Hatano Y., Katsumura Y., Mozumder A. (Eds.) Charged Particle and Photon Interactions with Matter - Recent Advances, Applications, and Interfaces

Подождите немного. Документ загружается.

761

UV Molecular Spectroscopy

from

Electron Impact for

Applications

to Planetary

Atmospheres

and Astrophysics

Joseph M. Ajello

California Institute of Technology

Pasadena,

California

Rao S. Mangina

California Institute of Technology

Pasadena,

California

Robert R. Meier

George Mason University

Fairfax,

Virginia

28.1 introduCtion

In the upper atmospheres and torus regions of the terrestrial and Jovian planets and in the inter-

stellar medium (ISM), an important mechanism for energy transfer and diagnostic spectroscopy

is electron collision processes with both neutral and ionic species leading to the emission of

electromagnetic radiation. Six of the planets (Earth, Mars, Jupiter, Saturn, Uranus, and Neptune)

are known to have internal magnetic elds that lead to particle acceleration and energy deposi-

tion into a planetary atmosphere (Bagenal etal., 2007). Mercury also has an intrinsic magnetic

Contents

28.1 Introduction .......................................................................................................................... 761

28.2 UV

Spectroscopy of Molecules in Planetary Atmospheres ................................................. 763

28.3

Apparatus

and Experimental Methods................................................................................. 767

28.4

Present

Status of H

, N

, CO, and SO

.................................................................................772

28.4.1 H

–UV ...................................................................................................................... 772

28.4.2 H

–VOIR................................................................................................................... 776

28.4.3 N

–EUV....................................................................................................................780

28.4.4 N

–FUV.................................................................................................................... 786

28.4.5 CO.............................................................................................................................792

28.4.6 SO

............................................................................................................................793

28.5 Conclusions........................................................................................................................... 796

Acknowledgments..........................................................................................................................797

References...................................................................................................................................... 797

762 Charged Particle and Photon Interactions with Matter

eld with a tenuous atmosphere that is a planetary exosphere. The ubiquitous presence of ener-

getic electron-excited ultraviolet (UV) dayglow and aurora in the solar system (Broadfoot etal.,

1979, 1981a,b, 1989; Sandel etal., 1979; Yung etal., 1982; Meier, 1991; Ajello etal., 1998b, 2001,

2005a; Gustin etal., 2002, 2004) has been studied spectroscopically over the last 30 years using

observations from interplanetary spacecraft beginning with the Voyager Grand Tour mission and

Mars and Venus Mariner missions, and the earth-orbiting satellites beginning with the Orbiting

Geophysical Observatories (OGOs) and International Ultraviolet Explorer (IUE). Simultaneously,

astronomical observations of H

Rydberg band emissions from Herbig-Haro and T-Tauri stellar

objects (Raymond etal., 1997; Herczeg etal., 2002, 2004; Bergin etal., 2004) and models of H

Rydberg band emissions generated in the interior of molecular clouds within the ISM (Gredel

etal., 1987, 1989; Liu and Dalgarno, 1996) have been achieved, rst with the Copernicus, which

was equipped with a UV telescope (Grewing etal., 1978, Snow, 1979), and followed at higher

spectral resolution by the Hubble Space Telescope (HST) with its corrected optics (Petersen and

Brandt, 1995).

The importance of electron impact excitation of molecules was dramatized during the Voyager

mission due to ndings of rich dayglow and auroral spectra at each of the outer planets with strong

magnetospheres and thick H

atmospheres. For example, the Voyager 1 spacecraft equipped with the

ultraviolet spectrometer (UVS), having a detection range of 50–170nm, arrived at Jupiter in January

1979 and at Saturn in November 1980, and made spectacular discoveries related to the physical

processes involving electron acceleration and corotating plasma that control the atmosphere and

magnetosphere (Broadfoot etal., 1979, 1981a,b, 1989; Sandel etal., 1979). Voyager discovered that

Jupiter’s UV aurora is the second brightest source of UV in the solar system (second to the Sun),

with an emission intensity of ∼10

13−14

W in the H

Rydberg bands. Its ultimate power source is the

rotational energy of the planet as well as plasma processes in the near corotating middle magneto-

sphere.

The breadth of objects studied with these telescopes is shown in Figure 28.1.

the early 1980s, the initial attempts in modeling the extreme ultraviolet (EUV) auroral spec-

tra obtained by Voyager and other missions were not very successful due to lack of spectroscopic

signatures and their reliable cross section data. At this point in time, reliable electron excitation

cross sections can only be provided through laboratory measurements, but seldom by theory. Useful

earlier reviews can be found for terrestrial auroral spectroscopy in Auroral Physics (Meng etal.,

1991), and for emission cross section measurements prior to 1998 for some of the planetary species

Avakyan etal., (1998).

The

analysis of observations of planetary atmospheres made by HST, Cassini and Far Ultraviolet

Spectroscopic Explorer (FUSE), and Earth-orbiting spacecraft Thermosphere Ionosphere

Mesosphere Energetics and Dynamics/Global Ultraviolet Imager (TIMED/GUVI), Midcourse

Space Experiment (MSX), Defense Meteorological Satellite Program (DMSP), POLAR spacecraft,

and Imager for Magnetopause-to-Aurora Global Exploration (IMAGE) require accurate collision

Figure 28.1 Typical types of objects that are studied by spacecraft equipped with UV observatories are

the

Earth (Terrestrial Objects), Jupiter (Jovian planets) and the Horse Head nebula (ISM).

UV Molecular Spectroscopy from Electron Impact for Astrophysics 763

cross sections. With the advent of the newest generation of high-resolution UV imaging space

instruments (e.g., Space Telescope Imaging Spectrograph (STIS) and Faint Object Spectrograph

(FOS) with UV spectral resolving power ≈10

onboard HST, the emissions of the outer planets

have been examined in much greater detail at better resolution and in different spectral regions

than Voyager.

28.2 uv speCtrosCopy oF moleCules in planetary atmospheres

To illustrate the key relationships among spectroscopic observations, molecular parameters,

and planetary constituents, we present the formalism for the Jovian aurora. Airglow processes

are similar except that precipitating particle uxes are replaced by photoelectron uxes. A

schematic of the Jovian magnetic eld interacting with the ionosphere is shown in the upper

panel of Figure 28.2 along with a model of the primary + secondary differential electron ux,

(electrons/s/cm

/eV) at various altitudes in the Jovian aurora shown in the lower panel (Ajello

etal., 2001, 2005a; Grodent etal., 2001). The auroral UV spectrum produced by the precipi-

tating electrons can be used to estimate Q, the precipitating electron energy ux of the pri-

mary particles, and E

, their characteristic energy (Strickland etal., 1995). The modeling of the

dynamical magnetosphere–ionosphere coupling producing the Jovian aurora has been recently

described (Bunce and Cowley, 2001; Cowley and Bunce, 2001; Hill, 2001). These authors sug-

gest that the auroral oval indicates the presence of a global-scale Birkeland current system that

maps to ∼30R

. This current system passes planetary angular momentum to the outward mov-

ing plasma sheet maintaining the middle magnetosphere in near corotation (Hill, 2001). In the

region of upward currents (downward electrons), eld-aligned potentials accelerate electrons to

3000 km

Primary

300 km

Differential flux

(electron/cm

/s/eV)

Energy (eV)

(a)

(b)

Figure 28.2 (a) The model of the Jupiter ionosphere–magnetosphere coupling showing the currents for the

aurora (Cowley and Bunce, 2001; Hill, 2001), and (b) the electron differential energy distribution of primary +

secondary electrons at four altitudes in the Jupiter aurora are distinguished by dashed or solid curves (Ajello

etal.,

2001, 2005a). The unmarked dotted curve is at 800

km.

764 Charged Particle and Photon Interactions with Matter

auroral energies of 10–100 keV (Mauk etal., 2002), which in turn generate UV and visible aurora

through electron–molecule collisions that are observed by spacecraft at the Jovian planets.

The volume emission rate, V

(z) (photons/s/cm

), for a simple molecule H

spectral line at wave-

length,

λ, emitted in the aurora of the Jovian planets can be written (Ajello etal., 2001, 2005a) as

V z g z T

j jm jλ

( ) ( ) ( ) )= −ω λ η(1 ( ) ,N z (28.1)

where g

(z) is the excitation rate (excitations per molecule per second) at altitude z into an upper

electronic state j characterized by α, v, and J which are principal quantum number, vibrational

level quantum number, and rotational level quantum number, respectively. The index m refers to

the quantum numbers β (principle), v (vibrational), and J (rotational) for the lower electronic state.

N(z,T) is the atmospheric H

density at altitude z with a local kinetic temperature T. The quantity η

is the nonradiative yield of predissociation and pre-ionization for the upper state. ω

(λ) is the emis-

sion

branching ratio for the transition j → m at wavelength λ and is given by

ω λ

( ) = (28.2)

where

is the Einstein spontaneous transition probability from the upper state j to a lower state m

is the total emission transition probability to all lower states

The excitation rate, g

(z), is proportional to the sum of the individual excitation rates from a

ground state rotational-vibrational level, X(v

). The individual excitation rates equal to the product

of the fraction of molecules in the initial ground state rotational-vibrational level, X(v

), times the

ne structure excitation cross section σ

from level i to level j for an electron impact energy, ε, times

the

electron ux. It is written as

g z F z f T d

j X

i ij

( ) , ) ( ) ( ,= )













∑

∫

(ε σ ε ε

(28.3)

where f

is the fraction of the H

molecules in the initial ground vibrational level v

and rotational

level J

(see Ajello et al., 2001, 2005a for notation). The sum extends over the ne structure rota-

tional branches for Σ and Π transitions of H

in an upper atmosphere. F(ε, z) is the precipitating

electron

ux distribution function in electrons/sec/cm

/eV at an energy ε and altitude z.

The rst step in modeling the Jovian aurora requires careful laboratory studies to determine

σ ε

( )

for each rotational line following electron impact excitation of H

(Liu etal., 1998; Glass-

Maujean etal., 2009). UV emissions from the outer planets observed by Voyager and IUE were

explained with electron transport models of high-energy electron impact (1–100keV primary elec-

tron ux at the top of the atmosphere) of H

that included all Rydberg states from principal quantum

number, n = 2–5 (Yung etal., 1982; Shemanksy and Ajello, 1983; Ajello etal., 1984). Modeling to

date (Jonin etal., 2000; Liu etal., 2000, 2002, 2003; Glass-Maujean etal., 2009) now takes into

account the rotational-electronic coupling among the n = 2 through 5 states, ungerade B, B′, B″, C,

D, D′, D″ (

) n = 2,3,4,5 → X

and can be used for accurate modeling of the aurora spectra.

We show in Figure 28.3a the rst model analysis (Yung etal., 1982) (Equations 28.1 through

28.3) of the Jupiter auroral far ultraviolet (FUV) spectrum from IUE. A comparison of this spectrum

obtained at 0.1nm resolution with a model provided convincing proof that most of the emission

features come from H

(Yung etal., 1982). Comparison of a Saturnian auroral spectrum obtained

by Voyager in the EUV at 3.3nm resolution with an auroral model is shown in Figure 28.3b that all

of the features could be identied with molecular hydrogen as well (Shemansky and Ajello, 1983).

Gustin et al. (2009) has recently analyzed the Cassin ultraviolet imaging spectrograph (UVIS)

UV Molecular Spectroscopy from Electron Impact for Astrophysics 765

auroral spectrum of Saturn at much higher resolution of 0.4nm FWHM. The IUE analysis provided

the magnitude of both E

and Q. The characteristic energy for the primary electron ux was esti-

mated to be ∼10keV. A typical characteristic energy ranging from 5 to 100keV was found by the

UVS in the Galileo orbiter mission to Jupiter (Ajello etal., 1998b, 2001, 2005a) and subsequently by

Cassini and HST analyses (Dols etal., 2000; Gustin etal., 2002, 2004; Grodent etal., 2003a,b; Ajello

etal., 2005a). The globally averaged electron energy ux precipitating into the Jupiter atmosphere

was 0.5–2erg/cm

/s at the time of the IUE observation (Yung etal., 1982).

Relative intensity

V1 Saturn aurora

darkside n.pole

Observed data

H+e

Lβ

Model

e+H

(all bands)

& e+H

Model

e+H

(B, C)

& e+H

Wavelength (Å)

600 800 1000 1200 1400 1600

1 keV

IUE

1 keV

IUE data

10 keV

100 keV Aurora

100 keV

1150

(a)

(b)

1200 1250 1300 1350 1400 1450

Wavelength (Å)

1500 1550 1600 1650

10 keV

Figure 28.3 (a) IUE observation of Jupiter with three auroral models (1, 10, 100keV primary electrons) in

the FUV (Yung etal., 1982). (b) Shown in dash are the Voyager 1 and 2 observation of prominent UV radiation

from the atmosphere of Saturn corresponding to auroral emission, which is concentrated in the polar regions,

and is excited by high-energy particle precipitation along the magnetic eld lines (Shemansky and Ajello,

1983). The strong disk and auroral feature at 1216Å, which runs off scale, is the hydrogen Ly α line. The H

band emission between 900 and 1130Å is the characteristic of the auroral region. We indicate two models of

the

observations: B and C-states (n = 2 only) and all Rydberg states (n = 2–4).

766 Charged Particle and Photon Interactions with Matter

The STIS performed medium- and high-resolution spectral image observations (FWHM= 10

−2

−3

nm) of Jupiter in the FUV between 115 and 170nm, revealing the rotational structure of the

principal thermospheric gas, H

, in the aurora oval, in the polar cap, and in the Io ux tube. A STIS

medium-resolution spectrum and H

model are shown in Figure 28.4 (Ajello etal., 2005a). The

medium-resolution (0.09nm FWHM, G140 M grating) FUV observations 129.5–134.5nm by STIS

on January 13, 2001 were analyzed using a recently developed high-spectral-resolution model for

Aurora oval

Jupiter STIS image in Cassini Campaign January 13, 2001

STIS aperture

250

(a)

(b)

Signal (kR/Å)

HST oval aurora data

Regression model

EF, GK,.. Cascade

B-direct

Atomic O lines

225

200

175

150

125

100

1290 1300 1310 1320

Wavelength (Å)

1330 1340 1350

Figure 28.4 (a) A Jupiter STIS image taken at 16:50 UT on January 13, 2001, showing the three auroral

zones of polar cap, auroral oval and limb with the 52 × 0.5 arcs

slit projected on the image (Ajello etal.,

2005a). (b) The HST STIS G140M medium-resolution grating relative photon intensity short wavelength spec-

trum (1295–1345Å) for the north aurora of January 13, 2001 tted in linear regression. The linear regression

analysis with independent vectors of (1) a direct excitation optically thin spectrum of the Rydberg bands of

by a monoenergetic electron distribution at 100eV, (2) a cascade excitation optically thin spectrum of the

Rydberg bands of H

by a monoenergetic electron distribution at 100eV, and (3) an atomic oxygen multiplet at

1304Å. The regression model to the observational data is based upon a linear regression analysis with these

three independent vectors and transmission through a slab of CH

. We were able to estimate the total contribu-

tions from direct excitation and cascading (Ajello etal., 2005a). The atomic O 130.4nm emission is an artifact

the Earth’s dayglow superimposed on the Jupiter observation.

UV Molecular Spectroscopy from Electron Impact for Astrophysics 767

the electron-excited H

rotational lines that considered the Lyman Band spectrum (B

→ X

)

to be composed of an allowed direct excitation component (X

→ B

) and an optically forbid-

den component (X

→ EF, GK, H

−

H,…

followed by the cascade transition EF, GK, H

−

H,…

→ B

). The medium-resolution spectral regions for the Jupiter aurora were carefully chosen

emphasize the cascade component (Ajello etal., 2005a).

28.3 apparatus and experimental methods

In response to the need for accurate collision cross sections to model spectroscopic observations

of the terrestrial and Jovian planetary systems, the Emission Spectroscopy Laboratory (ESL) at Jet

Propulsion Laboratory (JPL) has established ve unique instruments for routinely measuring the

absolute emission cross sections of stable and radical gases from the broad spectral region of the

UV to the Visible-Optical-Infrared (VOIR) (40–1100nm). Three of the ve instruments are shown

in Figure 28.5 and are identied as follows: (1) atomic O (Johnson etal., 2003a,b, 2005), (2) 3m

high resolution (Liu etal., 1995), and (3) atomic H (James etal., 1998a) apparatuses. Not shown

in the gure are the VOIR (Aguilar etal., 2008; Ajello etal., 2008; Mangina et al., 2010) and the

large chamber for studying the long-lived metastable emissions (Kanik etal., 2003). The 3m optical

spectrometer system is capable of high spectral resolution with a resolving power of λ/Δλ = 50,000

and is equipped with a Codacon 1340 × 400 array detector for studying the rotational structure and

kinetic energy (line proles) of excited fragments (Ajello and Ciocca, 1996a). Each UV spectrom-

eter has dual exit ports and dual grating holders to allow scanning over two wavelength ranges,

e.g., the extreme ultraviolet (EUV from 40 to 120nm), far ultraviolet (FUV from 110 to 310nm), or

visible-optical-near

IR (VOIR from 300 to 1100

nm),

without breaking the vacuum.

Using

this instrumentation, ESL has carried out measurements consisting of a calibrated primary

data set of optically thin UV and VOIR uorescence spectra (50–1100 nm) at spectral resolutions

between 0.002 and 1nm and absolute excitation cross sections at electron impact energies 0–2keV

for several stable gases such as H

(Ajello etal., 1995b, 1996b; James etal., 1998b; Liu etal., 1998,

2000, 2002, 2003; Dziczek etal., 2000; Jonin etal., 2000; Aguilar etal., 2008; Glass-Maujean etal.,

2009); HD (Ajello etal., 2005b); D

(Ciocca etal., 1997a; Abgrall etal., 1999); He (Shemansky

etal., 1985); Ar (Ajello etal., 1990); Ne (Kanik etal., 1996); CO (Ciocca etal., 1997b; Zetner etal.,

1998; Beegle etal., 1999); CO

(Kanik etal., 1993); H

O (Makarov etal., 2004), O

(Noren etal.,

2001b; Kanik etal., 2003; Terrell etal., 2004); N

(Ajello etal., 1998a; Liu etal., 2008, Mangina

etal., 2010; Young et al., 2010), SO

(Ajello etal., 2002a, 2008; Vatti Palle etal., 2004), NO (Ajello

etal., 1989a), NO

(Young etal., 2009), N

O (Malone etal., 2008), and for the radical atomic gases

H (James etal., 1997, 1998a) and O (Noren etal., 2001a; Johnson etal., 2003a,b, 2005a). A large

number of analyses of data from the wide variety of satellite missions listed above have used the

ESL measured cross sections and line proles (Hord etal., 1992; Ciocca etal., 1997a,b; Prange

etal., 1997; Feldman etal., 2001; Gustin etal., 2002; 2004; Esposito etal., 2004). Likewise, the

mission planning and instrument calibration phases of the UV instruments on board Cassini and

the Pluto New Horizons (Stern etal., 2008) interplanetary spacecraft depended on cross sections,

spectra,

and spectral line proles established by the ESL (Ajello etal., 1995a,b).

The

emission cross sections of the majority of neutral and single-ionized planetary gases have

been reviewed by Avakyan etal. (1998) and Majeed and Strickland (1997), and for molecular hydro-

gen and its isotopes by Tawara etal. (1990). For atomic oxygen, one of the most important planetary

gases, and oxygen-bearing molecule’s reviews of cross sections have been given recently by Johnson

etal.

(2005) and McConkey etal. (2008).

The

experimental technique developed at JPL for the measurements of electron-impact-induced

UV emission cross sections and spectra of stable atoms and molecules has been described in Ajello

etal. (1989b, 2002a,b) and references therein, and is shown schematically in Figure 28.6. In brief,

UV emission spectra are generated from collision of a collimated beam of energetic electrons with

768 Charged Particle and Photon Interactions with Matter

a beam of target gas (produced by a capillary array) in a crossed-beams geometry at a background

pressure that ensures optically thin, single-scattering conditions. The interaction volume of the

crossed beams is approximately 2mm

. Emitted photons corresponding to the radiative decay of

collisionally excited states of the target species are detected by the UV spectrometer with its optic

axis at 54.74° (magic angle) or 90° to the plane containing the crossed electron- and target molecular

beams.

XYZ

Manipulator

Discharge

O-Jet

The atomic O experiment

e-Beam

Cem

Pmt

UV/VIS

Spectrometer

hν

Microwave

discharge

e-Gun

To turbos

Spark

Helmholtz

coils

High resolution 3m UV spectrometer

(a)

(b) (c)

The atomic H experiment

Spectrometer

hν

Lβ

Lα

Cross

ectrum

e-GUN

Electrostatic

Polarization

XYZ

Manipulator

Discharge

Thermal

blanket

frame

Codecon

Cem

pmt

Back

focussing

mirror

Collision

chamber

Turbo

pump

Gas

beam

E-Beam

UV Light

source

UV Spectrometer

hν

Figure 28.5 The JPL instrumentation consisting of: (a) an atomic O radical source with a low-resolution

spectrometer, λ/Δλ = 1000 (Johnson et al., 2005); (b) a high-resolution 3 m imaging spectrograph with

Codacon

MCP detectors similar to the detector own on Cassini UVIS, λ/Δλ = 50,000 (Liu etal., 1995); and

(c)

the atomic H radical source with low-resolution spectrometer, λ/Δλ = 1000 (James etal., 1997).

UV Molecular Spectroscopy from Electron Impact for Astrophysics 769

In an electron-impact-induced emission experiment, a molecule in a ground electronic X-state

is excited to an electronic state α. A model of the irradiance from the interaction region for the

transition

│

α, v, J〉 →

│

X, v

, J

〉 from a target molecule, such as H

or CO, is based on the calculated

transition probabilities (e.g., for H

Lyman and Werner bands use Abgrall etal. (1993a,b,c, 1997,

1999, 2000) and the rotational line positions of Roncin and Launay (1994)). For the CO Fourth

Positive system, we use the transition probabilities and molecular constants given in Morton and

Noreau

(1994). The model for electron-excited H

molecules has been presented in previous papers

(Liu etal., 1995, 1998, 2000, 2003; Jonin etal., 2000). We will briey describe it here. The popula-

tions of the ground-state rotational levels are controlled by the gas temperature and nuclear spin

of the molecule. The ground-state molecules in vibrational and rotational thermal equilibrium are

excited into the various rovibronic states according to the excitation rate g′

(α, v, J, E

). The photo-

emission intensity into the various branches from rovibronic state │α, v, J〉 is partitioned according

to the emission branching ratio, predissociation yield, and pre-ionization yield. The photoemission

volume intensity, V′ (photons/s/cm

), at any given point (x,y,z) in the interaction region dened by

the intersection of electron beam of energy E

and molecular beam, in the laboratory, including self-

absorption

and predissociation, is given by

′

⋅V X v v J

v v J J

A X v J

v J E( , , , , , )

, , , , , )

( , , , )

( , , , )α ⋅ α

f f

A X

(

(( ( , , )) ( , , , , )1− ⋅η α

Trα v J X v J T (28.4)

Electron collision processes study of neutral species

Apparatus

Collison

chamber

VUV (40–500 nm)

Spectrometer

(λ/Δλ=50,000)

hν

Pulse

(a)

(b) (c)

(0–2 keV)

hν

(H, H

, O, …)

Spectrum (40–500 nm) Excitation function (0–2 keV)

e+M M*+e

M* M+

hν

spectrum

hν

excitation

function

Energy selected λ selected

Energy (eV)

(cm

)

λ (nm)

e+M M*+e

M* M+

Figure 28.6 (a) A schematic of the crossed electron beam-molecular-beam apparatus for (b) measuring

electron-impact-induced uorescence spectra and (c) emission cross sections as a function of excitation energy

measurement

by continually scanning electron energy, e.g., from 0 to 2

keV.