Competition between O–H and S–H Intermolecular Interactions in Conformationally Complex Systems: The 2-Phenylethanethiol and 2-Phenylethanol Dimers

Noncovalent interactions involving sulfur centers play a relevant role in biological and chemical environments. Yet, detailed molecular descriptions are scarce and limited to very simple model systems. Here we explore the formation of the elusive S–H···S hydrogen bond and the competition between S–H···O and O–H···S interactions in pure and mixed dimers of the conformationally flexible molecules 2-phenylethanethiol (PET) and 2-phenylethanol (PEAL), using the isolated and size-controlled environment of a jet expansion. The structure of both PET–PET and PET–PEAL dimers was unraveled through a comprehensive methodology that combined rotationally resolved microwave spectroscopy, mass-resolved isomer-specific infrared laser spectroscopy, and quantum chemical calculations. This synergic experimental–computational approach offered unique insights into the potential energy surface, conformational equilibria, molecular structure, and intermolecular interactions of the dimers. The results show a preferential order for establishing hydrogen bonds following the sequence S–H···S < S–H···O ≲ O–H···S < O–H···O, despite the hydrogen bond only accounting for a fraction of the total interaction energy.


Experimental and computational methods
Mass-Resolved Excitation Spectroscopy.The experimental set up has been described in detail in previous publications, 1 and consists of an in-house designed linear time-of-flight (ToF) mass spectrometer; two systems of Nd/YAG laser + dye laser + doubling unit (Quantel Brilliant B + Fine adjustment and Quantel Qsmart 850 + Qscan) and an OPO system (LaserVision) pumped by a Nd/YAG laser (Continuum mod Surelite), together with the required electronics for control and signal acquisition.
Briefly, the sample (2-phenylethanethiol and/or 2-phenylethanol) was deposited in a sample holder and inserted in the gas line feeding a pulsed valve (Jordan Inc.) attached to the ionization chamber of an in-house designed mass spectrometer.At each aperture of the valve, operated at 10 Hz, a supersonic expansion was created, producing an adiabatic cooling of the molecules down to 2-5 K of rotational temperature and 50 -100 K of vibrational temperature.Under such conditions, the molecules aggregate.Either He or Ne were used as buffer gas at typical pressures between ~2-5 bar.The species populating the beam were excited with UV photons that first excited the S1←S0 electronic transition and then ionize the molecules.The electric field, created by a pair of extraction/repulsion plates with 400 V of voltage difference between them, sent the ions towards the detector, accelerating them according to their charge-to-mass ratio.Before reaching the field-free flight path, the ions were further accelerated by an acceleration plate.During the travel through the field-free region of the mass spectrometer, the ions segregate according to their mass, arriving to the detector at different times.
The detector consisted of a pair of MCP plates that produced a ~4 mV pulse per ion.The signal produced by the ions was collected with the aid of a digital oscilloscope (Tektronix TDS 3032), integrated and routed to the same computer that controls the whole experiment.Synchronization between the lasers, valve and oscilloscope was handled using three SRS 645 (Standford Research Systems) pulse generators.
In the 2-color resonance enhanced multiphoton ionization (REMPI) experiments, the excitation laser was scanned through the region of interest, while the total ion current in a given mass-channel was recorded.In this way, the mass-resolved electronic excitation spectrum of the molecules of interest was recorded, with ~0.1 cm -1 spectral resolution.In the IDIRS (ion-dip infrared spectroscopy) experiments, an additional IR laser beam was used, fired ca.400 ns prior to the UV lasers.To record the spectra, the IR laser was scanned through the SH-OH stretching mode region, while the excitation laser was kept tuned to an electronic transition of the species of interest.When the IR photons were resonant with a vibrational transition of the species probed by the UV laser, a dip in the ion signal was produced.In this way, the isomer-specific mass-resolved IR spectra of the species in the beam was recorded.
Rotational spectroscopy.The rotational experiment used a chirped-pulse Fourier transform microwave (CP-FTMW) spectrometer [2][3][4] operating in the frequency range 2-8 GHz.The spectrometer uses a direct-digital design introduced by Pate, 5 implementing fast-passage 6 rotational broadband excitation through the use of microwave chirped pulses.The chirped pulses (typically 1-4 s length) are produced with an arbitrary waveform generator (25 GSamples/s), which are later amplified and broadcasted into a jet expansion using a ridged waveguide horn antenna.The first experiments used a power amplifier of 25 W, which was later upgraded to 250 W using a travelling-wave tube.The compounds of 2-phenylethanethiol (b.p. 217°C) and 2-phenylethanol (b.p. 220°C) were obtained commercially and used without further purification.The sample was located inside a heating reservoir attached to a solenoid-driven injector.Both compounds were vaporized at moderate temperatures of 50-75°C.The sample vapors were pressurized with an inert carrier gas, using initially neon (2 bar) and later He-Ar 1:1 (2 bar).The gas expanded supersonically when passing through a circular nozzle (ϕ = 0.8 mm) into the evacuated high-vacuum chamber (ultimate pressure 10 -7 mbar).The jet was oriented perpendicularly to the exciting radiation.Following the chirped-pulse excitation the sample returns to the initial equilibrium state by emission of a free-induced decay.The time-domain transient emission (40 s) is recorded was a digital oscilloscope (20 MSamples/s), averaged and Fourier transformed to give the frequency-domain spectrum.One to eight chirped pulses are used per gas pulse.The number of averaged spectra in each spectrum was 1 M cycles, at a repetition rate of 5 Hz.The uncertainty of the frequency measurements was estimated to be better than 20 kHz.All frequency components are referenced to a Rb standard.
Computational methods.The structural screening combined molecular mechanics, 7,8 semiempirical methods 9 and manual searches, covering a large dataset of initial structures.All starting geometries were re-optimized with hybrid (B3LYP 10 ) and double-hybrid (B2PLYP 11 ) density-functional theory methods incorporating D3 12 empirical dispersion corrections (Becke-Johnson 13 damping function).These computational models used the def2-TZVP basis set. 14One of the isomers of the PET-PET dimer in Table S2 did not converge.Frequency calculations were performed using the harmonic approximation at the same level of theory.The interaction energies were calculated considering the basis set superposition errors (BSSE). 15All calculations were performed using Gaussian 16, using tight conditions for geometry optimizations and the ultrafine default for integral calculations. 16The presence of non-covalent interactions was analysed using the NCIplot method, 17,18 based on a reduced gradient of electronic density.The physical contributions to the binding potential of the water clusters were estimated by energy decomposition analysis using second-order symmetry adapted perturbation theory 19 (SAPT), implemented in PSI4. 20

Figure S1.
The two observed conformations of the monomers of 2-phenylethanethiol (upper row) and 2-phenylethanol (lower row).The global minimum of the two molecules is conformer Gg(gauchegauche), orienting the polar groups to the  ring.However, the second most stable conformer is Ag (antiperiplanar-gauche) in 2-phenylethanetiol but (Cs) plane-symmetric At (anti-anti) in 2phenylethanol.Each conformer (except the symmetric At) has two equivalent enantiomeric species.
The relative energies (in kJ mol -1 ) in the drawing were calculated at the B3LYP-D3(BJ)/def2-TZVP level.probing the same UV transitions at 37627 cm -1 (Isomer Gg) and 37675 cm -1 (Isomer At) as in the hole burning and the simulations obtained using normal mode analysis (red traces) and anharmonic calculations (blue trace).A correction factor of 0.963 was used for the harmonic simulation.
(Lower panel) Similar comparison for the most abundant isomer of 2-phenylethanethiol, probing the UV transition at 37589 cm -1 (Isomer Gg).As can be seen, the S-H stretching band is too weak to be detected and therefore, the IDIR spectrum of the weaker isomer was not measured.

Figure S12.
The microwave spectrum of a co-expansion of 2-phenylethanethiol (PET) and 2phenylethanol (PEAL) in the 2-8 GHz frequency region (upper panel), and a section of the rotational spectrum illustrating the assignment of PET-PEAL (lower panel).The positive trace shows the experimental spectrum.The negative trace is the simulation using the fitted rotational constants for isomer 2 in Tables 1 and S3-S4.

Structural parameters
HBond donor: HSCC/ deg [b] 67.4 66.9 -61. [e] Standard errors in units of the last digit.e] Standard errors in units of the last digit. [e] Standard errors in units of the last digit.
Table S2.Conformational search for the homodimer of 2-phenylethanethiol (PET-PET) using the B2PLYP-D3(BJ) method (tight convergence, ultrafine integration) and the def2-TZVP basis set, and comparison with the experimental rotational parameters.

Experiment Isomer 1 Isomer 2 Isomer 3
GgGgLp GgGgLp GgGg Rotational parameters A / MHz [a] 390.86720(30) [e] 394.d] 289  / kHz 11.6 e] Standard errors in units of the last digit.e] Standard errors in units of the last digit.e] Standard errors in units of the last digit.

Figure S7 .
Figure S7.The most stable isomers of the 2-phenylethanethiol homodimer (PET-PET), classified by their intermolecular interactions and ordered by electronic energy (E, kJ/mol).Isomers in the orange and green rectangles exhibit a thiol-thiol S-H•••S hydrogen bond.Calculations were performed at the B2PLYP-D3(BJ)/def2-TZVP level.

Table S1 .
Conformational search for the homodimer of 2-phenylethanethiol (PET-PET) using the B3LYP-D3(BJ) method (tight convergence, ultrafine integration) and the def2-TZVP basis set, and comparison with the experimental rotational parameters.

Table S5 .
Atomic coordinates (principal inertial axes system) for isomer 1 of the homodimer of 2-

Table S6 .
Atomic coordinates (principal inertial axes system) for isomer 1 of the heterodimer of 2-

Table S7 .
Atomic coordinates (principal inertial axes system) for isomer 2 of the of the heterodimer

Table S8 .
List of observed rotational transitions of the homodimer of 2-phenylethanethiol (PET-PET)

Table S9 .
List of observed rotational transitions of the heterodimer of 2-phenylethanethiol and 2phenylethanol (PET-PEAL) and differences with the predicted transitions with the parameters of