Infrared Spectroscopy of a Small Ion Solvated by Helium: OH Stretching Region of He N -HOCO +

Messenger spectroscopy is a well-established method for recording infrared (IR) spectra of molecular ions. It relies upon the tagging of weakly bound atoms or molecules, known as the ‘messenger’, to the ion of interest. The ideal tag species is helium since it has the weakest possible interaction with any molecular ion and is consequently the least likely to alter their structure and function. However the attachment of a helium tag is challenging because of the exceptionally cold conditions that are inherently required. In this work, electron ionization of doped liquid helium nanodroplets has been used to create cations tagged with a variable number ( N ) of helium atoms. Mass-selective ion detection has made it possible to record IR spectra as a function of N , thus revealing the effect on the structure and charge distribution within the ionic core as solvation becomes more extensive. We illustrate this capability for protonated carbon dioxide tagged with up to 14 helium atoms, He N -HOCO + . The first atom preferentially binds to the proton and results in a substantial red-shift of 44 cm -1 for the OH stretching vibration, whilst the stepwise attachment of additional atoms up to N = 7 causes small and progressive blue-shifts, which are attributed to the gradual formation of a ring of helium around the carbon atom. The methodology described herein offers a new route to obtain IR spectra of He-tagged ions and provides an insight into ion-solvent interactions at the molecular level.


I. INTRODUCTION
The spectroscopic study of neutral molecules and clusters embedded in liquid helium nanodroplets is well established. [1][2][3] In a typical experiment, the dopant is added by flowing a beam of helium nanodroplets through a region of low pressure gas or vapor. Absorption spectra can then be recorded using a form of action spectroscopy, commonly referred to as depletion spectroscopy, which utilizes mass-selective ion detection. 4 In contrast to the extensive work on neutral species, far fewer spectroscopic studies have been carried out on charged species formed from doped helium nanodroplets. Of these, the majority have focused on ions that reside inside the droplets, which can be created by photoionization of the neutral dopant, as pioneered by the Drabbels group, [5][6][7] or alternatively by external production of ions followed by pickup by the droplets, as demonstrated by von Helden and co-workers. [8][9][10][11] Recently, Scheier and co-workers used electron ionization of doped helium nanodroplets to generate He-tagged C60 + cations, which were subsequently probed using electronic spectroscopy. [12][13][14] In the electron ionization process, the released energy strips helium from the cation but some molecular ions emerge with helium atoms still attached.
Photodissociation spectra were then measured mass-selectively, enabling the electronic spectrum of the ion to be recorded as a function of the number of helium atom tags. Since the ions must be cold in order to retain the weakly-bound tags, the result is a form of cold ion electronic predissociation spectroscopy.
In a recent study, we demonstrated that vibrational predissociation spectra can also be measured for He-tagged cations prepared from helium nanodroplets. In that case, we measured IR spectra for the protonated acetic acid monomer tagged with one and three helium atoms.
Our ab initio calculations predicted that an attached helium atom will shift the OH stretching vibration by <5 cm -1 relative to the bare ion, and ultimately this enabled the identification of two different structural isomers of the molecular ion. 15 Surprisingly, these were the first IR spectra to be measured for either conformer. The success of that study can be attributed to the This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset.

PLEASE CITE THIS ARTICLE AS DOI:10.1063/1.5124137
4 rapid cooling of neutral dopants to an equilibrium temperature of 0.37 K inside the nanodroplets, which led to the trapping of metastable dimer structures and, following electron ionization, this generated cold samples of the two ionic conformers tagged with helium.
There are several alternative and well-established methods that can be used to prepare ions tagged with a messenger species, the most common of which are supersonic plasma expansions and cold ion traps. These techniques are well-suited for preparing ions tagged with rare gas atoms or diatomics such as Ne, Ar and H2, but tagging with helium atoms is significantly more challenging because of the lower temperatures required to attach this weakly interacting species. Nevertheless, several groups have been successful in this endeavor and have reported IR spectra for He-tagged ions. [16][17][18][19][20][21][22] Of particular interest are test cases demonstrating that, unlike other tags, helium is a relatively innocent bystander that typically does not change the structure of the ion or the IR spectrum in any substantive way. 18,23,24 As such, helium is considered the ideal messenger species but the established techniques are limited to the attachment of only a few helium atoms. 24,25 In contrast, electron ionization of doped helium nanodroplets can be used to prepare ions tagged with between one and 100 helium atoms, or even more, which allows for a bottom-up approach to studying helium solvation effects. [12][13][14] Protonated carbon dioxide, HOCO + , is an interesting example of a He-tagged ion because it exhibits a substantial red-shift in the OH stretching fundamental from 3375.4 to 3328.3 cm -1 upon attachment of one helium atom. 26,27 This arises because, just like helium, CO2 has a low proton affinity and therefore the two species compete for the same proton. As a result, the attached helium atom subtly alters the OH bond and hence its vibrational frequency.
Because of this sensitivity to the local helium environment, HOCO + is a good candidate for probing more extensive helium solvation effects using IR spectroscopy, and this can be achieved by measuring the vibrational frequency shift as a function of the number of attached helium atoms. A similar protocol has previously been used to study helium solvation of small This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset. PLEASE CITE THIS ARTICLE AS DOI:10.1063/1.5124137 5 neutral molecules, such as OCS, CO2 and N2O. [28][29][30][31] However, an advantage of using ions is that the vibrational frequency shifts will typically be larger and more easily discerned because of the stronger electrostatic interactions with the tag.
In this work, we demonstrate that IR spectra of ions tagged with multiple helium atoms can be obtained using electron ionization of doped helium nanodroplets combined with vibrational predissociation spectroscopy. We describe the methodology used and illustrate the technique for HOCO + tagged with up to 14 helium atoms.

A. Experimental
The helium nanodroplet apparatus has been described previously. [32][33][34] A mean droplet size of ~5000 helium atoms was achieved using a nozzle temperature of 16 K and a helium pressure of 32 bar. The nanodroplet beam was skimmed and passed through a pickup cell to add the dopant. In this case, acetic acid was used as the neutral precursor and a partial pressure of 6 × 10 -6 mbar attained an average pickup of one dopant molecule per nanodroplet. After travelling along a 60 cm flight path, the nanodroplet beam entered a quadrupole mass spectrometer (Extrel, MAX-1000) with an electron ionization source operating at an energy of 90 eV. The ionization process generated a range of ions, including HOCO + , both with and without helium atoms attached. Spectra were recorded by directing the IR beam from a pulsed OPO/OPA laser (LaserVision, Belleview, WA) through the vacuum chamber in a counter-propagating direction relative to the helium nanodroplet beam. The laser beam path was chosen to provide optimal alignment with those droplets located between the electron filament and the center of the deflector that feeds ions into the quadrupole filter of the mass spectrometer, i.e. a short region of roughly 3 cm in length. We note that this alignment is different from that typically used to measure IR spectra for neutral dopants, for which the optimum configuration has the IR beam This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset. PLEASE CITE THIS ARTICLE AS DOI:10.1063/1.5124137 6 fully aligned along the entire helium droplet beam path. The OPO/OPA, which is pumped by an unseeded pulsed Nd:YAG laser (Continuum Surelite II-10), has a pulse energy of ~10 mJ and a linewidth of ~4 cm -1 . The wavelength was calibrated using a MIR polystyrene reference film (PerkinElmer). Time-dependent measurements of the mass-selected ion signal enabled discrimination of those ions arriving between 100 and 300 s after the firing of the Nd:YAG laser. The resulting signal was accumulated by pulse counting electronics and fed to a PC for subsequent data processing. This, in combination with computer-controlled scanning of the output wavelength of the OPO/OPA, was used to record IR spectra of He-tagged HOCO + .

B. Computational details
Ab initio calculations were performed at the MP2/aug-cc-pVTZ level of theory for HeN-HOCO + for N = 0 to 7 using Gaussian 16 software. 35 Possible binding sites for the first helium atom were identified using a geometry search procedure in which 26 initial geometries were generated by placing the helium atom at selected locations on the surface of a 8 × 8 × 8 Å cube centered on the cation. These locations are the 8 vertices, the 6 face centers and the midpoints of the 12 edges of the cube. The identified structures were then re-optimized using a 'verytight' convergence criterion. Harmonic vibrational frequency calculations were used to test whether the structures corresponded to true energy minima, after which corrections for the zero point energy (ZPE) and basis-set superposition error (BSSE) were made. Subsequently, for each of the three lowest energy structures for N = 1, an additional helium atom was placed at 26 different locations on the surface of a 8 × 8 × 8 Å cube, which generated a total of 78 initial geometries for the N = 2 search procedure. Each optimized structure was tested as described above, and the procedure was repeated for all values of N up to N = 7.
The harmonic vibrational frequencies for the OH stretch were empirically scaled by a factor of 0.9556 to allow comparison with the experimental spectra. This value is derived by scaling the computed value for untagged HOCO + so that it equals the known experimental This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset. value of 3375.4 cm -1 . 26 Anharmonic vibrational frequency calculations were also performed for N = 1 to 4 by using generalized second-order vibrational perturbation theory (GVPT2).

A. Mass spectrometry
A mass spectrum is presented in Fig. 1, which was obtained using electron ionization of helium nanodroplets doped with an average of one acetic acid molecule per droplet. The statistical nature of the pickup process means that some droplets will contain more than one molecule This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset.

B. IR spectroscopy by detecting bare HOCO +
The IR depletion spectra shown in Fig. 2 were measured by detecting wavelength-dependent changes in the mass-selected ion signal at m/z 45 (HOCO + ) under a variety of experimental conditions. A decrease in nozzle temperature from 16 K to 12 K leads to an increase in droplet size, whilst a reduction in the data acquisition gate delay from 600 s to 100 s favors the detection of species located close to the mass spectrometer when the laser pulse is triggered, which is the region where ions are produced. Both of these criteria have a considerable influence on the relative intensities of the IR peaks at 3337 and 3585 cm -1 . This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset.

PLEASE CITE THIS ARTICLE AS DOI:10.1063/1.5124137
In Fig. 2(a), the strong positive-going peak at 3585 cm -1 is assigned to the OH stretching mode of the neutral acetic acid monomer. The other positive-going peak at 3292 cm -1 , which is weak, corresponds to the stretching mode of a hydrogen-bonded OH group in a metastable neutral acetic acid dimer. 36 A slightly broader peak is observed at 3337 cm -1 that exhibits a negative depletion owing to an enhancement in ion signal at m/z 45. This clearly differentiates it from the positive-going peaks assigned to IR absorption by neutral dopants and we tentatively assign it to the OH stretching mode in HOCO + . The known frequency of this vibration for the bare ion is 3375.4 cm -1 , 26 whilst the value for He-HOCO + is 3328 cm -1 . 27 This suggests that the experimental peak observed at 3337 cm -1 arises from IR absorption by HOCO + tagged with one or more helium atoms originating from the nanodroplet and that these helium atoms redshift the OH stretching frequency (see later).
To further investigate the origin of the negative-going peak at 3337 cm -1 , it is desirable to measure IR spectra at the parent mass channels corresponding to HeN-HOCO + . The mass spectrum in Fig. 1 shows little evidence of ion signal originating from He-tagged HOCO + but a vertically expanded view in Fig. 3 allows us to identify such peaks for HOCO + complexed with between 1 and 14 helium atoms. These peaks, seen at m/z 49, 53, 57, etc., have intensities approximately two orders of magnitude lower than for the bare HOCO + ion at m/z 45. Despite the low ion yields, it has been possible to measure their IR spectra by scanning the IR wavelength while monitoring the ion signal intensity.
In Fig. 3, labels are also shown for those mass peaks that can contain ion signal from He-tagged protonated acetic acid, HeM-(CH3COOH)H + . These peaks lie at m/z 65, 69, 73, etc, which are at the same position as HeN-HOCO + for N ≥ 5. Due to the limited resolution of our mass spectrometer, we are unable to resolve these two distinct species in the mass spectrum.
However, we showed in a previous publication that He-tagged protonated acetic acid and HOCO + have very different IR spectra in the OH stretching region, 15 enabling the two ions to be readily distinguished (see Fig. S1).
This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset. This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset. This arises because the total binding energy of all of the helium atoms is lower than the photon energy, as shown by the ab initio calculations discussed in Section III E. In comparison, the depletion in HeN-HOCO + signal for each selected value of N represents IR absorption by that species only. The dominant feature observed at ~3330 cm -1 in Fig. 4(b) is assigned to the OH stretching mode in He-HOCO + , i.e., N = 1. This is confirmed by a comparison with the experimental and calculated values of 3328.3 and 3317.8 cm -1 obtained by Dopfer et al. 27 Simulations of the rotational profile have been performed using PGOPHER software 37,38 and This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset.

PLEASE CITE THIS ARTICLE AS DOI:10.1063/1.5124137
compared with the experimental band shape (see Fig. S2(a), Table S1). The agreement is rather poor, which suggests a non-equilibrium thermal distribution. This is not surprising given that ion cooling in helium nanodroplets is known to be highly nonthermal. 39,40 Nevertheless, a 'temperature' of between 20 and 60 K does reproduce some of the observed band broadening.
It is also possible that other processes contribute significantly to the broadening, such as large amplitude motion at the zero-point energy, particularly for the intermolecular bending and stretching vibrations. Further insight could be obtained by modelling this effect using path integral molecular dynamics simulations. An unusual feature in the N = 1 spectrum [see Fig. 4(b)] is a secondary absorption extending from ~3340 to ~3400 cm -1 , which is absent from the N = 2 and 3 spectra. This broad feature must therefore arise only from the binding of a single helium atom to HOCO + .
Interestingly, the earlier work on He-HOCO + by Dopfer et al., which utilized a pulsed supersonic expansion ion source, obtained a somewhat different spectral profile in the region immediately to the blue of the main feature at 3328 cm -1 . 27 A tail extending to ~3358 cm -1 was observed and attributed, at least in part, to rotational band structure arising from a contraction of the intermolecular bond upon excitation of the OH stretch fundamental. A weak peak at ~3364 cm -1 was also observed but not assigned. Even though there is some evidence of these features in the present work, they are somewhat dwarfed by the aforementioned broad band centered at ~3370 cm -1 .
The difference between our spectral profile and that of Dopfer et al. 27 might arise from the different methods used for generating the tagged ions. In gas-phase experiments that use a This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset. A similar observation is well documented for neutral clusters embedded in helium nanodroplets. 36,[41][42][43] To further explore the source of the broad feature at ~3370 cm -1 , we have performed ab initio calculations at the MP2/aug-cc-pVTZ level of theory for HOCO + , both with and without an attached helium atom. A geometry search procedure was used to find prospective stable structures for He-HOCO + and this led to the identification of three possible locations for the helium atom, as illustrated in Fig. 5, where the lowest energy structure (1A) is the proton-bound conformer previously identified by Dopfer et al. 27 The BSSE-and ZPE-corrected relative energies for the three calculated structures are 0.0, 1.28 and 1.88 kJ mol -1 , respectively, whilst scaled harmonic vibrational frequency calculations predict values of 3320, 3377 and 3376 cm -1 for the OH stretching fundamental (see Table S2). Comparison with the experimental spectrum shown in Fig. 4(b) therefore confirms that the most intense feature centered at ~3330 cm -1 derives from structure 1A, in which the helium atom is proton-bound, whilst the broad feature at ~3370 cm -1 , most likely derives from structures 1B and 1C.

PLEASE CITE THIS ARTICLE AS
This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset.  Table S1 in the Supplementary Material] are able to account for the spectral region between ~3360 to ~3380 cm -1 , as well as the spectral feature centered at 3397 cm -1 , whereby the latter is a close match to a rotational sideband in structure 1C. For the spectral region between 3340 and 3360 cm -1 , a possible assignment is to vibrational hot bands associated with structure 1A. Our anharmonic calculations predict that these will occur at 3343 and 3355 cm -1 [see Table S3 and Sec. I(ii) in the Supplementary Material].

PLEASE CITE THIS ARTICLE AS
However, the bands might be broadened due to the large amplitude motion of the thermally populated intermolecular bending vibration. We therefore propose that several different factors This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset.

PLEASE CITE THIS ARTICLE AS DOI:10.1063/1.5124137
contribute to the broad feature, including metastable isomers and vibrational hot bands. This conclusion is similar to that proposed by Bieske and co-workers to explain comparable broad features in ArN-HCO + , particularly for N = 3. 44 Delocalization of the helium atom and large amplitude motion at the zero-point energy for metastable structures 1B and 1C may also contribute to the band broadening.
In Fig. 4, a relatively weak feature is observed at ~3426 cm -1 in the spectra measured for N = 1 and 2, and which is just about discernible above the noise for N = 3. This feature is blue-shifted by ~96 cm -1 relative to the OH stretching fundamental and has previously been observed for N = 1 by Dopfer et al. 27 who tentatively assigned it to the ν1 + νs combination band, where ν1 is the OH stretching mode and νs is the low frequency He-H + intermolecular stretching mode. A comparable combination band was observed in the spectrum of Ar-tagged HOCO + , 27,45 but was barely perceptible for Ne-tagged HOCO + . 27 Harmonic calculations for the Ar-and Ne-tagged ions strongly supported the assignments to ν1 + νs; however equivalent calculations for He-tagged HOCO + showed poor agreement with experiment. As a result, Dopfer et al. 27 stated that an alternative assignment to an intramolecular combination band, such as ν2 + ν3 or ν2 + ν4, could not be ruled out. Selected results are presented in Table I. This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset. The calculated IR intensities for the ν1 + νs combination band are at least three orders of magnitude smaller than for the ν1 fundamental band, making the assignment of this transition to the 3426 cm -1 band seem unlikely. Calculated intensities for the ν2 + ν3 combination band are significantly higher but the calculated frequencies are >160 cm -1 to the blue of the experimental band at 3426 cm -1 . Instead, the anharmonic calculations indicate that the most likely assignment is to the ν2 + ν4 combination band, which has an even higher predicted intensity and a frequency lying within 40 cm -1 of the experimental value. We further note that the calculated frequency for the ν1 fundamental in HOCO + (N = 0) is 24 cm -1 lower than the known experimental value of 3375.4 cm -1 , indicating that the anharmonic calculations tend to underestimate the frequencies. If this is accounted for by blue-shifting the ν2 + ν4 band frequency by 24 cm -1 for N = 1 to 3 then the difference between experiment and calculations is reduced to ≤4 cm -1 , providing additional support for this assignment.

PLEASE CITE THIS ARTICLE AS
As mentioned earlier, the ν1 + νs combination band was previously observed in experimental IR spectra for Ar-and Ne-tagged HOCO + and it was presumed that the equivalent band should also be observed for He-HOCO + . To address this point we have performed anharmonic calculations for Ar-and Ne-tagged HOCO + and we have determined IR intensities of 313 and 15 km mol -1 , respectively, for the ν1 + νs band (see Table S4). These intensities are several orders of magnitude larger than predicted for the He-tagged ion (0.5 km mol -1 ), thus discounting the above presumption. However, we note that the GVPT2 method is better suited to small, rigid molecules than weakly bound complexes, and so the calculated structure deriving from intermolecular vibrations should be treated with some caution. Thus, while we favor an assignment to the ν2 + ν4 combination band, the ν1 + νs transition cannot be completely ruled out.
This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset. This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset.

PLEASE CITE THIS ARTICLE AS DOI:10.1063/1.5124137
To gain further insight into the band shift with increasing N, we have extended our ab initio calculations for HeN-HOCO + up to a value of N = 7. The local minima structures identified in this work are shown in Figs. S6 to S11, whilst calculated energies and vibrational frequencies are given in Tables S5 to S10. We note that such calculations will be indicative rather than exact because the weakly-bound complexes are likely to be very floppy owing to relatively flat potential energy surfaces, which means that nuclear quantum dynamics will be important for the helium atoms.
We have found that the number of local energy minima identified using our geometry search procedure increases dramatically from 9 at N = 4 to a total of 18 at N = 5, 6 and 7. Many of these structural isomers are near-isoenergetic and, as such, any helium atoms in locations with similar binding energies are likely to be delocalized. 46,47 This result confirms our assertion that it is not entirely realistic to assign a particular isomeric structure to the IR spectrum measured for each value of N, particularly for higher N. However it is still a useful exercise to calculate the trend in OH stretch frequencies as a function of N and to compare this trend with the experimental results. The most straightforward way to achieve this is by first considering the calculated global minimum structures.
The minimum energy structure for N = 1 was shown in Fig. 5(a), whilst those for N = 2 to 7 are presented in Fig. 7. For N = 1, the helium atom is located at a distance of just 1.69 Å from the proton but far from the oxygen and carbon atoms. The addition of a second helium through to the sixth helium atom leads to the stepwise formation of a five-membered ring around the carbon atom, whilst the first helium atom remains in an almost fixed location at ~1.71 Å from the proton. The distance between the carbon atom and helium ring atoms is ~2.8 Å and the ring lies in a plane roughly perpendicular to the O-C-O axis. The formation of a ring of helium atoms is reminiscent of previous IR spectroscopy studies of small neutral molecules solvated by helium, such as OCS, CO2 and N2O. [28][29][30][31] This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset. The seventh helium atom preferentially binds to the proton end of the ion, causing the helium atom closest to the proton to shift slightly, which results in an increased He-H + separation of 1.73 Å. As such, the seventh helium atom appears to play a role in connecting the first added helium atom with the ring atoms, perhaps beginning the process of forming a complete solvation shell around the cation. corrected energy of this structure is 0.40 kJ mol -1 higher than for the five-membered ring conformer shown in Fig. 8(c). This is attributed to the stiffness of the 6-member ring, which results in higher vibrational frequencies for various intermolecular modes and the OH torsional mode.  Fig. 9(a), there is good agreement between This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset. This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset.

PLEASE CITE THIS ARTICLE AS DOI:10.1063/1.5124137
In Fig. 9(b), an expanded view shows the subtle shifts in the experimental OH stretching frequency observed for 1 ≤ N ≤ 10. In this region, the largest shifts occur between N = 3 and 7.
A linear fit to the experimental data within this range indicates a blue-shift of 1.44 cm -1 for each added helium atom, compared with a predicted blue-shift of 1.90 cm -1 per atom from the scaled harmonic frequency calculations. However, it is important to note that the calculated results presented in Fig. 9 are for the global minimum structures only. Therefore, in a rudimentary attempt to take into account the delocalization of the helium atoms, we have considered a range of local energy minima structures that give alternative locations of the helium atoms relative to the ion. By considering the average value of the OH stretching frequencies for conformers with energies < 0.5 kJ mol -1 above the global minimum structure for each value of N (see Tables S2 and S5 to S10), we have calculated a blue-shift of 1.56 cm -1 per helium atom, in closer agreement with the experimental results.
In Table II, Table S11 and Fig. S12). Upon attachment of further helium atoms, a decrease in all intramolecular bond lengths and the COH angle is generally observed, along with an increase in the OCO angle. These structural changes in the ion core directly report on the ion-solvent interactions. Furthermore, our calculations indicate that the OH stretching vibration shows a larger N-dependent frequency shift than any other intramolecular fundamental mode (see Table S12 and Fig. S13) and therefore it is a more sensitive probe of helium solvation.
This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset.  The BSSE and ZPE corrected energy for removing a single helium atom, D0, decreases from a maximum value of 154 cm -1 at N = 1 to just 54 cm -1 at N = 7 (see Table II), and it is anticipated to be even smaller at higher values of N. These low D0 values support our earlier hypothesis that absorption of an IR photon at ~3330 cm -1 will strip all of the helium atoms from the ion for the values of N ≤ 14 investigated in this work, and most likely up to N = 70 or more.

PLEASE CITE THIS ARTICLE AS
The partial atomic charges in HOCO + have been calculated for each global minimum structure of HeN-HOCO + , where N = 1 to 7. These values correspond to the electrostatic potential (ESP) derived charges and were determined using the CHELPG scheme. 48 Even though bare HOCO + is often referred to as protonated CO2, our calculations show that the partial charge on the carbon atom is approximately double that on the hydrogen atom.
Furthermore, there is a strong correlation between the location of the positive partial charges and the preferred solvent binding sites, i.e., at the hydrogen and carbon atoms. This confirms that the dominant ion-solvent interactions are electrostatic in nature.
The partial charges show small but clear trends with N (see Table S13 and Fig. S14).
Attachment of the first helium atom leads to a decrease in the positive charge on the hydrogen This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset.
atom from 0.49e to 0.40e and an induced charge of 0.06e on the helium atom. This coincides with a lengthening of the OH bond and a redistribution of partial charge within the ion, including an increased positive charge on the carbon atom. Another notable change occurs when the sixth helium atom is added, resulting in a decrease in partial charge on the carbon atom from 0.83e to 0.77e and an increase in the total charge on the helium ring atoms from 0.04e to 0.07e. The significance of the sixth helium atom is that it completes the formation of a 5-membered ring around the carbon atom, providing a partial electrostatic shield between the positive charge on the carbon and any additional helium atoms that may subsequently be added.
Thus, the partial charge calculations shed light on the electrostatic ion-solvent interactions and the resulting structural changes in the ionic core. Furthermore, even though the total charge on HeN-HOCO + remains constant for all values of N, the magnitude of the partial charge on each atom in HOCO + is lower for N = 6 and 7 than for the unsolvated ion (N = 0). This suggests that any additional helium atoms beyond N = 6 will be less perturbed by the ionic impurity and has implications for the onset of superfluidity once a critical value of N is reached.
Finally we note that, despite our reservations, static ab initio calculations have been successful in modelling the spectroscopic behavior of HeN-HOCO + up to a value of N = 7 and have shown good agreement with experiment. However, this study and future experimental studies on the solvation of ions in helium would benefit from quantum molecular dynamics simulations.

F. Remarks on the experimental technique
The apparatus can be used to measure IR spectra for both neutral molecules and He-tagged ions. It is a relatively simple task to switch between the two modes of operation and this is primarily achieved by changing the gate time settings on the data acquisition electronics. This capability allows for a structural comparison of the neutral precursor and the fragment ion, and can provide information on the electron ionization process. 15 Furthermore, the ability to prepare This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset.

PLEASE CITE THIS ARTICLE AS DOI:10.1063/1.5124137
metastable neutral structures inside helium nanodroplets allows for the creation of new ionic structures that might be difficult or impossible to form using other techniques. However, despite these advantages, there is still scope for improvement. In particular, the present experimental set-up has only one mass-selective stage to record mass-selected IR spectra. In the case of He-tagged ions, an improvement would be to employ tandem mass spectrometry, where the first stage selects a particular ion and the second stage detects a particular photodissociation product. Such an arrangement would deliver a zero background ion signal at off-resonance wavelengths and so would massively improve the signal/noise ratio.
The benefit of mass-selectivity is particularly evident when comparing the present technique with previous IR spectroscopy studies of He-tagged neutral molecules, such as OCS, CO2 and NO2. [28][29][30][31] Those studies were limited to small molecules because of a lack of massselectivity and, as a result, substantial deconvolution of rotationally-resolved IR spectra was required in order to determine the N-dependent behavior. Fortunately, because the present technique does not suffer from the same restriction, it can readily be extended to the study of larger molecular ions. Now the limiting factor on the suitability of an ion will likely be whether there is a measurable frequency change in one of the vibrational modes when the number of attached messenger tags is varied. When there is negligible change, the measured spectrum will still be of considerable value because it will be representative of the bare ion. As such, our technique can be used to report on (i) the bare ion, 15 and/or (ii) ion solvation effects.
It is a relatively straightforward task to extend this technique to other weakly-bound messenger tags, such as, neon, argon or molecular hydrogen. This could be achieved by doping the helium nanodroplets with the new tag species prior to electron ionization. Because interactions with alternative tags will be stronger than for helium, the new tag will preferentially bind to the neutral dopant. Following electron ionization, ions tagged with the new species will be generated, which can subsequently be probed using mass-selective IR depletion spectroscopy.
This is the author's peer reviewed, accepted manuscript. However, the online version of record will be different from this version once it has been copyedited and typeset.

IV. CONCLUSIONS
We have demonstrated that electron ionization of neutral acetic acid embedded in helium nanodroplets can be used to generate a small ion, HOCO + , tagged with up to 14 helium atoms.
IR spectra of the mass-selected ions reveal that the OH stretching vibration is a sensitive probe of the gradual solvation of the ion in helium. In particular, the combined experimental and theoretical data indicate that the first helium atom preferentially binds to the proton and, as further helium atoms are added, a ring of helium is formed around the carbon atom. As well as subtly changing the structure of the ion core, the most strongly bound helium atoms alter the distribution of charge. This may have implications for the formation of the first few solvation layers around a dopant ion following its introduction into a helium nanodroplet. Finally, we note that the experimental technique described here provides an alternative to other techniques for recording IR spectra of helium-tagged ions and can in principle be applied to a wide range of ions, both large and small.

SUPPLEMENTARY MATERIAL
See supplementary material for details of rotational contour simulations of the OH stretching band in He-HOCO + ; ab initio calculations on local minima structures for HOCO + tagged with up to seven helium atoms, one argon atom and one neon atom; ab initio calculations for He-C2OH5 + , CO2H5 + and O3H + .