Infrared action spectroscopy of doubly charged PAHs and their contribution to the aromatic infrared bands
Shreyak Banhatti, Julianna Palotás, Pavol Jusko, Britta Redlich, Jos Oomens, Stephan Schlemmer, Sandra Brünken
AAstronomy & Astrophysics manuscript no. Banhatti_Arxiv © ESO 2021February 22, 2021
Infrared action spectroscopy of doubly charged PAHs and theircontribution to the aromatic infrared bands
S. Banhatti , J. Palotás , P. Jusko , B. Redlich , J. Oomens , , S. Schlemmer , and S. Brünken I. Physikalisches Institut, Universität zu Köln, Zülpicher Str. 77, 50937 Köln, Germanye-mail: [email protected] Radboud University, Institute for Molecules and Materials, FELIX Laboratory, Toernooiveld 7, 6525ED Nijmegen, the Netherlandse-mail: [email protected] Max Planck Institute for Extraterrestrial Physics, Gießenbachstraße 1, 85748 Garching, Germany van ’t Ho ff Institute for Molecular Sciences, University of Amsterdam, Science Park 908, 1098XH Amsterdam, the Netherlands15 February 2020 ; accepted manuscript
ABSTRACT
The so-called aromatic infrared bands (AIBs) are attributed to emission of polycyclic aromatic hydrocarbons (PAHs). The observedvariations toward di ff erent regions in space are believed to be caused by contributions of di ff erent classes of PAH molecules, that is tosay with respect to their size, structure, and charge state. Laboratory spectra of members of these classes are needed to compare themto observations and to benchmark quantum-chemically computed spectra of these species. In this paper we present the experimentalinfrared (IR) spectra of three di ff erent PAH dications, naphthalene + , anthracene + , and phenanthrene + , in the vibrational fingerprintregion 500 - 1700 cm − . The dications were produced by electron impact ionization (EI) of the vapors with 70 eV electrons, andthey remained stable against dissociation and Coulomb explosion. The vibrational spectra were obtained by IR predissociation ofthe PAH + complexed with neon in a 22-pole cryogenic ion trap setup coupled to a free-electron infrared laser at the Free-ElectronLasers for Infrared eXperiments (FELIX) Laboratory. We performed anharmonic density-functional theory (DFT) calculations forboth singly and doubly charged states of the three molecules. The experimental band positions showed excellent agreement with thecalculated band positions of the singlet electronic ground state for all three doubly charged species, indicating its higher stability overthe triplet state. The presence of several strong combination bands and additional weaker features in the recorded spectra, especiallyin the 10-15 µ m region of the mid-IR spectrum, required anharmonic calculations to understand their e ff ects on the total integratedintensity for the di ff erent charge states. These measurements, in tandem with theoretical calculations, will help in the identification ofthis specific class of doubly-charged PAHs as carriers of AIBs. Key words.
ISM: lines and bands - ISM: molecules - Techniques: spectroscopic - Methods: laboratory: molecular - Molecular data- Line: identification
1. Introduction
The widespread mid-infrared (mid-IR) emission features ob-served in many astrophysical objects, such as HII regions, plan-etary nebulae (PNe), reflection nebulae, and young stellar ob-jects, have been a topic of great interest since their detectionin the 1970s and 1980s. Strong bands are observed at 3.3, 6.2,7.7, 11.3, and 12.7 µ m along with several weak features, whichtogether make up the unidentified infrared (UIR) or aromatic in-frared bands (AIBs) (Sellgren 1984; Sellgren et al. 1985; Jour-dain de Muizon et al. 1990; Cohen et al. 1986). These bandsare hypothesized to arise from a family of polycyclic aromatichydrocarbons (PAHs) excited by absorption of ultraviolet (UV)radiation from nearby stars and their subsequent emission in themid-IR region (Leger & Puget 1984; Allamandola et al. 1989,1985; Hudgins et al. 1997)Depending on the conditions in the interstellar medium(ISM), di ff erent families of PAH molecules are proposed to ex-ist, such as neutral and ionic variants of di ff erent size, shape, andhydrogenation states, all of which have an e ff ect on the relativeband intensities and band positions in the observed UIR bands(Hony et al. 2001; Dartois & D’Hendecourt 1997). For example,the relative intensity variation between the 6.2, 7.7, 8.6, and 11.2 µ m bands has been attributed to the degree of ionization of PAHs (Galliano et al. 2008). Also, the 18.9 µ m band has been identifiedas a signature of multiply charged PAHs (Tielens 2008).Identifying di ff erent classes of PAHs as carriers of UIRbands requires laboratory data of their gas-phase IR spectra. Sev-eral schemes have been used over the past decades to recordthe IR spectra of PAHs. In particular for the study of chargedPAHs, the development of intense and widely tunable free-electron lasers in facilities such as the Free-Electron Lasers forInfrared eXperiments (FELIX) Laboratory (Oepts et al. 1995) and CLIO has made it possible to implement sensitive ac-tion spectroscopic schemes such as infrared multiple photondissociation (IRMPD). Gas-phase IR spectra of several small-to medium-sized cationic species have been recorded with theIRMPD scheme (e.g., naphthalene, phenanthrene, anthracene,coronene, and protonated naphthalene), pioneered by Oomensand coworkers (Oomens et al. 2000, 2001; Bakker et al. 2011;Lorenz et al. 2007). More recent works on the IR spectroscopyof gas phase PAH ions include larger species such as di-indenoperylene (C H + ), dicoronylene (C H + ), hexa-peri-hexabenzocoronene (C H + ) (Zhen et al. 2017, 2018), and ru-bicene (C H + ) (Bouwman et al. 2020) and PAH anions of http://old.clio.lcp.u-psud.fr/clio_eng/clio_eng.htm Article number, page 1 of 13 a r X i v : . [ a s t r o - ph . GA ] F e b & A proofs: manuscript no. Banhatti_Arxiv naphthyl, anthracenyl, and pyrenyl (Gao et al. 2014). Their spec-tra have been shown to contain vibrational modes which couldaccount for some of the features observed in the UIR bands.Pentagon-containing PAHs, for example, were shown to possessdistinct vibrational modes at around 1100 cm − ( ∼ . µ m), andthe ratio of the 9 . / . µ m band strength was used to estimatetheir relative abundance in several astronomical sources (Bouw-man et al. 2020). Experimental results also verified that proto-nated nitrogen-containing species (H + PANHs) might indeed beresponsible for the 6.2 µ m UIR emission (Alvaro Galué et al.2010).Another powerful action spectroscopic technique is infraredpredissociation (IRPD), which requires cold ion conditions totag the ions with a weakly bound rare-gas atom. Earlier IRPDstudies applied to PAH ions used a molecular beam cooled by su-personic expansion, and they provided gas phase spectra of coldprotonated (Ricks et al. 2009) and cationic (Piest et al. 1999)naphthalene and phenanthrene (Piest et al. 2001) via dissocia-tion of their weakly bound complexes with Ar. With the adventof cryogenic ion trapping techniques, IRPD spectra can now berecorded at temperatures as low as 4 K, allowing for the useof weaker bound rare-gas atoms such as He or Ne as taggingagents, introducing smaller shifts of vibrational bands in the ex-perimental spectra (Asmis et al. 2002; Jašík et al. 2014; Güntheret al. 2017; Gerlich et al. 2018; Jusko et al. 2019). The advantageof IRPD compared to multiphoton IRMPD spectroscopy is thatdissociation happens after absorption of a single photon, withapproximately constant e ffi ciency, resulting in experimental vi-brational spectra more closely resembling the linear absorptionspectrum of the ions, both in intensity and band positions. In thepast, we have successfully used this technique in our cryogenicion trap instrument coupled to the widely tunable free electronlasers at the FELIX Laboratory to record narrow-linewidth spec-tra of PAH cations and related species (Jusko et al. 2018a,b; Pan-chagnula et al. 2020), and here we apply it for the spectroscopiccharacterization of PAH dications.The presence of PAH dications and their formation in theISM has been discussed previously by Leach (1986). Theoreti-cal investigations on their stability and vibrational spectra havebeen carried out by Malloci et al. (2007a), Bakes et al. (2001a,b),and Bauschlicher & Langho ff (1997). However, most of the lab-oratory work done so far in regards to the vibrational spectraof PAHs has been focused on their neutrals and monocations.There is no spectroscopic laboratory data on doubly chargedPAHs in the IR regime except for an IRMPD study of hexa-peri-hexabenzocoronene, C H + (Zhen et al. 2017). In this workwe present the IR predissociation spectra of the following threedoubly charged PAHs using Ne-tagging in a cryogenic ion trap:naphthalene + (naph + , C H + ), anthracene + (anth + ), and itsisomer phenanthrene + (phen + , both C H + ).
2. Methods
The vibrational spectra of PAH dications were recorded us-ing the FELion cryogenic 22-pole ion trap setup coupled tothe free-electron laser FEL-2 at the FELIX Laboratory (Oeptset al. 1995). The FELion experimental setup has already beendescribed in detail in Jusko et al. (2019). To provide a briefoverview, the respective vapors of PAHs are ionized by electronimpact ionization (EI), which results in the formation of severalcharged fragments including the doubly charged naphthalenecation (C H + , mass-to-charge ratio of m / z =
64, naph + in the following) and the doubly charged phenanthrene and anthracenecations (C H + , m / z =
89, phen + , and anth + , resp.). In our ex-periments, we used electron energies of 50-70 eV, which is wellabove the appearance energy of naph + at 21.52 eV, and phen + and anth + around 20 eV (Holm et al. 2011; Malloci et al. 2007b;van der Burgt et al. 2018, 2019). The dication yield was foundto be the highest around 50 eV and remained constant at higherenergies, as can be seen in Fig. 1, showing the anth + (and anth + )yield as a function of electron energy. Fig. 2 shows a typical massspectrum upon EI of phenanthrene vapor.While naph + was produced e ffi ciently with an RF storagesource (Gerlich 1992), we could not produce high yields ofphen + or anth + ions in this way. This is probably due to thechemical quenching of initially produced dications by chargetransfer and reactions with the neutral precursor and fragmentsin the high pressure (10 − mbar) storage source over the typi-cal storage time of seconds. Hence, a non-storage EI source wasused for these dications. The ions produced in the source weremass selected to within m / z = ± H + or C H + (see fig A.1). The mass selected ions werethen complexed in situ with neon (Ne) in a 22-pole cryogenicion trap. Either pure Ne gas (in the case of anth + and phen + )or a 3:1 He:Ne gas mixture (in the case of naph + ) was pulsedinto the ion trap for 100-150 ms at the beginning of the storageperiod, with the trap held at a temperature of 15 K or 6.3 K, re-spectively. A second quadrupole mass filter with a mass resolu-tion of better than m / z = / z = +
10 higher than the PAH + inthe mass spectra, which was observed for all of the abovemen-tioned ions; the tagging yield was of the order 30 %. In addition,we also observed doubly and triply tagged ions at m / z = +
20 and +
30 (see Fig. A.1 in the appendix). The Ne-dication complexesare stored in the trap for typically 2.6 seconds where they are ir-radiated by IR radiation from the free-electron laser FEL-2. TheFEL was set to a narrow bandwidth (FWHM) of 0.4-0.5 % at10 Hz repetition rate and the spectrum was recorded in the rangeof 500 − − (20 − . µ m) with a typical laser pulse energyinside the trap of 10 −
40 mJ. The IRPD spectra were recordedby measuring the depletion of Ne-dication complex ion countsas a function of laser frequency. To account for fluctuations inthe number of complex ions and varying laser energy duringthe course of a single scan, the spectra were normalized to laserpulse energy and baseline corrected before averaging over mul-tiple scans. Saturation depletion measurements allowed us to de-code the isomer ratios of molecular ions having two or more sta-ble structures, see, for example, Jusko et al. (2018a). A detailedaccount of this technique is discussed in Jusko et al. (2019).
Geometry optimizations and frequency calculations were per-formed at the density functional theory (DFT) level. The Gaus-sian16 software package (Frisch et al. 2016) was used forall calculations as installed at the Cartesius supercomputer atSurfSARA, Amsterdam. The vibrational frequencies of doublycharged PAHs were calculated within the harmonic approxima-tion using the hybrid B3LYP functional with 6-311 + G(d,p) ba-sis set and scaled uniformly with 0.9679 (Andersson & Uvdal2005). Furthermore, anharmonic calculations within the vibra-
Article number, page 2 of 13. Banhatti, J. Palotás, P. Jusko, B. Redlich , J. Oomens, S. Schlemmer, and S. Brünken: Infrared action spectroscopy of doubly charged PAHsand their contribution to the aromatic infrared bands I o n c o un t s x anth anth + Fig. 1.
Ionization yield for m / z =
89 (anth + ) and m / z =
178 (anth + ) pro-duced in a non-storage electron ionization source.
80 90 100 110 120 130 140 150 160 170 180 190 m / z I o n c o un t s x C H C H C H +8 C H +8 C H +10 Fig. 2.
Mass spectrum using phenanthrene precursor in the mass rangem / z = / z = / z =
89, and the C H loss fragment at m / z =
152 are shown in thefigure. tional second-order perturbation level of theory (VPT2) wereperformed as implemented in Gaussian 16 with the same func-tional and basis set combination. We should note here that theVPT2 method implemented in Gaussian16 only provides a basictreatment of resonances. A more accurate calculation, which isbeyond the scope of the present work, would require an explicittreatment of resonating polyads with variational approaches, asdemonstrated earlier on the example of PAH molecules (Mackieet al. 2015, 2018; Chen 2018; Piccardo et al. 2015).To investigate the e ff ect of Ne-tagging on the vibrationalspectra, additional harmonic calculations were performed on thenaph + -Ne complexes. Several binding geometries of the Neatom to the PAH are conceivable with the Ne atom on top ofthe molecular plane and in plane with the molecule. The fivelowest energy conformers were optimized and harmonic vibra-tional frequencies were calculated (see appendix B.1). Here thedispersion corrected wB97XD functional with a cc-pVTZ basiswas used, which was found previously to work well for weakly-bound RG-ion complexes (Jusko et al. 2018a).
3. Results and discussion
The experimental IRPD spectrum of Ne-tagged naph + (Ne-IRPD) is shown in the top panel of Fig. 3. For comparison, thecalculated spectra of the singlet electronic ground state with A g symmetry and of the 0.31 eV higher lying triplet B g electronicstate are shown in the panels below. The peak positions fromthe fitted experimental spectrum and the corresponding band as-signments based on the calculations are listed in Table 1. Mostof the band positions in the Ne-IRPD spectrum can be readilyassigned to the singlet state of naph + , and they are in excellentagreement with the calculated anharmonic frequencies to within10 cm − for the fundamental bands.To account for the influence of the Ne-tag on the vibrationalband positions, we compared harmonic DFT calculations forbare and di ff erent conformers of Ne-tagged naph + . They re-vealed negligible band shifts of < − for most of the bands,with only a few bands maximally shifted by 5 cm − (see Ap-pendix B.1). The IRPD spectrum of the Ne-tagged species cantherefore be viewed as an excellent proxy for that of the bare ion.Only three of the experimental bands, at 1361 cm − ,1240 cm − , and 777 cm − , could not be assigned in a straight-forward manner. The band at 1362 cm − lies close to weak com-bination bands predicted at 1361 and 1365 cm − (not listed inTable 1). Several strong predicted combination bands were as-signed to a weak experimental feature at 1381 cm − , and theycould also be assigned to the 1362 cm − band if we assume largeanharmonic shifts unaccounted for by the calculations. Anotherfeature at 832 cm − assigned to a predicted band at 812 cm − shows a similarly large deviation. For combination bands, largerdeviations of calculated anharmonic band positions are expectedwhen using VPT2, so we tentatively assigned these features toone or a combination of the above combination bands.The rather strong feature at 1240 cm − is close to the strongC-H in plane bending vibrational mode at 1229 cm − , which al-most appears as a doublet. From the calculations, we can excludethat the symmetry breaking induced by the Ne-tag has a signif-icant influence on the spectrum, and that the presence of mul-tiple Ne-ion isomers are responsible for the observed splittingof the band. We therefore assume that the blue-shifted featureat 1240 cm − is due to a combination band of the strong C-Hbending mode with a vibration involving only the Ne-tag, a phe-nomenon often seen in rare-gas tagging experiments (Brünkenet al. 2019). The lowest harmonic fundamental frequencies in-volving the Ne-atom are shown in Fig. B.1. They fall in the5 −
60 cm − wavenumber range, and they are thus lying close tothe observed di ff erence between the two experimental features.However, we should note that our calculations also show a Fermiresonance of the 1229 cm − mode with a close-lying combina-tion band of the same symmetry, which might gain intensity dueto this interaction.The weak band at 777 cm − appears to be coincident with apredicted fundamental of triplet naph + at 764 cm − , but the ab-sence of a much stronger predicted triplet mode around 943 cm − in the experimental spectrum suggests otherwise. To test thepossible presence of the electronically excited triplet state, weperformed saturation depletion measurements on several strongbands assigned to singlet naph + , revealing an abundance of atleast 80 % of the ground electronic state. This indicates that upto 20 % of the formed dications could be in the electronicallyexcited triplet state or a di ff erent isomer (Leach et al. 1989a;Solano & Mayer 2015). Since we do not observe any other triplet Article number, page 3 of 13 & A proofs: manuscript no. Banhatti_Arxiv bands as expected from the calculated spectrum in Fig. 3 (bottompanel), the latter is more likely.Whereas the calculated, scaled harmonic band positions ofthe singlet state correlate well with those from the anharmoniccalculation, we can observe large shifts (up to 20 cm − ) be-tween both levels of theory, indicating the need to include modedependent anharmonic e ff ects to correctly describe the vibra-tional PAH dication spectra. Furthermore, the strong combina-tion bands between 1480-1490 cm − observed in the experimen-tal spectrum can be accounted for only with anharmonic calcula-tions. The narrow linewidths observed in our Ne-IRPD spectrumcan serve as a benchmark for anharmonic calculations of PAHdications just as previously shown for PAH neutrals (Mackieet al. 2015; Maltseva et al. 2015). The IRPD spectrum of phen + − Ne compares well with the cal-culated anharmonic spectrum of the A singlet electronic stateas seen in Fig. 4. Table 2 shows the peak positions with the as-signed anharmonic and harmonic frequencies with many of thepeak positions coinciding within 10 cm − of the calculated an-harmonic spectrum for the singlet ground state. Again, we donot observe any bands that could be assigned to the triplet B electronic state 0.44 eV higher (C.1).The bands below 800 cm − show larger shifts ( <
20 cm − )likely due to larger anharmonic e ff ects. Several combinationbands can be observed, most of which are weak and overshad-owed by the strong fundamental bands, except for the three iso-lated features at 1000, 1246, and 1355 cm − . Another strongfeature at 1427 cm − is a mix of fundamental and combinationbands making the overall feature appear broad.We observe a similar doublet feature in the phen + spec-trum at 1333 and 1343 cm − . As discussed above for naph + , wesuspect this blue-shifted feature to be due to a strong combina-tion band involving the Ne-tag. However, similar to naph + , ourcalculations reveal a Fermi resonance of the fundamental pre-dicted at 1337 cm − with two combination modes at 1350 and1378 cm − . The lower panels in Fig. 4 show the Ne-IRPD spectrum of anth + compared to calculated anharmonic spectra of its singlet A g electronic state. The band assignments are shown in Table 3.Due to the higher symmetry (D h ) of anth + compared to phen + (C v ), the spectrum shows fewer and less intense vibrationalbands. Once again, we do not see any trace of the energeticallyhigher-lying (0.79 eV) triplet B g electronic state (C.1).Several bands in the experimental spectrum coincide with thephen + bands (see caption in Fig. 4). These phen + bands appearto be due to contamination from the previous experiment wherephenanthrene was introduced in the source.The 1357 cm − band,for example, is a convolution of three bands, two of which areassigned to anth + with some contribution from phen + contam-ination of its 1355 cm − combination band.Since phen + is the more stable isomer based on our cal-culations (by 0.7 eV), it is possible that the phen + bands ob-served in the anth + spectrum are due to isomerization of anth + to phen + during the electron impact ionization process, as hasbeen predicted theoretically for the anthracene monocations (Jo-hansson et al. 2011). To verify this, we also recorded spectraat lower ionization energy (30 eV) expecting a change in the um ] IRPDHarmonic SingletAnharmonic Singlet
600 700 800 900 1000 1100 1200 1300 1400 1500 1600Wavenumber [cm ] I n t e n s i t y Anharmonic Triplet
Fig. 3.
IRPD spectrum of Ne-tagged naph + (top panel, blue) comparedwith calculated anharmonic and harmonic band positions of singletnaph + and anharmonic frequencies of triplet naph + . Both fundamentalmodes (blue) and combination modes (red) are shown. The calculatedspectrum was convoluted with a Gaussian lineshape function, where thewidth is given by the FEL bandwidth, and the area corresponds to thecalculated intensity in kmmol − . population of the phen + isomer and consequently a change inthe peak intensity of phen + bands relative to the anth + bands.This method was successfully implemented for benzylium andtropylium cations (C H + ), two isomers formed in dissociativeionization of toluene (Jusko et al. 2018a), but we did not ob-serve any change in the peak intensity for any of the phen + bands. Instead, the intensity of the phen + bands decreased overtime during the measurement campaign, leading us to concludethat they are due to contamination instead of isomerization. Rel-ative depletion values on the isolated phen + bands at 786 cm − and 854 cm − indicate a contamination of 40 − ff erence spectrum where the experimental (scaled)phen + spectrum was subtracted from the anth + spectrum to ac-count for this contamination. For this, we used an average 30%scaling factor over the whole spectral region, whereas the con-tamination varied from 40 −
20 % over the course of the measure-ments, as outlined above. This leads to some remaining phen + artifacts in the di ff erence spectrum, for example, several bandsbetween 1100 − − have not been completely removed,whereas several negative artifacts appear in the lower wavenum-ber range. Overall, however, this procedure allows for a bettercomparison to the calculated spectrum, as given in Table 3.
4. Astrophysical implications and conclusions
It is known that the charge state of the PAH plays an impor-tant role in the IR emission spectrum, especially in the 6-9 µ mregion where singly and multiply charged cations of PAH showstrong bands (Langho ff Article number, page 4 of 13. Banhatti, J. Palotás, P. Jusko, B. Redlich , J. Oomens, S. Schlemmer, and S. Brünken: Infrared action spectroscopy of doubly charged PAHsand their contribution to the aromatic infrared bands
Table 1.
Experimentally measured band positions v vib (cm − ), FWHM δ (cm − ), and relative intensities of the naphthalene dication comparedto DFT computed harmonic (scaled by 0.9679) and anharmonic fundamental and combination mode positions and intensities I (kmmol − ) > − . Values in brackets denote uncertainties (1 σ ) in units of the last significant digit. IRPD (This work) Anharm. calc. Harm. calc. symmv vib δ I rel v vib I I rel v vib I I rel
589 10.1 0.05 577 7.7 0.03 b u uc a
812 15 0.08 b u u u u u b a ua u u b a ua u - - - a u u ua u u a ua ua ua u u - - - a u Notes. ( a ) Combination bands for singlet naph + . ( b ) These bands are either too weak or broad due to closely located bands and cannot be fitted. ( c ) See text for discussion on this band. gions, they play an important role in the chemistry, which in turnis a ff ected by their stability. There has been some discussion inthe past on the stability of PAH dications, since their energiesare often higher than those of their singly charged fragmentsdue to Coulomb repulsion (Leach 1996). At higher internal en-ergies, 5 eV above the second ionization energy, for exampleprovided during the ionization process, fragmentation channelsvia covalent dissociation into a smaller dication fragment anda neutral also open up. We see this channel in the mass spec-trum for anthracene electron impact ionization above energies of23(2) eV with the appearance of a mass peak at m / z =
76, whichwe interpret as the C H fragment dication. However, often thedissociation is hindered by barriers in the potential energy sur-face along the dissociation coordinate, as has been shown in thecase of the benzene dication (Rosi et al. 2004; Jašík et al. 2014).Earlier mass-spectrometric studies reported a yield of roughly10 % for PAH dications upon ionization of the neutral, with onlymarginal variations for di ff erent PAH sizes and ionization meth-ods (e.g., EI and photo-ionization) (Leach et al. 1989a,b; Zhenet al. 2017, 2016). We see a similar behavior for the three PAHdications considered in this study. As exemplarily shown for theanthracene ions in Figs. 1 and 2, we reach dication to mono-cation ratios of up to 30 % at high electron impact energies,and a dication yield of around 10 % compared to all observedfragment plus parent ions. Evaluation of the abundance of PAHdications in the ISM, where they need to be produced by se-quential photo-ionization competing with dissociation, chemical reactions, and recombination processes, requires detailed mod-eling of the PAH evolution. In particular, the comparably smallPAH dications studied here are likely not stable in interstellarconditions (Montillaud et al. 2013; Zhen et al. 2015, 2016). Itwould therefore be interesting to extend the vibrational studiespresented here to the class of larger, astronomically more rele-vant PAH dications.To date, vibrational spectral information for PAH dicationscomes mainly from theoretical calculations, and we are awareof only one experimental IR study by Zhen et al. (2018). Mal-loci et al. did a comparative theoretical study on IR propertiesof 40 di ff erent PAH neutrals, monocations, and dications usingDFT and TD-DFT theoretical techniques. Based on these cal-culations, most of the dications were predicted to have a sin-glet ground state including the dications presented in this workand our recorded spectra support this finding. Another notablecharacteristic in the recorded IR spectra of all three dications isthe presence of several intense combination bands, especially inthe 5 − µ m region. Here, anharmonic calculations proved tobe crucial in assigning these combination bands. The presenceof combination bands has an e ff ect on the fraction of the totalintegrated intensity (InI) in the di ff erent spectral ranges 2.5-3.5,5-10, 10-15 µ m, and > µ m. Di ff erences in the relative inte-grated intensities in these spectral ranges serve as identifiers forthe charge state of PAHs in the ISM, as has been discussed indetail previously (Bauschlicher & Bakes 2000; Malloci et al.2007a). Whereas drastic changes in the integrated intensities Article number, page 5 of 13 & A proofs: manuscript no. Banhatti_Arxiv
Table 2.
Experimentally measured band positions v vib (cm − ), FWHM σ (cm − ), and relative intensities of the phenanthrene dication comparedto DFT computed harmonic (scaled by 0.9679) and anharmonic fundamental and combination mode positions and intensities I (kmmol − ) > − . Values in brackets denote uncertainties (1 σ ) in units of the last significant digit. IRPD (This work) Anharm. calc. Harm. calc.v vib σ I rel v vib I I rel v vib I I rel b a
873 9.8 0.07985(1) 9 0.58 983 72.7 0.55 962 83.5 0.351000(1) 12 0.28 a
996 7.8 0.061027(1) 10 0.65 1025 16.4 0.12 1008 22.9 0.101028 36.5 0.28 1010 39.6 0.17- - - a a a a a a a a d a,d a,d a a a a b a a a a a a a a a a a Notes. ( a ) Combination bands for singlet naph + . ( b ) These bands are either too weak or broad due to closely located bands and cannot be fitted . ( c ) See text for discussion on this band.Article number, page 6 of 13. Banhatti, J. Palotás, P. Jusko, B. Redlich , J. Oomens, S. Schlemmer, and S. Brünken: Infrared action spectroscopy of doubly charged PAHsand their contribution to the aromatic infrared bands
Table 3.
Experimentally measured band positions v vib (cm − ), FWHM σ (cm − ) and relative intensities of the anthracene dication compared to DFTcomputed harmonic (scaled by 0.9679) and anharmonic fundamental and combination mode positions and intensities I (kmmol − ) > − .Values in brackets denote uncertainties (1 σ ) in units of the last significant digit. IRPD (This work) Anharm. calc. Harm. calc.v vib σ I rel v vib I I rel v vib I I rel a
845 8.6 0.025 c c a c c a a a c a a c c a a a a c a a a a a a a a Notes. ( a ) Combination bands for singlet naph + . ( b ) These bands are either too weak or broad due to closely located bands and cannot be fitted. ( c ) See text for discussion on this band. were observed between neutral and cationic species, there weremuch fewer predicted variations between the singly and doublycharged PAHs. The question arises if the appearance of strongcombination bands changes this picture. A comparison is madein Appendix C showing the e ff ect of including anharmonicity inthe calculations on the integrated intensities of singly and doublycharged PAHs considered in this work. As the choice of the ba-sis set in theoretical calculations of vibrational spectra has beendiscussed previously (Andersson & Uvdal 2005; Bauschlicher &Langho ff / / ff ect on InIfor either the singly or doubly charged species. However, whenperforming anharmonic calculations (with the B3LYP / µ m region is 5-7% higherfor both singly and doubly charged PAHs compared to harmoniccalculations using the same functional. In contrast, in the 5-10 µ m region, InI predicted by anharmonic calculations is lowercompared to harmonic calculations by 28 % for naph + , 15% foranth + , and phen + , but only 2-5 % for doubly charged PAHs. Inthe spectral regions > µ m and 10-15 µ m, both singly and dou- Article number, page 7 of 13 & A proofs: manuscript no. Banhatti_Arxiv um ]Naph NePhen Ne * * * * * * * * *o Anth
Ne600 700 800 900 1000 1100 1200 1300 1400 1500 1600Wavenumber [cm ] I n t e n s i t y (Anth Ne - Phen
Ne) * phen + bands due to contamination, ◦ overlapping phen + and anth + bands Fig. 4.
IRPD spectrum of Ne-taggged dications naph + , phen + , and anth + compared to calculated anharmonic (B3LYP / − . The spectrum in the bottom panel is the anth + -Ne spectrum subtracted by30% of the weight of the relative intensity of the phen + -Ne spectrum to account for the contaminated bands. bly charged anth and phen show negligible ( < ff erences observed in the relative integratedintensities in each of the spectral regions and the presence ofcombination bands suggest that anharmonic calculations are nec-essary to predict accurate vibrational spectra for the family ofcomparatively small PAH cations as discussed here and also inprevious work (Maltseva et al. 2015; Lemmens et al. 2019).Thus, including these anharmonic e ff ects has consequences on the interpretation of relative UIR band intensities observed inthe ISM.Here, we have shown that cryogenic IRPD experiments usingNe-tagging prove to be a very powerful method to obtain vibra-tional spectra of PAH dications by providing narrow features andintensities being more comparable to the linear absorption crosssection. This makes it a very e ff ective tool to benchmark theoret-ical calculations. In order to validate our finding that anharmonice ff ects significantly influence the IR band positions and intensi-ties of PAH monocations and dications, future IRPD experimentstargeting larger and structurally di ff erent classes of PAH cationsshould be conducted. As discussed above, the double ionizationprocess competes with fragmentation channels. One of the mainfragmentation channels observed for PAHs is the loss of acety-lene, C H (Johansson et al. 2011; Simon et al. 2017; Ling & Lif- Article number, page 8 of 13. Banhatti, J. Palotás, P. Jusko, B. Redlich , J. Oomens, S. Schlemmer, and S. Brünken: Infrared action spectroscopy of doubly charged PAHsand their contribution to the aromatic infrared bands shitz 1998), as we have also observed in this study (see Fig. 2).Elucidating the structure of these (C H )-loss fragment ions, ashas been previously done for the case of naphthalene (Bouwmanet al. 2016), is another interesting application for the narrow-linewidth IRPD action spectroscopy that we have presented here. Acknowledgements.
This project is funded by the Marie Skłodowska Curie Ac-tions (MSCA) Innovative Training Networks (ITN) H2020-MSCA-ITN 2016(EUROPAH project, G. A. 722346). We are grateful for the experimental supportprovided by the FELIX team and acknowledge the Nederlandse Organisatie voorWetenschappelijk Onderzoek (NWO) for the support of the FELIX Laboratory.we also thank NWO Exact and Natural Sciences for the use of supercomputerfacilities (Grant nr. 2019.062). We thank the Cologne Laboratory Astrophysicsgroup for providing the FELion ion trap instrument for the current experimentsand the Cologne Center for Terahertz Spectroscopy (core facility, DFG grantSCHL 341 / References
Allamandola, L. J., Hudgins, D. M., & Sandford, S. A. 1999, ApJ, 511, L115Allamandola, L. J., Tielens, A. G. G. M., & Barker, J. R. 1985, ApJL, 290, L25Allamandola, L. J., Tielens, A. G. G. M., & Barker, J. R. 1989, ApJS, 71, 733Alvaro Galué, H., Pirali, O., & Oomens, J. 2010, A&A, 517, A15Andersson, M. P. & Uvdal, P. 2005, J. Phys. Chem. A, 109, 2937Asmis, K. R., Brümmer, M., Kaposta, C., et al. 2002, Phys. Chem. Chem. Phys.,4, 1101Bakes, E., Tielens, A., Bauschlicher Jr, C. W., Hudgins, D. M., & Allamandola,L. J. 2001a, ApJ, 560, 261Bakes, E. L. O., Tielens, A. G. G. M., & Bauschlicher, C. 2001b, ApJ, 556, 501Bakker, J. M., Redlich, B., van der Meer, A. F. G., & Oomens, J. 2011, ApJ, 741,74Bauschlicher, C. & Bakes, E. 2000, Chem. Phys., 262, 285Bauschlicher, C. W. & Langho ff , S. R. 1997, Spectrochim. Acta A Mol. Biomol.Spectrosc., 53, 1225 , ab Initio and Ab Initio Derived Force Fields: State ofthe ScienceBouwman, J., Boersma, C., Bulak, M., et al. 2020, A&A, 636, A57Bouwman, J., de Haas, A. J., & Oomens, J. 2016, Chem. Commun., 52, 2636Brünken, S., Lipparini, F., Sto ff els, A., et al. 2019, J. Phys. Chem. A, 123, 8053Chen, T. 2018, The Astrophysical Journal Supplement Series, 238, 18Cohen, M., Allamandola, L., Tielens, A. G. G. M., et al. 1986, ApJ, 302, 737Dartois, E. & D’Hendecourt, L. 1997, A&A, 323, 534Frisch, M. J., Trucks, G. W., Schlegel, H. B., et al. 2016, Gaussian~16 RevisionC.01, gaussian Inc. Wallingford CTGalliano, F., Madden, S. C., Tielens, A. G. G. M., Peeters, E., & Jones, A. P.2008, ApJ, 679, 310Gao, J., Berden, G., & Oomens, J. 2014, ApJ, 787, 170Gerlich, D. 1992, Inhomogeneous RF fields: A versatile Tool for the Study ofprocesses with Slow Ions, Vol. LXXXII (Wiley, New York), 1–176Gerlich, D., Jašík, J., Strelnikov, D. V., & Roithová, J. 2018, ApJ, 864, 62Günther, A., Nieto, P., Müller, D., et al. 2017, J. Mol. Spectrosc., 332, 8, molec-ular Spectroscopy in TrapsHolm, A. I. S., Johansson, H. A. B., Cederquist, H., & Zettergren, H. 2011, J.Chem. Phys, 134, 044301Hony, S., Van Kerckhoven, C., Peeters, E., et al. 2001, A&A, 370, 1030Hudgins, D., Allamandola, L., & Sandford, S. 1997, Advances in Space Re-search, 19, 999 , proceedings of the F3.2 Symposium of COSPAR ScientificCommission FHudgins, D. M. & Allamandola, L. J. 1999, ApJ, 513, L69Jašík, J., Gerlich, D., & Roithová, J. 2014, J. Am. Chem. Soc., 136, 2960Johansson, H. A. B., Zettergren, H., Holm, A. I. S., et al. 2011, J. Chem. Phys,135, 084304Johnson, D. R. 2019, NIST Computational Chemistry Comparison and Bench-mark Database, nIST Standard Reference Database Number 101, Release 20,August 2019, http: // cccbdb.nist.gov / .Jourdain de Muizon, M., D’Hendecourt, L. B., & Geballe, T. R. 1990, A&A,235, 367Jusko, P., Brünken, S., Asvany, O., et al. 2019, Faraday Discuss., 217, 172Jusko, P., Simon, A., Banhatti, S., Brünken, S., & Joblin, C. 2018a,ChemPhysChem, 19, 3173Jusko, P., Simon, A., Wenzel, G., et al. 2018b, Chem. Phys. Lett, 698, 206Langho ff , S. R. 1996, J. Phys. Chem. A, 100, 2819Leach, S. 1986, J Electron Spectros Relat Phenomena, 41, 427Leach, S. 1996, Zeitschrift für Physikalische Chemie, 195, 15Leach, S., Eland, J., & Price, S. 1989a, J. Phys. Chem. A, 93, 7575Leach, S., Eland, J., & Price, S. 1989b, J. Phys. Chem. A, 93, 7583Leger, A. & Puget, J. L. 1984, A&A, 137, L5 Lemmens, A. K., Rap, D. B., Thunnissen, J. M. M., et al. 2019, A&A, 628, A130Ling, Y. & Lifshitz, C. 1998, J. Phys. Chem. A, 102, 708Lorenz, U., Solcà, N., Lemaire, J., Maître, P., & Dopfer, O. 2007, AngewandteChemie International Edition, 46, 6714Mackie, C. J., Candian, A., Huang, X., et al. 2015, J. Chem. Phys, 143, 224314Mackie, C. J., Candian, A., Huang, X., et al. 2018, Phys. Chem. Chem. Phys.,20, 1189Malloci, G., Joblin, C., & Mulas, G. 2007a, Chem. Phys., 332, 353Malloci, G., Mulas, G., Cappellini, G., & Joblin, C. 2007b, Chem. Phys., 340,43Maltseva, E., Petrignani, A., Candian, A., et al. 2015, ApJ, 814, 23Montillaud, J., Joblin, C., & Toublanc, D. 2013, A&A, 552, A15Oepts, D., van der Meer, A. F. G., & van Amersfoort, P. W. 1995, Infrared Phys.Technol., 36, 297Oomens, J., Sartakov, B. G., Tielens, A. G. G. M., Meijer, G., & von Helden, G.2001, ApJ, 560, L99Oomens, J., van Roij, A. J. A., Meijer, G., & von Helden, G. 2000, ApJ, 542, 404Panchagnula, S., Bouwman, J., Rap, D. B., et al. 2020, Phys. Chem. Chem. Phys.,Piccardo, M., Bloino, J., & Barone, V. 2015, International Journal of QuantumChemistry, 115, 948Piest, H., Oomens, J., Bakker, J., von Helden, G., & Meijer, G. 2001, Spec-trochim. Acta A Mol. Biomol. Spectrosc., 57, 717Piest, H., von Helden, G., & Meijer, G. 1999, ApJ, 520, L75Ricks, A. M., Douberly, G. E., & Duncan, M. A. 2009, ApJ, 702, 301Rosi, M., Bauschlicher Jr, C. W., & Bakes, E. 2004, ApJ, 609, 1192Sellgren, K. 1984, ApJ, 277, 623Sellgren, K., Allamandola, L. J., Bregman, J. D., Werner, M. W., & Wooden,D. H. 1985, ApJ, 299, 416Simon, A., Rapacioli, M., Rouaut, G., & Gadéa, F. 2017, Philos. Trans. RoyalSoc. A, 375, 20160195Solano, E. A. & Mayer, P. M. 2015, J. Chem. Phys., 143, 104305Tielens, A. 2008, Annu. Rev. Astron. Astrophys., 46, 289van der Burgt, P. J. M., Dunne, M., & Gradziel, M. L. 2018, The European Phys-ical Journal D, 72, 31van der Burgt, P. J. M., Dunne, M., & Gradziel, M. L. 2019, J. Phys. Conf. Ser.,1289, 012008Wenzel, G., Joblin, C., Giuliani, A. A., et al. 2020, A&AWitt, A. N., Gordon, K. D., Vijh, U. P., et al. 2006, ApJ, 636, 303Zhen, J., Candian, A., Castellanos, P., et al. 2018, ApJ, 854, 27Zhen, J., Castellanos, P., Bouwman, J., Linnartz, H., & Tielens, A. G. G. M.2017, ApJ, 836, 28Zhen, J., Castellanos, P., Paardekooper, D. M., et al. 2015, The AstrophysicalJournal, 804, L7Zhen, J., Castillo, S. R., Joblin, C., et al. 2016, ApJ, 822, 113 Article number, page 9 of 13 & A proofs: manuscript no. Banhatti_Arxiv
Appendix A: Mass spectrum showing doubly charged anthracene tagged with Ne
85 87 89 91 93 95 97 99 101 103 105 107 109 m / z I o n c o un t s x C H
102 + C H
102 +
Ne C H Fig. A.1.
Mass spectrum showing anth + (C H + , m / z =
89) with one (C H + – Ne, m / z =
99) and two (C H + -Ne , m / z = ∆ m / z = + (at lower mass) and C substituted or hydrogenated phen + (higher masses), respectively, see also van der Burgt et al.(2018). The mass resolution of the second quadrupole mass filter is better than m / z = Appendix B: Comparison of vibrational spectra of bare and Ne-tagged Naph + um ] E = 0kJ/mol I n t e n s i t y E = 3.18kJ/molE = 4.07kJ/molE = 10.18kJ/mol
20 40 60
E = 4.09kJ/mol
800 900 1000 1100 1200 1300 1400 1500 1600Wavenumber [cm ] Fig. B.1.
Vibrational spectra of doubly charged bare naphthalene and neon-tagged naphthalene calculated using DFT within the harmonic approx-imation at the wB97XD / cc-pVTZ level. The respective energies of the five lowest conformers relative to the lowest energy conformer are stated.The frequency shifts between the di ff erent neon isomers and the bare ion are negligible. The three bands observed in the 0-60 cm − region are thebending and stretching vibrations involving the neon atom where we observe large shifts in frequencies among its isomers.Article number, page 11 of 13 & A proofs: manuscript no. Banhatti_Arxiv
Appendix C: Comparison between anharmonic and harmonic calculations
Anharmonic ( X + )B3LYP/6-311GAnharmonic ( X )B3LYP/6-311G Harmonic ( X + )B3LYP/6-311GHarmonic ( X )B3LYP/6-311G 0246810121416 Harmonic ( X + )B3LYP/6-311GHarmonic ( X )B3LYP/6-311G Harmonic ( X + )B3LYP/4-31GHarmonic ( X )B3LYP/4-31G 455055606570758085 % o f t o t a l i n t e g r a t e d i n t e n s i t y PhenanthreneAnthraceneNaphthalene 7072747678808284
PhenanthreneAnthraceneNaphthalene8121620242832
10 11 12 13 14No of Carbon atoms4681012 >15um
10 11 12 13 14No of Carbon atoms345678 >15um
Fig. C.1.
Integrated intensities in di ff erent regions of the mid IR spectra. Left Panel: Comparison made between harmonic calculations using theB3LYP / / / Table C.1.
DFT computed harmonic (scaled by 0.9679) fundamental mode positions v vib (cm − ) and intensities I (kmmol − ) > − of tripletelectronic ground state for naph + , anth + , and phen + . Naph + Anth + Phen + v vib I v vib
I v vib
I137 5.1 252 28.2 346 12407 17.6 410 26.7 401 8.1755 97.5 548 23.9 712 91.8922 216.9 715 87.6 787 22976 9.6 741 113.6 808 47.11064 57.4 781 153.8 849 13.71138 32.6 864 50.9 927 119.91186 15.2 880 31.9 991 9.61189 11.8 902 296.9 1019 20.61365 19.7 943 16.5 1046 9.41401 269.2 1133 13.1 1085 10.51416 250.9 1294 210.9 1114 183082 8 1349 386.7 1164 453089 94.7 1373 81.8 1170 168.53093 91.6 1376 269.5 1209 61393 42.5 1302 7.82982 7.5 1310 24.12989 6.1 1364 8.82998 41.8 1381 101.43003 48.2 1385 34.21422 59.41439 13.81441 87.21490 44.31509 25.62996 32.23002 493009 11.23022 10.4
Fig. C.2.
Naphthalene dication showing the orientation of the molecular axes used in the calculations ( II r representation), defining the symmetrylabels given in Table 1. Inertia moments: Iy < Iz <<