Fragmentation of doubly charged HDO, H2O, and D2O molecules induced by proton and monocharged fluorine beam impact at 3 keV S. Martin, L. Chen, R. Brédy, J. Bernard, and A. Cassimi

Citation: J. Chem. Phys. 142, 094306 (2015); doi: 10.1063/1.4913398 View online: http://dx.doi.org/10.1063/1.4913398 View Table of Contents: http://aip.scitation.org/toc/jcp/142/9 Published by the American Institute of Physics

THE JOURNAL OF CHEMICAL PHYSICS 142, 094306 (2015)

Fragmentation of doubly charged HDO, H2O, and D2O molecules induced by proton and monocharged fluorine beam impact at 3 keV S. Martin,1 L. Chen,1 R. Brédy,1 J. Bernard,1 and A. Cassimi2 1

Institut Lumière Matière, UMR5306 Université Claude Bernard Lyon 1-CNRS, Université de Lyon, 69622 Villeurbanne Cedex, France 2 CIMAP, GANIL Université de CAEN, Bd. H. Becquerel, Caen, France

(Received 21 October 2014; accepted 11 February 2015; published online 4 March 2015) Doubly charged ions HDO2+, H2O2+, and D2O2+ were prepared selectively to triplet or singlet excited states in collisions with F+ or H+ projectiles at 3 keV. Excitation energies of dications following two-body or three-body dissociation channels were measured and compared with recent calculations using ab initio multi-reference configuration interaction method [Gervais et al., J. Chem. Phys. 131, 024302 (2009)]. For HDO2+, preferential cleavage of O–H rather than O–D bond has been observed and the ratio between the populations of the fragmentation channels OD+_H+ and OH+_D+ were measured. The kinetic energy release has been measured and compared with previous experiments. C 2015 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4913398] I. INTRODUCTION

Understanding the fragmentation dynamics of small molecules, especially when two or three electrons are removed, is still a challenging issue. A recent work on the fragmentation of CO23+ shows a large variety of expected or unexpected dissociation processes.2 Fragmentation of doubly charged water or heavy water ions has been studied with a large variety of excitation methods using photons,3 electrons,4,5 slow6,7 or fast ion8,9 impact. Two-body and/or three-body dissociation schemes have been predicted in theoretical calculations for each molecular energy state.1 For doubly charged HDO molecules, isotope effect has been predicted and observed8 showing dominant OD+_H+ dissociation channel via the cleavage of O-H bond. A usual description of the process is based on the vertical transition for removing two electrons and the competition between the losses of H+ or D+ fragments. Indeed, in the repulsive electrostatic field, the lighter fragment H+ is more accelerated than D+. After a half period of vibration, H+ can reach a larger distance from the oxygen atom, while the D+ ion stays in average in closer vicinity of the oxygen. Therefore, H+ can be lost more easily leading preferentially to the formation of the OD+ stable molecule. The ratio of probabilities for the two dissociation channels OD+_H+ over OH+_D+ is called in the following the “isotopic ratio.” It was measured to be around 6.5 in previous collision experiments,8 while calculations showed a strong dependence on the energy and the multiplicity of the HDO2+ molecular state. For example, the isotopic ratio for the ground triplet state was calculated to be 8.6, whereas for the first and second singlet states the ratios were found to be 3 and 15.7, respectively.1 Up to now, the experimental value of the isotopic ratio has been measured without differentiation of the multiplicity and the selection of excited states. It corresponds then to an average value due to the population of the ground and several excited states and can be estimated taking into account of the weight for each state. In addition, in these previous experiments, the kinetic energy release (KER) for 0021-9606/2015/142(9)/094306/8/$30.00

each dissociation channel has been measured, which provided information on the validity of assumptions on the vertical transition and on the asymptotic states of the dissociation channels. The aim of this paper is to investigate in more details the dissociation dynamics of HDO2+, H2O2+, and D2O2+ by measuring the excitation energies of the triplet and singlet states, the KER associated to each fragmentation channel, and the isotopic ratio versus the multiplicity for HDO2+. The measurements have been performed in collisions with F+ or H+ projectiles at 3 keV. Using the CIDEC (Collision Induced Dissociation under Energy Control) method,10 the excitation energy deposited into target molecules was controlled and analyzed in coincidence with the detection of the fragmentation products. Due to the spin conservation rule in the collision, the triplet states were investigated using the monocharged fluorine projectile whereas the singlet states were studied using the proton projectile.

II. EXPERIMENTAL METHOD

The CIDEC method provides the internal excitation energy map of the target molecules prior to dissociation. It is originated from the double-charge-transfer (DCT) spectroscopy developed for studying energy levels of electronic-states of doubly charged molecular ions.11,12 DCT spectroscopy is based on the formation of anions by double electron capture in inelastic collisions between a singly charged incident projectile A+ and a neutral molecular target M, A+ + M → A− + M2+∗. By measuring the kinetic energy loss of the anion A−, the ground and excited energy levels of the doubly charged target could be determined. DCT spectroscopy has been applied for the water molecule using F+ and H+ projectiles in order to study selectively the doubly charged triplet and singlet excited states.3,7 The CIDEC method developed in our laboratory combined the DCT spectroscopy, fragment mass spectroscopy,

142, 094306-1

© 2015 AIP Publishing LLC

094306-2

Martin et al.

J. Chem. Phys. 142, 094306 (2015)

and coincidence measurement techniques. It aimed at studying the decay of target molecules as a function of the excitation energy deposited during ion-target collisions. According to the principle of this method, the kinetic energy loss ∆E of the anion was related to the internal excitation energy Eint of the doubly charged molecule gained in the charge exchange process.10 Since the anion can be detected only if it was formed in the ground state, the energy balance of the interaction A+ + M → A− + M2+∗ is given by I1(A) = −∆E − EA(A) + I1(M) + I2(M) + Eint. In this expression, I1(A) and EA(A) stand, respectively, for the first ionization potential and the electron affinity of the projectile A, which are precisely known for H and F atoms; I1(M) and I2(M) for the first and second ionization potential of the target M. For water neutral molecules, the ground electronic configuration is noted (1a1)2 (2a1)2 (1b2)2 (3a1)2 (1b1)2. The doubly charged water molecular ions have not been observed as stable or metastable entities even at the ground state (3a1)−1 (1b1)−1 3 B1 where two electrons are removed from the outermost orbitals (3a1)2 (1b1)2. Therefore, it is not possible to determine the internal excitation energy Eint with respect to the stable ground doubly charged state as in the previous studies on larger molecules using the CIDEC method. As a consequence, in this work, the excitation energy Eexc of the doubly charged molecular ions is defined with respect to the ground state of the neutral as in the DCT spectroscopy and is determined from the measured kinetic energy loss ∆E with the relation, Eexc = I1(A) + EA(A) + ∆E.

(1)

By using projectile ions of different I1(A) + EA(A) values, the excitation energy region can be controlled roughly. In the present experiment, two singly charged projectiles were selected H+ and F+, with I1(H) + EA(H) = 14.3 eV and I1(F) + EA(F) = 20.8 eV. Therefore, comparing to F+ projectile, the global excitation energy window is expected to shift to lower values by about 6.5 eV with H+ projectile. Furthermore, due to the spin conservation rule during the charge exchange process, doubly charged excited states were populated with selective multiplicity using these two projectiles.3,7,11,13 Indeed, the electron spin of the ground state of neutral water molecules is S = 0, that of the proton and the ground state of H− is also S = 0, the electronic states of doubly charged molecules populated in the following charge exchange process H+ + M → H− + M2+∗ should be singlet with S = 0. Differently, the ground state of F+ is triplet with S = 1 and that of F− is singlet with S = 0, therefore doubly charged molecules formed in collisions F+ + M → F− + M2+∗ should occupy triplet states with S = 1. Here, we have neglected the contribution of singlet metastable states of F+, for example, the lowest one 1 D at 3 eV. In fact, in ion beams delivered by an electron cyclotron resonance (ECR) source, the metastable yield was found usually in the order of a few percent.14

III. EXPERIMENTAL SETUP

The experimental setup shown in Fig. 1 has been described in previous papers.10,15,16 A mono charged fluorine or proton beam was extracted at 3 kV from an ECR Nanogan

FIG. 1. Experimental setup. The mixture of HDO, H2O, and D2O are sent through a needle into the interaction region.

III ion source. It collided on a molecular jet composed of a mixture of H2O, D2O, and HDO escaping from a small chamber through a needle placed perpendicularly and closely to the ion beam. HDO molecules were produced by spontaneous exchange between H and D atoms in a mixture of pure liquids H2O and D2O. The liquid containing a fraction of HDO resulting from H-D exchange was frozen and installed in the small chamber. A fast primary pumping was performed to remove the air just before connecting the chamber to the needle through an adjustable valve. At the room temperature, the mixture was melted and a small part of it was evaporated to fill the chamber. By adjusting precisely the valve, the molecular flux was controlled in order to reduce the probability for a single projectile ion to collide with more than one target molecule. In the interaction region, an electric field of 60 V/cm perpendicular to both the ion beam and the molecular jet was applied. The 3 keV ion beam of energy dispersion estimated to 1 eV was focused with an Einzel lens to about 0.2 mm in diameter at the crossing point with the molecular jet. After a two-electron-transfer collision event, the scattered anion passing through an electrostatic analyzer was detected with a channeltron placed just behind the exit slit of the analyzer. Due to the transverse extension of the incident ion beam along the electric field, additional kinetic energy dispersion estimated to 1 eV was induced for the scattered anions. The recoil ions, composed of charged fragments from transient parent ions H2O2+, D2O2+, or HDO2+, were extracted from the collision region along the transverse electric field, then accelerated to about 3 keV/charge and analyzed with a time-of-flight (TOF) spectrometer. After flying through the TOF tube, the recoil ions hit a detector equipped with multichannel plates and a 128-pixel multianode. With this detector, it was possible to differentiate two fragments arriving simultaneously at the detector according to the sites of the activated pixels. For each collision event, the detection time of the anion provided the reference trigger for the time-to-digital converter of the TOF spectrometer. As a consequence, the TOF of the fragments resulting from a collision event was recorded in coincidence with the corresponding scattered projectile anion. In our experiment, the voltage applied to the analyzer was related directly to the kinetic energy loss of the detected anion. It was scanned step by step. The TOF spectrum was recorded at each step and plotted in a 2D spectrum as a function of the

094306-3

Martin et al.

J. Chem. Phys. 142, 094306 (2015)

FIG. 3. Projection of the spot OD+ to the vertical axis of the 2D EX-RI spectra in Figs. 2 and 4. Squares: from Fig. 2 (F+); Circles: from Fig. 4 (H+).

FIG. 2. Top panel: 2D EX-RI spectrum for the collisions F+ + M → F− + M2+∗ (triplet states). M represents the mixture of HDO, H2O, and D2O. Horizontal axis: Time of flight of the recoil ions; vertical axis: Eexc defined as in Eq. (1). This 2D EX-RI spectrum is divided into three regions according to three ranges of Eexc. (a) Projection of the 2D EX-RI spectrum to the horizontal axis in low Eexc region corresponding to double collisions. The peak attributed to H2O+ is partly contributed by OD+. (b) Projection of the 2D spectrum to the horizontal axis in medium Eexc region. (c) Projection of the 2D spectrum to the horizontal axis in high Eexc region.

voltage. The voltage axis of the 2D spectrum was converted to the excitation energy Eexc according to Eq. (1) by use of argon gas as reference target for the kinetic energy loss calibration. In the following, the TOF spectra of recoil ions are presented as a function of the excitation energy Eexc in a 2D-coincidence spectrum, called EX-RI spectrum (EXcitation-Recoil Ion).

IV. RESULTS AND ANALYSES

The 2D EX-RI spectrum obtained in collisions F+ + HDO(H2O, D2O) → F− + HDO2+∗ (H2O2+∗, D2O2+∗) is shown in Fig. 2. In these interactions, doubly charged transient parent ions were prepared in triplet states, which followed different dissociation channels. The most intense spots in Fig. 2 are attributed to OD+, OH+, D+, and H+ fragments. First, one notes that the projection to the vertical axis of an individual spot of Fig. 2 gives the energy Eexc dependence of the yield for the corresponding fragmentation channels. As an example, the projection to the vertical axis of the spot OD+ is presented in Fig. 3 showing a maximum yield around Eexc = 40 eV. Other intense spots, OH+, D+, and H+ are observed around the same Eexc value. This suggests that in this energy range doubly charged molecules dissociated preferentially into two charged fragments, (OD+ or OH+)_(D+ or H+). However, different from the spots of heavy fragments OD+ and OH+, the spots of H+ and D+ show an extension toward higher Eexc values. This suggests the presence of three-

body channels at slightly higher energies involving two light charged fragments and the neutral oxygen. To analyze roughly the main features of the fragmentation channels at different excitation energy regions, the 2D EX-RI spectrum was divided into three parts according to Eexc. Partial projections onto the TOF axis of the 2D spectrum selected for low (50 eV) energy ranges are shown in Figs. 2(a), 2(b), and 2(c), respectively. Peaks observed in Fig. 2(a) in low energy region are mainly assigned to D2O+, HDO+, H2O+ (DO+), and OH+. The peak attributed to H2O+ is partly contributed by the fragment OD+, while those of D2O+ and HDO+ can be identified as intact singly charged molecules without ambiguity. They are attributed to one electron capture processes in double collisions, i.e., a projectile F+ collides successively on two molecules leading to F0 in the first collision and F− in the second collision. The energy Eexc obtained from the measured kinetic energy loss of the projectile using Eq. (1) corresponds to the summation of the excitation energy of the two singly charged molecules involving in the double collision process. The measured value is around 25 eV reaching the upper limit at 28 eV. As observed for water molecules in double collisions with DCT spectroscopy,7 the first two vertical ionization energies were found to be 12.62 and 14.74 eV. Our measured energy 25 eV corresponds well to twice the first ionization energy (12.62 eV) and the upper limit 28 eV is comparable to the summation of the first two vertical ionization energies, (12.62 + 14.74) eV. This shows that in double collision processes, one of the singly charged target molecules could be prepared in an excited state. In the present experiment, double collision events were reduced to less than 1% of the total collision events. Relative intensities of D2O+ and HDO+ peaks reflect the partial pressures of D2O and HDO molecules in the interaction region. We note that the density of HDO is larger than that of D2O demonstrating the efficiency of HDO formation via H-D exchange in the liquid initially composed of H2O and D2O. In Figure 2(b), the dominant peaks are assigned to OD+, + OH , D+, and H+. Due to the presence of H2O2+, D2O2+, HDO2+ transient parent ions in the collision region, each fragment peak is not related to a unique parent ion. For

094306-4

Martin et al.

example, the peak OD+ can be attributed to the dissociation of HDO2+ into OD+_H+ and/or the dissociation of D2O2+ into OD+_D+. So, from this spectrum, we are not able to get precise information on the dissociation of one particular molecular ion and to measure, for example, the isotopic ratio of HDO2+. For light fragments, the assignment is more difficult. For example, the peak H+ is a mixture of contributions from four dissociation channels, HDO2+ into OD+_H+ or H+_D+_O, H2O2+ into OH+_H+ or H+_H+_O. To identify precisely a dissociation channel, the coincident measurement of charged fragments from a single dissociation event is required. Although the peak O+ is very weak in Fig. 2(b) due to its relatively low yield comparing to other peaks in the spectrum, the corresponding spot can be clearly identified in the 2D spectrum with excitation energy around Eexc = 43.8 eV. The observation of the spot O+ is tentatively attributed to threebody dissociation channels, O+_D+_(H0 or D0) or O+_H+_(H0 or D0) from doubly charged parent molecular ions. A highenergy component of such three-body dissociation channels of doubly charged parent ions is observed in Figure 2(c). The corresponding O+ spot in the 2D spectrum was measured to be centered at Eexc = 62.4 eV. It is notable that in Fig. 2(c) only atomic ions O+, H+, and D+ are present showing that in this high-energy range two-body dissociation channels vanish and all molecular bonds are broken. In Fig. 2(c), the peak O+ appears broader than the heavy fragments resulting from the two-body fragmentation in the medium energy range, for example, the OH+ and OD+ peaks in Fig. 2(b). The peaks H+ and D+ in Fig. 2(c) are also broader than the corresponding peaks in Fig. 2(b). This suggests that the KER of the three-body fragmentation channels is larger at this higher Eexc region. The 2D EX-RI spectrum obtained in collisions H+ + HDO(H2O, D2O) → H− + HDO2+∗ (H2O2+∗, D2O2+∗) is shown in Fig. 4. By using the proton projectile, doubly charged

FIG. 4. Top panel: 2D EX-RI spectrum in H+ + M → H− + M2+∗ (singlet states) collisions. M represents the mixture of HDO, H2O, and D2O. Horizontal axis: Time of flight of the recoil ions; vertical axis: Eexc defined as in Eq. (1). (a) Projection of the 2D EX-RI spectrum to the horizontal axis for Eexc < 34 eV. (b) Projection of the 2D EX-RI spectrum to the horizontal axis for Eexc > 34 eV.

J. Chem. Phys. 142, 094306 (2015)

transient parent ions were prepared in singlet states. Comparing to the 2D EX-RI spectrum (Fig. 2) obtained in collisions with F+, the upper limit of this 2D spectrum (52 eV) is about 10 eV lower. In fact, according to the principle of the CIDEC method as explained previously, the excitation energy window using H+ impact is shifted by about 6.4 eV to lower values than in collisions with F+. Therefore, the probability to prepare parent ions in higher energy region, typically around 60 eV, is negligible. Except the upper limit, the two spectra in Figs. 2 and 4 are quite similar. The projection to the vertical axis of the spot OD+ is presented in Fig. 3. Comparing to the projection obtained in collisions with F+, the maximum yield with H+ is shifted to higher energy by about 1 eV. From similar analyses of the other spots, a global energy shift of 1 eV to higher energies is noticed in the case of H+ impact. This difference is closely related to the energy shift of the singlet state with respect to the triplet state of doubly charged water molecules at a given electronic configuration. Indeed, the lowermost singlet state of doubly charged water molecule (1b1)−2 1A1 is about 1 eV higher than the ground triplet state (3a1)−1 (1b1)−1 3 B1. In fact, the repulsive potential energy surfaces of the above singlet and triplet states of doubly charged parent ions are asymptotically correlated to the atomic states of neutral oxygen with the corresponding multiplicity and the lowest singlet state O(1D) is about 1.967 eV17 higher than the ground triplet state O(3P). In order to identify the parent ions related to measured fragments, we have analyzed the coincidence between fragments. Figures 5 and 6 show a part of coincidence maps obtained in collisions with F+ and H+, respectively. Correlation between a light fragment, H+ or D+, and a heavy fragment, OD+ or OH+, leads to four spots attributed to OD+_H+, OD+_D+, OH+_H+, and OH+_D+. Each spot is related to a well-defined parent ion. The spots OD+_H+ and OH+_D+ result from the two-body dissociation of HDO2+ parent ions.

FIG. 5. Coincidence map between the TOFs of the heavy and the light fragments for F+ + M → F− + M2+∗ collisions. The shape of the spots is due to the recoil ion collection restriction condition. Only ions with forward and backward trajectories were collected and detected.

094306-5

Martin et al.

J. Chem. Phys. 142, 094306 (2015)

FIG. 6. Coincidence map between the TOFs of the heavy and the light fragments for H+ + M → H− + M2+∗ collisions.

As shown in Figs. 5 and 6, the dominant dissociation channel of HDO2+ corresponds well to OD+_H+ in the two cases. By measuring the count numbers of the two spots OD+_H+ and OH+_D+, the ratio of the counts was estimated to be 4.5 ± 0.8 using H+ and 3.8 ± 0.8 using F+ projectiles. These two values correspond, respectively, to the isotopic ratios for singlet and triplet states of doubly charged parent ions populated in this experiment. The large uncertainty is mainly due to the correction of random coincidence count for the weak OH+_D+ spot. From the coincidence maps (Figs. 5 and 6), the KER has been also extracted for each fragmentation channel. By selecting only a small number of pixels around the position where the stable monocharged target ions hit the MCP detector, fragments with initial velocity forward and backward the detector were selectively recorded in the spectra. The shape of each spot is therefore typical and is carrying information on the kinetic energy of fragments of the corresponding dissociation channel. Here, diagonal partial projection of each spot is composed of two well-separated peaks. This suggests that the KER energy distribution and accordingly the Franck-Condon

factor of the vertical transition are not very broad. This is in qualitative agreement with the COLTRIMS8 (cold target recoil ions momentum spectroscopy) experiment, where the width of the KER distribution was measured to be around 2 eV. By measuring the shift between the backward and forward peaks, we have deduced the KER for each dissociation channel. Due to the poor resolution, no difference was observed using the two projectiles. Spots with low statistics corresponding to O+_H+ and O+_D+ can be noticed in the Figs. 5 and 6. The absence of the spot O+_D2+ in the coincidence spectrum and the absence of the spot O+_H2+ using pure water as the target show that the emission of molecular hydrogen H2+ or D2+ is negligible. This supports our attribution of spots O+_H+ and O+_D+ to three-body dissociation including a missed small neutral, H or D. The statistics for the above three-body processes were too poor to measure the KER in the present experiment. However, statistics for three-body processes leading to two charged small fragments and a neutral oxygen, H+_D+_O, H+_H+_O or D+_D+_O, were rather high. By analyzing the fragment coincidence spots, H+_D+, H+_H+, and D+_D+ (not shown in Figs. 5 and 6), we have extracted the KER for the corresponding threebody dissociation processes. The measured KER values are presented in Table I and compared with the data obtained in previous experiments. Our values are close to the KER measured by Richardson et al.3 where water molecules were ionized and excited via HeII photon absorption. Comparing to experiments of collisions using highly charged ions,8,18 Ni25+ at 11.7 MeV/u, S15+ at 13.5 MeV/u, and Ne10+ at 170 keV, and theoretical estimations,8 our values are about 2 eV lower than the measured KER and 1 eV lower than the theoretical ones. For measuring the mean excitation energy of parent ions corresponding to each fragmentation channel, we have extracted the excitation energy distribution in coincidence with the detection of two charged fragments. Figure 7 shows an example of this kind of analysis for experiment using proton projectile. In Fig. 7(a), the 2D EX-RI spectrum was built by selecting events where one of the two charged fragments was D+ and by plotting the TOF of the other charged fragment versus the excitation energy Eexc of the molecule. In Fig. 7(b), the 2D EX-RI spectrum was built in coincidence with H+. In the 2D spectra of Fig. 7, each spot is assigned to a dissociation channel without ambiguity.

TABLE I. Experimental KER (eV) measurements and comparison with other measurements. KER (eV) Muranaka18 Dissociation channel OD+_H+ OD+_D+ OH+_H+ OH+_D+ H+_D+_(O) H+_H+_(O) D+_D+_(O)

This experiment 5 5 5 4.5 5.5 5 5

Richardson3

Legendre8

4.7 4.8 4.5 4.7 4.5

7

6 8.5

4.3

7.5

S15+ 13.5 MeV/u

Ne10+ 170 keV

6.9 6.5 6.5 6.5

6.7 6.3 6.3 6.3

094306-6

Martin et al.

J. Chem. Phys. 142, 094306 (2015)

FIG. 7. In collisions with H+, 2D spectra built in coincidence with D+ (a) and H+ (b). Horizontal axis: Eexc; vertical axis: TOF of the correlated charged ion. Projections of individual spots of the 2D spectrum to the horizontal axis for O_H+_D+; OH+_D+; OD+_D+; OH+_H+; OD+_H+ channels.

The projection of a spot to the energy axis is well fitted with a Gaussian function and gives the yield of the corresponding fragmentation channel as a function of the excitation energy of the parent molecules. Similar analyses have been performed for the data obtained with fluorine projectile (not shown). For each dissociation channel, we have measured the total count and the center of the fitted Gaussian function defined as the parent ion mean excitation energy. The data were classified according to the well-defined multiplicity, triplet states for F+ impact (Table II), and singlet states for H+ impact (Table III). Taking into account of the statistics and calibration procedure, the error bars for the excitation energies have been estimated to be ±0.3 eV and ±0.5 eV for F+ and H+ projectiles, respectively. In the case of H2O2+, the measured mean parent ion excitation energies corresponding to the dissociation channels (OH+_H+) and (H+_H+_O) can

be compared with previous experiments obtained using DCT spectroscopy. For H+ impact, our measured values 41.6 and 44.2 eV are in good agreement with the two components 41.4 and 45.5 eV measured by Richardson et al.3 For F+ impact, Severs et al.7 have observed three components, two main peaks at 40.0 and 43.2 eV, and one weaker peak at 45.5 eV. Our measured values 40.8 and 44.1 eV are comparable to the two main peaks of the previous authors with a slight shift to lower energies, while their third peak was not resolved in our measurement. In a more recent theoretical work, B. Gervais et al.1 have calculated the energy levels and fragmentation schemes of doubly charged water molecules including HDO using ab initio multi-reference configuration interaction method. By comparing our measured energy values with the calculations, correspondences between the observed fragmentation

TABLE II. For each dissociation channel in F+ impact experiment: measured mean excitation energy of the HDO2+, H2O2+ or D2O2+ parent ions and the count number. Theoretical data for doubly charged triplet states. Values with asterik: estimated from the data of Ref. 19. Experiment (F+) Dissociation channel

Theory1

Energy (eV) ±0.3 eV

Count

OD+_H+ OD+_D+ OH+_H+ OH+_D+ O+_H+_(H or D) H+_D+_(O) H+_H+_(O) D+_D+_(O)

40.7 40.5 40.8 40.4 43.8 44.6 44.1 44.6

11 500 8 525 7 450 3 020 280 8 290 7 080 10 020

O+_H+_(H or D)

62.4

90

Electronic state

Energy (eV) (H2O2+)

3-body part (HDO2+)

(3a1)−1 (1b1)−1 3B1

40.33

4%

(1b2)−1 (1b1)−1 3A2

44.32

99%

(2a1)−1 (1b1)−1 3B1 (2a1)−1 (3a1)−1 3A1

61∗ 62∗

094306-7

Martin et al.

J. Chem. Phys. 142, 094306 (2015)

TABLE III. For each dissociation channel in H+ impact experiment: measured mean excitation energy of the HDO2+, H2O2+, or D2O2+ parent ions and the count number. For singlet states: theoretical energy level and the 3-body part predicted in the dissociation of HDO2+. This experiment (H+) Dissociation channel

Energy (eV) ±0.5 eV

Count

OH+_D+ OD+_D+ OH+_H+ OD+_H+ H+_D+_(O) H+_H+_(O)

41.3 41.4 41.6 42.2 45.6 44.2

2 920 7 650 8 700 13 100 11 410 11 720

channels and three groups of electronic states of the doubly charged water molecules can be established as shown in Tables II and III. In the case of F+ impact, for two-body dissociations, the mean excitation energy Eexc was measured to be in the range from 40.4 to 40.8 eV (Table II). Taking into account of the error bars, these values are in good agreement with the ground triplet energy level 40.33 eV of (3a1)−1 (1b1)−1 3 B1. The isotopic ratio measured in collisions with F+ should be attributed to the dissociation of the triplet ground state of HDO2+. However, the measured value of isotopic ratio 3.8 is much smaller than the theoretical prediction,1 8.6 for (3a1)−1 (1b1)−1 3B1. It should be noticed that the theoretical isotopic ratio was calculated for molecules initially in the ground vibrational and rotational state, while in the experiment the effusive molecular jet was formed at the room temperature. The discrepancy between the measured and theoretical values suggests probably that the initial vibrational and rotational states of neutral HDO play an important role in the dissociation dynamics of the doubly ionized molecules. In the case of H+ impact, two-body dissociation channels were observed in the energy range from 41.3 to 42.2 eV. This energy region corresponds well to the energy levels of the first singlet states, 41.37 eV for (1b1)−2 1A1 and 42.84 eV for (3a1)−1 (1b1)−1 1B1. The theoretical isotopic ratios were found to be 3 and 15.7 for the lower and the higher energy states (1A1 and 1B1), respectively. Similar as in the case of the triplet state, the isotopic ratio for each state is expected to be modified due to the initial vibrational and rotational states of neutral HDO. Furthermore, under the present experimental conditions, the two singlet states 1A1 and 1 B1 are not resolved. The experimental value of 4.5 should result from the combination of the contributions of these two states. From the Table III, one can notice that the mean energy of the channel OD+_H+ was measured to be about 42.2 eV and that of OH+_D+ to be about 41.3 eV. A slight shift to higher energy for the channel OD+_H+ is therefore observed comparing to OH+_D+. This is possible when both 1A1 and 1 B1 states are populated and the isotopic ratio for the state 1 B1 is much larger than for the state 1A1. This is in qualitative agreement with the theoretical prediction. For both multiplicities, the two-body dissociation channels can be related to the group of states corresponding to the lowest electronic configurations, (3a1)−1 (1b1)−1 and (1b1)−2. Furthermore, in the low energy region from 40 to 42 eV, only a

Theory1 Energy (eV) (H2O2+)

3-body part (HDO2+)

(1b1)−2 1A1

41.37

2%

(3a1)−1 (1b1)−1 1B1

42.84

6%

(3a1)−2 1A1 (1b2)−1 (1b1)−1 1A2

46.03 46.02

32% 99%

Electronic state

small number of population undergoes three-body dissociation as shown in the Figure 7, where a slight shoulder around 40 eV can be noticed for the projection of the spot D+_H+. Therefore, we confirm that in this energy region, molecules dissociate mainly into two fragments in agreement with the theoretical modeling, where the three-body dissociation part was given to be 4% for (3a1)−1 (1b1)−1 3B1 the triplet state (Table II) and 2% ((1b1)−2 1A1) and 6% ((3a1)−1 (1b1)−1 1B1) for the two singlet states (Table III). Three-body fragmentation channels become dominant in a slightly higher energy region. With F+, the main channels involving two small charged fragments and a neutral O were found in the range from 44.1 eV for (H+_H+_O) to 44.6 eV for (H+_D+_O) (Table II). The minor channel with an oxygen ion O+ representing about 1/100 of the other one was found around 43.8 eV. These values are near the triplet state (1b2)−1 (1b1)−1 3 A2 at 44.3 eV. In agreement with our observation, this state was predicted to dissociate mainly by three-body dissociation (99%).1 With proton H+, three-body channels with neutral oxygen were measured to be around 44.2 eV (H+_H+_O) and 45.6 eV (H+_D+_O) (Table III). In the case of HDO2+, the measured mean energy (45.6 eV) is comparable to the theoretical energy (46.0 eV) of (3a1)−2 1A1 and (1b2)−1 (1b1)−1 1 A2 for which the three-body dissociation part was estimated to be about 32% and 99%, respectively. Around this energy, the three-body channel was observed largely dominant comparing to the two-body dissociation channels. The slight shoulder at the high-energy side of the OD+_H+ peak (Fig. 7(b)) could be explained by a weak population of the state (3a1)−2 1A1. A part of the population of this state (32%) dissociates into three fragments leading to a relatively small contribution to the measured three-body channel. So, we attribute the observed (H+_D+_O) fragmentation channel mainly to the population of the state (1b2)−1 (1b1)−1 1A2. By comparing our measurements and the theoretical values, we can see that the three-body dissociation channels are mainly related to the population of the configuration (1b2)−1 (1b1)−1 at both the triplet 3A2 and singlet 1A2 states in the energy range from 44 to 46 eV. Three-body fragmentation channels are also observed in a much higher energy range, characterized by the high-energy component of O+ centered at Eexc = 62.4 eV in F+ impact experiment. This value is comparable to the energy levels of states (2a1)−1 (1b1)−1 3B1 and (2a1)−1 (3a1)−1 3A1 estimated from Auger electron spectroscopy.19 Indeed, using the more

094306-8

Martin et al.

recent ground state energy level, 40.33 eV for (3a1)−1 (1b1)−1 3 B1,1 these excited states are estimated, respectively, to 61 and 62 eV. The energy gap between this channel and the other dissociation channels is therefore related to the electronic structure of the doubly charged ions. In fact, in the energy region around 40 eV, the two electrons are captured from the outermost orbitals (1b1)2 and (3a1)2 which are composed of, respectively, 100% and 72% of O (2p). The capture of one of the electrons from the inner orbital (2a1)2, a mainly O (2s) orbital, leads to an energy jump to higher values of about 20 eV. Although this extra amount of energy is larger than the ionization energy of a hydrogen atom, it is not sufficient for a third electron to escape from the molecule via autoionization due to the strong coulomb attraction field of the remaining triply charged molecule. The peak O+ as well as H+ or D+ in this high energy region is interpreted as the results of three-body dissociation with two charged fragments and a small neutral. The absence of triple-hit events which might correspond to O+_H+_H+ or O+_D+_H+ or O+_D+_D+ confirmed this attribution. The energy level of the first singlet state with a 2a1 vacancy (2a1 1b1)−1 1B1 is about 8 eV19 higher than the triplet state (2a1)−1 (1b1)−1 3B1. It is too high to be populated in collisions with H+.

V. SUMMARY

We have studied the fragmentation of doubly charged HDO2+, H2O2+, and D2O2+ molecular ions produced by double charge transfer in collisions with proton and fluorine projectiles using the CIDEC method. Excitation energies of singlet and triplet states of parent ions with respect to the neutral ground state have been measured for different dissociation channels. Although the energy resolution using the CIDEC method is lower than the DCT spectroscopy, the fragmentation schemes of different electronic states were observed and analyzed selectively in the present work. In agreement with the theoretical predictions, the ground and the first excited configurations (3a1)−1 (1b1)−1 3B1, 1B1 and (1b1)−2 1A1 were found to decay preferentially via two-body dissociation and the excited configuration (1b2)−1 (1b1)−1, 3A2 and 1A2 decay mainly via three-body dissociations including a neutral O. Higher excited triplet states (2a1)−1 (1b1)−1 3B1 and (2a1)−1 (3a1)−1 3A1 populated via the capture of one of the two electrons from the inner orbital (2a1) were observed to undergo three-body dissociation involving O+, a small charged fragment and a small neutral. For HDO2+, we have confirmed that DO+_H+ is the dominant two-body dissociation channel. However, the comparison of the measured isotopic ratios with the theoretical predictions is not straightforward. It depends not only on the multiplicity but also on the precise electronic state populated in the experiment. In F+ impact case, only the ground state (3a1)−1 (1b1)−1 3B1 is involved in the twobody dissociation process; however, a pronounced difference

J. Chem. Phys. 142, 094306 (2015)

between the measured isotopic ratio (3.8 ± 0.8) and the predicted value (8.6) was observed. This discrepancy suggests that the initial vibrational and rotational states of neutral HDO may play also a role in the dissociation dynamics of the doubly ionized molecules. The good agreement between the measured excitation energy and the calculated electronic energy levels leads to the conclusion that in two electron transfer process using H+ and F+ projectiles, the molecules are prepared to electronic excited states by electron capture directly from inner orbitals without vibrational energy deposition. This is further confirmed by KER measurements. Indeed, the KER measured for both two and three body dissociation channels are quite comparable to values obtained in photon ionization experiments where doubly charged molecules were prepared uniquely via electronic excitation. ACKNOWLEDGMENTS

We acknowledge financial supports by the ANR “ANNEAU” 2010-042601 program and B. Gervais for stimulating discussion. 1B.

Gervais, E. Giglio, L. Adoui, A. Cassimi, D. Duflot, and M. E. Galassi, J. Chem. Phys. 131, 024302 (2009). 2N. Neumann, D. Hant, L. P. H. Schmidt, J. Titze, T. Jahnke, A. Czasch, M. S. Schoeffler, K. Kreidi, O. Jagutzki, H. Schmidt-Boecking, and R. Doerner, Phys. Rev. Lett. 104, 103201 (2010). 3P. Richardson, J. Eland, P. Fournier, and D. Cooper, J. Chem. Phys. 84, 3189 (1986). 4S. W. J. Scully, J. A. Wyer, V. Senthil, M. B. Shah, and E. C. Montenegro, Phys. Rev. A 73, 040701 (2006). 5K. Tan, C. Brion, P. Vanderleeuw, and M. Vanderwiel, Chem. Phys. 29, 299 (1978). 6H. Luna, A. L. F. de Barros, J. A. Wyer, S. W. J. Scully, J. Lecointre, P. M. Y. Garcia, G. M. Sigaud, A. C. F. Santos, V. Senthil, M. B. Shah, C. J. Latimer, and E. C. Montenegro, Phys. Rev. A 75, 042711 (2007). 7J. Severs, F. Harris, S. Andrews, and D. Parry, Chem. Phys. 175, 467 (1993). 8S. Legendre, E. Giglio, M. Tarisien, A. Cassimi, B. Gervais, and L. Adoui, J. Phys. B: At., Mol. Opt. Phys. 38, L233 (2005). 9A. M. Sayler, J. W. Maseberg, D. Hathiramani, K. D. Carnes, and I. BenItzhak, AIP Conf. Proc. 680, 48 (2003). 10S. Martin, L. Chen, A. Salmoun, B. Li, J. Bernard, and R. Brédy, Phys. Rev. A 77, 043201 (2008). 11F. M. Harris, Int. J. Mass Spectrom. Ion Processes 120, 1 (1992). 12O. Furuhashi, T. Kinugawa, S. Masuda, C. Yamada, and S. Ohtani, Chem. Phys. Lett. 337, 97 (2001). 13R. P. Grant, F. M. Harris, S. R. Andrews, and D. E. Parry, Int. J. Mass Spectrom. Ion Processes 142, 117 (1995). 14A. Denis, M. C. Buchet-Poulizac, J. Bernard, L. Chen, S. Martin, and J. Désesquelles, Phys. Scr. 61, 431 (2000). 15S. Martin, L. Chen, R. Bredy, G. Montagne, C. Ortega, T. Schlatholter, G. Reitsma, and J. Bernard, Phys. Rev. A 85, 052715 (2012). 16L. Chen, S. Martin, J. Bernard, and R. Brédy, Phys. Rev. Lett. 98, 193401 (2007). 17C. E. Moore, “NIST: Atomic Spectra Database: Tables of Spectra of Hydrogen, Carbon, Nitrogen, and Oxygen Atoms and Ions,” in CRC Series in Evaluated Data in Atomic Physics, edited by J. W. Gallagher (CRC Press, Boca Raton, FL, 1993). 18T. Muranaka, “Dynamique de la fragmentation de molécules tri-atomiques,” Ph.D. thesis, Université de Caen/Basse-Normandie, 2007. 19H. Siegbahn, L. Asplund, and P. Kelfve, Chem. Phys. Lett. 35, 330 (1975).