Kinetic energy distributions and line profile

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 109, A09303, doi:10.1029/2002JA009353, 2004

Kinetic energy distributions and line profile measurements of dissociation products of water upon electron impact Oleg P. Makarov,1 Joseph M. Ajello, Prahlad Vattipalle, and Isik Kanik Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California, USA

M. C. Festou Observatoire Midi-Pyreńeés, Toulouse, France

Anil Bhardwaj2 Space Physics Laboratory, Vikram Sarabhai Space Centre, Trivandrum, India Received 26 February 2002; revised 24 May 2004; accepted 23 June 2004; published 10 September 2004.

˚ ) and O I (1302 A ˚ and 1152 A ˚ ) resulting [1] The Doppler line profiles of H Ly-a (1216 A from electron impact dissociative excitation of H2O have been measured with a highresolution (l/Dl = 50,000) ultraviolet spectrometer. The line profiles are used to calculate the kinetic energy distribution of the hydrogen atoms produced in dissociative excitation and ionization of H2O at electron impact energies 25, 35, and 100 eV. Three distinct populations of H(2p) atoms were found. The kinetic energy of hydrogen atoms is found to have contributions from a low-energy component, with a mean energy of 0.2 eV at all the three electron impact energies. In addition, a medium-energy component appears with a mean energy of 2.0 eV for 35 eV electrons, and a high-energy component, 7 eV, ˚ ) and O I (1152 A ˚ ) line profiles for 100 eV electrons. The measurement of O I (1302 A indicate that the kinetic energy of excited O I atoms is very small (1 eV or less) at all electron impact energies. Most of the energy released in dissociation is found in the translational energy of the hydrogen atoms. The excitation functions of H Ly-a, O I ˚ ) feature of oxygen, and A(0) X(0) molecular band of hydroxyl near 3050 A ˚ (1302 A ˚ from threshold to 600 V were also measured. The spectrum (1.0 A FWHM) of the rotational structure of OH (A X) from electron impact dissociation indicates a high degree of rotational excitation, which is almost identical to the rotational structure from dissociative recombination of H2O+. The results presented in this paper have important applications to planetary bodies, like comets, icy satellites of outer planets, Saturn’s magnetosphere, and rings, all of which have H2O and its daughter products in large INDEX TERMS: 6005 Planetology: Comets and Small Bodies: Atmospheres—composition and amount. chemistry; 6020 Planetology: Comets and Small Bodies: Ice; 6210 Planetology: Solar System Objects: Comets; 6218 Planetology: Solar System Objects: Jovian satellites; 6280 Planetology: Solar System Objects: Saturnian satellites; KEYWORDS: comets, ultraviolet spectroscopy, electron collisions, cross sections, water Citation: Makarov, O. P., J. M. Ajello, P. Vattipalle, I. Kanik, M. C. Festou, and A. Bhardwaj (2004), Kinetic energy distributions and line profile measurements of dissociation products of water upon electron impact, J. Geophys. Res., 109, A09303, doi:10.1029/2002JA009353.

1. Introduction [2] Water is ubiquitous in a host of solar system bodies. It is found in telluric planet atmospheres, comets, asteroids, numerous giant planet satellites, the rings of Saturn, Centaurs, and transneptunian objects, including Pluto and Charon. Consequently, water and its dissociation products 1 Now at Physics Department, University of Connecticut, Storrs, Connecticut, USA. 2 Now at NASA Marshall Space Flight Center, Huntsville, Alabama, USA.

Copyright 2004 by the American Geophysical Union. 0148-0227/04/2002JA009353$09.00

are expected to be found in very different environments. Once water exists in gaseous form, it can be dissociated by UV photons and/or energetic particles into dissociation products, that is, H, O, OH, or H2, etc. Water is extremely difficult to directly observe from the ground since it does not have any strong optical resonance transitions. Therefore its daughter products are often used as proxies for determining the abundance of water in planetary bodies. Moreover, daughter products are also observed for their own importance and also because of the unique role they subsequently play in their environment, like satellites’ atmospheres and magnetospheres of the giant planets. The yields for the various dissociation channels and the energetics of the production of the end products of water

A09303

1 of 15

A09303

MAKAROV ET AL.: ELECTRON IMPACT DISSOCIATION OF WATER

destruction are thus of primary importance in various studies of solar system bodies. The Saturn magnetosphere and the atmosphere of comets deserve particular attention, since water and water dissociation products have been found in them in large quantities. With the ongoing Cassini tour of the Saturn planetary system beginning in mid2004 and the Rosetta mission to comet 67 P/ChuryumovGerasimenko, the electron impact studies of H2O are particularly relevant. This paper addresses the consequences of e-H2O collisions on the energetics and line emission profile of the daughter species produced in dissociative excitation processes. [3] Comets are water-dominated bodies [e.g., Wilkening, 1982; Hubner, 1990; Festou et al., 1993], which support a huge (105 – 106 km) atmosphere ‘‘coma’’ when closer to the Sun (3 AU for Halley-type comets). Hence aeronomical processes initiated due to e-H2O collisions become important in their coma. The electrons in the cometary coma can be of photoelectron origin or of solar wind origin. The importance of processes (e.g., excitation, dissociation, ionization) originating due to electron impact in cometary coma has been better realized in the post-Halley era [cf. Cravens et al., 1987; Bhardwaj et al., 1990, 1996; Gan and Cravens, 1990; Gringauz and Verigin, 1991; Haberli et al., 1996; Bhardwaj, 1999], particularly in view of intensive exploration of comet Halley in March 1986 by as many as six different spacecraft and by an extensive network of groundand space-based observations. The wealth of observations now available on comets clearly indicates that an accurate knowledge of the destruction paths of water and waterderived products and information about other collisional processes are required to understand and interpret numerous observations of comets [e.g., Festou et al., 1993; Itikawa et al., 2001]. [4] In the outer solar system, water and its main dissociation products, OH and H, play a major role in controlling the state of the Saturnian magnetosphere [cf. Richardson, 1998]. Atomic hydrogen was first observed in a rocket experiment by Weiser et al. [1977], and its presence was subsequently confirmed by Voyager 1 observations [Shemansky and Hall, 1992], while the existence of OH in Saturn’s magnetosphere was discovered by HST observations [Shemansky et al., 1993; Hall et al., 1996]. The sources of these neutrals in the Saturnian magnetosphere are the icy inner satellites, rings, and meteoroid ablation. The influx of water species into the upper atmosphere of Saturn can have major ramifications for the ionosphere and H+3 near-IR emissions [e.g., Majeed and McConnell, 1991; Moses and Bass, 2000; Bhardwaj and Gladstone, 2000]. The situation at Jupiter is very different. In the Jupiter magnetosphere, not only are icy particles lacking to provide the source of H and OH but also neutrals ionize very rapidly. In fact, near the orbit of Io, ions dominate while at the same distance in the Saturnian system, neutrals are more abundant relative to ions by orders of magnitude. Water is also an important constituent of the surfaces of three out of four main satellites of Jupiter: Europa, Ganymede, and Callisto. The presence of water on these satellites was determined from infrared spectra [Sill and Clark, 1982]. The water on these bodies can be a source of building up a substantial oxygen atmosphere due to sublimation [Kumar and Hunten, 1982] and more importantly

A09303

due to radiolysis of water-ice by energetic particles from the Jovian magnetosphere [e.g., Johnson, 1990; Sieger et al., 1998]. Oxygen atmospheres have been detected on Europa and Ganymede by HST observations of the atomic oxygen ˚ [Hall et al., 1995, 1998; Feldman lines at 1304 and 1356 A et al., 2000]. The ratio of OI line intensities suggest O2 as the source. Even atomic hydrogen has been detected on Ganymede by Galileo UV spectrometer [Barth et al., 1997] and HST-STIS [Feldman et al., 2000] observations of H Ly-a emissions from Ganymedian hydrogen exosphere, while charged particle measurements indicated an ion outflow of protons at Ganymede [Frank et al., 1997]; both observations imply an ongoing hydrogen gas production on the satellite. [5] The velocity of the OH fragment in photodissociation of water molecules by solar radiation has been a subject of some controversy [Festou, 1981; Crovisier, 1989; Wu and Chen, 1993; Budzien et al., 1994], and consequently so has the velocity of H atoms because the conservation of energy and momentum in the water dissociation reaction couple the velocities of H and OH. Therefore a measurement of the velocity of the H atom also yields that of the OH radical. In cometary coma the velocity of H atoms is derived from the measurements of the line profile of hydrogen Ly alpha ˚ or of hydrogen Balmer alpha emission (H Ly-a) at 1216 A ˚ . The H Ly-a line profile of emission (Ha) at 6563 A Kobayashi-Berger-Milon (C/1975 N1) measured with the Copernicus satellite showed that the H has two velocity components: 20 km s 1 (2 eV) and 6 – 8 km s 1 (0.3 eV) [Festou et al., 1979]. This is consistent with the two-step photodissociation of water predicted by Keller and Thomas [1975]; the high-velocity H component arising from H2O dissociation and the low-velocity component arising from dissociation of OH. More recent observations of H Ly-a profile in comet Hayakutake (C/1996 B2) with HST-GHRS showed that the H atom velocity peaks near 18 km s1 and has some components in excess of 20 km s1 [Combi et al., 1998]. The Ha emission line is excited through absorption of solar photons from the Lyman series, H Ly-b and higher, and subsequent cascade to the 22S excited state. The first cometary Ha line profile observation was made with a ground-based Fabry-Perot interferometer in comet Kohoutek [Huppler et al., 1975]. Later observations at Ha in comets have revealed the presence of multiple velocity components of H atom in the cometary coma [Kerr et al., 1987; Brown and Spinrad, 1993]. In particular, the Ha observations of comets Austin (C/1989 X1) and Levy (C/1990 K1) suggested the presence of H atoms with velocity up to 40 km s1 [Brown and Spinrad, 1993]. [6] The excited states of water lie above the lowest set of dissociation thresholds beginning with the lowest threshold at 5.1 eV which produces a pair of ground state H-OH fragments. The UV emission from electron impact induced fluorescence is solely from excited fragments (OH, O, H) [Avakyan et al., 1998]. Ajello [1984] have measured the UV ˚ and have identified spectrum of H2O from 400 to 1800 A 27 atomic features. Avakyan et al. [1998] have reviewed the availability of emission cross-section data. The only UV emissions for which excitation functions have been mea˚ ), and OH(A2S+ ! X2). sured are H Ly a, O I (1304 A ˚) The energy dependence of the H Ly a and O I (1304 A emission cross sections have been studied by Mohlmann

2 of 15

A09303


A09303

˚ FWHM) spectrum of electron impact dissociation of water by 200 eV Figure 1. High-resolution (1.0 A electrons. Peaks are identified in Table 1.

and De Heer [1978], Lawrence [1970], and Morgan and Mentall [1974]. The energy dependence of the excited OH fragment has been measured by Becker et al. [1980] and Beenakker et al. [1974]. In this paper we record the UV spectrum at higher spectral resolution than the work of Ajello [1984] and remeasure the energy dependence of the strongest excited fragment species cited above. [7] In this paper we have experimentally studied the collision of 25– 200 eV electrons with water molecules. A description of the instrumental setup is first given. Subsequently, we present the Doppler line profile measurements and describe the techniques used to transform line profiles into velocity distributions. High-resolution line profiles and excitation functions are obtained, from which velocity distributions of the H atoms are directly derived for the first time. A fluorescence spectrum of the OH (A X ) (0, 0) band resulting from electron impact dissociative excitation of water molecules is presented. The spectrum will allow us to estimate the internal rotational temperature of OH following prompt emission by electron impact versus the photodissociation process of water molecules, a major process in the region of cometary coma close to the nucleus. The prompt emission by OH is observed to date in only one comet [Budzien and Feldman, 1991]. The OH (A X ) emission is especially important as it directly traces the parent water molecule properties. The paper continues with a presenta˚ ), tion of the excitation functions of H Ly-a, O I (1302 A and OH (A1) from dissociative excitation of water. A general discussion of the results concludes the paper.

2. Experimental Setup and Results [8] The details of the experimental apparatus have been described earlier [Liu et al., 1995; Ajello et al., 1996]. In brief, the apparatus consists of an electron-impact collision chamber in tandem with a high-resolution Acton 3.0-m UV spectrometer. The UV line profiles of atomic oxygen and

hydrogen resulting due to the dissociative excitation of H2O are measured by crossing a magnetically collimated beam of electrons with a beam of water molecules formed by a capillary array at 90. The emitted photons were detected by the channel electron multiplier (CEM) coated with CsI to increase the quantum efficiency in the region above ˚ . The incident electron energy was kept fixed at 1100 A either 25, 35, or 100 eV during the line profile measurements. The spectra were measured at 30, 100, and 200 eV electron energies. The excitation functions (cross sections) were measured at fixed wavelength from threshold up to 600 eV electron energy. For excitation function measure˚ ments the spectral width of the spectrometer was set to 6 A FWHM by setting the entrance and exit slits to 2 mm. A resolving power of 46,000 (34,000) is achieved by operating the spectrometer in third (second) order. Owing to restriction of the spectrometer that allowed a maximum grating rotation angle of 8, the maximum wavelength that ˚ . Thus the H Ly-a and O I could be measured was 3700 A ˚ ) line profiles of oxygen could be measured in third (1152 A ˚ ) line profile was measured in second order. The O I (1302 A order. The line shapes were measured under experimental conditions that ensured the linearity of signal with electron beam current and gas pressure. The spectra were measured in the crossed beam mode, while the cross sections were measured in the static gas mode. [9] The spectra of electron-impact dissociation of water are shown in Figures 1 and 2 at electron impact at 200 eV and 100 eV, respectively. The FWHM of the spectrum in ˚ , which was achieved by setting the width Figure 1 is 1.0 A of the entrance and exit slits of the spectrometer to 340 mm. The spectra were modified in order to correct for nonuniform sensitivity of the system consisting of a grating and channeltron. The relative calibration curve, which is shown in Figure 3, was obtained by measuring the intensities of the emission from electron impact dissociation of argon in ˚ range and molecular hydrogen in 800– 1700 A ˚ 500– 950 A

3 of 15

A09303


A09303

˚ FWHM) spectrum of electron impact dissociation of water by 100 eV Figure 2. High-resolution (0.5 A electrons. Featured are the prominent emission peaks of O I and H I. The spin-forbidden doublet features ˚ are multiplied by a factor of 80 for better visualization. of O I near 1356 A range and comparing them with known (Ar) and model (H2) intensities [Ajello et al., 1988]. The two models ratios of ˚ . The resulting intensities were normalized at 800– 950 A data was fitted with a sixth degree polynomial. The fit is shown in Figure 3. ˚ [10] The emission peaks appearing in 500 to 1400 A wavelength region in Figure 1 are multiplets of O II and O I, mostly from dissociative excitation of H2O, together with the Lyman series of hydrogen emission np 1s, n = 2 8. Some of the features of O II appear in the Figure 1 in second order and they are also identified. [11] Figure 2 shows the relative calibrated spectrum of water at 100 eV electron impact energy featuring emission lines which are the focus of this investigation. These features are taken with twice the resolution of that of ˚ . Shown are a singlet Figure 1, that is, with FWHM = 0.5 A 1 1 o ˚ , an H Ly-a line at line O I ( D2 D2) at 1152.151 A ˚ , triplet lines of O I near 1302 A ˚ , and doublet 1215.672 A ˚ . The lines of spin-forbidden transitions of O I near 1356 A spin-forbidden doublet features are plotted after multiplying by a factor of 80 for better visualization. Finally, all features ˚ feature were normalized to unit peak value of the 1152 A (arbitrary units), and the peak of H Ly-a was reduced by a factor of 35 to fit in the plotting area. [12] Detailed specification of peaks in Figures 1 and 2 are given in Table 1, which also lists cross sections of featured emission lines of water. The cross sections were put on an absolute scale based on H Ly-a cross section for electron impact dissociation of water at 200 eV of Ajello [1984]. The H Ly-a cross section result of Ajello [1984], which was based on the Shemansky et al. [1985] calibration standard

of H Ly-a from H2, was revised by the ratio 7.16/8.18 using the recent results of Liu et al. [1998]. This correction results in a new value of the H Ly-a cross section of 5.5 1018 cm2 at 200 eV impact energy. The cross sections for 100 eV electrons were obtained by summing individual resolved peaks in Figure 2 and were put on an absolute scale based on the cross section of H Ly-a line at 200 eV and the ratio of 100-eV to 200-eV cross sections of H Ly-a of 1.28 obtained from the excitation function measurements for Ly-a presented later in this work.

Figure 3. Relative inverse sensitivity of the 3 m spectrometer with Channel Electron Multiplier detector; sixth degree polynomial fit of the experimental data.

4 of 15


A09303

Table 1. Features of Emission Spectrum of Water From 400 to ˚ Resulting From Dissociative Excitation Upon Electron 1400 A Impact at 100 eV and 200 eV Feature Number 1 2 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20, 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37

Wavelength ˚ Range, A 537.83 – 538.32 539.09 – 539.85 555.06 – 555.12 600.58 – 600.59 616.30 – 617.06 644.15 – 644.16 672.95 – 673.77 718.46 – 718.61 796.63 – 796.68 832.76 – 834.47 877.80 – 879.55 922.01 926.226 930.748 935.192 937.800 948.69 – 950.73 949.74 481.59 – 481.76 (second order) 972.54 971.74 – 973.89 988.58 – 990.80 990.50 499.88 – 500.35 (second order) 1025.72 1025.76 – 1028.16 1039.23 – 1041.69

s, in 1019 cm2 Transition(s) O II(2s 2Do 2p 2P) O II(2p 4So 3s 4P) O II(2p 2Do 3s 2D) O II(2p 2Po 3s 2D) O II(2p 2Do 3s 2P) O II(2s 2Po 2p 2S) O II(2p 2Po 3s 2P) O II(2s 2Do 2p 2D) O II(2s 2Po 2p 2D) O II(2s 4So 2p 4P) O I(2p 3P 3s 3Po) O I(2p 1D 3d 1Fo) H I(1s 2S 8p 2Po) H I(1s 2S 7p 2Po) O I(2p 1D 4s 1Do) H I(1s 2S 6p 2Po) O I(2p 3P 5d 3Do) H I(1s 2S 5p 2Po) O II(2p 2Do 3d 2D) H I(1s 2S 4p 2Po) O I(2p 3P 4d 3Do) O I(2p 3P 2s 3Do) O I(2p 1D 3s 1Po) O II(2p 2Po 4s 2P)

H I(1s 2S 3p 2Po) O I(2p 3P 3d 3Do) O I(2p 3P 4s 3So) second order of feature 1 1083.13 – 1086.03 O II(2p2 4P 2s3p 4So) second order of feature 2 1128.07 – 1132.97 O II(2p2 4P 2s3p 4Po) 1152.151 O I(2p 1D 3s 1Do) second order of feature 4 1215.672 H I(1s 2S 2p 2Po) 1217.65 O I(2p 1S 3s 1Po) second order of feature 5 second order of feature 6 1302.17 – 1306.03 O I(2p 3P 3s 3So)

1355.60 – 1358.51

second order of feature 7 O I(2p 3P 3s 5So)

200 eV 100 eV 0.31 0.20 0.12 0.43 0.05 0.08 0.53 0.04 0.59 0.26 0.07 0.08 0.18 0.08 0.39 0.85

1.8 1.6

7.0 0.15 0.05 0.02 1.9

1.9

55

70

2.3

1.7a 1.0a 0.3a 0.7b

a

The data for 100 eV electron impact energy show the cross sections for individual resolved lines of the triplet emission. b This is the approximate value based on the comparison of measured ˚ emission (see partial cross section and the full cross section of the 1356 A text).

˚ ) emission is not a prompt emission, [13] The O I (1356 A and an unknown fraction of excited atoms drifts away from the field of view (FOV) of the spectrometer resulting in reduced signal. The radial outward diffusion of excited atoms prevented us from putting the cross section of the ˚ ) emission on the absolute scale. However, one O I (1356 A ˚ ) emission can estimate the cross section of the O I (1356 A multiplet based on our earlier work on electron impact on oxygen molecules, results of which have appeared elsewhere [Kanik et al., 2003]. In that work we measured the fraction escaping from the 3 m spectrometer FOV at 100 eV by normalizing the measurement of the partial cross section ˚ ) to the full absolute cross section measured of O I (1356 A in a large chamber with a detector that accounted for

A09303

photons emitted at large distance (2 m) from the electron beam. The comparison yielded the fraction of the reciprocal of 80, that is, 1/80 of the oxygen atoms in a 5So2 state emitted photons while in the FOV of the spectrometer. Also, the velocity distribution of oxygen atoms in a 5So2 state had a peak at 1.2 eV. In this work it is estimated that the velocity distribution of oxygen atoms results in the same peak, around 1.25 –1.75 eV. Consequently, the fraction of the oxygen atoms in a 5So2 state that emits photons, while in the FOV of the spectrometer, is also expected to be 1/80. The cross section of the O I(2p 3P 3s 5So) transition was ˚ fine calculated by multiplying the area of the 1356 A structure peaks in Figure 2 by 80 and comparing the result with the area and cross section of the O I(2p 3P 3s 3So) transition at 100 eV impact energy. The approximate value ˚ was calculated to be 7 1020 cm2. of the 1356 A [14] The high-resolution experimental line profiles for H Ly-a emission from dissociative excitation and dissociatsive ionization of H2O are shown in Figure 4 at three electron impact energies: 25 eV (Figure 4a), 35 eV (Figure 4b), and 100 eV (Figure 4c). Figure 4 also shows the profiles of the slit function taken with the same resolving power as the line profile. The slit function was taken by illuminating the spectrometer in zeroth order. The resulting data, which is shown in Figure 4 as crosses, exhibit very good signal-tonoise (S/N) ratio, S/N > 10 in counting statistics for both peak values and measured data, and was not further filtered. The experimental data, which is shown in Figure 4 as triangles, show more scatter as the energy of the impacting electron decreases. The line profiles of H Ly-a produced from water by electron impact were measured for the first time and show good signal-to-noise ratio in general. [15] Figure 5a shows the high-resolution line profile and associated slit function profile for the O I (2p 3P2 3s 3So1) ˚ measured in second order emission feature at 1302.168 A resulting from the electron impact dissociative excitation and ionization of water by 100 eV electrons. Figure 5b shows the 100-eV electron impact induced line profile of O I ˚ measured in third order (2p 1D 3s 1Do) at 1152.15 A produced in dissociative excitation of H2O. The associated ˚ (24 mA ˚ ) in second FWHM of slit function was 32 mA (third) order. The apparent FWHM of the line profiles was ˚ for O I (1302 A ˚ ) and 30 mA ˚ for only slightly larger, 33 mA ˚ ). The reason why the measured line profile O I (1152 A for oxygen atoms shows the same magnitude of the FWHM as the slit function is that the oxygen atoms acquire very little velocity when water molecule breaks up into one oxygen and two hydrogen atoms. From conservation of momentum and energy, it follows that for the most favorable case of 100 eV electron impact energy, the energy apportioned to the oxygen atoms is 0.25 times the energy of the individual hydrogen atoms. In the next section we will show that the peak energy of the hydrogen atom reaches 5 – 7 eV, which leaves 1.25– 1.75 eV for the oxygen atom. A value of 1.5 eV peak energy of the velocity distribution of ˚ FWHM of Doppleroxygen atom translates into 15 mA shifted measured line profile. Assuming that the measured line profile, slit function, and true line profile are all Gaussian, pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffithe measured line profile then is equal to ˚ for O I (1302 A ˚ ), which is close to 322 þ 152 = 35 mA pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ˚ the measured line profile of 33 ± 2 mA, and 242 þ 152 = ˚ for O I (1152 A ˚ ), which is close to 30 ± 2 mA ˚ . The 28 mA

5 of 15

A09303


A09303

measurement [Ajello, 1984] of the H Lyman-a cross section from water (22%), the uncertainty of the recent H Lyman-a reference cross section [Liu et al., 1998] from molecular hydrogen (10%), the signal statistics (1%), and pressure and electron beam current uncertainties (5%). The cumulative error in the absolute cross section of each water feature is 25%.

3. Line Profile and Kinetic Energy Distribution Analysis [17] The observed line profile is a convolution of the true emission profile and instrumental slit function. The instru˚ in third order. mental slit function has a FWHM of 24 mA The value is comparable to the FWHM values of the measured line profiles. As a result, a fast Fourier transform (FFT) technique was used to recover the actual line profile [Press et al., 1986]. Expressed mathematically, the measured line profile I(l) is given by the convolution integral Z

I ðlÞ ¼

T ðl0 ÞAðl l0 Þdl0 ;

ð1Þ

where T(l0) is the true line profile at wavelength l0 and A(l l0) is the instrumental response function. The

Figure 4. High-resolution experimental line profile ˚ (triangles spectrum of H I (1s 2S 2p 2Po) at 1215.672 A show measured data; dash-dotted line shows filtered and symmetrized data) and the instrumental slit function (plus signs show measured data; dashed line shows filtered and ˚ FWHM in third order symmetrized data) at 27 mA (a) 25 eV, (b) 35 eV, and (c) 100 eV. The step size was ˚ /channel. The background pressure was 9 2.667 mA 106 Torr with an electron beam current of 200 mA. velocity of oxygen atoms corresponding to smaller kinetic energies of hydrogen atoms, for example at 25 and 35 eV electron impact energies, is even smaller, which results in a smaller width of the oxygen line profile. Owing to this small value of the true line profile further analysis of the oxygen line profiles was not performed. [16] The principle uncertainties contributing to the experimental cross section values are the uncertainty of our 1984

Figure 5. High-resolution experimental line profile ˚ and spectrum of (a) O I (2p 3P 3s 3So) at 1302.168 A 1 1 o ˚ (b) O I (2p D 3s D ) at 1152.151 A. The background pressure was 9 106 Torr with an electron beam current of 200 mA.

6 of 15

A09303


instrumental response function is determined from the slit function shown in Figure 4. The mathematical FFT technique has been explained in our previous publications [Ajello et al., 1995a, 1995b]. We selected a step filter to remove high-frequency noise from fast Fourier transforms of the measured signal, IT, and from the instrumental response function, AT. Then each data set (measured signal and instrumental slit function) was symmetrized around the central peak by taking two points equidistant from the peak and replacing them both with their average value. (The resulting line profile in effect appears as if shifted by a fraction of the wavelength step used in the measurement and is otherwise almost indistinguishable from the initial line profile.) The low-pass step filter, FT, was then applied to the ratio AT/IT to obtain the fast Fourier transform of the true line profile, TT. The low-pass step filter was modified by a Welch window [Press et al., 1986] in order to eliminate spectral leakage into side lobes of the deconvoluted line profiles. Filtered and symmetrized data are shown in Figure 4 as smoothed lines for measured line profiles (dash-dotted line) and slit functions (dashed line). [18] The S/N ratio at the peak of the line profile is greater than 10 for all line profiles. The true line profile, T(l), the measured line profile, I(l), at low energies, and the slit function are all approximately Gaussian in form. The square root of the sum of squares of the FWHM of the true line and the slit function should approximately be equal to the FWHM of the measured profile. This was found to be the ˚ for the H Ly-a line case with the discrepancy of 1 to 5 mA profiles. [19] The true line profiles of H Ly-a emission at three electron impact energies are shown in Figure 6. The FWHM ˚ for values of the deconvoluted true line profiles were 69 mA ˚ for 35 eV, and 136 mA ˚ for 100 eV electron25 eV, 81 mA impact energies. The wide wings of the true line profile of H Ly-a at 100 eV indicate that there is a high velocity component in the velocity distribution of hydrogen atoms. Deconvoluted data of Figure 6 was used to calculate the kinetic energy distribution of hydrogen atom fragments from electron impact dissociative excitation and ionization of water. For each of the line profiles in Figure 4, the

Figure 6. Deconvolution of the 25 eV (dashed), 35 eV (dash-dotted), and 100 eV (solid) line profiles from the measured profiles of Figure 4 using FFT technique.

A09303

Figure 7. Kinetic energy distribution of H I(2p 2Po) atoms at 25 eV (dashed), 35 eV (dash-dotted), and 100 eV (solid) from dissociative excitation of H2O. corresponding kinetic energy distribution of the fragments, P(E), is given by Pð E Þ ¼ k

dT ; dl

ð2Þ

where k is a multiplicative constant [Ogawa and Higo, 1979, 1980; Ogawa et al., 1992]. [20] The kinetic energy distributions of the H atoms for three cases of electron impact energy are shown in Figure 7 as a solid line for 100 eV electrons, a dash-dotted line for 35 eV electrons, and a dashed line for 25 eV electrons. The figure also shows the low-energy region in more detail in the insert. We identified three peaks, labeled as I, II, and III in the figure, which indicate the three components of the energy distribution functions. All three electron impact energy cases have velocity distribution components which are identified by peak I. The maxima of the distribution I, which are seen in the insert, have values lying in the range 0.3 ± 0.1 eV for the three electron energies studied, 25, 35, and 100 eV. In terms of the velocity of hydrogen atoms, the measured maximum corresponds to 6 –8 km/s. These values are the same within the measurement uncertainty of ±0.2 eV. The fact that the velocity distribution component I is present in all the three cases of electron impact energy indicates that the very low energy UV photons or lowenergy (20 eV) charged particles are sufficient to cause the fragmentation of water molecule into constituents (hydrogen atoms) with low velocity. The adiabatic thresholds are listed in Table 2. Furthermore, the 0.1 eV lowenergy component, component I, must come from the total fragmentation of H2O into 2H-O since fragmentation into OH-H produces very fast hydrogen atoms (2 eV) [Wu and Chen, 1993]. An alternative explanation is that there is a breakup of H2O into OH-H through access of optically forbidden repulsive states of H2O. However, the increase of the component I velocity distribution with increasing electron energy favors the complete fragmentation of H2O. [21] Region II in Figure 7 indicates the component of the velocity distribution of hydrogen atoms with peak values at 1.6 eV and 2.2 eV for 35 eV and 100 eV electron impact

7 of 15


A09303

A09303

Table 2. Adiabatic Threshold Energies for UV Features of H2O by Electron Impact Energy, eV Dissociation Channels 2

2

+

H* + OH(X ), OH*(A S ) H* + H(1s 2S) + O(2p 3P, 1D, 1S) H* + O(2p 3P) + H*, H+ H* + O+ + H+ O* + H2 O* + 2H(1s 2S) O* + H(1s 2S) + H+ O* + H2 O* + 2H(1s 2S) O* + H(1s 2S) + H+ OH* + H(1s 2S) OH* + H+

Theory

AP

Transitions (Wavelength)

15.31, 19.36 19.70, 21.67, 23.89 29.90, 33.30 46.92 14.55 19.03 32.62 15.79 20.27 33.86 9.16 22.76

16.5 24.5 35.0 ... 15.2 21.0, 24.0 34.5 ... ... ... 9.2 ...

H I (1s 2S 2p 2Po) ˚) (1215.672 A

energies, respectively. The velocities of hydrogen atoms that correspond to the peaks for distribution component II are 17.5 and 20.5 km/s for 35 eV and 100 eV electron impact cases, respectively. The 25 eV impact energy case does not show a clearly identified distribution component II in the 1.6– 2.2 eVenergy range. (The peak, which appears at 3.5 eV for 25 eV energy electrons, is most likely the result of applying FFT techniques to a set of data with finite size and therefore was excluded from data analysis. The same considerations were applied for 35 eV data for peaks that appear at 4.5, 8.5, and 11.5 eV.) The absence of the distribution component II in the case of 25 eV electrons indicates that the low-energy photons or charged particles do not have sufficient energy to cause the breakup of water molecules into hydrogen atoms with such large velocities. According to Wu and Chen [1993], this type of distribution reflects fragmentation into OH-H components. [22] Finally, region III on Figure 7 shows the velocity distribution component of the hydrogen atoms produced from electron impact dissociative excitation and ionization of H2O at energy 100 eV. The distribution component III appears to be very broad, at approximately 5 eV FWHM, and extends from 3 eV to 12 eV. Clearly, the distribution III appears only in the case of 100 eV electrons. The peak of the distribution III appears approximately at energy 7 eV, which corresponds to velocity 37 km/s of the hydrogen atoms. This component can be expected for dissociative ionization process listed in Table 2.

O I (2p 3P 3s 3So) ˚) (1302.17 A O I (2p 1D 3s 1Do) ˚) (1152.15 A OH (A2S+ X2) ˚) (3063.6 A

and OH A2 Sþ ! OH X 2 þ hn:

ð4Þ

The results of Zipf showed that H2O+ dissociative recombination preferentially populated the A2S+ (v = 0) level and that the hydroxyl bands exhibited a high degree of rotational excitation. [24] With an experimental setup similar to this work, the optical emission spectrum of OH(A X ) bands in the ultraviolet region from electron impact on H2O has been studied before by other groups [Mu¨ller et al., 1993; Tokue et al., 1994; Darrach and McConkey, 1991]. The broad rotational envelope found by Mu¨ller et al. [1993] is similar

˚ Band System 4. Rotational Analysis of the 3064 A of OH(A X ) [23] Figure 8 shows the comparison of our data of the ˚ system bands spectra of the OH, (A X ) (0, 0), 3064 A emission upon electron impact at 100 eV (Figure 8a) and the spectra of (0, 0) bands excited by dissociative recombination of H2O+ (E. C. Zipf, private communication, 1995). Data of this work were taken with entrance and exit slits of the spectrometer at 340 mm, which corresponded to ˚ FWHM of the instrumental slit function. It is evident, 1A by comparing the Figures 8a and 8b, that Zipf’s data has the ˚ . The data of Zipf is from same spectral resolution of 1 A dissociative recombination of H2O+ ions, that is, from the process H2 Oþ þ e ! OH A2 Sþ þ H

ð3Þ

˚ system of Figure 8. High-resolution spectrum of 3064 A ˚ OH(A(0) – X(0)): (a) this work (1 A FWHM, 100 eV electrons) and (b) microwave discharge spectrum from experimental results of E. C. Zipf (private communication, 1995).

8 of 15

A09303


to our results (see Figure 8a) showing the same high level of thermal population. They showed that a nonthermal rotational population was observed for the diatomic fragments and concluded that neither a Boltzmann distribution with a single Trot nor a biexponential distribution described the measured spectra satisfactorily. Both Mu¨ller et al. [1993] and E. C. Zipf (private communication, 1995) showed that the formation of OH(A2S+) and H(1s) ~ 1A1) state. These fragments correlate with the linear H2O(B authors concluded that the fragments are produced by optically allowed and spin-forbidden excitations of the H2O molecule. The optically allowed excitation of the bent H2O ~ 1A1 state is the main precursor (X~ 1A1) state into the linear B for the formation of highly rotationally excited OH(A2S+) fragments.

Table 3. EUV Emission Cross Sections for H2O Energy, eV

Cross Section, H I ˚ ), 1018 cm2 (1216 A

Cross Section, O I ˚ ), 1019 cm2 (1302 A

16.0 17.5 20.0 22.5 25.0 27.5 30.0 32.5 35.0 37.5 40.0 42.5 45.0 47.5 50.0 52.5 55.0 57.5 60.0 62.5 65.0 67.5 70.0 72.5 75.0 100.0 150.0 200.0 250.0 300.0 350.0 400.0 450.0 500.0 550.0 600.0

... 0.21 0.43 0.59 0.85 1.20 1.54 1.91 2.30 2.72 3.15 3.56 3.96 4.34 4.69 5.01 5.31 5.57 5.81 6.02 6.20 6.36 6.49 6.61 6.71 7.04 6.37 5.49 4.73 4.14 3.68 3.31 3.00 2.73 2.50 2.29

0.08 0.14 0.22 0.32 0.40 0.54 0.68 0.85 1.01 1.18 1.36 1.52 1.69 1.85 2.00 2.14 2.27 2.39 2.50 2.61 2.69 2.78 2.85 2.91 2.96 3.15 2.77 2.31 1.94 1.66 1.45 1.28 1.15 1.04 0.94 0.86

5. Excitation Function Analysis [25] Absolute emission cross sections have been presented for all emission features of H2O from 500 to ˚ at 200 eV electron impact energy (cf. Table 1). 1400 A All the emission features can be attributed to the process of dissociative excitation of H2O. The absolute emission cross ˚ system of OH, section is also presented for the 3064 A A2S+ X2. [26] Further measurements of the excitation function have been conducted at those wavelengths where the features are the strongest for the two types of atomic fragments (H I and O I) and also at the band head wavelength of A2S+(v0 = 0) X 2(v00 = 0) transition of ˚ . The wavelengths and spectral resolutions OH at 3064 A are chosen such that they are well resolved for each type of transition. Two energy ranges were scanned, covering the energy range 0 – 50 eV (the threshold region) and from 0 to 600 eV (the Bethe-Born region). [27] It is important to separate the underlying dissociation channels. For example, the basic processes observed in single collision of electrons and H2O leading to O I ˚ ) are as follows: Dissociative excitation (1302 A H2 O þ e ! H2 þ O* þ e;

ð5Þ

H2 O þ e ! 2H þ O* þ e:

ð6Þ

Dissociative ionization excitation H2 O þ e ! H þ Hþ þ O* þ e:

ð7Þ

Many other thresholds exist involving excited states (denoted by an asterisk) of two or more fragments. The excited states may involve rotational, vibrational, and electronic energy. In addition, fragments may be ionized. [28] Similar equations can be written for the H Ly-a ˚ ), O I (1152 A ˚ ), and for the OH excited fragments, (1216 A which were observed in this experiment. The adiabatic threshold energies for the measured UV features of water, ˚ ), namely, H I (1s 2S 2p 2Po) Lyman-a (1215.67 A ˚ , O I(2p 1D O I(2p 3P 3s 3So) multiplet at 1302.17 A ˚ , and OH (A2S+(v0 = 0) 3s 1Do) singlet at 1152.15 A 2 00 X (v = 0)), are given in Table 2 where some of the thresholds are grouped together for compactness. Also

A09303

presented in Table 2 are the observed appearance potentials (AP). A comparison of the appearance potentials with the adiabatic thresholds, which lie at lower energies, gives the identity of the process involved and the mean kinetic energies of the fragments. The change in slope of the cross-section curve with energy gives an estimate of the relative importance of each process. In Table 3 we present ˚ ) features cross-section data for the H Ly-a and O I (1302 A as a function of energy; the corresponding cross-section data ˚ system of OH(A X ) is presented in Table 4. for the 3064 A ˚ ), [29] The excitation functions of H Ly-a, O I (1302 A and OH (A2S+ (v0 = 0) X 2 (v00 = 0)) lines are shown in Figures 9– 11 on the absolute scale. The excitation functions were measured from threshold to 600 eV with 1024 steps of the impact energy for all features and were accumulated over the period of time sufficient to produce statistically significant results. The threshold region of the excitation function together with labeled appearance potentials of the thresholds are shown in inserts. Smoothed curves of experimental cross sections are also shown in these figures as solid lines. [30] Table 3 lists the absolute emission cross sections of ˚ ) and O I (1302 A ˚ ) lines at selected points H Ly-a (1216 A from 0 to 600 eV. The relative cross sections were put on an absolute scale by using the previously measured H Ly-a emission cross section from dissociative excitation of water at 200 eV of Ajello [1984]. As mentioned before in this

9 of 15


A09303

˚ System of OH for Table 4. Absolute Cross Sections of 3064 A Electron Impact on Water Cross Section, s (1018 cm2) Energy, eV

This Worka

10.0 12.5 15.0 17.5 20.0 22.5 25.0 27.5 30.0 32.5 35.0 37.5 40.0 42.5 45.0 50.0 60.0 70.0 80.0 90.0 100.0 120.0 140.0 170.0 200.0 250.0 300.0 350.0 400.0 450.0 500.0 600.0 700.0 800.0 900.0 1000.0

0.81 5.09 8.30 9.85 10.3 10.1 9.64 9.11 8.60 8.16 7.79 7.50 7.27 7.08 6.93 6.69 6.28 5.87 5.44 5.02 4.64 4.01 3.52 3.00 2.61 2.13 1.74 1.40 1.10 0.84 0.59

Beenakker et al. [1974]

6.63 6.03 5.66 5.35 5.08 4.83 4.64 4.32 4.00 3.64 3.39 2.97 2.66 2.40 2.24 2.06 1.96 1.79 1.58 1.49 1.31 1.21

a ˚ , normalized to results of This work, measurement from 3060 to 3078 A ˚ Beenakker et al. [1974] at 100 eV for full Dv = 0 sequence of the 3064 A system.

A09303

based on the internal energy distribution of OH(X2i) radicals produced at solar H Ly-a. In both cases, there are two groups of H atoms velocity distributions with 1 – 2 km/s and 15– 24 km/s mean velocities v for hydroxyl radicals and hydrogen atoms, respectively. Slow hydroxyl radicals dissociate into slow products giving rise to the slow component in the velocity distribution of hydrogen atoms (see also Keller and Thomas [1975]). We can confirm these same distributions, which we labeled I and II in Figure 7, from our studies of electron impact dissociation of H2O at 25 and 35 eV. In addition, we find a third group, labeled III, from high-energy (50 eV) electron impact dissociation of H2O. [32] In fact, the high-velocity H atoms are produced in photodissociation of H2O mainly by EUV photons (by the bright lines in the solar EUV spectrum), with a velocity as high as 40 km s1 [Wu and Chen, 1993]. Such highvelocity H atoms have indeed been observed in comets Austin (C/1989 X1) and Levy (C/1990 K1) through highspectral-resolution ground-based measurements of Ha ˚ ) line profiles [Brown and Spinrad, 1993]. Recent (6563 A ˚ ) line profile in comet measurements of H Ly-a (1216 A Hayakutake (C/1996 B2) with HST/GHRS [Combi et al., 1998] do not support the presence of high-velocity H atoms in the cometary coma. However, the noise level in the HST observations may not be as good as in the ground-based Ha observations [Brown and Spinrad, 1993], which could have resulted in nondetection of high-velocity H component in the HST data. [33] The hot H atoms produced in the dissociation of H2O by electron impact can interact with other molecules and thereby can have implications for the neutral chemistry in the cometary coma, Saturnian magnetosphere, and other water-containing planetary bodies. For example, in the case of water the reactions could be [Roessler, 1992] H þ H2 O ! OH þ H þ H

ð8Þ

H þ H2 O ! OH þ H2 :

ð9Þ

and 18

2

paper, the old value of Ajello [1984], 6.3 10 cm , was corrected based on a new value of H Ly-a emission cross section from dissociative excitation of H2O at 200 eV. Thus ˚ ) and O I (1302 A ˚) new absolute values for H Ly-a (1216 A emission cross section were obtained; for example, the emission cross sections for H Ly-a are 7.04 1018 cm2 and 5.49 1018 cm2 for 100 and 200 eV electron impact ˚ ) are 3.15 energies, respectively, and those for O I (1302 A 1019 cm2 and 2.31 1019 cm2 at the same energies.

6. Discussion [31] The velocity distributions of hydrogen atoms and hydroxyl radicals produced through solar photodissociation of water were calculated by Wu and Chen [1993]. Using the available absolute partial cross sections for the production of H and OH in photodissociation of H2O and the solar flux ˚ range, including peak fluxes at the bright in 500– 1860 A solar lines, they calculated the first moments of the velocity distribution v for H fragments to be 21.5 – 24.0 km/s for case I and 14.4– 24.0 km/s for case II. For OH fragments the values are 1.48 – 1.65 km/s for case I and 0.95 – 1.65 km/s for case II. The authors distinguish between cases I and II

[34] Equation (8) requires that the H atom energy be higher than 5.11 eV. The results of the present study has already shown (see Figure 7) that H atoms with energy >5 eV are produced in e-H2O dissociation when the impacting electron energy is ^100 eV. It is worth pointing out here that energetic (keV) electrons have been observed in the coma of comet Halley [Gringauz and Verigin, 1991; Bhardwaj et al., 1990]. Equation (9) requires that the H atom energy be higher than 0.93 eV. [35] Bo¨se and Sroka [1973] studied the dissociation process in H2O. They studied excitation functions and obtained the cross section values for several atomic emission features of H and O. For example, they obtained the H Ly-a cross section value of 7.5 1018 cm2 and the O I ˚ ) cross section value of 2 1019 cm2 at 100 eV. (1152 A The ratio of these two cross sections, 37.5:1, is close to the ratio obtained in this experiment, 36.8:1 (see features 29 and 31 in Table 1, where the H Ly-a cross section is 7.0 ˚ ) is 1.9 1018 cm2 and the cross section of O I (1152 A 1019 cm2 at 100 eV electron impact energy).

10 of 15

A09303


A09303

Figure 9. Absolute cross section of Ly-a line from dissociative excitation of water by electron impact from 15 to 600 eV impact energy (dots show experimental data; solid line shows smooth of data). The inset shows the threshold behavior from 10 to 50 eV impact energy. Appearance potentials are shown in the inset.

[36] We show in Figure 12 the difference in spectra at low-energy dissociative excitation below 40 eV and high energy greater than 40 eV that includes ionization excitation. The spectrum from pure dissociative excitation shown at 30 eV is very simple, of neutral atomic H and O and with ˚ . The spectrum from dissociative no feature below 800 A excitation and ionization at 200 eV features neutral atomic

emissions from H and O as well as O+ emission features ˚ (see the identification of H I, O I, and O II below 800 A emission features in Figure 1). [37] Absolute emission cross sections of OH(A2S+ ! ˚ system for 40– 1000 eV electron impact of X) 3064 A water were measured by Beenakker et al. [1974]. Their result was presented for the whole Dv = 0 sequence of the

˚ ) line from dissociative excitation of water by electron Figure 10. Absolute cross section of O I (1302 A impact from 12 to 600 eV impact energy (dots show experimental data, solid line shows smooth of data). The inset shows the threshold behavior from 10 to 50 eV impact energy. Appearance potentials are shown in the inset. 11 of 15

A09303


A09303

Figure 11. Absolute cross section of OH (A2S+(v0 = 0) X 2(v00 = 0)) line from dissociative excitation of water by electron impact from 0 to 600 eV impact energy. The inset shows the threshold behavior from 5 to 45 eV impact energy. The appearance potential is shown in the inset. ˚ system of OH, which extends from 3060 to 3500 A ˚. 3064 A Becker et al. [1980] also presented the apparent excitation ˚ system of function of the (0, 0) band of the 3064 A ˚ OH extending from 3061 to 3120 A. Their data was measured from threshold to 100 eV electron energies. However, it is difficult to resolve the (0, 0) band from other members of the Dv = 0 sequence in their data. They did not ˚ band. In the present work we have resolve the 3064 A

measured the excitation function of the portion of the band ˚ system of OH centered at 3070 A ˚ and head of the 3064 A ˚ limited by the bandpass of the spectrometer set to 8 A, that ˚ . However, since all ro-vibrational is, from 3062 to 3078 A levels of the Dv = 0 sequence have nearly the same threshold energy, within 0.5 eV, the excitation function of a narrow portion of the band is identical to the excitation function of the whole Dv = 0 sequence.

Figure 12. The calibrated spectra of electron impact dissociative excitation and dissociative ionization of H2O at 30 and 200 eV (FWHM = 0.5 nm). 12 of 15


A09303

A09303

˚ system of Figure 13. Comparison of the emission cross sections of the Dv = 0 sequence of the 3064 A OH (A X).

[38] Following the assumption stated above, we scaled ˚ system the excitation function of the subband of the 3064 A to the absolute scale based on the results of Beenakker et al. [1974] at 100 eV. Our results are normalized to the ˚ band. To find the excitation function of the entire 3064 A scaling factor, we measured the areas of the calibrated spectrum in two wavelength regions. We found that the area of the bandhead measured in our experiment is 14% of the entire band. Thus the scaled results of this work, obtained from the subband excitation function and the scaling factor, together with the results of Beenakker et al. [1974], are presented in Table 4. The plots of the cross sections are shown in Figure 13. [39] These results are important for the remote sensing instruments of the Cassini (Saturn) and Rosetta (comet 67 P/Churyumov-Gerasimenko) missions. The Cassini ultraviolet spectrometer and Cassini solid-state imaging system will be making observations of the Saturn’s icy satellites beginning in 2004. In addition, high-resolution line profiles of H Ly-a by Hubble Space Telescope from astronomical objects containing H2O will help distinguish electron impact dissociation from solar photodissociation.

˚ [42] 2. We have obtained medium-resolution (0.5 A ˚ , 1304 A ˚ and 1356 A ˚) FWHM) spectra of the O I (1152 A multiplets showing the relative intensities of the fine structure lines. ˚ [43] 3. We have measured the high-resolution (27 mA FWHM) line profiles of H Ly a at 25, 35, and 100 eV and have identified three distinct populations of H(2p) atoms with mean kinetic energies of approximately 0.2, 2.0, and 7 eV. [44] 4. The measurement of the high-resolution O I ˚ ) line profiles at 24 and 32 mA ˚ (1152 and 1304 A FWHM, respectively, indicate an upper limit to the mean kinetic energies of the excited O (1D and 3S) atoms of 1 eV. ˚ FWHM) of [45] 5. The medium-resolution spectrum (1 A ˚ the OH(3050 A) band at 100 eV indicates a high degree of rotational excitation, similar to that found in dissociative recombination of H2O+. ˚) [46] 6. The excitation functions of H Ly a, O I (1304 A 2 + 0 2 00 and the OH (A S (v = 0) X (v = 0)) molecular ˚ were measured from 0 to emission band near 3050 A 600 eV.

7. Conclusion

[47] Acknowledgments. The research described in this paper was carried out at the Jet Propulsion Laboratory, California Institute of Technology, and was sponsored by the U.S. Air Force Office of Scientific Research (AFOSR), the Aeronomy Program of the National Science Foundation (ATM-9320589), and NASA Planetary Atmospheres, Space Astrophysics Research and Analysis, and Space Physics Program Offices. Two of us (O.M. and P.V.) are supported by the National Research Council through a Resident Research Associateship at the Jet Propulsion Laboratory, California Institute of Technology. [48] Arthur Richmond thanks Tom Slanger and two other reviewers for their assistance in evaluating this paper.

[40] We have examined at 100 eV and 200 eV electron impact energies the UV fluorescence spectra of H2O over ˚ at low, medium, the wavelength region from 400 to 1400 A and high spectral resolutions. The salient conclusions are the following: [41] 1. We can identify 37 spectral features of atomic H, ˚ FWHM). The O, and O+ at low spectral resolution (1 A atomic emissions are produced by dissociation excitation. We present emission cross sections for each spectral feature normalized to the H Ly a cross section (7.0 1018 cm2 at 100 eV and 5.5 1018 cm2 at 200 eV), the strongest feature in the vacuum UV.

References Ajello, J. M. (1984), The EUV spectrum of H2O by electron impact, Geophys. Res. Lett., 11, 1195 – 1198.

13 of 15

A09303


Ajello, J. M., et al. (1988), Simple ultraviolet calibration source with reference spectra and its use with the Galileo orbiter ultraviolet spectrometer, Appl. Opt., 27, 890 – 914. Ajello, J. M., S. M. Ahmed, I. Kanik, and R. Multari (1995a), Kinetic energy distribution of H(2p) atoms from dissociative excitation of H2, Phys. Rev. Lett., 75, 3261 – 3264. Ajello, J. M., I. Kanik, S. M. Ahmed, and J. T. Clarke (1995b), Line profile of H Lyman a from dissociative excitation of H2 with application to Jupiter, J. Geophys. Res., 100, 26,411 – 26,420. Ajello, J. M., S. M. Ahmed, and X. Liu (1996), Line profile of H Lyman-b emission from dissociative excitation of H2, Phys. Rev. A, 53, 2303 – 2308. Avakyan, A. V., R. N. Il’in, V. M. Lavrov, and G. N. Ogurtsov (1998), Collision Processes and Excitation of UV Emission from Planetary Atmospheric Gases, p. 151, Gordon and Breach, Newark, N. J. Barth, C. A., C. W. Hord, A. I. F. Stewart, W. R. Pryor, K. E. Simmons, W. E. McClintock, J. M. Ajello, K. L. Naviaux, and J. J. Aiello (1997), Galileo ultraviolet spectrometer observations of atomic hydrogen in the atmosphere at Ganymede, Geophys. Res. Lett., 24, 2147 – 2150. Becker, K., B. Stumpf, and G. Schulz (1980), Crossed-beam investigations of the OH(A2S+ ! X2i) emission spectrum after dissociative excitation of water by electron impact, Chem. Phys. Lett., 73, 102 – 105. Beenakker, C. I. M., F. J. De Heer, H. B. Krop, and G. R. Mo¨hlmann (1974), Dissociative excitation of water by electron impact, Chem. Phys., 6, 445 – 454. Bhardwaj, A. (1999), On the role of solar EUV, photoelectrons, and auroral ˚ in electrons in the chemistry of C(1D) and the production of C I 1931 A the inner cometary coma: A case for comet P/Halley, J. Geophys. Res., 104, 1929 – 1942. Bhardwaj, A., and G. R. Gladstone (2000), Auroral emissions of the giant planets, Rev. Geophys., 38, 295 – 353. Bhardwaj, A., S. A. Haider, and R. P. Singhal (1990), Auroral and photoelectron fluxes in cometary ionospheres, Icarus, 85, 216 – 228. Bhardwaj, A., S. A. Haider, and R. P. Singhal (1996), Production and emissions of atomic carbon and oxygen in the inner coma of comet Halley: Role of electron impact, Icarus, 120, 412 – 430. Bo¨se, N., and W. Sroka (1973), Dissociative excitation processes in H2O, Z. Naturforsch., 28a, 22 – 26. Brown, M. E., and H. Spinrad (1993), The velocity distribution of cometary hydrogen: Evidence for high velocities?, Icarus, 104, 197 – 205. Budzien, S. A., and P. D. Feldman (1991), OH prompt emission in Comet IRAS-Araki-Alcock 1983 VII, Icarus, 90, 308 – 318. Budzien, S. A., M. C. Festou, and P. D. Feldman (1994), Solar flux variability and the lifetimes of cometary water and OH, Icarus, 107, 164 – 188. Combi, M. R., M. E. Brown, P. D. Feldman, H. U. Keller, R. R. Meier, and W. H. Smyth (1998), Hubble Space Telescope ultraviolet imaging and high-resolution spectroscopy of water photodissociation products in comet Hyakutake (C/1996 B2), Astrophys. J., 494, 816 – 821. Cravens, T. E., J. U. Kozyra, A. F. Nagy, T. I. Gombosi, and M. Kurtz (1987), Electron impact ionization in the vicinity of comets, J. Geophys. Res., 92, 7341 – 7353. Crovisier, J. (1989), The photodissociation of water in cometary atmospheres, Astron. Astrophys., 213, 459 – 464. Darrach, M., and J. W. McConkey (1991), Rotationally resolved laserinduced fluorescence studies of OH (X, v = 0) produced by electronimpact dissociation of jet-cooled H2O and CH3OH, Chem. Phys. Lett., 184, 141 – 146. Feldman, P. D., M. A. McGrath, D. F. Strobel, and H. W. Moos (2000), HST/STIS ultraviolet imaging of polar aurora on Ganymede, Astrophys. J., 5353, 1085 – 1090. Festou, M. C. (1981), The density distribution of neutral compounds in cometary atmospheres, Astron. Astrophys., 95, 69 – 79. Festou, M. C., E. B. Jenkins, H. U. Keller, E. S. Barker, J. L. Bertaux, J. F. Drake, and W. L. Upson (1979), Lyman-alpha observations of comet Kobayashi-Berger-Milon (1975 IX) with COPERNICUS, Astrophys. J., 232, 318 – 328. Festou, M. C., H. Richman, and R. M. West (1993), Comets, Astron. Astrophys. Rev., 4, 363; 5, 37. Frank, L. A., W. R. Paterson, K. L. Ackerson, and S. J. Bolton (1997), Outflow of hydrogen ions from Ganymede, Geophys. Res. Lett., 24, 2151 – 2154. Gan, L., and T. E. Cravens (1990), Electron energetic in the inner coma of comet Halley, J. Geophys. Res., 95, 6285 – 6303. Gringauz, K. I., and M. I. Verigin (1991), Permanent and nonstationary plasma phenomena in comet Halley’s head, in Cometary Plasma Processes, Geophys. Monogr. Ser., vol. 61, edited by A. D. Johnstone, pp. 107 – 116, AGU, Washington, D. C. Haberli, R. M., K. Altwegg, H. Balsiger, and J. Geiss (1996), Heating of the thermal electrons in the inner coma of comet P/Halley, J. Geophys. Res., 101, 15,579 – 15,589.

A09303

Hall, D. T., D. F. Strobel, P. D. Feldman, M. A. McGrath, and H. A. Weaver (1995), Detection of an oxygen atmosphere on Jupiter’s moon Europa, Nature, 23, 677 – 679. Hall, D. T., P. D. Feldman, J. B. Holberg, and M. A. McGrath (1996), Fluorescent hydroxyl emissions from Saturn’s ring atmosphere, Science, 272, 516 – 518. Hall, D. T., P. D. Feldman, M. A. McGrath, and D. F. Strobel (1998), The far-ultraviolet oxygen airglow of Europa and Ganymede, Astrophys. J., 499, 475 – 481. Hubner, W. F. (Ed.) (1990), The Physics and Chemistry of Comets, Springer-Verlag, New York. Huppler, D., R. J. Reynolds, F. L. Roesler, F. Schlerb, and J. Trauger (1975), Observations of Comet Kohoutek (1973f) with a ground-based Fabry-Perot spectrometer, Astrophys. J., 202, 276 – 282. Itikawa, Y., S. J. Kim, Y. H. Kim, and Y. C. Minh (2001), Molecular emissions from the atmospheres of giant planets and comets: Needs for spectroscopic and collision data, Adv. Atom. Molec. Phys., 47, 129 – 162. Johnson, R. E. (1990), Energetic Charged Particle Interactions With Atmospheres and Surfaces, Springer-Verlag, New York. Kanik, I., C. Noren, O. Makarov, P. Vatti Palle, J. Ajello, and D. Shemansky (2003), Electron impact dissociative excitation of O2: 2. Absolute emission cross sections of OI (130.4 nm) and OI (135.6 nm), J. Geophys. Res., 108(E11), 5126, doi:10.1029/2000JE001423. Keller, H. U., and G. E. Thomas (1975), A cometary hydrogen model: Comparison with OGO-5 measurements of comet Bennett (1970II), Astron. Astrophys., 39, 7 – 19. Kerr, R. B., C. A. Tepley, R. P. Cageao, S. K. Atreya, T. M. Donahue, and I. M. Cherchneff (1987), Observations of comet Halley at H a and ˚ , Geophys. Res. Lett., 14, 53 – 56. 6300 A Kumar, S., and D. M. Hunten (1982), The atmospheres of Io and other satellites, in Satellites of Jupiter, edited by D. Morrison, pp. 782 – 806, Univ. of Ariz. Press, Tucson, Ariz. Lawrence, G. M. (1970), Dissociative excitation of some oxygen-containing molecules: Lifetimes and electron impact cross sections, Phys. Rev. A, 2, 397 – 407. Liu, X., S. M. Ahmed, R. A. Multari, G. K. James, and J. M. Ajello (1995), High-resolution electron-impact study of the far-ultraviolet emission spectrum of molecular hydrogen, Astrophys. J., 101, suppl., 375 – 399. Liu, X., D. E. Shemansky, S. M. Ahmed, G. K. James, and J. M. Ajello (1998), Electron-impact excitation and emission cross sections of the H2 Lyman and Werner systems, J. Geophys. Res., 103, 26,739 – 26,758. Majeed, T., and J. C. McConnell (1991), The upper ionospheres of Jupiter and Saturn, Planet. Space Sci., 39, 1715 – 1732. Mohlmann, G. R., and F. J. De Heer (1978), Production of H, D (2, 2p) by electron impact (0 – 2000 eV) on simple hydrogen containing molecules, Chem. Phys., 28, 331 – 341. Morgan, H. D., and J. E. Mentall (1974), VUV dissociative excitation of H2O, NH3, CH4 by electron impact, J. Chem. Phys., 60, 4734 – 4739. Moses, J. I., and S. F. Bass (2000), The effects of external material on the chemistry and structure of Saturn’s ionosphere, J. Geophys. Res., 105, 7013 – 7052. Mu¨ller, U., T. Bubel, and G. Schulz (1993), Electron impact dissociation of H2O: Emission cross sections for OH*, OH+*, H*, and H2O+* fragments, Z. Phys. D, 25, 167 – 174. Ogawa, T., and M. Higo (1979), Analysis of the translational energy distribution of H* produced in the e – H2 collision, Chem. Phys. Lett., 65, 610 – 612. Ogawa, T., and M. Higo (1980), Translational energy distribution and production mechanism of excited hydrogen atoms produced by controlled electron impact on H2, Chem. Phys., 52, 55 – 64. Ogawa, T., S. Ihara, and K. Nakashima (1992), Fano plots for the slow and fast groups of excited hydrogen atoms produced in e – H2 collisions, Chem. Phys., 161, 509 – 513. Press, W. H., B. P. Flannery, S. A. Teukolsky, and W. T. Vetterling (1986), Numerical Recipes—The Art of Scientific Computing, 818 pp., Cambridge Univ. Press, New York. Richardson, J. D. (1998), Thermal plasma and neutrals gas in Saturn’s magnetosphere, Rev. Geophys., 36, 501 – 524. Roessler, K. (1992), Hot atom chemistry in space, in Handbook of Hot Atom Chemistry, edited by T. Matsura, Kodansha, Tokyo. Shemansky, D. E., and D. T. Hall (1992), The distribution of atomic hydrogen in the magnetosphere of Saturn, J. Geophys. Res., 97, 4143 – 4161. Shemansky, D. E., J. M. Ajello, and D. T. Hall (1985), Electron impact excitation of H2: Rydberg band systems and the benchmark dissociative cross section for H Lyman-alpha, Astrophys. J., 296, 765 – 773. Shemansky, D. E., P. Matheson, D. T. Hall, H.-Y. Hu, and T. M. Tripp (1993), Detection of the hydroxyl radical in the Saturn magnetosphere, Nature, 363, 329 – 331.

14 of 15

A09303


Sieger, M. T., W. C. Simpson, and T. M. Orlando (1998), Production of O2 on icy satellites by electronic excitation of low-temperature water ice, Nature, 394, 554 – 556. Sill, G. T., and R. N. Clark (1982), Composition of the surfaces of the galilean satellites, in Satellites of Jupiter, edited by D. Morrison, pp. 174 – 212, Univ. of Ariz. Press, Tucson, Ariz. Tokue, I., M. Kobayashi, S. Suzuki, and Y. Ito (1994), Rotational distributions of OH(A2S+) produced by electron-impact dissociation of jet-cooled H2O and CH3OH, Chem. Phys., 179, 503 – 511. Weiser, H., R. C. Vitz, and H. W. Moos (1977), Detection of Lyman a emission from the Saturnian disk and from the ring system, Science, 197, 755 – 757. Wilkening, L. L., (Ed.) (1982), Comets, 766 pp., Univ. of Ariz. Press, Tucson, Ariz.

A09303

Wu, R., and F. Z. Chen (1993), Velocity distribution of hydrogen atoms and hydroxyl radicals produced through solar photodissociation of water, J. Geophys. Res., 98, 7415 – 7435.

J. M. Ajello, I. Kanik, O. P. Makarov, and P. Vattipalle, Jet Propulsion Laboratory, California Institute of Technology, 4800 Oak Grove Rd., M/S 183-601, Pasadena, CA 91109, USA. ( [email protected]; isik. [email protected]; [email protected]; prahlad.vattipalle@jpl. nasa.gov) A. Bhardwaj, NASA Marshall Space Flight Center, NSSTC, SD50, Huntsville, AL 35805, USA. ([email protected]) M. C. Festou, Observatoire Midi-Pyreńeés, Av. Edouard Belin 14, F-31400, Toulouse, France. ([email protected])

15 of 15