PAPERS | JULY 01 2022

Motion of a charged particle in the static fields of an infinite

straight wire 
Joel Franklin ; David J. Griffiths ; Nelia Mann

Am. J. Phys. 90, 513–519 (2022)

https://doi.org/10.1119/5.0077042

CrossMark

 
View Export
Online Citation

18 December 2023 12:44:30

Motion of a charged particle in the static fields of an infinite
straight wire
Joel Franklina)
Department of Physics, Reed College, Portland, Oregon 97202
David J. Griffithsb)
Department of Physics, Reed College, Portland, Oregon 97202
Nelia Mannc)
Department of Physics, Union College, Schenectady, New York 12308
(Received 29 October 2021; accepted 28 February 2022)
The nonrelativistic trajectory of a point charge q in the magnetic field of a steady line current I is
characterized by the three components of its initial velocity. The motion is periodic in the
_ and z,
cylindrical coordinates s, /, _ describing, in the generic case, a kind of “double helix,” with
one helix serving as a guide while the other winds around it. A positive charge “drifts” in the
direction of the current (a negative charge goes the other way). The inclusion of a uniform line
charge k (coinciding with the current) does not alter the motion qualitatively, but it does change the
drift velocity, and can even reduce it to zero, collapsing the trajectory to the surface of a toroid.
The relativistic treatment modifies and illuminates these results. # 2022 Published under an exclusive
license by American Association of Physics Teachers.
https://doi.org/10.1119/5.0077042

I. INTRODUCTION so the equations of motion are

In this paper, we study the motion of a charged particle in 2 z_ € þ 2s_ /_ ¼ 0; ½3 €z ¼ a s_ ;

½1 s€  s/_ ¼ a ; ½2 s/ (5)
the magnetic field of a long straight steady current. The s s
motion is (perhaps surprisingly) periodic in the radial cylin-

18 December 2023 12:44:30

drical coordinate and also in the longitudinal and azimuthal where
velocities. l0 qI
In Sec. II, we obtain the nonrelativistic equations of a (6)
2pm
motion and their first integrals. In Sec. III, we examine three
special cases and adopt useful conventions for numerical is a constant characterizing the strength of the force.
work. In Sec. IV, we endow the wire with a uniform static Question: What is the trajectory of the particle?1–6
charge (in addition to its current); this introduces an electric Multiplying Eq. (5-[2]) by s, we get
force, but the main effect is to alter the drift velocity. In Sec.
€ þ 2ss_ /_ ¼ d 2_ ‘
V, we present a relativistic treatment, including both electric s2 / ðs /Þ ¼ 0 ) s2 /_ ¼ (7)
and magnetic fields. Some of these results have appeared in dt m
the recent literature, as indicated in the endnotes, but we is a constant of the motion (physically, ‘ is the z component
have endeavored to make our discussion self-contained. of the angular momentum). Meanwhile, Eq. (5-[3]) indicates

II. EQUATIONS OF MOTION d d

ðzÞ
_ ¼ ða ln sÞ ) z_ ¼ a ln ðs=s0 Þ; (8)
dt dt
The magnetic field of an infinite straight wire carrying a
steady current I, in cylindrical coordinates (s; /; z), with I where s0 is a second constant of the motion—not to be con-
along the z axis, is fused with s(0) (physically, s0 is related to the momentum
_ it is a lit-
conjugate to z). Instead of using Eq. (5-[1]) to get s,
l0 I ^ tle quicker to exploit the fact that magnetic forces do no
B¼ /: (1)
2p s work, so the kinetic energy is conserved:
The force on a charge q moving with velocity 1 1 2
E ¼ mv2 ¼ mðs_ 2 þ s2 /_ þ z_ 2 Þ; (9)
2 2
v ¼ s_ ^s þ s/_ /
^ þ z_ ^z (2)
or
is
l qI 1 ‘2 1
F ¼ qðv  BÞ ¼ 0 ðs_ ^s þ s/_ / ^ þ z_ ^z Þ  /
^ E ¼ ms_ 2 þ þ ma2 ½ln ðs=s0 Þ2 : (10)
2ps 2 2ms2 2
l qI (3)
¼ 0 ðs_ ^z  z_ ^s Þ ¼ ma: Serendipitously, this is nothing but the equation of motion
2ps for a mass m in the one-dimensional potential
The acceleration (in cylindrical coordinates) is
‘2 1
a ¼ ð€
2
s  s/_ Þ ^s þ ðs/
€ þ 2s_ /Þ
_ /^ þ €z ^z ; (4) VðsÞ ¼ 2
þ ma2 ½ln ðs=s0 Þ2 : (11)
2ms 2

513 Am. J. Phys. 90 (7), July 2022 http://aapt.org/ajp # 2022 Published under an exclusive license by AAPT 513
 Fig. 1. The potential in Eq. (11).

This potential is plotted in Fig. 1. The equilibrium radius

se is given by dV=ds ¼ 0:
 2  
‘ se
¼ ln : (12)
mase s0

Provided that
"  2 #
‘2 ‘
E > Vðse Þ ¼ 1þ ; (13)
2ms2e mase
Fig. 2. The helix, case A.

18 December 2023 12:44:30

the particle slides back and forth, periodically, between the
turning points smin and smax . But if s(t) is periodic, so too are rffiffiffiffi sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
/_ (Eq. (7)) and z_ (Eq. (8)). This means that /ðtÞ and z(t) k a2 2‘2 a pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
x¼ ¼ 2
þ 2 4¼ 1 þ 2 ln ðse =s0 Þ: (17)
both take the “quasiperiodic” form7 m se m se se

j t þ f ðtÞ; (14) The trajectory is a kind of double helix, with a small coil
winding around the larger helix of Sec. III A (see Fig. 3).
where j is a constant and f(t) is periodic (f ðt þ TÞ ¼ f ðtÞ,
with T the period of s(t)). Except for a constant azimuthal C. Zero angular momentum
velocity and a constant longitudinal drift, all three coordi-
nates oscillate (at the same frequency). The entire motion is If ‘ ¼ 0; /_ ¼ 0, and the motion is confined to a plane con-
a kind of wobbling deformed helix, or “helicoid.”3 taining the z axis, the equilibrium radius is se ¼ s0 ((Eq.
(12)), and the frequency for small oscillations is x ¼ a=s0
((Eq. (17)).6 The trajectories are cycloid-like (see Fig. 4).
III. SPECIAL CASES AND NUMERICAL SOLUTIONS
A. The Helix D. Numerical solutions
If the amplitude of the radial oscillations is zero, then it is The general solution depends on the parameter a ((Eq.
a true helix (/_ and z_ are constants):8 (6)), with the dimensions of speed. For numerical work, we
might as well set9
_ ‘ ‘2
sðtÞ ¼ se ; /ðtÞ ¼ 2; _ ¼
zðtÞ ; (15) a  1; (18)
mse am2 s2e

(see Fig. 2). this just defines a convenient unit for velocity. We will also
set m ¼ 1, establishing the unit of mass. Six initial conditions
determine the trajectory:
B. Small amplitudes
(1) s(0) (the initial radial position); we can use this as the
The second derivative at the equilibrium point (the “spring unit of length, so
constant” for small oscillations) is
 sð0Þ  1 (19)
d 2 V  ma2 2‘2
k  2  ¼ 2 þ 4; (16)
ds se se mse (the unit of time, then, is sð0Þ=a).
(2) /ð0Þ (the initial azimuthal angle); we might as well set
so the (angular) frequency for small amplitudes is this to zero

514 Am. J. Phys., Vol. 90, No. 7, July 2022 Franklin, Griffiths, and Mann 514
 /ð0Þ  0 (20)
(this just establishes the orientation of the axes).
(3) z(0) (the initial value of z); again, we might as well set
this to zero
zð0Þ  0 (21)
(that just picks the position of the origin along the z
axis).
_
(4) sð0Þ (the initial radial velocity): Fig. 4. The quasi-cycloid, case C.
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
_
sð0Þ ¼ 2E  ‘2  ðln s0 Þ2 : (22) _ is
The longitudinal drift velocity (the average value of z)
_
(5) /ð0Þ (the initial angular velocity): Dz
_ vd ¼ ; (26)
/ð0Þ ¼ ‘: (23) T
_
(6) zð0Þ (the initial longitudinal velocity): where
_
zð0Þ ¼ ln s0 : (24) Dz ¼ 2½zðsmax Þ  zðsmin Þ (27)
Thus, each trajectory is characterized by three numbers,
_
which we can take to be E, ‘, and s0 (or, if you prefer, sð0Þ; is the change in z during one full cycle:12
_
/ð0Þ, _
and zð0Þ). In Figs. 2–4, we illustrate each of the trajec- ð ð
2 smax dz 2 smax z_
tories discussed.10,11 vd ¼ ds ¼ ds
T smin ds T smin s_
ð
E. Period and drift velocity 2a smax ln ðs=s0 Þ
¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ds: (28)
T smin
For a given trajectory, the radius is localized (smin  s ð2E=mÞ  ð‘=msÞ2  ½a ln ðs=s0 Þ2
 smax ), but the z coordinate goes from 1 to 1. The
period of the oscillations is The drift velocity is in the direction of the current if the
ðT ð smax charge is positive. (If it is negative, vd is the opposite to I.)

18 December 2023 12:44:30

ds
T ¼ dt ¼ 2 Here is the proof.5
0 smin s_ Proof: Rewriting Dz (Eq. (27)), we get
ð smax
ds ðT
¼2 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi : (25)
smin ð2E=mÞ  ð‘=msÞ2  ½a ln ðs=s0 Þ2 Dz ¼ z_ dt: (29)
0

From Eqs. (5) and (7),

2 2
s  s/_ Þ ¼ s2 /_  s€
az_ ¼ sð€ s
 2
‘ d
¼  ðssÞ _ þ s_ 2 ; (30)
ms dt

so
"  2 # 
ðT
1 2 ‘ 1 T
Dz ¼ s_ þ dt  ss_  : (31)
a 0 ms a 0

But the last term is zero, since s(t) is periodic, and the
remaining integral is positive definite (except for the trivial
case of a particle at rest or with no charge), so Dz shares the
sign of a (positive if qI > 0, minus if qI < 0).

IV. ADDING AN ELECTRIC FIELD

Suppose the wire carries a uniform line charge k in addi-
tion to the steady current I. There is now a static electric
field
1 k
E¼ ^s ; (32)
2p0 s
Fig. 3. The “double helix,” case B. as well as the magnetic field (Eq. (1)), and the net force is

515 Am. J. Phys., Vol. 90, No. 7, July 2022 Franklin, Griffiths, and Mann 515
 F ¼ qðE þ v  BÞ; sem ðtÞ ¼ sm ðtÞ; /em ðtÞ ¼ /m ðtÞ;
qk l qI zem ðtÞ ¼ zm ðtÞ þ cgt; (43)
¼ ^s þ 0 ðs_ ^s þ s/_ / ^ þ z_ ^z Þ  /;
^
2p0 s 2ps
    (Eq. (37)). For example, in case A (the helix), if the charge
l0 qI kc2 density is
¼ zþ
s_ ^  z_ ^s ¼ ma; (33)
2ps I
pffiffiffiffiffiffiffiffiffi 2p0 2
where c ¼ 1= 0 l0 is the speed of light. The first equation k¼ ‘ ; (44)
qms2e
of motion is now
then (Eq. (15))
2 ðz_  gcÞ
½10  s€  s/_ ¼ a ; (34)
s z_ em ¼ z_ m þ cg ¼ 0; (45)
where and the helix collapses to a circle (Fig. 5); the magnetic force
_ satisfies
disappears entirely, and the orbital speed ðv ¼ se /)
kc
g : (35)
I qk mv2
 ¼ ; (46)
The other two are unchanged (Eq. (5)), and hence so too are 2p0 se se
the first integrals (7) and (8):
as it should (the electric attraction supplying the centripetal
‘ force). In the generic case, we can “stop the drift”—produc-
s /_ ¼
2
and z_ ¼ a ln ðs=s0 Þ: (36) ing a localized “donut” trajectory (Fig. 6)—by subtracting vd
m
from vz ð0Þ and introducing an electric field (Eq. (32)) with
However, for future convenience, we write the latter as
k am
¼ vd : (47)
s 0 Þ þ cg;
z_ ¼ a ln ðs=~ (37) 2p0 q

s 0  s0 ecg=a .
where ~ If all this comes as a surprise, remember that in relativity, a
Conservation of energy now includes the electric potential current in longitudinal motion acquires an electric charge.14

18 December 2023 12:44:30

energy:13 Conversely, introducing k amounts to adopting a moving ref-
erence frame. We will explore this further in Sec. V.
1 qk
E ¼ mv2  ln ðs=~
s0Þ
2 2p0 V. RELATIVISTIC TREATMENT
(38)
1 2
¼ mðs_ 2 þ s2 /_ þ z_ 2 Þ  mcag ln ðs=~
s 0 Þ; Thus far, we have treated the motion nonrelativistically.
2 In this section, we consider the fully relativistic case, and—
for a change—we will use a Lagrangian approach. The rela-
or, introducing a new energy E~ by subtracting off the irrele-
tivistic Lagrangian for a point charge in electromagnetic
vant constant,
fields specified by the potentials V and A is15
1 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
E~  E  mc2 g2 L ¼ mc2 1  v2 =c2  qðV  v  AÞ: (48)
2
(39)
1 ‘2 1 In our case
¼ ms_ 2 þ 2
s 0 Þ2 :
þ ma2 ½ln ðs=~
2 2ms 2
k l0 I
V¼ ln ðs=aÞ; A¼ ln ðs=aÞ ^z ; (49)
The radial motion is that of a particle of energy E~ in a one- 2p0 2p
dimensional potential well
where a is an arbitrary reference point (which we might as
‘2 1 well take to be the same for V and A). Thus,
VðsÞ ¼ s 0 Þ 2 ;
þ ma2 ½ln ðs=~ (40)
2ms2 2 rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

1 2
the same form as we had without the electric field (compare L ¼ mc 1  2 s_ 2 þ s2 /_ þ z_ 2
2
c
Eq. (11))! qk ql I
This implies that there is a mapping between solutions to þ ln ðs=aÞ  0 z_ ln ðs=aÞ: (50)
the original problem and solutions of the generalized system. 2p0 2p
Say that without the electric field, we have a solution
Since / and z are ignorable coordinates, the associated
fsm ðtÞ; /m ðtÞ; zm ðtÞg: (41) canonical momenta are constants of the motion:
@L _
From this, we can construct a solution with the electric field p/  ¼ cms2 /; (51)
_
@/
fsem ðtÞ; /em ðtÞ; zem ðtÞg (42)
@L ql I
~ such that pz  ¼ cmz_  0 ln ðs=aÞ; (52)
(with s0 ! s~0 and E ! E), @ z_ 2p

516 Am. J. Phys., Vol. 90, No. 7, July 2022 Franklin, Griffiths, and Mann 516
 c2 p/
/_ ¼ ; (57)
s2 ðE þ agmc ln ðs=aÞÞ
c2 ðpz þ am ln ðs=aÞÞ
z_ ¼ : (58)
ðE þ agmc ln ðs=aÞÞ

_ Eq. (55) indicates

As for s,

1 v2 1 2
2
¼ 1  2
¼ 1  2 ðs_ 2 þ s2 /_ þ z_ 2 Þ
c c c
m 2 c4
¼ ; (59)
ðE þ agmc ln ðs=aÞÞ2

so substituting in Eqs. (57) and (58),

ðmcÞ2 þ ðp/ =sÞ2 þ ðpz þ am ln ðs=aÞÞ2

s_ 2 ¼ c2  c4 : (60)
ðE þ agmc ln ðs=aÞÞ2
Fig. 5. The helix (Fig. 2) collapses to a circle.
This can be written in a more suggestive way
where
1 2 1 2
1 mc ¼ ms_ þ Veff ðsÞ; (61)
c  pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi : (53) 2 2
1  v2 =c2
where
In addition, we have the conserved total energy16 h i
2 2 2
qk 1 4 ðmcÞ þ ðp/ =sÞ þ ðpz þ am ln ðs=aÞÞ
E ¼ cmc2  ln ðs=aÞ: (54) Veff  mc :
2p0 2 ðE þ a g mc ln ðs=aÞÞ2

18 December 2023 12:44:30

(62)
From Eq. (54), we solve for c:
Equation (61) can be read as a (non-relativistic!) one-
1 dimensional problem in which a particle of total energy
c¼ ðE þ agmc ln ðs=aÞÞ; (55)
mc2 1=2mc2 rocks back and forth in the potential well Veff .
In particular, if g ¼ 0 (the purely magnetic case), then
where (as before)
h i
l qI kc mc4 2 2 2
a 0 ; g : (56) Veff ¼ ðmcÞ þ ðp/ =sÞ þ ðpz þ am ln ðs=aÞÞ
2pm I 2E2
B
Expressing /_ and z_ in terms of s (and the conserved ¼ A þ 2 þ C ½ln ðs=s1 Þ2 ; (63)
s
quantities):
which (except for the irrelevant constant A) has the same
form as Eq. (11); again s(t) is periodic, and so too are /_ and
_ Qualitatively, the trajectories are the same as in the nonrel-
z.
ativistic case. The situation is more complicated, when g is
nonzero (i.e., with the electric field included), for now
Veff ðsÞ does not go to infinity as s ! 0 and s ! 1. Instead,
if p/ 6¼ 0
8
>
< 1; as s ! 0;

Veff ! 1 1 2 (64)
>
: g2 2 mc ; as s ! 1;

whereas if p/ ¼ 0
 
1 1 2
Veff ! 2 mc (65)
g 2
in both directions. Moreover, Veff has an infinite spike at

ssp ¼ aeE=agmc ; (66)

Fig. 6. The double helix shrinks to a donut. where the denominator vanishes. In terms of ssp ,

517 Am. J. Phys., Vol. 90, No. 7, July 2022 Franklin, Griffiths, and Mann 517
Fig. 7. Typical graphs of Veff ðsÞ, for (i) p/ 6¼ 0 and ag > 0; (ii) p/ ¼ 0 and ag < 0; (iii) p/ 6¼ 0 and ag < 0. The horizontal line is at 1=2mc2 . In the first graph,
the inner region is inaccessible, the electric force is repulsive, and the charge flies off to infinity; in the second graph, the outer region is inaccessible, the elec-
tric force is attractive, and the charge falls to the center; in the third graph, outer region is inaccessible, the electric force is attractive, and the charge oscillates
(in s). The heavy line indicates the physical region.

ag are really just three distinct regimes here, and if we can find
c¼ ln ðs=ssp Þ; (67)
c the trajectories for the “seed” configurations:
and since for physically possible motion c > 1, it follows (i) I ¼ 0, kq > 0,

18 December 2023 12:44:30

that
if ag > 0; then s > ssp ;
if ag < 0; then s < ssp :
In the first case kq > 0, the electrical force is repulsive, and
the motion is unbounded; in the second case kq < 0, the
electrical force is attractive, and the motion is bounded (in
s). Typical plots are shown in Fig. 7.
An illuminating way to think about these results is to
imagine starting with a combination (k; I), and performing a
Lorentz boost along the z direction. Because
 
l ck I
J ¼ ðcq; JÞ ¼ ; ; (70)
A A
(where A is the cross-sectional area of the “wire”) constitutes
a four-vector, ðcqÞ2  J 2 , and hence ðckÞ2  I 2 is Lorentz
invariant. It could be time-like (or shall we call it “charge-
like”?), or spacelike (“current-like”), and in the former case
“forward” (k > 0) or “backward”:17
(i) charge-like, forward ðckÞ2  I2 > 0, k > 0,
(ii) charge-like, backward ðckÞ2  I 2 > 0, k < 0,
(iii) current-like, ðckÞ2  I2 < 0.
By suitable Lorentz boost a forward charge-like configura-
tion can be transformed to any other forward charge-like
configuration ðk0 ; I0 Þ with ðckÞ02  I 02 ¼ ðckÞ2  I2 , a back-
ward charge-like configuration can be transformed into any
other backward charge-like configuration with ðckÞ02 I02
¼ ðckÞ2  I2 , and with a boost plus a rotation a current-like
configuration can be transformed into any other current-like
configuration with ðckÞ02  I02 ¼ ðckÞ2  I2 . Therefore, there
(ii) I ¼ 0, kq < 0,
(iii) k ¼ 0, I 6¼ 0,
(68)
we can construct the trajectory for any ðk; IÞ by Lorentz
(69) transformation. These are the regimes illustrated in Fig. 7: (i)
runaway repulsion to s ¼ 1, (ii) attraction to the axis, and
(iii) quasi-oscillatory motion with bounded s.
ACKNOWLEDGMENTS
The authors thank Nicholas Wheeler, Raymond Mayer,
and Kirk McDonald for illuminating discussions and a
sharp-eyed referee for calling attention to Eqs. (68) and (69)
and for clarifying the Lorentz boost method and its relation
to Minkowski geometry.
AUTHOR DECLARATIONS
Conflict of Interest
There are no conflicts of interest to declare.
a)
Electronic address: jfrankli@reed.edu, ORCID: 0000-0002-9073-5507.
b)
Electronic address: griffith@reed.edu, ORCID: 0000-0002-4406-4781.
c)
Electronic address: mannn@union.edu
1
M. M€ uller and K. Dietrich, “Classical motion of a charged particle in the
magnetic field of a rectilinear current,” Z. Phys. D 33, 101–107 (1995).
This paper allows the particle to carry a magnetic dipole moment in addi-
tion to its charge.
2
D. Yafaev, “A particle in a magnetic field of an infinite rectilinear
current,” Math. Phys., Anal. Geom. 6, 219–230 (2003).
3
J. Aguirre, A. Luque, and D. Peralta-Salas, “Motion of charged particles in
magnetic fields created by symmetric configurations of wires,” Physica D
239, 654–674 (2010).
4
A. Prentice, M. Fatuzzo, and T. Toepker, “Charged particle dynamics in
the magnetic field of a long straight current-carrying wire,” Phys. Teach.
53, 34–37 (2015). This paper treats the two-dimensional example (case C).
In an “Addendum” (Phys. Teach. 53, 262 (2015)) the authors check that
kinetic energy is conserved in this problem.

518 Am. J. Phys., Vol. 90, No. 7, July 2022 Franklin, Griffiths, and Mann 518
5 11
H. Essen and A. B. Nordmark, “Drift velocity of charged particles in mag- See the supplementary material at https://www.scitation.org/doi/suppl/
netic fields and its relation to the direction of the source current,” Eur. 10.1119/5.0077042 for the Mathematica code used to create the figures.
12
Phys. J. D 70, 198–208 (2016). In Case C (‘ ¼ 0) these integrals can be evaluated in closed form: T ¼
6 2ps0 I1 ðv=aÞ
M. Asadi-Zeydabadi and C. S. Zaidins, “The trajectory of a charged parti- a I0 ðv=aÞ; vd ¼ v I0 ðv=aÞ ; where In is the modified Bessel function of order n,
cle in the magnetic field of an infinite current carrying wire in the nonrela- and v is the (constant) speed of the charge. These results were obtained by Essen
tivistic limit,” Results Phys. 12, 2213–2217 (2019). This paper treats the and Nordmark (Ref. 5), but with unfortunate typos (their Eqs. (73) and (74));
two-dimensional example (Case C). they were derived in a different way by Asadi-Zeydabadi and Zaidins (Ref. 6).
7
This from integrating the Fourier expansion /ðtÞ _ ¼ c0 13
We might as well use s~0 as the zero for electric potential. You can use a
P follows 2pn
þ 1 n¼1 nc cos T t þ d n sin 2pn
T t . different reference point if you like: the equations will be a little more
8
D. J. Griffiths, Introduction to Electrodynamics, 4th ed. (Cambridge U. P., cumbersome, but the result is the same.
14
Cambridge, 2017), Problem 5.39. E. M. Purcell and D. J. Morin, Electricity and Magnetism, 3rd ed.
9
What if a is negative? Rotating the apparatus by 180 would reverse the (Cambridge U. P., Cambridge, 2013), Sec. 5.9.
15
sign of the current and render a positive, while carrying one possible tra- J. D. Jackson, Classical Electrodynamics, 3rd ed. (John Wiley & Sons,
jectory into another, so there is no loss of generality in assuming a is New York, 1999), Eq. (12.12).
16
positive. Jackson (Ref. 15), Eq. (12.17).
10
To draw the figures we used A. Chambliss and J. Franklin, “A 17
The “light-like” case (ðckÞ2  I2 ¼ 0) is qualitatively similar, with the
magnetic velocity Verlet method,” Am. J. Phys. 88, 1075–1082 exception of the trivial limit k ¼ 0, I ¼ 0, for which the trajectory is of
(2020). course a straight line, with constant velocity.

18 December 2023 12:44:30

519 Am. J. Phys., Vol. 90, No. 7, July 2022 Franklin, Griffiths, and Mann 519