of 7

Transient Negative Species in Supercritical Carbon Dioxide:  Electronic Spectra and Reactions of CO 2 Anion Clusters †

All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Transient Negative Species in Supercritical Carbon Dioxide:  Electronic Spectra and Reactions of CO 2 Anion Clusters †
  Transient Negative Species in Supercritical Carbon Dioxide: Electronic Spectra andReactions of CO 2  Anion Clusters † Kenji Takahashi,* ,‡,§ Sadashi Sawamura, ‡ Nada M. Dimitrijevic, § David M. Bartels, § andCharles D. Jonah* ,§  Argonne National Laboratory, 9700 South Cass A V  enue, Argonne, Illinois 60439, and Di V  ision of Quantum Energy Engineering, Hokkaido Uni V  ersity, Sapporo 060-8628, Japan Recei V  ed: June 19, 2001; In Final Form: October 2, 2001 Transient absorption spectra following ionization of supercritical CO 2  have been investigated using the pulseradiolysis technique. Absorption spectra measured from 400 to 800 nm suggest that at least two transientspecies absorb. We have previously reported that one species is (CO 2 ) 2 + . In the near UV region, we observeda transient species of which the lifetime and reactivity are different from the dimer cation. We assign thisspecies to a dimer anion, (CO 2 ) 2 - , or an anion - molecule complex, (CO 2 - )(CO 2 )  x . Comparison with thephotobleaching of CO 2  anion clusters in solid rare gas matrixes and their reactivity with H 2  and O 2  confirmthe assignment. Theoretical calculations, in which solvation is taken into account, are consistent with theseassignments. It is well-established that the adiabatic electron affinity of CO 2  is negative, but the adiabaticelectron affinity of CO 2  dimer has been calculated to be 0.89 eV for  D 2 d   symmetry (CO 2 ) 2 - in the gas phase.The calculations predict that CO 2 - in a model continuum solvent is stable to autodetachment. 1. Introduction There have been extensive efforts to understand the charac-teristics of supercritical fluids as solvents. 1,2 Energy-transfer andelectron-transfer reactions have also been used as a probe toelucidate the solvent characteristics. 3,4 It has been a challengeto utilize CO 2  in reaction schemes because it is difficult toactivate CO 2 . 5 One potential route to activate CO 2  involvesreduction of CO 2 . 6,7 This could be achieved by an electrontransfer to CO 2  to yield the corresponding anion radical. Forthis reason, we wanted to explore the existence of the anionand its reactivity.There have been extensive studies of ion - molecule reactionsin gas-phase CO 2 . It is well-established that the primary positiveion in the radiolysis of CO 2  gas is CO 2 + and that CO 2 + formsclusters rapidly with CO 2 . 8 - 10 The atomic oxygen radical anion,O - , which is produced by dissociative electron attachment toCO 2 , is known to be the primary negative ion in the high-energyradiolysis. 11 - 17 To dissociate CO 2  to CO  +  O - , about 4.0 eVof energy is required (calculated from the bonding energy(O - CO)  )  5.453 eV, and the electron affinity(O)  )  1.461 eV). 18 O - will react rapidly with CO 2  to form CO 3 - . The reactionsthat electrons with less energy than 4 eV undergo are not yetknown. While several studies of negative species in the gasphase exist, there are few reports of the observation of CO 2 anions from the ionization of CO 2  in the gas phase. Even in aliquid and high-density CO 2 , some thought that CO 2 - wasunstable and thus not observable. 19,20 In a mobility experimentby Jacobsen and Freeman, electron attachment to a molecularcluster of CO 2  was assumed. 21 When the density of CO 2  isgreater than 14 × 10 25 molecules/m 3 in the coexistence vapor,the negative and positive charges have the same mobility, so itis expected that the electrons are permanently attached to CO 2 .CO 2 - has also been observed in a low-pressure glow discharge. 22 Recently, it has been reported that CO 2 - can be formed bydouble electron transfer to CO 2 + in the gas phase. 23 However,to the best of our knowledge, until the recent photolyticmeasurements by Shkrob and Sauer, there is no other clearevidence that CO 2  anion is formed in condensed or gas-phaseCO 2 . 24 The most widely accepted value for the adiabatic electronaffinity of CO 2  in the gas phase is  - 0.6  (  0.2 eV. 25 The abinitio calculations also predict a negative electron affinity. 26 - 29 The absence of CO 2 - in a low-pressure environment arises fromthe poor Franck  - Condon overlap between the ground states of the neutral CO 2  and its anion. Ground-state CO 2 - is bent withan O d C d O angle of about 135 ° , whereas the neutral ground-state CO 2  is linear. Linear CO 2 - , formed by vertical electronattachment, will be vibrationally excited by 2 - 3 eV. Thisspecies will either autodetach or dissociate in a few femtosec-onds. Collisional stabilization of this ion cannot occur in low-pressure conditions.There have been many studies on carbon dioxide cluster anionformation in molecular beams or van der Waals clusters. 30 - 36 For example, Klots and Compton 31 demonstrated anion clusterformation in a CO 2  molecular jet, where the dimer anion isformed, presumably by evaporative electron attachment. Theion-signal intensity as a function of incident electron energyhas shown a maximum around 3 eV. It is believed that themolecular cluster binds an excess electron differently than themonomer. 37 Although there are still open questions in theformation mechanism of CO 2  anion clusters, 18,38 - 40 it is gener-ally concluded that a single CO 2  molecule cannot bind an excesselectron, whereas a weakly bound CO 2  cluster is able to bindan electron. Several calculations for CO 2 -dimer anion have beenmade, and three stable isomers with  D 2 d  ,  D 2 h , and  C  s  structureare predicted. 41 - 43 * To whom correspondence should be addressed. E-mail for K.T.:kenji@pleiades.qe.eng.hokudai.ac.jp. Fax for K.T.: + 81-11-706-6675. E-mailfor C.D.J.: CDJonah@anl.gov. Fax for C.D.J.: 630-252-4993. † Work was performed under the auspices of the Office of Basic EnergySciences, Division of Chemical Science, US-DOE under contract numberW-31-109-ENG-38. § Argonne National Laboratory. ‡ Hokkaido University. 108  J. Phys. Chem. A  2002,  106,  108 - 11410.1021/jp012340s CCC: $22.00 © 2002 American Chemical SocietyPublished on Web 12/05/2001  According to these experimental facts and theoretical predic-tions, stable CO 2  anion or anion clusters could be formed incondensed or supercritical CO 2 . We have recently studied thereduction of   p -benzoquinone in supercritical CO 2  using pulseradiolysis. 44 That work clearly demonstrated that an electrondonor has been produced by ionization of supercritical CO 2 .We tentatively assigned the electron donor to the CO 2  clusteranion; however, we did not observe any spectra that could beattributed to the anion. This is partly because in the solution astrong absorption by  p -benzoquinone and its anion overlaps theregion where the absorption of the CO 2  anion cluster might beexpected. In addition, the absorption of the anion might be quitesmall.In recent experiments using rare gas matrixes, stable CO 2 monomer anion and its dimer have been observed. 43,45,46 In thesestudies, the infrared absorption assigned to (CO 2 ) 2 - wasdecreased by irradiation with visible light. This suggests thatan absorption band of (CO 2 ) 2 - in rare gas solid is located inthe visible region. These results led us to reexamine spectraproduced in supercritical CO 2 .In this paper, we will discuss absorption spectra of transientspecies that we assign to the CO 2  anion clusters, which wereproduced by electron beam irradiation in supercritical CO 2 . Wealso examined the reactions of the anion cluster with H 2  andO 2 . To determine whether the CO 2  anion will autoionize, abinitio calculations were performed. We believe that these arethe first measurements of absorption spectra of CO 2  anionclusters in supercritical CO 2 .Very shortly after our absorption measurements were per-formed, Shkrob and Sauer 24 studied the mobility of an anion insupercritical CO 2 . Electrons were created in a supercriticalsolution by photoionizing benzene or anthracene. To get thearomatics into the supercritical solution conveniently in theappropriate quantities, the aromatics were dissolved in a smallamount of hexane and injected into the system. The concentra-tions of the aromatic and the carrier hexane were sufficientlysmall that no effect would be expected on the reaction kinetics.They reported evidence that the resulting electron attaches toCO 2  to form the solvent anion. The photodetachment spectrumobserved is reported as similar to photoelectron spectra of the(CO 2 ) n - cluster ( n  )  6 - 9) and consistent with our results forabsorption spectrum of the anion clusters. These results can bea bridge between radiolysis and photolysis and between electronsin neat CO 2  and CO 2  solutions. It can also advance ourunderstanding of the formation mechanism of the CO 2  anioncluster in supercritical CO 2  and CO 2  solution. 2. Experimental Section The CO 2  was supercritical fluid chromatography (SFC) gradefrom Scott Specialty Gases. Impurities in this gas reported byScott are H 2 O ( < 3 ppm), O 2  ( < 2 ppm), CO ( < 5 ppm), H 2  (5ppm), CH 4  (2 ppm), N 2  (50 ppm), and Ar (5 ppm). It was usedas received. The H 2  (Aldrich, 99.99 + %) and O 2  (AGA, 99.9%)were also used as received.Experiments were performed in a stainless steel high-pressurecell. Two suprasil windows (1-cm thick, 3-cm diameter) weremounted to the cell using Teflon O-rings and passed theanalyzing light and electron beam. The optical path length is 5cm, and the effective diameter is 1.2 cm (an F/#  )  4). Theoptical densities shown in the figures were not divided by theoptical path length. All experiments were done at a constanttemperature of 40.1  (  0.1  ° C (a reduced temperature of 1.03 ° C) unless otherwise stated. The temperature was controlled andmonitored using a temperature controller (Omega, model CN1001 RTD). The pressure in the cell was adjusted using anHPLC pump (JASCO, model PU-980) and monitored with adigital pressure meter (Cole-Parmer, model 7350-38). Measure-ments have been carried out at 104 bar, at which the density is0.65 g/cm 3 , and at 53 bar, at which the density is 0.13 g/cm 3 .The experimental arrangements for supercritical fluids weresimilar to those described previously. 44 Pulse radiolysis experiments were performed using theArgonne 20 MeV linear accelerator with electron pulses of 4ns. For optical detection, interference filters (bandwidth 40 nm)and a silicon photodiode (EG&G FND100) were used. Thisdetection method was chosen over our monochromator/photo-multiplier tube system because a better signal-to-noise (S/N)ratio can be obtained. Because the accumulation of radiolyticproducts prevents extensive signal averaging, a good S/N ratiois quite important. Most of the signals presented here were theaverage of two single-shot signals. The electron beam and theanalyzing light were collinear. This arrangement was chosento increase the optical path length. However, the electron pulseirradiation to the 10-mm-thick optical windows mounted on thehigh-pressure cell caused a small absorption. This absorptionintensity was about 0.003 at 400 nm, and the half-lifetime wasabout 200 ns. The window absorption intensities depend onabsorbed dose, which is sensitive to the linac beam conditions.Hence, we measured absorption of the optical window byelectron beam irradiation before starting the experiments everytime, and the absorption signal was corrected by subtractingthe signal from the window absorption.Radiolytic products, mainly, carbon monoxide, ozone, andoxygen, accumulate after the irradiation of several pulses andinfluence the measurements. To prevent this, the sample wasfreshly prepared after each irradiation. In some experiments,the measurements were performed under flowing conditions,in which the pressure was controlled by a backpressure regulator(TESCOM 26 - 1700). However, it was difficult to control thetemperature at 40.1  ° C and a pressure of 104 bar. Hence, thismethod was applied only for a measurement at lower temper-ature and pressure. 3. Results and Discussion In this section, we will first describe the spectra that oneobserves in pure CO 2 . We then discuss measurements using H 2 ,which reacts with positive ions, to determine the spectrum thatmight be associated with an anion. We compare the observedspectrum with results from photobleaching experiments of matrix-isolated anions. Experiments in the presence of oxygengive further information on the species that are observed. Abinitio calculations were used to estimate the energetics of thedifferent species; these results were used to confirm theplausibility of the suggested reactions. 3.1. Transient Spectra in Pure CO 2 .  We have alreadyreported that an absorption spectrum can be observed in pureCO 2  within about 200 ns after the electron pulse. 47 The transientabsorption spectra are shown in Figure 1. The absorption spectralshape is broad and the maximum is located around 700 nm.We previously attributed the observed absorption spectrum to(CO 2 ) 2 + by comparing with the photodestruction cross section 48 of (CO 2 ) 2 + and by considering the reactivity of this species withseveral cation scavengers. Although the absorption spectrumcorresponds relatively well to the photodestruction cross section,there is a discrepancy, especially in the blue region. Hence, wethought it possible that another species also was absorbing.Confirming this hypothesis is the shoulder or growth inabsorption from 450 to 400 nm that can be seen in Figure 1.Negative Species in Supercritical Carbon Dioxide  J. Phys. Chem. A, Vol. 106, No. 1, 2002  109  The 400-nm absorption decays with different kinetics than doesthe 700-nm absorption.Figure 2 shows a very clear difference in the kinetics at lowertemperature (35  ° C). Note that the measurements in this figurewere done at a pressure of 53 bar (density  )  0.13 g/cm 3 ), sothe diffusion-controlled reaction of the cation, which is observedat 700 nm, is much faster than the data in Figure 1 (104 bar,0.65 g/cm 3 ). The absorption signal at 700 nm decays faster thanthe signal at 400 nm. This result clearly demonstrates that thetransient species around 700 nm is different from that around400 nm. We suggest that the spectrum at 400 nm is predomi-nantly due to (CO 2 ) 2 - or (CO 2 - )(CO 2 )  x . To establish thisassignment, we isolated this spectrum by the addition of compounds that react with the cation. We then compared theresulting spectrum with what is known about the anion spectrumfrom the literature. 3.2. Reactions and Spectra in H 2  /CO 2  Mixture.  Reactions. We have tried to isolate the spectrum of the species that absorbsin the near UV region from the overlapping absorption of the(CO 2 ) 2 + . Using a specific cation scavenger is one way to achievethis. The scavenger and its reacted product should have noabsorption in the wavelength region of interest. There are manyorganic molecules that can be used as cation scavengers, butthese molecules and their cations have strong absorptions inthe near UV region. The hydrogen molecule does not sufferfrom this problem.In the gas phase, the reaction of CO 2 + with H 2  has been wellstudied. CO 2 + reacts at close to collision-limited rates as 9,49,50 For (CO 2 ) 2 + , we can expect a similar hydrogen-atom abstractionreaction. 9,51 We can hope that the reaction between H 2  and theanion will be much slower. It is known that the reaction of CO 3 - with H 2  is quite slow in the gas phase ( k   <  5  ×  10 6 M - 1 s - 152 ); however, there is no report for the reaction of H 2  with CO 2 anion cluster. Hence, this technique may allow us to extractthe anion spectrum.Figure 3 shows transient spectra measured in a H 2  /CO 2 solution with 160 mM of H 2 . The spectrum immediately afterthe electron pulse is the same as that measured in pure CO 2 .The spectrum after the (CO 2 ) 2 + decay ( t  > 50 ns) shows a broadabsorption from 650 nm to the UV region and suggests at leastone species that is not the cation dimer. As we will discussbelow, this spectrum is consistent with photobleaching of infrared spectra of (CO 2 ) 2 - . Spectra.  We have been unable to find direct data on the nearUV spectrum of (CO 2 ) 2 - . However, there is indirect evidencethat comes from the photobleaching of the infrared spectra of (CO 2 ) 2 - in various matrixes. In one study, 43 mercury-arcphotolysis with a 470-nm cutoff filter (essentially 546 and 578nm mercury emission) decreased the (CO 2 ) 2 - infrared bands inan argon matrix about 50%. Irradiation through a 380-nm cutoff filter decreased the infrared bands about 90%, while photolysiswith the full arc (240 - 580 nm) totally removed the (CO 2 ) 2 - infrared bands.Another confirmation arises from the results of Thompsonand Jacox. 45 They measured vibrational spectra of CO 2 + , C 2 O 4 + ,CO 2 - , and (CO 2 ) 2 - (which were formed by Penning ionizationof a Ne/CO 2  mixture) and observed three different types of photobleaching. The infrared absorption bands for CO 2 - werediminished by photodetachment after tungsten-lamp irradiationthrough a 695-nm cutoff. This is interesting because in water Figure 1.  Transient absorption spectra observed at 20, 80, and 150 nsafter the electron pulse in pure CO 2  at a temperature of 40.1  ° C andpressure of 104 bar (density  )  0.6 g/cm 3 ). Figure 2.  Comparison of decay kinetics at 700 nm with those at 400nm. The temperature is 35  ° C, and the pressure is 53 bar (density of 0.13 g/cm 3 ). Figure 3.  Transient spectra observed at four different times in H 2  / CO 2  mixture. CO 2 + + H 2 f  HCO 2 + + H  k  ) 8.4 × 10 11 M - 1 s - 1 110  J. Phys. Chem. A, Vol. 106, No. 1, 2002  Takahashi et al.  the absorption spectrum of CO 2 - is located in the UV region.The IR absorption of the molecular complex (arising from theweak interaction of CO 2 - with CO 2 ) increases on irradiationwith light that passes through a 780-nm cutoff filter anddecreases with shorter wavelength irradiation. They explainedthis photobleaching by assuming that the photodetachment of CO 2 - occurs by irradiation through a 780 nm cutoff filter andthe resulting electrons may be captured by pairs of adjacentCO 2  molecules to form a CO 2  anion - molecule complex, whichin turn photodetaches at a higher photon energy. Anotherprominent feature in the infrared, which was assigned to the(CO 2 ) 2 - species, increased after successive tungsten-lampirradiation through 780-, 695-, and 630-nm cutoff filters andwas destroyed by irradiation with a mercury-arc lamp through420-nm cutoff filter.Because we expect no significant medium effect from anargon or neon matrix on the absorption spectrum, both experi-ments suggest that the absorption spectrum of the CO 2 - anioncluster is in the visible. We expect the spectra in sc-CO 2  to besimilar. Possible Role of CO 3 - .  One possible intermediate that couldappear in the spectrum is CO 3 - , a species that is important inthe gas phase. This species is formed in the gas phase throughthe reaction O - +  CO 2 . It is known that O - reacts with H 2 with a reaction rate of 4.4  ×  10 10 M - 1 s - 1 in the gas phase. 53 Because O - is a precursor for CO 3 - , if the spectrum observedin the near UV region were CO 3 - , the initial absorption intensitywould be decreased by adding H 2  into CO 2 . The reportedreaction rate of O - with CO 2  in the gas phase is 1.6  ×  10 11 M - 1 s - 1 . 54 Using these reaction rates and the concentration of H 2  and CO 2 , a reaction probability of O - with H 2  in CO 2  underthe highest H 2  concentration is calculated to be about 0.002.(This assumes that the ratio of the reaction rates of O - with thetwo species is similar in sc-CO 2  and the gas phase.) Therefore,we did not expect H 2  to scavenge O - under the presentexperimental conditions. 3.3. Reactions in O 2  /CO 2  Mixture.  We have already reportedthe reaction rate of (CO 2 ) 2 + with O 2  and found that the reactionrate is 5  ×  10 10 M - 1 s - 1 in 104 bar of supercritical CO 2 . 44 Inthe gas phase, a rate of 9 × 10 10 M - 1 s - 1 was reported for thereaction. Because of this relatively fast reaction rate of (CO 2 ) 2 + with O 2 , we might hope that the spectrum near the UV regioncould be extracted by eliminating the absorption overlap by thedimer cation. One difference from the H 2  /CO 2  mixture is thatO 2  will act also as an anion scavenger, which means that thedecay kinetics in the near UV might become faster. A seconddifference is the possible formation of CO 4 - .In Figure 4, the transient spectra and the decay kinetics at700 and 400 nm are shown. The concentration of O 2  isapproximately 5 mM, and under this condition, the decay at700 nm is much faster than that at 400 nm. This situationprovides us another way to extract the near-UV spectrumwithout interference from the (CO 2 ) 2 + absorption. The observedspectra are quite similar to the spectra measured in H 2  /CO 2 mixture, suggesting that the same species are present.In the presence of O 2 , one might form CO 4 - through thereaction of CO 2 - with O 2 . The photodestruction of CO 4 - insolid neon has been previously reported. Vestal and Mauclaire 55 found small a ion yield in the photodestruction of CO 4 - between600 and 305 nm. However, Moseley and co-workers  56,57 failedto reproduce the ion signal reported by Vestal and Mauclaireand concluded that the photodestruction cross section of CO 4 - is small over the entire 840 to 350 nm range. Jacox andThompson 58 found that infrared absorption bands of CO 4 - insolid neon could be photobleached by irradiation with a 260-nm cutoff filter, suggesting that absorption of CO 4 - is locatedat a shorter wavelength region than the absorption band of (CO 2 ) 2 - .Some further evidence that the observed spectrum is not CO 3 - has been obtained by examining the reaction rate with O 2 . Fromthe decay kinetics at 400 nm as a function of the O 2  concentra-tion, the second-order reaction rate has been estimated as 2 × 10 9 M - 1 s - 1 . This is similar to but slower than the rate measuredby Shkrob and Sauer of 8  ×  10 9 M - 1 s - 1 . 24 By comparing areported reaction rate for CO 3 - with O 2  in the gas phase 59 ( k  < 3.6 × 10 6 M - 1 s - 1 ), we conclude that the observed rate constantis not that for the reaction of O 2  with CO 3 - . Thus, the absorptionspectrum that we measured in the near UV region is expectedto arise from (CO 2 ) 2 - (or multimer) and not CO 4 - or CO 3 - .So far, we have used (CO 2 ) 2 - to describe the dimer anion.Ab initio calculations predict three different structures for thedimer anion, namely, the symmetrical  D 2 d   and  D 2 h  structuresand the asymmetrical  C  s  structure. 43,60 For the  D 2 d   and  D 2 h structures, the excess charge is equally distributed on the twoCO 2  molecules, while in the  C  s  asymmetric structure the excesscharge is localized in a bending CO 2  molecule. One wouldexpect that these isomers, which have different charge distribu-tions, would show different electronic absorption spectra.Further, a study by Zhou and Andrew 43 concluded that the1665.5 cm - 1 band observed on annealing in solid neon is a vander Waals complex of linear CO 2  and bent CO 2 - and(CO 2 )  x (CO 2 - ) and not the  C  s  anion structure characterized bycalculation. However, in the photobleaching of those anionclusters in matrixes, there is no clear difference between thosedimer anions and a complex because the photobleaching hasbeen done using wide-band cutoff filters. Hence, it is possiblethat the visible absorption spectra that we measured couldinclude contributions from CO 2 -anion clusters with differentstructures. This could be a reason that the spectra measured arebroad and asymmetric.In experiments in which photoionization was used to generatecharged species in sc-CO 2  and conductivity was used to observe Figure 4.  Transient spectra observed in O 2  /CO 2  mixture. The insetshows decay kinetics at 400 and 700 nm. Negative Species in Supercritical Carbon Dioxide  J. Phys. Chem. A, Vol. 106, No. 1, 2002  111  the chemistry, Shkrob and Sauer observed a high conductivityspecies that they assigned to a CO 2 -anion. 24 They found thatthey could photobleach this high conductivity species to generatean even more mobile conduction band electron. The actionspectrum for this photoprocess is very similar to that observedfor the blue - near-UV species that we have assigned to the dimeranion. 3.4. Calculation.  There have been many ab initio calculationsreported for CO 2 - . 26 - 29 Recently, extensive calculations havebeen carried out by Gutsev et al. 29 using the infinite-ordercoupled-cluster method (CCSD) and density functional theory(DFT). The calculated electron affinity of monomeric CO 2 depends on the model and basis set adopted and ranges between - 0.48 and  - 0.69 eV. (DFT calculation with the B3LYPfunctional and 6-311 + G(2d) basis sets predicts the highestelectron affinity ( - 0.48), and the lowest value ( - 0.69) wasobtained by CCSD and 6-311 + G(3df) basis sets.) These canbe compared to the adiabatic electron affinity of monomericCO 2 , which has been experimentally determined as  - 0.6  ( 0.2. 25 These calculations have been done for the gas phase.Because we already know that we cannot neglect solvation whenwe consider the stability of CO 2  anion, we performed calcula-tions including the effects of solvation. In this section, we wishto investigate the energetics of the monomer and dimer anionand the CO 3 - and see whether these energetics are consistentwith the mechanisms suggested above.Geometry optimizations for the CO 2  monomer and dimeranion were performed using the Gaussian 94 program. 61 TheB3LYP functional and 6-311 + G* basis sets for C and O atomswere used. After the geometry optimization of the ground state,the potential energy surfaces of CO 2 - were calculated. For thedimer anion, three different anion structures are known, namely,the symmetrical  D 2 d   and  D 2 h  anion and the  C  s  asymmetricalmolecule - anion complex. 41 - 43 We adopted the self-consistent isodensity polarized continuummodel (SCI-PCM) to consider solvation effects. 62 In this method,the solvent is modeled as a continuum of uniform dielectricconstant. The effects of solvation are folded into the iterativeSCF computations for the full coupling between the cavity andthe electron density. It is well-known that in supercritical fluidsthere can be considerable clustering of the solvent moleculesnear a solvent, 63,64 so the uniform dielectric constant may be apoor approximation. The value one should choose for thedielectric constant is not obvious. The bulk value of 1.5 for thedielectric constant is certainly an underestimate. For this reason,we report calculations using 2.0. We have also done thecalculations for 1.5, and the conclusions are qualitatively thesame.Figure 5 shows the potential energy surface of CO 2  and CO 2 anion as a function of O d C d O angle. The zero of energy wasselected so that the ground-state CO 2  is 0 eV. Also includedare results for solvent effects using the SCI-PCM model. Thebond lengths were optimized at each angle.In the gas phase, the energy of CO 2 - in its ground state (bent)is 0.37 eV above the ground state of linear neutral CO 2 . Thelowest point of crossing between the energy surfaces of CO 2 - and CO 2  occurs at a bond angle of about 152 ° . CO 2 - can beformed if its vibrational excitation energy exceeds approximately0.58 eV. However, the activation energy for autodetachment is0.21 eV. This low activation energy for the autodetachment isthe reason that the lifetime of CO 2 - is quite short and CO 2 - isunstable in the gas phase.On the other hand, in a dielectric continuum solvent, theenergy of CO 2 - in its ground state is 0.76 eV below the groundstate of neutral CO 2 . Because a system that has a zero dipolemoment will not exhibit a solvent effect for the SCI-PCM model,the calculations for neutral CO 2  are the same in the solvent andthe gas phase. The lowest point of crossing between the surfacesoccurs at bond angle of 171 ° , significantly closer to the linearO d C d O angle than that calculated for the gas phase. Anotherimportant difference from the gas-phase result is that theactivation energy for autodetachment is 0.89 eV. This meansthat the solvated CO 2 - is stable to autodetachment rather thanmetastable. These calculation results are consistent with ex-perimental observation of CO 2 - in nonpolar organic liquid 65 and supercritical ethane 66 and explain why CO 2 - is stable insuch media. There is no reason that we cannot expect similarstabilization of CO 2 - in supercritical CO 2 .The adiabatic electron affinity, EA ad , for the dimer can becalculated within the Born - Oppenheimer approximation aswhere  ZPE   is zero point vibrational energy and can be estimatedwithin the harmonic approximation. The energy of the neutralCO 2  dimer at the B3LYP/6-311 + G* level is calculated to be - 377.271 006 hartree, and the zero point energy is 0.023 602hartree, where the neutral dimer structure is the  slipped parallel geometry. 67,68 It is known that the  D 2 d   C 2 O 4 - is more stablethan the  C  s  and  D 2 h  forms, so we only calculated for the  D 2 d  symmetry. From its total energy ( - 377.303 144 928) and zeropoint energy (0.023 127), the electron affinity of   D 2 d   C 2 O 4 - iscalculated to be 0.89 eV. The experimentally determined valueof 0.8 eV has been reported by Quitevis and Herschbach, 40 whomeasured the charge-transfer reaction from alkali atoms to theweakly bound CO 2  cluster. Surprisingly, the calculated andmeasured values are identical. Therefore, one can expect thateven in the gas phase, CO 2  dimer can capture an electron andform a stable dimer anion. We expect that the dimer anion willbe more stable in the supercritical fluid, much like the monomeranion calculated above.We can put some experimental limits on the electron affinityfor the dimer in a sc-CO 2  solution. We previously studied the Figure 5.  Potential energy surfaces of CO 2 , CO 2 - in the gas phaseand CO 2 - in a solvent as a function of the OCO angle. The energiesare scaled so that the ground state of CO 2  is 0 eV. EA ad )  E  total (neutral) + ZPE(neutral) -  E  total (anion) - ZPE(anion) 112  J. Phys. Chem. A, Vol. 106, No. 1, 2002  Takahashi et al.
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks