A Barrierless Pathway Accessing the C9H9 and C9H8 Potential Energy Surfaces via the Elementary Reaction of Benzene with 1-Propynyl

The crossed molecular beams reactions of the 1-propynyl radical (CH3CC; X2A1) with benzene (C6H6; X1A1g) and D6-benzene (C6D6; X1A1g) were conducted to explore the formation of C9H8 isomers under single-collision conditions. The underlying reaction mechanisms were unravelled through the combination of the experimental data with electronic structure and statistical RRKM calculations. These data suggest the formation of 1-phenyl-1-propyne (C6H5CCCH3) via the barrierless addition of 1-propynyl to benzene forming a low-lying doublet C9H9 intermediate that dissociates by hydrogen atom emission via a tight transition state. In accordance with our experiments, RRKM calculations predict that the thermodynamically most stable isomer – the polycyclic aromatic hydrocarbon (PAH) indene – is not formed via this reaction. With all barriers lying below the energy of the reactants, this reaction is viable in the cold interstellar medium where several methyl-substituted molecules have been detected. Its underlying mechanism therefore advances our understanding of how methyl-substituted hydrocarbons can be formed under extreme conditions such as those found in the molecular cloud TMC-1. Implications for the chemistry of the 1-propynyl radical in astrophysical environments are also discussed.


Experimental Results
Laboratory frame. We monitored potential products formed from the reactive scattering of the 1-propynyl radical (CH 3 CC; 39 amu) with benzene (C 6 H 6 ; 78 amu) along Θ CM at m/z = 117 (C 9 H 9 + ) and 116 (C 9 H 8 + ) to assess the formation of a persistent reaction intermediate (adduct) and/or hydrogen loss reaction product, respectively. Signals were observed for each m/z value. However, the TOF spectra recorded at each m/z depict an identical pattern after scaling and thus originate from the same channel, namely the formation of C 9 H 8 (116 amu) by elimination of atomic hydrogen (1 amu). The signal at m/z = 117 therefore arises from the natural distribution of carbon atom isotopes yielding 13 CC 8 H 8 occurring at level of about 9.9%. Following confirmation of the atomic hydrogen loss product channel, we recorded 5.6 × 10 6 TOF spectra of nascent C 9 H 8 in the LAB frame at m/z = 116 in 2.5° intervals from 21.75° to 58.25° (Fig. 3). The TOF spectra were then normalized with respect to Θ CM and integrated to yield the laboratory angular distribution (Fig. 3), which is nearly symmetric around Θ CM thereby suggesting the presence of indirect reaction dynamics via one or more relatively long-lived C 9 H 9 intermediate(s) preceding dissociation to C 9 H 8 plus atomic hydrogen. The laboratory data were fit by forward convolution using a single reaction channel, namely CH 3 CC (39 amu) + C 6 H 6 (78 amu) → C 9 H 8 (116 amu) + H (1 amu) with a reaction cross section proportional to E C −1/3 for a barrier-less reaction within the line-of-centre model (section 5) 28 . To trace the source of the hydrogen atom emission, we substituted the benzene reactant with D6-benzene (C 6 D 6 ) and performed a second experiment with CH 3 CC (39 amu) plus C 6 D 6 (84 amu) to determine if the atom is lost from the 1-propynyl or benzene reactant. We probed the atomic hydrogen and atomic deuterium loss www.nature.com/scientificreports www.nature.com/scientificreports/ channels along Θ CM at m/z = 122 (C 9 H 2 D 6 + ) and 121 (C 9 H 3 D 5 + ), respectively. A very strong signal was recorded at m/z 121 while a weaker signal was discernible at m/z = 122. By comparison with the ratios recovered in the CH 3 CC -C 6 H 6 reaction for the 13 CC 8 H 8 /C 9 H 8 product(s), we conclude that the TOF spectrum at m/z = 122 owes to 13 C-enrichment in the form of 13 CC 8 H 3 D 5 and that only the atomic deuterium loss channel to form C 9 H 3 D 5 is observed. Importantly, the TOF recorded at Θ CM can be fit using the CM functions derived from the hydrogenated system, which suggests the products formed in each experiment are isotopologues (Fig. 4).
Centre-of-mass frame. In the laboratory frame, it is clear that the reaction product(s) are formed via the loss of an aromatic hydrogen atom forming one or more C 9 H 8 isomers. Details of the reaction coordinate are encoded in the product distribution and can be revealed in the centre-of-mass (CM) frame, from which we now examine via the CM translational energy P(E T ) and angular T(θ) flux distributions (Fig. 5). The P(E T ) terminates at 198 ± 27 kJ mol −1 and represents the maximum energy available E avail to the reaction system. Taken together with the collision energy E C of 40.8 ± 0.5 kJ mol −1 , we obtain a reaction energy for this process of ∆ r G = −157 ± 27 kJ mol −1 , where these quantities have been related through energy conservation via E avail = E C − ∆ r G. While the most probable E T occurs near 28 ± 4 kJ mol −1 , nascent C 9 H 8 products carry an average translational energy of 57 ± 8 kJ mol −1 , suggesting that only about 29 ± 6% of E avail is disposed into translational degrees of freedom. The high fraction of energy appearing as rovibrational excitation and nonzero peaking of the P(E T ) are markers for a reaction mechanism that forms products indirectly via activated C 9 H 9 intermediate(s) that must overcome a barrier to disscoiation 28 . The symmetry of the T(θ) distribution about θ = 90° along with the nonzero intensity at all angles together indicate the reaction proceeds through a long-lived C 9 H 9 intermediate complex. Furthermore, the maximum in the distribution at θ = 90° strongly suggests the C 9 H 9 intermediate decomposes via a transition state that emits atomic hydrogen perpendicular to its rotational plane, i.e. parallel to the total angular momentum vector as defined by the initial conditions of the experiment 28,29 .

Discussion
We now combine the experimental findings with electronic structure calculations to elucidate the underlying reaction dynamics and the nature of the structural isomer(s) produced by considering three aromatic C 9 H 8 isomers. These are indene (p1), 1-phenyl-1-propyne (p2), and 1-phenyl-1,2-propadiene (p3) with computed reaction energies of −262 ± 6, −151 ± 6, and −142 ± 6 kJ mol −1 , respectively. Considering the experimentally derived reaction energy of −157 ± 27 kJ mol −1 , the computational data support the formation of the C 9 H 8 isomers p2 and/ or p3 via hydrogen atom loss. Continuing on the basis of energetics, indene (p1) can be reasonably excluded from consideration among the detected reaction products where the formation of the indene isomer would increase the maximum translational energy beyond the error limits (±27 kJ mol −1 ) of the P(E T ) to about 300 kJ mol −1 . By examining the reaction potential energy surface (PES), we can gain deeper insight into the underlying reaction mechanism to form C 9 H 8 isomers via the bimolecular reaction of 1-propynyl with benzene ( Fig. 6; Supplemental Information).
The 1-propynyl radical can add to the π-electron system of the benzene molecule without an entrance barrier (Fig. S1) forming a C 9 H 9 intermediate i1 featuring the propynyl substituted to the six-membered ring that lies 208 kJ mol −1 below the energy of the separated reactants. Intermediate i1 can dissociate via ts2 by eliminating the newly-formed sp 3 hydrogen atom on the six-membered ring (Fig. S2) nearly perpendicular to the orbital plane of the molecule at an angle of 88.5° (Fig. 7) to form 1-phenyl-1-propyne (p2). The ts2 exit transition state is characterized by the restoration of aromaticity and hence is rather tight at 44 kJ mol −1 above the p2 product channel. Alternatively, the sp 3 hydrogen atom on i1 can migrate to the C1 exocyclic position over a 148 kJ mol −1 barrier (ts1) to form i2 gaining an additional 76 kJ mol −1 in stability. From i2, each of the three C 9 H 8 isomers p1, p2, and p3 can be accessed (Fig. S2). Elimination of the exocyclic C1 or methyl hydrogen atoms results in the formation of p2 or 1-phenyl-1,2-propadiene (p3) via ts3 or ts4, respectively. Unlike the i1 → p2 + H pathway where the hydrogen atom is ejected nearly parallel to the total angular momentum vector via ts2, the exit transition states  www.nature.com/scientificreports www.nature.com/scientificreports/ ts3 and ts4 connecting i2 → p2 + H and i2 → p3 + H eliminate the hydrogen atom at angles of 20° and 0° with respect to the orbital plane of the molecule (Fig. 7). The minimum energy path to p1, not explicitly considered here since it was computed previously, involves extensive isomerization through an additional 8 intermediates and 9 transition states 1,5 .
Although both isomers p2 and p3 can account for the experimental translational energy of C 9 H 8 , a close inspection of the C 9 H 9 PES alongside our isotopically labelled study permits a clarification of the experimental reaction dynamics. The CH 3 CC -C 6 D 6 reactive scattering experiment exposed the exclusive formation of C 9 H 3 D 5 via elimination of atomic deuterium originating at the benzene reactant. Isotopologues of indene p1, if formed, would give signals for atomic hydrogen (C 9 H 2 D 6 ) and deuterium loss (C 9 H 3 D 5 ) at a ratio of about 1:1 via a decomposing bicyclic intermediate 1,5 , whereas D6-p3 (C 6 D 5 CDCCH 2 ) would form by elimination of a methyl hydrogen atom from D6-i2. Hence, only D5-p2 (C 6 D 5 CCCH 3 ) can account for this observation via dissociation of intermediates D6-i1 or D6-i2 (Fig. 8). Recalling that the best-fit T(θ) distribution peaked sideways at 90°, we note that of the three dissociation channels considered, only the i1 → ts2 → p2 + H path supports this finding where the hydrogen emitted by ts2 occurs nearly parallel to the principal axis with the greatest moment of inertia (I c = 1.16 × 10 −44 kg m 2 ) at an angle of 1.5°. For an asymmetric top decomposing at the top of an exit barrier, microcanonical transition state theory has shown that hydrogen atom emission from a long-lived complex along the principal axis of the greatest moment of inertia (C) results in sideways scattering 30,31 , as observed in this experiment (Fig. 5). The i2 → ts3 → p2 + H and i2 → ts4 → p3 + H pathways do not give rise to sideways scattering where ts4 dissociates in the AB plane and ts3 emits its hydrogen atom nearly perpendicular to its C axis. We also note the average E T for C 9 H 8 this experiment is 57 ± 8 kJ mol −1 and is a close match to the exit barrier of ts2 i1 → p2 + H ts3 i2 → p2 + H ts4 i2 → p3 + H 89°0°F igure 7. Computed geometries of the exit transition states leading to the formation of 1-phenyl-propyne (p2) and 1-phenyl-1,2-propadiene (p3). The angle for each departing hydrogen atom is given with respect to the rotational plane of the decomposing complex.
p2-d 5 + D i2-d6 p3-d 6 + H www.nature.com/scientificreports www.nature.com/scientificreports/ 44 ± 5 kJ mol −1 for the p2 + H product channel. At the experimental collision energy of 40.8 ± 0.5 kJ mol −1 each of the p1-p3 product channels are open, however, considering the barriers to isomerization on the C 9 H 9 PES, the low energy route to 1-phenyl-1-propyne (p2) via intermediate i1 is expected to dominate in a statistically-conforming molecular system, where the i1 → i2 barrier (ts1) is 47 kJ mol −1 greater than that required for dissociation to p2 + H via ts2. This is supported by our RRKM calculations which suggest that p2 accounts for more than 99% of the total C 9 H 8 yield. Combined, our computational analysis and experimental data indicate the formation of 1-phenyl-1-propyne (p2) as the predominant product formed in the bimolecular reaction of 1-propynyl with benzene via a simple addition/elimination mechanism, i.e. CH 3 CC + C 6 H 6 → i1 → p2 + H.
The temperature-dependent rate constant for the barrierless reaction of 1-propynyl (CH 3 CC) with benzene (C 6 H 6 ) was evaluated using the long-range transition state theory, where the capture rate is assessed considering dipole-quadrupole, dipole-induced dipole and dispersion interactions between the reactants. The calculations gave the value of the rate constant increasing from 3.7 × 10 −10 cm 3 molecule −1 s −1 at 10 K to 5.4 × 10 −10 and 6.5 × 10 −10 cm 3 molecule −1 s −1 at 100 and 300 K, respectively. These results corroborate our hypothesis that the reaction of the 1-propynyl radical with benzene should be fast even in cold molecular clouds and the rate constant can be included in astronomic kinetic models.
Lastly, we compare our results to the analogous phenyl (C 6 H 5 ; X 2 A 1 ) plus methylacetylene (CH 3 CCH; X 1 A 1 ) reaction, where a hydrogen atom has been transferred from benzene to 1-propynyl -that, despite its obvious similarity with the title reaction, follows completely distinct reaction dynamics. In a low-pressure (300 Torr) high temperature (1200-1500 K) chemical reactor the phenyl reacts with methylacetylene to form the C 9 H 8 isomers indene (p1), 1-phenyl-1-propyne (p2), 1-phenyl-1,2-propadiene (p3) at fractions of 10, 82, and 8% respectively 1 ; in a high-energy (E C = 140 kJ mol −1 ) crossed molecular beams experiment, a similar outcome was observed with the formation of 1-phenyl-1-propyne (p2) via an addition/elimination reaction mechanism via a short-lived reaction adduct 8 . In the low-energy (E C = 45 kJ mol −1 ) case, however, the C 6 H 5 plus CH 3 CCH reaction results in the formation of indene (p1) as confirmed through a series of isotopic experiments and supported by RRKM calculations in the zero-pressure limit suggesting the indene isomer accounted for 81-91% of all hydrogen-loss products formed 5 . Although the 1-phenyl-1-propyne (p2) isomer was predicted to occur at a level of 7-10%, it could not be explicitly traced in the crossed molecular beam experiments. The most recent theory describing the C 6 H 5 plus CH 3 CCH reaction suggests the phenyl addition to methylacetylene is inhibited by at least a 6 kJ mol −1 entrance barrier 1 , and results in the formation of the geometric isomer of i2 with the methyl-group set trans to the phenyl ring that undergoes cis-trans isomerization to i2 via a barrier of only 20 kJ mol −1 ; all remaining transition states are energetically lower than the separated reactants and the system proceeds spontaneously to indene (p1) via annulation at the phenyl moiety. Isomerization to i1, the initial intermediate formed in the title reaction, is inhibited by a rather large barrier of 221 kJ mol −1 and thus is not competitive with the low-energy path to indene formation. Hence the nature of the entrance channel (barrier vs barrierless), initial intermediate formed, and the high-energy i1-i2 transition state (ts1) underlie the distinct reaction dynamics followed in the analogous CH 3 CC + C 6 H 6 and CH 3 CCH + C 6 H 5 reactions under similar experimental conditions. conclusion Crossed molecular beams (CMB) experiments, combined with electronic structure and statistical calculations, were exploited to investigate the reaction dynamics of the of 1-propynyl (CH 3 CC; X 2 A 1 ) reaction with benzene (C 6 H 6 ; X 1 A 1g ) under single-collision conditions. The reaction was found to proceed indirectly and is initiated by the barrierless addition of 1-propynyl to a carbon atom on the benzene ring forming a low-lying intermediate i1 that eliminates the hydrogen atom to form 1-phenyl-1-propyne (C 6 H 5 CCCH 3 ; p2) in an overall exoergic reaction (experimental, −157 ± 27 kJ mol −1 ; computational, −151 ± 6 kJ mol −1 ). Our RRKM analysis suggests that this pathway should account for more than 99% of C 9 H 8 products. Interestingly, differences in the entrance channel for the title reaction as compared to phenyl (C 6 H 5 ) addition to methylacetylene (CH 3 CCH) give rise to distinct reaction products under similar experimental conditions, with the latter forming the PAH indene (C 9 H 8 ; p1) as the major hydrogen-loss product. Compared to prior experiments probing access to the C 9 H 9 potential energy surface, such as with phenyl addition to methylacetylene and benzyl (C 6 H 5 CH 2 ) addition to acetylene (C 2 H 2 ), the reaction mechanism uncovered here is distinguished by being the first barrierless path to C 9 H 8 where the 1-propynyl radical can disrupt aromaticity at benzene without a penalty. Despite the advantage, this pathway is not likely to compete with PAH producing reaction schemes in high temperature combustion environments where the availability of 1-propynyl radicals are likely linked to the decomposition of and/or hydrogen abstraction from the hydrocarbon methylacetylene. Under these conditions, methylacetylene is more likely to -in the case of H atom reactions -form allene via hydrogen-assisted isomerization, or react producing acetylene by methyl displacement or the propargyl radical (CH 2 CCH) by hydrogen abstraction, rather than form 1-propynyl by elimination or abstraction of its acetylenic hydrogen atom 32,33 .
The isolobal ethynyl (C 2 H; X 2 Σ + ) radical, where the methyl group has been replaced by a hydrogen atom, presents an analogous chemistry to that uncovered for the 1-propynyl radical. Using the D1-ethynyl (C 2 D) and D6-benzene isotopologues, molecular beams experiments demonstrated the formation of D6-phenylacetylene (C 6 D 5 CCD) in a mechanism similar to the title reaction where dissociation immediately follows the barrierless formation of a low-lying C 8 D 7 intermediate 34,35 . In each case, the reactivity is localized to the radical end of the alkyne, rendering the D/CH 3 substituents mere spectators in the reaction coordinate. The addition/elimination reaction mechanism via a (pseudo)barrierless entrance channel is also observed in the isoelectronic boronyl (BO; X 2 Σ + )-benzene and cyano (CN; X 2 Σ + )-benzene reaction systems. Addition of the boronyl radical followed by hydrogen atom loss results in the overall exoergic formation of phenyl oxoborane (C 6 H 5 BO) 36,37 ; also nearly 20 years before its detection in the interstellar medium 38 , molecular beams experiments demonstrated that cyanobenzene (C 6 H 5 CN) is formed via a low-energy, barrier-less pathway involving two neutral reactants that, www.nature.com/scientificreports www.nature.com/scientificreports/ in theory, proceeds spontaneously at 0 K and proposed the aromatic molecule to be present in cold molecular clouds 39,40 .
Although the 1-propynyl radical has not yet been detected in astrophysical sources, methylacetylene -a potential precursor -is both abundant and prolific in different regions of space 41,42 , and could provide a source of 1-propynyl radicals in the photodissociation region of molecular clouds or near carbon-rich AGB stars via photolytic cleavage of the acetylenic C-H bond (CH 3 CCH + hν → CH 3 CC + H) [24][25][26] . To establish the activity of the 1-propynyl radical in astrochemistry, laboratory characterizations of its structure are needed to aid observation efforts toward its detection and ultimately its inclusion in astrochemical modelling networks alongside its low-energy isomer, propargyl (CH 2 CCH). Considering that all barriers are below the energy of the separated reactants, 1-propynyl addition to benzene is a plausible reaction scheme for the low temperature extremes that persist in cold molecular clouds and strongly suggests the existence of 1-phenyl-1-propyne (C 6 H 5 CCCH 3 ) in molecular clouds, though additional spectroscopic data are required to support its assignment among existing astronomical data 43 .

Methods
Experimental. The reactions of 1-propynyl (CH 3 CC; X 2 A 1 ) with benzene (C 6 H 6 ; X 1 A 1g ) and D6-benzene (C 6 D 6 ; X 1 A 1g ) were carried out under single collision conditions exploiting a universal crossed molecular beams machine at the University of Hawaii [44][45][46][47][48][49] . The pulsed 1-propynyl radical beam was produced by photodissociation of 1-iodopropyne (CH 3 CCI; TCI, 99%+) seeded at a level of 0.5% in helium (99.9999%; AirGas) at 193 nm (Complex 110, Coherent, Inc.) at 30 Hz and 20 mJ per pulse in the primary source chamber [50][51][52][53] . This gas mixture was stored in a Teflon-coated sample cylinder and was introduced into a piezoelectric pulsed valve operating at 60 Hz, pulse widths of 80 μs, peak voltages of −400 V, and at 760 Torr backing pressure. The pulsed 1-propynyl beam passes through a skimmer and is then velocity selected by a four-slot chopper wheel rotating at 120 Hz. On-axis (Θ = 0°) characterization of the primary beam determines a peak velocity v p of 1658 ± 12 m s −1 and speed ratio S of 7.1 ± 0.3. This section of the primary beam collides perpendicularly with a supersonic beam of (D6)-benzene prepared in the secondary source chamber. A pulsed valve in the secondary source operated at a repetition rate of 60 Hz, a pulse width of 80 μs, and a peak voltage of −400 V, generated a pulsed molecular beam of (D6)-benzene [Aldrich Chemistry; ≥99.9% (>99.9% atom D)] seeded in argon (99.9999%, AirGas) at fractions of 10% at 550 Torr. The (D6)-benzene peak velocities were determined to be v p = 622 ± 10 ms −1 with S = 19.3 ± 0.6 (v p = 623 ± 9 ms −1 , S = 19.3 ± 0.4) resulting in a nominal collision energy E C of 40.8 ± 0.5 kJ mol −1 (41.8 ± 0.6 kJ mol −1 ) and centre-of-mass angle Θ CM of 36.9 ± 0.5° (39.0 ± 0.5°) ( Table 1). We note that any propargyl (CH 2 CCH; X 2 B 1 ) radicals formed in the photodissociation of 1-iodopropyne as the result of isomerization from 1-propynyl do not react under our experimental conditions with benzene, where addition to benzene is inhibited by a barrier of 65-75 kJ mol −1 54 ; this barrier cannot be overcome at collision energies of 40.8 ± 0.5 kJ mol −1 in our present experiments.
The crossed molecular beams machine employs two beam sources affixed at 90 degrees and a triply differentially pumped universal detector that is rotatable in the plane defined by the reactant beams 55 . The detector houses an electron impact ionizer coupled to a quadrupole mass spectrometer that permits filtering of ionized (80 eV) products according to mass-to-charge ratio. Ions are ultimately registered by a photomultiplier tube and filed into bins by multichannel scaling according to time of arrival to produce a time-of-flight (TOF) spectrum. The laboratory data are forward-convoluted to the centre-of-mass (CM) frame and the resulting CM translational energy P(E T ) and angular T(θ) flux distributions are analysed to inform the reaction dynamics 56,57 . Errors of the P(E T ) and T(θ) functions are determined within the 1σ error limits of the accompanying LAB angular distribution while maintaining a good fit of the laboratory TOF spectra.
theoretical. Geometries of the reactants, intermediates, transition states, and products of the CH 3 CC -C 6 H 6 system were optimized at the density functional B3LYP/6-311 G(d,p) level of theory 58,59 . Vibrational frequencies were computed at the same theoretical level and were used for the evaluation of zero-point vibrational energy corrections (ZPE) and in calculations of rate constants. Energies were refined by single-point calculations using the model chemistry G3(MP2,CC)//B3LYP/6-311 G(d,p) level of theory [60][61][62] . This composite approach normally provides a chemical accuracy of 3-6 kJ mol -1 for the relative energies and 0.01-0.02 Å for bond lengths as well as 1-2° for bond angles 62 . The ab initio calculations were performed using the GAUSSIAN 09 63 and MOLPRO 2010 64 program packages. Rate constants of all pertinent unimolecular reaction steps on the C 9 H 9 PES following initial association of 1-propynyl with benzene were computed using Rice-Ramsperger-Kassel-Marcus (RRKM) theory [65][66][67] , as functions of available internal energy of each intermediate or transition state, where numbers and densities of states were obtained within the harmonic approximation using B3LYP/6-311 G(d,p) computed frequencies. The internal energy was taken as a sum of the collision energy and a negative of the relative energy of a species with respect to the reactants (the chemical activation energy). One energy level was considered Beam v p (m s −1 ) S E C (kJ mol −1 ) Θ CM (degree) CH 3 CC (X 2 A 1 ) 1658 ± 12 7.1 ± 0.3 C 6 H 6 (X 1 A 1g ) 622 ± 10 19.3 ± 0.6 40.8 ± 0.5 36.9 ± 0.5 C 6 D 6 (X 1 A 1g ) 623 ± 9 19.3 ± 0.4 41.8 ± 0.6 39.0 ± 0.5 Table 1. Peak velocities (v p ) and speed ratios (S) of the 1-propynyl (CH 3 CC), benzene (C 6 H 6 ), and D6-benzene (C 6 D 6 ) beams along with the corresponding collision energies (E C ) and center-of-mass angles (Θ CM ). www.nature.com/scientificreports www.nature.com/scientificreports/ throughout as at a zero-pressure limit. Then, RRKM rate constants were utilized to compute product branching ratios by solving first-order kinetic equations within steady-state approximation 68 using a newly developed computer code 52 . Dipole and quadrupole moments and isotropic polarizabilities of CH 3 CC and C 6 H 6 required for the long-range transition state theory 69 calculations of the entrance channel rate constant were computed at the density functional ωB97XD level of theory 70 with Dunning's correlation-consistent cc-pVTZ basis set 71 .

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding authors on reasonable request.