The Excited State in Atmospheric Chemistry George Marston Department of Chemistry, University of Reading, Whiteknights, PO Box 224, Reading RG6 6AD, UK 1 Introduction In recent years, the chemistry of the Earth's atmosphere has pro- vided a focus for much scientific activity. Interest has been stimu- lated by concerns about the environmental consequences of Man's activities in general, and phenomena such as Global Warming, Acid Rain and the Antarctic Ozone Hole in particular. Of course, any attempt to understand the causes of these effects necessarily requires a detailed knowledge of the chemistry of the unpolluted atmosphere, and the result is that significant advances in our knowl- edge of atmospheric chemistry have been made in the last two decades. To a reasonable extent, it is possible to discuss the chem- istry of the atmosphere in terms of the behaviour of its components in thermal equilibrium with their surroundings; i.e.for a given chemical species, relative populations of a particular energy level can be calculated using the Boltzmann equation: where nl and n,, are the concentrations of the ith energy level and ground energy level, g, is the degeneracy of the ith level, E, is its energy, k is the Boltzmann constant and T is the temperature. However, just as the atmosphere is not in chemical equilibrium, there are many examples of excited atoms, molecules, radicals and ions that are present in the atmosphere in concentrations far greater than would be expected on the basis of equation (I).It is towards the chemistry of these excited states that this review addresses itself. Much of the discussion will centre on vibrationally and electroni- cally excited species; at typical atmospheric temperatures, the energy required to access these states is, as a general rule, too great to be achieved in thermal processes. For translational energy and rotational energy, the separations of the energy levels are small compared to kT and at thermal equilibrium, a range of levels is excited. However, we shall see that circumstances exist where translational and rotational temperatures of a particular species may be very much greater than the thermodynamic temperature of the surroundings. Ultimately, the energy for excitation within the atmosphere comes from the Sun, either in the form of solar radiation or charged particles (the solar wind).Excitation mechanisms include direct George Marston wlas born in Newcastle upon Tyne, England in 1961. He studied chemistry at the University of Oxford where he received his BA (1984) and, under the superi*ision of Richard Wayne, his DPhil (1987). He then spent two years at NASAIGoddurd Space Flight Center with Lou Stief before returning to Oxford. In 1993 he was appointed Lecturer in Earth Observation Science and Chemistry at the University of Leicester and in October 1995 took up a Lectureship in Chemistry at the University of Reading. His research work involves the application of lab- oratory measurements in spec- troscopy and kinetics to problems in atmospheric chem- istry.absorption of radiation, photodissociation, product formation in exothermic reactions and the interaction of molecules with charged particles. Once formed, these species can influence atmospheric chemistry in many ways. As we shall see in Section 2,O(lD) plays a fundamentally important role in the chemistry of both the tropo- sphere and the stratosphere. The reactivity of O(lD) is significantly different from that of the ground state, O('P), and other examples of this behaviour also exist. Excited atomic oxygen is also impor- tant in the aurorae and airglow and these phenomena are discussed. Radiation emitted from vibrationally excited molecules can be detected in space, allowing remote sensing of many species, while chemical reactions leading to vibrationally excited products can lead to significant problems in the retrieval of concentrations from such measurements.Infrared raditation emitted from the Earth's surface can excite vibrations in a number of trace gases in the atmosphere leading to radiation trapping and the greenhouse effect; in the middle atmosphere, emission from vibrationally excited molecules provides a mechanism for atmospheric cooling. The role of nitric oxide in the thermosphere is a very interesting one; mech- anisms for its formation seem to involve translational, rotational, vibrational and electronic excitation. It is not possible in a review of this length to cover all aspects of the chemistry of excited states in the atmosphere with the detail that they deserve.In particular, it should be noted that almost all atmos- pheric chemistry is driven by photochemical processes and that such processes involve the promotion of molecules to excited states. The dynamics of the reorganisation of atoms following these excita- tions are not considered in this article. Similarly, there IS not room to discuss the notation used to describe electronically excited states [e.g.O(lD) and 02(a1A,)] and the interested reader is referred to an introductory text on spectroscopy.' I do not intend to give a detailed description of the chemistry of the atmosphere or its structure; some of the standard chemistry2 is dealt with in Section 2, and the atmos- phere is divided into regions according to temperature gradient.These regions are (with very rough altitude ranges): the troposphere (< 18 km); the stratosphere (18-50 km); the mesophere (SO-90 km); and the thermosphere (>90 km).2 2 The Importance of O(lD) The first excited state of atomic oxygen, O(ID), is one of the most important trace constituents in the Earth's atmosphere, playing crit- ical roles in many areas of atmospheric chemistry. In this section, its importance in the troposphere and stratosphere is discussed. In chemical terms, the troposphere can be thought of as a low-temper- ature combustion system. One of the most important overall pro- cesses in this region is the conversion of methane, a by-product of many natural processes, to carbon dioxide and water by a reaction that, stoichiometrically, can be represented very simply: CH, + 20, -CO, + 2H,O (1) However, just as in the case of a methane flame, the mechanism for this process is complex and involves a number of radical species.In a combustion system, the energy for the formation of radicals is thermal, while in the troposphere, the energy source for radical formation is solar radiation and the oxidation is photochemical. The critical process is the photolysis of ozone at wavelengths between 290 and 3 10 nm: 33 The atomic fragment, O(ID), then goes on to generate the important radical species OH in the reaction O(lD) + H,O -OH + OH (3) The OH radicals thus formed can then abstract a hydrogen atom from methane molecules OH + CH, -H,O + CH, (4) Once this reaction occurs, the rest of the oxidation can proceed It is outside the scope of this article to discuss in detail this mech- anism, but the important point is that the oxidation of methane (and other hydrocarbons) in the troposphere only occurs after the forma- tion of an excited oxygen atom If excited atomic oxygen were not formed in reaction (2a), hydrocarbon oxidation in the troposphere would occur by a very different mechanism What should also be clear is that the oxidising capacity of the troposphere is strongly dependent on the rate at which photolysis of ozone occurs The rate of photolysis depends on the absorption cross-section, quantum yield and the solar flux in the troposphere Because of the absorp- tion of solar UV radiation by ozone in the stratosphere, the absorp- tion cross-section for ozone photolysis at the wavelengths that reach the troposphere is very small The solar flux increases as wave- length increases, but above h = 310 nm, energy restrictions do not allow formation of O(lD) in a spin-allowed process Nevertheless, experiments carried out by Hancock and coworkers3 indicate that in this region of the spectrum, O(lD) may be generated with a signif- icant quantum yield and the information obtained from these experiments may have important implications for our under-standing of tropospheric chemistry The chemistry of the stratosphere is also strongly influenced by O(lD) In this region of the atmosphere, the dominant chemical species is ozone, the formation and destruction of this molecule can be described by a set of oxygen-only reactions known as the Chapman scheme 0, + Izv-0 + 0 O+ 0, + M -0, + M o,+hv-.O,+O 0 + 0,-0, + 0, Qualitatively, this scheme describes the layered structure of ozone in this region quite well, but predicts ozone concentrations that are greater than those actually observed The origin of the discrepancy is that catalytic cycles of the type x +o,-xo+o, xo+o-x +o, allow trace constituents to control the concentration of ozone because the net effect of reactions (8) and (9) is to convert one oxygen atom and one ozone molecule into two oxygen molcules, while regenerating the species X These trace species are reactive radicals such as C1, OH and NO Chlorine atoms have a direct photochemical source, being formed in the photolysis of chlori- nated organic compounds that reach the stratosphere CF,Cl, + hv+CF,Cl + Cl (10) However, another minor source involves reaction with O(ID) CF,Cl, + O(lD) -C10 + CF,CI (1 1) followed by reaction of C10 with 0 atoms Furthermore, the strong- est source of stratospheric NO is the reaction O(lD) + N,O +NO + NO (12) and the only sources of OH also involve O( D) CHEMICAL SOCIETY REVIEWS.1996 O(lD) + H,O -OH + OH (3) O(lD) + CH, -OH + CH, (13) The effect of these active chlorine, nitrogen and hydrogen species in the natural atmosphere is to reduce stratospheric 0, levels by a factor of two and we see that, just as in the troposphere, O(ID) plays a central role in the chemistry of the region 3 The Troposphere Aside from the chemical change brought about by O(ID) described in the previous section, excited states do not play a very important part in the chemistry of troposphere, although they are of critical importance in determining the temperature of the atmosphere and chemical reactions can be strongly temperature-dependent Quenching of excited species by 0, and N, competes effectively with unimolecular processes such as radiative decay in this region More importantly, the high-energy photons and charged particles that cause excitation in the upper atmosphere have been absorbed and there are few processes with sufficient energy to generate excited states in the troposphere One of the reasons that excited states are important in the atmosphere is that those that are access- ible from the ground state in an allowed transition determine the radiation field throughout the atmosphere In the optical region of the spectrum, excited states can be located using conventional tech- niques,' while in the far-ultraviolet, electron energy loss spec- troscopy4 ? has proved useful The main excitation mechanism in the troposphere is, as we have seen, the photolysis of ozone to generate O(lD) and O,(alA,) The excited atomic fragment is rapidly quenched, but generates ground- state 0 atoms that in turn react with molecular oxygen to form ozone and thus start the cycle again, i e quenching removes O(lD) from the atmosphere only temporarily The molecular product of the dissociation, O,(alA,), is quenched only slowly dt atmospheric pressure, and is present in the troposphere at relatively high concentrations It was thought6 that this species could contribute to the oxidation of hydrocarbons, particularly alkenes, but quantitative studies have shown it to be very unreactive O,(alAJ in solution is known to be important in, for example, biological system^,^ and recent evidence suggests that it may play a minor role in atmos- pheric oxidation in cloud droplets 5, Ozone may also lead to excited state formation 1zu reaction with alkenes These reactions are 1,3-dipolar cycloadditions and the adduct falls apart to give a carbonyl and a carbonyl oxide known as a Criegee intermediate, RIRT-0-0 In the gas phase, the inter- mediate is believed to be generated with significant vibrational excitation and may either decompose or be stabilised and take part in bimolecular reactions Atkinson and coworkers'() have measured OH yields from these reactions and it seems that the OH is gener- ated in the decomposition of the Criegee intermediate A correlation exists between the fractional yield of OH and the standard enthalpy changes for the reactions" of ozone with a series of alkenes The more energy released, the greater the OH yield, as might be expected if the OH is generated from a Criegee intermediate with vibrational excitation However, Hatakeyama et ul l2 have shown that, although the yield of stabilised Criegee intermediates is pres- sure dependent, the yield at the high-pressure limit for (E)-but-2- ene is significantly less than unity They have suggested that the Criegee intermediate is generated not only with vibrational excita- tion, but that some is formed on an unbound electronically excited surface It is clear that much experimental and theoretical work is still required to understand the detailed mechanism of these reac- tions 4 The Aurorae and Airglow The most spectacular manifestation of the existence of excited states in the Earth's atmosphere comes in the form of the Aurora Borealis (The Northern Lights) and the Aurora Australis (The Southern Lights) These phenomena appear in the polar night as shimmering flames that light up the sky, with electronically excited THE EXCITED STATE IN ATMOSPHERIC CHEMISTRY-G.MARSTON states of atomic oxygen playing key roles. These dramatic features result from the interaction of the solar wind with atmospheric constituents at altitudes greater than 100 km, the solar wind having been deflected by the magnetic field of the Earth. Indeed, the Earth's magnetosphere has been described as a gigantic cathode-ray tube that focuses solar electrons onto the Earth's poles, with the atmos- phere behaving as a fluorescent screen.I3 It might be better, in this context, to describe the atmosphere as a phosphorescent screen, because, as will be described, the major transitions giving rise to the emitted radiation are forbidden by the electric dipole selection rules.Incoming solar electrons with energies in the region of 10000eV interact with molecules in the upper atmosphere of the polar regions, leading to processes such as electron impact ionization, dissociative electron attachment and energy transfer. I4 The products of these interactions may emit radiation themselves, or take part in reactions leading to excited entities. Although a number of excited species are generated during aurorae, the most prominent visible features arise from atomic oxygen. The strongest observed emission comes from the O(lS) +O(ID) transition, the so-called 'auroral green line' at A = 557 nm, but a weaker emission in the red occurs from the O(ID)-O('P) transition at A = 630 nm. These excited states are believed to be generated in the dissociative recombination process: l4 Many of the transitions that give rise to the aurorae are formally for-bidden by the electric dipole selection rules, and the extent to which they are forbidden can have a significant effect on the location of auroral features.For example, the origin of the auroral green line is at an altitude of approximately 100 km, while emission from O(ID) occurs at altitudes up to about 400 km. Part of the reason for this difference lies in the different radiative lifetimes of the two states. In order for emission to be observed, the emission rate must compete with the rate of quenching by atmospheric molecules.Emission from O(lS), with a radiative lifetime of 0.7 seconds, can compete with quenching at much lower altitudes (higher pressures) than can O(lD), with a radiative lifetime of 200 seconds. The quenching rates of the two species are also important, O(lS) being quenched much more slowly than O(ID). Emission from O(IS) is also a strong feature of the airglow, vari- ously called the dayglow, twilightglow or nightglow depending on when it is observed. The airglow and the aurorae share many common emission features, but the aurorae are sporadic, intense and concentrated in the polar regions, while the airglow is continu- ous, very weak and can be observed at all latitudes. Excitation mechanisms for particular species in the airglow often differ from those in the aurora.For example, the auroral green line is excited by energy transfer from electronically excited molecular oxygen in the nightglow. A much-simplified mechanism for its formation is: where 0; is an electronically excited state of molecular oxygen. This two-step process is known as the Barth mechanism15 for excitation of the green line. Other emitters in the airglow are primarily electronically excited species, although the hydroxyl radical, OH, with vibrational excita- tion up to 1' = 9 makes a significant contribution to the nightglow.2 The reaction between H atoms and ozone: H + 0, +OH + 0, (17) has just sufficient energy to populate the 9th vibrational levelI6 and is the dominant source of OH(i*) at night.The reaction: 0 + HO, +OH + 0, (18) which has enough energy to populate OH (v =6), has been consid- ered and rejected by KayeI7 as a secondary source of vibrationally Figure 1 Discharge flow-Fourier transform spectrometer apparatus. B, beamsplitter; D,, germanium detector; D,, laser detector; F, flow tube; I, observation region; L, laser; MI, moving mirror; M,, fixed mirror; S, stepper motor. The orange glow emanating from the flow tube is the Lewis-Rayleigh afterglow produced from discharged N,. The major con- tnbutor to the glow is from the N,(B3rIg-A%:) transition, which is observed in aurorae. excited OH. However, laboratory evidence has shown that the reac- tion can give rise to emission from OH (v d 6). Figure 1 shows a discharge-flow apparatus coupled to a Fourier transform infrared spectrometer that has been used to study airglow mechanisms in the laboratory in the near-infrared region of the spectrum.I8 Using this apparatus, Lunt et al.detected emission from vibrationally excited OH formed in reaction 18,and an emission spectrum recorded from the products of reaction 18 is illustrated in Figure 2. A quantitative laboratory study, comparing the strength of this emission to that produced by reaction 17, is required to assess the importance of reaction 18 to the nightglow. Not only does vibrationally excited OH contribute to the airglow, but it can also take part in chemical reactions that are not possible for ground-state hydroxyl radicals. The reaction between 0 atoms and OH to give 02(alAg) as a product is only energetically possible if the OH radical possesses at least one quantum of vibrational energy.0 + OH(v 3 1) -H + 02(a1Ag) (19) 02(aIAg)is one of the strongest airglow emitters, despite the fact that the transition to the ground state 02(X3Z;) is forbidden by three electric dipole selection rules; its emission spectrum'' (measured in the laboratory) is illustrated in Figure 3 at medium resolution (ca.8 cm-I). Evidence that reaction 19occurs comes from field, modelling and laboratory investigations. Gattinger and Vallance-Jones2" have observed correlations in rocket-borne experiments between 02(alA,) emissions and emissions from vibrationally excited OH, while, on the basis of a modelling study, Krasnopolsky2I suggested that if 20% of the overall process generated 02(a1Ag), the observed altitudexoncentration profiles for O,(aIA,) could be explained.In the laboratory, Lunt et al.18have shown that reaction 19does occur, but with an efficiency of only ca. 2.5%. Their results seem to indi-cate that reaction 2 can make only a small contribution to the night time emission, but uncertainties in calibration and the difficulties of modelling a reaction system with OH in a range of vibrational levels mean that further experiments are needed on this reaction. Current interest in O,(aIA,) nightglow is centred on 0-atom recombination followed by energy transfer,22 in a mechanism that is closely related to the Barth mechanism for the formation of O(lS) referred to above.It is worth mentioning at this point a recent study by Miller et al.,23following from work by Slange~-,~~ that illustrates again how vibrational excitation can lead to enhanced reactivity. CHEMICAL SOCIETY REVIEWS, 1996 r 0-OH <o8 *)X It 0)a.-ro v -1 (r" '4 I 1 I 1 I I 1800 1600 1400 1200 1000 Unm Figure 2 Vibrational emission from OH formed in the reaction of 0atoms with HO,. The numbers identify the upper and lower vibrational quantum numbers involved in the transition and the arrows locate the positions of the Q branches. The laser lines are from scattered light from the HeNe laser shown in Figure 1. Resolution, 8 cm-I. (Reproduced with permission from J. Chem.Soc., Faraday Trans. 2, 1988,84, 899) z7800 8000 v/cm-' Figure 3 Emission from the transition O,(a'A, -X-1C;). Resolution, 8 cm-I. Approximately 10% of the photolysis of ozone between 200 and 310 nm occurs to give 0 and 0, in their electronic ground states: 0, + hv -02(X3Z;) + O(3P) (2b) the remaining 90%leading to excited singlets (reaction 2a). In reac- tion 2b, a significant fraction of the residual energy is channelled into vibrational excitation in the molecular fragment. Vibrationally excited 0, cannot contribute to the airglow as 0, does not possess a dipole moment, but reaction with 0, to give 0, is energetically feasible: O,(v 2 26) + 0, -0 + 0, (20) As yet, evidence is circumstantial, but the reaction may affect ozone concentrations by as much as 10% at some altitudes, and further lab- oratory investigation is urgently required. Continuing for the moment with the subject of ozone photolysis, the major daytime contributor to the airglow is O,(a'A,), formed in reaction 2a: 0, + hv-, O,(a'A,) + O(lD) (2a) As discussed in Section 2, there are some uncertainties in the wave- length dependence of the quantum yield for this process., Nevertheless, the broad features of O,(alA,j formation and quench- ing rates are sufficiently well understood in the daytime atmosphere that measurements of emission from O,(a'A,) made by rocket- borne instruments have been used successfully to generate [O,]-altitude profiles.2s These measurements are a slightly unusual example of how emission from excited states can be used to detect species in the atmosphere and this topic makes up the next section of this review.5 Remote Sensing A number of satellite-borne instruments have measured the concentrations of atmospheric constituents by observing emission from their vibrationally excited states. For example, the Limb Monitor of the Stratosphere (LIMS j instrument onboard the Nimbus 7 satellite obtained vertically-resolved concentration pro- files with almost global coverage for O,, NO,, H,O and HNO,. More recently, the Improved Stratospheric and Mesopheric Sounder (ISAMS) and the Cryogenic Limb Array Etalon Spectrometer (CLAES j on the Upper Atmosphere Research Satellite (UARS) made observations on a range of molecules using this method.Concentrations are retrieved from measured radiances in appropriate regions of the spectrum and it is assumed that the species of interest is in thermal equilibrium, i.e. that Boltzmann's THE EXCITED STATE IN ATMOSPHERIC CHEMISTRY-G MARSTON H-0 filter 105 10-~ 1300 1400 1500 1600 1700 vJcm ’ Figure 4 Spectral responses of LIMS 6 2 pm (NO,) and 6 9 pm (H,O) filters and absorption coefficients of H,O v, and NO, v3 fundamental bands (Reproduced with permission from B J Kerridge and E E Remsberg, J Geophys Res 1989,94, 16 323) equation is obeyed This assumption is known as the Local Thermodynamic Equilibrium (LTE) approximation Absolute concentrations can be obtained for the vibrationally excited mole- cule if the instrument is accurately calibrated in the spectral region of interest and the Einstein A factor is known for the vibrational transition Unfortunately, the LTE approximation can break down if mech- anisms other than thermal excitation exist for the formation of the vibrationally excited state that is being observed For example, Solomon et ul 26 have considered whether Nonlocal Thermodynamic Equilibrium (NLTE) effects could influence the retneval of ozone concentrations from radiance measurements by the LIMS instrument at 9 6 Fm The v, fundamental band of ozone emits in this region of the spectrum, but laboratory studies showed that an appreciable fraction of the reaction 0 + 0, + M-0, + M (21) leads to excitation of ozone in this mode In addition, the i31= vibrational level can be excited directly by absorption of infrared radiation emitted from the Earth’s surface From a detailed model- ling study using laboratory data on the kinetics of the formation and quenching of vibrationally excited ozone, Solomon et ul came to the conclusion that reaction 21 had a substantial effect on ozone retrievals at altitudes above 50 km While these interferences cer- tainly increase the difficulty of determining ozone concentrations, these authors point out that the measurements also provide informa- tion about nonthermal processes in the atmosphere Kerridge and Remsberg,’ have considered the possibility that NLTE in the t3 mode of NO, may affect both NO, and H,O retrievals from the LIMS instrument Figure 4 shows the spectral responses of the LIMS filters for H,O and NO, along with spectra for these species between 1300 and 1700 cm I The figure shows (\v3that direct excitation to NO, = 1) by absorption of infrared radiation followed by re-radiation might affect the NO, measure-ments, but could only have a minimal effect on the water retrievals, because the overlap between the NO, (v3 = 1) emission spectrum and the H,O filter is very small However, processes leading to the excitation of higher vibrational levels of NO, could lead to emission in the region of the water filter For example the 1 ,= 6 to \ = 5 transition occurs at the peak transmission of the H,O filter Kerridge and Remsberg considered two possible mech- anisms for the formation of highly vibrationally-excited NO, Absorption of visible solar radiation at wavelengths greater than the dissociation threshold for NO, (400 nm) leads to excitation to the first excited state of NO, Very strong coupling exists between the ground (X2Al)and first excited state (A2B2) leading to vibra-tional excitation in the molecular ground state The reaction of NO with 0, NO + 0, -NO, + 0, (22) is known to form both electronically and vibrationally excited NO, Clough and Thrush published emission spectra of the Av, = A1 ,= -1 combination bands at 3 7 pm, and detected some emission at 6 2 pm, but did not publish the spectrum 28 Laboratory evidence supporting the role of reaction 22 comes from work carried out at the Rutherford Appleton Laboratory, where emission resulting from products of the reaction was recorded using a high-resolution Fourier transform spectrometer 2o Figure 5 shows a spectrum of the emission observed on reaction of NO with 0, The position and structure of the band reveal that it cor- responds to the 1, = 1 to I?, = 0 transition of NO,, showing that reaction 22 could well affect the retrieval of NO, from the LIMS data Furthermore, at lower resolution, emission can be detected at slightly lower energy, as is illustrated in Figure 6 This emiwon appears to come from Av, = -I transitions of NO, with significant amounts of vibrational energy, providing evidence for Kerridge and Remsberg’s suggestion that reaction 22 could influence the retrieval of H,O data There is one further point regarding the formation of excited state products in the reaction of NO with 0, that is of some importance The reaction was once considered as a possible source of nighttime O,(alAg) in the mesosphere, thus contributing to the nightglow However, laboratory evidencelY indicates an upper limit of 3 X 10 for the fraction of reaction 22 that forms O,(aIA,) At the same time, electronically excited NO,(A*B,) appears to be formed in the reaction These two observations are, at first sight, contradictory The reaction of NO with 0,can, as a result of the degeneracy of NO, proceed on two reaction surfaces and the two lowest-lying sets of products are N02(X2Al) + O,(XiC,) and N0,(X2A,) + O,(aIA,) Even if we restrict the transition state to a planar configuration, these two sets of products correlate with reactants, and it is difficult to see how N0,(A2B,) can be accessed Redpath et ul w have tried to reconcile the problem by suggesting that the reaction proceeds through a transition state where the NO approaches the central 0 atom of the ozone molecule, maintaining C, symmetry Under these circumstances, O,(alA,) does not correlate with reactants, wherea5 an excited state of NO, (but not the A2B, state) does However, this explanation suggests that the reaction proceeds through a transition state with a very restricted geometry AdIer-Golden3I has suggested that NO, is actually generated in the ground electronic state, but with high vibrational excitation The vibrationally excited NO, can then cross onto the A2B, state, thus explaining how this state can be formed while 02(alAg) is not Adler-Golden’s description of the reaction mechanism also explains some work carried out by Clough and Thrush,28 who dis- covered that the temperature dependences of the emission from vibrationally and electronically excited NO, formed in reaction 22 were identical These authors proposed that vibrationally excited NO, was formed I-la the electronically excited state, rather than the other way around The difficulty with this mechanism is that it implies that none of the excess energy in the ground-state channel CHEMICAL SOCIETY REVIEWS, 1996 1500 1550 1600 1650 17 10 G/cm-’ Figure 5 Spectrum showing emission from vibrationally excited NO, generated in the reaction of NO with 0,.Resolution, 1 cm-I.I I I 1400 1800 2200 2600 3000 tr/cm-’ Figure 6 Spectrum showing emission from vibrationally excited NO, generated in the reaction of NO with 0,.Resolution, 8 cm-I. is released into the vibrational modes of NO,; because NO, contains the newly formed bond in the reaction, this description of the reac- tion is unconvincing. From the point of view of the LIMS retrievals, the details of the mechanism are very important; only a small frac- tion of the reaction generates electronically excited NO, and Clough and Thrush’s mechanism would suggest that reaction 22 could not significantly affect vibrational populations in the atmos- phere.In Adler-Golden’s interpretation, the experimental evidence allows a significant fraction of the reaction to generate vibrationally excited NO, in the atmosphere. 6 Heat Balance in the Atmosphere In order that satellite instruments such as LIMS and ISAMS are able to measure concentrations of atmospheric constituents, energy, in the form of infrared radiation, must be transferred from the atmos- phere to space. Excited states are clearly linked to this process and have an important part to play in both the heating and the cooling of the atmosphere. In the lower atmosphere, absorption of infrared radiation emitted from the Earth’s surface excites vibrations and leads to warming of the atmosphere. At higher altitudes, vibration- ally or electronically excited species may be generated by a number of mechanisms, and the subsequently-emitted radiation can escape this optically-thin region of the atmosphere and lead to cooling or reduced heating efficiency.It is possible to calculate the temperature of the Earth assuming that it is in thermal equilibrium with the Sun and that both behave as black bodies., The calculated temperature is approximately 256 K, in reasonable agreement with the temperature measured from space on the basis of the total thermal radiation emitted by the THE EXCITED STATE IN ATMOSPHERIC CHEMISTRY-G. MARSTON Earth. However, surface temperatures on Earth are on average around 288 K, significantly higher than both the calculation and measurement from space.The origin of this discrepancy is that while the Earth may be treated, at least approximately, as a black body, the Earth and the atmosphere combined may not. Viewed from space, the total amount of radiation emitted by the Earth gives a temperature of 256 K according to Stefan's Law, but the frequency distribution of radiation is closer to that of a black body with a tem- perature of 288 K with radiation at certain wavelengths not present. These wavelength regions with no radiation result from the absorp- tion of radiation by infrared-active molecules such as CO, and H,O, the greenhouse gases. These molecules are excited into one of their vibrational modes and this energy is ultimately converted into heat, thus increasing the temperature of the lower atmosphere.At high altitudes, collisions can excite vibrations in infrared-active mole- cules (mainly CO,), which can then emit radiation. If the atmos- phere is optically thin, trapping will not occur and the radiation can escape to space, causing atmospheric cooling. The ultimate source of heat for the Earth and its atmosphere is the Sun. While in the lower atmosphere, atmospheric heating occurs mainly indirectly, following warming of the Earth's surface, at higher altitudes heating occurs directly. A good example of the results of direct heating is the temperature inversion observed in the stratosphere., This inversion occurs because ozone is present in rel- atively high concentrations in this part of the atmosphere, and absorbs solar UV radiation which causes a warming effect.The resultant temperature gradient, with warm air above cold, is extremely stable to vertical motion and thus has a very important effect on the atmospheric dynamics of the region. At altitudes below the stratosphere, the efficiency of conversion of solar radiation to heat is very close to unity, but it should be borne in mind that the mechanism for the conversion is complex and at higher altitudes, the efficiency may be very much reduced. Absorption converts solar radiation into atomic and molecular internal energy, and, if bonds are broken, chemical potential energy. Mlynczak and Solomon32 have considered the mechanism of middle atmosphere heating following the photolysis of ozone.As we have already seen, this process occurs through more than one channel: 0,+ hu-O,(aIAJ +O(lD) 0,+ hv-O,(XYZ;) + o(3~) Rate constants for quenching of O,(alA,) by atmospheric gases are very small (ca. 10-20 cm3 molecule-* s-l) and emission at h = 1270 nm can transfer energy from the atmosphere; above 40 km, virtually all photons emitted in the upward direction escape and many photons emitted in the opposite direction will heat parts of the atmosphere far removed from the photolysis event. Mlynczak and Solomon make the point that this process does not constitute atmos- pheric cooling because the energy has not been converted to heat when it is released; rather, the escape leads to a reduction in heating efficiency of the absorbed radiation.Quenching of O(lD) by N, is very rapid and emission from atomic oxygen does not affect heating efficiencies in the lower thermosphere. However, vibrationally excited N, formed in the quenching collision can transfer its energy to CO, in the v3mode which can then emit at h = 4.3 pm. Mlynczak and Solomon used a detailed model to conclude that below about 50 km the heating efficiency of the singlet channel was virtually loo%, but may drop to as low as 65% at 100km. The triplet channel could also have a reduced efficiency if the vibrationally excited 0, were not completely converted to heat. These authors concluded that quenching was efficient, but did not consider the possibility of the reaction of vibrationally excited 0, (Section 3, reaction 20).The effect of chemical potential energy on the heating of the atmosphere was also considered', by Mlynczak and Solomon. This energy is released following exothermic chemical reaction and such reactions may occur a long way removed from where the photoly- sis happened, thus providing a way of transferring energy through the atmosphere; warm polar winters in the mesosphere have been attributed to this behaviour. Mlynczak and Solomon presented evi- dence to show that the reaction of H atoms with 0, is an important source of heat in the middle atmosphere, although their analysis was hampered by uncertainties in the radiative and quenching lifetimes of vibrationally excited OH. It has been assumed that the reaction of 0 atoms with HO, efficiently converts chemical potential energy into heat. However, as we saw in Section 4,this reaction appears to generate vibrationally excited OH and its heating efficiency may be reduced by radiative emission from these species.Cooling by emission from vibrationally excited CO, is the most important mechanism for disposing of heat from the middle atmos- phere. However, at very high altitudes, this process becomes increasingly inefficient and a more important mechanism is via transitions between the fine-structure components of ground-state atomic oxygen, O(3P).O('P) is made up of three components with different values of J, the quantum number for total (orbital plus spin) electronic angular momentum; the components are, in increas- ing energy, 'P,, ?P,and 3P,, with degeneracies of 5,3 and 1, respec-tively. The transition: is believed to make a significant contribution to the cooling of the thermosphere, at least for as long as the fine-structure components exist in thermal equilibrium.The sparsity of collisions in the upper thermosphere is such that local thermodynamic equilibrium is not maintained3, (for z >400-600 km depending on solar activity), and this cooling mechanism is lost as the J = 2 level becomes over- populated relative to the other components. This effect is an inter- esting example as it shows that the excited state can be important because it is not occupied when, on the basis of the Boltzmann dis- tribution, it should be.7 Nitric Oxide in the Thermosphere Another species that contributes to the dissipation of energy from the thermosphere is nitric oxide, NO, which appears to moderate the temperature during periods of high solar and geomagnetic activity. A discussion of the behaviour of this molecule in the thermosphere is particularly pertinent to this review because we must consider translational, rotational, vibrational and electronic excitation! Thermospheric nitric oxide is of great importance because during the long polar winter it is transported down through the atmosphere and affects stratospheric ozone concentration^.^^ The molecule is formed in the reaction of ground-state N(4S) atoms with 0,: N(4S) + 0, -NO + 0 (24) This reaction is slow at room temperature,35 and even at thermo- spheric temperatures is not a major source because NO can be destroyed in the 'cannibalistic' reaction: N(4S) + NO -N, + 0 (25) A more important source of NOJ4involves the reaction of the first electronically excited state of N atoms, N(,D), with 0,: This species is generated in a number of processes, the main lower thermospheric source34 being the dissociative electron attachment of NO+: NO+ + e -N(2D) + 0 (27) while electron impact dissociation of N, is also important: N, +e -N(4S) +N(,D) + e (28) The electrons involved in reaction 28 must possess a significant amount of energy in order to effect the dissociation and are either photoelectrons or secondary auroral electrons.A second excited state of atomic nitrogen, N(,P), is also formed36 in the electron- impact dissociation of N,, but is quickly quenched to N(2D) Barth’4 concluded in a modelling study that NO concentrations in the lower thermosphere were highly sensitive to the branching ratio for N(2D) in reactions 27 and 28 However, there are other uncertainties associated with thermo- spheric NO production It has been suggested that one channel of reaction 26 may lead to the formation of O(‘D) and thus contrib- ute to auroral or dayglow emission at 630 nm GCrard36 suggests that there is no evidence one way or the other for this supposition and that the problem will not be resolved until direct laboratory evidence is obtained Recent measurements indicate that NO is present in the thermosphere in high vibrational states, but also with rotational energy significantly greater than expected from thermal processes 77 Reaction 26 is energetically capable of accessing the observed levels, but angular momentum constraints seem to require significant translational energy in the reactants 37 Theoretical calculations on the dynamics of reaction 24 suggest that translationally hot ground-state N(4S) atoms may react with 0, significantly faster than N atoms with a thermal velocity distri- bution and be an important source of NO, particularly in the daytime thermosphere M Emirsion from electronically excited NO is also detected in the thermosphere,39 one source being the recombination of N and 0 atoms The Sand y bands of NO have been detected by rocket-borne spec- trorneterc3’ and used to derive nighttime N(5) altitude+oncentra- tion profiles (during the day, fluorescent emission from electronically excited NO interferes with such measurements, but has been used to determine NO concentrations) Emission from NO(b4C ) generated in reaction 29 has been detected in the near- infrared,40 and it can be assumed that a considerable fraction of the reaction leads to NO(a4n) formation, as only this state and the ground state correlate with the reactant atoms It is likely that these 5tates have some role to play in the heating efficiency of reaction 29, doublets formed in the reaction will emit radiation to space, ulti- mately giving ground-state NO(X2n), emission from NO(a4n) is forbidden by the electronic dipole selection rules, and so energy rtored in this state should be efficiently converted to heat 8 Concluding Remarks In this paper the role of the excited state in the atmosphere has been reviewed It is clear that excited states are important in many aspects of atmospheric chemistry and physics Some parts of the science are quite well understood, the chemistry of O(ID) in the troposphere is a good example, although it should be remembered that there are uncertainties in the quantum yield for its formation from ozone photolysis at wavelengths that are important in the lower atmos- phere Advances in satellite technology allow excited species to be observed in the remotest regions of the atmosphere and enormously fast digital computers can be used to interpret the measurements through modelling studies These models require a detailed knowl- edge of the fundamental physical and chemical processes that occur in the atmosphere and such knowledge comes from laboratory experiments Many questions concerning these fundamental pro- cesses remain unanswered and even details from well-studied systems (e g ozone photolysis) can prove to have important conse- quences Determining quantitative details for the mechanisms of formation of excited species and their losses through quenching, radiative decay and reaction require sophisticated experimental methods and still provide a significant challenge for laboratory sci- entists CHEMICAL SOCIETY REVIEWS, 1996 9 References 1 J M Hollas, Mode? n Spectroscopy, 2nd edn ,Wiley, Chichester, 1992 2 R P Wayne, Chemistry of Atmospheres, 2nd edn , Clarendon Press, Oxford, 1991 3 S M Ball and G Hancock, Geophys Res Lett, 1995,22, 1213 4 W Johnstone, N J Mason, W R Newell, P Biggs, G Marston and R P Wayne, J Phys B At Mol Opt Phys , 1992,253873 5 J E Davies, N J Mason, G Marston and R P Wayne, .I Phys B At Mol Opt Phys , 1995,28,4179 6 A U Khan, J N Pitts, Jr ,and E B Smith, En\ iron Sci Techno1 ,1967, 1,656 7 R V Benasson, E J Land and T G Truscott, Excited States and Free Radicals in Bioloqy and Medicine, Oxford University Press, Oxford, 1993 8 F Bohm, G Marston, T G Truscott and R P Wayne, .I Chem Soc Faraday Trans , 1994,90,2453 9 B C Faust and J M Allen, J Geophys Res , 1992,97, 12 913 10 R Atkinson, S M Aschmann, J Arey and B Shorees, J Geophys Res , 1992,97,6065 11 G Marston, unpublished results 12 S Hatakeyama, H Kobayashi and H Akimoto, J Phys Chem , 1984, 88,4736 13 S I Akasofu, Am Scientist, 1981,69,492 14 M H Rees, Physics and Chemistrq of the Upper Atmosphere, Cambndge University Press, Cambndge, 1989 15 C A Barth,Ann Geophys 1964,20, 182 16 G E Streit and H S Johnston, J Chem Phys , 1976,64,95 17 J A Kaye, J Geophys Res , 1988,93,285 18 S T Lunt, G Marston and R P Wayne, .I Chem Soc Faraday Trans 2, 1988,84,899 19 G Marston, D Phil Thesis, Oxford, 1987 20 R L Gattinger and A Vallance-Jones, in Physics and Chemistry of Upper Atmospheres, ed B M McCormac, D Reidel, Dordrecht, 1973 21 V A Krasnopolsky, Planet Space Sci , 1985,34,511 22 R P Wayne, Res Chem Intermed , 1994,20, 395 23 R L Miller,A G Suits,P L Houston, R Toumi, J A MackandA M Wodtke, Science, 1994,265, 183 1 24 K 0 Patten, Jr, P S Connell, D E Kinnison, D J Wuebbles, T G Slanger and L Froidevaux, J Geophys Res, 1994,99, 12 1 1 25 J F Noxon, Planet Space Sci , 1982,30,545 26 S Solomon, J T Kiehl, B J Kerndge, E E Remsberg snd J M Russell 111,J Geophys Res , 1986,91,9865 27 B J Kerridge and E E Remsberg, J Geophys Res , 1989,94, 16 323 28 P N Clough and B A Thrush, Trans Faraday Soc , 1969,65,23 29 G Marston, J Ballard, B J Kerndge and R P Wayne, manuscnpt in preparation 30 A Redpath, M Menzinger and T Carnngton, Chem Phys , 1978, 27, 409 31 S M Adler-Golden, J Phys Chem ,1989,93,684 S M Adler-Golden, J Phys Chem , 1989,93,691 32 M G Mlynczak and S Solomon, J Geophys Res -Atmospheres, 1993, 98, 10517 33 R Sharma, B Zygelman, F von Esse and A Dalgarno, Geophys Res Lett, 1994,21, 1731 34 C A Barth, Planet Space Sci , 1992,40,315 35 A J Bamett, G Marston and R P Wayne, J Chem Soc Faiaday Ttaizs 2, 1987,83, 1453 36 J C Gerard, Planet Space Sci , 1992,40,337 37 P S Armstrong, S J Lipson, J A Dodd, J R Lowell, W A M Blumberg and R M Nadile, Geophys Res Lett, 1994, 21,2425 38 J W Duff, F Bien and D E Paulson, Geophys Res Lett, 1994, 21, 2043 39 P D Tennyson, P D Feldman, J F Hartig and R C Henry, J Geophys Res , 1986,29,595 40 P Biggs, G Marston and R P Wayne, J Mol Spect , 188,129,236