Spectroimaging observations of H3 on Jupiter

1
Spectro-imaging observations of H3 on Jupiter
Emmanuel Lellouch
Observatoire de Paris, France
2
H3 in planetary atmospheres

• Discovered in 1988 in Jupiters auroral regions
  (Drossart et al. 1989) from its 2??2 emission
• Since then, also discovered in Saturn and Uranus,
  and in Jupiters low latitude regions
• General goals on the observations (esp. Jupiter)
• Characterize the morphology and variability of
  the emission
• Interpret the emission rates in terms of H3
  column density and temperature 
• Measure winds from Doppler shifts on H3
  emissions
• Bands observed so far on Jupiter ??2, 2?2, 2?2 -
  ??2
• H3 traces and drives the energetics and dynamics
  of Giant Planets upper atmospheres
• (see Steve Millers talk)

3
Outline

• Detection of the 3 ??2 - ??2 band on Jupiter
• The question of LTE and the multiplicity of H3
   temperatures 
• The spatial variations of the H3 emission do
  they trace variations in the H3 column or
   temperature ?

4
Spectro-imaging observations

• Imaging FTS  BEAR  at Canada-France-Hawaii
  Telescope
• Sept. 1999 Oct. 2000
• 2 filters
• 2.09 µm 4760-4805 cm-1
• 2.12 µm 4698-4752 cm-1
• 14 data cubes  (2 spatial, 1 spectral dimension)
  sampling either the Northern or the Southern
  auroral region of Jupiter at various longitudes
• See detailed report in Raynaud et al. Icarus,
  171, 133 (2004)

5
Example of spectrum 2.09 µm range
2?2 band
6
Example of spectrum 2.12 µm range
2?2 and 3?2-?2 bands H2 S1(1)
7
The two 3??2 - ?2 lines

• Obs. freq. (cm-1) Calc. Freq. (cm-1)
  Assignment E (cm-1)
• (Neale et al. 1996)
• --------------------------------------------------
  -------------------------------
• 4721.79 4722.383 3?23
  - ?2 R(6,7) 7993.60
• 4749.66 4749.937 3?23
  - ?2 R(5,6) 7998.64
• --------------------------------------------------
  -------------------------------

8
Temperature/column density determinations

• Classical LTE formulation
• This effectively assumes complete LTE (Trot
  Tvib Tkin)
• Justified for rotational distribution (??rad
  1000 sec, ?coll 10-3 sec)
• For vibrational distribution, ??rad 10-2 sec ?
  non-LTE however earlier studies (Y.H. Kim et
  al. 1992) have found that
• All the nv2 levels are underpopulated w.r.t.
  ground state
• But their relative populations are close to LTE
  distribution  quasi-LTE  distribution
• Here
• 2.09 µm observations (2 ?2 lines only) determine
  Trot in v2 2 level
• 2.12 µm observations (including the 3 ?2 - ?2
  lines) determine  Tvib (relative populations of
  v2 2 and v2 3 levels)

9
Temperature results

• Tvib
• Trot

Central meridian longitude
In average, Trot 1170/-75 K gt Tvib 960 /-
50 K ? underpopulation of v2 3 relative to ?2
2 with respect to LTE case
10
Trot and N (H3) results

• We get Trot (2?2) 1170/-75 K, and N
  (4-8)x1010 cm-2
• Quite different from Lam et al. 1997 Trot (?2)
  700-1000 K, and N 1012 cm-2
• Interpretation
• The 2 bands probe different levels in an
  atmosphere with strong positive temperature
  gradient

Grodent et al. 2001
11
A non-LTE H3 model (Melin et al. 2005)

• Based on detailed balance calculations (Oka and
  Epp 2004) and a physical (temperature,density)
  model of Jupiters upper atmosphere
• Calculate line production profile in atmosphere
  and compare with LTE situation
• Results
• The 2?2 band probes higher and hotter atmospheric
  levels than ?2
• Non-LTE effects are minor for ?2 but much more
  significant for 2 ?2 lines
• Thus, using 2 ?2 lines leads to too low a column
  density

Melin et al. 2005
12
Trot vs. Tvib

• We find Trot 1170/-75 K gt Tvib 960 /- 50
  K, i.e. an underpopulation of v2 3 relative to
  v2 2 with respect to LTE case
• Interpretation
• non-LTE effects are even more significant for
  3?2 - ?2 (and higher overtones) than for 2?2
  lines ? the temperature determined by assuming
  QLTE for v2 2 and v2 3 underestimates Tkin

Melin et al. 2005
13
Beyond the QLTE hypothesis

• Main conclusion non-LTE effects are severe
• ? May induce large errors in T and especially N
  (H3) determined from overtone and hot bands (up
  to 2 orders of magnitude underestimate for N
  (H3) !)
• Future studies rather than  temperature/column
  density retrievals , better to perform  forward
  modelling , i.e. test Tkin(z), n(H3) profiles
  directly against data

14
Spatial distribution of emissions H2 vs. H3
H3 2?22 R(7,7) 4732 cm-1
H2 S1(1) 4712 cm-1
H3 3?23- ?2 R(6,7) 4722 cm-1
 Hot spot  near ? 70N, LIII 160
15
Variations in H3 emission rates variations in
temperature or column density ?
1-5 five bins of increasing emission (5  hot
spot  near LIII 160)

• Except in  hot spot , emission variations
  mostly due to variations in N(H3)
• Confirmed by search for correlations between
  intensities/temperatures/columns on individual
  pixels

16
Emission variations are mostly due to variations
in N(H3)

• In agreement with Stallard et al. 2002
• Little or no temperature variations thermostatic
  effect from H3
• H3 cooling dominant above homopause
• e.g. T(0.01 µbar) 1300 K. Would be 4800 K if no
  H3 cooling (Grodent et al. 2001)
• Large variations of input energy radiated by
  modest increases of temperature
• E.g.  diffuse  and  discrete  aurora
  (differing by amount of hard electron
  precipitation) have similar (within 100 K)
  temperatures
• Exception the  hot spot , actually hotter
  than other regions by 250 K

17
Conclusions

• We have detected the 3?2 - ?2 band of H3 on
  Jupiter
• Our temperature/column density determinations
  differ with those obtained from other bands. This
  can be understood from
• Strong non-LTE effects on combination and
  overtone bands
• The temperature profile in Jupiters auroral
  ionosphere
• Spatial variations in the H3 emission generally
  trace variations in H3 columns and not in
  temperature
• The increased temperature in the  hot spot 
  (also visible at other wavelengths but NOT in H2
  emission) remains a mystery

18
The H3 Northern auroral  hot spot 

• Located at 70N, LIII 160
• Has Tvib 250 K warmer than
• other regions
• Region peculiar at other ?
• Thermal IR emission of several
• hydrocarbons
• Far UV (footprint of polar cusp)
• X-ray emission
• But not in H2 S1(1) !
• Origin?
• Impact of very energetic (gt 100 keV) electron (cf
  origin of FUV features)?
• Would rather produce deep (cold) H3
• Increased vertical mixing due to increased
  precipitation, resulting in elevated homopause?
• Increased CH4 ? reduces the deep cold H3
  component ? increase of mean H3 temperature
• But is it consistent with increase of H3
  emission?
• Why not seen in H2 ?