A rotational and vibrational investigation of phenylpropiolonitrile (C_6H_5C_3N)
Zachary Buchanan, Kin Long Kelvin Lee, Olivia Chitarra, Michael C. McCarthy, Olivier Pirali, Marie-Aline Martin-Drumel
AA rotational and vibrational investigation of phenylpropiolonitrile (C H C N) Zachary Buchanan a,b, ∗ , Kin Long Kelvin Lee c , Olivia Chitarra a , Michael C. McCarthy c , Olivier Pirali a,d ,Marie-Aline Martin-Drumel a a Université Paris-Saclay, CNRS, Institut des Sciences Moléculaires d’Orsay, 91405 Orsay, France b Department of Chemistry, The University of California Davis, Davis, CA, USA c Center for Astrophysics | Harvard & Smithsonian, Cambridge, Massachusetts 02138, United States d SOLEIL Synchrotron, AILES beamline, l’Orme des Merisiers, Saint-Aubin, 91190 Gif-sur-Yvette, France
Abstract
The evidence for benzonitrile (C H CN) in the starless cloud core TMC–1 makes high-resolution studiesof other aromatic nitriles and their ring-chain derivatives especially timely. One such species is phenylpro-piolonitrile (3-phenyl-2-propynenitrile, C H C N), whose spectroscopic characterization is reported here forthe first time. The low resolution (0.5 cm − ) vibrational spectrum of C H C N has been recorded at far- andmid-infrared wavelengths (50–3500 cm − ) using a Fourier Transform interferometer, allowing for the assign-ment of band centers of 14 fundamental vibrational bands. The pure rotational spectrum of the species hasbeen investigated using a chirped-pulse Fourier transform microwave (FTMW) spectrometer (6–18 GHz), acavity enhanced FTMW instrument (6–20 GHz), and a millimeter-wave one (75–100 GHz, 140–214 GHz).Through the assignment of more than 6200 lines, accurate ground state spectroscopic constants (rotational,centrifugal distortion up to octics, and nuclear quadrupole hyperfine constants) have been derived from ourmeasurements, with a plausible prediction of the weaker bands through calculations. Interstellar searches forthis highly polar species can now be undertaken with confidence since the astronomically most interestingradio lines have either been measured or can be calculated to very high accuracy below 300 GHz. Keywords: pure rotation, vibration, astrophysical species, PAH derivative, phenylpropiolonitrile
1. Introduction
The ubiquity of aromatic molecules is closely-correlated to their stability and lack of reactivity,with functionalized aromatics serving as a commonmotif in biological chemistry. Polycyclic aromatichydrocarbons are a prominent class of aromatics;they are well-known constituents in the outflows ofcertain evolved carbon stars and common byprod-ucts in incomplete combustion processes [1, 2, 3].From a purely spectroscopic viewpoint, a numberof simple derivatives of benzene —the prototypicalaromatic ring C H — have either not been charac-terized at all or at insufficient resolution to under-take an astronomical search in the coldest most qui-escent molecular clouds in space. The recent discov-ery of benzonitrile —the simplest aromatic nitrile(C H CN or PhCN hereafter)— using radio obser-vations towards Taurus Molecular Cloud (TMC-1) ∗ [email protected] [4] has reignited the interest in nitrogen-containingaromatics generally and CN-functionalized aromat-ics specifically [5]. The subsequent identificationof cyanocyclopentadiene, C H CN, in the samecloud [6] has only intensified this interest. Ni-triles are also known to be important constituentsin the chemistry of Titan’s atmosphere [7], andthey are very prominent in the interstellar medium(ISM), accounting for roughly 20% of the 220 or somolecules (47) detected in the ISM to date, includ-ing cyanopolyynes as long as HC N [8].The presence of a nitrile group often imparts amolecule with a large permanent dipole momentand an intense rotational spectrum. In the case ofPhCN, replacing a single H atom in benzene with aCN group transforms an otherwise highly symmet-ric ring into a highly polar species ( µ a = 4 . D;[9]) thereby greatly aiding its detection both inthe laboratory and in space. Whilst the rotationalspectrum of PhCN has been known for more than
Preprint submitted to Journal of Molecular Spectroscopy January 28, 2021 a r X i v : . [ a s t r o - ph . GA ] J a n alf-a-century [10], its interstellar detection wasgreatly aided by measurements at very high accu-racy ( i.e. , at a resolving power f / ∆ f > ) at cen-timeter wavelengths [4, 9]. In light of this finding,new —or in many cases improved— high-resolutionstudies of molecules closely related in structure andcomposition are worth pursuing.In this work, we report a combined pure ro-tational and vibrational investigation of a deriva-tive of PhCN, phenylpropiolonitrile (3-phenyl-2-propynenitrile, C H C N, abbreviated as PhC N inthe following) where the nitrile group is replaced bya longer chain variant (C N) (Fig. 1). This specieshas previously been identified as a possible prod-uct from the reaction between the cyano radicaland phenylacetylene (C H CCH) [11]. To the bestof our knowledge, however, spectroscopic investiga-tions have been limited to experimental and the-oretical vibronic spectroscopy [12, 13] while high-resolution, rotationally-resolved studies are appar-ently lacking. In light of the large permanent elec-tric dipole moment calculated here (5.9 D), labora-tory measurement of rotational frequencies wouldallow astronomical searches for this ring-chain tobe undertaken with little or no ambiguity. If foundin space, the abundance of PhC N would provide akey test for models of aromatic chemistry which arepoorly constrained at present. The infrared spec-trum of PhC N is also of interest as a point of com-parison with other benzene derivatives whose vi-brational spectra are often plagued by a myriad ofperturbations and resonances, notably Fermi andDarling-Dennison in the C − H and C −−−
C stretchingregions [14, 15, 16], in addition to Fermi and Cori-olis interactions for low frequency ( ∼ cm − ),large amplitude modes that are prominent at roomtemperature [17, 18, 19]—even under astronomicalconditions. Figure 1: Molecular structure of PhC N; a and b principalinertial axes are indicated (in red and green, respectively);the c -axis extends out of the molecular plane and is notshown.
2. Experimental and computational meth-ods
Calculations were performed using the Gaus-sian’16 suite of electronic structure programs [20].The goal of these calculations was to provide accu-rate estimates of spectroscopic parameters, includ-ing rotational constants, dipole moment, rotation-vibration corrections, and fundamental vibrationalfrequencies using second-order vibrational pertur-bation theory (VPT2). Geometry optimizations ofPhC N and four of its isomers—namely the iso-cyanide analog PhCCNC and three variants whereCN is substituted for a H atom on the ring ofphenyl acetylene to yield ortho-, meta-, and para -cyanoethynylbenzene (CEB)—were carried out atthe ω B97X-D/cc-pVQZ level on an ultrafine in-tegration grid, in which optimized structures arethose in which convergence to better than − ofthe root-mean-squared (RMS) value of the gradi-ent has been achieved. For PhC N, we also per-formed (an)harmonic frequency analysis, obtainingboth the harmonic and anharmonic vibrational fre-quencies and intensities at the same level of the-ory; with the latter, the rotation-vibration interac-tion constants α were computed. Cartesian coor-dinates of the optimized equilibrium structures canbe found in Tables S1-S5 in the Supporting Infor-mation.In addition to spectroscopic parameters, we havealso performed rudimentary thermochemical calcu-lations on the relative energies of PhC N and its iso-mers using the G3//B3LYP composite method [21],which has been shown to provide near-chemical ac-curacy at excellent computational cost [22]. Giventhat semi-empirical methods typically perform beston closed-shell molecules with relatively simplerelectronic structure, as the ones studied here, webelieve these calculations provide a quantitativedetermination of relative stabilities, accurate to ± kJ/mol. The gas-phase vibrational spectrum of PhC Nwas recorded in the far-infrared (far-IR) and mid-infrared (mid-IR) using the Bruker IFS 125 FT in-terferometer located at the AILES beamline of theSOLEIL synchrotron facility (no synchrotron radia-tion was used in the present study) [23]. For the far-IR measurements, the spectrometer was equippedwith a µm mylar-silicon composite beamsplitter2nd a liquid helium-cooled silicon bolometer. AKBr beamsplitter and a sensitive HgCdTe detec-tor, equipped with a cryogenically cooled entranceiris and optical filters [24], were used in the mid-IRregion. Vapor of PhC N was injected in a White-type multipass cell aligned to yield a 150 m opti-cal path length [25] and isolated from the interfer-ometer by 50 µm-thick polypropylene windows inthe far-IR and wedged ZnSe windows in the mid-IR range. In both spectral regions, the interfer-ometer was continuously evacuated to − mbarto minimize absorption from residual water. Spec-tra were recorded at a resolution of 0.5 cm − us-ing a globar light source and an entrance iris of4 mm, and consist of 100 and 500 co-added inter-ferograms for the far-IR and mid-IR regions, re-spectively. Both spectra were recorded at a samplepressure of 5 µbar. It is worth noting that the rota-tional structure within the vibrational bands couldnot be resolved even at the highest resolution of thespectrometer (0.001 cm − ). Microwave measurements were performed usinga chirped-pulse Fourier transform microwave (CP-FTMW) spectrometer located at the Center forAstrophysics [26] which operates between 8 and18 GHz. About 0.3 g of solid PhC N was intro-duced into a reservoir located behind the pulsednozzle and Ne carrier gas (at a flow of ∼
20 sccm atstandard temperature and pressure) passed throughthe reservoir; the resulting gas mixture was then in-jected into the vacuum chamber by operating thepulsed valve at a very low repetition rate (5 Hz).Because the vapor pressure of PhC N is relativelyhigh at room temperature (several mbar), it wasnot necessary to heat the reservoir to observe rota-tional lines in the CP-FTMW spectrum with goodsignal-to-noise ratios.Approximately 17,000 nozzle pulses, each probedby 10 microwave chirps, were acquired. Additionaldetails on the experimental set-up are provided inRef. [26]. The resulting spectrum, with electronicsartifacts and well-known contaminant lines ( e.g. ,acetone) removed, is displayed in Fig. 2. Transi-tions with ≤ J (cid:48)(cid:48) ≤ and K (cid:48)(cid:48) a ≤ are visible. In parallel to the present investigation, purerotational transitions of PhC N were also iden-
Figure 2: Experimental CP-FTMW spectrum (in black) ofPhC N in comparison with a simulation at T rot = 1 K (inpurple) using the calculated set of rotational constants; thesimulation is plotted here with negative intensity to moreeasily compare the two spectra.
Left panel: the full spec-trum, noting that a few lines originating from impurities arevisible around 10 GHz;
Right panel: expanded view aroundthe q R(7) transitions, as indicated by the gray rectangle onthe left panel. The simulation was performed using PGO-PHER [27]. tified by several of the co-authors of this paperwhile analyzing the discharge products of a ben-zene/nitrogen (N ) mixture [28] . Using a cavity-enhanced FTMW (CE-FTMW) spectrometer oper-ating between 6 and 40 GHz, transitions of PhC Nwere measured at roughly ten times higher spec-tral resolution than can be achieved with the CP-FTMW instrument. In the following, we onlybriefly summarize the experimental conditions rel-evant to this work; further details about the CE-FTMW instrument and the discharge mixture ex-periments are presented in Refs. [26] and [28], re-spectively.PhC N was synthesized by subjecting a mix-ture of C H and N diluted heavily in Ne to ahigh-voltage discharge (800 V); typical flow rateswere 14, 12, and 20 sccm respectively. Comparedto the CP-FTMW measurements where PhC Nwas the precursor, the discharge experiment resultsin a slightly higher rotational temperature, i.e. , ∼ K and produces rotational lines with demon-stratively lower signal-to-noise ratio. Nevertheless,at the highest resolving power ( ∼ ) the nitro-gen nuclear-quadrupole structure for several low- K a transitions ( K (cid:48)(cid:48) a = 0 − ) in the 6–20 GHz regionwas partially resolved. The room temperature spectrum of PhC N wasrecorded at ISMO with an absorption spectrome-ter in which a frequency multiplication scheme is3sed to generate millimeter-wave radiation [29]. Bycombining the output of a radiofrequency synthe-sizer (Rohde & Schwarz) operating in the 2–20 GHzregion with one of two amplifier / multiplier chains,it is possible to produce broadly tunable radia-tion with modest power (of a few mW) through-out the millimeter band; a Radiometer PhysicsGmbH (RPG) SMZ110 for 75–97 GHz, and a Vir-ginia Diodes Inc. (VDI) for 141–214 GHz.The millimeter-wave radiation was collimated us-ing a 10 mm focal length Teflon lens into a 2 mlong Pyrex absorption cell and further focused ontoa Schottky diode detector from VDI, using a sec-ond identical lens. The input radiation was fre-quency modulated at a frequency of 48.2 kHz anda commercial lock-in amplifier (Ametek) demodu-lated the signal at the second harmonic. The spec-trum was recorded using 50 kHz frequency steps,a time constant of 50/100 ms, and an FM devi-ation of 200/250 kHz (where the two values referthe lower/higher frequency measurements, respec-tively).A flow of PhC N, not exceeding a pressure of2 µbar, was pumped through the cell and evacu-ated by a turbomolecular pump. Above this pres-sure, significant pressure broadening of the rota-tional lines was observed.
3. Results and discussion
PhC N belongs to the C symmetry group andpossesses a ˜X A electronic ground state with 39modes of vibration following the irreducible repre-sentation Γ = 14 A ⊕ ⊕ ⊕
13 B . All vibra-tional modes except those of A symmetry are IRactive; A and B modes correspond to in-plane vi-brations (respectively a - and b -type bands) while B modes are out-of-plane vibrations ( c -type bands).PhC N is a prolate asymmetric top molecule witha permanent dipole moment of 5.9 Debye along the a axis (axis of the C N bonds, Fig. 1), according toour calculations. PhC N also has nitrogen nuclearquadrupole hyperfine structure, but this splitting isonly partially resolved at low frequency in the CE-FTMW measurements. Given that the molecule is C symmetry, we also include statistical weightsfor equivalent exchangeable nuclei for correct tran-sition intensities. Identical to PhCN, there are twosets of equivalent hydrogen atoms ( I = I = ) that give rise to Fermi-Dirac statistics for symmet-ric (even K ) and antisymmetric (odd K ) rotationalstates with a ratio of
10 : 6 [4].
The infrared spectrum of PhC N is presented inFig. 3 together with a simulation of the vibrationalfundamentals predicted by our anharmonic quan-tum chemical calculations. At our experimentalresolution, the rotational contour of most bands isevident in the spectrum. Because the simulation isin very good agreement with the experiment, nu-merous assignments can be made with confidence.For ambiguous assignments, two criteria can be in-voked: (i) c -type bands (B symmetry) exhibit asharp Q-branch; and (ii) a -type bands (A ) are usu-ally narrower than the others (as can be seen on thesimulated spectrum). Out of the 36 active infraredmodes of PhC N, 14 can be assigned with little orno ambiguity while 9 other bands have tentative as-signments at this juncture; the remaining bands areeither weak or are predicted in crowded regions ofthe spectrum (see Table 1 for a detailed list of theproposed assignments). Experimental band centersare taken as the frequency of the Q -branch whenone exists ( c -type bands), equidistant between the P - and R -branches, or at the maximum of the enve-lope when no clear contour is visible. Considering c -type Q -branches are several wavenumbers wide,an accuracy of ± cm − can be expected for theband centers with B symmetry. For the others, aconservative value of ± cm − is proposed.A few of the assignments proposed in Fig. 3 andTable 1 warrant further discussion. While the as-signment of the ν band (out-of-plane backbonemotion) is relatively secure thanks to the presenceof a sharp Q -branch, the band appears strongerthan expected from our calculations. A possibleexplanation is that ν (C −−− C − C −−− N rocking), pre-dicted roughly at the same frequency as ν , con-tributes to the experimental band profile, thus en-hancing absorption. In absence of any specific fea-ture arising from ν on the experimental spec-trum, we have chosen to assign this band to thesame band center as ν , at 69 cm − . Slightlyhigher in frequency are ν and ν , which in-volve similar nuclear motion as ν and ν andare also predicted to lie close in energy (at 192and 221 cm − , respectively), which in combinationlikely yield the feature observed around 200 cm − .It is not straightforward, however, to unambigu-ously assign each band center, in part because ν able 1: Fundamental vibrational bands (position and intensity) of PhC N from the quantum chemical calculations performedin this work at the harmonic and anharmonic levels of theory, and proposition of assignments. Modes are numbered followingthe anharmonic calculations frequency order. Modes energy are in wavenumbers, intensities are in km/mol, δ values are in %.The experimental assignments are split in two categories, the relatively secure ones (column “Assign.") and the tentative ones(column “Prop."); δ values of the latter are reported in italics. Mode Harm. Calc. Anharm. Calc. Exp. ν Sym. Energy Int. Energy Int. Assign. Prop. δ a -4.9/-3.0 -0.8/-3.1
12 A
981 0.1 960 0.013 A
706 3.4 693 3.514 A
371 0.8 362 0.515 A
876 0.0 862 0.017 A
413 0.0 412 0.018 B
967 2.8 953 2.4 920 -3.520 B
792 38.7 784 32.7 757 -3.421 B
716 36.7 706 42.7 687 -2.622 B
576 13.2 568 12.4 532 -6.323 B
523 0.2 515 0.324 B
384 2.8 377 2.8 359 -4.925 B
195 2.6 192 2.5 185 -3.4
26 B
74 0.6 71 0.7 69 -3.327 B
28 B
32 B -0.9
33 B
644 0.0 635 0.036 B
584 0.5 574 0.437 B
511 5.3 504 5.6 482 -4.338 B
226 2.3 221 2.4 209 -5.6
39 B
71 1.1 71 1.2 69 -2.7 a δ = (Exp . − Anharm . Calc . ) / Anharm . Calc . ×
500 1000 1500 2000 2500 3000Wavenumber / cm
100 200 300 400 500 0.000.01
19 10 9 734 30
11 32 31 Wavenumber / cm C a l c u l a t e d i n t e n s i t y / a r b . u . T r a n s m i tt a n c e Experiment SimulationA B B Figure 3: Experimental (top traces, in black) and simulated (bottom traces, at 300 K, where different colors correspond todifferent symmetries) vibrational spectrum of PhC N. The three bottom panels are expanded portions of the full spectrum,as indicated by gray rectangles. Simulation performed using the PGOPHER software [27] and the results of the anharmoniccalculations (band centers, intensities, and rotational constants) and normalized to the strongest IR active mode ( ν ). Thesimulations are inverted relative to the observed spectrum solely for comparison purposes. Secure band assignments areindicated by dash lines; additional labelled bands are those for which an assignment is proposed; simulated bands without anylabels remain unassigned. is the only band of B symmetry with a sizeabletransition moment that does not exhibit a sharpQ-branch. Tentative assignments are proposed inTable 1, but their accuracy should be taken withcaution as the actual bands centers could lie in a10 to 20 cm − window from the proposed assign-ments. Concerning ν (ring symmetric breathing)and ν (C −−− C − C −−− N stretch)—predicted at 1012and 2265 cm − , respectively—two bands lie system-atically close to the expected energy, with reason-able band profiles, thus in each case both assign-ments are reported in Table 1. In the case of ν ,the lowest frequency assignment, i.e. , 2154 cm − appears most likely as the relatively strong intensityof the observed band would indicate a fundamen-tal that could otherwise not be predicted. How-ever the shape of the band lying at 2198 cm − iscloser to that expected for an a -type band, assum-ing that our aforementioned criterion remains validat these frequencies (i.e. hot bands and combina-tion bands could significantly affect the simplisticfundamentals-only picture).The most difficult analysis lies in the ∼ − cm − region, where C − H stretching motionsare observed. From previous studies on similarmolecules like phenylacetylene [14] and naphthalene[16], it is well-known that this frequency range isplagued by anharmonic resonances which can signif-icantly complicate assignment. For PhC N, threemodes of A ( ν , ν , ν ) and two of B ( ν , ν )symmetries (Table 1) very close in energy (of or-der of ∼
10 cm − for the harmonic frequencies) andcan mix strongly, thus qualitatively shifting the fun-damental energies and altering band intensities. Insuch cases, an approximate deperturbation analysiswas performed using the generalized VPT2 calcula-tions to identify and treat anharmonic resonances,e.g. those arising from Fermi (so-called “1–2” res-onances) and Darling-Dennison (“1–1” interactionsin the present case). These resonances are identifiedbased on small differences in the energies of statesand a model variational calculation [30]; the formeris a zeroth order estimate for resonances, while thelatter tests for the magnitude of the coupling [31].The VPT2 routines in Gaussian treat the prob-lem of state-to-state coupling as effective × ν : ν is “dark” in the harmonic approximation,and it gains intensity primarily through borrowingintensity from the stronger ν and ν bands. Ina general resonance picture, this interaction sharesintensity, and to render a mode completely inac-tive is extremely rare if not unheard of. Given theVPT2 treatment here is only approximate, our ef-fective deperturbation analysis is likely inadequateto properly treat these bands, and a fully coupledmodel involving fundamentals, combination bands,and overtones, is required instead.In light of this preliminary analysis, however, aswell as the fact that the ∼ cm − region is heav-ily congested, we assume all three A modes are IRactive. Measurement under cold conditions—eitherin a supersonic jet or buffer gas cell—will help fu-ture analysis of this molecule by eliminating thepossibility of combination bands and overtones, inaddition to minimizing lineshape blending from ro-tational contours. Similarly, selective deuterationmight clarify the assignment of some features [14]. Table 2: Strong anharmonic resonances and their corre-sponding off-diagonal matrix elements for bands in the ∼ − cm − region. Darling-Dennison (DDR) andFermi (FR) resonances are indicated; in this table, the for-mer corresponds to − type DDR, referring to the numberof quanta for states involved. State 1 State 2 Type Coupling v = 1 v = 1 DD -6.6 v = 1 v = 1 DD 14.5 v = 1 v = 1 + v = 1 FR -23.2 v = 1 v = 1 + v = 1 FR -29.3 v = 1 v = 1 DD 9.4
Using the ground state rotational and quarticcentrifugal distortion constants from the anhar-monic calculation, 65 strong lines of the CP-FTMWspectrum were assigned in a straightforward fashionusing the PGOPHER software [27] (Fig. 2). The de-rived values are extremely close to those predictedby the calculation (the weighted frequency differ-ence δ is less than 1 % for the rotational constants, see Table 3). This initial set of constants was thenused to assign the millimeter-wave data. Loomis-Wood diagrams were produced by means of theLWWa software from Lodyga et al. [32] to aid inthe assignment of the high- J transitions. In total,6151 a -type transitions (3780 different frequenciesas a result of unresolved asymmetric splitting) ofPhC N in its ground vibrational state were assignedin the millimeter-wave spectrum, with values of J (cid:48)(cid:48) up to 199 and K (cid:48)(cid:48) a up to 42.The SPFIT/SPCAT suite of programs [33] us-ing the Watson S -reduced Hamiltonian in the I r representation was used to determine best-fit spec-troscopic constants. All transitions were weightedaccorded to their expected experimental accuracy, i.e. , 2 kHz and 25 kHz for the CE-FTMW and CP-FTMW transitions, respectively, and 50 kHz for themillimeter-wave transitions. To reproduce the datato their experimental accuracy, inclusion of severalsextic and octic centrifugal distortion constants wasrequired. Finally, the CE-FTMW transitions—theonly ones for which the nuclear quadrupole split-ting was resolved—were added to the fit to deter-mine the χ (N) terms. All 57 hyperfine componentswere reproduced to the measurement uncertaintyby adjusting only χ aa (N) and χ bb (N). Both param-eters are close to those expected from calculation(to within about 15 %, Table 3). When a transitionwas observed by both CP-FTMW and CE-FTMWspectroscopy, only the latter was retained in the fit,owing to the higher frequency accuracy of the cavityinstrument. The rotational constants derived froma fit to all of the assigned rotational transitions arereported in Table 3. The final 49 kHz RMS valueof the fit, corresponding to a reduced standard de-viation σ = 1 . , indicates that our present modeladequately reproduces the ground state rotationalspectrum of PhC N.Figure 4 shows a portion of the millimeter-wave spectrum in comparison with a simulation ofPhC N in its vibrational ground state using the ex-perimentally determined best-fit parameters fromTable 3. As illustrated in this figure, many linesremain unassigned, but most of these likely arisefrom vibrational satellites, for which no attemptat assigning was made in the present study. Al-though longer integration times would allow a morein-depth analysis of these satellites, the spectrumis already fairly dense, implying that we may beclose to the confusion limit. Indeed, the 2 µbar pres-sure used in this study, although quite low, was ac-tually a compromise between reasonable signal-to-7 T = 300 K Figure 4: Portions of the millimeter-wave spectrum of PhC N in comparison with a simulation of the pure rotational transitionsin the ground vibrational state using the best-fit set of spectroscopic constants (Table 3). The simulation has been performedusing the PGOPHER software and the resulting trace was then post-processed with a second derivative to allow a morestraightforward comparison with the experimental spectrum. The line density in the experimental trace is far greater than oursimulation, very likely because of lines from vibrational satellites. noise ratio and pressure broadening. Even at thispressure many lines are broader then expected fromthe effects of pressure broadening alone, and conse-quently may in fact be a spectral superposition ofseveral transitions.Regarding hyperfine splitting, its magnituderapidly collapses with increasing J , as expectedfrom this type of interaction. No splitting isobserved in the CP-FTMW nor in millimeter-wave measurements, and in the cavity experi-ments it is marginally resolved above J (cid:48)(cid:48) ≈ ( ∼ GHz). Even at the lowest- J transitions mea-sured here ( J (cid:48)(cid:48) = 5 , ∼ GHz), the splitting dueto hyperfine and Doppler effects are comparable(Fig. 5). The experimentally derived value of χ aa (N) [ − . (77) MHz; Table 3] is very similarto that reported for PhCN [-4.23738(36) MHz [9]],and the relative magnitudes are in agreement withthe theoretical values of χ aa (N) calculated at the ω B97X-D/cc-pVQZ level of theory for PhC N (-4.959 MHz) and PhCN (-4.962 MHz). The smallchanges in χ (N) are an indication that the localelectronic structure of the nitrogen atom is rel-atively insensitive to the distance from the aro-matic ring. Equivalently, this finding implies that electron delocalization through ring conjugationis very poorly coupled to the chain regardless oflength. Similar behavior is seen for cyanopolyynechains, where the value of eQq equivalent to χ aa (N)(around − . MHz) is relatively invariant with thelength of the chain as well [34].
Detection of molecules in space by radio astron-omy is heavily dependent on the magnitude of theirpermanent dipole moment. In comparison withPhCN, our theoretical predictions suggest PhC Nis substantially more polar (5.9 D vs. of 4.5 D,where the former value has statistical uncertaintyof ± . D based on our prior benchmarking at the ω B97X-D/cc-pVQZ level of theory [35]). Currently,PhCN is theorized to form in cold, dark clouds viaa barrierless reaction between C H and CN rad-ical [4, 36]. We thus speculate that PhC N couldbe formed via C insertion to PhCN or throughan analogous C −−− N addition reaction between CNradical and phenylacetylene (PhC H), or C N andC H . If the latter mode is operative, then theabundance ratio of PhCN/PhC N will be depen-dent on CN/C N, assuming similar reaction cross-8 able 3: Spectroscopic constants (rotational, centrifugal dis-torsion, and nuclear quadrupole constants) of PhC N in itsvibrational ground state (in MHz) and relevant fit parame-ters. Numbers in parenthesis are σ uncertainties expressedin the unit of the last digit. Parameters in brackets werekept fixed to the calculated values. Constant Calc. a Exp. δ b A . .
722 (15) B . . C . . D J × . . D JK × .
78 0 . D K × .
42 0 .
400 (68) -4.8 d × − . − . d × − . − . H J × − .
957 (11) H JK × . H KJ × − . h × .
575 (15) h × . L JJK × − .
295 (20) L JK × . L KKJ × − .
512 (36) χ aa − . − .
219 (77) -15 χ bb .
39 2 .
114 (68) -12 χ cc .
57 2 . c -17 N d J (cid:48)(cid:48) max , K (cid:48)(cid:48) a max σ e a ω B97XD/cc-pVQZ level of theory, Bayesian corrected for A , B , and C , and anharmonic values for the centrifugal distortionconstants, and equilibrium values for the hyperfine constants b δ = ( B exp . − B calc . ) /B calc . × (in %) c Derived value d Total number of lines in the fit / Number of different frequen-cies / Number of lines with resolved nuclear quadrupole struc-ture e Reduced standard deviation, unitless sections. The CN + PhC H process has been stud-ied in crossed-beam experiments by Bennett et al.[11], where the authors identify PhC N as a po-tential reaction product, albeit not definitively sodue to the lack of isomer specificity and at collisionenergies well in excess of interstellar cloud condi-tions ( ∼ kJ/mol). This finding suggests other iso-mers, namely ortho, meta, para -CEB, might plausi-bly be formed from this reaction. While we have notattempted to experimentally characterized thesespecies, estimates of their spectroscopic parametersare provided here. By scaling the experimental pa-rameters for PhC N (Table 3) to correct for vibra- I n t e n s i t y / a r b . u . J K a K c J K a K c = 6 Figure 5: Example of a transition showing resolved hyperfinestructure on the CE-FTMW spectrum, and comparison with10 K simulations using the final set of spectroscopic parame-ters, with and without taking into account the Doppler split-ting (simulations performed using the PGOPHER software,assuming a Lorentzian lineshape). As before, the simulationwith Doppler splitting is inverted relative to the observedspectrum solely for comparison purposes. tional and electronic effects (Table S6 in the Sup-porting Information), in conjunction to Bayesianuncertainties obtained from benchmarking [35], re-liable constraints of these constants should aid fu-ture experimental searches for these species.In terms of astronomical detection, although thehyperfine structure does not take a large part thepresent rotational analysis, this splitting is partiallyresolved with cavity measurements up to 16 GHz.In cold, dark clouds such as TMC-1 where thelinewidths are comparable to those measured withour CE-FTMW spectrometer, it is thus necessaryto consider hyperfine splitting, as recent work onPhCN [4] demonstrates. This is particularly truefor a relatively heavy molecule like PhC N whosestrongest lines should lie at centimeter-wavelengthsat low temperatures (Fig. 6): at 10 K, the strongestfeatures correspond to J (cid:48)(cid:48) = 21 around 23 GHz,while at 6 K —a typical temperature for moleculesin TMC-1— the strongest features are near 15 GHz( J (cid:48)(cid:48) = 14 ). Thus, the X / K u (8–12/12–18 GHz)bands appear to be the most promising to de-tect PhC N in TMC–1. In sources with some-what warmer temperatures, the intensity of indi-vidual transitions is significantly decreased due tothe larger partition function, and the peak inten-sity, although relatively flat, falls in the W (75–110GHz) and N (100–200 GHz) bands.Detection of multiple isomers is also a powerfultool for constraining physical and chemical condi-tions in astrophysical environments. To assist in9
50 100 150 200 250Frequency / GHz 10 K100 K * 5300 K * 10
Figure 6: Calculated rotational spectrum of PhC N at 10 K(purple), 100 K (blue), and 300 K (orange). For the pur-poses of display, these relative intensities for the 100 K plothave increased by a factor of 5, while for the 300 K plot, theincrease is 10. this process, we have calculated the relative en-ergy of the aforementioned ortho, meta, and para -CEB isomers, along with the isocyanide isomer,PhCCNC. As shown in Fig. 7, apart from the iso-cyanide isomer, placement of the acetylenic unit ondifferent parts of the ring produces isomers withcomparable stability to PhC N, i.e., they are effec-tively degenerate at the level of uncertainty affordedby G3//B3LYP ( ± kJ/mol). As such, a determi-nation of their relative abundances would providea sensitive test of thermodynamics vs. kinetics inmolecule formation. To aid further laboratory andhopefully astronomical efforts, Table S6, in additionto providing estimates of rotational constants, alsoreports calculated dipole moments at the ω B97X-D/cc-pVQZ level of theory using two methods ofempirical scaling to correct for vibrational effectsand deficiencies in the electronic structure method.We note that the Bayesian scaling factors obtainedin Ref. [35] applied to PhC N —where we nowhave accurately determined parameters— exceedthe performance of the purely theoretical VPT2corrections, and bring the theoretical predictionswithin a few MHz of the experimentally measuredones (see also Fig. S1). We expect a similar degreeof precision and uncertainty for other isomers.
4. Conclusion
Using a combination of gas-phase measurementsand quantum chemical calculations, the fundamen-tal vibrational frequencies of most of the strongestIR-active fundamentals of PhC N and over 6000pure rotational transitions in its ground state be-
Figure 7: 0 K energetics of isomers of interest calculated atthe G3//B3LYP level of theory, relative to PhC N. Energiesto the nearest kJ/mol are annotated above each bar. tween 6.5 and 220 GHz have been measured. Theassignment of spectra in different regions wasguided by theoretical predictions: in the infrared,anharmonic calculations helped to assign mostbands, with the exception of the most congestedand perturbed region around cm − ; in the ra-dio domain, estimates of centrifugal distortion andhyperfine terms proved useful. Comparisons withPhCN suggest that the local electronic structureof the terminal nitrogen —as probed through itsquadrupole moment χ aa (N)— is relatively unaf-fected by chain lengthening, indicating a similarelectric field gradient for nitrogen in both PhCNand PhC N. In addition, we provide accurateand reliable predictions of the thermochemistryand spectroscopic parameters for several isomers ofPhC N, which should prove useful in guiding fu-ture laboratory experiments. These isomers includethe isocyanide isomer of PhC N and the three cya-noethynylbenzenes alluded to in previous work [11];their discovery along with PhC N in the ISM wouldprovide sensitive measurements of local chemicaland physical interstellar environments.With highly precise measurements of the restfrequencies and corresponding spectroscopic con-stants, a search for PhC N can now be undertakenwith considerable confidence in the ISM. The largepermanent dipole moment (predicted to be 5.9 D),in addition to its chemical and structural similarityto astronomical PhCN makes PhC N an excellentcandidate for detection towards cold, dark molecu-10ar clouds such as TMC-1. Hyperfine-resolved mea-surements are expected to be highly relevant in apotential discovery, given that at low temperaturesthe strongest transitions lie in the X / K u bandswhere the splitting is comparable to the sourcelinewidth. Acknowledgements
O.C., O.P., and M.-A.M.-D. acknowledge fund-ing support from the Région Ile-de-France throughDIM-ACAV + , from the Agence Nationale de laRecherche (ANR-19-CE30-0017-01), from the “
In-vestissements d’Avenir ” LabEx PALM (ANR-10-LABX-0039-PALM), and from the Programme Na-tional “Physique et Chimie du Milieu Interstellaire”(PCMI) of CNRS/INSU with INC/INP co-fundedby CEA and CNES. K.L.K.L. and M.C.M. acknowl-edge funding support from NSF grant AST-1908576and NASA grant 80NSSC18K0396. Z.S.B. acknowl-edges support from the Chateaubriand Fellowshipof the Office for Science & Technology of the Em-bassy of France in the United States. The authorsare thankful to the AILES beamline staff for pro-viding access to the FTIR interferometer.
References [1] M. Frenklach, E. D. Feigelson, Formation of polycyclicaromatic hydrocarbons in circumstellar envelopes, TheAstrophysical Journal 341 (1989) 372.[2] T. P. Snow, A. N. Witt, The Interstellar Carbon Budgetand the Role of Carbon in Dust and Large Molecules,Science 270 (1995) 1455–1460.[3] A. Tielens, Interstellar Polycyclic Aromatic Hydrocar-bon Molecules, Annual Review of Astronomy and As-trophysics 46 (2008) 289–337.[4] B. A. McGuire, A. M. Burkhardt, S. Kalenskii, C. N.Shingledecker, A. J. Remijan, E. Herbst, M. C. Mc-Carthy, Detection of the aromatic molecule benzoni-trile (c-C H CN) in the interstellar medium, Science359 (2018) 202–205.[5] D. McNaughton, M. K. Jahn, M. J. Travers,D. Wachsmuth, P. D. Godfrey, J.-U. Grabow, Labora-tory rotational spectroscopy of cyano substituted poly-cyclic aromatic hydrocarbons, Monthly Notices of theRoyal Astronomical Society 476 (2018) 5268–5273.[6] M. C. McCarthy, K. L. K. Lee, R. A. Loomis, A. M.Burkhardt, C. N. Shingledecker, S. B. Charnley, M. A.Cordiner, E. Herbst, S. Kalenskii, E. R. Willis, C. Xue,A. J. Remijan, B. A. McGuire, Interstellar detection ofthe highly polar five-membered ring cyanocyclopenta-diene, Nature Astronomy (2020). In press.[7] T. Gautier, N. Carrasco, A. Buch, C. Szopa,E. Sciamma-O’Brien, G. Cernogora, Nitrile gas chem-istry in Titan atmosphere, Icarus 213 (2011) 625. [8] R. Loomis, et al., The third time’s a charm? a rig-orous investigation of spectral line stacking techniquesand application to the detection of HC N, submitted(2020).[9] K. Wohlfart, M. Schnell, J.-U. Grabow, J. Küpper, Pre-cise dipole moment and quadrupole coupling constantsof benzonitrile, Journal of Molecular Spectroscopy 247(2008) 119–121.[10] B. Bak, D. Christensen, W. B. Dixon,L. Hansen–Nygaard, J. Rastrup–Andersen, Ben-zene Ring Distortion by One Substituent. MicrowaveDetermination of the Complete Structure of Benzoni-trile, The Journal of Chemical Physics 37 (1962)2027–2031. Publisher: American Institute of Physics.[11] C. J. Bennett, S. B. Morales, S. D. L. Picard, A. Canosa,I. R. Sims, Y. H. Shih, A. H. H. Chang, X. Gu,F. Zhang, R. I. Kaiser, A chemical dynamics, kinet-ics, and theoretical study on the reaction of the cyanoradical (CN; X Σ + ) with phenylacetylene (C H CCH;X A ), Phys. Chem. Chem. Phys. 12 (2010) 8737–8749.[12] K. M. Jawad, T. S. Zwier, L. V. Slipchenko,C. I. Viquez Rojas, The exotic excitedstate behavior of 3-phenyl-2-propynenitrile,http://hdl.handle.net/2142/97063, 2017. Interna-tional Symposium on Molecular Spectroscopy.[13] C. I. Viquez Rojas, L. V. Slipchenko, T. S.Zwier, K. M. Jawad, Computational modeling ofelectronic spectroscopy of 3-phenyl-2-propynenitrile,http://hdl.handle.net/2142/97102, 2017. InternationalSymposium on Molecular Spectroscopy.[14] G. King, S. So, Ethynylbenzene; The vibrational spec-tra of some deuterated isomers, Journal of MolecularSpectroscopy 36 (1970) 468–487. ZSCC: 0000063.[15] C.-H. Chang, G. Lopez, T. J. Sears, P. M. Johnson,Vibronic Analysis of the S —S Transition of Pheny-lacetylene Using Photoelectron Imaging and SpectralIntensities Derived from Electronic Structure Calcula-tions, The Journal of Physical Chemistry A 114 (2010)8262–8270. ZSCC: NoCitationData[s0].[16] O. Pirali, M. Vervloet, G. Mulas, G. Malloci, C. Joblin,High-resolution infrared absorption spectroscopy ofthermally excited naphthalene. Measurements and cal-culations of anharmonic parameters and vibrational in-teractions, Physical Chemistry Chemical Physics 11(2009) 3443. ZSCC: 0000046.[17] O. Pirali, Z. Kisiel, M. Goubet, S. Gruet, M. A. Martin-Drumel, A. Cuisset, F. Hindle, G. Mouret, Rotation-vibration interactions in the spectra of polycyclic aro-matic hydrocarbons: Quinoline as a test-case species,The Journal of Chemical Physics 142 (2015) 104310.Publisher: American Institute of Physics.[18] Z. Kisiel, A. Kraśnicki, The millimetre-wave rotationalspectrum of phenylacetylene, Journal of MolecularSpectroscopy 262 (2010) 82–88.[19] M. A. Zdanovskaia, B. J. Esselman, H. S. Lau, D. M.Bates, R. C. Woods, R. J. McMahon, Z. Kisiel, The103–360 GHz rotational spectrum of benzonitrile, thefirst interstellar benzene derivative detected by radioas-tronomy, Journal of Molecular Spectroscopy 351 (2018)39–48.[20] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.Scuseria, M. A. Robb, J. R. Cheeseman, G. Scal-mani, V. Barone, G. A. Petersson, H. Nakatsuji, X. Li,M. Caricato, A. V. Marenich, J. Bloino, B. G. Janesko,R. Gomperts, B. Mennucci, H. P. Hratchian, J. V. Or- iz, A. F. Izmaylov, J. L. Sonnenberg, D. Williams-Young, F. Ding, F. Lipparini, F. Egidi, J. Go-ings, B. Peng, A. Petrone, T. Henderson, D. Ranas-inghe, V. G. Zakrzewski, J. Gao, N. Rega, G. Zheng,W. Liang, M. Hada, M. Ehara, K. Toyota, R. Fukuda,J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Ki-tao, H. Nakai, T. Vreven, K. Throssell, J. A. Mont-gomery, Jr., J. E. Peralta, F. Ogliaro, M. J. Bearpark,J. J. Heyd, E. N. Brothers, K. N. Kudin, V. N.Staroverov, T. A. Keith, R. Kobayashi, J. Normand,K. Raghavachari, A. P. Rendell, J. C. Burant, S. S.Iyengar, J. Tomasi, M. Cossi, J. M. Millam, M. Klene,C. Adamo, R. Cammi, J. W. Ochterski, R. L. Martin,K. Morokuma, O. Farkas, J. B. Foresman, D. J. Fox,Gaussian 16 Revision A.01, 2016.[21] A. G. Baboul, L. A. Curtiss, P. C. Redfern,K. Raghavachari, Gaussian-3 theory using density func-tional geometries and zero-point energies, The Journalof Chemical Physics 110 (1999) 7650–7657.[22] J. M. Simmie, K. P. Somers, Benchmarking CompoundMethods (CBS-QB3, CBS-APNO, G3, G4, W1BD)against the Active Thermochemical Tables: A LitmusTest for Cost-Effective Molecular Formation Enthalpies,The Journal of Physical Chemistry A 119 (2015) 7235–7246.[23] J.-B. Brubach, L. Manceron, M. Rouzières, O. Pirali,D. Balcon, F. Kwabia-Tchana, V. Boudon, M. Tudorie,T. Huet, A. Cuisset, P. Roy, in: WIRMS 2009, volume1214 of AIP Conference Proceedings , pp. 81–84.[24] M. Faye, M. Bordessoule, B. Kanouté, J.-B. Brubach,P. Roy, L. Manceron, Improved mid infrared detec-tor for high spectral or spatial resolution and syn-chrotron radiation use, Review of Scientific Instruments87 (2016) 063119.[25] O. Pirali, M. Goubet, T. R. Huet, R. Georges,P. Soulard, P. Asselin, J. Courbe, P. Roy, M. Vervloet,The far infrared spectrum of naphthalene character-ized by high resolution synchrotron FTIR spectroscopyand anharmonic DFT calculations, Physical ChemistryChemical Physics 15 (2013) 10141–10150.[26] K. N. Crabtree, M.-A. Martin-Drumel, G. G. Brown,S. A. Gaster, T. M. Hall, M. C. McCarthy, Microwavespectral taxonomy: A semi-automated combination ofchirped-pulse and cavity fourier-transform microwavespectroscopy, The Journal of Chemical Physics 144(2016) 124201.[27] C. M. Western, PGOPHER: A Program for Simulat-ing Rotational, Vibrational and Electronic Spectra, J.Quant. Spectrosc. Radiat. Transfer 186 (2017) 221–242.[28] M. C. McCarthy, K. L. K. Lee, P. B. Carroll, J. P.Porterfield, B. Changala, J. H. Thorpe, J. F. Stanton,Exhaustive product analysis of three benzene dischargesby microwave spectroscopy, J. Phys. Chem. A (2020).[29] O. Pirali, M. Goubet, V. Boudon, L. D’Accolti,C. Fusco, C. Annese, Characterization of isolated1-aza-adamantan-4-one (C H NO) from microwave,millimeter-wave and infrared spectroscopy supportedby electronic structure calculations, J. Mol. Spectrosc.338 (2017) 6 – 14.[30] J. M. L. Martin, T. J. Lee, P. R. Taylor, J. François,The anharmonic force field of ethylene, C H , bymeans of accurate ab initio calculations, The Jour-nal of Chemical Physics 103 (1995) 2589–2602. ZSCC:NoCitationData[s0].[31] J. Bloino, M. Biczysko, V. Barone, Anharmonic Ef- fects on Vibrational Spectra Intensities: Infrared, Ra-man, Vibrational Circular Dichroism, and Raman Op-tical Activity, The Journal of Physical Chemistry A 119(2015) 11862–11874. ZSCC: 0000070.[32] W. Lodyga, M. Kręglewski, P. Pracna, ˇo. Urban, Ad-vanced graphical software for assignments of transitionsin rovibrational spectra, Journal of Molecular Spec-troscopy 243 (2007) 182 – 188. PRAHA2006, The 19thInternational Conference on High Resolution MolecularSpectroscopy.[33] H. M. Pickett, The Fitting and Prediction of Vibration-rotation Spectra with Spin Interactions, J. Mol. Spec-trosc. 148 (1991) 371–377.[34] M. C. McCarthy, W. Chen, M. J. Travers, P. Thaddeus,Microwave Spectra of 11 Polyyne Carbon Chains, TheAstrophysical Journal Supplement Series 129 (2000)611.[35] K. L. K. Lee, M. McCarthy, Bayesian Analysis of Theo-retical Rotational Constants from Low-Cost ElectronicStructure Methods, The Journal of Physical ChemistryA 5 (2020) 898–910.[36] K. L. K. Lee, B. A. McGuire, M. C. McCarthy, Gas-phase synthetic pathways to benzene and benzonitrile:a combined microwave and thermochemical investiga-tion, Physical Chemistry Chemical Physics 21 (2019)2946–2956.calculations, The Jour-nal of Chemical Physics 103 (1995) 2589–2602. ZSCC:NoCitationData[s0].[31] J. Bloino, M. Biczysko, V. Barone, Anharmonic Ef- fects on Vibrational Spectra Intensities: Infrared, Ra-man, Vibrational Circular Dichroism, and Raman Op-tical Activity, The Journal of Physical Chemistry A 119(2015) 11862–11874. ZSCC: 0000070.[32] W. Lodyga, M. Kręglewski, P. Pracna, ˇo. Urban, Ad-vanced graphical software for assignments of transitionsin rovibrational spectra, Journal of Molecular Spec-troscopy 243 (2007) 182 – 188. PRAHA2006, The 19thInternational Conference on High Resolution MolecularSpectroscopy.[33] H. M. Pickett, The Fitting and Prediction of Vibration-rotation Spectra with Spin Interactions, J. Mol. Spec-trosc. 148 (1991) 371–377.[34] M. C. McCarthy, W. Chen, M. J. Travers, P. Thaddeus,Microwave Spectra of 11 Polyyne Carbon Chains, TheAstrophysical Journal Supplement Series 129 (2000)611.[35] K. L. K. Lee, M. McCarthy, Bayesian Analysis of Theo-retical Rotational Constants from Low-Cost ElectronicStructure Methods, The Journal of Physical ChemistryA 5 (2020) 898–910.[36] K. L. K. Lee, B. A. McGuire, M. C. McCarthy, Gas-phase synthetic pathways to benzene and benzonitrile:a combined microwave and thermochemical investiga-tion, Physical Chemistry Chemical Physics 21 (2019)2946–2956.