Summer School
Radiative Transfer and Numerical MHD
Institute of Theoretical Astrophysics
Oslo, 19 – 29 June 2007
Polarimetry: Index, references and exercises
Jorge Sánchez Almeida
– Kitt Peak Synoptic maps
– Line ratio method
– Broad Band Circular Polarization of Sunspots
– Quiet Sun Magnetic fields
Examples of Solar Magnetometry:
Inversion Techniques:
– General ingredients
– Examples, including the magnetograph equation
– Stokes parameters, Jones parameters, Mueller matrixes and
Jones matrixes
– Equation of radiative transfer for polarized light
– Zeeman effect
– Selected properties of the Stokes profiles, ME solutions, etc.
Instrumentation:
– Polarimeters, including magnetographs
– Instrumental Polarization
Radiative Transfer for Polarized Radiation.
Summary – Index (1):
– Hanle effect based magnetometry
– The use of lines with hyperfine structure
– He 1083nm chromospheric magnetometry
– Polarimeters on board Hinode
Advanced Solar magnetometry.
Summary – Index (2):
goto end
refs. on magnetometry
1
References on Solar Magnetometry
This short list does not try to be comprehensive. It just contains the works
cited during the lecture on magnetometry, plus a few general monographs that
may be useful as reference books.
1. Polarization of the light and its various representations
• Chandrasekhar (1950). Re-introduces the Stokes parameters defined
some 100 years before by G. G. Stokes (1852, Trans. Camb. Phil.
Soc. 9, 399). Short readable summary.
• Shurcliff (1962). Old but classical text-book.
• Azzam & Bashara (1989). Technical but complete monograph.
2. Introduction to radiative transfer for polarized light applied to Solar physics.
• Rees (1987). Gentle introduction, as the title says
• Landi Degl’Innocenti (1992). Complete introduction.
• Jefferies et al. (1989). Comprehensive derivation of the radiative
transfer equations from classical electrodynamics.
• Landi Degl’Innocenti & Landi Degl’Innocenti (1972) and Landi Degl’Innocenti
(1983). Quantum derivation.
• Stenflo (1994). Full monograph with many specific applications.
• Trujillo-Bueno et al. (2002). Proceedings of a school on Astrophysical
Spectropolarimetry.
• Landi Degl’Innocenti & Landolfi (2004). Monograph. It has it all!
3. Historical Papers
• Unno (1956). Original derivation of the so-called Unno-Rachkovsky
equations.
• Rachkovsky (1962). Original paper, difficult to find.
4. Mueller matrix of a weakly polarizing optical system. Sánchez Almeida &
Vela Villahoz (1993)
5. Equivalent Zeeman triplet: Beckers (1969), Landi Degl’Innocenti (1982).
6. Weak magnetic field approximation. Landi Degl’Innocenti & Landi Degl’Innocenti
(1973). Jefferies & Mickey (1991).
7. Milne-Eddington solutions of the radiative transfer equations for polarized
light. Landolfi & Deglinnocenti (1982), Landi Degl’Innocenti & Landi
Degl’Innocenti (1985).
refs. on magnetometry
2
8. Weakly polarized medium approximation. Sánchez Almeida & Trujillo
Bueno (1999)
9. Continuum polarization. Kemp (1970).
10. Polarimeters. Baur (1981)
11. Instrumental Polarization. Sánchez Almeida et al. (1991). Capitani et al.
(1989).
12. Seeing induced cross-talk. Lites (1987).
13. Cross-over effect. Grigorjev & Katz (1972). Sánchez Almeida & Lites
(1992).
14. Inversions: Milne-Eddington, Skumanich & Lites (1987). MISMA, Sánchez
Almeida (1997) PCA, Rees et al. (2000).
15. Line-Ratio Method. Stenflo (1973)
16. Numerical simulations of complex magnetic fields. Cattaneo (1999).
17. Broad-Band Circular Polarization in Sunspots. Illing et al. (1974a), Illing
et al. (1974b), Makita (1986).
18. Magnetometry with lines having hyperfine structure. López Ariste et al.
(2002), Landi Degl’Innocenti (1975).
19. Hanle effect based magnetometry. Stenflo (1982), Faurobert-Scholl (1993).
References
Azzam, R. M. A., & Bashara, N. M. 1989, Ellipsometry and Polarized Light
(Amsterdam: North-Holland)
Baur, T. G. 1981, Optical Ingeneering, 20, 2
Beckers, J. M. 1969, A Table of Zeeman Multiplets, AFCRL-69-0115 (Bedford,MA: Office of Aerospace Research USAF), revise
Capitani, C., Landi Degl’Innocenti, E., Cavallini, F., Ceppatelli, G., & Landi
Degl’Innocenti, M. 1989, SPh, 120, 173, revise
Cattaneo, F. 1999, ApJ, 515, L39
Chandrasekhar, S. 1950, Radiative Transfer (Oxford: Oxford University Press)
Faurobert-Scholl, M. 1993, A&A, 268, 765
Grigorjev, V. M., & Katz, J. M. 1972, SPh, 22, 119
Illing, R. M. E., Landman, D. A., & Mickey, D. L. 1974a, A&A, 35, 327
refs. on magnetometry
3
—. 1974b, A&A, 37, 97
Jefferies, J. T., Lites, B. L., & Skumanich, A. 1989, ApJ, 343, 920, revise
Jefferies, J. T., & Mickey, D. L. 1991, ApJ, 372, 694
Kemp, J. C. 1970, ApJ, 162, 169+, check
Landi Degl’Innocenti, E. 1975, A&A, 45, 269
—. 1982, SPh, 77, 285
—. 1983, SPh, 85, 3
Landi Degl’Innocenti, E. 1992, in Solar Observations: Techniques and Interpretation, ed. F. Sánchez, M. Collados, & M. Vázquez (Cambridge: Cambridge
University Press), 71
Landi Degl’Innocenti, E., & Landi Degl’Innocenti, M. 1972, SPh, 27, 319
—. 1973, SPh, 31, 299
—. 1985, SPh, 97, 239
Landi Degl’Innocenti, E., & Landolfi, M. 2004, Astrophysics and Space Science
Library, Vol. 307, Polarization in Spectral Lines (Dordrecht: Kluwer)
Landolfi, M., & Deglinnocenti, E. L. 1982, SPh, 78, 355, revise
Lites, B. W. 1987, Appl. Optics, 26, 3838
López Ariste, A., Tomczyk, S., & Casini, R. 2002, ApJ, 580, 519
Makita, M. 1986, SPh, 106, 269
Rachkovsky, D. N. 1962, Izv. Krymsk. Astrofiz. Obs., 27, 148
Rees, D. E. 1987, Numerical Radiative Transfer, ed. W. Kalkofen (Cambridge:
CUP), 213, check
Rees, D. E., López Ariste, A., Thatcher, J., & Semel, M. 2000, A&A, 355, 759
Sánchez Almeida, J. 1997, ApJ, 491, 993
Sánchez Almeida, J., & Lites, B. W. 1992, ApJ, 398, 359
Sánchez Almeida, J., Martinez Pillet, V., & Wittmann, A. D. 1991, SPh, 134,
1, revise
Sánchez Almeida, J., & Trujillo Bueno, J. 1999, ApJ, 526, 1013
Sánchez Almeida, J., & Vela Villahoz, E. 1993, A&A, 280, 688
refs. on magnetometry
4
Shurcliff, W. A. 1962, Polarized Light (Cambridge MA: Harvard University
Press)
Skumanich, A., & Lites, B. W. 1987, ApJ, 322, 473
Stenflo, J. O. 1973, SPh, 32, 41
—. 1982, SPh, 80, 209
—. 1994, Solar Magnetic Fields, ASSL 189 (Dordrecht: Kluwer)
Trujillo-Bueno, J., Moreno-Insertis, F., & Sánchez, F., eds. 2002, Astrophysical
spectropolarimetry (Cambridge: Cambridge University Press)
Unno, W. 1956, PASJ, 8, 108
J. Sánchez Almeida,. Oslo, June 2007.
1
Exercises on Solar Magnetometry
1
Problem
Derive the Mueller matrix of a linear retarder using the radiative transfer equation for polarized light.
Help: set to zero all elements of the absorption matrix, except those representing a selective retardance in the x-axis. Set to zero the emission terms.
Integrate the resulting differential equation. The solution renders a linear relationship between the input and the output Stokes vectors, i.e., the Mueller
matrix of the retarder.
Solution:
1
0
Mueller Matrix ≡ 
0
0

0
1
0
0
0
0
cos δ
− sin δ

0
0 
,
sin δ
cos δ
where δ is the total retardance.
2
Problem
Derive the Mueller matrix of a linear polarizer that transmits linear polarization
in the x-axis.
Help: set to zero all elements of the absorption matrix, except those representing a selective absorption in the y-axis. Set to zero the emission terms.
Integrate the resulting differential equation.
Solution:
1
1 1
Mueller Matrix ≡ 
2 0
0

3
1
1
0
0
0
0
0
0

0
0

0
0
Problem
Demonstrate that a change of the magnetic field azimuth by 180o does not
modify the polarization produced by a constant magnetic field.
Question: What is the effect of this 180o-ambiguity on the measurements of
solar magnetic field azimuths?
2
Exercises on Solar Magnetometry
4
Problem
Show that in LTE and if the velocity is zero, the Stokes I, Q, U profiles are
symmetric whereas the Stokes V profile is anti-symmetric,
I(λ) = I(−λ),
Q(λ) = Q(−λ),
U (λ) = U (−λ),
V (λ) = −V (−λ).
The symbol λ stands for the wavelength referred to the central wavelength of
the transition.
Help: All Zeeman patterns are symmetric, i.e., for each red-shifted σ+ component there is a blue-shifted σ− with the same amplitude. Similarly, each redshifted π-component is paired off with a blue-shifted π-component. Keep also
in mind that the absorption profile of each component is symmetric, whereas
the retardance profile is anti-symmetric.
5
Problem
The observed Stokes V profiles are never anti-symmetric, implying the existence
of spatially unresolved structures in the resolution elements. The resolution
elements of present solar observations are much thinner in the direction along
the line-of-sight (some 100 km) than in the directions perpendicular to the
line-of-sight (some 725 km, for 1" resolution). Demonstrate that if an observed
Stokes V profile shows net-circular polarization (NCP), the unresolved structure
is smaller than the smallest dimension of the resolution element (smaller than,
say, 100 km).
The NCP is defined as the mean Stokes V averaged over the line profile,
N CP =
Z
λf
V (λ) dλ.
λi
The absorption produced by the spectral line is assumed to be zero outside the
range limited by λi and λf .
Help: Work out the NCP produced by unresolved velocity sub-structure which
varies across the line-of-sight but is constant along the line-of-sight.
3
Exercises on Solar Magnetometry
6
Problem
In order to explain the Broad-Band Circular polarization observed in Sunspots,
Makita (1986, Solar Phys. 106, 269) proposed the generation of circular polarization by spectral lines formed in a transverse magnetic field (i.e., a magnetic
field perpendicular to the line-of-sight). Is it possible or should Makita attend
a school on solar magnetometry before proposing such things?
Help: Makita does not need to attend the school on magnetometry.
7
Problem
Demonstrate that there is no instrumental polarization (IP) in the primary focus
of a telescope. More precisely, the telescope does not produce IP on the optical
axis at that primary focus.
Procedure: First, write down the Jones matrix corresponding to the reflection
on a flat mirror. Be careful with the reference systems in which it is expressed.
Then divide the entrance aperture of the telescope into small sub-apertures.
The Jones matrix of each one of these sub-apertures is the Jones matrix of a
flat mirror. Add up the contributions off all sub-apertures to compute the net
effect produced by the primary mirror of the telescope.
Help: If an optical device is rotated by an angle α, the Jones matrix of the
rotated device m(α) is just
m(α) = R(−α) m(0) R(α)
with
R(α)
cos α sin α
− sin α cos α
.
The new Jones matrix is expressed in the same reference systems in which the
original Jones matrix m(0) was expressed.
Questions:
1. What is the IP in a Cassegrain focus?
2. What is the IP in a Coude focus? Does is vary with time?
3. Why was the average carried out using Jones matrices rather than Mueller
matrices?
4. What happens if one measures off the optical axis?
4
Exercises on Solar Magnetometry
8
Problem
Calibration of a real magnetogram.
The term calibrate a magnetogram means transforming the observed degree of
circular polarization into magnetic flux densities along the line-of-sight. Use the
so-called magnetograph equation,
V (λ) = −k Bz λ20 gef f
dI(λ)
,
dλ
(1)
which provide the relationship between the circular polarization V at each wavelength λ, and the magnetic flux density along the line-of-sight Bz . The symbols
gef f and λ0 stand for the effective Lande factor and the central wavelength of
the transition, respectively. When I and V have the same units, the calibration
constant is
k = 4.67 × 10−13 Å−1 G−1 .
(2)
Needs: IDL. The data are stored as an IDL saveset. It is called mag problem.save
and it contains the following variables
magneto: raw magnetogram, i.e., an image made out of Stokes V
signals.
stokesi: Stokes I profile containing the spectral line used to produce the magnetogram.
wave: wavelength of stokesi, in Å and refereed to 6302.5 Å.
wave0: central wavelength of the line used to produce the magnetogram.
filter wave: central wavelength of the color filter used to obtain
the magnetogram (in Å and referred to wave0).
filter width: band-pass of the color filter (in Å).
Information: The magnetogram has been taken in the blue flank of the
Fe i line at 6302.5 Å, which has an effective Lande factor of 2.5. The color
filter has a box-like shape, i.e., it is zero within the band-pass and zero elsewhere. The pixel size of the magnetogram is 280 km × 280 km. Surface area of
the Sun: 6.09 × 1022 cm2 .
Questions:
1. What is the calibration constant of the magnetogram, i.e., the constant
that transforms the degree of polarization into G units.
2. Compute the magnetic flux of the full observed region.
3. Compute the unsigned magnetic flux density of full observed region.
Exercises on Solar Magnetometry
5
4. If the full sun were covered by magnetic fields like the ones in this magnetogram, what would be the unsigned flux that they carry? Active regions
during the solar maximum carry some 8 × 1023 Mx (1 Mx=1G cm2 ).
9
Problem
Determine the magnetic field strength in the umbra of a sunspot using a MilneEddington inversion.
Needs: IDL. The data are stored as an IDL saveset named sunspot 6302.save.
It contains a set of Stokes profiles of Fe i 6303.5 Å observed in an umbra,
si0: Stokes I profile, referred to the continuum intensity of the quiet
Sun.
sq0: Stokes Q profile.
su0: Stokes U profile.
sv0: Stokes V profile.
wave0: wavelength in Å and refereed to 6302.5 Å.
Procedure: Use the IDL routine plot me.pro to synthesize the Stokes profiles
of Fe i 6302.5 Å. By trial and error, vary the free parameters of this line up to
a point where the observed polarization is reproduced.
Help: You may need to neglect Stokes I when fitting the polarization. Why?
Information: The effective Lande factor of Fe i 6302.5 Å is 2.5.
Questions:
1. What are the magnetic field strength, inclination and azimuth of this
umbra? Check the 180o ambiguity.
2. What is the macroscopic velocity?
3. Have you used Stokes I when fitting the Stokes profiles? If the answer is
no, explain why.
6
Exercises on Solar Magnetometry
10
Problem
The so-called line-ratio-method (LRM) was introduced by Stenflo (1973, Solar
Phys., 32, 41) to show that the magnetic field strengths in plage regions are
strong (kG), despite the fact that magnetograph observations find them to be
weak (say, 100 G). The LRM models the ratio between the magnetograph signals
observed in two spectral lines that are identical in unmagnetized atmospheres
but whose Zeeman splitting is very different (Fe i 5250 Å, gef f = 3, and Fe i
5247 Å, gef f = 2). Should the magnetic field were weak, the ratio between the
magnetograph signals observed in these two lines would be given by the ratio
of effective Landé factors, i.e.,
VF eI
5250 (λ)/VF eI 5247 (λ)
= 1.5
However, plage observations show
VF eI
5250 (λ)/VF eI 5247 (λ)
≃ 1.1
when λ ≃ 50 mÅ (λ: wavelength refereed to the line center). Use the LRM to
show that this observational fact implies kG field strengths in plage regions.
Explain why magnetographs render apparent magnetic fields that are much
weaker than the intrinsic ones.
Procedure: Use the solutions of the radiative transfer equation for polarized light in a Milne-Eddington atmosphere with longitudinal magnetic field.
They yield Stokes V profiles given by
V (λ) ∝
1
1
−
1 + η0 f (λ + ∆λB ) 1 + η0 f (λ − ∆λB )
Keep in mind that
∆λBF eI
5250
= 38.6 mÅ
B
,
1000 G
∆λBF eI
5247
= 25.7 mÅ
B
.
1000 G
and
Compute the ratio VF eI 5250 /VF eI 5247 at λ = 50 mÅ for a range of magnetic
field strengths. Choose the one that renders the observed ratio. For these
lines, the line-to-continuum absorption coefficient ratio η0 is some 2, whereas
the absorption profile f (λ) can be a Gaussian function with full-width-halfmaximum 90 mÅ and area equals one.
J. Sánchez Almeida,. Oslo, June 2007.