Ubiquitous Suprathermal Tails on the Solar Wind and
Pickup Ion Distributions
George Gloeckler
Department of Physics and IPST, University of Maryland, College Park, Maryland 20742, USA
Abstract. Using time-of flight mass spectrometers on Ulysses and ACE we measure the velocity distributions of particles in the poorly explored suprathermal range that extends from about twice to fifty times the solar wind speed. We
find that H+, He++ and He+ are always present in this energy (or speed) range at variable intensity levels that depend on
solar wind conditions. Contrary to most common expectations, even during most quiet times, in the absence of shocks
and other interplanetary disturbances, suprathermal power-law tails on solar wind distributions are observed that extend
to the highest energies measured. Suprathermal tails are observed in distributions of solar wind and interstellar pickup
ions, not just for protons but also in distributions of all heavy ions that can be measured. Tails are seen in the quiet and
disturbed slow wind, as well as in the super-quiet fast wind from polar coronal holes. Instead of diminishing rapidly
with heliocentric distance, tails are found to persist to at least 5.4 AU and become harder. We suggest that preaccelerated particles contained in these tails may well be the population that is injected for further acceleration by
shocks. It remains unknown what processes produce these tails and whether these tails persist to very large distances.
INTRODUCTION
Pre-accelerated particles contained in these tails may
well be the population that is injected for further
acceleration by shocks. We review observations from
Ulysses and ACE of energy spectra of solar wind and
pickup ions observed in the fast and slow quiet solar
wind, and contrast these with those found in
association with heliospheric shocks. Explaining the
origin of these ubiquitous tails remains a challenge.
There is little dispute that shocks are among the
prime accelerators of particles in and beyond our solar
system. A prime example is Anomalous Cosmic Rays
that are most likely accelerated to their highest energies (hundreds of MeV) by the heliospheric termination shock. Yet a major difficulty with the shock
acceleration mechanism is the injection problem.
Shocks alone may not be able to efficiently accelerate
particles below some threshold energy of about 100
keV/nuc. Observations of heliospheric suprathermal
ions begin to provide some answers. We find that in
the quiet solar wind, in the absence of all kinds of
interplanetary shocks and strong turbulence associated
with such shocks, in the absence of waves and compression regions, etc., solar wind and interstellar
pickup ion velocity distributions have pronounced and
persistent high-energy tails. The distribution functions
of these ions are non-maxwellian with strong powerlaw tails to the highest energies observed. These tails
are not generated in the process of solar wind acceleration because they also appear in distributions of
interstellar pickup ions that are absent near the Sun.
They are observed not just for protons but also in
distributions of all heavy ions that were measured.
OBSERVATIONS OF SUPRATHERMAL
TAILS
Measurements presented here were made using the
Solar Wind Ion Composition Spectrometer (SWICS)
instruments on Ulysses (1) and ACE (2). SWICS is an
ion mass vs. mass/charge spectrometer in which
energy/charge analysis, followed by post-acceleration,
is combined with a time-of-flight and energy measurement to determine the mass/charge, mass and
energy of ions from ~0.6 to 60 (Ulysses) and ~100
(ACE) keV/charge. The Ulysses orbit spans heliocentric distances between 1.4 and 5.4 AU and all latitudes
from +80° to –80°. ACE orbits around L-1 at ~1 AU
in the ecliptic plane.
CP679, Solar Wind Ten: Proceedings of the Tenth International Solar Wind Conference,
edited by M. Velli, R. Bruno, and F. Malara
© 2003 American Institute of Physics 0-7354-0148-9/03/$20.00
583
F(W) Phase Space Density
(s3/km6)
SWICS Ulysses
10
9
10
7
H
3
1
10-1
He++
2
He
+
Ftail = 6•(W-1) - 5.6
60
Ftail = 0.4•(W-1) - 5.1
10
-3
1
2
4
W Ion Speed/Solar Wind Speed
7
10
14
10
10
10
12
He
10
8
O
10
6
Fe
10
4
10
2
100
SWICS
10
3
F=Const•E -1.75
ULEIS
10
4
10
5
SIS
10
6
10
7
Kinetic Energy (eV/nucleon)
CRIS
10
8
10
9
such as shocks and compression regions of CIRs that
are known to produce distributions with strong suprathermal tails (4, 5) and to accelerate particles (see
e.g. reference 6). These accelerated particles, however, would rapidly lose energy by adiabatic cooling in
the expanding solar wind and the suprathermal tails of
their distributions would decay in a matter of days.
+
Ftail = 24•(W-1) - 5.1
10
-5
Rave = 5.26 AU
th
10
1016
FIGURE 2. Fluence of He, O, and Fe during a 33-month
long time period at 1 AU during nearly the same phase in the
solar cycle as in Figure 1. (From reference 3). These are the
average spectra seen by the bow shock of Earth. No time
periods were excluded from these averages.
V = 18.6 km/s
5
He, O, and Fe Fluences
(10/97 to 6/00)
1018
102
VSW = 403 km/s
10
10
Fluence (particles/cm sr-MeV/nucleon)
In order to establish a baseline against which other
spectra will be compared, we show in Figure 1 the
average velocity distributions of H+, He+ and He++
that were obtained during a two-year time period when
Ulysses was in the slow wind, near its aphelion at ~5
AU, and close to the ecliptic plane. The spectra are
complicated, being a mixture of several distinct
components: (a) the solar wind which peaks at W ≈ 1
and has a non-maxwellian distribution, (b) the
relatively flat spectra of the interstellar pickup ions
with a cutoff at W ≈ 2, and (c) well developed
suprathermal tails above W ≈ 2 that are reasonably
good power laws in the solar wind frame with indices
of 5.1 for H+, He++ and 5.6 for He+. In these tails there
is more He+ than He++ and the H+/He++ ratio of 60
indicates acceleration occurs at W ~ 1.5 (where this
ratio is also 60) and not the at the bulk SW, at W = 1,
where the H+/He++ ratio is considerably larger.
In order to establish the time history of particle
densities in the suprathermal tail regions we show in
Figure 3 the count rate of all ions (mostly protons)
with speeds above 2.3 times the ambient solar wind
speed (i.e. W > 2.3). The top panel is for a time period
during the declining phase from solar maximum when
forward-reverse shocks and compression regions of
CIRs were common. Acceleration of suprathermal
ions by shocks (indicated by solid and dotted vertical
lines) is clearly visible especially for the strongest
shocks and in CIRs between forward-reverse shock
pairs, such as the one identified by the shaded area in
the panel. Sometimes there is modest acceleration
when no shocks are observed, for example, between
DOY91 ~110 to ~115, ~175 to 185 and 192 to 200.
However, there are few, if any, 6-hour intervals when
no suprathermal ions are seen.
10
FIGURE 1. Phase space density of H+, He++ and He+ vs. W,
the normalized ion speed. These baseline distributions,
measured with SWICS on Ulysses, were averaged over a two
year period (1997.108-1999.108). No time periods were
excluded. These represent the average typical spectra at the
bow shock of Jupiter during a 2-year period.
The distributions of every species measured show
power law tails extending to 50 times the solar wind
speed. This is illustrated in Figure 2 (from reference
3) which shows the fluence of He, O and Fe measured
with mass spectrometers on ACE. In these long-term
averaged spectra we see the slow and fast solar wind,
interstellar pickup He (the small bump at about
2•104 eV), particles accelerated in CIRs and by traveling shocks, and solar energetic particles. Suprathermal tails extend to energies beyond tens of MeV and
are power laws between ~20 to 2,000 keV/nuc. The
power law index is ~1.75 which is equivalent to 5.5 for
phase space density vs. ion speed. The power law index is about the same for He, O and Fe. These distributions were averaged over long time periods and thus
include contributions from particles of various sources,
In the middle panel we show a 110 day time period
in 1998 when Ulysses was at low latitudes, at ~5 AU
in the slow wind during the ascending phase from
solar minimum. CMEs were predominant during that
phase of the solar cycle. The maximum rate observed,
indicated by the arrow head on the left, is almost a
factor of 10 lower than the maximum rate in 1991.
During counterstreaming electron events, most likely
CMEs, the rates are low. The highest rates are
584
–
–
–
–
*
**
SWICS Ulysses
*
*
–
**
–
*
**
–
*
*
4
**
10
(a)
102
Counts in 6 hours
100
100
120
140
DOY 1991
160
180
200
4
10
(b)
2
10
0
10
280
300
4
10
320
DOY 1998
340
360
380
North Polar Coronal Hole
W > 2.3
2
10
(c)
0
10
140
160
180
DOY 1995
200
220
240
FIGURE 3. Counts in 6 hours of ions (mostly protons) in the suprathermal regions (W > 2.3) of the velocity distributions in
three 110-day long periods: (a) during the declining phase from solar maximum in 1991 at low latitudes between ~3 and ~4 AU;
(b) during the ascending phase from solar minimum in 1998 at low latitudes and ~5 AU; and (c) at solar minimum in 1995 at
high latitudes in the north polar coronal hole. In panel (a) forward shock times are indicated by dotted vertical lines and reverse
shocks by solid lines (from a list in reference 7). Approximate shock strengths are indicated by ‘**’ for Ms (sonic Mach
number) ≥ 2, ‘*’ for 1.5 ≤ Ms < 2 and ‘–’ for Ms < 1.5. In panel (b) counterstreaming electron events (J. Gosling, private
communication) are indicated by the hatched vertical bars.
observed behind shocks (e.g. around DOY98.331).
There are hardly any 6-hour intervals when no
suprathermal ions (W > 2.3) are seen.
tails as is illustrated in Figure 4 (see also references 4
and 5). For all three species the distribution functions
between the forward–reverse shock pairs of a CIR are
quite similar and the suprathermal tails are power laws
(in the solar wind frame), each having the same index
of 4.5. The spectra directly behind the forward shock
of November 27, 1998 (most likely associated with the
CME that follows it (see panel (b) of Figure 3) are also
similar to one another. The suprathermal tails (above
W ≈ 2.3) are again power laws. The indices, while
close to 4.5, are somewhat different among the three
ion species. The tail spectra of all the species in both
the CIR and CME associated shocks are harder (less
steep) than the tails in the 1997-1999 baseline
distributions shown in Figure 1.
The bottom panel of Figure 3 shows the counting
rate of suprathermal ions during a 110-day period in
1995 when Ulysses was at high latitudes in the north
polar coronal hole at solar minimum. No shocks were
recorded during the entire time period. The rates are
very low but steady. Detailed mass vs. mass/charge
analysis of the triple coincidence pulse-height data
clearly shows that the counts are real (H+, He++ and
He+) and not due to some instrumental background.
Clearly suprathermal particles are present in the superquiet solar wind, far removed from any shocks.
In Figure 5 we compare the baseline proton distribution from Figure 1 and that measured from 1991.088
to 1993.108 during the declining phase from solar
maximum, to the CIR shock-associated H+ spectrum
shown in the top panel of Figure 4. As expected, the
Velocity Distribution Behind Shocks
Shocks are powerful particle accelerators. They
heat the solar wind and produce strong suprathermal
585
SWICS Ulysses
10
4
Between F-R Shocks
R = 3.54 AU
Vth = 47.4
V = 550 km/s
SW
(s3/km6)
2
10
0
Phase Space Density
60
10
10-2
10
6
10
4
10
2
10
0
F
tail
= Const•(W-1)
- 4.5
H+
60
He
+
He
++
1998.331.00:00-333.12:00
After F- Shock
R = 5.25 AU
Vth = 40.1
V = 498 km/s
F
SW
tail
10-2
(s3/km6)
6
F(W) Phase Space Density
10
SWICS Ulysses
1991.185.00:00-188.18:00
= Const•(W-1)
+
He
γ = 4.9
0.6
0.8
1
+
3
5
1991.185.00-188.18
between F-R Shocks
γ = 4.5
103
101
10-1
V
SW
1991.088-1993.108
= 500 km/s V = 21.3 km/s
th
γ = 4.5
10-3
10-5
1997.108-1999.108
V = 403 km/s V = 18.6 km/s
SW
1
th
2
γ = 5.1
4
W Ion Speed/Solar Wind Speed
7
10
only those times which have less than 4 counts in any
six-hour period. This selection threshold is indicated
by the dotted horizontal lines in all panels of Figure 3.
It is evident from Figure 3 that most of such time
periods are far removed from shocks.
He
γ = 4.0
W Ion Speed/Solar Wind Speed
10
H+
FIGURE 5. Two-year averaged baseline distributions of H+
(1991.088-1993.108 and 1997.108-1999.108) and the proton
spectrum in the CIR compression region from Figure 4.
-γ
H
γ = 4.4
++
10
7
5
The proton velocity distributions during quiet times
in the CME-dominated slow wind (1997-99) and in the
super-quiet wind of the two polar coronal holes are
shown in Figure 6. The intensity in the tail of the
quiet-time slow wind is down by about a factor of 5 to
6 from the baseline (dashed curve) tail intensity, but
the spectral shape is the same for both. In the polar
coronal holes a power-law suprathermal tail is also observed. However, the tail found in polar coronal holes
is very weak and soft (steep) with an index of about 8.
It seems extremely unlikely that these quiet-time tails,
especially in the polar coronal holes, are remnants of
particles accelerated by remote shocks.
FIGURE 4. Averaged velocity distribution of H+, He+ and
He++ associated with CIR forward and reverse shocks (top
panel) and a CME driven shock (bottom panel). The
averaging time periods are indicated in each panel and
correspond to the dark shaded regions in Figure 3. The
location of Ulysses and average solar wind bulk and thermal
speeds are also listed in each panel.
most intense (strongest) tails are found in the shock associated proton distribution which is also characterized
by the highest thermal speed. The 1991-93 baseline
tail has the same power law index as the CIR tail but is
about a factor of 10 weaker. The 1997-99 baseline
distribution has the steepest and weakest tail, and the
lowest solar wind thermal speed compared to the two
other populations. This implies that the CIR-dominated solar wind of 1991 to 1993 is more effective in
heating the solar wind and producing suprathermal
tails than the CME-dominated solar wind of 1997 to
1999 following solar minimum.
Quiet-Time Tails at 1 AU and 5.3 AU
We examine changes in the quiet-time suprathermal
tails with heliocentric distance by comparing in Figure
7 the spectra of protons measured at 1 AU by SWICS
on ACE in 1999, with that measured at 5.26 AU by
SWICS on Ulysses during 1997.108-1999.108, the
CME-dominated phase of the solar cycle. The
velocity distributions differ in several respects. The
solar wind at 5.26 AU is cooler than it is at 1 AU.
Pickup H+, invisible at ACE, is clearly seen at 5.26
AU for 1.4 < W < 2.2. But the most interesting result
is that the quiet-time proton tail spectrum above
W ≈ 2.3 at 5.26 AU is stronger and harder than at 1
AU. Evidently, suprathermal proton tails, rather than
Suprathermal Tails During Quiet Times
In the super-quiet period in the polar coronal hole
wind during solar minimum the count rate of ions
above W = 2.3 never exceeded 3 counts in a six-hour
interval as was shown in Figure 3(c). We use this
maximum rate to select quiet times in the turbulent
slow solar wind (see top two panels of Figure 3) to be
586
diminishing due to adiabatic cooling, continue to be
regenerated in the expanding solar wind, even during
quiet times in the absence of shocks.
SWICS Ulysses and ACE
H+
107
Quiet-time Slow Solar Wind
F(W) Phase Space Density
3
6
(s /km )
109
DISCUSSION
F(W) Phase Space Density
(s3/km6)
We have shown that everywhere in the heliosphere
where measurements could be made, suprathermal
tails, extending to the highest measurable energies, are
always present. They exist not only in the baseline
solar wind, but also during quiet times far removed
from shocks that are well known to accelerate
particles. Suprathermal tails are seen even in the super-quiet solar wind of polar coronal holes. They are
observed in solar wind and pickup ion distributions,
10
H
6
+
Baseline
Quiet
102
100
10-2
10-4
10
-6
Polar Coronal Holes
γ = 8.0
Ftail = Const•(W-1) - γ
1
2
4
W Ion Speed/Solar Wind Speed
7
101
10-1
10
-3
10
-5
5.26 AU
γ = 5.1
60 quiet days in 1999
1 AU
γ = 6.2
Ftail = Const•(W-1) - γ
1
2
4
7
W Ion Speed/Solar Wind Speed
10
the way to the termination shock, then ions in these
tails would be the seed populations for the Anomalous
Cosmic Rays (ACRs) accelerated by the termination
shock. In Figure 8 we show that this may indeed be
the case. The quiet-time He+ spectrum measured at ~5
AU was projected to the termination shock assumed to
be at 100 AU. A least- squares power-law fit to the
suprathermal tail between 5 and 19 keV/nuc, extrapolated to 1 MeV/nuc connects to the lowest two He
points measured at 77 AU by Voyager 1 (9) which are
most likely the high-energy end of the He+ tail at 77
AU rather than the He ACRs. Presence of suprathermal tails at the termination shock would solve the injection problem for termination shock acceleration of
ACRs, especially in the strong- shock case.
Slow Solar Wind
1997.108-1999.108
γ = 5.1
104
1997.108-1999.108
(x27.7)
103
FIGURE 7. Quiet-time spectra of protons at 1 AU and 5.26
AU in 1999. The Ulysses spectrum is that from Figure 6
multiplied by R2 = 27.7. During the CME-dominated phase
the tail at Ulysses is harder (less steep).
SWICS Ulysses
108
105
10
FIGURE 6. Quiet-time proton spectra in the CME-dominated slow solar wind (1997.108-1999.108) and in polar
coronal holes (1994.001-365 and 1995.090-1996.210). The
1997-99 baseline proton suprathermal tail (from Figure 1) is
indicated by the dashed curve.
SUMMARY AND CONCLUSIONS
and for light and heavy ions. These tails are powerlaws in speed (or energy/nucleon) with similar indices
for different species. The hardest spectra and the
strongest tails are found in shock-associated distributions, the softest and weakest tails in the super-quiet
wind of solar-minimum polar coronal holes.
Long term averages of velocity distributions (baseline distributions) of H+, He+, and He++ in the slow
solar wind show well developed tails that are
approximately power-laws in speed W (in the solar
wind frame) with indices between ~5 and ~5.5. These
baseline distributions include particles accelerated in a
variety of ways. In the tails, the intensity of He+ is
higher than that of He++ implying that ions in the tails
come predominantly from outside the thermal part of
the solar wind distribution (W ≈ 1), e.g. from portions
of the distribution where W > ~1.5. Power-law tails of
baseline He, O and Fe distributions at 1 AU have the
same indices, ~5.5, and extend to energies of several
MeV/nuc (speeds up to 50 times the solar wind speed).
The ions contained in the quiet-time tails are most
likely to be the particles accelerated most efficiently
by CIR and CME shocks, as well as bow shocks of
planets such as Earth and Jupiter. Since the tails extend to beyond 10s of keV, they provide a pre-accelerated population, thus solving the shock injection
problem. If tails continue to be generated in the quiet
wind beyond the ~5 AU of present observations, all
587
-1
2
Differential Flux (cm s sr MeV/nuc)
ACKNOWLEDGMENTS
Pickup He
1997.108-1999.108
x(1/20)
+
10
5
10
3
10
1
10
-1
10
-3
10
-5
The essential contributions of the many individuals
(see reference 1 and 2) at the Universities of Maryland
and Bern, the Max-Planck-Institut für Aeronomie, and
the Technical University of Braunschweig who
contributed to the success of the SWICS experiments
on Ulysses and ACE are most gratefully acknowledged. I thank L. A. Fisk, N. A. Schwadron and T. H.
Zurbuchen for stimulating discussions, G. M. Mason
for kindly provided the time periods of the 60 quiet
days in 1999, and Christine Gloeckler for her essential
help with data reduction. This work was supported in
part by NASA/Caltech grant NAG5-6912 and
NASA/JPL contract 955460.
Baseline
strong
shock
Quiet-time
Ftail = A•E•(E 1/2 - 1)-6.9
ACR
He
dj/dE ≈ B•E -2.45
0.001
0.01
0.1
weak
shock
1
Energy (MeV/nuc)
10
100
REFERENCES
FIGURE 8. Differential intensity vs. energy per nucleon of
baseline He+ (from Figure 1) and quiet-time He+ during 1997
to 1999, and ACR He measured by Voyager 1 at ~75 AU in
1998-1999.182 (8), open circles, and in 2000 (9), open
squares. Dashed and dotted curves are respectively strong
and weak shock model fits to the 1998-99 ACR He (9). The
He+ distributions were reduced by a factor of 20 to account
for the 1/R decrease in density of interstellar pickup helium
at the termination shock at ~100 AU. The solid curve is the
fit of the form indicated in the figure to the quiet-time He+
spectrum from 5 to 19 keV/nuc, extended to 1 MeV/nuc.
1. Gloeckler, G., et al., Astron. Astrophys. Suppl. Ser. 92,
267-289 (1992).
2. Gloeckler, G., et al., Space Sci. Rev. 71, 79-124 (1995).
3. Mewaldt, R. A., et al., “Long-Term Fluences of
Energetic Particles in the Heliosphere” in Solar and
Galactic Composition, edited by R. F. WimmerSchweingruber, AIP Conference Proceedings 598, New
York: American Institute of Physics, 2001, pp. 165-169.
Tails are observed during quiet time periods in the
slow solar wind. These periods are far removed from
shocks, waves, CIR compression and high turbulence
regions which are known to accelerate particles. The
intensity of quiet-time tails is lower by a factor of 5 to
6 from baseline but their power-law index is the same
as baseline. Power-law tails are present in polar
coronal holes, where no shocks, waves or CIRs are
observed. Compared to baseline distributions, these
super-quiet tails are weaker by factors of 100 or more
and steeper (index ~8). Quiet-time tails persist undiminished to distances of at least 5 AU (as sampled by
Ulysses) and become harder (index decreases).
Because quiet-time tails persist to large distances and
are seen also for pickup He, we can rule out that they
are formed during solar wind acceleration close to the
Sun. What mechanism produces the quiet-time
suprathermal tails still remains an open question.
4. Gloeckler, G., Space Sci. Rev. 89, 91-104 (1999).
5. Gloeckler, G., Fisk, L. A., Zurbuchen, T. H., Schwadron,
N. A., “Sources, Injection and Acceleration of
Heliospheric Ion Populations” in Acceleration and
Transport of Energetic Particles Observed in the
Heliosphere, edited by R. A. Mewaldt et al., AIP
Conference Proceedings 528, New York: American
Institute of Physics, 2000, pp. 221-228.
6. Mason, G. M., and Sanderson, T. R., Space Sci. Rev. 89,
77-90 (1999).
7. Balogh, A., Gonzales-Esparza, J., A., Forsyth, R. J.,
Burton, M. E., Goldstein, B. E., and Bame, S. J., et al.,
Space Sci. Rev. 72, 171-180 (1995).
8. Cummings, A. C., Stone, E. C., and Steenberg, C. D.,
Astrophys. J. 578, 194-210 (2002).
Quiet-time suprathermal tails are most likely the
seed population for particles accelerated by CIR and
CME shocks and bow shock of planets. Should these
tails continue to the outer regions of the heliosphere,
as appears to be the case, then they could well be the
source of ACRs accelerated by the termination shock.
9. Krimigis, S. M., Decker, R. B., Hamilton, D.C., Hill, M.
E., and Gloeckler G., “Survey of Energetic Particles
Observed at Voyager 1 and 2 During 1999-2001” in
Proceedings of the 27th ICRC, pp. 3607-3610.
588