Strong and selective isotope effect in the vacuum ultraviolet photodissociation branching ratios of carbon monoxide

Rare isotope (13C, 17O and 18O) substitutions can substantially change absorption line positions, oscillator strengths and photodissociation rates of carbon monoxide (CO) in the vacuum ultraviolet (VUV) region, which has been well accounted for in recent photochemical models for understanding the large isotopic fractionation effects that are apparent in carbon and oxygen in the solar system and molecular clouds. Here, we demonstrate a strong isotope effect associated with the VUV photodissociation of CO by measuring the branching ratios of 12C16O and 13C16O in the Rydberg 4p(2), 5p(0) and 5s(0) complex region. The measurements show that the quantum yields of electronically excited C atoms in the photodissociation of 13C16O are dramatically different from those of 12C16O, revealing strong isotope effect. This isotope effect strongly depends on specific quantum states of CO being excited, which implies that such effect must be considered in the photochemical models on a state by state basis.

The measurements presented seem to be of good quality. However, they of a small subset of CO bands in a moderately congested region of the CO spectrum. Since C(1D) formation is energetically allowed up to ~ 100 nm, it would make vastly more sense to present experimental results for isotopic branching ratios for some of the bands between 94 and 100 nm. It is impossible to draw conclusions about the importance of the 1.2 nm region they have studied for the simple reason that ii contributes only a small fraction to the total rate of CO photodissociation.
Second, the rather extensive discussion of O isotopes fractionation, and the experiments of the Thiemens group, is completely unsupported by the data presented. If the authors wish to make a meaningful contribution to the understanding O isotope fractionation in CO dissociation, then they should perform experiments using C16O, C17O, and C18O. The discussion of O isotopes they have presented is merely conjecture, and does not acknowledge the role of other factors such as linewidth for the short wavelength CO bands. Line overlap has an important effect on O isotope fractionation.
For the above two reasons I cannot recommend this paper for publication in its present form.
Additional comments: 1. The excited state C'1Sigma+(v'=7) is referred to repeatedly throughout the text. Is this a new state assignment (and if so please give the citation)? Usually this state is simply the 1Sigma+(v'=2) state. Table 1, however there is no discussion whatsoever about uncertainties in the text. The authors should include a brief description of the uncertainty levels in the text itself and how they came to determine those uncertainties.

Uncertainty estimates are included in the results presented in
5. The grammar in the manuscript needs a careful review by the authors and/or the editors of the journal. Multiple awkward phrasings detract from the quality of the science being presented.

Regards, Glenn Stark
Reviewer #3 (Remarks to the Author): The authors report on a compelling experimental study that characterizes a strong and selective isotopic effect of photodissociation branching ratios of carbon monoxide in the VUV absorption range. They use VUV excitation generated by two photon resonant enhanced four wave mixing coupled to a time-slice velocity-map ion imaging apparatus (TSVMI) to measure the branching ratios in the Rydberg 4p(2), 5p(0) and 5s(0) complex region for three indirect predissociative pathways for the two isotopomers : 12C16O or 13C16O : one leading to the atoms in their ground triplet (3P)states, one leading to C(1D) + O(3P) and one leading to C(3P) and O(1D). They report on a significant isotopic effect for the predissociation channel leading to C(1D) + O(3P) for the absorption bands 1Π(v´=2) and C´ 1Σ+(v´=7). For these bands, 12C16O primarily dissociates in the lowest channel while 13C16O dissociates to the next channel producing to C(1D) atoms. Significant variations of the branching ratios are also reported for the other absorption bands in this spectral region. To explain the strong experimental effect, the authors adopt the explanation proposed in ref 34 : due to the mass effect, different vibronic states are excited in the two isotopomers : in the case when the predissociation process is indirect and involves an electronic coupling between a Rydberg and a valence state, the ladder of rovibrational states in the two electronic states are shifted by the change in reduced mass, more in the case of the shallow valence state, which leads to different efficiencies of the electronic coupling for the two isotopomers due to the fact that accidental resonances occur at slightly different energies in the two isotopomers. As noted already in ref 34 and reinstated by the authors, this mechanism is general. In this sense, I recommend that the authors tone down the qualification 'new', 'novel', 'for the first time', that they use in several places in the manuscript. It is very important to document experimental the ubiquity of the mechanism proposed in ref 34 for N2 in case of indirect predissociation for other diatomic molecules but it can no longer be qualified to be novel. The authors might want to consider to also refer to more recent work of the authors of ref 34, for example J. S. Ajay, K. G. Komarova, F. Remacle and R. D. Levine, Proceedings of the National Academy of Sciences, 2018, 115, 5890.
The paper is clearly written and reports on important experimental results that will stimulate more theoretical work. I recommend its publication in nature communication.
First, we really want to thank all the three reviewers for carefully reading our manuscript, and give us many valuable and important suggestions in such short time. We have carefully read all the comments, and revised the manuscript accordingly. The detailed replies point-by-point are listed as follows:

Reviewer #1 (Remarks to the Author):
The authors present measured branching ratios for CO dissociation at VUV wavelengths (92.8 to 94.0 nm). They demonstrate preferential formation of 13C(1D) for a couple of the bands in this region, and argue that such a preference for 13C formation may be important for understanding the difference in the 12C/13C ratio for the Sun versus planetary materials. They apply a simple diabatic model to explain the preferential formation of 13C(1D) in their experiments. They then expand their diabatic arguments to O isotopes branching ratios in CO dissociation.

Our reply:
The reviewer is right about the logic that we are using in the manuscript, that we measured several typical absorption bands of 13 C 16 O in a small energy range, and found that the branching ratios are very different from that of 12 C 16 O, then we argue that the underlying mechanism can be general for other bands of 13 C 16 O that are not measured yet, and also for 12 C 17 O and 12 C 18 O. We believe that this logic generally makes sense, which I hope the reviewer can also agree with.
We clearly noted that this isotope effect on the photodissociation branching ratio is very random, it could produce more C( 1 D), it could also produce less C( 1 D), depending on the specific absorption band, this has been clearly shown in the manuscript. We are also clear that the overall effect of this isotope effect cannot be quantified without measuring all the relevant absorption bands and quantitative modeling, thus we did not make any conclusions anywhere in the manuscript about if this observation will prefer or not prefer the overall 13 C( 1 D) formation. The only conclusions that we have emphasized are: isotope substitution can strongly affect the photodissociation branching ratios, and this isotope effect depends on specific absorption band, and it need be considered in photochemical models on a state-by-state basis. To this point, we think the amount of experiment that we present in the manuscript is enough to support the conclusions that we made.
The measurements presented seem to be of good quality. However, they of a small subset of CO bands in a moderately congested region of the CO spectrum. Since C(1D) formation is energetically allowed up to ~ 100 nm, it would make vastly more sense to present experimental results for isotopic branching ratios for some of the bands between 94 and 100 nm. It is impossible to draw conclusions about the importance of the 1.2 nm region they have studied for the simple reason that ii contributes only a small fraction to the total rate of CO photodissociation.

Our reply:
The reviewer is correct that we have only measured a very small range for 13 C 16 O, while as I said above, we are not trying to make any "real final conclusions" about the possible impacts on the astrophysical models, because to do this, we need finish measuring all the ~ 30 absorption bands in the range 94-100 nm, and also higher energy range 91.1-92.8 nm, and then perform a quantitative modeling considering the self-shielding effect and also other isotopic effects, for example the isotope dependent dissociation rates, like that has been done by Lyons in Ref. 25. This is very large amount of experimental and modeling work, which we think is beyond the scope of a single communication paper like this.
We are now on the way to measure the branching ratios of all the strong absorption bands of 13 C 16 O, 12 C 17 O and 12 C 18 O step by step, while this will take very long time. Since the submission of this manuscript, we have now measured the branching ratios of 13 C 16 O for most of the strong absorption bands in the range 94-100nm (not completed yet), and have obtained some preliminary data. The branching ratios for higher energy range 91.1-92.8 nm have also been planned, but not measured yet. The current preliminary data in the range 94-100nm we obtained showed that several absorption bands have enhanced C( 1 D) productions by 20%-30%, like W(3sσ) 1 Π(v´=2), L(4pπ) 1 Π(v´=0) and K(4pσ) 1 Π(v´=0); there are also bands with reduced C( 1 D) productions, like W(3sσ) 1 Π(v´=1), by about 15%-20%. Even though we have not completed the measurements in the 94-100nm range so far yet, this observation already confirms the results and conclusions we made in the present manuscript. We hope that the reviewer can agree with us on not putting these new data which are still very preliminary into the present manuscript, because it will make our data set very fragmented, and result in major re-organization of the manuscript, and to our opinion this effort will not make the present manuscript stronger, as it is still not enough to make any real conclusions about the overall production rates of 13 C( 1 D) without measurements in the range 91.1-92.8 nm and any quantitative modeling by considering the self-shielding effect and the dissociation cross sections of all these absorption bands. We would prefer to presenting the new data in 94-100 nm wavelength range in subsequent publications until we finish all the data analysis.
To make things more clear, we have added one more sentence in the second paragraph of the section "Possible impacts on photochemical models" as "Although without systematic branching ratio measurements like that in the present study for all the relevant 13 C 16 O absorption bands and quantitative photochemical modeling as that by Lyons and coworkers 25 , it is not possible to predict how the current observation could affect the modeling outcomes". We have also changed the phrase "definitely necessary" to "could improve", to tone down the possible importance of the branching ratio measurement.
Second, the rather extensive discussion of O isotopes fractionation, and the experiments of the Thiemens group, is completely unsupported by the data presented. If the authors wish to make a meaningful contribution to the understanding O isotope fractionation in CO dissociation, then they should perform experiments using C16O, C17O, and C18O. The discussion of O isotopes they have presented is merely conjecture, and does not acknowledge the role of other factors such as linewidth for the short wavelength CO bands. Line overlap has an important effect on O isotope fractionation.
Our reply: In fact, we are initially motivated by the O isotopic fractionation effects in the Solar system and wanted to measure the branching ratios for 12 C 18 O at the beginning, while the sample of enriched 12 C 18 O is super expensive ( 12 C 17 O is even more expensive) and not easy to obtain in China, then we decide to measure the branching ratios of 13 C 16 O which is much cheaper, to see if there are any isotopic effects. It is also because that we think the C and O problems associated with the photodissociation of CO are closely related to each other, and the current measurements on 13 C 16 O could strongly imply that 12 C 17 O and 12 C 18 O can also have different branching ratios compared with that of 12 C 16 O. We are further motivated by this experiment, and have already ordered a bottle of 12 C 18 O from USA which is already on the way. In the introduction section, we mentioned both C and O for generally emphasizing the importance of carefully measuring the isotope dependent dissociation branching ratios, which is also the initial motivation of the whole project which is being supported by Chinese National Science Foundation. We agree with the reviewer that self-shielding effect still plays the major role on O isotope fractionation process, and other factors can affect the results, and the idea we want to present here is that the branching ratio may be also one of these other factors, and this is our motivation for measuring the branching ratio of 12 C 17 O and 12 C 18 O in the next step. To make this clear, we have rewritten the first paragraph of the section "Possible impacts on the photochemical models", please refer to the revised manuscript for details. We have emphasized the importance of the self-shielding effect and other isotopic effects as requested by the reviewer, and hope that this can be accepted by the reviewer.
For the above two reasons I cannot recommend this paper for publication in its present form.
Additional comments: 1. The excited state C'1Sigma+(v'=7) is referred to repeatedly throughout the text. Is this a new state assignment (and if so please give the citation)? Usually this state is simply the 1Sigma+(v'=2) state.
Our reply: This state was assigned as C'1Sigma+(v'=7) in Ref. 8. We have added a footnote under Table I to make it clear.
2. The potential energy curves in Figure 1b are not really useful in this cartoon format. A more accurate and properly labeled set of PECs is needed.
Our reply: There are two reasons that we use a simple carton picture for illustrating the dissociation mechanisms here. First, accurate PECs for CO which include the information of mutual coupling strengths among different electronic states in such high energy region are still not available, thus it is not really possible to specifically label all the interacting potential curves; second, Nature Communications are dedicated to very broad scopes of readers from different areas, a simple carton picture which is illustrative for readers who are not familiar with molecular photodissociation would be very helpful. I hope the reviewer can agree on this.

Reviewer #2 (Remarks to the Author):
I recommend the paper for publication with only minor revisions.
The new results -concerning the strong isotope-dependence of photodissociation branching ratios in carbon monoxide -are relevant to ongoing studies of predissociation mechanisms in CO. In addition to the intrinsic merit of a more detailed understanding of the interactions of high-lying electronic states in CO, there are important astrophysical applications that rely on a quantitative understanding of predissociation in CO and its isotopologues. This paper highlights the previously under-appreciated role of branching ratio differences among the CO isotopologues as a possible explanation for the strong fractionation signatures of carbon and oxygen isotopes in the solar system. The qualitative analysis presented in the paper -a coupled-channel model of diabatic states leading to indirect predissociation -is a sensible explanation for the trends seen in the experimental results. The experimental methods are state-of-the-art and the level of experimental detail presented is appropriate.
Our reply: Thank you for the very positive comment on our work! Below I have indicated some suggested minor revisions, but there is no need for further review.
1. page 1, title: please replace "on the vacuum…" with "in the vacuum…" Our reply: Fixed.
2. page 2, line 41: in the introductory discussion, the term "To test the self-shielding models" will be confusing to anyone who is not intimately familiar with the works being cited. The authors should define this term -a single sentence would help to clarify the introductory remarks.
Our reply: We have added the follow two sentences to the introduction section: Self-shielding process happens when a light beam with broad spectral distribution penetrates a sample of mixed gaseous molecules which have slightly different absorption line positions. Due to the different column densities of the gaseous components, which result in different attenuation speeds according to Beer's law, the relative amount of light absorption for different molecular species will change along the light propagation direction.
3. page 7, line 200: the phrase "the substitution of 12C by 13C must have dramatically changed the relative coupling strengths among different electronic states" should be re-phrased. As the authors later mention, the coupling between electronic states is independent of isotope. It is the coupling between individual rovibronic levels that depends on isotope (not the coupling between electronic states) -because of the energy shifts that the authors describe.
Our reply: We have changed "among different electronic states" to "among different rovibronic states".
4. Uncertainty estimates are included in the results presented in Table 1, however there is no discussion whatsoever about uncertainties in the text. The authors should include a brief description of the uncertainty levels in the text itself and how they came to determine those uncertainties.
Our reply: A footnote for describing the error bars is added to Table I. 5. The grammar in the manuscript needs a careful review by the authors and/or the editors of the journal. Multiple awkward phrasings detract from the quality of the science being presented.
Our reply: We have reviewed our manuscript for many times, and have tried our best to make all grammars correct before we submit it. We will do this again.

Regards, Glenn Stark
Reviewer #3 (Remarks to the Author): The authors report on a compelling experimental study that characterizes a strong and selective isotopic effect of photodissociation branching ratios of carbon monoxide in the VUV absorption range. They use VUV excitation generated by two photon resonant enhanced four wave mixing coupled to a time-slice velocity-map ion imaging apparatus (TSVMI) to measure the branching ratios in the Rydberg 4p (2), 5p(0) and 5s(0) complex region for three indirect predissociative pathways for the two isotopomers : 12C16O or 13C16O : one leading to the atoms in their ground triplet (3P)states, one leading to C(1D) + O(3P) and one leading to C(3P) and O(1D). They report on a significant isotopic effect for the predissociation channel leading to C(1D) + O(3P) for the absorption bands 1Π(v´=2) and C´ 1Σ+(v´=7). For these bands, 12C16O primarily dissociates in the lowest channel while 13C16O dissociates to the next channel producing to C(1D) atoms. Significant variations of the branching ratios are also reported for the other absorption bands in this spectral region. To explain the strong experimental effect, the authors adopt the explanation proposed in ref 34 : due to the mass effect, different vibronic states are excited in the two isotopomers : in the case when the predissociation process is indirect and involves an electronic coupling between a Rydberg and a valence state, the ladder of rovibrational states in the two electronic states are shifted by the change in reduced mass, more in the case of the shallow valence state, which leads to different efficiencies of the electronic coupling for the two isotopomers due to the fact that accidental resonances occur at slightly different energies in the two isotopomers. As noted already in ref 34 and reinstated by the authors, this mechanism is general. In this sense, I recommend that the authors tone down the qualification 'new', 'novel', 'for the first time', that they use in several places in the manuscript. It is very important to document experimental the ubiquity of the mechanism proposed in ref 34 for N2 in case of indirect predissociation for other diatomic molecules but it can no longer be qualified to be novel.