SVENSKA GEOFYSISKA FORENINGEN

VOLUME 10, NUMBER 4 Tetlus

NOVEMBER 1958
A QUARTERLY JOURNAL OF GEOPHYSICS
The Helicoidal Structures in the Cosmical Electrodynamics
By SYUN-ICHI AKASOFU, Geophysical Institute, Tahoku University, Sendai, Japan
l
(Manuscript received December 12, 1957)
Abstract
The general-type solution of the equation [curl (curl H x H)] = 0 are studied in the cylindrical
systems, according to LUNDQUIST'S and CHANDRASEKHAR'S methods, In general, the magnetic
fields are of helicoidal type in the cylindrical systems. Several examples are studied in the
cosrnical fields,
()tJ I
-+curlvxv=- cud HxH-
()t 4ne
- grad ~ ~ v
2
+V), (6)
In the cosmical dimensions, such configura-
tions in the cylindrical system seem to appear
in the following examples: (i) Force-free dis-
charges, (ii) Magnetic fields of spiral arm, (iii)
Current jets, (iv)Isoroattionand theftlamentary
structures of diffuse nebulae. In these examples,
P and T functions are reduced to the solutions
of (3) and (4).
2. Magnetohydrostatic cases
The simultaneous equations governing the
velocity field and the magnetic field in an
incompressible, inviscid and conductive fluid
are (in e.m.u.)
- (}H= curl - ~ - curl H - v x H), (5)
(jt 4na
div V= 0, div H=o,
1 Address after January, 1959: Geophysical Institute,
University of Alaska',
Tellus X (1958), 4
1-807638
I. Introduction
CHANDRASEKHAR (1956a) has shown that a
magnetic field none of whose components are
azimuthal can always be expressed as the sum
of a toroidal T field and a poloidal P field. In
this paper, we study several configurations of
magnetic fields which satisfy the following
equations under the restriction that none of
their components are azimuthal;
curl (curl H x H) = o. (I)
In cylindrical co-ordinate systems (ru, lfJ, z), H
with the above-mentioned restriction has a
form
H = - (U ~ P r., + luTI", + ~ :_ (iu
2
P) I
z
, (2)
(IZ (f) ot
(j2p 1 (jP ()2P ~ G (-2 )_m.(-2P) ()
')-2 + - ',- + '1 2 + -2 ((J P - ':l' W , 3
uW (f) aco uZ W
and the poloidal function P and the toroidal
function T which satisfy (I) and (2) are the
solutions of the following differential equations,
respectively.
410
SYUN-ICHI AKASOFU
(16)
(15)
(
. 1 d ~ 2 P )
+ r: ~ (I) + H
o
1'1"
W aco
curl H
o
= o.
curl H = iil ; ~ 1 w- ciJL1
s
P 1 q> +
+ ~ d_ (w2T) I .
W dw 'I'
H = H
o
+ h = - WdP I - + iilT I +
dZ W 'I'
L1
sP=
-IXT,
rt dP
dz = IX dZ'
I d _ (I d )
rz: ~ (w
2
T) = IX -:=- ~ (w
2P)
+ H
w d(l) W dw
where K is a constant. But, K must be zero
in this case, according to the requirement
that T be regular on the z-axis. Then,
IXH
T=IXP+_o.
2
Equation (21) can be reduced to
Comparing each component on both sides,
we can obtain the following relations;
or
Let us study this case along the method
developed by CHANDRASEKHAR for the axi-
symmetric case. In cylindrical co-ordinate
system, H and curl Hare
(10) Then, (13) becomes
CUrlh=IX(Ho+h), (17)
(II )
curl hx (Ho+h)=o,
(i)
and
curl H=IX'H or jxH=o,
where
grad ~ + V) = 0,
(ii) IX = 0.
(i) Beltrami field
In the first case, we fmd
IX' = 4nelX.
In general, such a configuration in a vector
field is called the Beltrami field. This case
means that all the terms in the equation of
motion (6) are zero, independently. LUND-
QUIST (1950) and CHANDRASEKHAR (1956a)
studied this problem for the axisymmetric case.
The latter author stated in a separate paper
(1957) that the magnetic field gets altered in
such a way that its influence on the motions
is progressively reduced and a stationary state
is eventually reached. The writer is not sure
that such a configuration appears in the in-
dependent system. For example, from where
come the electric currents in the independent
system?
However, if the electric discharge occurs
along the magnetic lines of force in the solar
atmosphere and j x (H
o
+ h) becomes larger
than any other terms in (6), we can expect the
following configuration;
The two cases appear as the special solutions,
where the various notations are the generally
used ones (cf. CHANDRASEKHAR, 1956b).
Let us study some static cases. In this case,
(6) becomes
_1_ curl H x H = grad (E + V). (8)
4
n
e e
Taking the curl of this, we obtain (I). As was
pointed out by LUNDQUIST (1950, 1952), the
solution of (8) satisfies
Hx grad (E + V)
_1_ curl H = IXH+ e (9)
4ne H2
Tellus X (1958). 4
HELICOID AL STR UCTURES IN COSMICAL ELECTRODYNAMICS 41 I
Substituting (23) into (19), the differential
equation which must be satisfied by P is
(X2H
L1
5
P= - (X2p - _0. (24)
2
This is the special case of (3). With suitable
boundary conditions, we can expect the heli-
coidal structure of the magnetic field which
will be studied in the next section. In fact,
DUNGEY and LOUGHHEAD (1954) studied the
stability of this type of field having the impres-
sion that some prominences show filaments
being twisted, though, they have not verified
why such a configuration appears in such a
system.
(ii) Complex-lamellar field
The second case is that
Hxgrad (25)
Taking the dot product with H, we can get
H curl H = o. (26)
In general, such a configuration in a vector
field is called the complex-lamellar field.
The magnetic field which satisfies (26) can be
expressed by the two scalar functions "P and X;
H = "P grad x.
For axisymmetric case, this becomes
dX dX
H = "P ") _ lUi + "P ":) I
z

oeo uZ
Comparing (28) with (2), we find
T=o,
and also
I J _ JX
H = -- - (w
2
P) =111-
iu iiJ Jz T JiiJ'
H
_ I J (-2P) _ dX
z r:": ")_ W -"PT-
(t) at dt
This is the case which is studied by FERRARO
(1954) in the case of equilibrium of the mag-
netic star.
Tellus X (1958). 4
(iii) Magnetic fields of spiral arm
The Beltrami field and the complex-lamellar
field are only the special solution of equation
(8). In this part, we discuss the most simplest
general-type solution of (8). This discussion is
quite analogous with the method developed by
CHANDRASEKHAR and PRENDERGAST (1956) and
PRENDERGAST (1956).
Taking the curl of (8), we have the equa-
tion (I),
curl (curl H x H) = 0 (31)
As the simplest example, we consider the mag-
netic fields of the spiral arm. The force-free
discharge discussed above can be studied quite
analogously along this line. CHANDRASEKHAR
and FERMI (1953) assumed the following
equilibrium condition in the spiral arm,
pgrav. = pkin. +pmag. (32 )
They used this equation for somewhat different
purpose. In that paper, they have supposed
that the direction of the galactic magnetic
field is roughly parallel to the direction of the
spiral arm.
When we equate the gravitational pressure
in the arm to the sum of the gas pressure and
the magnetic pressure and the density in the
arm is assumed to be constant, the similar
equation to (32) reduced to (I). Using (I)
and (16), the poloidal function which must
satisfy (I) becomes (cf. PRENDERGAST (1956)
eq. (IS))
J2P 3 JP J2P
") - 2 +-;:- ") - + ';}2 + (X2P = X,
aio (t) at uZ
and the toroidal function which also must
satisfy (I) becomes
__d_ (iiJ2 '['2) = 2 (X2ii)2P (34)
d(ciJ2
p
) ,
where we assume
G= iX
2
iiJ2p x= constant),
tP = x = constant,
in the equations (3) and (4), respectively. It
must be noticed that (33) has the same form
with (24).
SYUN-ICHI AKASOFU
T=rxP.
, Jp=o
at OJ = R, lP= 0 d dP (44)
an d{V = 0;
and this is satisfied, only if
When we take large rx, the toroidal field be-
comes larger compared with the poloidal field
and then it seems that this configuration be-
comes unstable, as was pointed out by ALFVEN
(1950 a) .
Substituting the final form of P (45) into (2),
the magnetic f1dd in the spiral arm is
The another requirement is
from which wc can obtain the final form of P
]2(rxR)=0.
The first root of (47) is
rx
l
R = 5. I 35
(,nc H
-- I, X =
4n F
_ (d(V
2P)
+ ()(V
2P))
=
471 ()n {)z
- grad ({,)2P). (3 8)
4
n
jxH=
curl H = 471} = rxH - e,na'l"
- grad (V
2P)
= grad p + (! grad V, (40)
4n
Rewriting (8) in the following form
j x H = grad p + e grad V, (39)
and substituting (38) into (39), we obtain
and
Then, it is easily verified that
where we assume ()/Jz = o.
If we take x = -rx
2
H
o
/2, this equation is
exactly the same as (24). The solutions of (34)
and (35) which are regular on the z-axis arc
p =!:_ + A]l(rxiv)
rx
2
ev
412
Then, (33) becomes
d
2P
3 dp
-+- --+rx
2
p =x
dev
2
/iJ d{v '
or
(wP)=p+nV (41)
4n
On our case, we take
(
In the simplest case, the boundary conditions
to be satisfied at the surface of the spiral arm are
_ {the pressure is zero,
at lU = R, the magnetic field is zero;
where R is the radius of the spiral arm. Then,
this is
The lines of forcc resemble a helix wrapped on
the cylindrical surface.
The boundary conditions adopted here are,
of course, very scanty. (d. SPITZER (1956)).
Moreover, this is not the pure magnetostatic
problem. However, it may be noticed that
SHAJN (1957) obtained the systematic devia-
tion of the general field from the galactic
plane about 18 in the solar vicinity.
(iv) Current jets
The current-jet theory of filaments is a more
general case of the force-free discharge. It
seems that the force-free discharge occurs in
the high solar atmosphere, but the general case
may occur in the chromospheric regions where
j x H term can be balanced by the sufficient
Tellus X (1958), 4
HELICOIDAL STRUCTURES IN COSMICAL ELECTRODYNAMICS 4
13
curl V= w.
where
curl (curl H x H) = 0,
curl (curl v x v)= 0,
and
0= curl (_1_ curl H-v x H) (64)
4na
As we are concerned with the steady state,
we can expect v and H progressively reduced
to the following two cases.
In this case, we can determine the velocity
and magnetic fields by the following equations,
l( )
ldw/dt= 0 (steady vorticity),
cur III x v = 0 orl
and curl (curl H x H) = o.
(59)
Then, at each regular point, we can determine
a plane which we may call the Lamb surface
ana whose normal is parallel to the Lamb
vector 111+ o, A necessary and sufficient con-
dition for the existence of the Lamb surface is
that the Lamb vector be complex-lamellar and
non-vanishing (cf. TRuEsEDELL), 1954). That is,
(IIIXV).CUrl(wxv)=o, wxv#o. (57)
Substituting (55) into (57), we have
(III x v). (_1_ curl (curl H x H) _ dW) = o.
4ng dt
(58)
E. + v
2
+ V = constant. (61)
g 2
Another Lamb surface which is everywhere
normal to the vectorj x H exists, if
Then, the Lamb surface exists, if
(curl H x H). curl (curl H x H) = o. (60)
From (55), it appears that if curl (wx v)=0,
then curl (j x H) = 0 in the case of steady
vorticity. That is, two types of the Lamb
surfaces co-exist in the fluid in this case.
Then, we can imagine the line which is the
intersection of two Lamb surfaces. On this
line, a curvilinear Bernoulli theorem can exist.
dW 1
+curl (wx v) = - curl (curl H x H),
u 4ng
(55)
From the boundary conditions, we can de-
termine A, B and cx. It is easily shown that K
must be zero. Substituting (52) and (53) into
(2), we can obtain the magnetic field and the
current system, which are to be far
more complex than COWLING s model.
3. Lamb surface
Rewriting (5) into the following form (54)
and taking the curl of this equation, we have
dV 1
- +W x v = - curl H x H -
()t 4ng
- grad v
2
+ V), (54)
From (34), we obtain a toroidal function
pressure gradient. This has been studied as the
constriction of discharge by ALFVEN (1950b).
The current-jet theory assumes the form
- grad p +j x H = o. (50)
Taking the curl of this equation, we find also
curl (curl H x H) = o. (51)
Thus, we consider the solution of (33) and (34)
under the boundary conditions that the mag-
netic fields are zero except b<iii <a. That is,
the magnetic field is enclosed between two
concentric cylindrical surfaces.
COWLING (1957) criticized the current-jet
theory, because such a field implies surface
currents flowing in opposite directions on (I) = b
and iil = a and such oppositely directed currents
are difficult to set going. He probably says
that this configuration is too artificial. How-
ever, COWLING'S estimation is only a special
solution of (I) and in general, such a simple
configuration does not occur.
In this case, a poloidal function which
satisfies (33) or (35) has a form
p= z., All (CX(u) (52)
cx
2
iil (u
Tellus X (1958), 4
SYUN-ICHI AKASOFU
cosmic field, some diffuse nebulae may be
expected to reach fmally this state. The equa-
tions (62) and (63) have the same form with
(1) and we can expect that some filamentary
nebulae have the helicoidal structure in their
fine structure. SHAJN (1956) suggested the close
relationships between the shape of these nebulae
and the magnetic fields. SEVERRNY (1956) point-
ed out that the knots of eruptive prominences
are moving along spirals as if it were a motion
of an isolated charge. The latter type field
will be studied in a separate paper.
The writer wishes to express his thanks to
Professor Y. KATO for his interest and to
Professor S. CHANDRASEKHAR for his kind
comments. The writer has benifited much from
discussions with Mr. T. TAMAO throughout
this study.
(0'-+00),
(65)
(0' is finite). (66) (ii) (62), (63) and (64)
414
(i) (62), (63) and curl (v x H) = 0
For the first case, we obtain FERRARO'S iso-
rotation (1937), as the special case. In general,
v becomes parallel to H and in such a case,
there is no interaction between the velocity
field and the magnetic field and further two
Lamb surfaces are parallel to each other. If v
is not parallel to H, then, in general, dH/dt
does appear, the components of which do not
satisfy (62) and (63) and v and H fields vary
with the progress of time in such a manner
that v and H become parallel.
As the concept of infinite conductivity comes
mainly from very large linear dimension in
REFERENCES
ALFVEN. H. 1950a: Discussion of the origin of the
terrestrial and solar magnetic fields. Tel/us. 2, pp.
74-82.
-. 1950b: Cosmical Electrodynamics. London. Oxford
Univ. Press, p. 63 and p. 157.
CHANDRASEKHAR. S. 1956 a: On force-free magnetic
fields. Proc. Nat. Acad. Sci. 43. pp. 1-5.
-, 1956b: Axisymmetric magnetic fields and fluid mo-
tions. Astrophys. ] . u4. pp. 232-243.
-. 1957: On cosmic magnetic fields. Proc. Nat. Acad. Sci.
43, pp. 24-27.
-, and FERMI. E. 1953: Magnetic fields in spiral arm.
Astrophys. ]., lIS, pp. Il3-lI5.
-. and PRENDERGAST. K. H., 1956: The equilibrium of
magnetic stars. Proc. Nat. Acad. Sci., 42, pp. 5-9.
COWUNG, T. G., 1957: Magnetohydrodynamics. New
York. Interscience Pub. INC., p. 26.
DUNDEY, J. W., and LOUGHHEAD, R. E., 1954: Twisted
magnetic fields in conducting fluids. Aust.]. Sci. 7.
pp. 5-13.
FERRARO. V. C. A. 1937: The non-uniform rotation of
the sun and its magnetic field. Mon. Not., 97, pp.
458-472.
-, 1954: On the equilibrium of magnetic stars. Astro-
phys.]., Il9, pp. 407-412.
LUNDQUIST, S., 1950: Magnetost atic fields. Ark. f Fysik
2. pp. 361-365.
-. t952: Studies in magneto-hydrodynamics. Ark. f.
Fysik, 5. pp. 297-347
PRENDERGAST, K. H . 1956: The equilibrium of a self-
gravitating incompressible fluid sphere with a mag-
netic field, 1. Astrophys.]., 123, pp. 498-507.
SEVERRNY, A. B., 1956: The processes in active regions of
the sun and electromagnetics. Presented paper at
I.A. V. symposium on electromagnetic phenomena in
cosmical physics, Stockholm, Cambridge Univ. Press,
1957
SHAJN, G. A., 1956: On the magnetic fields in the inter-
stellar space and in diffuse nebulae. Presented paper
at LA. V. symposium on electromagnetic phenomena in
cosmical physics, Stockholm. Cambridge Univ. Press.
1957
-, 1957: The orientation to the galactic equator of the
general magnetic field of the galaxy in the solar .
vicinity. Astrophys.]. USSR.. 34, pp. 1-7.
SPITZER. L. JR. 1956: On a possible interstellar galactic
corona. Astrophys.].. u4, pp. 20-34.
TRUESDELL, C., 1954: The kinematics of Vorticity. Bloom-
ington, Indiana Univ. Press, p. 132.
Tellus X (1958), 4