Suppression of Multiphoton Ionization of Aniline in Large Superfluid Helium Droplets

Abstract

We report suppression of multiphoton ionization (MPI) of aniline doped large superfluid helium droplets containing over 5 × 10⁶ atoms. In contrast, surface-bound sodium atoms and dimers are readily desorbed and ionized. Adequacy of the experimental conditions is also confirmed from ejection of embedded aniline cations from smaller droplets containing multiple cations, and MPI of gaseous aniline. The photoelectrons have a mean-free-path of less than 1 nm and a thermalization distance of 10 nm. In a droplet with a diameter of over 70 nm, effective charge recombination within the droplet is expected.

Graphical abstract

Introduction

Superfluid helium droplets have captured the fascination of many experimentalists and theorists alike.^1–5 Early activities have mostly exploited the superior cooling capabilities of small droplets, on the order of thousands of helium atoms/droplet, for spectroscopic and dynamic studies of dopants and dopant-helium interactions.^1–3 More recently, large droplets containing millions of atoms/droplet have been investigated, motivated by the desire of growing unique nanostructures and cooling macromolecules such as proteins.^4–7 Regardless of its size, an inescapable effect of any matrix is associated with the solvent cage.^8–10 Although in experiments of “vibrational transition moment angles” pioneered by the Miller group,¹¹ the effect of the solvent cage is limited to line-broadening, in electronic transitions, the solvent is known to split the zero phonon line and produce extensive phonon wings.^{2, 12, 13} In photodissociation, translational energy release is affected by collisions between the fragments and the helium atoms,^{8, 14} and the yield of dissociation is decreased because of fragment recombination.¹⁵ Under conditions of field induced alignment, the rotational coherence of the dopant is partially scrambled by the solvent, and the observed revivals of molecular alignment are damped faster than in the gas phase.¹⁶

In spite of the caging effect of helium droplets, non-thermal ejection of cations and molecules has been reported from large and small droplets.^{17, 18} Upon resonant vibrational or electronic excitation, instead of energy dissipation through evaporation of surface atoms, the solvent cage suffers significant disruption, resulting in ejection of the dopant out of the droplet. While in small droplets a single IR photon is sufficient for desorption,¹⁷ in large droplets with ionic dopants, multiple photons are required.¹⁹

The possibility of caging by the helium atoms within a droplet did not deter the attempt of single-photon^{20, 21} or multiphoton ionization (MPI) of dopant molecules including those in the interior^22-25 and on the surface.^{26, 27} Under synchrotron radiations in the vacuum UV above the first excited state of helium,^{20, 21} the predominant event is direct excitation or ionization of helium atoms, and the embedded molecules are ionized through either Penning ionization or charge transfer. In the case of MPI in the UV/VIS region, only dopant molecules can absorb and ionize in the laser field, and the effect of the helium solvent manifests in the reduced and broadened kinetic energy distributions of the photoelectrons.¹⁰ In all cases, the droplets are generally small in size (< 10⁵), and the ionization event occurs through escape of photoelectrons, while the resulting cations retain a number or most of the helium atoms.

In this work, we report an effect of the helium environment in suppressing MPI of aniline (A) embedded in droplets containing over 5 × 10⁶ helium atoms/droplet, using a range of laser powers under both resonant and non-resonant conditions. Different from the non-thermal ejection process of Filsinger et al in large droplets,¹⁹ our experiment begins with neutral aniline instead of A⁺, and different from the previous reports on MPI of doped droplets,^22-25 our experiment is the first attempt of MPI in large droplets. We confirm the viability of the experimental setup using MPI of surface-bound sodium atoms and sodium dimers, ejection of embedded A⁺ under laser excitation from small droplets containing more than one cation, and MPI of gaseous aniline. We interpret this result in terms of the mean-free-path (1 nm) and the thermalization distance (< 10 nm) of the low energy photoelectrons, and the diameter of the large droplets (> 70 nm). The photoelectrons have a negligible probability of escaping from the droplet and ultimately they will recombine with the cations. Our result offers a cautionary tale in employing superfluid helium droplets as an ultra-cold non-interacting medium: large droplets can exhibit bulk-like behaviors and participate in physical and chemical processes.

Experimental Setup

The experimental apparatus is shown in Fig. 1.^{7, 28, 29} Helium droplets are generated from a pulsed valve (General Valve, Parker, series 99) with an orifice of 0.5 mm at stagnation pressures between 20 and 30 atm. The pulsed valve is cooled by a cryostat (Sumitomo, SRDK-415D2) and heated by a cartridge heater to maintain temperatures in the range of 4 to 16 K. The duration of the electrical opening signal to the pulsed valve is between 110 and 140 μs although the exact duration of the opening time of the valve is unknown. The size distribution of the droplets, after picking up a green fluorescent protein (GFP), has been determined to be bimodal,⁷ on the order of 3 to 6 × 10⁶ and above 10⁸ helium atoms/droplet (diameters: 63 - 80 and larger than 204 nm). Not knowing the binding energy of GFP with the surrounding helium atoms in a droplet, we can calculate a lower limit on the number of evaporated helium atoms from the kinetic energy and the heat release of room temperature GFP containing ~2000 atoms. This value of 2 × 10⁶ atoms should then be added to the measured size after doping, and the resulting size of 5 × 10⁶ atoms/droplet for the smaller group, should then be the lower limit of the droplets directly from the source. Changes in the source temperature result in changes in the relative abundance of the two groups of droplets, while the average size of each group varies only slightly. Although detailed reproducibility of the valve is known to be questionable, the presence of two groups of droplets with two different sizes has been reliable throughout the experiment. While it would be desirable, this source cannot produce droplets smaller than 5 × 10⁶ atoms/droplet. Immediately downstream from the skimmer (Beam Dynamics, 5 cm long and 2 mm in diameter), the droplet beam encounters a biased grid, and above and below the grid are two filaments, one bare that can supply hot electrons for electron impact ionization (EI) of neutral droplets, and the other coated with a zeolite paste rich in sodium. Further downstream room temperature gaseous aniline is routed into the chamber though a nozzle with a diameter of 0.5 mm controlled by a needle valve. A flat plate with a hole of 1 cm in diameter separates the doping region and the analysis region. The bottom (repeller) electrode of the time-of-flight (TOF) mass spectrometer (MS) can be pulsed to 500 V to detect charged species in the droplet beam. Further downstream, a retardation plate and a Daly dynode detector can be used for measurements of mass-to-charge ratios (m/z) of detected droplets.^{7, 28, 29} The vacuum level in the source chamber is 2 × 10⁻⁷/5 × 10⁻⁶ Torr when the droplet beam is off/on, the doping chamber is 1 × 10⁻⁷/1 × 10⁻⁶ Torr, and the analysis chamber is 1 ×10⁻⁷ Torr regardless of the droplet condition.

Figure 1:

Experimental setup showing the bare filament for hot electrons, the paste coated filament for Na, the grid for energizing the hot electrons for ionization, and the retardation electrode for mass analysis of the ion doped droplets. The distances are labeled in units of cm, and the gray line represents a flat plate with a central hole separating the doping region from the analysis region.

While the small TOF is used to measure ions containing a few hundred He atoms with a limited mass resolution, the Daly detector can measure masses up to millions of atomic units (u). Only m/z can be determined from these measurements, although for the convenience of discussion, sometimes the number of helium atoms/droplet, related to the m/z by a factor of 4 assuming z = 1, are used in this manuscript. Fig. 2(a) shows the mass spectrum of a neat droplet beam obtained by setting the grid voltage to 300 V and pulsing the repeller at 500 V. Only helium clusters containing 100 – 400 atoms can be detected. The lower mass limit at 400 u is due to inefficient transmission through the beam path of 33 cm from the ionization grid to the TOF, since smaller ions are likely lost in stray electric fields or when colliding with residual gases in the vacuum chamber. The upper mass limit at 1600 u is determined by the deflection power of the repeller electrode, and heavy ions with velocities perpendicular to the TOF (velocity of the droplet beam) can collide with the wall of the TOF. Fig. 2(b) shows two traces recorded on the channeltron of the Daly detector when the Dynode was biased at −20 keV: the ionizing grid was at 150 and 300 V respectively and the intensities are normalized for comparison. The horizontal axis is the arrival time of the ionized droplets from the TOF to the Daly dynode and is not directly reflective of the mass of the detected droplets. Instead, the size information can be derived from retardation measurements^{7, 28, 29} based on the speed derived from the arrival time and the flight distance. In Fig. 2(b), two major groups of ions are observed with two different velocities and sizes at 150 V, but only one group with an average size of 5 × 10⁵ atoms/droplet has a substantial intensity at 300 V. We note that an energy of 300 eV can only remove 5 × 10⁵ atoms, hence the mechanism of size reduction is not just evaporation of surface helium atoms. Multiple charging and droplet fragmentation are both possible mechanisms of reducing the mass-to-charge ratios of the large droplets.^{30, 31}

Fig. 2.

Ion signal from electron impact ionization of pure superfluid helium droplets recorded using (a) the time-of-flight mass spectrometer at an ionization energy of 300 V with a pulsed repeller voltage of 500 V; (b) the Daly detector biased at −20 keV (the electrodes of the TOF are all grounded) at ionization energies of 150 and 300 V. The mass limit of (a) is imposed by the transmission efficiency for small ions and by the geometry and voltage setting of the TOF for large ions. The vertical lines label the corresponding number of helium atoms for the given m/z assuming z = 1. The arrival time in (b) is not directly related to the sizes of the ions. Rather the size is derived from a retardation experiment.⁷ The source temperature was 7 K and stagnation pressure was 30 atm.

Depending on the experiment, three different light sources are used, all operated at 10 Hz. A Nd: YAG (Precision II 8000, Continuum) pumped optical parametric oscillator (OPO, Panther, Continuum) is used to cover the range from 410 to 620 nm, with pulse energies from 200 μJ - 6 mJ, a pulse duration of ~8 ns, and a beam diameter of 30 μm, and the resulting peak intensities are 3 × 10⁹ – 1 × 10¹¹ W/cm². A Nd:YAG (Spectra Physics, GCR 230) pumped dye laser (Laser Analytical Systems, LD 2051) is used between 405 and 415 nm, with a power range from 100 μJ to 3 mJ and a diameter of 60 μm, corresponding to peak intensities from 5 × 10⁸ to 2 × 10¹⁰ W/cm². The 2^nd to 4^th harmonic of the Nd:YAG at 532, 355 and 266 nm are also used for ionization with a power range from 3 to 80 mJ and a beam diameter from 20 μm (266 nm) to 40 μm (532 nm).

To distinguish ions ejected from doped droplets from the gaseous background, including ions from photoionization of diffused aniline and sodium, all spectra related to the doped droplet were recorded under active background subtraction conditions.³² We set the droplet source at 5 Hz and the laser at 10 Hz, and the difference signal obtained when the droplet was present and absent is considered the effective signal due to doped droplets. We acknowledge that even with this precaution, interactions of excited helium species such as He* from the droplet with gaseous species cannot be completely excluded.

Results

Aniline doped droplets

Fig. 3 trace (a) shows the mass spectrum of gaseous aniline recorded from the diffused sample in the laser ionization region. This result confirms the adequacy of the laser wavelength and power density for MPI of aniline. As evidence to the spatial and temporal overlap between the droplet beam and the laser beam, Fig. 3 trace (c) shows the result from laser excitation of small multiply charged droplets doped with aniline.³³ The droplets are first ionized at 300 V, resulting in a size distribution as illustrated in Fig. 2(b). These droplets are then doped with aniline, and charge transfer from one of the He₂⁺ in the droplet forms A⁺. Whenever the laser wavelength scans through a multiphoton resonance of A⁺, A⁺ can be observed. For both traces (a) and (c), the laser was at 420 nm, slightly different from the exact resonant wavelength but still within the broad excitation range of A⁺. The broad feature of Fig. 3(c) in the large mass region corresponds to the cationic clusters of helium, and it is reproduced in the inset on the right, trace (e), for a detailed comparison with that from neat (undoped) droplets, trace (d), reproduced from Fig. 2(a). The similarity between the two spectra for He_n⁺ confirms the current assignment.

Figure 3:

Mass spectra under different conditions: (a) MPI of gaseous aniline, (b) MPI of aniline embedded in small neutral droplets, and (c) laser excitation of aniline embedded in ionized small droplets. The laser for traces (a) and (c) was at 420 nm with an intensity of 2 × 10¹⁰ W/cm², and the source temperature of the droplet beam for (c) was 4.5 K. The size of the droplets in (c), reduced by the ionization electrons at 300 V, is 5 × 10⁵ atoms/droplet, as shown in Fig. 2(b). Trace (b) was recorded using a different apparatus with a different pulsed valve at 266 nm, 3 × 10¹⁰ W/cm², and an average size of 3 × 10⁴ atoms/droplet. The label C_n only indicates the number of carbon atoms n in each group of ions, while the number of attached hydrogen atoms, as shown from the inset on the left, is omitted. The inset on the right compares the distributions of the helium cluster ions with and without aniline doping: trace (e) is a scaled duplicate of (a), and trace (d) is a duplicate of Fig. 2(a).

To observe MPI of embedded neutral aniline, the electron source and the grid were turned off, and to further minimize potential complications due to timing issues, the repeller was biased at a constant voltage of 500 V. With the same settings for the laser, the droplet, and aniline, we failed to observe any ions either from the TOF MS or the Daly detector. We have also increased the laser intensity to 10¹² W/cm² (80 mJ/pulse) at 355 and 532 nm, and switched the laser wavelength to 266 nm at 3 × 10¹⁰ W/cm², but still failed to observe any aniline related masses on the MS or the Daly detector. We are forced to conclude that no ionization occurred from embedded neutral aniline. We have observed the same behavior under different source temperatures up to 16 K, within a range of doping pressures between 3 × 10⁻⁷ – 5 × 10⁻⁶ Torr, and laser wavelengths from 410 nm to 600 nm, in addition to 355 and 266 nm.

Limited by the droplet source, we could not reduce the droplet size below 5 × 10⁶ atoms/droplet, so we repeated the experiment using a different apparatus²⁴ that is known to only produce smaller droplets because of a different pulsed valve and cryostat. At 266 nm and 3 × 10¹⁰ W/cm², the mass spectrum from droplets containing 3 × 10⁴ atoms/droplet is shown in Fig. 3 trace (b). The lower voltage on the repeller of this mass spectrometer results in a much higher mass resolution but a diminished sensitivity for large masses, and the fine structures surrounding each major carbon-containing species are due to the difference in the number of attached hydrogen atoms as shown in the inset on the left. Using this apparatus, we have also observed that with increasing droplet sizes up to 2 × 10⁵ atoms/droplet, all aniline-related ions decrease in intensity.

In the low mass region, trace (b) is similar to trace (c), with the parent ions being the most abundant. Trace (a) from the gas phase, however, shows intensive fragmentation. Although the ionization mechanisms are different between traces (b) and (c), the cooling capability of the surrounding helium atoms in a droplet seems universally effective in preventing fragmentation, as reported in the literature.³⁴

Sodium doped droplets

We suspect that the lack of MPI from aniline doped droplets could be related to the caging effect of large droplets on dopants located inside. If the dopant is located on the surface, no such effect should exist. Heated molecular sieve pastes are a source of sodium atoms,²⁹ and neutral sodium atoms can be attached to the surface of helium droplets.³⁵ As a corollary experiment, we turned of both the electron source and the grid voltage, and heated up the filament coated with molecular sieve. The top panel of Fig. 4 shows the observed Na⁺ and Na₂⁺ in the MS from MPI of doped neutral droplets obtained using the OPO laser in the intensity range of 3 × 10⁹ – 1 × 10¹¹ W/cm². Diffusion of neutral sodium atoms from the heated paste into the laser field results in the excitation and ionization spectrum of gaseous sodium recorded without the droplet beam, as shown in the bottom panel. The lowest energy electric dipole allowed transition via single-photon excitation is the 3p ← 3s transition, as labeled in the figure.³⁶ All other transitions in the shorter wavelength range are two-photon transitions to nd Rydberg states with the principal quantum number n labeled on the comb. The sudden rise in the ionization yield at 483 nm corresponds to the two-photon ionization threshold,³⁶ above which the ionization yield monotonically decays.

Figure 4:

Spectra of Na⁺ and Na₂⁺ from doped helium droplets generated at a source temperature of 7 K and stagnation pressure of 30 bar (top panel). The ionization spectrum of gaseous Na recorded when the droplet source was off is shown in the bottom panel for comparison. The top panel is the difference between the ion signals with and without the droplet, hence whenever there is a strong resonance for the gaseous sample, negative or positive abnormalities in the difference spectrum can be seen. The comb in the bottom panel labels the nd Rydberg series through two-photon transitions, and the converging limit is the two-photon ionization threshold of Na at 483 nm. The single-photon transition 3p ← 3s, i. e., the D line of Na, is also labeled.

Excitation and desorption of sodium atoms located on the surface of a droplet have been reported extensively in the literature.^{35, 37, 38} The desorption process happens in picoseconds³⁹ and is followed by absorption of one extra photon, leading to Na⁺.⁴⁰ The resonances seen in the top panel show correlations to those of gaseous Na, and the sharp positive and negative peaks are due to incomplete subtraction of Na⁺ from the strong diffused sodium atoms. The ionization thresholds of surface-bound Na and Na₂ show red-shifts compared with that of Na in the gas phase.⁴¹ The value of the shift of Na, 0.5 nm from Fig. 4, is consistent with fluorescence measurements⁴² but smaller than that reported by Loginov and Drabbels⁴³. It is worth noting that Na₂⁺ is only present in doped droplets, and is undetectable from the diffused gaseous sodium under our conditions. The observations of Fig. 4 are consistent in the temperature range of the droplet source from 4 K to 14 K, and above 14 K, the signal intensity from droplet related sodium dropped below our detection limit.

The broad resonance around 570 nm in the upper panel is tentatively assigned as the two-photon transition of surface-bound Na to the 4d or 4f level. The two transitions in bare atoms are separated by ~20 cm⁻¹, and the current assignment corresponds to a blue shift of ~ 650 cm⁻¹ relative to that of a bare atom.³⁶ The center of the transition is close to the 4f ← 3s transition observed by Loginov et al.⁴⁴ from single-photon excitation, which would be forbidden without symmetry breaking in a droplet environment. The authors also compared the one- and two-photon transitions to the 4p level, and reported that the blue shift of a two-photon transition is 400 cm⁻¹ whereas that of a one-photon transition is 150 cm⁻¹. In this context, the current observed blue shift is in accordance with the previous report.⁴⁴ A puzzling consequence of this assignment is the missing broad transitions to the corresponding higher d and f levels. We suspect that the transition of the bare atom to the 4d level might be saturated in the bottom panel, and the actual intensity to the higher d levels might be much weaker than that to the 4d level. The resonance near 425 nm is unique to doped droplets, and we have limited information on its assignment.

Discussion

Two experiments including ejection of A⁺ from small multiple cation-doped droplets and MPI of Na-doped droplets both confirm the spatial and temporal overlap between our laser beam and the droplet beam. Both the intensity and wavelength of the laser are adequate for ionization of gaseous aniline. The lack of MPI from aniline doped droplets has to be attributed to trapping of photoelectrons in large superfluid helium droplets. Unlike the surface-bound alkali metal atoms,³⁵ aniline is expected to reside inside a helium droplet, and ionization of embedded aniline entails escape of the photoelectrons from the droplet. The ionization threshold of aniline is 7.72 eV,¹⁰ requiring 2 photons at 266 nm, 3 or 4 photons in the range of 322 – 644 nm. The resulting photoelectrons have kinetic energies less than 3 eV across the range from 266 nm to 644 nm. The mean-free-path of elastic scattering of electrons in a helium droplet ranges from 0.75 nm at 1 eV to 0.8 nm at 3 eV.⁴⁵ However, the much smaller mass of an electron than that of a helium atom results in minimal energy loss after each elastic collision.⁴⁶ Hence even if the photoelectron is directed back to the cation after the first few collisions, immediate recombination is still impossible. In fact, the low energy electron can remain mobile for ~1 ps, go through more than a thousand collisions, and travel an average distance — the thermalization distance – of 10 nm.^{46, 47} For a droplet containing 5 × 10⁶ helium atoms with a radius of 38 nm, a photoelectron has less than 2% chance to escape from the droplet. The lack of observed cations in this experiment should therefore be related to the limited detection sensitivity of our system. We note here that the escape of the aniline cation from an ionized small multiply charged droplets in Fig. 3(c) is different from the escape of the photoelectrons: in the former case, the size of the droplet is reduced to 5 × 10⁵, and A⁺ is already close the surface of a droplet because of the Coulomb repulsion from the coexisting He₂⁺.³³

The experiment using a different apparatus²⁴ with much smaller droplet sizes is also in agreement with this assessment. In the size range of less than 2 × 10⁵ atoms/droplet, we can observe not only MPI from aniline doped droplets, but also decreases in the ion yield with increasing droplet sizes. This latter observation is complicated by the possibility of a decreased amount of aniline transported to the laser beam by the droplets, due to a possible decrease in the total number of droplets when the size of each droplet is larger. However, Fig. 3(c) implies that even when the droplet sizes are over 10⁶ atoms/droplet, once they are reduced by EI at over 70 V in the grid region, A⁺ is readily observable upon laser excitation. This fact confirms that there should be a sufficient amount of aniline transported to the laser beam for observation within our sensitivity. Hence the complete lack of A⁺ from MPI of large droplets should be related to the failed escape of photoelectrons in large droplets.

The caging effect of photoelectrons is limited to dopants located in the interior of large droplets, hence no such effect is expected for surface dopant such as alkali metal atoms and clusters nor for small droplets.³⁵ In fact, Loginov et el. reported observations of photoelectrons from MPI of aniline doped droplets in the size range of 10³ - 10⁴ atoms/droplet.¹⁰ The authors noticed that electrons with energies less than 1230 cm⁻¹ were caged inside the droplets.

One undesirable property of large droplets is their tendency to pick up residual gas contaminants along the flight path. These contaminants can complex with the aniline dopant and could potentially increase the ionization threshold. The mass spectrum of Fig. 3b shows that at 266 nm in a laser field of 3 × 10¹⁰ W/cm², in addition to ionization (a 2-photon process), multiple C-C bonds in aniline are also broken, hence the laser intensity is more than sufficient to induce 3-photon absorption in aniline. As long as the increase in ionization threshold due to complexation is within the energy of a single photon at 266 nm, i.e. 4.6 eV, ionization should still be achievable even for the aniline-contaminant complex. Hence the lack of ionization from our experiment should not be related to contaminants.

Concluding remarks

The current report on the lack of MPI in aniline doped large droplets offers an interesting and a cautionary tale on the effect of the helium matrix. Multiphoton ionization of dopant molecules with UV/VIS photons necessarily produces low energy electrons. For species adsorbed on the surface of a large droplet or for any species embedded inside small droplets containing thousands of helium atoms, escape of the photoelectron is not significantly affected, and near gas-phase behaviors can be expected. However, with more than several million atoms in a droplet, the photoelectron can thermalize and recombine with the cation before it could escape from the center to the surface of a droplet. The net result is suppression of ionization. This negative result offers a word of caution when exploiting the ultracold environment of superfluid helium droplets for spectroscopic and other related studies.

Highlights for Review:

1. Suppression of multiphoton ionization by large superfluid helium droplets

2. Surface-bound species are not affected by droplet sizes