Physical Chemistry Chemical Physics

Introduction

Proton transfer is an important primary process involved in many chemical reactions and biological processes. In the gas phase, numerous studies have been performed to characterize this process in both neutral and ionic clusters. Excited state proton transfer (ESPT) has been observed and well characterized in two model systems: naphthol-ammonia 1-10 and phenol-ammonia. ^11-18 In these systems, the ESPT results from the coupling of the first covalent excited state [phenol(S₁)-(NH₃)_n] with the ion pair excited state [PhO ^*-(NH₄⁺)(NH₃)_{n 1}] (phenolate anion excited state-protonated ammonia cluster).
The first observation of the ESPT appearance as the cluster size increases was carried out by monitoring the characteristic fluorescence of the naphtholate anion. ^1,2 Later, ionization potential measurements through a two-photon ionization scheme were achieved for both phenol ^11,12 and naphthol ⁸ ammonia clusters, where the strong ionization potential decrease for n=4 in the phenol-(ammonia)_n clusters (and n=3 in naphthol-(NH₃)_n) was assigned to a signature of ESPT.
In benchmark time resolved experiments, the ESPT dynamics was studied using a picosecond pump-probe scheme with ion detection (in both naphthol^5-10 and phenol clusters^13-18). Picosecond decays were observed on [NaphOH-(NH₃)₃]⁺ as well as on [PhOH-(NH₃)_n5]⁺ and were ascribed to the ESPT dynamics.
The most important features of the time resolved experiments on phenol-ammonia clusters are as follows. (i) The very clear picosecond decays observed at the mass of PhOH-(NH₃)_n=5-7 (ref. 13-17) in the pump-probe experiments have been shown to depend on the probe laser wavelength. (ii) A corresponding a rise time is observed on the protonated ammonia clusters (NH₄⁺)(NH₃)_n>4. (iii) This dynamics is not observed for clusters with less basic molecules (water, methanol, . . .).
So far these data have been rationalized as follows. ^13-17 (i) The starting point is the non-proton-transferred cluster in the neutral ground state and, at time t=0, the cluster is excited to the non-transferred part of the excited state potential energy surface (Fig. 1, arrow 1). (ii) The major dynamical process which is evidenced is the proton transfer from PhOH* to (NH₃)_n on a time scale of 60 to 70 ps, while a longer time component is assigned to solvent reorganization. (iii) This process is detected by ionizing the excited phenol-ammonia cluster. If the excited cluster is ionized before proton transfer, then the ion observed is (PhOH-(NH₃)_n)⁺ (Fig. 1, arrow 2). In contrast, if ionization takes place after ESPT, the PhO-NH₄⁺(NH₃)_{n 1} part of the ionic surface is reached and because of the excess energy in the PhO-NH₄⁺(NH₃)_{n 1} ion, fragmentation into PhO+NH₄⁺(NH₃)_{n 1} occurs: the PhO-NH₄⁺(NH₃)_{n 1} ion signal is no longer detected, and the NH₄⁺(NH₃)_{n 1} fragment becomes observable (Fig. 1, arrow 3). This accounts for the decay observed on the (PhOH-(NH₃)_n)⁺ masses as well as for the growth observed on the fragment NH₄⁺(NH₃)_{n 1} masses.

Fig. 1 Schematic potential energy diagram for a PhOH-(NH3)n cluster (e.g. n=5). These curves represent the case where proton transfer can occur in the excited state. The ionic state presents a high barrier between non-proton-transferred and proton-transferred species as required for the observation of picosecond decays on the [PhOH-(NH3)n]+ mass peak and rise times on the NH4+(NH3)n m mass peaks (see text).

The above interpretation implies the existence of a significant barrier on the ionic surface between the non-proton- transferred and the proton-transferred species. This is necessary to ensure a selectivity in the ionic dissociation pathway, depending on the part of the ionic surface reached through ionization of the excited cluster. Before ESPT, the ionization photon brings the system into the (PhOH-(NH₃)_n)⁺ well of the ionic potential energy surface and due to the barrier, the fragmentation in PhO+ NH₄⁺(NH₃)_{n 1} is impossible whereas after ESPT, the PhO-NH₄⁺(NH₃)_{n 1} well is reached which leads to fragmentation.^13-17
One should keep in mind that the observed signal reflects not only the dynamics in the excited state but also the way to probe it. In other words, if the dynamical process is induced by the pump photon, there must be a selectivity in the ionization process by the probe photon to observe a time dependent signal.
Since these very good studies on the picosecond dynamics in phenol and naphthol ammonia clusters,^5-10,13-17 other investigations have been carried out on these systems, which have lead to question this initial interpretation.
(i) A picosecond decay was recorded by Bernstein and co-workers ¹⁸ upon detection of the (PhOH-(NH₃)₂)⁺ mass, which is a cluster size where ESPT is a priori energetically forbidden. This has been interpreted as due to ESPT in larger clusters (n5), followed by evaporation of NH₃ units (three or more) in the ionic state. This result provides the hint that evaporation of ammonia units in the ion has to be seriously taken into account.
(ii) Another recent experiment performed by Jacoby et al. ¹⁹ showed that the fingerprint of PhOH*-(NH₃)_n (n=1-4) complexes, i.e. the existence of resolved vibrational structures in the excitation spectra, can also be recorded on the NH₄⁺(NH₃)_{n 1} fragment masses when ionizing with 2 photons of the same color, which corresponds to a total energy of 8.9 eV. For the small sizes (n=1-3), no ESPT is expected although the appearance of protonated fragments is necessarily linked to the proton transfer mechanism either in the excited or in the ionic state.
(iii) A barrier to proton transfer in the ionic state of the 1-1 complex had been postulated by Mikami et al. to explain their results on the PhOH-NH₃⁺ dissociation: ²⁰ when the 1-1 complex is ionized through the S₁ state, the non-transferred part of the ionic surface is reached. Around 1 eV above the ionization threshold, two dissociation channels are open at the same energy:

(PhOH-NH3)+PhOH++NH3

(1 )

(PhOH-NH3)+PhO+NH4+

(2 )

Only the first dissociation channel has been observed even when ionizing with a total energy of 9 eV, i.e. 1.3 eV above the ionization threshold or 0.3 eV above the thermodynamical limit for these processes. The second channel (PhO evaporation) is supposed to be observed only if proton transfer has occurred in the ionic state and may be regarded as the signature of this process. From the non-observation of NH₄⁺, the authors deduced that the height of the barrier in the 1-1 ionic state is at least 1.3 eV. However, the absorption spectrum of the (PhOH-(NH₃)_n=1,2)⁺ cations, later performed by the same group, has been assigned to the phenoxy radical, indicating that the proton transfer reaction has occurred in the ionic state and that the proton-transferred structure is the most stable one. ^21,22
(iv) The ground state proton transfer in phenol-(NH₃)_n clusters has also been investigated through the measurement of one VUV photon ionization thresholds, using synchrotron radiation. ²³ As for the excited state proton transfer, an important decrease of the ionization threshold, which occurs for n=6 ammonia molecules, may be regarded as the signature of the absorption from a proton-transferred structure in the ground state. This threshold cluster size is in qualitative agree ment with the n=5 size suggested for ground state proton transfer in naphthol-(NH₃)_n clusters, ¹⁸ naphthol being a stronger acid than phenol. This experiment is also in good agreement with theoretical calculations.^24-25 The major problem raised by this experiment is that if proton transfer occurs in the ground state for n=6 or 7 ammonia molecules, then the picosecond dynamics observed in the excited state for n=6 or 7 cannot be proton transfer.
(v) Calculations have been performed for the phenol-(NH₃)₅ in the ground ^24,25 and excited ²⁶ states which confirm that the proton-transferred structure is metastable in the ground state, whereas it is the most stable structure in the excited state (by 0.17 eV) with a barrier from PhOH*-(NH₃)₅ to PhO ^*-NH₄⁺(NH₃)₄ of 0.14 eV. The authors agree to assign the fast decay component of the picosecond signal to proton tunneling through a barrier but calculate the solvent rearrangement rate to be much shorter than the long time component observed in the experiments. On the other hand calculations on the ionic (PhOH-NH₃)⁺ cluster ²⁷ predict no barrier between PhOH⁺-NH₃ and PhO-NH₄⁺, hence no barrier to dissociation in PhO+NH₄⁺.
To summarize, experimental results and theoretical calculations do not seem to give conclusive evidence for an important barrier to proton transfer in the ionic state, especially for medium cluster sizes. This lead us to test the presence of a large barrier to proton transfer in the ionic state of phenol-(NH₃)_n clusters, the barrier leading to a specific fragmentation, since it is one of the main implicit assumptions in the interpretation of picosecond dynamics.
Before starting with clusters, it may be useful to recall the photophysical processes occurring in the excited states of bare phenol, which can also be relevant to the dynamics of phenol embedded in an ammonia cluster. As many substituted benzenes, phenol has a weak fluorescence quantum yield (0.08 in a non-polar solvent ²⁸ ) and consequently a short fluorescence lifetime (1.5-2 ns gas phase value ²⁹ ) as compared to the radiative lifetime of 30 ns, ref. 28), because of strong couplings with the triplet (ISC) and ground (IC) states.
Since the triplet state is around 1 eV below S₁ (E(S₁-T₁)=7744 cm ¹ in free phenol), intersystem crossing will lead to fragmentation of the clusters as shown in the case of the phenol-water complex. ^28,29 An ISC rate constant of k_isc=5×10⁷ s ¹ has been measured for free phenol and does not change significantly in the water complex: k_isc=2.3×10⁷ s ¹. ²⁹ The sum of fluorescence and ISC yields varies between 0.16 and 0.46 depending on the authors. ^29,30 These values, far from 1, evidence an important yield for interconversion and eventually for photochemistry, since the dissociation channel PhOHPhO+H is lower in energy (by 0.66 eV) than S₁, ³¹ so that S₁ excitation of phenol can also lead to the dissociation into phenoxyl radical and a hydrogen atom.
The best way to disentangle the excited state proton transfer itself from the fragmentation processes in the ionic (or excited) state is to use a two-color ionization scheme. The small clusters (n=1-4) can be selectively excited by the pump photon ¹⁹ and the energy imparted to the ionic state can be partially controlled by varying the probe laser wavelength.
Fig. 2 shows the energetic diagram with the different energy levels in the ionic phenol-ammonia clusters. It has been built by scaling the two dissociation channels of the 1-1 complex ion: PhOH⁺+NH₃ and PHO+NH₄⁺ with known thermodynamical values: phenol gas phase acidity (_acidH=1460.9±8 kJ mol ¹), ^31-33 phenol ionization potential (8.505 eV), ³⁴ PhO electron affinity (EA=217.4±0.6 kJ mol ¹), ³⁵ ammonia proton affinity (8.85 eV), ³² and PhOH-NH₃ ground state binding energy (D₀=0.22 eV experimental value ²⁰ or D₀=0.303 eV calculated value ³⁶ ). These values lead to:

Fig. 2 Energy diagram for the phenol-ammonia ionic clusters. This energy diagram takes the phenol ground state as the zero of energy. (1-n) correspond to the energies of clusters in their ground states. (1-n*) correspond to the clusters in the excited states (v=0). 19 (1-n+) correspond to the vertical ionization thresholds. 11,20,23 The boxes correspond to the thresholds for dissociation in PhO+NH4+(NH3)n 1, the height of each box representing the uncertainty on the thermodynamical values. 37 The boxes are labeled (NH3)nH+ instead of PhO+NH4+(NH3)n 1 for visibility.

taking the energy scale origin on the bare phenol ground state. For larger clusters, (NH₃)_n NH₄⁺ energy levels are deduced from the proton affinities ³⁷ of ammonia clusters. The PhOH-(NH₃)_n vertical ionization thresholds are experimental measurements. ^11,20,23
This energy diagram helps us to design the required experiment, i.e. to choose the best probe wavelength for which the total energy brought into the ion is just above the PhO+NH₄⁺(NH₃)_{n 1} dissociation channel for either the 1-2 or the 1-3 complex. No proton transfer in the excited state is expected for these cluster sizes, thus according to the previous model, the PhO+NH₄⁺(NH₃)_{n 1} dissociation channel should not be observed because a high barrier to proton transfer is assumed in the ionic state. If we use a pump photon around 281.5 nm to excite the cluster (in the absorption range of 1-2 and 1-3 clusters) ¹⁹ and a 355 nm photon to ionize, the total energy in the complex will be 7.90 eV. In the 1-3 cluster, this corresponds to an excess energy 0.25 eV larger than the dissociation energy of (PhOH-(NH₃)₃)⁺ in PhO+NH₄⁺(NH₃)₂. On the other hand, the 1-2 complex can be ionized but the dissociation channel leading to NH₄⁺NH₃ is energetically closed, which will be a good way to estimate the possible effects of multiphoton processes.
It will be seen in the present study that although ESPT is not energetically allowed for clusters with two or three ammonia molecules, the NH₄⁺(NH₃)_1,2 fragments are observed in REMPI experiments, even when a delay as large as 350 ns is applied between excitation and ionization lasers. These fragments may come from dissociation in the ionic state: in that case, excited PhOH-(NH₃)_n clusters must live much longer than the initially excited S₁ state and triplet states must be involved. The NH₄⁺(NH₃)_n fragments may alternatively come from ionization of stable NH₄(NH₃)_n produced through intracluster hydrogen transfer in the excited S₁ state: PhOH*(S₁)-(NH₃)_nPhO+(NH₄)(NH₃)_{n 1}. This mechanism is analogous to the reaction occurring in excited ammonia clusters: (NH₃)_n^*NH₂+(NH₄)(NH₃)_{n 2}. ^38,39
These two possibilities will be analyzed and discussed in comparison with other experimental data from different groups.

Experimental

The clusters are produced by expanding a mixture of Ne seeded with NH₃ (0.5%) and passing over a reservoir contain ing phenol at room temperature. The backing pressure used is typically 1.5-2 bar and the nozzle diameter is 300 m. Under these conditions no large clusters are observed and a clean spectrum of the smaller clusters (1 to 4) can be obtained as in ref 12. The pump laser is an unfocused (4 mm² spot) frequency doubled excimer pumped dye laser, whose power is kept low (around 80 J). The probe ionizing laser is either the third harmonic of a YAG laser (355 nm) kept at very low power (from 1 to 5 mJ) in a 25 mm² spot or a frequency doubled YAG pumped dye laser (200 J). The two lasers are electronically synchronized with 1 ns precision but due to the lasers' temporal width (10 ns), the effective resolution cannot be better than a few ns (2 ns). The ions are detected in a 1 m time of flight mass spectrometer.

Results

Discussion

We will first discuss the case of the 1-2 and 1-3 complexes and in particular the observation of the protonated ammonia products when the delay between pump and probe is large (up to 350 ns). This result is surprising since so far, PhOH*(S₁)-(NH₃)_2,3 have been thought to be non-reactive species. Thus ionization by the second laser should occur during the lifetime of PhOH*(S₁)-(NH₃)_2,3. The phenol (S₁) lifetime is short (2 ns, ref. 29 and 30) and we do not expect the PhOH*(S₁)-(NH₃)_2,3 lifetime to be much greater (it is 15 ns for PhOH*(S₁)-H₂O, ref. 29). Thus, to explain the results, we have to find either a long-lived state or an alternative way to produce NH₄⁺(NH₃)_n other than fragmentation in the ionic state.
For the sake of clarity, Fig. 8 presents the different states that can be involved in the case of PhOH(NH₃)₃. Two main hypotheses can be postulated: intersystem crossing and hydrogen transfer.

Fig. 8 Crude representation of the potential curves involved in the PhOH-(NH3)3 two-photon two-color ionization. The reaction coordinate being PhOH···(NH3)3 or PhO···(NH4)(NH3)2.

Conclusion

The dynamics of excited phenol-ammonia clusters seems to be more complicated than previously expected: a new channel, the hydrogen atom transfer is introduced in the present study to rationalize the nanosecond pump-probe observations. The evolution of the excited state dynamics as the cluster's size increases can be summarized including this new channel (although the exact cluster size might still be subject to further investigation) as follows. (i) For the 1-2 and 1-3 clusters, the most important reactive channel is H transfer. This process is slow enough in the 1-2 complex to observe fluorescence and probably faster in the 1-3 complex. (ii) For the 1-4, 1-5, 1-6 clusters, in addition to the H transfer channel, the proton transfer channel becomes energetically allowed, enhancing the intersystem crossing (towards triplet states). (iii) For larger clusters (n7) the proton transfer already occurs in the ground state, and the H transfer reaction is closed.
The case of the 1-1 complex is still under investigation: its strong fluorescence indicates that if hydrogen transfer occurs, it is a slow process. NH₄ has a very short lifetime (15 ps) and therefore its concentration will remain very small in the jet. Its ionization threshold (4.66 eV) is some 0.25 eV above the S₁S₀ transition of the phenol-ammonia clusters so that the probe laser also acts as a pump leading to a low signal-to-noise ratio. For these reasons, no conclusions can be drawn at present.
The previous analysis of time resolved experiments was perfectly reasonable in view of the experiments available at the time. It is only because new experiments 18,19,21-23 were in contradiction with the previous interpretation^13-17 that we have reached the conclusion that the dynamics in this system has to be revisited, including new channels.
With the new hypothesis of hydrogen transfer, much more work has to be done to understand this simple phenol-ammonia system in which many photophysical and photochemical processes are observed to compete and evolve with the cluster size.