The Astrophysical Journal, 648:1285Y1290, 2006 September 10 # 2006. The American Astronomical Society. All rights reserved. Printed in U.S.A.
THEORETICAL MODELING OF FORMIC ACID (HCOOH), FORMATE (HCOO), AND AMMONIUM ( NHþ 4 ) VIBRATIONAL SPECTRA IN ASTROPHYSICAL ICES Jin-Young Park and David E. Woon Molecular Research Institute, 2495 Old Middlefield Way, Mountain View, CA 94043; [email protected]
Received 2006 February 7; accepted 2006 May 23
ABSTRACT Ions embedded in icy grain mantles are thought to account for various observed infrared spectroscopic features, particularly in certain young stellar objects. The dissociation of formic acid (HCOOH ) in astrophysical ices to form the formate ion (HCOO) was modeled with density functional theory cluster calculations. Like isocyanic acid (HOCN ), HCOOH was found to spontaneously deprotonate when sufficient water is present to stabilize charge transfer complexes. Both ammonia and water can serve as proton acceptors, yielding ammonium (NHþ 4 ) and hydronium (H3O+) counterions. Computed frequencies of weak infrared features produced by stretching and bending modes in both HCOO and HCOOH were compared with experimental and astronomical data. Our results confirm laboratory assignments that a band at 1381 cm1 can be attributed to the CH bend in either HCOO or HCOOH, but a band at 1349 cm1 corresponds to CO stretching in HCOO. Another feature at 1710 cm1 (5.85 m) can possibly be assigned to a CO stretching mode in HCOOH, as suggested by experiment, but the agreement is less satisfactory. In addition, we examine and analyze spectroscopic features associated with NHþ 4 , both as a counterion to HCOO or OCN and in isolation, in order to compare with experimental and astronomical data in the 7 m region. Subject headingg s: astrochemistry — ISM: molecules — molecular processes Online material: color figures OCN. There are a few additional tentative identifications: deuterated water (HDO), the formate ion (HCOO) or acetaldehyde (CH3CHO), and ammonia (NH3). Similar inventories have been observed in other sources (Gibb et al. 2004). The identification of the cyanate ion (OCN) was the subject of uncertainty and debate for a number of years after the initial observation of a broad feature at 2165 cm1 (4.62 m) in W33A by Soifer et al. (1979) and subsequently by other observers in various YSOs and other objects. While Grim & Greenberg (1987) conjectured quite early that the band was due to OCN with ammonium (NHþ 4 ) serving as a counterion, experimental confirmation did not follow until somewhat later (Schutte & Greenberg 1997; Demyk et al. 1998; Hudson et al. 2001; Novozamsky et al. 2001). Even then, some uncertainty remained. In prior work (Park & Woon 2004a, 2004b), we used quantum chemical theory to characterize the behavior of HNCO and HOCN dissociation in ice mantles and found that + OCN in charge transfer complexes with NHþ 4 and H3O reproduced the strong feature at 2165 cm1, as well as two weaker features. Furthermore, we were also able account for shifts of the principal band due to distinct isotopic substitutions of C, N, O, and H (Bernstein et al. 2000; Palumbo et al. 2000; Novozamsky et al. 2001). While there may be circumstances in which a feature arises at 4.62 m due to another carrier (Fraser et al. 2005), the OCN assignment appears to be very secure when ices are likely to be present and certain compositional constraints are met. Another anion that is believed to be present in ice mantles in W33A and elsewhere is HCOO. Schutte et al. (1997, 1999) proposed that a weak band at 1349 cm1 (7.41 m) could be assigned to the CO stretch of HCOO, while a band at 1381 cm1 (7.24 m) could be assigned to the CH bending motion of HCOOH or HCOO. They demonstrated that HCOO /NHþ 4 charge transfer complexes formed when the dilute mixture H2 O/ HCOOH/NH3 ¼ 100/3:6/3:6 was deposited at 10 K and warmed to 120 K. New features formed that were assigned to HCOO or NHþ 4 , while HCOOH and NH3 features lost intensity. Formate ion features appeared upon deposition, indicating that
1. INTRODUCTION Ice mantles may coat interstellar or circumstellar dust grains under certain conditions, providing sites where the chemistry can be quite different from that which prevails in the gas phase (Herbst 1993; Ehrenfreund & Fraser 2003). Characterizing the range of chemical behavior that can occur on or within ice mantles and the spectroscopy of ice-bound species is essential to a full understanding of the nature and evolution of both the gas- and condensed-phase composition in young stellar objects ( YSOs) and elsewhere. The implications of gas-grain chemistry affect both cloud modeling and the astrobiological search for plausible precursors to life. Much of what is known about the condensed-phase component of interstellar and circumstellar clouds is derived from vibrational spectroscopic features in the mid-infrared (2.5Y25 m) region (Sandford 1996). With limited windows available to groundbased instruments due to atmospheric interference, the Infrared Space Telescope ( Kessler et al. 1996) and its successor, the Spitzer Space Telescope ( Werner et al. 2004), finally enabled high-resolution observation of the entire mid-infrared spectrum. The astronomical data are interpreted by comparisons with the spectra of astrophysical ice analogs obtained in the laboratory at low temperatures and under high or ultrahigh vacuum conditions. The literature is extensive, ranging from compendia of pure and multicomponent ices (e.g., Hagen et al. 1983; Hudgins et al.1993; Ehrenfreund et al. 1996a, 1996b; Boudin et al. 1998) to detailed studies on particular features, behavior, or issues (e.g., Elsila et al. 1997; Fraser et al. 2005). Spectroscopic predictions can also be generated computationally using quantum chemical theory, which allows additional insight into molecular level behavior not easily accessible to experimental study. The massive deeply embedded protostar W33A has been thoroughly assessed to determine which molecules are present in the form of ices (Gibb et al. 2000). Observed species include H2O, CO, CO2, 13CO2, CH4, CH3OH, H2CO, HCOOH, OCS, and 1285
PARK & WOON
proton transfer to NH3 is not impeded by a barrier. While they observed that the 1381 cm1 band could arise from the CH bending mode of either HCOOH or HCOO, they assigned the astronomical feature to HCOOH on the basis of their measured values for the intensities of these bands and their relationship to features with unambiguous assignments. In addition to the two bands described above, they found another, somewhat more intense band for HCOO at 1580 cm1 that is also due to CO stretching. In subsequent work, Keane et al. (2001) concluded that a band at 1710 cm1 (5.85 m) could be assigned to HCOOH. In this work, we investigate the formation of charge transfer complexes involving the deprotonation of formic acid to form HCOO and examine the band positions and intensities for the CO stretching and the CH bending modes. In parallel with our work, Allouche and coworkers have also recently examined the behavior of HCOOH in ice (Allouche 2005; Bahr et al. 2005). Relevant results from this work are described in context below. Unresolved questions also persist regarding the spectra and behavior of NH3 and NHþ 4 embedded in ice mantles. Keane et al. (2001) discuss the possible assignment of a feature at 1460 cm1 (6.85 m) to NHþ 4 , but they note that there is a lack of evidence supporting a correlation between the presence of OCN and this band, as one would expect if it arises as a counterion in acidbase reactions in grain mantles. In the experiment of Schutte et al. (1999) described above, a broad band develops centered on 1490 cm1, which they attribute to NHþ 4 counterions. Subsequently, Schutte & Khanna (2003) performed further laboratory studies on NHþ 4 production in irradiated ices and reported a second feature at 3067 cm1 (3.26 m). Using results from the HCOO complexes as well as unreported predictions from our previous work on OCN ( Park & Woon 2004a, 2004b), we also analyzed the spectroscopic properties of NHþ 4 as an embedded ion. Additional calculations were performed for the NHþ 4 ion alone, with no counterion present, and for other related species that might account for the observed band. 2. METHODOLOGY Molecules embedded within a condensed phase medium exhibit different spectroscopic behavior than species in the gas phase, particularly if the medium is composed of a strongly interacting, polarizing species such as water. At cryogenic temperatures, most rotation is inhibited in water ice by the network of intermolecular bonds and by steric hindrance. Vibrational motion is likewise affected by the constraints of neighboring molecules. Furthermore, polarization distorts molecular charge distributions with respect to the gas phase throughout the material, also impacting vibrational motion. Specific interactions between molecules can induce even larger changes, ranging from a small amount of charge transfer in hydrogen bonds to large chemical changes, such as the stabilization of charge transfer pairs observed in the present work. Vibrational modes may also couple widely, leading to motion that can be very delocalized. The quantum chemical cluster approach is well suited for modeling the spectroscopic properties of molecules and ions embedded within amorphous ice matrices. The size scaling of density functional theory (DFT) methods makes it possible to routinely treat clusters of fifteen or more small molecules. This is large enough to account for most of the relevant short-range near-neighbor interactions that affect the chemical nature and spectroscopic behavior of an embedded species. Longer range, bulk effects are not incorporated, but our results indicate that the model is sufficient. If the active vibrational carriers in an ice mantle are located near the surface, a continuum approach to incorporating bulk effects may actually include too much polarization.
Fig. 1.— Structures of formic acid, formate ion, and ammonium ion with bond lengths in angstroms as computed at the B3LYP/6-31+G level. [See the electronic edition of the Journal for a color version of this figure.]
All calculations were performed with the GAUSSIAN 03 suite (Frisch et al. 2003) of electronic structure programs using DFT at the B3LYP level (Becke 1993; Lee et al. 1988) and 6-31+G basis sets (Hariharan & Pople 1973; Clark et al. 1983). Second derivatives were computed to verify that structures are true stationary points and to provide zero-point energy ( ZPE) corrections. Computed gas-phase structures for HCOOH, HCOO, and NHþ 4 are shown in Figure 1. The natures of the CO stretches in HCOOH and HCOO in the gas phase are somewhat different. The distinct C=O (carbonyl) and COH bonds of HCOOH become equivalent bonds in HCOO when the proton is removed, resulting in the typical symmetric and asymmetric couplings. The symmetry of the CO bonds is broken when HCOO is embedded in ice, but the CO stretching frequencies of HCOOH and HCOO remain quite distinct. By contrast, the CH in-plane bends of HCOOH and HCOO are very similar both in the gas phase and in ice. Due to symmetry, there is only one IR-active deformation mode in NHþ 4 in the gas phase. As we show below, the breaking of symmetry in ice results in three NH deformation peaks, which is consistent with the pronounced broadness of the peak assigned to NHþ 4 . This mode consists of various combinations of HNH bending motions. Computed harmonic frequencies for the cluster calculations discussed below were scaled to correct for anharmonicity and methodological error. A single scaling factor is often used for a given combination of method and basis set. Studies such as that of Scott & Radom (1996) indicate that the rms standard deviation for a single scaling factor can be 30 cm1 or more for B3LYP. We have attempted to improve the accuracy of our predictions by deriving mode-specific scale factors whenever possible from empirical molecular vibrational frequencies for HCOOH, HCOO, and NHþ 4 taken from the NIST Chemistry WebBook database (Shimanouchi 2005).1 The following scale factors were applied to HCOO: symmetric CO stretch, 0.983; asymmetric CO stretch, 0.973; and CH in-plane bend, 0.978. Due to the absence of experimental data for the CH in-plane bend, the average of the two CO stretches was used for that mode. The scale factors used for HCOOH were CO( H ) stretch, 0.975; C=O stretch, 0.972; and 1
No. 2, 2006
FORMATE AND AMMONIUM IONS IN ICES
Fig. 2.— Small cluster structures of HCOOH and HCOO with selected intermolecular bond lengths in angstroms as computed at the B3LYP/6-31+G level. [See the electronic edition of the Journal for a color version of this figure.]
CH in-plane bend, 0.989. As mentioned above, the symmetry of NHþ 4 is broken when the ion is embedded within an ice matrix. The scale factor derived for the single NH deformation mode observed in the gas phase (0.968) was used for all three bands observed in the cluster calculations. Note that all of the scale factors are in the range 0.97Y0.99, indicating that the corrections are modest. The agreement between our predictions and available experimental data suggests that the errors associated with modespecific scaling factors are somewhat less than those associated with a single scaling factor. 3. RESULTS AND DISCUSSION 3.1. HCOOH and HCOO Proton transfer from formic acid in an ice matrix was studied in clusters of various sizes with and without NH3 present as a base. All structures were optimized at the B3LYP/6-31+G level. Representative small clusters that include 2Y6 H2O molecules are depicted in Figure 2; large clusters with 14 or 15 H2O molecules are shown in Figure 3. Proton transfer does not occur in HCOOHNH3-2H2O, but it occurs spontaneously in the case in which a third H2O bridges between the acid and the base and in all larger clusters that include this minimal coordination around the product HCOO and NHþ 4 ions. In the larger clusters, we see the typical hydrogen bonding network consisting of rings with varying numbers of water molecules. As in OCN ( Park & Woon 2004a, 2004b), the polar end of HCOO has six coordination sites. Even in the largest clusters, the hydrophobic H of HCOO remains un-
connected to the hydrogen bonding network due to the low polarity of the CH bond. Also as in OCN, we found that HCOOH will transfer a proton to water even when no NH3 is present as an acceptor if there is sufficient water included. During the optimization of that case, the proton once again shifted to a water not directly coordinated to HCOO (see the HCOO-H3O+-14H2O cluster in Fig. 3). We also optimized small cluster structures in which HCOOH does not lose a proton to H2O as well as a large cluster in which no counterion was present (HCOO-15H2O) as part of our assessment of the spectroscopy of HCOO in ice. The intensities and scaled frequencies of CO stretching modes and the CH bending mode of HCOOH or HCOO from eight cluster structures are collected in Table 1. For HCOO, the symmetric CO stretching frequency is near 1300 cm1 in small clusters. As the coordination approaches that of the full ice matrix, the values rise to 1346 and 1351 cm1 for the respective cases in + which NHþ 4 and H3O are present as a counterion. These are both very close to the experimental band position at 1349 cm1. For the CH bend, the predictions with only a few water molecules present are closer to the large cluster cases, where the frequency is 1380 cm1 for either counterion. This is exceptionally close to the measured band at 1381 cm1. Note that the counterion appears to be necessary to match experiment: in HCOO-15H2O, the frequencies of both CO stretches and the CH bend are shifted to somewhat higher values, particularly for the former. Our calculations with a counterion present also reproduce the band due to asymmetric CO stretching that Schutte et al. (1999) found at + 1580 cm1, falling at 1580 and 1584 cm1 for NHþ 4 and H3O ,
PARK & WOON
Fig. 3.— Large cluster structures of HCOO as computed at the B3LYP/6-31+G level. The location of the H3O+ ion in the middle structure is indicated with an arrow. [See the electronic edition of the Journal for a color version of this figure.]
respectively. When a counterion is present, the computed intensities of the three bands are quite comparable to those reported by Schutte et al.: our values are 2.4Y2.8, 0.5Y 0.7, and 9Y11 ; 1017 cm molecule1 for the symmetric CO stretch, the CH bend, and the asymmetric CO stretch, respectively, compared with the experimental values of 1.7, 0.8, and 10 ; 1017 cm molecule1. The consistency of the intensity predictions therefore supports their conclusion that the 1380 cm1 feature in W33A is due to CH bending in HCOOH, not HCOO. Table 1 also includes our predictions for the analogous modes in HCOOH in clusters of limited size (since proton transfer oc-
curred in the largest cluster). In a 5H2O cluster (not shown in Fig. 2), HCOOH is more closely coordinated with water than it is in the 6H2O case, where HCOOH is bound more tenuously to a water hexamer ring that might represent a surface binding configuration. The frequencies of the two cases are somewhat different as a consequence. There is better agreement with experiment for the CH bend when HCOOH has more coordination with water (1385 vs. 1411 cm1). The C=O stretching frequency is 1744 and 1734 cm1 for the five- and six-water cases, respectively. This is near the band at 1710 cm1 that Keane et al. (2001) attribute to HCOOH, but the discrepancy is too large to conclude that our
No. 2, 2006
FORMATE AND AMMONIUM IONS IN ICES
TABLE 1 Intensities and Scaled Computed Frequencies of HCOOH and HCOO Modes in Various Clusters Symmetric CO Stretch
Asymmetric CO Stretch
HCOO-NHþ 4 -3H2O .............. HCOO-NHþ 4 -5H2O .............. HCOO-NHþ 4 -14H2O ............ HCOO-H3O+-14H2O ........... HCOO-15H2O...................... HCOOH-NH3-2H 2O ............. HCOOH-5H2O....................... HCOOH-6H2O.......................
1314 1300 1346 1351 1396 1198 1166 1205
2.64 5.26 2.81 2.42 4.50 4.13 5.03 4.34
1373 1378 1380 1380 1396 1381 1385 1411
0.52 0.19 0.49 0.73 0.15 1.25 0.11 0.35
1578 1607 1584 1580 1595 1698 1744 1734
8.71 7.20 9.22 11.16 4.00 16.28 8.46 7.74
In units of 1017 cm molecule1.
predictions confirm the experiment. However, our computed intensities of 8:5 ; 1017 and 7:7 ; 1017 cm molecule1 are close to the value of 6:7 ; 1017 cm molecule1 that Keane et al. quote from prior work by Mare´chal (1987). Recently, companion experimental (Bahr et al. 2005) and theoretical (Allouche 2005) studies considered the adsorption of HCOOH on ice. While the emphasis of the computational work is on binding energies and clustering of formic acid, they also report a structure that is very similar to our HCOOH-6H2O cluster, with formic acid bridging across a hexameric H2O ring. 3.2. NHþ 4 Our focus above and in previous studies ( Park & Woon 2004a, 2004b) has been on the formation and spectroscopy of the OCN and HCOO anions, using NH3 as a proton acceptor. However, the spectroscopy of NH3, NHþ 4 , and related carriers embedded in ice is also of interest. To enable us to expand the analysis beyond the previous cluster calculations, several additional structures were optimized with no counterion present: NHþ 4 with seven water molboth with nine water molecules. These ecules and NH3 and NHþ 3 are depicted in Figure 4. Intensities and scaled frequencies for nine cluster structures are gathered in Table 2. In all of the clusters with NHþ 4 , three peaks are present near 1500 cm1. The spread between them varies from 23 to 109 cm1, depending on cluster size and the presence and location of a negative counterion. For the three largest clusters with HCOO or OCN, the center band falls at 1504, 1501, and 1500 cm1. This is close to the band at 1490 cm1 reported by Schutte et al. (1999) that they assign to NHþ 4 . The single, broad laboratory feature appears to arise from three overlapping, symmetry-broken NH deformation peaks that possess similar intensities of about (1Y2) ; 1017 cm molecule1. We certainly see the NHþ 4 features if HCOO or OCN are present. The intensities are similar to that of the 1350 cm1 CO stretching mode of HCOO. However, it is much weaker than the 2165 cm1 band of OCN, which we predicted to have an intensity of about 2 ; 1016 cm molecule1 (Park & Woon 2004a, 2004b), an order of magnitude larger than the NHþ 4 band. It is possible that recognition of this mismatch will affect the conclusion about the nature of the carrier for the 6.85 m band. While Schutte & Khanna (2003) also assigned a band at 3067 cm1 (3.26 m) to NHþ 4 , we are not able to match this with our calculations. The ammonium NH and water OH stretching modes at that wavelength tend to couple unpredictably in the different clusters, yielding peaks that vary over a range of 200 cm1 or more.
þ Fig. 4.— Cluster structures of NHþ 4 , NH3, and NH3 as computed at the B3LYP/6-31+G level. [See the electronic edition of the Journal for a color version of this figure.]
PARK & WOON TABLE 2 Intensities and Scaled Computed Frequencies of Deformation Modes for NHþ 4 and Related Species in Various Clusters Deformation Mode 1
Deformation Mode 2
Deformation Mode 3
NHþ 4 -7H2O........................................ HCOO-NHþ 4 -5H2O ......................... HCOO-NHþ 4 -14H2O ....................... b OCN-NHþ 4 -8H2O .......................... þ OCN -NH4 -14H2Oc......................... b OCN-NHþ 4 -8H2O .......................... þ OCN -NH4 -14H2Oc......................... NH3-9H 2O........................................ NHþ 3 -9H2O........................................
1463 1426 1489 1471 1466 1485 1479 1188 1584
1.17 2.54 1.41 1.76 1.79 1.68 1.72 2.77 1.45
1486 1525 1504 1512 1501 1514 1500 1666 1587
2.05 1.18 1.37 1.50 1.37 1.43 1.57 0.33 0.27
1539 1535 1512 1519 1528 1515 1523 1674 ...
4.36 0.64 1.26 1.32 1.59 1.49 1.71 0.08 ...
a b c
In units of 1017 cm molecule1. OCN from HNCO ( N coordinated to NHþ 4 ). OCN from HOCN (O coordinated to NHþ 4 ).
Two other candidate ice species of interest are ammonia and its cation, NHþ 3 , which can be formed through photoionization. In prior work ( Woon 2004; Woon & Park 2004) we found that ionization energies of various molecules can be decreased by 1Y 4 eV in ice compared to the gas phase. This makes photoionization much more favorable in ices than in the gas phase, to the point that it may compete with photolysis when an ice mantle is exposed to ultraviolet radiation. Some molecules lose a proton during this process (such as methanol, where the hydrated CH3OH+ ion rearranges to yield the methoxy radical CH3O and H3O+), while in other cases the ion remains intact (such as benzene and polycyclic aromatic hydrocarbons). Ammonia falls into the latter category. The deformation modes of NH3 fall at 1188, 1666, and 1673 cm1, but the two higher frequency peaks are both quite weak. The NHþ 3 cation has just two deformation modes, with frequencies near 1585 cm1, which is not a good match to the observed feature.
4. CONCLUSIONS Quantum chemical density functional theory calculations were performed for formate and ammonium ions embedded in clusters of varying sizes. If more than three water molecules are present, HCOOH spontaneously transfers a proton to NH3. Large cluster calculations in which a counterion is present reproduce experimental bands of HCOO at 1349, 1381, and 1580 cm1. The computed intensities of these features agree well with experiment. We also found features associated with NH deformation modes 1 that of NHþ 4 that are consistent with a broad band at 1490 cm is observed experimentally.
The support of NASA Exobiology grant NAG5-13482 is gratefully acknowledged.
REFERENCES Allouche, A. 2005, J. Chem. Phys., 122, 234703 Hudgins, D. M., Sandford, S. A., Allamandola, L. J., & Tielens, A. G. G. M. Bahr, S., Borodin, A., Ho¨fft, O., Kempter, V., & Allouche, A. 2005, J. Chem. 1993, ApJS, 86, 713 Phys., 122, 234704 Hudson, R. L., Moore, M. H., & Gerakines, P. A. 2001, ApJ, 550, 1140 Becke, A. D. 1993, J. Chem. Phys., 98, 5648 Keane, J. V., Tielens, A. G. G. M., Boogert, A. C. A., Schutte, W., & Whittet, Bernstein, M. P., Sandford, S. A., & Allamandola, L. J. 2000, ApJ, 542, 894 D. C. B. 2001, A&A, 376, 254 Boudin, N., Schutte, W. A., & Greenberg, J. M. 1998, A&A, 331, 749 Kessler, M. F., et al. 1996, A&A, 315, L27 Clark, T., Chandrasekhar, J., Spitznagel, G. W., & Schleyer, P. v. R. 1983, Lee, C., Yang, W., & Parr, R. J. 1988, Phys. Rev. B, 37, 785 J. Comput. Chem., 4, 294 Mare´chal, Y. 1987, J. Chem. Phys., 87, 6344 Demyk, K., Dartois, E., d’Hendecourt, L., Jourdain de Muizon, M., Heras, Novozamsky, J. H., Schutte, W. A., & Keane, J. V. 2001, A&A, 379, 588 A. M., & Breitfellner, M. 1998, A&A, 339, 553 Palumbo, M. E., Pendleton, Y. Y., & Strazzulla, G. 2000, ApJ, 542, 890 Ehrenfreund, P., Boogert, A. C. A., Gerakines, P. A., Jansen, D. J., Schutte, Park, J.-Y., & Woon, D. E. 2004a, ApJ, 601, L63 W. A., Tielens, A. G. G. M., & van Dishoeck, E. F. 1996a, A&A, 315, L341 ———. 2004b, J. Phys. Chem. A, 108, 6589 Ehrenfreund, P., & Fraser, H. J. 2003, in Solid State Astrochemistry, ed. Sandford, S. A. 1996, in ASP Conf. Ser 97, Polarimetry of the Interstellar V. Pirronello, J. Krelowski, & G. Manico` ( Dordrecht: Kluwer), 317 Medium, ed. W. Roberge & D. C. B. Whittet (San Francisco: ASP), 29 Ehrenfreund, P., Gerakines, P. A., Schutte, W. A., van Hemert, M. C., & Schutte, W. A., & Greenberg, J. M. 1997, A&A, 317, L43 van Dishoeck, E. F. 1996b, A&A, 312, 263 Schutte, W. A., Greenberg, J. M., van Dishoeck, E. F., Tielens, A. G. G. M., Elsila, J., Allamandola, L. J., & Sandford, S. A. 1997, ApJ, 479, 818 Boogert, A. C. A., & Whittet, D. C. B. 1997, Ap&SS, 255, 61 Fraser, H. J., Bisschop, S. E., Pontoppidan, K. M., Tielens, A. G. G. M., & Schutte, W. A., & Khanna, R. K. 2003, A&A, 398, 1049 van Dishoeck, E. F. 2005, MNRAS, 356, 1283 Schutte, W. A., et al. 1999, A&A, 343, 966 Frisch, M. J., et al. 2003, Gaussian 03 Package ( Rev. C.02; Pittsburgh: Gaussian) Scott, A. P., & Radom, L. 1996, J. Phys. Chem., 100, 16502 Gibb, E. L., Whittet, D. C. B., Boogert, A. C. A., & Tielens, A. G. G. M. 2004, Shimanouchi, T. 2005, Vibrational Frequency Data, in NIST Chemistry WebApJS, 151, 35 Book, ed. P. J. Linstrom & W. G. Mallard ( NIST Stand. Ref. Database 69; Gibb, E. L., et al. 2000, ApJ, 536, 347 Gaithersburg: NIST), http://webbook.nist.gov Grim, R. J. A., & Greenberg, J. M. 1987, ApJ, 321, L91 Soifer, B. T., Puetter, R. C., Russell, R. W., Willner, S. P., Harvey, P. M., & Hagen, W., Tielens, A. G. G. M., & Greenberg, J. M. 1983, A&AS, 51, 389 Gillett, F. C. 1979, ApJ, 232, L53 Hariharan, P. C., & Pople, J. A. 1973, Theor. Chim. Acta, 28, 213 Werner, M. W., et al. 2004, ApJS, 154, 1 Herbst, E. 1993, in Dust and Chemistry in Astronomy, ed. T. J. Millar & D. A. Woon, D. E. 2004, Adv. Space Res., 33, 44 Williams ( Bristol: IOP), 183 Woon, D. E., & Park, J.-Y. 2004, ApJ, 607, 342