Chinese Journal of Chemical Physics  2020, Vol. 33 Issue (1): 91-100

#### The article information

Hong Gao, Yu Song, William M. Jackson, Cheuk-Yiu Ng

Photodissociation Branching Ratios of $^{\textbf{12}}$C$^{\textbf{16}}$O from 108000 cm$^{\textbf{-1}}$ to 113200 cm$^{\textbf{-1}}$ Measured by Two-Color VUV-VUV Laser Pump-Probe Time-Slice Velocity-Map Ion Imaging Method: Observation of Channels for Producing O($^\textbf{1}$D)

Chinese Journal of Chemical Physics, 2020, 33(1): 91-100

http://dx.doi.org/10.1063/1674-0068/cjcp1911199

### Article history

Received on: November 12, 2019
Accepted on: December 3, 2019
Photodissociation Branching Ratios of $^{\textbf{12}}$C$^{\textbf{16}}$O from 108000 cm$^{\textbf{-1}}$ to 113200 cm$^{\textbf{-1}}$ Measured by Two-Color VUV-VUV Laser Pump-Probe Time-Slice Velocity-Map Ion Imaging Method: Observation of Channels for Producing O($^\textbf{1}$D)
Hong Gaoa,b , Yu Songb , William M. Jacksonb , Cheuk-Yiu Ngb
Dated: Received on November 12, 2019; Accepted on December 3, 2019
a. Beijing National Laboratory for Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190, China;
b. Department of Chemistry, University of California, Davis CA 95616, USA
Abstract: The photoabsorption and photodissociation of carbon monoxide (CO) in the vacuum ultraviolet (VUV) region is one of the most important photochemical processes in the interstellar medium, thus it has attracted numerous experimental and theoretical studies. Here, we employed the two-color VUV-VUV laser pump-probe time-slice velocity-map ion imaging method to measure the relative branching ratios [C($^3$P$_0$)+O($^1$D)]{[C($^3$P$_0$)+O($^3$P)]+ [C($^3$P$_0$)+O($^1$D)]} and [C($^3$P$_2$)+O($^1$D)]{[C($^3$P$_2$)+O($^3$P)]+[C($^3$P$_2$)+O($^1$D)]} in the VUV photoexcitation energy range of 108000$-$113200 cm$^{-1}$. Here, one tunable VUV laser beam is used to excite CO to specific rovibronic states, and a second independently tunable VUV laser beam is used to state-selectively ionize C($^3$P$_0$) and C($^3$P$_2$) for detection. State-selective photoionization through the 1VUV+1UV/visible resonance-enhanced multiphoton ionization scheme has greatly enhanced the detection sensitivity, which makes many new weak absorption bands observable in the current study. The branching ratio measurement shows that the spin-forbidden channels C($^3$P$_0$)+O($^1$D) and C($^3$P$_2$)+O($^1$D) only open at several discrete narrow energy windows. This might be caused by certain accidental resonance-enhanced spin-orbit interactions between the directly excited Rydberg states and valence states of triplet type which finally dissociate into the spin-forbidden channels.
Key words: Carbon monoxide    Photodissociation    Vacuum ultraviolet    Branching ratio
Ⅰ. INTRODUCTION

The photochemistry, spectroscopy and indirect predissociation dynamics of CO has been studied intensively [1-3] because it has important practical applications in the fields of astrophysics and astrochemistry [4-6], yet there still remains many things that are not known and are not completely understood. There have been numerous spectroscopic studies on all the CO isotopologues using various experimental methods [7-11]. Direct detection of the photodissociation fragments of $^{12}$C$^{16}$O has been reported, and it was found that $^{12}$C$^{16}$O can dissociate into the triplet channel C($^1$D)+O($^3$P) with rates comparable with that into the lowest dissociation channel C($^3$P)+O($^3$P) at certain absorption bands [12, 13]. Accompanying these experimental efforts has been a great deal of theoretical work done to understand the complicated spectroscopic structures of CO and the predissociation dynamics that involves numerous Rydberg-Rydberg, Rydberg-valence interactions [14-19]. Despite this large amount of experimental and theoretical work, the detailed predissociation dynamics and the complicated coupling schemes between electronic states of different symmetries and different bonding types are still not completely understood, especially in the high energy region where there is a high density of quantum states [18, 20]. To completely unravel the predissociation dynamics of all the CO isotopologues, further efforts from both experimentalists and theoreticians are needed.

We have recently combined a time-slice velocity-map ion imaging (TS-VMI) setup with a table-top tunable vacuum ultraviolet (VUV) laser radiation source generated by the two-photon resonance-enhanced four-wave mixing (TPRE-FWM) scheme to measure the photodissociation branching ratios of $^{12}$C$^{16}$O among the channels C($^3$P)+O($^3$P), C($^1$D)+O($^3$P) and C($^3$P)+O($^1$D) [21-25]. These measurements showed that substantial amounts of excited carbon [C($^1$D)] and oxygen [O($^1$D)] atoms are produced in the photodissociation process, that depend on the specific rovibronic states of $^{12}$C$^{16}$O that are excited. The branching ratios provide valuable information for understanding the predissociation dynamics of CO that is not available from any pure spectroscopic studies [17, 26]. At the same time, knowing the amount of excited atomic products that are generated in the photodissociation of $^{12}$C$^{16}$O are important to our understanding of the photochemical processes that occurred during the formation of the solar system [27, 28]. To measure the photodissociation properties of small molecules like CO in more detail, we have recently built a second independently tunable VUV laser radiation source, which can be used to state-selectively photoionize the atomic photofragments. The state-selective photoionization method not only provides more detailed studies of photodissociation dynamics, but also greatly enhances the detection sensitivities. With this two-color VUV-VUV laser pump-probe time-slice velocity-map ion imaging setup, we have studied the state-to-state photodissociation dynamics of several important small molecules, like CO [29], N$_2$ [30, 31] and CO$_2$ [32-34]. Here, we employ this same method for systematically measuring the photodissociation branching ratios of $^{12}$C$^{16}$O from 108000 cm$^{-1}$ to 113200 cm$^{-1}$, which is close to its ionization threshold. In the current experiment, one tunable VUV beam is used to excite the $^{12}$C$^{16}$O molecules to a specific rovibronic state that then predissociates; a second independently tunable VUV beam is used for state-selectively ionizing C($^3$P$_2$) or C($^3$P$_0$) in a 1VUV+1UV/visible resonance-enhanced multiphoton ionization (REMPI) scheme. This greatly enhances the detection efficiency so that numerous rovibronic transitions can be identified in the C atom photofragment spectra, with some of them observed for the first time. The time-slice velocity-map ion imaging method is used to measure the branching ratios [C($^3$P$_0$)+O($^1$D)]{[C($^3$P$_0$)+O($^3$P)]+[C($^3$P$_0$)+O($^1$D)]} and [C($^3$P$_2$)+O($^1$D)]{[C($^3$P$_2$)+O($^3$P)]+[C($^3$P$_2$)+ O($^1$D)]} at each of the rovibronic peaks observed in the photofragment spectra. By plotting the branching ratios versus the VUV photoexcitation energies over the studied energy range, it can be seen that the spin-forbidden channels that produce excited oxygen atoms [O($^1$D)] are only open at several narrow energy windows and outside of these windows the percentages are less than 5%. This provides hints of where on the potential energy curves of CO the spin-orbit coupling is strong. This is not only important for understanding the predissociation dynamics, but also for its applications in the fields of astrophysics and astrochemistry [27].

Ⅱ. EXPERIMENTS METHODS

A detailed description of the experimental setup has been presented previously [29, 30, 35], thus only experimental details that are relevant to this study will be mentioned here. The setup consists of a TS-VMI section and two independently tunable VUV laser radiation sources generated by the TPRE-FWM method. To generate the VUV pump beam in the photon energy range 108000-113200 cm$^{-1}$, the two-photon resonant transition of Kr at the UV wavelength 212.556 nm (2$h\nu $$_{ \rm{UV}} = 94092.8 cm ^{-1} ) is used in the TPRE-FWM process. The visible laser radiation is tuned in the wavelength range 523-724 nm. The UV and visible laser beams are generated by two dye lasers which are pumped by a single Nd:YAG laser. The pump VUV (the sum-frequency component) beam enters the photodissociation and photoionization (PD/PI) region and crosses the supersonically expanded ^{12} C ^{16} O beam perpendicularly. There ^{12} C ^{16} O molecule is excited to a specific excited rovibronic state. To state-selectively ionize C( ^3 P _0 ) and C( ^3 P _2 ) for detection, we first excite them to the intermediate Rydberg states 2s ^2 2p3d ^3 D _1 and 2s ^2 2p3d ^3 D _3 by absorbing a single VUV photon (the probe VUV) at 78293.5 cm ^{-1} and 78274.9 cm ^{-1} , respectively; the C atoms in the intermediate Rydberg states can be ionized by absorbing a second UV or visible photon which are co-propagating with the probe VUV beam. The probe VUV beam (the difference-frequency component) counter propagates to the pump VUV beam. It is also generated by using the two-photon resonant transition of Kr at the UV wavelength 212.556 nm, and the corresponding visible laser is at \sim 633 nm. To eliminate the possible unnecessary signal generated by the sum-frequency component from the probe VUV beam, a MgF _2 window is inserted between the PD/PI region and the four-wave mixing chamber of the probe beam, which blocks the sum-frequency component. We optimize the relative intensities and delays between the pump and probe laser beams, and also any relevant parameters of the ion detector and signal amplifier, so that the ion signal generated by either individual VUV beam is negligible. The C ^+ ions formed in the PD/PI region are accelerated by a set of electrostatic ion lens and fly to a 75 mm diameter micro-channel plate (MCP) detector. There are two different data acquisition modes in our case, namely the spectroscopy mode and TS-VMI mode. In the spectroscopy mode, we fix the probe VUV beam to state-selectively ionize C( ^3 P _0 ) and C( ^3 P _2 ) for detection, and simultaneously scan the pump VUV beam in the range 108000-113200 cm ^{-1} , the time-of-flight (TOF) signal from the MCP goes to the input of a Stanford Research Systems (SRS) Boxcar data acquisition system which is gated so that only the C ^+ ion is integrated. In this mode, the C ^+ photofragment spectrum of ^{12} C ^{16} O is obtained, which help us identify the rovibronic states of ^{12} C ^{16} O that are excited by the pump VUV beam. In the TS-VMI mode, the pump VUV beam is fixed at a specific rovibronic state of ^{12} C ^{16} O identified in the C ^+ photofragment spectra, and the probe VUV beam is scanned back and forth to cover the whole Doppler profile caused by the recoil velocities of the C photofragments. The kinetic energy distribution of the C ^+ ions can be measured, from which the branching ratios of different dissociation channels can be obtained. Ⅲ. EXPERIMENTAL RESULTS As described above, the following two specific branching ratios are measured in the current experiment by using the TS-VMI setup through state-selective detections of C( ^3 P _0 ) and C( ^3 P _2 ):  \begin{eqnarray} \rm{Ratio-I}& = &\frac{[ \rm{C}(^3 \rm{P}_0) + \rm{O}(^1 \rm{D})]}{ [ \rm{C}(^3 \rm{P}_0) + \rm{O}(^3 \rm{P})] +[ \rm{C}(^3 \rm{P}_0) + \rm{O}(^1 \rm{D})]} \\ \rm{Ratio-II} & = &\frac{[ \rm{C}(^3 \rm{P}_2) + \rm{O}(^1 \rm{D})]}{ [ \rm{C}(^3 \rm{P}_2) + \rm{O}(^3 \rm{P})] + [ \rm{C}(^3 \rm{P}_2) + \rm{O}(^1 \rm{D})]} \end{eqnarray} The overall plots of Ratio-I and Ratio-II versus the VUV photoexcitation energy measured in the current study are shown in FIG. 1 (a) and (b), respectively; the C ^+ ion photofragment spectrum by detecting C( ^3 P _0 ) is shown in FIG. 1(c). Since there is not a single laser dye that can cover the whole energy range, several different dyes are used. Thus the obtained several pieces of spectra are connected to each other without rescaling their relative intensities and no normalization according to the VUV intensities is performed either, thus only the peak positions and widths are meaningful in FIG. 1(c). Due to the very high detection sensitivity owning to the 1VUV+1UV/visible REMPI method used in the current experiment, many new absorption bands of ^{12} C ^{16} O are observed for the first time. Detailed analysis and accurate assignments of the complicated spectrum as shown in FIG. 1(c) are not trivial, we will thus mainly focus on presenting and discussing the dependence of photodissociation branching ratios on the VUV photoexcitation energy as shown in FIG. 1 (a) and (b) in the current work.  FIG. 1 (a) Ratio-I versus the VUV photoexcitation energy, the C( ^3 P _0 ) is ionized for detection through the 1VUV+1UV/visible REMPI method using the Rydberg transition of 2s ^2 2p3d ^3 D _1$$ \leftarrow $$^3 P _0 at 78293.5 cm ^{-1} ; (b) Ratio-II versus the VUV photoexcitation energy, the C( ^3 P _2 ) is ionized through the 1VUV+1UV/visible REMPI method using the Rydberg transition of 2s ^2 2p3d ^3 D _3$$ \leftarrow $$^3 P _2 at 78274.9 cm ^{-1} ; (c) C ^+ ion photofragment spectra of ^{12} C ^{16} O in the energy range 108000 cm ^{-1} -113200 cm ^{-1} , photoionization of C( ^3 P _0 ) is monitored by fixing the probe VUV at 78293.5 cm ^{-1} and scanning the pump VUV from 108000 cm ^{-1} to 113200 cm ^{-1} . The relative intensities of the spectra are not normalized according to the VUV intensities. The drop line shows the position of the ionization energy of ^{12} C ^{16} O As shown in FIG. 1 (a) and (b), Ratio-I and Ratio-II show interesting dependences on the VUV photoexcitation energy. They are close to zero (less than 5%) at most of the photon energy positions, which means that the channel for generating both atomic fragments in the electronic ground state dominates the photodissociation process at most part of the absorption spectrum. Only at several discrete narrow energy windows does the channel for producing O( ^1 D) open. These energy windows are 108350-108590 cm ^{-1} , 108970- 109170 cm ^{-1} , 109350-109600 cm ^{-1} , 110160-110360 cm ^{-1} , 111050-111420 cm ^{-1} , 112140-112310 cm ^{-1} and 112600-113000 cm ^{-1} . In the following, we zoom into all these energy windows and show them in detail. A. 109350-109600 cm ^{\textbf{-1}} ^{12} C ^{16} O shows very rich absorption structures in this energy region, which has attracted many spectroscopic studies [8, 10, 20]. The detailed C ^+ ion photofragment spectra obtained in the current study are presented in FIG. 2(c), and the corresponding Ratio-I and Ratio-II are shown in FIG. 2(a). To demonstrate how much the detection sensitivity has been improved by using the 1VUV+1UV/visible REMPI method through a second independently tunable probe VUV beam, we also present in FIG. 2(b) the C ^+ ion photofragment spectrum obtained in our previous study on the single VUV beam experimental setup for comparison [21]. Due to the much higher detection sensitivity in the current study, the signal to noise ( S / N ) ratio of the spectra has been dramatically improved, which enables us to clearly resolve the rotational lines up to J' = 19 in the P-branch of the absorption band of the (4s \sigma ) ^1$$ \Sigma $$^+ ( v' = 4) state, as shown in FIG. 2(c). In the absorption band of the (4p \pi ) ^1$$ \Pi$($v'$ = 3) state, rotational lines up to $J'$ = 11 have been resolved compared with our previous study with a single VUV beam, where the highest resolved rotational level was $J'$ = 4 [22].

 FIG. 2 Zoom-in view of the energy window 109300-109630 cm$^{-1}$ in FIG. 1. (a) Ratio-I (red dots) and Ratio-II (blue dots) versus the VUV photoexcitation energy; (b) C$^+$ ion photofragment spectrum obtained by using a single VUV beam, which is adopted from Ref.[21]; (c) C$^+$ ion photofragment spectra obtained by using the 1VUV+1UV/visible REMPI method as described in the text. Parts of the spectra in (b) and (c) are multiplied by constant factors as labeled in the figure. The two downward pointing arrows show the positions of local irregularities

Using these well rotationally resolved spectra, two local irregularities have been observed as indicated by the two downward pointing arrows in FIG. 2(c), where the rotational lines are either missing or with reduced intensities. In the two-photon ionization spectra obtained by Eidelsberg et al., the P(22) transition line of the absorption band of the (4s$\sigma$)$^1 $$\Sigma$$ ^+$($v'$ = 4) state was missing due to an accidental perturbation [10], which is confirmed in the current study (not shown in FIG. 2(c)). Besides P(22), the P(16) line is also missing as shown in FIG. 2(c), which did not show any irregularities in either the jet absorption or the two-photon ionization spectra obtained by Eidelsberg et al. [10]. The reasons for the discrepancy between these two spectroscopic studies are still not clear.

We observed irregular intensity reduction at R(4) and R(5) of the (4p$\pi$)$^1 $$\Pi ( v' = 3) state, that was not mentioned in the study by Eidelsberg et al. [10], but a small intensity reduction from R(5) can be noticed in their jet absorption and also the two-photon ionization spectra (see FIG. 15 in Ref.[10]). As analyzed previously by Eidelsberg et al., the states (4s \sigma ) ^1$$ \Sigma^+$($v'$ = 4), (4p$\sigma$)$^1 $$\Sigma$$ ^+$($v'$ = 3) and (5p$\pi$)$^1 $$\Pi ( v' = 1) are strongly coupled with each other, and all of them interact with the nearby C'$$ ^1 $$\Sigma^+ ( v' = 10) valence state [10]. The complicated coupling scheme can also be revealed by the nontrivial dependences of Ratio-I and Ratio-II on the rotational levels as shown in FIG. 2(a). The percentage of O( ^1 D) formed decreases as the rotational quantum number increases for all the four identified vibronic states studied here. This might be caused by the coupling between the Rydberg ^1\Pi states and the D'^1\Sigma^+ state which dissociates into the ground channel C( ^3 P)+O( ^3 P). The strength of this coupling, and thus the rates into the ground channel increase linearly with J' ( J' +1), which on the other hand will decrease the percentages of O( ^1 D) produced [22, 23, 25]. To quantitatively explain the rotational dependence of Ratio-I and Ratio-II as observed here, accurate theoretical calculations are needed. B. 108350-108590 cm ^{\textbf{-1}} The detailed structure of the C ^+ ion photofragment spectrum is shown in FIG. 3(c), and the corresponding Ratio-I and Ratio-II are shown in FIG. 3 (a) and (b), respectively. In FIG. 3(c), the diffuse absorption band of the (5d \sigma ) ^1$$ \Sigma $$^+ ( v' = 0) state is also shown as a reference, for which both Ratio-I and Ratio-II are close to zero. The photodissociation branching ratios of this absorption band were measured before in the single VUV beam experiment, and the percentages of the two spin-forbidden channels C( ^1 D)+O( ^3 P) and C( ^3 P)+O( ^1 D) were found to be less than 1%, which is consistent with the present study [22]. At the higher energy side of the diffuse band, two extremely weak absorption bands at 108505 cm ^{-1} and 108575 cm ^{-1} are observed, which were absent from our previous single VUV beam experiments [21, 22]. The absorption band at 108505 cm ^{-1} has well resolved rotational structures with the characteristics of a ^1$$ \Pi $$\leftarrow$$ ^1 $$\Sigma^+ transition. Its rotational constant and rotational transition lines are listed in Table Ⅰ. It has a rotational constant of 1.92 cm ^{-1} which is almost the same as the vibronic ground state of ^{12} C ^{16} O. This results in the typical completely unresolved Q branch as observed in FIG. 3(c). To our knowledge, this band has never been reported before. The diffuse absorption band at 108575 cm ^{-1} was observed before by Eidelsberg et al., and the upper state was assigned to the valence state C'$$ ^1\Sigma^+$($v'$ = 9) [10]. Both of these two weak absorption bands show substantial predissociation into the channels for producing O($^1$D) as shown by the relatively large values of Ratio-I and Ratio-II in FIG. 3 (a) and (b). In the single VUV beam experiment, the slightly lower vibrational state $C'^1\Sigma^+$($v'$ = 7) was identified, and significant predissociation into the two spin-forbidden channels C($^1$D)+O($^3$P) and C($^3$P)+O($^1$D) was observed [24, 36]. These observations indicate relatively strong interactions between the valence $C'^1\Sigma^+$ state and the triplet states which correlate with the spin-forbidden dissociation limits.

 FIG. 3 Zoom-in views of the energy windows 108350-108590 cm$^{-1}$ (a, b, c) and 108970-109170 cm$^{-1}$ (d, e, f) in FIG. 1. Some weak absorption bands in the spectra are multiplied by a constant factor to be shown in the figures, as marked by the numbers
Table Ⅰ Rotational constants and frequencies of the rotational transition lines for several newly observed absorption bands in the present study. The values in the parentheses are from Casey's PhD thesis [37]
C. 108970-109170 cm$^{\textbf{-1}}$

The detailed C$^+$ ion photofragment spectra in this energy window are presented in FIG. 3(f), and the corresponding Ratio-I and Ratio-II are shown in FIG. 3 (d) and (e), respectively. The two strong absorption bands of the $^1\Pi$($v'$ = 2) and (6p$\sigma$)$^1\Sigma^+$($v'$ = 0) states are also presented in FIG. 3(f) as references. These two absorption bands were studied before by using the single VUV beam setup, the percentages for generating O($^1$D) were found to be $\sim$58% for the absorption band of the $^1\Pi$($v'$ = 2) state, and less than 5% for the absorption band of the (6p$\sigma$)$^1 $$\Sigma^+ ( v' = 0) state [21, 22]. These observations are qualitatively consistent with the current measurements as shown in FIG. 3 (d) and (e). Besides the above two strong absorption bands, two weak absorption bands at 108987 and 109036 cm ^{-1} are observed in FIG. 3(f). To our knowledge, neither of these two bands is reported before. The band at 108987 cm ^{-1} is diffuse, and the percentage for generating O( ^1 D) is \sim 20%. The band at 109036 cm ^{-1} has well resolved rotational structures, its rotational constant and rotational transition frequencies are listed in Table Ⅰ. The upper state has a symmetry of ^1\Pi , and a rather small rotational constant of 1.50 cm ^{-1} . The branching ratios of this state for generating O( ^1 D) are almost the same as those of the nearby ^1\Pi ( v' = 2) state as shown in FIG. 3 (d) and (e). This implies relatively strong ^1$$ \Pi$-$^3 $$\Pi interactions in this energy region, which finally leads to the formation of the spin-forbidden channel C( ^3 P)+O( ^1 D). D. 110160-110360 cm ^{\textbf{-1}} The energy window 110160-110360 cm ^{-1} mainly covers the v' = 0 levels of Rydberg states 7s \sigma , 7p \sigma and 7p \pi . The detailed structure of the C ^+ ion photofragment spectrum is shown in FIG. 4(c), and the corresponding Ratio-I and Ratio-II are shown in FIG. 4 (a) and (b), respectively. The Rydberg state (7s \sigma ) ^1\Sigma^+ ( v' = 0) was observed before by Casey [37], the rotational constant and rotational transition lines measured in the current study as listed in Table Ⅰ are consistent with Casey's measurements within the line width of the VUV laser radiation. The percentages of producing O( ^1 D) at P(1), P(2), R(0) and R(1) are close to zero for (7s \sigma ) ^1$$ \Sigma $$^+ ( v' = 0), while the values of Ratio-I and Ratio-II at R(2) and R(3) are very large, and their intensities are also anomalously high. This is caused by overlapping with a state of ^1$$ \Pi$ symmetry which has not been observed before. The rotational constant and rotational transition lines of this new state are listed in Table Ⅰ. Its Q(1) and R(0) lines overlap with the R(2) and R(3) lines of (7s$\sigma$)$^1\Sigma $$^+ ( v' = 0), respectively; its Q(2) and R(1) lines are marked with stars in FIG. 4(c). To our knowledge, this ^1\Pi state has never been observed before. It predominantly dissociates into the channel for producing O( ^1 D), and this is the reason for the irregularities of transition line intensities and branching ratios observed for the state (7s \sigma ) ^1\Sigma^+ ( v' = 0) as described above.  FIG. 4 Zoom-in views of the energy windows 110160-110360 cm ^{-1} (a, b, c) and 111050-111420 cm ^{-1} (d, e, f) in FIG. 1. Some weak absorption bands in the spectra are multiplied by a constant factor to be shown in the figures, as marked by the red numbers. The stars in (c) represent rotational peaks belonging to the newly observed absorption band The Rydberg states (7p \sigma ) ^1\Sigma^+ ( v' = 0) and (7p \pi ) ^1\Pi ( v' = 0) have been observed before by Casey [37], their rotational constants and rotational transition lines determined in the current study are listed in Table Ⅰ. The line positions of these two studies are consistent except that the R-branch of the state (7p \pi ) ^1\Pi ( v' = 0) was shifted by one rotational quantum in Casey's assignment. The branching ratios of producing O( ^1 D) are close to zero for the state (7p \sigma ) ^1\Sigma^+ ( v' = 0), and 10%-30% for the state (7p \pi ) ^1$$ \Pi$($v'$ = 0), as shown in FIG. 4 (a) and (b). Between these two Rydberg states, there is an additional diffuse absorption structure, which has similar Ratio-I and Ratio-II values to those of the state (7p$\pi$)$^1\Pi$($v'$ = 0). The position of this structure is close to the state $C'^1\Sigma^+$($v'$ = 11) as predicted by Eidelsberg et al. [10], thus we tentatively assign this band as absorption from the ground vibronic state of $^{12}$C$^{16}$O to the valence state $C'^1\Sigma^+$($v'$ = 11). The relatively large values of Ratio-I and Ratio-II (10%-30%) again indicate that the interaction of $C'^1\Sigma^+$ with triplet states correlating with the spin-forbidden channels is quite strong, which is in accordance with that observed for $C'^1\Sigma^+$($v'$ = 7 and 9) as shown above.

E. 111050-111420 cm$^{\textbf{-1}}$

The energy window 111050-111420 cm$^{-1}$ mainly covers the $v'$ = 0 levels of Rydberg states 8p$\sigma$, 8p$\pi$, 9p$\sigma$, 9p$\pi$; and the $v'$ = 1 levels of Rydberg states 6p$\sigma$ and 6p$\pi$. The detailed structure of the C$^+$ ion photofragment spectra is shown in FIG. 4(f), and the corresponding Ratio-I and Ratio-II are shown in FIG. 4 (d) and (e), respectively. The amounts of O($^1$D) produced at the $v'$ = 0 levels of Rydberg states 8p$\sigma$ and 9p$\sigma$ are close to zero, and the corresponding $\pi$ components have slightly larger yields for O($^1$D), which are between 10% and 20%. The energy region for the $v'$ = 1 levels of Rydberg states 6p$\sigma$ and 6p$\pi$ is from $\sim$111330 cm$^{-1}$ to $\sim$111420 cm$^{-1}$, in which complicated rotational structures can be observed as presented in FIG. 4(f). Komatsu et al. studied this spectroscopic region by using a triple-photon resonance excitation ion-dip method, and detailed analysis using the $l$-uncoupling Hamiltonian model was performed [38]. According to their analysis, the absorption peaks at $\sim$111400 cm$^{-1}$ should belong to rovibronic transitions from the ground vibronic state to the Rydberg state (6p$\pi$)$^1\Pi$($v'$ = 1), which dissociates predominantly into the channels for producing O($^1$D); those absorption peaks at $\sim$111350 cm$^{-1}$ are mixtures of transitions into the Rydberg state (6p$\sigma$)$^1\Sigma^+$($v'$ = 1) and several other $^1\Sigma^+$ states of non-Rydberg type, this together with several local accidental perturbations makes the rotational structure of this region complicated [38]. This can also be revealed from the drastic variations of Ratio-I and Ratio-II in this narrow energy window as shown in FIG. 4 (d) and (e). Besides the Rydberg states described above, there is a weak absorption band with an upper state of $^1\Sigma^+$ symmetry located at $\sim$111239 cm$^{-1}$. The rotational constant and rotational transition lines determined in this study are listed in Table Ⅰ. Casey also observed a $^1\Sigma^+$ state at the same position, which was assigned as the Rydberg state ($B^2\Sigma^+$)3s$\sigma$($v'$ = 1) [37]. Casey determined its band origin at 111240.9 cm$^{-1}$ and a rotational constant of 1.764 cm$^{-1}$, compared with 111239.5 and 1.821 cm$^{-1}$ respectively as determined in the current study. This state is found to dissociate with substantial quantum yields into the channels for producing O($^1$D) as indicated in FIG. 4 (d) and (e).

F. 112140-112310 cm$^{\textbf{-1}}$ and 112600-113000 cm$^{\textbf{-1}}$

These two energy windows are close to the ionization threshold of $^{12}$C$^{16}$O at 113031.3 cm$^{-1}$ [39], thus high density of Rydberg states is supposed to be present in this region. The detailed structures of the C$^+$ ion photofragment spectra in these two energy windows are shown in FIG. 5 (c) and (f), and the corresponding Ratio-I and Ratio-II are shown in FIG. 5 (a, b), and (d, e), respectively. Accurate assignments of all these Rydberg states are not trivial. In the energy window 112140-112310 cm$^{-1}$, only the absorption peak at 112300 cm$^{-1}$ shows substantial dissociation into the channels for producing O($^1$D). The upper state of this band is tentatively assigned to be the Rydberg state 13p ($v'$ = 0) according to the spectroscopic study by Okazaki et al. [13]. In the energy window 112600-113000 cm$^{-1}$, there are two clusters of absorption peaks at $\sim$112670 cm$^{-1}$ and $\sim$112800 cm$^{-1}$, where relatively large values of Ratio-I and Ratio-II are observed. The peaks at 112670 cm$^{-1}$ are at the expected positions for the $v'$ = 0 levels of Rydberg states with $n$ = 18, and the peaks at 112800 cm$^{-1}$ are at the $v'$ = 0 levels of Rydberg states with $n$ = 22-24. From this study, we see that high Rydberg states close to the ionization threshold can still dissociate into the channels for producing O($^1$D) with substantial quantum yields.

 FIG. 5 Zoom-in views of the energy windows 112140-112310 cm$^{-1}$ (a, b, c) and 112600-113000 cm$^{-1}$ (d, e, f) in FIG. 1
Ⅳ. DISCUSSION AND SUMMARY

By using a single tunable VUV laser radiation source and the TS-VMI setup, we have measured the photodissociation branching ratios among the three channels C($^3$P)+O($^3$P), C($^1$D)+O($^3$P) and C($^3$P)+O($^1$D) for $^{12}$C$^{16}$O in the energy region 102500-110500 cm$^{-1}$ [21-23, 25], and a general trend of the branching ratios has been derived and summarized in a recent study [24]. For the directly excited singlet states of $^{12}$C$^{16}$O to dissociate into the spin-forbidden channels C($^1$D)+O($^3$P) and C($^3$P)+O($^1$D), spin-orbit couplings must have played an important role in the process, especially the $^1\Pi$-$^3\Pi$ type. Generally, excited states of $^1\Pi$ symmetry have large chance to go into the spin-forbidden channels due to the possible direct spin-orbit couplings with $^3\Pi$ states; excited states of $^1\Sigma^+$ symmetry usually need couple with a $^1\Pi$ state as an intermediate step to dissociate into the spin-forbidden channels, thus have relatively smaller chance. In this study, we have taken advantage of the extremely high detection efficiency provided by the 1VUV+1UV/visible REMPI method using a second independently tunable VUV laser radiation source, and extended the branching ratio measurements to much higher photoexcitation energies close to the ionization threshold of $^{12}$C$^{16}$O. Our measurements prove that the general trend described above is also applied at higher energy range: states of $^1\Pi$ symmetry have much larger chance to dissociate into the spin-forbidden channels than states of $^1\Sigma^+$ symmetry due to the strong $^1\Pi$-$^3\Pi$ interactions in the photodissociation process of $^{12}$C$^{16}$O. Most of the absorption bands identified in the current study that show substantial quantum yields for producing O($^1$D) have upper states of $^1\Pi$ symmetry, such as the absorption bands at 108505, 109017, 109036, 109480, 109565, 110175, 110350, 111400 cm$^{-1}$. For Rydberg states with large $n$ values, the $^1\Sigma^+$ and $^1\Pi$ components are usually strongly mixed due to the $l$-uncoupling interactions, this is the case for the states observed in the energy windows 112140-112310 cm$^{-1}$ and 112600-113000 cm$^{-1}$ as described in the previous section.

In summary, we have combined a TS-VMI setup with two independently tunable VUV laser radiation sources generated by TPRE-FWM to measure the two branching ratios [C($^3$P$_0$)+O($^1$D)]{[C($^3$P$_0$)+O($^3$P)]+[C($^3$P$_0$)+O($^1$D)]} and [C($^3$P$_2$)+O($^1$D)]{[C($^3$P$_2$)+O($^3$P)]+[C($^3$P$_2$)+ O($^1$D)]} in the VUV photoexcitation energy range 108000-113200 cm$^{-1}$. Our measurements tell at which energy windows the channels for producing O($^1$D) open. This is important not only for understanding the complicated photodissociation dynamics of $^{12}$C$^{16}$O, but also for practical applications in astrophysics and astrochemistry, as the production of excited oxygen atoms from VUV photodissociation of CO is believed to play an important role in explaining the anomalous oxygen isotope distribution observed in the Solar system [27, 40, 41].

Ⅴ. ACKNOWLEDGMENTS

This work was supported by the National Natural Science Foundation of China (No.21803072), the Program for Young Outstanding Scientists of Institute of Chemistry, Chinese Academy of Science (ICCAS), and Beijing National Laboratory for Molecular Sciences (BNLMS). This material is based upon work supported by the National Aeronautics and Space Administration Award #80NSSC18K0592 and National Science Foundation under CHE-1763319. Yu Song and William. M. Jackson gratefully acknowledge the support of NSF under grants CHE-1301501 and AST-1410297.

Reference
 [1] R. Hakalla, M. L. Niu, R. W. Field, A. N. Heays, E. J. Salumbides, G. Stark, J. R. Lyons, M. Eidelsberg, J. L. Lemaire, S. R. Federman, N. de Oliveira, and W. Ubachs, J. Quant. Spectrosc. Radiat. Transf. 189, 312(2017). DOI:10.1016/j.jqsrt.2016.12.012 [2] R. Hakalla, T. M. Trivikram, A. N. Heays, E. J. Salumbides, N. de Oliveira, R. W. Field, and W. Ubachs, Mol. Phys. 117, 79(2019). DOI:10.1080/00268976.2018.1495848 [3] P. Cacciani, F. Brandi, J. P. Sprengers, A. Johansson, A. L'Huillier, C. G. Wahlström, and W. Ubachs, Chem. Phys. 282, 63(2002). DOI:10.1016/S0301-0104(02)00622-5 [4] R. N. Clayton, Nature 415, 860(2002). [5] H. Yurimoto, and K. Kuramoto, Science 305, 1763(2004). DOI:10.1126/science.1100989 [6] J. R. Lyons, and E. D. Young, Nature 435, 317(2005). DOI:10.1038/nature03557 [7] M. Ogawa, and S. Ogawa, J. Mol. Spectrosc. 41, 393(1972). DOI:10.1016/0022-2852(72)90212-3 [8] K. S. E. Eikema, W. Hogervorst, and W. Ubachs, Chem. Phys. 181, 217(1994). DOI:10.1016/0301-0104(94)85026-7 [9] M. Komatsu, T. Ebata, T. Maeyama, and N. Mikami, J. Chem. Phys. 103, 2420(1995). DOI:10.1063/1.469665 [10] M. Eidelsberg, F. Launay, K. Ito, T. Matsui, P. C. Hinnen, E. Reinhold, W. Ubachs, and K. P. Huber, J. Chem. Phys. 121, 292(2004). DOI:10.1063/1.1756579 [11] M. Eidelsberg, J. L. Lemaire, S. R. Federman, G. Stark, A. N. Heays, Y. Sheffer, L. Gavilan, J. H. Fillion, F. Rostas, J. R. Lyons, P. L. Smith, N. de Oliveira, D. Joyeux, M. Roudjane, and L. Nahon, A & A 543, A69 (2012). [12] A. Okazaki, T. Ebata, T. Sutani, and N. Mikami, J. Chem. Phys. 108, 1765(1998). DOI:10.1063/1.475553 [13] A. Okazaki, T. Ebata, and N. Mikami, J. Chem. Phys. 114, 7886(2001). DOI:10.1063/1.1363672 [14] S. O'Neil, and H. F. Schaefer, J. Chem. Phys. 53, 3994(1970). DOI:10.1063/1.1673871 [15] A. N. Heays, M. Eidelsberg, G. Stark, J. L. Lemaire, L. Gavilan, S. R. Federman, B. R. Lewis, J. R. Lyons, N. de Oliveira, and D. Joyeux, J. Chem. Phys. 141, 144311(2014). DOI:10.1063/1.4897326 [16] M. Majumder, N. Sathyamurthy, G. J. Vázquez, and H. Lefebvre-Brion, J. Chem. Phys. 140, 164303 (2014). [17] H. Lefebvre-Brion, and A. Kalemos, J. Chem. Phys. 145, 166102(2016). DOI:10.1063/1.4966687 [18] H. Lefebvre-Brion, and A. Kalemos, J. Chem. Phys. 144, 134302(2016). DOI:10.1063/1.4945117 [19] H. Lefebvre-Brion, and M. Majumder, J. Chem. Phys. 142, 164306(2015). DOI:10.1063/1.4918910 [20] K. P. Huber, Philos. Trans. R. Soc. London, Ser. A 355, 1527(1997). DOI:10.1098/rsta.1997.0074 [21] H. Gao, Y. Song, L. Yang, X. Shi, Q. Yin, C. Y. Ng, and W. M. Jackson, J. Chem. Phys. 135, 221101(2011). DOI:10.1063/1.3669426 [22] H. Gao, Y. Song, L. Yang, X. Shi, Q. Z. Yin, C. Y. Ng, and W. M. Jackson, J. Chem. Phys. 137, 034305(2012). DOI:10.1063/1.4734018 [23] H. Gao, Y. Song, Y. C. Chang, X. Shi, Q. Z. Yin, R. C. Wiens, W. M. Jackson, and C. Y. Ng, J. Phys. Chem. A 117, 6185(2013). DOI:10.1021/jp400412n [24] X. Shi, H. Gao, Q. Z. Yin, Y. C. Chang, R. C. Wiens, W. M. Jackson, and C. Y. Ng, J. Phys. Chem. A 122, 8136(2018). DOI:10.1021/acs.jpca.8b08058 [25] P. Jiang, X. Chi, Q. Zhu, M. Cheng, and H. Gao, Phys. Chem. Chem. Phys. 21, 14376(2019). DOI:10.1039/C8CP07620K [26] H. Gao, and C. Y. Ng, Chin. J. Chem. Phys 32, 23(2019). DOI:10.1063/1674-0068/cjcp1812290 [27] X. Shi, Q. Z. Yin, H. Gao, Y. C. Chang, W. M. Jackson, R. C. Wiens, and C. Y. Ng, Astrophys. J. 850, 48(2017). DOI:10.3847/1538-4357/aa8ee7 [28] J. R. Lyons, E. Gharib-Nezhad, and T. R. Ayres, Nat. Commun. 9, 908(2018). DOI:10.1038/s41467-018-03093-3 [29] H. Gao, Y. Song, W. M. Jackson, and C. Y. Ng, J. Chem. Phys. 138, 191102(2013). DOI:10.1063/1.4807302 [30] Y. Song, H. Gao, Y. C. Chang, D. Hammoutene, H. Ndome, M. Hochlaf, W. M. Jackson, and C. Y. Ng, Astrophys. J. 819, 23(2016). DOI:10.3847/0004-637X/819/1/23 [31] Y. C. Chang, K. Liu, K. S. Kalogerakis, C. Y. Ng, and W. M. Jackson, J. Phys. Chem. A 123, 2289(2019). DOI:10.1021/acs.jpca.8b11691 [32] Z. Lu, Y. C. Chang, H. Gao, Y. Benitez, Y. Song, C. Y. Ng, and W. M. Jackson, J. Chem. Phys. 140, 231101(2014). DOI:10.1063/1.4883515 [33] Z. Lu, Y. C. Chang, Q. Z. Yin, C. Y. Ng, and W. M. Jackson, Science 346, 61(2014). DOI:10.1126/science.1257156 [34] Z. Lu, Y. C. Chang, Y. Benitez, Z. Luo, A. B. Houria, T. Ayari, M. M. Al Mogren, M. Hochlaf, W. M. Jackson, and C. Y. Ng, Phys. Chem. Chem. Phys. 17, 11752(2015). DOI:10.1039/C5CP01321F [35] Y. Song, H. Gao, C. Y. Ng, and W. M. Jackson, EAS Publications Series 58, 295(2012). DOI:10.1051/eas/1258049 [36] P. Jiang, X. Chi, Q. Zhu, M. Cheng, and H. Gao, Nat. Commun. 10, 3175(2019). DOI:10.1038/s41467-019-11086-z [37] M. Casey, PhD thesis, Dublin: University of College, (1978). [38] M. Komatsu, T. Ebata, and N. Mikami, J. Chem. Phys. 99, 9350(1993). DOI:10.1063/1.465519 [39] P. Erman, A. Karawajczyk, E. Rachlew-Källne, C. Strömholm, J. Larsson, A. Persson, and R. Zerne, Chem. Phys. Lett. 215, 173(1993). DOI:10.1016/0009-2614(93)89283-N [40] K. D. McKeegan, A. P. A. Kallio, V. S. Heber, G. Jarzebinski, P. H. Mao, C. D. Coath, T. Kunihiro, R. C. Wiens, J. E. Nordholt, R. W. Moses, D. B. Reisenfeld, A. J. G. Jurewicz, and D. S. Burnett, Science 332, 1528(2011). DOI:10.1126/science.1204636 [41] R. N. Clayton, Proc. 74th Annual Meteoritical Society MeetingHouston, TX: LPI 5010, (2011).

a. 中国科学院化学研究所，北京分子科学国家研究中心，北京 100190;
b. 加利福尼亚大学戴维斯分校化学系，戴维斯 95616