Vol 454 | 3 July 2008 | doi:10.

1038/nature07029

LETTERS
Magnetic fields at the solar wind termination shock
L. F. Burlaga1, N. F. Ness2, M. H. Acuña1, R. P. Lepping1, J. E. P. Connerney1 & J. D. Richardson3

A transition between the supersonic solar wind and the subsonic The moderate strength of TS-3 might have been produced by the
heliosheath was observed by Voyager 1, but the expected termina- pickup protons and other suprathermal ions that were present18, as
tion shock was not seen owing to a gap in the telemetry1–4. Here we considered in models9,19–21.
report observations of the magnetic field structure and dynamics The magnetic field strength oscillated quasi-periodically within
of the termination shock, made by Voyager 2 on 31 August–1 the ramp (Fig. 3). The ramp moved past Voyager 2 in ,90 s, and
September 2007 at a distance of 83.7 AU from the Sun (1 AU is the each oscillation passed the spacecraft in ,12.3 s on average. The
Earth–Sun distance). A single crossing of the shock was expected, shock speed gives a characteristic length of the fluctuations in the
with a boundary that was stable on a timescale of several days. But ramp, l < 1,000 km 5 0.15 c/vpi,u. An oscillating ramp containing
the data reveal a complex, rippled, quasi-perpendicular supercrit- substructures with a scale of 0.2 c/vpi,u was observed at 1 AU (ref. 22),
ical magnetohydrodynamic shock of moderate strength under- where shocks are not influenced by pickup protons.
going reformation on a scale of a few hours. The observed A few hours after crossing TS-3, Voyager 2 crossed the termination
structure suggests the importance of ionized interstellar atoms shock (TS-4) back into the solar wind. The re-entry into the solar
(‘pickup protons’) at the shock. wind might have been caused by a ripple propagating on the shock’s
This Letter discusses three termination shock crossings (TS-2, TS-
3 and TS-4) that were observed directly by Voyager 2 (Fig. 1). At least TS-2 TS-3 TS-4
two additional termination shock crossings (TS-1 and TS-5) a 0.3
Heliosheath Solar wind Heliosheath SW
occurred when there were gaps in the telemetry. The multiple cross-
48 s 0.2
ings imply motions of the termination shock, possibly caused by its

B (nT)
large-scale motion or ripples propagating along it5–7. The times
between crossings are of the order of magnitude expected for ripples 0.1
propagating on the shock.
Observations by the magnetic field experiment on Voyager 2 (ref. 0.0
b 0.08
8) show no significant change in the direction of the magnetic field B 0.06 4 min
SD4min

across any of the three termination shock crossings in Fig. 1, indi- 0.04
cating that the shock was quasi-perpendicular at each encounter. 0.02
This geometry is necessary to provide the return of reflected ions
from the shock9,10, which ultimately produce some of the heating c 0.00 360

l (deg)
48 s
behind a high Mach number (supercritical) shock. 270
The structure of TS-3 is shown in Fig. 2. The foot, ramp and d 90 180
d (deg)

overshoot of TS-3 (ref. 9) moved past Voyager 2 in ,23 min, 0

,1.5 min and ,17 min, respectively. There is an increase in the –90
e
magnetic field strength B in the foot, where B is compressed as the 300 192 s
V (km s–1)

solar wind slows down11. The foot contains gyrating protons that are
reflected from the shock9,12,13. A further abrupt decrease in speed to its 200
lowest value occurs at the ramp, owing to the electric potential assoc-
100
iated with cross-field currents and gradients in B in the ramp14,15.
18:00 20:00 22:00 00:00 02:00
Plasma waves were observed at the ramp of TS-3 (ref. 16). The ratio of Time (h:min)
the maximum B in the overshoot to B downstream of the shock is
,2.5, indicating that the ratio of the magnetosonic Mach number Day 243 of 2007 Day 244 of 2007
Mms to the critical fast Mach number (Mc < 2) is ,5 (ref. 17), sug- Figure 1 | Three crossings of the termination shock, illustrating
gesting that Mms < 10 for TS-3, consistent with a supercritical shock. reformation and the variability of its structure. The figure shows 48-s
From the passage time of the ramp (1.5 min) and the estimated averages of magnetic field strength B (a) and the standard deviation of B for
shock speed (68 6 17 km s21; ref. 11), the thickness of the ramp is 4-minute intervals (b), azimuthal angle l (c) and elevation angle d (d), and
,6,000 km. The ramp size is usually expressed in terms of the ion the 192-s average of solar wind speed V (e) as a function of time measured in
inertial length c/vpi,u, where c is the speed of light and vpi,u is the days from the beginning of 2007. (Here the directions l and d of B are in
plasma frequency of an ion upstream. For TS-3, c/vpi,u < 6,000 km heliographic coordinates.) The time between the first and second of the
observed termination shock crossings (TS-2 and TS-3, respectively) is
and the thickness of the ramp is ,1 c/vpi,u. ,3.7 h, and that between the second and third of the observed crossings (TS-
The strength of TS-3, measured by the ratio of the B in the solar 3 and TS-4, respectively) is ,2.8 h. The Larmor frequency of a proton with
wind upstream of the shock (B1) to that behind the shock (B2) is the solar wind speed 300 km s21 gyrating in the magnetic field upstream of
B2/B1 5 1.7 6 0.1. The corresponding density ratio is N2/ the termination shock is VL,u21 < 17 min. The two enhancements in B at TS-
N1 5 1.4 6 0.2 < B2/B1, consistent with a shock strength of 1.6 6 0.2. 2 were separated in time by ,VL,u21. The uncertainty in B is 60.03 nT.
1
NASA/Goddard Space Flight Center, Greenbelt, Maryland 20771, USA. 2The Catholic University of America, Washington DC 20064, USA. 3Massachusetts Institute of Technology,
Cambridge, Massachusetts 02139, USA.
75
©2008 Macmillan Publishers Limited. All rights reserved
LETTERS NATURE | Vol 454 | 3 July 2008

surface5,23. The structure of TS-4 (Fig. 1) is different from that of TS- TS-3
a 0.3
3. In the foot region, B(t) (where t is time) appears as a narrow peak, 0.48 s
and the overshoot is smaller than that of TS-3. Such a peak evolves
from the foot as a result of bunching of reflected solar wind ions 0.2

B (nT)
where they are turned back towards the shock6,12. The amplitude of
the overshoot decreases while the step-like foot evolves to a peak. 0.1
These changes were observed from TS-3 to TS-4 (Fig. 1). The struc-
ture of the termination shock evolved significantly within 2.7 h. The Foot Ramp Overshoot
small peak in B at the front of TS-4 is expected to evolve to a new 0.0 360
b

l (deg)
ramp on this timescale, as part of the shock reformation pro- 270
cess10,12,24,25.
Reformation of the local structure of a supercritical quasi-perpen- c 90 180
45

d (deg)
dicular shock was predicted by both hybrid and full particle simula- 0
tions10,13,25 for a shock with large Mms and/or a low b, where b is the –45
ratio of thermal pressure to magnetic pressure. Recall that we esti- –90
00:09 00:10 00:11 00:12 00:13
mated that Mms < 10 for TS-3. Neglecting pickup protons, b 5 0.04 Time (h:min, day 244 of 2007)
in the solar wind upstream of TS-3, so that b might be small even if
the pickup protons contribute significantly to it. Reformation is a Figure 3 | The internal structure of the ramp of TS-3. The structure is based
on observations of the magnetic field strength B (a) and its directions l
patchy cyclic shock reformation process with a characteristic time of (b) and d (c) at 0.48-s intervals. The magnetometer on Voyager 2 sampled
the order of the downstream gyroperiod25. the vector magnetic fields in the termination shock at a rate of
Further evidence for local reformation of the termination shock is 2.08 samples s21, and the spacecraft was able to transmit all of this
provided by the qualitative transformation of the shock structure information, making it possible to determine the complex internal structure
during the ,3.9-h interval between TS-2 and TS-3 (Fig. 1). At the of the ramp shown here.
front of TS-2, the bulk speed V increased continuously, rather than in
a step-like form. Instead of a simple ramp-overshoot structure in B,
there were two narrow enhancements in B resembling solitons, in Received 19 February; accepted 15 April 2008.
each of which there was a change in B comparable to that in the ramp 1. Burlaga, L. F. et al. Crossing the termination shock into the heliosheath: Magnetic
of TS-3. A shock structure at 1 AU resembling that of TS-2 was fields. Science 309, 2027–2029 (2005).
reported in fig. 5 of ref. 26. 2. Decker, R. B. et al. Voyager 1 in the foreshock, termination shock, and heliosheath.
Science 309, 2020–2024 (2005).
3. Gurnett, D. A. & Kurth, W. S. Electron plasma oscillations upstream of the solar
wind termination shock. Science 309, 2025–2027 (2005).
4. Stone, E. C. et al. Voyager 1 explores the termination shock region and the
TS-3 heliosheath beyond. Science 309, 2017–2020 (2005).
0.30
a Ramp 5. Winske, D. & Quest, K. B. Magnetic field and density fluctuations at perpendicular
0.25 Overshoot supercritical collisionless shocks. J. Geophys. Res. 93, 9681–9693 (1988).
Foot Undershoot 6. Lembege, B. et al. Selected problems in collisionless shock physics. Space Sci. Rev.
0.20 110, 161–226 (2004).
Solar Heliosheath
B (nT)

wind 7. Burgess, D. & Scholer, M. Shock front instability associated with reflected ions at
0.15 the perpendicular shock. Phys. Plasmas 14, 012108 (2007).
0.10 8. Behannon, K. et al. Magnetic field experiment for Voyager-1 and Voyager-2. Space
Sci. Rev. 21, 235–257 (1977).
0.05 9. Goodrich, C. C. in Collisionless Shocks in the Heliosphere: Reviews of Current Research
(eds Tsurutani, B. T. & Stone, R. G.) 153–168 (Geophys. Monogr. Ser. Vol. 35,
0.00 American Geophysical Union, Washington DC, 1985).
b
48 s 10. Scholer, M., Shinohara, M. I. & Matsukiyo, S. Quasi-perpendicular shocks: Length
l (deg)

scale of the cross-shock potential, shock reformation, and implications for shock
270 surfing. J. Geophys. Res. 108 (A1), doi:10.1029/2002JA009515 (2003).
11. Richardson, J. D., Kasper, J. C., Wang, C., Belcher, J. W. & Lazarus, A. J. Cool
180 heliosheath plasma and deceleration of the upstream solar wind at the
c termination shock. Nature doi:10.1038/nature07024 (this issue).
45 12. Biskamp, D. & Welter, H. Numerical studies of magnetosonic collisionless shock
d (deg)

0 waves. Nucl. Fusion 12, 663–666 (1972).

–45 13. Phillips, P. E. & Robson, A. E. Influence of reflected ions on the magnetic structure
of a collisionless shock. Phys. Rev. Lett. 29, 154–157 (1972).
–90 14. Leroy, M. Structure of perpendicular shocks in collisionless plasmas. Phys. Fluids
d
300 26, 2742–2753 (1983).
192 s 15. Woods, L. C. On the structure of collisionless magneto plasma shock waves at
V (km s–1)

supercritical Alfvén Mach numbers. J. Plasma Phys. 3, 435–442 (1969).

16. Gurnett, D. A. & Kurth, W. S. Intense plasma waves at and near the solar wind
200 termination shock. Nature doi:10.1038/nature07023 (this issue).
17. Livesey, W. A. et al. ISEE 1 and 2 observations of magnetic field overshoots in
quasi-perpendicular bow shocks. Geophys. Res. Lett. 9, 1037–1040 (1982).
100 18. Decker, R. B. et al. Mediation of the solar wind termination shock by non-thermal
23:31 00:00 00:28 00:57 01:26 ions. Nature doi:10.1038/nature07030 (this issue).
Time (h:min) 19. Whang, Y. C., Burlaga, L. F. & Ness, N. F. Locations of the termination shock and
heliopause. J. Geophys. Res. 100, 17015–17023 (1995).
Day 243 of 2007 Day 244 of 2007 20. Zank, G. P. et al. Interstellar pickup ions and quasi-perpendicular shocks:
Implications for the termination shock and interplanetary shocks. J. Geophys. Res.
Figure 2 | TS-3 is a supercritical quasi-perpendicular shock. The 48-s
101, 457–477 (1996).
averages of the magnetic field strength B (a) and its directions l (b) and d 21. Lipatov, A. S. & Zank, G. P. Pickup ion acceleration at low b perpendicular shocks.
(c) are shown here together with the 192-s average of the solar wind speed V Phys. Rev. Lett. 82, 3609–3612 (1999).
(d) across TS-3. The magnetic field strength profile shows the classical 22. Newbury, J. A., Russell, C. T. & Gedalin, M. The ramp widths of high-Mach-
features of a supercritical quasi-perpendicular shock: a ‘foot’, ‘ramp’, number quasi-perpendicular collisionless shocks. J. Geophys. Res. 103 (A12),
‘overshoot’, ‘undershoot’ and smaller oscillations, in that order. 29581–29593 (1998).
76
©2008 Macmillan Publishers Limited. All rights reserved
NATURE | Vol 454 | 3 July 2008 LETTERS

23. Lowe, R. E. & Burgess, D. The properties of rippling in quasi-perpendicular Acknowledgements We thank T. McClanahan and S. Kramer for support in the
collisionless shock fronts. Ann. Geophys. 21, 1–9 (2003). processing of the data. We also thank D. Berdischevsky for computing the
24. Lembege, B. & Dawson, J. M. Self consistent study of a perpendicular collisionless instrument zero level corrections for the data in this paper, and for helping to solve
and nonresistive shock. Phys. Fluids 30, 1767–1788 (1987). the problems created by the erroneous decoding of a spacecraft systems command
25. Lembege, B. & Savoini, P. Non-stationarity of a 2-D quasi-perpendicular supercri- sent to Voyager 2 in 2006. N.F.N. was partially supported by a NASA grant to CUA.
tical collisionless shock by self-reformation. Phys. Fluids 4, 3533–3548 (1992).
26. Bale, S. D. & Mozer, F. S. Measurement of large parallel and perpendicular electric Author Information Reprints and permissions information is available at
fields on electron spatial scales in the terrestrial bow shock. Phys. Rev. Lett. 98, www.nature.com/reprints. Correspondence and requests for materials should be
205001 (2007). addressed to L.F.B ([email protected]).

77
©2008 Macmillan Publishers Limited. All rights reserved