Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://star.arm.ac.uk/preprints/435.ps Äàòà èçìåíåíèÿ: Mon Jan 17 18:42:02 2005 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 21:27:45 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: coronal hole

Astronomy & Astrophysics manuscript no. Gl203 January 17, 2005
(DOI: will be inserted by hand later)
Line Broadening of EUV lines Across the Solar limb: A Spicule
Contribution?
J. G. Doyle 1 , J. Giannikakis 1,2 , L. D. Xia 1,3 and M.S. Madjarska 4,5
1 Armagh Observatory, College Hill, Armagh, BT61 9DG, N. Ireland
2 Sect. of Astrophysics, Astronomy and Mechanics, Dept. of Physics, Univ. of Athens, Athens 15783, Greece
3 School of Earth and Space Sciences, University of Science and Technology of China, Hefei, Anhui 230026, China
4 Max­Planck­Institut f˜ur Sonnensystemforschung # , Max­Planck­Str. 2, 37191 Katlenburg­Lindau, Germany
5 Department of Solar Physics, Royal Observatory of Belgium, Av. Circulaire 3, B­1180 Bruxelles, Belgium
Received date, accepted date
Abstract. Spectral lines formed in the solar transition region show an increase in the line width, peaking at #10,000 km above
the limb. Looking at a region o#­limb with no obvious spicules, the non­spicule region has a significantly smaller line width
above 6,000 km compared those taken in a spicule region. We suggest that this increase in line broadening is not due to small
scale random motions but rather to unresolved line shifts due to spicules and/or macro­spicules activity.
Key words. Sun: atmosphere -- transition region -- o#­limb -- line broadening -- spicules
1. Introduction
Line width measurements can provide important details on
small­scale mass motions and ion temperatures, and if coro­
nal lines are used, informations on coronal heating may be ob­
tained. Several authors have searched for disk center to limb
changes (Chae et al. 1998, Erdelyi et al. 1998, Doyle et al.
2000) finding a small variation. In o#­limb data, Banerjee et al.
(1998), Doyle et al. (1999), Harrison et al. (2002) and O'Shea
et al. (2003) have all used data relating to lines formed in the
corona, finding a small increase in the line width before reach­
ing a turn­over point. The data of Harrison et al. showed a sig­
nificant narrowing of a coronal line above 50,000 km. which
the authors suggested was related to the dissipation of wave
energy. However, O'Shea et al. (2005) has shown that the line
widths start to show a decrease in their values at exactly the
same location where the dominant excitation changes from
being collisionally to radiatively dominant. For lines formed
around 100,000 to 300,000 K, several authors, e.g. Mariska et
al. (1979), Peter & Vocks (2003), have noted an increase in
the line width at 10 to 15 ## above the limb. Mariska et al. sug­
gested that this broadening was unlikely to be simply due to
an increase in the wave flux above the limb and proposed that
inhomogeneous structures could be the cause. More recently,
Peter & Vocks (2003) interpreted the increase as evidence of a
large increase in the ion temperature to more than 3 â 10 6 K
just above the limb. Here, we look at raster and time series data
Send o#print requests to: J.G. Doyle, email: jgd@arm.ac.uk OR
http://star.arm.ac.uk/preprints/
# former Max­Planck­Institut f˜ur Aeronomie
from lines formed around 200,000 K, suggesting an explana­
tion in terms of spicules. In Sect. 2 we discuss the observational
data which consists of both rasters and a time series, with the
results presented in Sect. 3.
2. Observational Data
2.1. Rasters
We used a raster sequence of the north solar limb (PCH) taken
by the spectrometer SUMER on­board the SoHO satellite. The
capabilities and specifications of the SUMER instrument were
described by Wilhelm et al. (1995, 1997) and Lemaire et al.
(1997). The observation was performed on 1996 August 10
from 00:03 to 16:09 UT. The target was the north polar coronal
hole region with a constant SoHO solar Y at 950 ## and SoHO
solar X moving from --699 ## to 721 ## . The exposure time was
60 s using slit 2 (i.e. 1 ## x 300 ## centered) with a step size of
1. ## 5. Detector A was used for producing the four 50 spectral
pixel windows at the wavelengths corresponding to the 2nd or­
der spectral lines: Mg # 624.94 å, O # 629.73 å and to the 1st
order: N # 1238.82 å, Fe ### 1242.01 å. Here, we select only
the O # (T#250000 K) transition region line.
We used the standard SUMER data reduction procedures to
apply all the corrections needed for the data. These corrections
are dead time and local gain correction, flat field subtraction,
and a correction for geometrical distortion. Since our interest
in this study was focused on the line widths we did not per­
form a wavelength calibration. Additionally a correction for the
spectral line shift caused by thermal deformations of the optical
bench of SUMER was applied (Dammasch et al. 1999).

2 Doyle et al.: Line Broadening of EUV lines O#­limb
Fig. 1. A sub­set of the image as obtained in O # 629 å on 10 August 1996 in the coronal polar region showing the position of the three data
plots given in Fig. 2. The scale on both axes are in arcsec.
Fig. 2. Non­thermal velocities as derived from the O # 629 å line calculated for three positions along the raster. Each point was derived by
averaging 21 pixels in the X­direction and a running mean of 9 pixels in the Y­direction. Data is only plotted up to pixel 220 along the slit. Also
included are plots of the continuum region close to the O # line position. The vertical line shows the position of the continuum limb.
For the line of interest i.e. O # 629 å, we performed a
one line Gaussian fit using the automated SolarSoft routine
XCFIT BLOCK. As a result, a set of Gaussian line parameters
(intensity, FWHM and position) was available for each pixel
within the raster. For studying the variations of the line width
as we approach the limb from the disk and also the behavior
in the o# limb areas we analyzed vertical stripes (parallel to
the slit) producing plots which show these variations versus the
Solar Y­coordinate. In order to increase the counts in the line
profile, we averaged 21 pixels in the X­direction and a running
mean of 9 pixels in the Y­direction.
2.2. Time Series
The data selected for this study were obtained as a time se­
ries in a polar coronal hole by SUMER/SoHO on 25 February
1997 starting at 00:03 UT. During the observation, the SUMER
slit was fixed at positions solar X = 0 ## and y=­950. ## 25. Slit 2
(1 ## â300 ## ) and detector B were used. The slit width determines
the spatial resolution along the X­direction, while the resolu­
tion element along the slit in the Y­direction (north­south; pos­
itive towards north) is approximately 1 ## , given by the pixel size
of the detector. The exposure time was 60 s. The spectral line
observed was N ## 765 å (T#140000 K).
In addition to the data analysis steps already mentioned,
we used a di#erent method to deduce the line parameters (ra­
diance, central position of the spectral line and width). This
method is useful when dealing with reduced counts or large
datasets. The procedure has being frequently used to obtain
SUMER Dopplergrams (see details in Dammasch et al. 1999)
and the results are statistically consistent with those obtained
by using standard Gaussian fitting program (Xia 2003). Here
the central position for every pixel is derived by integrating the
line radiance across a certain spectral window and determining
subsequently the location of the 50 % level with sub­pixel ac­
curacy. As a check, we also used this procedure in the raster
data, finding a similar result to that obtained from the Gaussian
fits.
For Doppler shifts of the N ## 765 å line, the zero velocity
is set to the value averaged over the whole period of the obser­
vation (794 time steps) at a fixed spatial pixel. The limb posi­
tion is defined as that derived based on the continuum short­
ward of the N ## line (see Xia et al. 2005 for more details).

Doyle et al.: Line Broadening of EUV lines O#­limb 3
Fig. 3. A selection of spicules and macro­spicules showing the varia­
tion of the N ## 765 å intensity, non­thermal velocities and line­shift
against height above the limb. The PCH was observed on 25 February
1997 between 00:03 and 13:58 UT. The times shown beside the curves
are related to the starting time of the observation. Those at t=452 &
735 min are macro­spicules, while the others are spicules.
3. Results
In Fig. 2 we plot the non­thermal velocities at three locations
along the X­direction in the raster as shown in Fig. 1; i.e. posi­
tion --133 ## , --59 ## and +15 ## , with the data being averaged over
21 pixels in X and 9 in Y. Here, we assume ionization equi­
librium and that the ion temperature is identical to the electron
temperature where the FWHM of the line is given by
FWHM =
# (## inst ) 2 + 4ln2 # 2
0
c 2
2kT
M + # 2 # (1)
## inst is the instrumental width, # 0 is the unshifted wavelength,
c the speed of light, k the Boltzmann constant, T the ion tem­
perature, M the atomic mass and # the non­thermal velocity.
The line was corrected for instrumental broadening using the
SolarSoft routine: CON WIDTH FUNCT 3.
In each plot, we clearly see a peak in the non­thermal veloc­
ity at #15 ## above the limb as seen in the continuum short­ward
of O # 629 å. In the 450 and 500 plots, we see an additional
broadening at # 25--30 ## . In order to gain some further insight
into the nature of this o#­limb broadening, we must look at the
time series data. In Fig. 3 we show the velocity profiles (non­
thermal and Doppler shift) derived from the N ## 765 å line as
a function of height above the limb. Despite the fact that N ##
is formed at around 140,000 K compared to O #'s 250,000 K,
the non­thermal velocity variation is similar. It reaches maxi­
mum around 5 ## o#­limb and remains at this value until around
18 ## , shows a slight decrease before rising again around 25 ##
o#­limb.
Like the line radiance, the Doppler velocities are highly
structured with a time scale down to 1 minute. Among them
two examples (t = 452 min and t = 735 min) were identified as
macro­spicules (Xia et al. 2005). Others (t =34, 261, 546, 630
min) are deduced as being `normal' spicules.
In Fig. 3, one finds that the Doppler shifts of all selected
structures are small (around ±5 km s -1 or smaller) just above
the limb, then quickly increases with height. After an initial
acceleration, the velocity reaches a rather constant value, al­
though with some fluctuation. The low velocity in this early
stage of spicule evolution has also been found with CDS ob­
servations (Pike & Harrison 1997, Pike & Mason 1998). We
suggest that the observed increase in the line broadening is not
due to small scale motions but rather to unresolved line shifts
due to spicules around 10 to 15 ## , and then macro­spicules fur­
ther o#­limb.
Fig. 4 shows a plot of the non­thermal velocity above the
limb, taken from a region without obvious spicules (dotted line)
and the whole observed data averaged (solid line). The solid
line is the non­thermal velocity averaged from all the data, i.e.,
the line profile at every Y pixel is averaged across the entire
794 time series, then getting the line parameter from this re­
binned profile. The dotted line is the non­thermal velocity av­
eraged across a dark region from 554 min to 557 min. Again,
after getting an average line profile at every Y pixel, then the
line width. The non­spicule region has a peak non­thermal ve­
locity between 7 and 10 ## o#­limb, and shows a significantly
smaller non­thermal velocity above 10 ## o#­limb than that from
the spicule region.
Note that the non­thermal velocities shown in Figs. 3 and
4 (obtained by SUMER detector B) are systematically larger
than those shown in Fig. 2 (obtained by the SUMER detector

4 Doyle et al.: Line Broadening of EUV lines O#­limb
A). This is possibly because of an insu#cient subtraction of
the instrumental broadening of the detector B, as discussed by
Popescu et al. (2004).
Fig. 4. A plot of non­thermal velocities above the limb, taken from
a time­region without obvious spicules (dotted line) and the whole
observed data averaged (solid line).
4. Discussion
There are many suggestions for the excess broadening of tran­
sition region lines, e.g. acoustic waves, Alfven waves, opac­
ity, turbulence, etc. Dere (1989) showed that the power in un­
resolved velocity variations (from line width measurements)
was greater than that predicted from the extrapolated power of
the resolved velocity variations (from line shift measurements),
therefore suggesting the unresolved motions could be driven by
a process that is di#erent from those producing the line shifts.
The idea behind the present study was not to explain the gen­
eral broadening in excess of the thermal width, but rather to
explain the additional increase in broadening seen in transition
region lines about 10,000 km above the limb. This seems to be
confined to a region of #3,000 km.
Tu & Marsch (1997) suggested that ion­cyclotron is an im­
portant process in the solar wind. Peter & Vocks (2003) have
more recently suggested that ion­cyclotron could be a possible
mechanism to explain this additional line broadening above the
limb. Although this is an interesting idea, it is di#cult to under­
stand why it should be confined to such a small region. In the
analysis of transition region lines, Chae et al. (1998) and Doyle
et al. (2000) both noted a 2--3 km s -1 di#erence in the line
width from disk center to the limb. This could be explained via
an increase in opacity from zero at disk center to unity at the
limb. Doyle & McWhirter (1980) showed many years ago that
some transition region lines were slightly e#ected by opacity at
the limb. However, to produce a 10 km s -1 increase via opacity
would imply unrealistic high optical depths. The above authors
also showed that the center­to­limb increase in line width could
be reproduced assuming the presence of mass flows with a most
probable speed of 5 km s -1 .
The present results suggest that spicule flows could play
a role in line broadening. Macro­spicules (assumed to be the
large­scale version of spicules) come in two types; erupting
loops and spiked­jets. Yamauchi et al. (2004) found that 43%
are of the erupting­loop type while 49% were the single­
column spiked jet. However, even the erupting­loop type pro­
duces two columns when the loop top rises and probably recon­
nects with open­field structures. The velocities of both types of
macro­spicules are in the range 32 to 42 km s -1 . It is expected
that the velocities in spicules are smaller than these values. This
is consistent with the observations that the spicules velocity
just above the limb is small and quickly under­goes accelera­
tion just above the limb. Tanaka (1972) found that 30% of H#
spicules produced a double­column structure, hence adding an
increasing amount of line shift. The spicule contribution to the
line widths is confirmed in Fig. 4 which shows that the line
width taken from a region without obvious spicules is substan­
tially smaller above 10 ## than that from a region with spicules.
Acknowledgements. Research at the Armagh Observatory is grant­
aided by the N. Ireland Dept. of Culture, Arts and Leisure.
LDX is grateful for a PRTLI research grant for Grid­enabled
Computational Physics of Natural Phenomena (Cosmogrid) and JG
to PPARC for funding via the Armagh Observatory's visitors grant
PPA/V/S/1999/00628. This work was also supported in part by
PPARC grant PPA/G/S/2002/00020. We thank Georgia Tsiropoula for
valuable comments on an earlier draft.
References
Banerjee, D., Teriaca, L., Doyle, J.G. &Wilhelm, K., 1998, A&A 339,
208
Chae, J., Schuhle, U. & Lemaire, P., 1998, ApJ 505, 957
Dammasch, I.E., Wilhelm, K., Curdt, W. & Hassler, D.M., 1999,
A&A, 346, 285
Dere, K.P., 1989, ApJ 340, 599
Doyle, J.G. & McWhirter, R.W.P., 1980, MNRAS 193, 947
Doyle, J.G., Teriaca, L. & Banerjee, D., 1999, A&A 349, 956
Doyle, J.G., Teriaca, L. & Banerjee, D., 2000, A&A 356, 335
Erdelyi, R., Doyle, J.G., Perez, M.E. &Wilhelm, K., 1998, A&A 337,
287
Harrison, R.A., Hood, A.W. & Pike, C.D., A&A 392, 319
Lemaire, P., Wilhelm, K., Curdt, W., et al. 1997, Sol. Phys., 170, 105
Mariska, J.T., Feldman, U. & Doschek, G.A., 1979, A&A 73, 361
O'Shea, E., Banerjee, D. & Poedts, S., 2003, A&A 400, 1065
O'Shea, E., Banerjee, D. & Doyle, J.G., 2005, A&A (submitted)
Peter, H. & Vocks, C., 2003, A&A 411, L481
Pike, C.D. & Harrison, R.A., 1997, Sol Phys 175, 457
Pike, C.D. & Mason, H.E., 1998, Sol Phys 182, 333
Popescu, M.D., Doyle, J.G. & Xia, L.D., 2004, A&A 421, 339
Tanaka, K., 1972, Big Bear Obs. Rep. 125
Tu, C.­Y. & Marsch, E., 1997. Sol Phys 171, 363
Wilhelm, K., Curdt, W., Marsch, E., et al. 1995, Sol. Phys., 162, 189
Wilhelm, K., Lemaire, P., Curdt, W., et al. 1997, Sol. Phys., 170, 75
Xia, L.D., 2003, PhD Thesis, Georg­August­Univ., G˜ottingen
Xia, L.D., Popescu, M.D., Doyle, J.G. & Giannikakis, J., 2005, A&A
(submitted)
Yamauchi, Y., Moore, R.L., Suess, S.T., Wang, H. & Sakurai, T., 2004,
ApJ 605, 511