An Infrared study of the Jovian aurorae

Matthew Pitkin
M.Sci. Astrophysics
ASTR 4C10: Astronomy Project

Supervisors: Dr Steve Miller
Dr Tom Stallard


Infrared spectra and images were taken of Jupiter on 6th Feb. 2001, using the SpeX spectrometer on the NASA IRTF. The wavelength region observed contained H3+ emission lines. These were used to calculate the parameters of temperature, column density and total emission of the H3+ for the Jovian ionosphere, within the auroral regions, to a high level of spatial detail in comparison to most previous studies.

The spectra were reduced, and processed using an IDL procedure. This was used to obtain positional information and to cut down the spectra so that only the Q-branch emission lines of H3+ were passed to the fitting program. To calculate the parameters the spectra were fitted to theoretical versions, using a FORTRAN program. A varying background, of reflected light, was used in the fit.

The northern aurora was mainly studied, as more of it was displayed. Profiles of the different parameters as they varied across the aurora were produced, with one particular profile giving the best view of the auroral structure. Longitude and latitude maps of the parameters were produced, to show how they varied across the whole aurora.

The parameters were found to be generally consistent with previous studies, and showed that H3+ has a major role in cooling the ionosphere, and is also a very good indicator of energy deposition. Various anomalies suggested possible evidence of heat transport within the aurora, and/or layered H3+ production at various altitudes in the ionosphere.

1 Introduction
On 6th February 2001 spectra were taken of Jupiter using the SpeX spectrometer on the NASA Infrared Telescope Facility (IRTF). The wavelength range of each spectrum was ~ 3.5 – 4µm, chosen because emissions from H3+ lie above a strong methane absorption of background light in this region. In this study, the spectra will be used to calculate three parameters (temperature, column density and total emission of H3+) in the auroral regions of the Jovian ionosphere. Along with each spectrum an image of Jupiter was taken, covering most of the L’ wavelength window, centred at a wavelength of ~ 3.8µm. The images will be used to show auroral morphology, and provide links between the intensity of the aurorae and the parameters shown above. The emission lines of the ion H3+ are being used because H3+ plays an important role in the energy regulation of Jupiter’s upper atmosphere, which will be discussed in more detail later. To be able to cover the entire north and south aurorae in a reasonable time, the spectrograph slit was oriented in the north-south plane of the planet, and was scanned west to east with each exposure. The rotation of the planet over the period that the spectra were taken also allowed a greater proportion of the aurorae to be covered. In all 21 spectra were taken on 6th Feb, and all will be used to construct an overall map of the aurorae. These aurorae are important, in that they are the main regions of energy deposition, in the form of charged particles, in the atmosphere.

Over the past decade, or so, the emission by H3+ on Jupiter have been much studied. It was accidentally discovered in 1988 by Drossart et al. (1989), while attempting to look at H2 lines, and since then it has been the main probe for upper atmospheric conditions. These conditions have been measured by various groups, covering all latitudes of Jupiter, with improvements in the spectral and spatial resolution being made over the years. The other main spectral region used to study the Jovian aurorae is the ultraviolet (UV), where emission lines from H and H2 exist. This current study is complementary to previous studies, and will attempt to show the role that H3+ plays within the auroral regions. It is also the highest spatial resolution study yet completed, which will hopefully allow us to pinpoint auroral features more accurately.

1.1 Jupiter’s Atmosphere
Jupiter is composed of elements with almost solar abundances. Therefore, hydrogen (~90%) and helium (~10%) are its main constituents, with only about 1% of its elemental abundance being from other elements. Due to the dominance of hydrogen in the atmosphere most of the other elements only exist in the form of hydrogen compounds (e.g. methane). Due to Jupiter being a gaseous body, a level has to be set as a reference from which the bottom of the atmosphere can be defined, which is set where the pressure is at 1 bar. Therefore, the atmosphere above this level is split into various regions, comparable with other planetary atmospheres. These regions are the troposphere, stratosphere, mesosphere, and the thermosphere/ionosphere in ascending order. The temperature and pressure vary through these regions depending on their compositions and the various energy transport processes that go along with those compositions.

1.2 Jovian Ionosphere
In this report the major area of interest is the ionosphere, in which the temperature increases with altitude above a base pressure of < 1mbar. In this region, which is above what is called the homopause, only hydrogen and helium species exist, due to them being the lightest species. Below the homopause, in the homosphere, vertical mixing homogenises the chemical composition. The ionosphere is generally coincident with the thermosphere, consisting of a large number of ions (e.g. H+, H2+, H3+) and free electrons, due to the ionisation rate being higher than the recombination rate. The ionosphere is linked to the magnetic field lines of Jupiter, and is thus connected to the surrounding space environment.

1.3 Jovian Magnetosphere
Jupiter has a huge magnetic field (~10 000 times the strength of Earth’s) produced by a dynamo effect in its massive metallic H core. This, in combination with Jupiter’s rapid rotation, creates a very large magnetic moment. This magnetic field displaces the solar wind forming a giant cavity, with the magnetotail stretching back as far as Saturn’s orbit. Due to this interaction with the solar wind and it not rotating along Jupiter’s axis of rotation, the magnetic field cannot be modelled as a dipole. Several successful, although obviously more complex, models of the magnetosphere have been made.

Jupiter’s innermost moon, Io, orbits at 5.9 Jovian radii (RJ) and is very volcanic because of tidal interactions with Jupiter. This volcanic activity throws out large amounts of material into the orbital plane. It forms a plasma torus around Io, and an extensive plasmasheet stretching out along the equatorial plane of Jupiter. The plasmasheet is linked to the ionosphere via the magnetic field lines (see Fig. 1), which forms current systems between them. These currents keep the plasmasheet in co-rotation with Jupiter by dragging it along with the planet. This co-rotation starts to break down further out from the planet when the inertia of the plasmasheet begins to become too great for the weaker currents to pull along. Where this co-rotation breaks down is not well delineated, but is often defined as ~30 RJ. Distances out on the plasmasheet can be mapped onto the ionosphere, with field lines from further out mapping to higher latitudes on the planet. Lines of equipotential magnetic field strength map out as ovals at high latitudes in the northern hemisphere of Jupiter, and are more circular in the southern hemisphere.


Fig. 1. This figure shows how the magnetic field lines map to the ionosphere and currents associated with them. This figure is taken from with permission from Stallard (2001)

1.4 Jovian Aurorae
Jupiter, like Earth, has aurorae associated with energetic particle precipitation along magnetic field lines at high latitudes. The auroral emissions cover a broad range of emission lines over a wide wavelength spectrum (X-ray, UV, visible, IR, radio), arising from different species/processes at different levels in the atmosphere. The aurorae arise from energetic particles (ions and electrons) impacting on the atmosphere of Jupiter in the regions of highest magnetic field strength (i.e. at the high latitudes around the magnetic poles). The impacting particles excite molecules/atoms/ions into higher energy states, which then relax and emit radiation. The particles also cause ionisation, leading to recombination lines and the production of more ions. The depth to which particles precipitate depends upon their energy, with higher energy particles getting further down into the atmosphere.

The most studied wavelength regions of the aurorae are the UV and the IR. The UV emissions generally studied are the H2 Lyman and Werner bands and H Lyman a line. These emissions are mainly used to study the auroral morphology, due to there being a direct one-to-one correspondence with particle precipitation and emission, because of the timescale between excitation and radiative de-excitation being small. The H2 studies are, however, not good for indicating temperatures within the aurorae. The IR emissions studied are mainly those from the molecular ion H3+. Particular emission lines are very good indicators of temperature within the aurorae, making them very useful in the study of the auroral (or more generally, ionospheric) energy balance. This shall be discussed more later. The morphology of the UV and IR aurora show the same basic features, although the IR emission will be a less accurate indicator of the point of particle precipitation. The structures of the aurorae relate to the structure of the magnetosphere and where the field lines cross the plasmasheet.

1.5 H3+
H3+ is a molecular ion and is the simplest polyatomic molecule there is, consisting of 3 protons and 2 electrons in an equilateral triangle structure when in equilibrium. It was discovered in 1911, although it was not fully believed to exist until electronic structure calculations, done in the 1930s, proved it was stable.

H3+ was predicted to play an important role in interstellar chemistry as a protonating agent, because it is normally very reactive. It was thus initially looked for in the interstellar medium. Its first actual astronomical detection was on Jupiter in 1988 (Drossart et al, 1989), and up until this time it had only been studied by laboratory spectroscopists. After its initial detection it was quickly realised how important H3+ would be in studying the Jovian ionosphere, and since then there have been many observations and studies produced on this subject. For more on the history of H3+ see Miller and Tennyson (1992).

H3+ formation is through the reaction sequence:

H2 + photon (or e*) = H2+ + e (+ e)
H2+ + H2 = H3+ + H

where the ionising agent can either be an UV photon or, as is the major agent in the Jovian aurorae, an energetic electron. The second reaction is strongly exothermic. As previously stated, H3+ is highly reactive, but in the Jovian ionosphere where only H and He exist H3+ cannot react. The lifetime of H3+ is therefore controlled by dissociative recombination:

H3+ + e = H2 + H = H + H + H

where this reaction is controlled by the electron density.

H3+ has no permanent dipole. This means that there is no allowed rotational spectrum (i.e. microwave spectrum). As shown above in the dissociative recombination, it has unstable electronic states and therefore no UV or visible spectrum. H3+ has two vibrational modes. One is a symmetric mode (called nu1), which produces no dipole and therefore no spectrum. The other is a bending mode (called nu2), which varies the position of the atoms and creates a dipole that can allow rotational-vibrational (ro-vibrational) modes. This means that the different vibrational energy levels are split into more rotational levels. Transitions among these ro-vibrational modes produce an infrared spectrum. In this study the transitions that are used will be the fundamental emissions, which are transitions between rotational levels from the V’ = 1 to the V = 0 vibrational levels (see Fig. 2).


Fig. 2. This is a Grotrian diagram of the vibrational levels of H3+ and their splitting into rotational levels, taken with permission from Stallard (2001)

1.6 H3+ in the Jovian aurora
H3+ in the aurorae is mainly produced just above the homopause, through ionisation from particle precipitation, with a small contribution from solar EUV radiation. Below this the concentration falls off rapidly, with a less rapid decline above this level (see Fig. 3). The H3+ fundamental lines are being studied to provide information on the temperature, column density and total emission of H3+ within the ionosphere.

1.7 Temperature
The temperature can be calculated by ratioing emission intensities from lines arising from different excitation levels. The reason H3+ is used for temperature calculations is that the rotational levels are in local thermodynamic equilibrium (LTE) with the surrounding atmosphere (Miller et al. 1990).

1.8 Column Density
The column density is the density of a column of material, in this case H3+, through the atmosphere. It is determined by comparison of the calculated emission of a single


Fig. 3. This shows the concentration of H3+ in the auroral regions (solid line) compared to the equator (dashed line) as produced using the Jovian Ionospheric Model (JIM) over a range of pressures, taken from Stallard (2001)

molecule at a given temperature, with that observed (i.e. actual intensity divided by calculated intensity gives number density). This parameter is useful in that it shows the abundance of H3+ in a particular region, indicating production rates.

1.9 Total Emission
The total emission of H3+, is the emission from all the molecules over all wavelengths. It is determined by calculating the total emission from a single molecule at a given temperature and multiplying it by the column density. The total emission was first used by Lam et al. (1997) as a useful tool in describing the total power output attributable to H3+ in the aurora. It also shows regions of cooling and gives valuable insight into the energy balance of the ionosphere.

1.10 Morphology
The H3+ aurora can be divided into different latitudinal regions based on where lines of equipotential magnetic field strength enter the ionosphere. The H3+ morphology corresponds well with that seen in UV studies. The main regions are the Io footprint, the main oval, and the polar caps. The Io footprint originates from field lines that map out to Io’s orbit, at 5.9 RJ, and provides a ground truth for models of the magnetosphere. The main oval originates from between ~12-30 RJ, and produces the largest proportion, ~25%, of the auroral emissions (Satoh and Connerney, 1999). The polar caps come from further than 30 RJ where the solar wind could be interacting with the ionosphere. The polar cap is seen to have a yin-yang structure, with a bright polar region (BPR) on the dusk side of the aurora, and a dark polar region (DPR) on the dawn side. These features will be discussed later, with reference to the results found in this study.

1.11 Previous studies
As stated before, the H3+ emissions from the Jovian aurorae have been extensively studied over the past decade or so, with improvements to the data and what it can tell us being made continuously. The period of study can also be used to give insight into the variation of the aurora over relatively large timescales. In the first detection of H3+ on Jupiter a rotational temperature for the aurora of 1099 ± 100K was found (Drossart et al, 1989). The next observations (Oka and Geballe, 1990) gave a dramatically different temperature of 670 ± 100K with column densities calculated to be in the range 0.1 – 1.0*1011 cm-2, which they attributed to the time variation of the H3+ emissions. Miller et al. (1990) produced a ro-vibrational temperature of 1100 ± 100K, very similar to that of Drossart, indicating that excitation to upper energy levels was by purely thermal processes. Later studies have produced temperatures generally in the range 900 ± 100K with column densities in the same range as that above, indicating time variation of H3+ production. The variations of the aurorae were studied on a short timescale with reference to the solar wind control (Baron et al, 1996), which showed a good correlation between daily fluctuations in auroral intensity and solar wind pressure. Lam et al. (1997) introduced the new parameter of total H3+ emission to characterise the energy balance within the aurorae better, due to the fact that they found a strong anti-correlation between the temperature and column density. There have also been studies of the temperature profile across all latitudes of Jupiter, with Miller et al. (1997) giving mid-to-low latitude temperatures of 800K and column densities of order 1011 cm-2. This value of temperature was far higher than could be expected from a purely solar energy input, which should give non-auroral ionospheric temperatures of ~200-300K. This strongly suggested H3+ transport from the auroral regions to lower latitudes, or particle precipitation at lower latitudes. Recent studies (see Miller et al, 2000) suggest that a mechanism for this energy transport could be in the form of an H3+ auroral electrojet. The most recent study (Stallard, 2001, hereafter referred to as Paper I) provided the first ever astronomical detection of a hotband emission line (see Fig. 2), meaning the vibrational temperature could be calculated. This was due to the very high resolution of the spectrometer used. The average temperatures found varied over the nights of observation between ~ 950 – 1050K. This work also provided the best spatial resolution for the parameters yet made meaning fairly detailed maps of the parameters could be made. The existence of an ion wind, called an electrojet, and other complex wind patterns was also shown in Paper I, by obtaining ion wind speeds from the study of Doppler shifts in the emission lines.

In this study the ro-vibrational temperature will again be measured. The spatial resolution of the spectrometer used will give one of the best resolutions yet seen, at 0.15” per pixel covering the whole of Jupiter, which had an angular diameter of ~ 42” on the night of observation. This gave approximately 280 pixels covering Jupiter. The spectral resolution of the spectrometer is R ~ 1000-2000 meaning that it was not possible to view the hotband line again. The images associated with each spectrum, and used for positional information, have a resolution of 0.12” per pixel giving a good comparison for the auroral morphology.

top | data reduction | results | discussion | conclusion

home | about me | research | links | travels | pictures | e-mail me