Article pubs.acs.org/JPCA
Charge-Transfer Complexes and Photochemistry of Ozone with Ferrocene and n‑Butylferrocene: A UV−vis Matrix-Isolation Study Laura F. Pinelo, Roger W. Kugel, and Bruce S. Ault* Department of Chemistry, University of Cincinnati, P.O. Box 210172, Cincinnati, Ohio 45221-0172, United States S Supporting Information *
ABSTRACT: The reactions of ozone with ferrocene (cp2Fe) and with n-butylferrocene (n-butyl cp2Fe) were studied using matrix isolation, UV−vis spectroscopy, and theoretical calculations. The codeposition of cp2Fe with O3 and of nbutyl cp2Fe with O3 into an argon matrix led to the production of 1:1 charge-transfer complexes with absorptions at 765 and 815 nm, respectively. These absorptions contribute to the green matrix color observed upon initial deposition. The charge-transfer complexes underwent photochemical reactions upon irradiation with red light (λ ≥ 600 nm). Theoretical UV−vis spectra of the charge-transfer complexes and photochemical products were calculated using TD-DFT at the B3LYP/6-311G++(d,2p) level of theory. The calculated UV−vis spectra were in good agreement with the experimental results. MO analysis of these long-wavelength transitions showed them to be n→ π* on the ozone subunit in the complex and indicated that the formation of the charge-transfer complex between ozone and cp2Fe or n-butyl cp2Fe affects how readily the π* orbital on O3 is populated when red light (λ ≥ 600 nm) is absorbed. 1:1 complexes of cp2Fe and n-butyl cp2Fe with O2 were also observed experimentally and calculated theoretically. These results support and enhance previous infrared studies of the mechanism of photooxidation of ferrocene by ozone, a reaction that has considerable significance for the formation of iron oxide thin films for a range of applications.
■
or between O(3P) and n-butyl cp2Fe, resulting in the formation of the photochemical products observed in the infrared spectra. Several studies have investigated the generation of atomic oxygen by photodissociation upon red irradiation of small charge-transfer complexes involving O3. For example, red irradiation of the charge-transfer complex O3:Br2 led to the identification of new bromine oxide compounds.14 Also, evidence of the long wavelength photolysis of ozone was reported by Bahou et al.,15 who observed oxygen isotope scrambling when 18O3 isolated in a matrix of 16O2 at 10 K was irradiated with light from a dye laser at 693 nm. In addition, the photochemistry of O3 in the visible range, specifically that of the Chappuis band (420 ≤ λ ≤ 700 nm), has been of particular interest due to its application in atmospheric O3 monitoring.16 In the present work, the reactions of ferrocene and nbutylferrocene with ozone were investigated using UV−visible spectroscopy to identify the species responsible for the green matrices and subsequent photochemistry. In addition, because some O2 is always present in O3 samples,17 O2 complexes with cp2Fe and n-butyl cp2Fe were also investigated. This study was carried using the low-temperature matrix-isolation technique combined with UV−vis spectroscopy. Theoretical calculations
INTRODUCTION
Metal oxide thin films have many important technological applications, including flat-panel displays, energy-saving window coatings, and solar cells.1−3 Thin films containing iron oxide have been of particular interest and have been used in photochemical cells, batteries, and gas sensors.4−10 These films are often formed through the processes of chemical vapor deposition (CVD) or atomic layer deposition (ALD). Despite the importance of metal oxide thin films, little is known about the chemical reactions of ozone with the organometallic compounds used to make these films.11 Recently, infrared argon matrix isolation studies12,13 of the reactions of O3 with ferrocene (cp2Fe) and n-butylferrocene (nbutyl cp2Fe) reported product formation as a result of a photochemical reaction initiated with red light (λ ≥ 600 nm). Prior to irradiation with red light, vivid green matrices were formed. These green matrices were proposed to be the results of charge-transfer complexes forming between O3 and cp2Fe and n-butyl cp2Fe. This conclusion was based primarily on the green color of the matrix, suggesting strong red and blue absorptions in the visible spectrum and a slightly red-shifted O3 ν3 infrared absorption characteristic of perturbed O3. The proposed reaction mechanism involved the photochemical dissociation of O3 to ground-state dioxygen and ground-state atomic oxygen, followed by reaction between O(3P) and cp2Fe © XXXX American Chemical Society
Received: July 28, 2015 Revised: September 15, 2015
A
DOI: 10.1021/acs.jpca.5b07292 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A were used to locate energy minima on the ground-state and excited-state surfaces, to determine transition energies and oscillator strengths, and to investigate the molecular orbitals involved in the observed transitions. The results of the present study support the mechanistic conclusions of the previous infrared studies and provide a conclusive interpretation of the green color of the matrices observed therein.
■
EXPERIMENTAL SECTION
Ozone was produced by the Tesla coil discharge of O2 (Wright Brothers) while being condensed with liquid nitrogen. Excess O2 was pumped off prior to the ozone being warmed to room temperature. For experiments with O2, molecular oxygen was used without further purification. The O3 (or O2) was then transferred to a mixing can, where it was diluted with argon (Ar) (Wright Brothers, 99.998%) in an Ar/O3 mole ratio of ∼90 (and an Ar/O2 mole ratio of ∼60). Stock ferrocene (Eastman Kodak) and n-butylferrocene (Strem Chemicals, 99%) were purified by sublimation. Either cp2Fe or n-butyl cp2Fe was placed in a sample holder and attached by an UltraTorr tee to the deposition line. The sample holder was heated with a sand bath to a temperature of ∼45 °C, generating a small vapor pressure (∼50 mTorr). Vaporized cp2Fe or n-butyl cp2Fe was mixed with argon flowing from a reservoir to the cold cell. All experiments were conducted using a standard matrixisolation system and the twin-jet configuration as previously described.12,18 The matrix was deposited in the dark onto a CsI cold window over 60 min, and deposition was stopped prior to the collection of the first spectrum. The baseline-corrected spectra were collected with a Varian Cary 4000 UV−vis spectrophotometer from 200 to 900 nm with a spectral bandwidth of 2.00 nm, an average scan time of 0.10 s/point, and a data interval of 1.00 nm. The strong intensity of the Hartley band19 of O3 at 254 nm precluded the observation of products at wavelengths ≤300 nm. Thus, all scans of codeposited samples were collected from 300 to 900 nm. To minimize the exposure of the matrix to the spectrometer source prior to irradiation, the outer windows were covered during deposition. The matrix was then irradiated using the output from an incandescent light bulb filtered through a red glass cutoff filter (λ ≥ 600 nm) in 15 min increments. Spectra were collected after a total of 15, 30, and 45 min of irradiation.
■
Figure 1. Dark TJ-deposition spectra in green, the reactants subtracted from them (O3 in blue and cp2Fe or n-butyl cp2Fe in red), and resulting difference spectra in black. (a) Ar/cp2Fe/O3 and (b) Ar/nbutyl cp2Fe/O3.
Figure 2. Calculated structures and the ground-state energy relative to the ground-state parent species of two charge-transfer complexes of O3 with cp2Fe (a,b).
COMPUTATIONAL DETAILS
■
Theoretical calculations were performed with the Gaussian 09 suite of programs20 using density functional theory (DFT) with the Becke, three-parameter, Lee−Yang−Parr (B3LYP) functional and the 6-311++G(d, 2p) basis set. Optimized geometries and energies for all ground-state species were obtained at this level of theory. A comparison of the charge distribution from natural bond orbital (NBO) analyses of the optimized species allowed for the identification of the proposed complexes as charge-transfer complexes.21 The energies of the excited states and the absorption spectra of the previously optimized ground-state species were calculated using timedependent density functional theory (TD-DFT).22−26 All calculations were performed at the Ohio Supercomputer Center.
RESULTS AND DISCUSSION Prior to codeposition experiments, UV−vis spectra were recorded of samples of O3, ferrocene (cp2Fe), and n-butyl ferrocene (n-butyl cp2Fe) each alone in solid argon at 14 K and shown in Figure 1a,b. The resultant spectra were consistent with literature spectra for these compounds.27−32 Then, a sample of Ar/O3 was codeposited with a sample of Ar/cp 2Fe under similar conditions using twin jet deposition in the dark. This resulted in the observation of a green matrix, as has been seen and previously reported in the infrared studies.12,13 While some changes relative to the blank were visibly apparent, difference spectra (also shown in Figure 1a) clearly revealed a new broad B
DOI: 10.1021/acs.jpca.5b07292 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A
In the previous O3/cp2Fe infrared study, no distinct infrared absorptions were detected upon initial twin jet dark deposition despite the formation of green matrices.12 The only new feature noted was a slightly perturbed antisymmetric stretching mode of O3, which shifted a few wavenumbers to the red of parent O3. This shift is characteristic of the formation of a molecular complex between O3 and cp2Fe, and it suggested that the green color the matrix might be due to a molecular complex. Consequently, theoretical calculations were carried out using DFT methods to determine whether a molecular complex was a stable species and, if it was, what the electronic spectrum of the complex would be. Many possible configurations are possible for a molecular complex, and a range of these was explored. Energy minima were identified for several complexes, including one in which the O3 resides over one of the cp rings (hereafter cp2Fe-O3A) and a second in which the O3 spans the two cp rings with the Fe in the middle (hereafter cp2Fe-O3B). These are shown in Figure 2. Despite their apparently different configurations, these two were calculated to have quite similar electronic properties and spectra. It is very likely that there are additional stable configurations possible for a 1:1 complex of O3 with cp2Fe. These probably will have similar electronic properties and spectra. Thus, the 765 nm band is assigned to a 1:1 molecular complex of O3 with Ar/cp2Fe. At the same time, because the ozone concentration is relatively high in these experiments, we cannot rule out a minor contribution from higher order (e.g., 2:1) complexes. Natural bond orbital (NBO) analysis of the calculated ground states of cp2Fe-O3A and cp2Fe-O3B showed an increase in charge density on the terminal oxygens of O3 and a decrease in charge density on ferrocene, with a net transfer of about 0.1 units of charge from ferrocene to ozone in the ground state of the complex. Thus, the description of these complexes as being charge transfer in nature is appropriate. Increased O−O bond lengths in complexed ozone relative to free ozone were also noted in the calculations and shown in Table 1. Electronic states of these two complexes were then calculated using TD-DFT methods. Because a spin flip is unlikely during the formation of these complexes, only singlet excited states were considered for O3 and the O3 complexes, while only triplet states were calculated for O2 and the O2 complexes. cp2Fe-O3A was calculated to have an electronic transition at 804 nm with an oscillator strength of 0.0280, while cp2Fe-O3B was calculated to have a similar transition at 821 nm with an oscillator strength of 0.0010, as shown in Table 2 and Figure 3.
Table 1. Calculated Oxygen−Oxygen Bond Lengths of O3 (O1O2O3)a uncomplexed O3 O3
(Å)
O1O2 O2O3 charge-transfer complexes
Fc-O3A
O1O2 O2O3 O1O2 O2O3 O1O2 O2O3 O1O2 O2O3
Fc-O3B nBuFc-O3A nBuFc-O3B a
1.256 1.256 (Å) 1.263 1.264 1.268 1.268 1.264 1.268 1.269 1.269
Computed at the B3LYP/6-311++G(d,2p) level of theory.
absorption centered at 765 nm with a second band centered at 361 nm and a shoulder at 313 nm. Because all of the samples of O3 contain at least some O2 due to slow decomposition of O3 in the vacuum line prior to deposition, twin jet codeposition experiments were carried out with samples of Ar/O2 and Ar/cp2Fe to identify possible absorptions due to the presence of O2. The spectrum of this matrix showed a strong absorption centered at 366 nm. Comparing the results of the O2 and O3 experiments, we can attribute the broad 765 nm band to the presence of O3 and cp2Fe, while the absorption around 361 nm with a 313 nm shoulder is likely due to the presence of O2 and cp2Fe. The initial experiment was repeated multiple times, varying the concentrations of O3 and of cp2Fe. Similar results were obtained throughout. A similar set of experiments was conducted involving the twin jet dark codeposition of samples of Ar/O3 and Ar/n-butyl cp2Fe to mirror the parallel infrared studies. The results for this system were comparable to the O3/cp2Fe system with some red shifting, with a broad absorption centered at 816 nm, a second product peak at 382 nm, and a shoulder at 328 nm as shown in Figure 1b. Then, twin jet codeposition experiments were carried out with samples of Ar/O2 and Ar/n-butyl cp2Fe. The spectra of these samples showed a strong absorption centered at 388 nm. Similar to the O3/cp2Fe system, the long wavelength 816 nm band is associated with the presence of O3 and n-butyl cp2Fe, while the 382 nm band with the 328 nm shoulder is associated with the presence of O2 and n-butyl cp2Fe.
Table 2. Calculated Absorption Bands and the Corresponding Excitation Energies of O3, cp2Fe-O3 Charge-Transfer Complexes, and n-Butyl cp2Fe-O3 Charge-Transfer Complexesa transition O3 cp2Fe-O3A cp2Fe-O3B n-butyl cp2Fe-O3A n-butyl cp2Fe-O3B
a
n n n n n n n n n n
(I) → π* (II) → π* (I) → π* (II) → π* (I) → π* (II) → π* (I) → π* (II) → π* (I) → π* (II) → π*
absorption (nm)
excitation energy (kcal/mol)
oscillator strength
602 536 804
47.5 53.3 35.6
