Content uploaded by Marijke Haverkorn
Author content
All content in this area was uploaded by Marijke Haverkorn on Jul 20, 2018
Content may be subject to copyright.
arXiv:astro-ph/0111381v1 20 Nov 2001
Parsec-scale structure in the warm ISM from polarized
galactic radio background observations
M. Haverkorn∗, P. Katgert∗and A. G. de Bruyn†
∗Leiden Observatory, P.O. Box 9513, 2300 RA Leiden, the Netherlands
†ASTRON, P.O. Box 2, 7990 AA Dwingeloo, the Netherlands
Abstract. We present multi-frequency polarization observations of the diffuse radio synchrotron background
modulated by Faraday rotation, in two directions of positive latitude. No extended total intensity Iis observed,
which implies that total intensity has no structure on scales smaller than approximately a degree. Polarized
intensity and polarization angle, however, show abundant small-scale structure on scales from arcminutes to
degrees. Rotation Measure (RM) maps show coherent structure over many synthesized beams, but also abrupt
large changes over one beam. RM’s from polarized extragalactic point sources are correlated over the field in
each of the two fields, indicating a galactic component to the RM, but show no correlation with the RM map of
the diffuse radiation. The upper limit in structure in Iputs constraints on the random and regular components
of the magnetic field in the galactic interstellar medium and halo. The emission is partly depolarized so that the
observed polarization mostly originates from a nearby part of the medium. This explains the lack of correlation
between RM from diffuse emission and from extragalactic point sources as the latter is built up over the entire
path length through the medium.
INTRODUCTION
Synchrotron radiation emitted in our galaxy provides a diffuse radio background, which is altered by Faraday rotation
and depolarization when it propagates through the galactic halo and interstellar medium. Multi-frequency observations
of the linearly polarized component of this radio background allow determination of the Rotation Measure (RM) of
the medium along many contiguous lines of sight, from which the structure of the small-scale galactic magnetic field,
weighted with electron density, can be probed. After the discovery paper ([1]), many other high resolution radio
background observations have shown intriguing polarization structure, mostly in the galactic plane ([2], [3], [4], [5]).
Here, we focus on regions of positive galactic latitude to evade complexities such as HII regions, supernova remnants
etc. Furthermore, these are sensitive and multi-frequency observations so that accurate RM values can be determined.
OBSERVATIONS
With the Westerbork Synthesis Radio Telescope (WSRT) we mapped the polarized radio background in two fields
of over 50 square degrees each in the second galactic quadrant at positive latitudes. All four Stokes parameters I,
Q,U, and Vwere imaged in five frequency bands centered at 341, 349, 355, 360, and 375 MHz simultaneously
(bandwidth 5 MHz), at a resolution of ∼4′. The first field, in the constellation Auriga, is 7◦×9◦in size and centered on
(l,b) = (161,16)◦; the second field, in the constellation Horologium, is 8◦×8◦wide and centered on (l,b) = (137,7)◦.
No total intensity Iwas detected (besides point sources) in either field down to ∼0.7 K, which is <1.5% of the
expected sky brightness in these regions, indicating that Idoes not vary on scales detectable to the interferometer,
i.e. below about a degree. However, linearly polarized intensity Pand polarization angle ϕshow abundant small-scale
structure. Other fields observed with the WSRT at a single frequency around 350 MHz also show small-scale structure
in polarized intensity and polarization angle, but of very different topologies ([6]).
FIGURE 1. Left: polarized intensity Pat 349 MHz in the Auriga field at 4′resolution. White denotes a maximum Tb,pol ≈18 K.
Right: RM in the Auriga field. Very high or low RM values (|RM| ≈ 30 −60 rad m−2) in the field have been removed from the
maps (see text).
ANALYSIS OF THE AURIGA FIELD
The structure in polarized intensity Pin the Auriga field shows a wide variety in topology on several scales, as shown in
the left plot of Fig. 1, where Pat 349 MHz is mapped at 4′resolution. The typical polarization brightness temperature
is Tb,pol ≈6 - 8 K, with a maximum of ∼18 K. From the Haslam continuum survey at 408 MHz ([7]), the I-background
at 408 MHz in this region of the sky is ∼33 K. Extrapolating this to our frequencies with a temperature spectral index
of −2.5 between 341 MHz and 408 MHz ([8]), the total intensity background is ∼41 - 52 K. So the maximum degree
of polarization pmax ≈35%, with an average pof 15%.
In addition, a pattern of black narrow wiggly canals is visible (see e.g. the canal around (α,δ) = (92.7, 49 - 51)◦).
These canals are all one synthesized beam wide and have been shown to separate regions of fairly constant polarization
angle ϕwhere the difference in ϕis approximately 90◦(±n180◦,n=1,2,3...), which causes beam depolarization
([9]). The angle changes are due to abrupt changes in RM. Hence, the canals reflect specific features in the angle
distribution. Other angle (and RM) changes within the beam cause less or no depolarization, so that they do not leave
easily visible traces in the polarized intensity distribution.
The RM of the Faraday-rotating material can be derived from ϕ(λ2)∝RMλ2(see [10] for details and pitfalls). The
right plot in Fig. 1 gives a 4′resolution RM map of the Auriga field. The average RM ≈ −3.4 rad m−2, and in general
|RM|<15 rad m−2. Very high or low RM values (|RM| ≈ 30−60 rad m−2) in the field occur only at positions where
polarized intensity Pis very low, so noise errors in polarization angle are very large. Therefore the RM’s at these
positions are not reliable and have been removed from the maps. RM’s show structure on scales of many beams (up
to degree scales), but also abrupt changes from one beam to another. This is illustrated in the left map of Fig. 2, where
a small part of the Auriga field is shown. Here, the lines are graphs of polarization angle against wavelength squared,
so that the slope of the line is RM. Each graph is an independent synthesized beam. The greyscale denotes polarized
intensity Pat 349 MHz, five times oversampled. Large sudden RM changes occur: e.g. at (α,δ) = (94.68,53.15)◦, RM
changes from −9 rad m−2to 7 rad m−2, and at (α,δ) = (94.60,53.00)◦from 3 rad m−2to −11 rad m−2. A change in
sign of RM indicates in general a change in direction of the galactic magnetic field along the line of sight, although
FIGURE 2. Left: Graphs of polarization angle against the square of the wavelength, so that the slope is RM. Each graph denotes
an independent beam. The greyscale denotes polarized intensity at 349 MHz, five times oversampled. RM’s range from ∼−10 to
10 rad m−2, ignoring the one anomalously large negative RM on the left side. Abrupt changes in Rotation Measure over one beam
(4′) are coherent along several beams. Right: RM’s of observed polarized extragalactic point sources in the Auriga field. The radii
of the circles are scaled with magnitude of RM, where filled circles are positive RM’s. Maximum (minimum) RM is 19.5 (−13.6)
rad m−2.
numerical models of propagation of polarized radiation through a mixed (synchrotron emitting and Faraday-rotating)
medium show a change of sign of RM also without a reversal in the magnetic field ([11]).
We detected seventeen polarized extragalactic sources in the Auriga field at a higher resolution (∼1′), with RM’s
from −13.6 to 19.5 rad m−2. The right plot in Fig. 2 shows the RM’s and positions of the sources, where the sizes of the
circles are proportional to RM, and open (filled) circles denote negative (positive) RM’s. The RM’s of the extragalactic
sources exhibit a clear gradient across the field of ∼5 rad m−2per degree roughly in the direction of galactic latitude,
indicating a galactic component to the RM’s of the sources. The change of sign over the field means a (local) reversal
in the magnetic field parallel to the line of sight. We estimate a RM component intrinsic to the source of <5 rad m−2,
consistent with earlier estimates ([12]). The RM structure of the diffuse galactic radiation is independent from the
observed RM of the extragalactic sources (see below).
ANALYSIS OF THE HOROLOGIUM FIELD
The left map in Fig. 3 shows polarized intensity at 349 MHz in the Horologium field. The average polarization
brightness temperature is ∼5 K, and Tb,pol in the ring-like structure is ∼11-15 K. The average degree of polarization
is 5% and the maximum 25%, again derived using the Haslam survey ([7]). The ring with diameter ∼2.7◦is visible in
all frequency bands, although the ring becomes more diffuse and smeared out towards higher frequencies and the left
side is clearer than the right side. Beam-wide depolarized canals are again caused by beam depolarization. In the lower
right, the depolarized canals are aligned along constant latitude. Caution is required in interpreting the Pmap, as the
Horologium field is imbedded in a region of very high constant polarization ([13]). Due to missing small spacings, we
cannot detect this large-scale component in Qand/or U, which causes a distorted image in Pas well as RM. Possible
solutions to this problem will be given in a forthcoming paper. A ring in Pwas also detected by Verschuur ([14]) at
40′resolution with a single dish, although not as an enhancement but as a deficiency in P.
In the RM map of the Horologium field (the right hand map of Fig. 3), a circular structure is clearly visible, which is
slightly bigger than the ring in polarized intensity. Inside this RM disk, RM’s decrease from the edge of the ring to the
center, from ∼0 to −10 rad m−2. Outside the disk, RM’s vary around zero without a clear gradient, with a maximum
of ∼7 rad m−2. Note that if Pis low outside the disk, the influence of undetected large-scale polarization becomes
larger and RM’s may not be well-determined. The left and center plots of Fig. 4 give horizontal cross sections through
the center of the RM disk and through the Pring at 349 MHz at the same position. The decrease in RM within the
RM disk can be modeled with a homogeneous sphere of constant electron density and magnetic field. For values of
electron density ne=0.03 cm−3and parallel magnetic field Bk=2µG, the distance to the sphere is about 300 pc.
In the Horologium field, we detected 18 polarized extragalactic point sources as shown in the right plot of Fig. 4.
FIGURE 3. Left: polarized intensity Pat 349 MHz in the Horologium field at 4′resolution. White denotes a maximum
Tb,pol ≈15 K. Right: RM in the Horologium field. Very high or low RM values (|RM| ≈ 30 −60 rad m−2) in the field have
been removed from the maps (see text).
FIGURE 4. Left and center: horizontal cross sections through the RM map and the Pmap of the Horologium field in Fig. 3 (at
4′resolution) through the center of the RM disk. The dashed line in the RM plot is a fit to the decrease in RM, calculated for a
sphere of constant thermal electron density and line-of-sight magnetic field. In the Pplot, the same fit is indicated at the bottom.
Right: RM’s of observed polarized extragalactic point sources in the Horologium field. The radii of the circles are again scaled with
magnitude of RM, but with radii twice as small as in Fig. 2. Maximum (minimum) RM is 7.4 (−67.9) rad m−2.
RM’s of these sources are in the range of −67.9 to 7.4 rad m−2. These RM values are not correlated with the RM’s
from the diffuse radiation, similar as in the Auriga field. Extragalactic source RM’s are higher than in the Auriga field,
which is most likely due to the lower latitude of the Horologium field.
INTERPRETATION OF THE OBSERVATIONS
The galactic synchrotron emission is thought to originate from two separate domains centered on the galactic plane:
a thin and a thick disk, with scale heights of 180 pc and 1.8 kpc respectively ([15]). The thick disk emits 90% of
the synchrotron radiation, and the thin disk coincides approximately with the stellar disk and the HI disk. Pulsar RM
observations have shown that the random magnetic field in the thin disk Bran ≈5µG is of the same order or larger than
TABLE 1. Three schematic domains with different characteristics in the galactic ISM and halo.
domain components scale height (pc) relevant constituents Bran/Breg
III Thick 1800 nrel,B≤0.1
disk
II Reynolds 1000 nrel,B,nth ≤0.1
layer
I Thin 180 nrel,B,nth ≥1
disk
the regular magnetic field ([16],[17]).
The thermal electrons that cause the Faraday rotation of the synchrotron background are contained in the Reynolds
layer, with a scale height of about a kpc ([18]). The thermal electrons are more confined to the galaxy than the
relativistic electrons, so there is a “halo” above the Reynolds layer that only contains relativistic but no thermal
electrons.
These two mediums thus define three domains in the galactic ISM and halo with different characteristics, as sketched
in Table 1. The thin synchrotron disk (domain I) extends to a few hundred parsec, and is mixed with the lower parts
of the Reynolds layer and the thick disk. The Local Bubble is not taken into account. The upper part of the Reynolds
layer is also mixed with the thick synchrotron disk (domain II), whereas the highest part of the thick synchrotron disk
is so high above the galactic plane that it doesn’t contain a significant amount of thermal electrons anymore (domain
III). Our observations, in particular the very low upper limit on small-scale structure in total intensity I, put strict
constraints on the characteristics of these three domains.
The first constraint is a small scale height of the thin disk. Due to the large random magnetic field component in
the thin disk, the emissivity Ihas a fluctuating component. If the observed integrated emissivity of this fluctuating
component is only a few Kelvin, the I-structure can be averaged out if it is on small enough scales, i.e. if there are
enough “turbulent cells” along the line of sight. Therefore only a thin layer is allowed where the Bran component is
large, with structure on small enough scales to average out the small-scale structure in Ithat is created in this layer.
Typically, in a layer with a scale height of 200 pc, Bran-structure on scales of 10 to 20 pc can smooth Idown to the
observed limits.
The second constraint is a constraint on the magnetic field in the layers above the thin synchrotron disk. Small-scale
structure in Iemitted in these layers has to be negligible. Assuming equipartition in energy between the relativistic
electrons and magnetic field, the synchrotron emissivity ε ∝ B2. As the fluctuation in Irequires ∆ε <1%, this implies
that Bran/Breg <0.1. Therefore, the absence of small-scale structure in Idictates a regular magnetic field dominating
over the random magnetic field component by more than a factor ten in the halo of the galaxy, i.e. above a scale height
of a few hundred parsecs, at least for the componentof the magnetic field perpendicular to the line of sight.
The observed structure in polarized intensity is made by depolarization on small scales in the layers containing
thermal electrons, by three mechanisms. The first mechanism is beam depolarization, as discussed above. Second,
linearly polarized radiation emitted at different depths is Faraday-rotated by different amounts (differential Faraday
rotation [19]) and third, small-scale structure in thermal electron density and/or magnetic field causes spatial structure
in Faraday rotation (internal Faraday dispersion [20]). These depolarization mechanisms define a wavelength depen-
dent Faraday depth for the polarized radiation, which indicates that most of the observed polarized intensity comes
from the nearer part of the medium. This explains the difference between RM structure from the diffuse emission
and from extragalactic point sources, as the latter is built up over the whole path length through the Faraday-rotating
medium (domains I and II). A more quantitative discussion can be found in Haverkorn et al. ([10])
CONCLUSIONS
We observed two fields of over 50 square degrees, both located in the second galactic quadrant at positive latitude,
in the constellations Auriga and Horologium. All four Stokes parameters were derived from observations done at five
frequencies 341, 349, 355, 360, and 375 MHz simultaneously. The total intensity Iemission is featureless on scales
smaller than approximately a degree, while linear polarizations Qand Ushow abundant structure on arcminute to
degree scales. Polarized intensity has a maximum Tb,pol ≈15 - 18 K.
The observed structure in polarized intensity Pis ’cloudy’ on scales from arcminutes up to a degree. Long canals
of one synthesized beam wide where no polarization is detected are caused by beam depolarization in a beam which
separates two regions where the polarization angle changes abruptly by 90◦(or 270◦, 540◦etc).
Values of RM in the two fields are in general small: |RM| <15 rad m−2. RM maps show coherent structure in
RM over several independent beams up to a degree, but also sudden large RM changes across one beam of more than
100%. Not only the magnitude of the RM changes but regularly also the sign, which is most easily explained by a
change in direction of Bk. In the Horologium field, a ring-like structure in polarized intensity coincides with a disk of
radially increasing RM, although the RM disk is slightly larger than the ring in P.
The lack of observed small-scale structure in total intensity Iputs constraints on the medium (the galactic ISM
and halo). In the thin disk of synchrotron radiation, comparable to the stellar and HI disk, the random magnetic field
component Bran is comparable to or larger than the regular field component Breg. As Bran causes fluctuating I, the layer
cannot be thicker than a few hundred parsecs, and the structure has to be on small enough scales to average out the
created Ifluctuations. Furthermore, in the thick synchrotron disk above the thin disk, the random magnetic field has
to be very small: Bran ≤0.1Breg. Because of depolarization in the synchrotron emitting and Faraday-rotating medium
most of the polarized emission we observe originates from the nearby medium. As the RM of an extragalactic point
source is built up along the entire path length through the medium, the RM structure of the diffuse emission can be
uncorrelated with the structure in RM values of extragalactic point sources, as is observed.
ACKNOWLEDGMENTS
The Westerbork Synthesis Radio Telescope is operated by the Netherlands Foundation for Research in Astronomy
(ASTRON) with financial support from the Netherlands Organization for Scientific Research (NWO). This work is
supported by NWO grant 614-21-006.
REFERENCES
1. Wieringa, M. H., de Bruyn, A. G., Jansen, D., Brouw, W. N., and Katgert, P., Astronomy and Astrophysics,268, 215 (1993).
2. Uyanıker, B., Fürst, E., Reich, W., Reich, P., and Wielebinski, R., Astronomy and Astrophysics Supplement Series,138, 31
(1999).
3. Gray, A. D., Landecker, T. L., Dewdney, P. E., Taylor, A. R., Willis, A. G., and Normandeau, M., Astrophysical Journal,514,
221 (1999).
4. Gaensler, B. M., Dickey, J. M., McClure-Griffiths, N. M., Green, A. J., Wieringa, M. H., and Haynes, R. F., Astrophysical
Journal,549, 959 (2001).
5. Duncan, A. R., Haynes, R. F., Jones, K. L., and Stewart, R. T., Monthly Notices of the Royal Astronomical Society,291, 279
(1997).
6. Katgert, P., and de Bruyn, A. G., “Small-scale structure in the diffuse polarized radio background: WSRT observations at λ≈
90 cm”, in New Perspectives on the Interstellar Medium, edited by A. R. Taylor, T. L. Landecker, and G. Joncas, Astronomical
Society of the Pacific, 1999, p. 411.
7. Haslam, C. G. T., Stoffel, H., Salter, C. J., and Wilson, W. E., Astronomy and Astrophysics Supplement Series,47, 1 (1982).
8. Roger, R. S., Costain, C. H., Landecker, T. L., and Swerdlyk, C. M., Astronomy and Astrophysics Supplement Series,137, 7
(1999).
9. Haverkorn, M., Katgert, P., and de Bruyn, A. G., Astronomy and Astrophysics,4, L245 (2000).
10. Haverkorn, M., Katgert, P., and de Bruyn, A. G., Astronomy and Astrophysics (2002), submitted.
11. Sokoloff, D. D., Bykov, A. A., Shukurov, A., Berkhuijsen, E. M., Beck, R., and Poezd, A. D., Monthly Notices of the Royal
Astronomical Society,299, 189 (1998).
12. Leahy, J. P., Monthly Notices of the Royal Astronomical Society,226, 433 (1987).
13. Brouw, W. N., and Spoelstra, T. A. T., Astronomy and Astrophysics Supplement Series,26, 129 (1976).
14. Verschuur, G. L., The Observatory,88, 15 (1968).
15. Beuermann, K., Kanbach, G., and Berkhuijsen, E. M., Astronomy and Astrophysics,153, 17 (1985).
16. Rand, R. J., and Kulkarni, S., Astrophysical Journal,343, 760 (1989).
17. Ohno, H., and Shibata, S., Monthly Notices of the Royal Astronomical Society,262, 953 (1993).
18. Reynolds, R. J., Astrophysical Journal,339, 29 (1989).
19. Gardner, F. F., and Whiteoak, J. B., Annual Review of Astronomy and Astrophysics,4, 245 (1966).
20. Burn, B. J., Monthly Notices of the Royal Astronomical Society,133, 67 (1966).
