Direct observation of photoinduced sequential spin transition in a halogen-bonded hybrid system by complementary ultrafast optical and electron probes

A detailed understanding of the ultrafast dynamics of halogen-bonded materials is desired for designing supramolecular materials and tuning various electronic properties by external stimuli. Here, a prototypical halogen-bonded multifunctional material containing spin crossover (SCO) cations and paramagnetic radical anions is studied as a model system of photo-switchable SCO hybrid systems using ultrafast electron diffraction and two complementary optical spectroscopic techniques. Our results reveal a sequential dynamics from SCO to radical dimer softening, uncovering a key transient intermediate state. In combination with quantum chemistry calculations, we demonstrate the presence of halogen bonds in the low- and high-temperature phases and propose their role during the photoinduced sequential dynamics, underscoring the significance of exploring ultrafast dynamics. Our research highlights the promising utility of halogen bonds in finely tuning functional properties across diverse photoactive multifunctional materials.

Since the initial submission the authors have substantially revised the paper and have clearly tried to take on board the queries of myself and the other reviewers.I am particularly pleased to see that they have undertaken an additional spectroscopic study on an ultrathin sample that strongly suggests that the sample retains its composition despite being so thinly sliced.
However, there are still a few points to be addressed.In their response the authors state that they have calculated the %HS to be 7.1%, how was this number calculated?
The authors also indicate that they've created a new movie called extended data movie 3, but I can only see two movies.Have the authors forgotten to upload it or is this the new black and white Laplacian movie?
Minor points 1. Lines 59-60, the ligand is negatively charged and as per IUPAC rules it should be 5-iodo-N-(8'quinolyl)-salicylaldiminate.This error is present throughout the manuscript.2. Line 124, DAS is used but not explained.3. Figure 1 remains more or less the same as before.Both myself and the other referee suggested that arrows or similar indicators be used to show the changes that occur upon photoexcitation.4. A constant frustration for me in the ESI is having to move backwards and forwards to find the figures referred to in the supplementary text.As this is the ESI, it would be far better to place the figures near the text where they are referred to. 5. Extended figure 3, the caption use the word 'blu', when it should be blue.6.In extended figure 12 it's very hard to read the labels for the atoms.I'd suggest that they are made larger or a different colour is used to make them more visible.This in fact applies to the new black and white movie they've created too.
Overall, I am happy with the changes made by the authors and I think the manuscript now more accurately reflects the results and clarifies the areas of confusion.Provided that they undertake these revisions I would be happy to recommend the paper for publication in Nature Communications.
Reviewer #2 (Remarks to the Author): The paper is a new version of the paper by Yifeng Jiang et al, previously submitted to , is about ultra-fast pump-probe study of photoinduced phenomena in a material made of a photoactive spin crossover molecular unit, connected through halogen-bonds to (dmit)2 units exhibiting strong or weak dimerization.The authors used ultrafast optical and infra-red spectroscopies and complementary ultrafast electron diffraction, to gain knowledge on ultrafast and out-of-equilibrium photoinduced dynamics.The experimental data reveal ultrafast changes within 500 fs and slower dynamics within 50 ps.In this new version of the paper, as well as Supp.
Info, the authors have considered the most important remarks and have improved the presentation of the results.The analysis of the results and the discussion are more relevant and rigorous in the present version, with several additional discussions about the method or the data analysis, and the associated limitations, such as about the role of the volume change and the halogen bonds.The extended data, especially the one about Halogen bond, make the discussion stronger.Now it is clear why the two different timescales can be associated with two different dynamics and how they couple.This work is very promising for exploring enhanced control of multifunctional through halogen bond and it can attract a broad community of researchers.
The reviewer recommends publication of the paper in its present form, considering the minor points below: 1-It is indicated line 111 "Fig.2b shows an approximate match of all the transient spectra of after +1 ps time delay with the characteristic of the formation of the HS state from the temperaturedependent measurement."However, few lines above it is indicated "OD of the HT phase is lower than that of the LT phase at wavelengths shorter than λ0 = 530 nm and higher at wavelengths longer than λ0 in the visible region" Therefore, it seems from Fig2b that it is only after 100 ps that OD transient spectra represent an approximate match characteristic of the formation of the HS state."I also found a typo: "transient spectra of after +1 ps time delay" ◊ "transient spectra after +1 ps time delay" Reviewer #2 (Remarks on code availability): Dear editor, I'm sorry, I am not able to review the code.I have no electron data to run or test.
points raised by the referees, we, coauthors, believe all problems pointed out have been 4 addressed, and we apologize for the typo and inconsistence in our manuscript.

5
During considering the response to referees' comments, we noticed there was typos in Fig. 4f 6 and the labels are inconsistent with the main text.We corrected the typos, from ξ to p. Due to 7 this correction, inconsistency between main text and Fig. 4f disappeared.to see that they have undertaken an additional spectroscopic study on an ultrathin sample that 17 strongly suggests that the sample retains its composition despite being so thinly sliced.

18
However, there are still a few points to be addressed.In their response the authors state that 19 they have calculated the %HS to be 7.1%, how was this number calculated?

20
The %HS mentioned here is the excitation fraction (fractional population of the excited HS 21 state).As described in Lines 288-292 of Method in the main manuscript, the excitation fraction 22 in this work was calculated by two complementary methods.First, the excitation fraction was 23 calculated from crystal structure and optical absorption at 400 nm.Then, the excitation fraction However, few lines above it is indicated "OD of the HT phase is lower than that of the LT phase 96 at wavelengths shorter than λ0 = 530 nm and higher at wavelengths longer than λ0 in the visible 97 region" Therefore, it seems from Fig2b that it is only after 100 ps that OD transient spectra We thank the referee for pointing out this.We agree that the OD transient spectra after +100 ps 103 has a better approximate match of the characteristics of the formation of the HT phase spectra.

104
We noticed our description in line 111 had been a misleading one.We changed "+1 ps" to "+100 105 ps" as suggested by referee.In addition, we changed "formation of the HS state" to "HT phase".

106
As written in the next paragraph, we assigned HS state formation as a very fast dynamics with 107 the time constant of 0.16 ps.After +100 ps, the system is after the sequential active of the dimer 108 softening process and it is in a more equilibrium state like the thermally-induced phase Jiang, Takahashi, Ishikawa, Miller and co-workers explore an iron(III) SCO 11 material where dimerization of the [Ni(dmit)2]-unit is coupled to spin crossover and mediated 12 by halogen bonding.The authors explores this using an ultra-fast pump-probe system to explore 13 the changes that occur in this multifunctional iron(III) SCO material.14Sincethe initial submission the authors have substantially revised the paper and have clearly 15 tried to take on board the queries of myself and the other reviewers.I am particularly pleased 16 24 was also calculated from the change of the Bragg peak intensities at a long-time delay and the 25 changes of the Bragg peak intensities during thermally induced phase transition.26 The authors also indicate that they've created a new movie called extended data movie 3, but I 27 can only see two movies.Have the authors forgotten to upload it or is this the new black and 28 white Laplacian movie?29 Yes, the new black and white Laplacian movie has been labelled as Extend Data Mov.3 in the halogen bonds.The extended data, especially the one about Halogen bond, make the discussion 86 stronger.Now it is clear why the two different timescales can be associated with two different 87 dynamics and how they couple.This work is very promising for exploring enhanced control of 88 multifunctional through halogen bond and it can attract a broad community of researchers.89 90 The reviewer recommends publication of the paper in its present form, considering the minor 91 points below: 92 1-It is indicated line 111 "Fig.2b shows an approximate match of all the transient spectra of 93 after +1 ps time delay with the characteristic of the formation of the HS state from the 94 temperature-dependent measurement."95 98represent an approximate match characteristic of the formation of the HS state."99 I also found a typo: "transient spectra of after +1 ps time delay" 100 "transient spectra after +1 ps time delay" 101 102 109 transition.As shown in Extend data Fig 5 and Extended Data Fig. 6, the main difference 110 between transient spectra after +100 ps and spectra after +1 ps and +5 ps is two additional