Matrix Infrared Spectroscopic and Computational Investigations of the

Feb 18, 2011 - 1913 dx.doi.org/10.1021/jp111592e |J. Phys. Chem. A 2011, 115, 1913- ..... for the relative energies of the states at the B3LYP level w...
1 downloads 0 Views 2MB Size
Download PDF
ARTICLE pubs.acs.org/JPCA

Matrix Infrared Spectroscopic and Computational Investigations of the Lanthanide-Methylene Complexes CH2LnF2 with Single Ln-C Bonds Xuefeng Wang,†,‡ Han-Gook Cho,†,§ Lester Andrews,*,† Mingyang Chen,^ David A. Dixon,*,^ Han-Shi Hu,# and Jun Li*,# †

Department of Chemistry, University of Virginia, Charlottesville, Virginia 22904-4319, United States Department of Chemistry, Tongji University, Shanghai, 200092 China § Department of Chemistry, University of Incheon, 177 Dohwa-dong, Nam-ku, Incheon, 402-749, South Korea ^ Department of Chemistry, The University of Alabama, Tuscaloosa, Alabama 35487-0336, United States # Department of Chemistry and Key Laboratory of Organic Optoelectronics and Molecular Engineering of Ministry of Education, Tsinghua University, Beijing 100084, China ‡

bS Supporting Information ABSTRACT: Laser-ablated lanthanide metal atoms were condensed with CH2F2 in excess argon at 6 K or neon at 4 K. New infrared absorption bands are assigned to the oxidative addition product methylene lanthanide difluorides on the basis of deuterium substitution and vibrational frequency calculations with density functional theory (DFT). Two dominant absorptions in the 500 cm-1 region are identified as lanthanide-fluoride stretching modes for this very strong infrared absorption. The predominantly lanthanidecarbon stretching modes follow a similar trend of increasing with metal size and have characteristic 30 cm-1 deuterium and 14 cm-1 13C isotopic shifts. The electronic structure calculations show that these CH2LnF2 complexes are not analogous to the simple transition and actinide metal methylidenes with metal-carbon double bonds that have been investigated previously, because the lanthanide metals (in the þ2 or þ3 oxidation state) do not appear to form a π-type bond with the CH2 group. The DFT and ab initio correlated molecular orbital theory calculations predict that these complexes exist as multiradicals, with a Ln-C σ bond and a single electron on C-2p weakly coupled with fx (x = 1 (Ce), 2 (Pr), 3(Nd), etc.) electrons in the adjacent Ln-4f orbitals. The Ln-C σ bond is composed of about 15% Ln-5d,6s and 85% C-sp2 hybrid orbital. The Ln orbital has predominantly 6s and 5d character with more d-character for early lanthanides and increasing amounts of s-character across the row. The Ln-F bonds are almost purely ionic. Accordingly, the argon-neon matrix shifts are large (13-16 cm-1) for the ionic Ln-F bond stretching modes and small (∼1 cm-1) for the more covalent Ln-C bond stretching modes.

’ INTRODUCTION There is an extensive literature of methylidene-metal complexes and their reactions, which are important in catalytic metathesis reactions. These transition metal complexes employed group four to six metals in the initial development and then evolved to also include the later transition metals.1-4 Although there is evidence for agostic Ln-H-C interactions in lanthanide neopentoxide complexes,5 there are mysteriously no lanthanide alkylidene complexes analogous to those extensively researched for the transition metals.6-8 Lanthanide N-heterocyclic carbene complexes appear to be simple Lewis base adducts with no significant metal-carbon double bond character7 as expected due to the very large singlet-triplet splitting in the carbene.9 The possibility of cerium-main group multiple bonding has been explored recently using density functional theory (DFT) calculations in combination with experiment for some ligand-supported complexes.8 We have r 2011 American Chemical Society

previously investigated reactions of laser-ablated transition metal atoms with methane and methyl halides and found simple insertion products (CH3-MX, X = H, F, Cl, Br) and methylidene products (CH2dMHX, X = H, F, Cl, Br). The methylidene structures are predicted to have agostic interactions, which lead to geometry distortions that vary with the metal and halogen substituent.10 This work has been extended to the actinide metals Th and U, where agostic interactions are also found in the actinide methylidene complexes.11-14 The reactions of groups 4 and 6 and Th metals with methylene halides give a symmetric C2v or Cs CH2dMX2 product without agostic distortion, but the analogous uranium species are predicted to be distorted based on DFT calculations.15-18 Here we report similar experiments and Received: December 6, 2010 Revised: January 17, 2011 Published: February 18, 2011 1913

dx.doi.org/10.1021/jp111592e | J. Phys. Chem. A 2011, 115, 1913–1921

The Journal of Physical Chemistry A

ARTICLE

DFT calculations for the reactions of lanthanides with CH2F2, which are shown to form methylene lanthanide difluorides with single lanthanide-carbon bonds and no evidence of agostic interactions leading to geometry distortions. Extensive calculations benchmarked by experiments on CH2CeXY (X,Y = H, F, Cl) using the DFT, CCSD(T), and CASPT2 methods have predicted, despite the fact that ubiquitous tetravalent Ce(IV) complexes are known, that cerium-methylene molecules do not form CedC double bonds. Instead, the CH2CeXY molecules had open-shell singlet ground states with a single Ce-C bond and two antiferromagnetically coupled single electrons forming H2C 3 - 3 CeXY biradicals.19

’ EXPERIMENTAL AND COMPUTATIONAL METHODS The laser-ablated lanthanide atoms were reacted with small molecules during condensation at 4-6 K. The experimental apparatus and procedure for these reactions have been described previously.10,20-22 The Nd:YAG laser fundamental (1064 nm, 10 Hz repetition rate with 10 ns pulse width) was focused onto a freshly cleaned lanthanide metal target (Johnson-Matthey) mounted on a rotating rod. The laser energy was varied from 1 to 5 mJ/pulse. Accordingly, laser-ablated lanthanide metal atoms were codeposited with 3-4 mmol of argon (Matheson, research) containing 0.5% methylene fluoride (DuPont) or CD2F2 at slightly lower concentrations (prepared by heating CD2I2 (MSD Isotopes) on HgF2 in a stainless steel finger (0.5 in. dia, 8 in. long) at 100 C for 2 h and distilling under vacuum from NaOH pellets to remove CO2)23 onto a CsI cryogenic window for 60 min. Several experiments were performed with 13CH2F2 (prepared by heating 13CH2Cl2 (Cambridge Isotope) on HgF2 in another stainless steel finger (0.5 in. dia, 6 in. long) at 150 C for 18 h, evacuating at 77 K to remove 12,13CO, and distilling under vacuum from NaOH pellets to remove 12,13CO2). Complementary neon matrix experiments were performed with six metals to investigate product matrix interactions. FTIR spectra were recorded at 0.5 cm-1 resolution on a Nicolet 750 FTIR instrument with 0.1 cm-1 accuracy using a HgCdTe range B detector. Matrix samples were annealed at different temperatures, and selected samples were subjected to broadband photolysis by a medium-pressure mercury arc street lamp (Philips, 175 W) with the outer globe removed. Following the example of transition and actinide metal CH2F2 activation products15-17 and extensive computational studies of metal oxides at Alabama,24 complementary electronic structure calculations, mostly at the DFT levels, were carried out using the Gaussian 03/09 and ADF 2006 programs.25,26 The geometries were optimized, and second energy derivatives were calculated to predict the vibrational frequencies of CH2FLnF and CH2LnF2 for all lanthanides and a wide range of spin states. The DFT calculations were performed with the B3LYP hybrid functional27 with the DFT optimized DZVP2 basis set on H, C, and F,28 and the Stuttgart segmented small core relativistic effective core potential (ECP) with the corresponding basis set on the lanthanide elements.29 The energies for the reaction CH2F2 þ Ln f CH2LnF2 for multiple spin states of CH2LnF2 were calculated, and the ground spin state for each lanthanide reaction product molecule was predicted. All open shell species, including singlet, were calculated at the spin unrestricted level. Additional ab initio electron correlation calculations were performed at the coupled cluster U/UCCSD(T) level30,31 with the aug-cc-pVDZ basis set32 on the C, H, and F and the Stuttgart ECP and basis set as given above for the lanthanide using the

Figure 1. Infrared spectra for the major product of the reaction of laserablated Ce atoms with methylene fluoride during cocondensation in excess argon at 6 K. (a) Ce codeposition with CH2F2, (b) after annealing to 20 K, (c) after >220 nm irradiation, (d) after annealing to 30 K, (e) Ce codeposition with CD2F2, (f) after annealing to 20 K, (g) after >240380 nm irradiation, (h) after annealing to 30 K.

Gaussian 09 program. Complete active space self-consistent field (CASSCF) calculations with the unpaired electrons on the Ln and one unpaired electron in the out-of-plane carbon 2p-orbital were performed for the high spin coupling and for the coupling with one electron on the Ln and the single electron on the C spin paired for some of the lanthanides. Here the CASSCF calculations have (n þ 1) electrons in (n þ 1) orbitals, where n is the number of singly occupied 4f orbital on the lanthanide metal. These calculations were performed with MOLPRO 2008.1 using the atomic basis sets described previously.33

’ RESULTS AND DISCUSSION Infrared spectra of the reaction products of lanthanide metals and methylene fluoride in solid argon and computational results to assist in their identification, vibrational assignment, and electronic structure are described. Six complementary neon matrix investigations were also performed. Transient species derived from precursor irradiation from the ablation plume such as CHF2 and CF2 were observed in every experiment.34-36 Experimental Infrared Spectra. Several argon matrix experiments were performed with laser-ablated Ce atoms using different laser energies, and the diagnostic metal-fluoride stretching region for the lower power investigation is shown in Figure 1. The most prominent new product absorptions, labeled with arrows, were observed at 504.8 and 491.0 cm-1 just below the precursor band at 532 cm-1 using CH2F2, and the CD2F2 reaction gave a single stronger band slightly shifted to 490.4 cm-1 also below the precursor band now at 525 cm-1. These new product absorptions, labeled with arrows, sharpen on annealing and increase slightly on ultraviolet irradiation. Annealing also produced two sharp, weak bands at 488.3 and 483.7 cm-1, which are in agreement with 488 and 484 cm-1 bands assigned to the strong, degenerate Ce-F stretching fundamental of CeF3.37 Two neon matrix experiments were also performed using different laser energies, and the stronger band was observed at 509.7 cm-1 with the weaker band masked by the tail of the precursor band at 528.6 cm-1. Annealing also produced weak 505.2 and 501.2 cm-1 satellite bands, which are in agreement 1914

dx.doi.org/10.1021/jp111592e |J. Phys. Chem. A 2011, 115, 1913–1921

The Journal of Physical Chemistry A

Figure 2. Infrared spectra for the major product of the reaction of laserablated Pr atoms with methylene fluoride during cocondensation in excess argon at 6 K. (a) Pr codeposition with CH2F2, (b) after annealing to 20 K, (c) after >220 nm irradiation, (d) after annealing to 30 K, (e) Pr codeposition with CD2F2, (f) after annealing to 20 K, (g) after >240380 nm irradiation, (h) after annealing to 30 K.

with the previous 505.3 and 501.8 cm-1 neon matrix counterparts assigned to CeF3. Several reactions of Pr with methylene fluoride produced weaker product absorptions at 519.7 and 511.8 cm-1, which increased about 30% on ultraviolet irradiation, and again only a single band was observed on deuterium substitution, slightly shifted to 511.3 cm-1, as shown in Figure 2. A weak PrF3 band increased on annealing at 458 cm-1.37-39 The reaction of Nd and CH2F2 gave two broader bands at 518.6 and 507.8 cm-1. The reaction of Sm yielded stronger absorptions at 510.4 and 456 cm-1 and started a trend of products with increasing frequencies going across the lanthanide series, as shown in the spectra for all of the nonradioactive lanthanide metals in Figure 3. Some of the strongest bands have a higher frequency partner, which are marked together in the figure. These pairs must be considered as candidates for the symmetric and antisymmetric Ln-F stretching modes. Table 1 summarizes the observed bands for the 13 stable lanthanide metals with deuterium counterparts listed in parentheses and 13C shifted frequencies in brackets for selected metals. The lanthanide metal reactivity in these experiments appeared to be comparable with the possible exception of slightly less reactivity for Pr and Dy, as suggested by the product yields in the infrared spectra shown in Figure 3. This is consistent with the substantial reaction exothermicities that are predicted as described below. Figure 4 compares the major Tb and CH2F2 reaction product band at 529.5 cm-1, just below CH2F2 at 532.0 cm-1, which shifted to 528.5 cm-1 with CD2F2 while the precursor band shifted more to 525 cm-1, the weak upper band at 543.7 cm-1 shifted to 542.3 cm-1, and the weak lower band at 467 cm-1 shifted to 436 cm-1. With 13 CH2F2 the precursor band shifted to 529.6 cm-1 and covered the major product band, but the weak upper band shifted to 543.0 cm-1 and the lower band to 453.7 cm-1. The reactions of Dy with CH2F2, 13CH2F2, and CD2F2 lead to isotopic shifts, which are small for the strongest bands but larger for the lowest bands at 472, 458, and 442 cm-1, respectively, as shown in Figure 5. The reactions of Dy were also performed in a solid neon matrix, and deuterium shifts were small for the upper bands at 559 cm-1 (558 cm-1) and 553 cm-1 (552 cm-1) but larger for the lower bands at 471 cm-1 (443 cm-1).

ARTICLE

Figure 3. Infrared spectra for the major lanthanide metal atom reaction products with CH2F2 at 0.5% in argon during codeposition in excess argon at 6 K for 60 min. (a) La, (b) Ce, (c) Pr, (d) Nd, (e) Sm, (f) Eu, (g) Gd, (h) Tb, (i) Dy, (j) Ho, (k) Er, (l) Tm, (m) Yb, and (n) Lu. Arrows denote CH2LnF2 product absorptions. The sharp, weak peak at 552.7 cm-1 is common to these and other metal experiments with CH2F2.

The reaction of Er in excess argon gave one strong band at 541.5 cm-1 together with a sharp common band at 553.0 cm-1 in the region where the weaker band is expected. However, the use of a neon matrix allowed the weaker Er-F stretching mode to be observed at 566.6 cm-1 owing to its blue shift above the 553.0 cm-1 common band described below while the stronger Er-F mode was observed at 557.4 cm-1 and the lower Er-C absorption was detected at 483 cm-1. The Er reaction with CD2F2 in argon revealed both bands at 553.5 and 540.5 cm-1 with the common product band deuterium shifted to 550.3 cm-1. Likewise for the Yb product bands, Figure 6 shows the bands from the reaction of Yb with CH2F2 at 557.8, 548.0, and 486 cm-1, which approximately doubled on ultraviolet irradiation, with the CD2F2 counterparts at 557.3, 546.8, and 457 cm-1, and the 13 CH2F2 product at 557.4, 547.8, and 471 cm-1. Corresponding neon matrix spectra for the Yb product absorptions are illustrated in Figure 7. The three CH2F2 product bands sharpened with matrix site splittings at 573.4, 571.1 cm-1; 564.6, 562.9 cm-1; and 487.7, 485.9 cm-1. These bands increased by 30% on >220 nm irradiation together with a 6% decrease in the adjacent CH2F2 product peak, and the CD2F2 product counterparts at 572.9, 571.0 cm-1; 563.8, 562.5 cm-1; and 459.1, 456.9 cm-1. These bands also increased by 30% on >220 nm irradiation with a slight decrease in the adjacent CD2F2 peak at 522.0 cm-1. The neon and argon matrix spectra exhibit the expected relationship,37 and the same ground-state product is expected to be trapped in both solids. The reaction of Lu was investigated in both argon and neon. The former matrix gave bands at 559.4, 554.7, and 494 cm-1, and the latter has weaker bands at 575.2, 570.2, and 493 cm-1 (see Figure 8). Because La can be considered as the beginning of the rare earth sequences and has similar chemical properties to the lanthanide metals, the La reaction was also investigated. The argon matrix spectrum at the bottom of Figure 3 shows a strong new band at 481.6 cm-1 with a weak partner at 499.7 cm-1. The sharp, weak 478.9 cm-1 band, which increased on annealing, is due to LaF3.37 1915

dx.doi.org/10.1021/jp111592e |J. Phys. Chem. A 2011, 115, 1913–1921

The Journal of Physical Chemistry A

ARTICLE

Table 1. Infrared Absorptions (cm-1) Observed for Lanthanide Metal Atom Reaction Products with Methylene Fluoride in Solid Argon at 6 K (Deuterium Substituted Values in Parentheses and 13C Frequencies in Braces for Selected Metals)a GS spin/ metal

Ln-F symmetric str

Ln-F antisymmetric str

Ln-C str

ΔErxnb

ES spinc

ΔEspind

La Ce

499.7, 520.3e [516 (517)] 504.8 [515 (515)]

481.6, 499.7e [504 (502)] 491.0, 509.7e (490.4) [501 (499)]

[432 (404)] [437 (409)]

-157.9 -155.0

2 3/1

0.4

Pr

519.7 [535f (543)]

511.8 (511.3) [521 (520)]

[446 (416)]

-122.0

2/4

0.2

Nd

518.6 [533 (533)]

507.8 [522 (520)]

[446 (418)]

-126.1

5/3

1.7

Pm

[546 (556)]

[529 (529)]

[456 (427)]

-105.5

4/6

0.2

Sm

[533 (537)]

510.4 [522 (520)]

456 [433 (405)]

-81.0

7/5

0.9

Eu

[525 (524)g, 523 (521)h]

512.6 [513 (519)g, 527 (523)h]

459 [362 (339)g, 416 (392)h]

-58.8

8/6

2.1

Gd

536.7 [534 (537)]

520.7 [528 (521)]

466 [415 (390)]

-98.7

9/7

0.7

Tb

543.7 (542.3){543.0} [560f(562){561}]

529.5 (528.5){n. o.} [559 (557){559}]

466.9 (436){453.7} [473 (440, 443i){459}]

-212.7

8/6

0.6

Dy

546.0 (544.6){545.8}

538.0 (537.8){537.5}

472.0 (442){457.6} [474

-162.9

7/5

0.8

[563 (564){562}]

[559 (557){559}]

(442){460}]

Ho

549.0 [565 (564)]

538.8 [557 (554)]

477 [478 (447)]

-129.3

4/6

2.3

Er

-, (553.5) [570 (569)]

541.5 (540.5) [566 (564)]

483 [483 (451)]

-106.5

3/5

5.4

Tm

558.3 [568 (568)]

547.0 [563 (560)]

486 [487 (445)]

-86.8

4/2

0.2

Yb

557.8, 572(557.3){557.4}

548.0, 564(546.8){547.8}

486, 487(457){471}

-68.2

1/3

0.5

Lu

[577 (577){577}] 559.2, 575.2 (559.1) [571 (571)]

[570 (567){570}] 554.7, 570.2 (553.5) [570 (566)]

[486 (455){471}] 494, 493 (467) [494 (462)]

-145.1

2

a

Calculated frequencies using B3LYP/DVZP2/Stuttgart in square brackets (D values in parentheses and 13C frequencies in braces) for the most stable spin state. Calculated reaction energies and ground state-excited state energy differences. b Energy of the reaction Ln þ CH2F2 f CH2LnF2 in kcal/mol. c Ground-state spin state/excited state spin state. d Energy splitting between the ground spin state and the excited spin state in kcal/mol. e Neon matrix frequencies given here in bold type and in the text. f CH2 wagging and LnF2 sym stretching are coupled. g Eu octet, Cs0 . h Eu sextet, Cs. i CH2 wagging and Ln-C stretching are coupled.

These bands sharpened upon annealing as shown in Figure 8. The corresponding neon matrix experiment produced these bands in lower yield, blue-shifted and split at 500.6, 499.7 cm-1; at 521.8, 520.3 cm-1; and at 497.2 cm-1. Annealing favors the higher frequency neon matrix site split absorption and ultimately decreases the 497.2 cm-1 LaF3 band.37 In contrast the CH2F2 band and the weak common peak red-shifted in neon. The fundamental frequency of the CeF molecule in the gas phase is approximately 550 cm-1, and CeF3 in solid argon has a very strong absorption at 488 cm-1.37,40 Thus, the 500 cm-1 region is appropriate for Ce-F bond vibrations in product molecules. The strongest absorption, the antisymmetric Ln-F stretching mode, increases from 458 cm-1 for PrF3 to 519 cm-1 for GdF3, to 545 cm-1 for TmF3, and to 552 cm-1 for LuF3 in solid argon.37-39,41,42 This is almost the same as the pattern for the major product absorptions summarized in Figure 3, increasing steadily from the La product at 481 cm-1 to the Lu product at 555 cm-1 with a small inflection for the Pr product at 512 cm-1. It is significant that the strongest SmF2, EuF2, and YbF2 difluoride absorptions at 435, 435, and 462 cm-1, respectively, are substantially lower in frequency than SmF3 and EuF3 at 508 and 511 cm-1, respectively.43 The near coincidence of the strongest product Ln-F and reported LnF3 stretching frequencies suggests that the present products are likely Ln(III) species. Note that 10 of the 13 Ln-F product absorptions (plus La products) are paired with a weaker counterpart as is appropriate for difluoride molecules. Some of our product bands are within a few cm-1 of the known trifluoride spectra but differ by 1, 2, 3, and 6 cm-1 in nine cases and by 54 cm-1 for the Pr products. However, PrF3 itself may be responsible for this discrepancy owing to vibronic perturbation with a low lying electronic state.39

The weaker bands between 456 and 495 cm-1 provide an important diagnostic vibrational mode. These absorptions shift by 27-31 cm-1 upon deuterium substitution. This shows clearly that the product molecules involve deuterium, which is reinforced by the small deuterium shifts observed for the Ln-F stretching modes and substantiated by our quantum chemistry calculations discussed below. In addition these absorptions shift 13-15 cm-1 upon 13C substitution. This is almost the amount (17 cm-1) calculated for a simple harmonic diatomic Tb-C oscillator at 467 cm-1. These modes can be described as Ln-(CH2) stretching modes with isotopic substitution giving the predicted mass effect. Thus, the 456-495 cm-1 bands are assigned to the important diagnostic Ln-C stretching modes, which are in excellent agreement with calculated values and their characteristic isotopic counterparts (Table 1). The results in summary show the formation of CH2LnF2; the theoretical calculations described below are consistent with this assignment. In addition, the low Ln-C stretching frequencies are consistent with our description of the Ln-C bond as a single bond rather than a more-or-less anticipated double bond, as in the case of H2Ce-CXY (X, Y = H, F, Cl).19 Calculation Results. DFT and ab initio calculations were performed for the methylene lanthanide difluorides. The DFT results are shown in Table 1 and the Supporting Information and include the electronic ground states of the products. The calculated exothermicities at the DFT level for the formation of the ground state of CH2LnF2 range from ∼-60 kcal/mol for Eu to as high as ∼-210 kcal/mol for Tb (Table 1). These calculations for the relative energies of the states at the B3LYP level without spin-orbit corrections are probably accurate to (2-3 kcal/mol for the early lanthanides where spin orbit corrections are not as 1916

dx.doi.org/10.1021/jp111592e |J. Phys. Chem. A 2011, 115, 1913–1921

The Journal of Physical Chemistry A

Figure 4. Infrared spectra for the major product of the reaction of laserablated Tb atoms with methylene fluoride during cocondensation in excess argon at 6 K. (a) Tb codeposition with CH2F2, (b) after annealing to 20 K, (c) after >290 nm irradiation, (d) after >220 nm irradiation, (e) after annealing to 30 K, (f) Tb codeposition with CD2F2, (g) after >290 nm irradiation, (h) after annealing to 30 K. (i) Tb codeposition with 13 CH2F2, (j) after annealing to 30 K, (k) after >220 nm irradiation, and (l) after annealing to 35 K.

ARTICLE

Figure 6. Infrared spectra for the major product of the reaction of laserablated Yb atoms with methylene fluoride during cocondensation in excess argon at 6 K. (a) Yb codeposition with CH2F2, (b) after annealing to 20 K, (c) after >220 nm irradiation, (d) after annealing to 30 K, (e) Yb codeposition with CD2F2, (f) after annealing to 20 K, (g) after >240380 nm irradiation, (h) after annealing to 30 K. (i) Yb codeposition with 13 CH2F2, (j) after >220 nm irradiation, (k) after another >220 nm irradiation, and (l) after annealing to 30 K.

Figure 5. Infrared spectra for the major product of the reaction of laserablated Dy atoms with methylene fluoride during cocondensation in excess argon at 6 K. (a) Dy codeposition with CH2F2, (b) after annealing to 20 K, (c) after >220 nm irradiation, (d) after annealing to 30 K, (e) Dy codeposition with CD2F2, (f) after annealing to 20 K, (g) after >240380 nm irradiation, (h) after annealing to 30 K. (i) Dy codeposition with 13 CH2F2, (j) after >290 nm irradiation, (k) after >220 nm irradiation.

Figure 7. Infrared spectra for the major product of the reaction of laserablated Yb atoms with methylene fluoride during cocondensation in excess neon at 4 K. (a) Yb codeposition with CH2F2, (b) after >220 nm irradiation, and (c) after annealing to 8 K. (d) Yb codeposition with CD2F2, (e) after >220 nm irradiation, and (f) after annealing to 8 K.

large. The alternative products H2FCLnF and HFCLnHF are at least 40 kcal/mol higher in energy than CH2LnF2, so that under thermodynamic control, CH2LnF2 should be the dominant product of the insertion reaction. The following trend in the reaction energies for the formation of the lowest energy product was found. The magnitude of the reaction energy decreases as the metal 4f orbitals are individually filled until 4f orbitals are halffilled for Ln = Eu. The reaction energy increases for Gd, which is

unique in the lanthanides with the s2df7 electron configuration. After Gd, the d orbital is not occupied in the ground state of the atom and the f orbitals start being doubly occupied. The maximum reaction energy is found for Tb and the reaction energies decrease as the f orbitals are doubly occupied until the f shell is fully filled for Yb. Although Lu ([Xe]4f145d6s2) is in the lanthanide series, its electron configuration is more like a transition metal, and Lu also does not follow the trend. 1917

dx.doi.org/10.1021/jp111592e |J. Phys. Chem. A 2011, 115, 1913–1921

The Journal of Physical Chemistry A

Figure 8. Comparison of the La and Lu reaction products with CH2F2 in excess neon and argon at 4-6 K. (a) La after codeposition in Ne, (b) after >290 nm irradiation, and (c) after annealing to 10 K. (d) La after codeposition in Ar, (e) after >220 nm irradiation, and (f) after annealing to 30 K. (g) Lu after codeposition in Ne, (h) after >220 nm irradiation, and (i) after annealing to 8 K. (j) Lu after codeposition in Ar, (e) after >220 nm irradiation, and (f) after annealing to 35 K. Neon spectra are amplified 2 .

The CCSD(T) calculations (Table S2, Supporting Information) are in general consistent with the spin state ordering given by the DFT calculations with the most notable difference being for Eu where the octet is lower in energy than the sextet for DFT by about 2 kcal/mol and the order is reversed at the CCSD(T) level with the sextet ∼2.5 kcal/mol more stable than the octet. The CCSD(T) and B3LYP reaction energies are qualitatively consistent but may differ by as much as 38 kcal/ mol, revealing the potential for significant errors in the B3LYP reaction energies for these lanthanide insertion reactions. Additional CASSCF calculations show that the high spin state is preferred over the state with the single electron on the out-ofplane C 2p orbital low-spin coupled to an unpaired electron on the lanthanide. The state splittings range from 0.2 kcal/mol for Yb to 3.6 kcal/mol for Tb. For Ce, the high spin state is preferred at the CASSCF level, and previous calculations at CASPT2 level showed the singlet state is slightly more stable than the triplet by about 1 kcal/mol.19 As discussed below, the best way to assign the observed spin state is to compare the calculated and experimental vibrational spectra, especially the splitting of the symmetric and antisymmetric LnF2 stretches. The ground state geometries are C2v if the molecule is planar, Cs if either the C or Ln is pyramidal, and C1 in a few cases. For Eu, one structure has the CH2 group rotated about the Ln-C bond to give a structure labeled as Cs0 to denote this rotation. Our results show that the rotational barrier about the Ln-C bond is small, on the order of a few kcal/mol, suggesting that the Ln-C bond is a single bond (see below). An important question that we were trying to address with the theoretical calculations is the nature of the Ln-C bond in terms of its single or double bond character, i.e., does the out of plane p orbital on the CH2 moiety couple with the f orbitals on the Ln metal atom to form a π bond with a significant π bond energy? Previous calculations of H2CCeXY (X = H, F, Cl) showed that the structures with a CedC double bond (r(CedC) ∼2.01 Å) is higher in energy by 9.4 kcal/mol at the CCSD(T) level of theory than the single bonded (r(Ce-C) ∼ 2.4 Å) diradical species with

ARTICLE

an unpaired electron on the Ce and another on the C.19 As shown in Table 1, most of the structures have two approximately degenerate states with spins differing by two, with the only exceptions being CH2LaF2 and CH2LuF2 which have only a single d electron unpaired and occupy the beginning and end of the lanthanide series. This arises because the out of plane orbital on the CH2 group is singly occupied and the metal forms a σ bond with the sp2 orbital remaining on C after the two C-H bonds are formed; thus, CH2LnF2 resembles a substituted methyl radical with the LnF2 as the substituent. The Ln can be considered to form two Ln-F bonds with the two electrons in the 6s orbital on the lanthanide if the Ln-F bonds are covalent or as is more appropriate, the Ln can be considered to be in the þ3 oxidation state with two ionic Ln-F bonds and one Ln-C bond. One of the Ln f orbitals (fz2) mixes with some Ln d orbtials to form the C-Ln bond leaving (n - 1) f-electrons in a high spin grouping on the Ln, which can couple high spin or low spin with the electron in the out of plane p on the CH2-Ln group. For La and Lu, a Ln d orbital forms the Ln-C bond, as there is no valence f orbital available. The coupling of the electrons on the Ln with the single electron on C is weak due to low overlap so two states, one high spin coupled, and one low spin coupled, which are essentially degenerate, are formed. For example, the calculated DFT spin density diagrams (Figure 9) show that one electron is localized in the out of plane p orbital on C and the other electron is localized in an f orbital on Yb in CH2YbF2 and in CH2CeF2 This is consistent with the molecular geometries and the low rotational barriers about the Ln-C bond and further shows that no additional Ln-C π-bond is formed. The out-ofplane carbon p orbital has only a very weak interaction with the Ln orbitals showing that no π bond is formed. We note that CH2LaF2 and CH2LuF2 are doublets with no electron spin on the metal site. This single electron interaction between the p orbital on the C with the Ln more closely resembles the dative π bonds formed in compounds like BH2NH2 and BH2PH2.44 In some cases, it is possible to optimize a structure with a short LndC bond that resembles a lanthanide-carbon double bond. Such structures correspond to excited states for these molecules. For Ce we were able to optimize an excited diamagnetic singlet state having paired electrons with a short Ln-C bond length only for CH2CeF2 with r(Ce-C) = 2.102 Å as compared to r(Ce-C) = 2.454 Å for the ground-state triplet, but this state is ∼17 kcal/mol higher in energy at the DFT level. This latter structure with the shorter double bond is observed for CH2ThH2.45 We performed additional CASSCF and CASPT2 calculations on the high-spin (S = 9) and low-spin (S = 7) states of CH2GdF2. The CASPT2 calculations show that the high-spin is more stable than the low-spin by some 3.6 kcal/mol. The calculated frequencies for the lowest energy spin state for most molecules at the DFT level support the assignments of the experimental infrared spectra. The calculated LnF2 symmetric and antisymmetric stretching frequencies are roughly 20 cm-1 greater than the experimental values. The symmetric stretch is ∼10-15 cm-1 higher than the antisymmetric stretching mode for La to Eu, ∼5-10 cm-1 larger for for Gd to Ho, and the splitting is less than 5 cm-1 for most of the remaining lanthanides from Er to Lu. Most of the calculated symmetric-antisymmetric splittings of the Ln-F stretch are slightly smaller than the experimental splittings, but the trend that this splitting decreases as the atomic number and the number of f electrons increases is found both experimentally and computationally. The calculated Ln-C stretching frequencies of the heavier lanthanides (Tb to Lu) 1918

dx.doi.org/10.1021/jp111592e |J. Phys. Chem. A 2011, 115, 1913–1921

The Journal of Physical Chemistry A

Figure 9. Spin density for the (a) triplet ground state CH2YbF2 and (b) the triplet state of CH2CeF2.

are within 5 cm-1 of the experimental values. For CH2GdF2, the calculated Gd-C stretch is too low by about 40 to 50 cm-1 as compared to experiment. This could arise because we do not have the correct mixing of the d and f orbitals. We predict the Ln-C stretch for the compounds containing the lighter lanthanides (La-Sm) to be lower than the heavier ones, which is consistent with the limited experimental observations, because the Ln-C stretching modes are very weak and difficult to observe at the low frequency end of the detector range. In the case of CH2YbF2, the comparison of calculated and observed frequencies provides a basis for preference of the ground state as triplet, despite the fact that our DFT calculation predicts that the singlet state is lower in energy by 0.5 kcal/mol. First, shifts in the the Yb-C stretching mode observed in solid argon at 486 cm-1, 13CH2YbF2 at 471 cm-1, and CD2YbF2 at 457 cm-1 are in excellent (2 cm-1) agreement with the calculated values for the triplet state. In contrast, the calculated singlet state Yb-C stretching mode is lower at 460 cm-1, which is not the relationship expected for calculated and observed frequencies.46,47 In addition, the computed relative intensities and the symmetric stretch falling below the antisymmetric Yb-F stretch for the singlet state (Table S1, Supporting Information) are not in agreement with the observations. The calculated triplet state frequencies and (intensities) are 577 cm-1 (66 km/mol) and 570 (194), which match the argon matrix values 557.8 cm-1 (1 relative absorbance unit) and 548.0 cm-1 (4 relative absorbance units) very well and the neon matrix values 571.1 cm-1 and 562.9 cm-1 even better! For CH2CeF2, the experimental value of the symmetric-antisymmetric mode splitting of the Ln-F stretches is 14 cm-1 with the symmetric mode higher. The calculated values for the triplet state have the symmetric stretch higher than the antisymmetric by 14 cm-1 just as found experimentally. In contrast the calculated spin unrestricted DFT values for the singlet state are reversed with the antisymmetric stretch 15 cm-1 higher than the symmetric stretch contrary to the experimental observation. Thus the DFT calculations of the IR spectra are consistent with the assignment of a triplet ground state. For comparison, CASPT2 calculations

ARTICLE

predict the antiferromagnetic singlet CH2CeF2 state to be 1 kcal/mol lower energy than that of the triplet,19 whereas our B3LYP calculations predict the triplet state to be 0.4 kcal/mol lower. At the CASPT2 level, the singlet state frequencies also have the symmetric state higher than the antisymmetric state. Unfortunately, there is little basis to select the spin state of the other Ln products based on comparison of calculated and observed frequencies, as the Ln-F stretching frequencies are nearly the same for each spin state. In the case of Sm and Eu, the Ln-C stretching modes observed suggest the S = 5 and 6 spin state products, respectively. As discussed above, there is a single electron on the carbon atom. This should behave like a typical substituted primary alkyl radical and be able to form an additional bond. As an example, we computationally reacted an H atom with CH2HoF2 to form CH3HoF2 which would have one less unpaired electron. The sextet state of CH2HoF2 yielded a quintet ground state for CH3HoF2 with the triplet and septet states 43 and 61 kcal/mol higher in energy, respectively. The reaction energy for the reaction 6 CH2HoF2 þ H f 5CH3HoF2 is -96 kcal/mol and is the negative of the C-H bond dissociation energy (BDE). This can be compared to the C-H BDE in the simplest alkane of ∼105 kcal/mol for CH4 and 99 kcal/mol for the C-H BDE in CH2F2.48 Bonding in the CH2LnF2 Molecules. In most cases, the Ln-C bond is localized on the C with about 85% on C and 15% on the Ln (Supporting Information). The C-Ln bond is formed by an sp2 hybrid orbital on C. The Ln component of the bond has predominantly s and d character with more d character earlier in the period and the amount of s character increasing across the period. The Ln-F bonds are almost purely ionic. The single electron on C is located in the out of plane p orbital remaining after the sp2 hybrid orbitals are formed. Some electron density from this singly occupied orbital delocalizes onto the Ln center. As an example, the isosurfaces and occupation numbers of the natural orbitals of F2GdCH2 with spin-multiplicities of 7 and 9 are shown in Figure 10. Clearly the Gd remains as an f7 center with one 4f-orbital weakly interacting with a C 2p orbital, leaving essentially two unpaired electrons located on adjacent Gd 4f and C 2p orbitals. In the early part of the period, the weak amount of dative overlap in the π space is predominantly d with some p and f. For Sm, there is additional f character. Consistent with the simple picture of the bonding, the argon-neon matrix shifts are large (13-16 cm-1) for the Ln-F stretching modes and small (1 cm-1) for the Ln-C stretching modes. This arises because the polar Ln-F bonds interact more strongly with the polarizable matrix host, and thus the stretching vibrations of ionic bonds sustain larger matrix shifts than those of covalent bonds. Such is the case for the LnF3 molecules.37 A similar situation has been observed for the ionic molecule (Liþ)(O2-) where the interionic stretching modes sustained large argon-neon matrix shifts and the intraionic mode (within the covalent O2- moiety) shifted only slightly.49

’ CONCLUSIONS The reactions of laser-ablated lanthanide metal atoms with CH2F2 in argon or neon matrixes led to the oxidative addition product CH2LnCF2. The structures are assigned by a combination of infrared spectral measurements and electronic structure calculations, mostly at the DFT level. The diagnostic infrared absorptions are the two strong Ln-F stretching modes in the 500 cm-1 region and the predominantly Ln-C stretching mode with a characteristic 30 cm-1 deuterium and 14 cm-1 13C 1919

dx.doi.org/10.1021/jp111592e |J. Phys. Chem. A 2011, 115, 1913–1921

The Journal of Physical Chemistry A

ARTICLE

Virginia), and Grant No. DE-FG02-03ER15481 (University of Alabama). D.A.D. thanks the Robert Ramsay Fund at the University of Alabama for partial support. J. L. is supported by NKBRSF (2011CB932400) and NSFC (20933003, 11079006) of China. The calculations were partially performed on Linux computer clusters at The University of Alabama and at the Alabama Supercomputing Center and by using supercomputers at the Computer Network Information Center, Chinese Academy of Sciences, Tsinghua National Laboratory for Information Science and Technology, and the Shanghai Supercomputing Center.

’ REFERENCES

Figure 10. The isosurfaces and occupation numbers of the natural orbitals of F2GdCH2 with spin-multiplicities of 7 and 9.

isotopic shift. The symmetric stretch is higher than the antisymmetric stretching mode. The CH2LnF2 complexes are not the same as the transition and actinide metal methylidenes with metal-carbon double bonds that have been investigated previously. The Ln metals do not form a π-type bond with the CH2 group to form a metallocarbene type species. Rather, a Ln-C σ bond is formed and the two electrons that would form the π bond remain unpaired as a single electron on C 2p weakly coupled with a single electron in the Ln 4f orbitals. This leads to biradical centers in which the unpaired electron on C is high spin or low spin coupled to an f electron on the Ln if such electrons are present. For doublet CH2LaF2 and CH2LuF2, where there are no open shell electrons on the Ln, the single electron is localized on the C. Thus, the best description of these CH2LnF2 species is as a substituted methyl radical with an LnF2 substituent bonded to the C by an Ln-C single σ bond. The LnF2 substituent may or may not have unpaired electron density.

’ ASSOCIATED CONTENT

bS Supporting Information. Tables of calculated geometries, frequencies, energies, spin densities, and NBO analysis, and a figure of orbital component spin density plots for triplet CH2YbF2. This material is available free of charge via the Internet at http://pubs.acs.org. ’ AUTHOR INFORMATION Corresponding Author

*E-mail: (L.A.) [email protected]; (D.A.D.) [email protected]; (J.L.) [email protected].

’ ACKNOWLEDGMENT We gratefully acknowledge financial support from DOE Office of Science, Basic Energy Sciences Grant No. DE-SC0001034 and NCSA computing Grant No. CHE07-0004N (University of

(1) Schrock, R. R. Chem. Rev. 2002, 102, 145. (2) Herndon, J. W. Coord. Chem. Rev. 2005, 249, 999. (3) (a) Herndon, J. W. Coord. Chem. Rev. 2006, 250, 1889. (b) Herndon, J. W. Coord. Chem. Rev. 2007, 251, 1158. (4) (a) Herndon, J. W. Coord. Chem. Rev. 2009, 253, 86. (b) Herndon, J. W. Coord. Chem. Rev. 2009, 253, 1517. (5) Barnhart, D. M.; Clark, D. L.; Gordon, J. C.; Huffman, J. C.; Watkin, J. G.; Zwick, B. D. J. Am. Chem. Soc. 1993, 115, 8461. (6) (a) Pool, J. A.; Scott, B. L.; Kiplinger, J. L. J. Am. Chem. Soc. 2005, 127, 1338. (b) Burns, C. J. Science 2005, 309, 1823. (7) (a) Arnold, P. L.; Liddle, S. T. Chem. Commun 2006, 3959. (b) Giesbrecht, G. R.; Gordon, J. C. Dalton Trans. 2004, 2387. (8) Clark, D. L.; Gordon, J. C.; Hay, P. C.; Poli, R. Organometallics 2005, 24, 5747. (9) (a) Dixon, D. A.; Arduengo, A. J., III. J. Phys. Chem. 1991, 95, 4180. (b) Dixon, D. A.; Arduengo, A. J., III J. Phys. Chem. A 2006, 110, 1968. (10) (a) Andrews, L.; Cho, H.-G. Organometallics 2006, 25, 4040 and references therein. (b) Cho, H.-G.; Andrews, L. Organometallics 2007, 26, 633. (c) Cho, H.-G.; Andrews, L. J. Phys. Chem. A 2007, 111, 2480. (11) Andrews, L.; Cho, H.-G. J. Phys. Chem. A 2005, 109, 6796. (12) Lyon, J. T.; Andrews, L. Inorg. Chem. 2005, 44, 8610. (13) Lyon, J. T.; Andrews, L. Inorg. Chem. 2006, 45, 1847. (14) Lyon, J. T.; Andrews, L.; Malmqvist, P. A.; Roos, B. O.; Wang, T.; Bursten, B. E. Inorg. Chem. 2007, 46, 4917. (15) Lyon, J. T.; Andrews Inorg. Chem. 2007, 46, 4799. (16) Lyon, J. T.; Cho, H.-G.; Andrews, L. Organometallics 2007, 26, 6373. (17) Lyon, J. T.; Andrews, L. Eur. J. Inorg. Chem. 2008, 1047. (18) Lyon, J. T.; Andrews, L.; Hu, H.-S.; Li, J. Inorg. Chem. 2008, 47, 1435. (19) Hu, H.-S.; Li, J.; Wang, X.; Andrews, L. Unpublished work, 2010. (20) Andrews, L.; Citra, A. Chem. Rev. 2002, 102, 885 and references therein. (21) Andrews, L. Chem. Soc. Rev. 2004, 33, 123 and references therein. (22) Willson, S. P.; Andrews, L. J. Phys. Chem. A 2000, 104, 1640. (23) Andrews, L.; Willner, H.; Prochaska, F. T. J. Fluorine Chem. 1979, 13, 273. (24) (a) Li, S.; Dixon, D. A. J. Phys. Chem. A 2007, 111, 11908. (b) Li, S.; Dixon, D. A. J. Phys. Chem. A 2007, 111, 11093. (c) Zhai, H.-J.; Li, S.; Dixon, D. A.; Wang, L.-S. J. Am. Chem. Soc. 2008, 130, 5167. (d) Li, S.; Dixon, D. A. J. Phys. Chem. A 2008, 112, 6646. (e) Li, S.; Hennigan, J. M.; Dixon, D. A.; Peterson, K. A. J. Phys. Chem. A 2009, 113, 7861. (f) Li, S.; Zhai, H.-J.; Wang, L.-S.; Dixon, D. A. J. Phys. Chem. A 2009, 113, 11273. (g) Li, S.; Dixon, D. A. J. Phys. Chem. A 2010, 114, 2665. (h) Li, S.; Peterson, K. A.; Dixon, D. A. J. Chem. Phys. 2008, 128, 154301. (i) Craciun, R.; Picone, D.; Long, R. T.; Li, S.; Dixon, D. A.; Peterson, K. A.; Christe, K. O. Inorg. Chem. 2010, 49, 1056. (j) Wang, T.-S.; NavarreteLopez, A. M.; Li, S.; Dixon, D. A.; Gole, J. L. J. Phys. Chem. A 2010, 114, 7561. 1920

dx.doi.org/10.1021/jp111592e |J. Phys. Chem. A 2011, 115, 1913–1921

The Journal of Physical Chemistry A (25) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, Jr., J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, N. J.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, € Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; S.; Daniels, A. D.; Farkas, O.; Fox, D. J. Gaussian 09, Revision A.2, Gaussian, Inc., Wallingford, CT, 2009. (26) ADF 2006.01, SCM Theoretical Chemistry, Vrije Universiteit, Amsterdam, The Netherlands. http://www.scm.com. (a) te Velde, G.; Bickelhaupt, F. M.; van Gisbergen, S. J. A.; Fonseca Guerra, C.; Baerends, E. J.; Snijders, J. G.; Ziegler, T. J. Comput. Chem. 2001, 22, 931–967. (b) Fonseca Guerra, C.; Snijders, J. G.; te Velde, G.; Baerends, E. J. Theor. Chem. Acc. 1998, 99, 391–403. (27) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Lee, C.; Yang, Y.; Parr, R. G. Phys. Rev. B 1988, 37, 785. (28) Godbout, N.; Salahub, D. R.; Andzelm, J.; Wimmer, E. Can. J. Chem. 1992, 70, 560. (29) (a) Dolg, M.; Stoll, H.; Preuss, H. J. Chem. Phys. 1989, 90, 1730. (b) Cao, X.; Dolg, M. J. Chem. Phys. 2001, 115, 7348. (30) (a) Purvis, G. D., III; Barlett, R. J. J. Chem. Phys. 1982, 76, 1910. (b) Raghavachari, K.; Trucks, G. W.; Pople, J. A.; Head-Gordon, M. Chem. Phys. Lett. 1989, 157, 479. (c) Watts, J. D.; Gauss, J.; Bartlett, R. J. J. Chem. Phys. 1993, 98, 8718. (d) Barlett, R. J.; Musial, M. Rev. Mod. Phys. 2007, 79, 291. (31) (a) Noga, J.; Bartlett, R. J. J. Chem. Phys. 1987, 86, 7041. (b) Scuseria, G. E.; Schaeffer, H. F., III. Chem. Phys. Lett. 1988, 146, 23. (32) (a) Dunning, T. H., Jr. J. Chem. Phys. 1989, 90, 1007. (b) Kendall, R. A.; Dunning, T. H., Jr.; Harrison, R. J. J. Phys. Chem. 1992, 96, 6796. (c) Dunning, T. H., Jr. J. Phys. Chem. A 2000, 104, 9062. (33) Werner, H.-J.; Knowles, P. J.; Lindh, R.; Manby, F. R.; Sch€utz, M.; Celani, P.; Korona, T.; Rauhut, G.; Amos, R. D.; Bernhardsson, A.; Berning, A.; Cooper, D. L.; Deegan, M. J. O.; Dobbyn, A. J.; Eckert, F.; Hampel C.; Hetzer, G.; Lloyd, A. W.; McNicholas, S. J.; Meyer W.; Mura, M. E.; Nicklass, A.; Palmieri, P.; Pitzer, R.; Schumann, U.; Stoll, H.; Stone, A. J.; Tarroni, R.; Thorsteinsson, T. MOLPRO, a package of ab initio programs. See http://www.molpro.net. (34) Milligan, D. E.; Jacox, M. E. J. Chem. Phys. 1968, 48, 2265. (35) Carver, T. G.; Andrews, L. J. Chem. Phys. 1969, 50, 5100. Andrews, L.; Prochaska, F. T. J. Chem. Phys. 1979, 70, 4714. (36) Jacox, M. E. J. Phys. Chem. Ref. Data 1994, Monograph 3; 1998, 27 (2), 115; 2003, 32, 1. (37) Hastie, J. W.; Hauge, R. H.; Margrave, J. L. J. Less-Common Met. 1971, 23, 359; 1975, 39, 309. (38) Wesley, R. D.; DeKock, C. W. J. Chem. Phys. 1971, 55, 3866. (39) Lesiecki, M.; Nibler, J. W.; DeKock, C. W. J. Chem. Phys. 1972, 57, 13852. (40) Bloch, J. C.; McCarthy, M. C.; Field, R. W.; Kaledin, L. A. J. Mol. Spectrosc. 1996, 177, 251. (41) (a) Bencze, L.; Feltrin, A.; Nunziante-Cesaro, S.; Popovic, A. Rapid Commun. Mass Spectrom. 1996, 10, 1248. (b) Lanza, G.; Minichino, C. J. Phys. Chem. A 2005, 109, 2127. (42) Kovacsa, A.; Konings, R. J. M. J. Phys. Chem. Ref. Data 2004, 33, 337 and references therein. (43) (a) Hastie, J. W.; Hauge, R. H.; Margrave, J. L. High Temp. Sci. 1971, 3, 56. (b) DeKock, C. W.; Wesley, R. D.; Radtke, D. D. High Temp. Sci. 1972, 4, 41. (44) Grant, D.; Dixon, D. A. J. Phys. Chem. A 2006, 110, 12955. (45) Andrews, L.; Cho, H.-G. J. Phys. Chem. A 2005, 109, 6796. (46) Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502.

ARTICLE

(47) Andersson, M. P.; Uvdal, P. L. J. Phys. Chem. A 2005, 109, 2937. (48) Luo, Y.-R. Comprehensive Handbook of Chemical Bond Energies; CRC Press: Boca Raton, FL, 2007. (49) Wang, X.; Andrews, L. Mol. Phys. 2009, 107, 739.

1921

dx.doi.org/10.1021/jp111592e |J. Phys. Chem. A 2011, 115, 1913–1921