) and one Ru(II) complex into the modified ditopic ligand. Therefore, the reactivity of two tpy units at the two ends of TOsRuT may differ little bit due to presence of two heterometal ions. In other way, the TOsRuT can be imagined as a structure with one side as head and another side as tail.
Thus, when TOsRuT reacts with Fe(II) to make the polymer, it could be anticipated that the resultant polymer (polyOsRuFe) may be formed in regular head-to-tail structure (Fig. 2a) or regular structure with some irregular head-to-head/tail-to-tail structure (Fig. 2b). That means the sequence of Fe(II)/Ru(II)/Os(II) in polyOsRuFe may be repeated in alternate fashion or the sequence may be repeated with small irregularity." to "It should be noted that if we look the chemical structure of the modified ditopic ligand TOsRuT closely, an asymmetric structure can be considered for this ligand due to presence of one Os (II) and one Ru(II) complex into the modified ditopic ligand. If the reactivity of two tpy units at the two ends of TOsRuT is different due to presence of two heterometal ions, the TOsRuT can be imagined as a structure with one side as head and another side as tail. Thus, when TOsRuT reacts with Fe(II) to make the polymer, it could be anticipated that the resultant polymer (polyOsRuFe) may be formed in regular head-to-tail structure (Fig. 2a) or regular structure with some irregular head-to-head/tail-to-tail structure (Fig. 2b). For example, if the Ru side binds Fe a lot better than the Os side, the first step of the assembly process would be the formation of the OsRuFeRuOs dimers which then would lead to an ordered (OsRuFeRuOs-Fe-OsRuFeRuOs)n polymeric sequence. However, if the reactivity of two tpy units at the two ends of TOsRuT is same, the complexation with Fe would lead to polyOsRuFe with a complete random sequence."

Conclusion (the first sentence):
"We succeeded in sequence control of three different transition metals in metallo-supramolecular polymer for the first time and achieved the quad-color electrochromism." was changed to "We succeeded in stepwise introduction of three different transition metals in metallo-supramolecular polymer for the first time and achieved the quad-color electrochromism.".
Finally, thank you very much once again for taking precious time from your work and helping us in improving this manuscript. We have tried our best to address your concern. We sincerely hope that the current fully revised manuscript would meet your concern.
Our reply to Reviewer-3 Remark: The authors carefully addressed all comments raized by me and all other referees. I would suggest publishing this manuscript in its current version.
Our reply: Many thanks for your recommendation for publication.