See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/228886362

Creation of a high-temperature plasma through merging and compression of

supersonic field reversed configuration plasmoids

Article in Nuclear Fusion · May 2011

DOI: 10.1088/0029-5515/51/5/053008

CITATIONS READS

51 5,205

3 authors, including:

John Slough George Votroubek

MSNW, Llc. Helion Energy, Inc.
175 PUBLICATIONS 1,571 CITATIONS 34 PUBLICATIONS 239 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by George Votroubek on 03 June 2014.

The user has requested enhancement of the downloaded file.

 Supersonic Merging and Compression of FRCs

Creation of a High Temperature Plasma through

Merging and Compression of Supersonic Field
Reversed Configuration Plasmoids

John Slough, George Votroubek, and Chris Pihl

MSNW LLC, 8551 154th Avenue NE, Redmond, WA 98052

Email: [email protected]

Abstract: A new device, the Inductive Plasma Accelerator, was employed to simultaneously form
and accelerate two oppositely directed Field Reversed Configurations (FRCs) where the relative
velocity (600 km/sec) of the plasmoids was much larger than their internal thermal motion. Upon
collision all of the FRC directional energy was observed to be rapidly thermalized concurrent with
complete magnetic reconnection of the two FRCs. Upon merging, the resulting FRC was
compressed to kilovolt ion temperatures exhibiting a configuration lifetime better than predicted
by past scaling of in situ formed FRCs. With the improved FRC confinement scaling, a pulsed
plasma device based on this approach capable of achieving fusion gain is examined. For an FRC
with a poloidal flux 20 mWb or greater, the fusion energy yield per pulse exceeds the plasma
energy for compression fields of 10 T or more. The scaling is insensitive to the compression
chamber radial scale, providing for the possibility of a very compact fusion neutron source.

Keywords: Field Reversed Configuration, FRC, magnetic compression, merging, pulsed

fusion

PACS Nos.28.52.cx, 52.30.-q, 52.35.Vd, 52.50.Lp, 52.55.Lf

1. Introduction
The field reversed configuration (FRC) is a plasmoid with a symmetric toroidal geometry in
which the confining magnetic field is provided by toroidal plasma currents [1]. The plasma
pressure is contained by the encompassing magnetic pressure and magnetic tension with the result
that the plasma energy is comparable to the FRC field energy (the ratio is commonly referred to
as the plasma  and is ~ 1 for the FRC). An illustration of the FRC along with the key equations
describing the FRC equilibrium can be found in figure 1. These features make the FRC the
geometrically simplest, most compact, and highest  of all magnetic confinement schemes. The
simply connected nature of the magnetic field with regard to the containment vessel and the linear
confinement geometry, allow for the translation of the FRC over large distances. These attributes
make the FRC especially attractive as a means to contain thermonuclear plasmas. These unique
qualities however are obtained at a price. The topological simplicity makes the generation and
sustainment of the large diamagnetic currents very challenging. The configuration has net bad
magnetic curvature and is susceptible to MHD interchange and kink modes. When isolated from
the vessel wall by an external axial magnetic field, as is typically the case, the FRC poloidal field
represents essentially an anti-aligned dipole with regard to the external field and is therefore
disposed to tilt instability.

1
 Supersonic Merging and Compression of FRCs

Figure 1. (Top) field lines and pressure contours for the Field Reversed
Configuration (FRC) with the characteristic scaling parameters indicated.
(Bottom) FRC equilibrium constraints and the diagnostic measurements that
together with the equilibrium relations that are employed to determine the basic
parameters of the FRC equilibrium.

Despite these daunting issues, stable high-temperature FRC plasmoids have been readily
formed where the requisite plasma heating and current generation was produced by rapid reversal
of the axial magnetic field in cylindrical coil geometry. Once formed, the FRC is observed to be
stable and the plasma well confined as long as the plasma remains in a kinetic regime. This
regime is characterized by S*, the ratio of the FRC separatrix radius, rs and the ion collisionless
skin depth c/pi. Past FRC scaling has shown that stable FRCs can be formed at low flux where
ion kinetic effects dominate. Both stability and transport are observed to rapidly deteriorate when
the kinetic condition S*/ > 5 was exceeded. Here  is the FRC separatrix elongation  (= ls/2rs)
[2]. In any case the FRC decays on a resistive time scale that is anomalous. FRC confinement in
the kinetic regime was initially observed to scale roughly as  ~ r2/i [3] where i is the ion
Larmour radius at the FRC separatrix. Since the FRC has primarily only a poloidal magnetic
field, the plasma pressure at the null must equal the radial pressure exerted by the external field in
equilibrium,

Be2  20 n0 k Te  Ti  , (1)

where the zero subscript refers to the value at the magnetic null radius R (= rs / 2 ). With Ti ~ Te
one has 1/i ~ n1/2 inferring that the diffusion coefficient for the FRC is independent of radial
scale and has only a positive scaling with density. Later results on LSX [4] and other experiments
indicated further dependences with the FRC elongation,  and the ratio of FRC separatrix radius,
rs to coil radius rc, with this ratio designated as xs.
The observed particle confinement, stated in terms of directly measured quantities that can be
accurately measured across all experiments yields the following scaling [5]:

 N  3.2  1015  0.5 xs0.8 rs2.1 n0.6 , (2)

where xs is the ratio of the FRC separatrix radius, rs to coil radius, rc. With reasonable
assumptions for the FRC relative size and shape ( ~ 15 and xs = 0.6), this scaling together with
kinetic condition, determine the plasma radius and density required to satisfy the Lawson criteria

2
 Supersonic Merging and Compression of FRCs

for fusion gain i.e., n0  1.5x1023 m-3 and rs  7.0 cm. The high plasma energy density implied by
these constraints prescribes a small, pulsed fusion regime for the FRC. These FRC parameters,
together with the required temperature and confinement however have not been obtained by any
of the formation methodologies that have been employed in past experiments.

2. Experiment
The ability of the FRC plasmoid to be translated several meters allows for the FRC formation
and kinetic energy for heating to be realized outside of the burn chamber and breeding blanket.
The high energy density state can then be obtained through the rapid conversion of the FRC axial
kinetic energy. The experiments to be described here were instigated to examine the physics of
this process. To accomplish this, a new device was constructed and is referred to as the Inductive
Plasma Accelerator (IPA) [6]. The initial configuration was equipped with an enlarged merging
chamber as illustrated in figure. 2.

Figure 2. Initial configuration of the IPA device and flux contours from
numerical calculations based on experimental conditions for FRC dynamic
formation and merging.

To minimize the formation time, as well as maximize the FRC acceleration, the IPA device was
constructed so that a new formation methodology could be employed, which is referred to as
dynamic formation [7]. The monolithic theta pinch coil, utilized for FRC formation in virtually all
previous FRC experiments [1] was replaced with a set of electrically isolated and independently
triggered formation/acceleration coils. In the initial setup five such coils were employed, and all
were supplied with an initial reverse bias field ~ 0.06 T. A forward bias was applied to the end
coils as well as the coils in the merging/compression section. For the compression experiments,
the formation section was increased in radius to provide for greater initial FRC flux and energy.
This was followed at smaller radius by a set of pure accelerator coils with positive bias for the
addition of kinetic energy after FRC formation. The merging/compression chamber was reduced
to this smaller radius as well (see figure 3). For both configurations the formation coils were
energized sequentially to both form and accelerate the FRC simultaneously. By executing field
reversal in this sequenced fashion the FRC flux can be maintained throughout the entire process

3
 Supersonic Merging and Compression of FRCs

[7]. The terminal FRC velocity was 210 km/sec with the initial configuration and as high as 300
km/sec in the final configuration with additional accelerator coils.

Figure 3. IPA configuration for FRC acceleration, merging and compression. The
flux contours from the MHD calculations correspond to the initial magnetic field
structure (0), the peak axial compression (8) and magnetic compression (16).

The initial plasma was generated by an annular array of 16 small coaxial plasma guns located
near the quartz vacuum tube wall under the first formation coil (see figure 3). These devices were
each inductively isolated and pulsed at high current (3-6 kA) and produced a fully ionized plasma
flowing at roughly 2 cm/sec. The plasma guns were based on the magnetoplasmadynamic
(MPD) thruster [8] which produces a fully ionized plasma, and the absence of significant neutral
flow prior to initiation was characterized in separate experiments. The ejected plasma coalesced
to form an annular plasma as indicated by end-on framing pictures. This provided for good flux
trapping on reversal as indicated by the large excluded flux signal during reversal, and the
formation of a more oblate FRC for acceleration. The use of plasma guns in this manner also
provided for a means to keep the merging chamber and accelerator at high vacuum. This was
important as a significant neutral density considerably affected FRC behavior that prevented both
FRC merging and the establishment of an equilibrium of any significant duration as indicated by
the diamagnetic signals in the compression section.
In the initial experimental configuration, the FRC was dynamically formed and injected into a
larger chamber to study the merging process without the complications of compression and
additional acceleration. The merging chamber consisted of a 0.14 m radius quartz cylinder lined
with a 0.3 m long cylindrical, thin tungsten liner. A suitable bias field of 0.1 - 0.2 T was
embedded inside this liner to radially confine the FRC plasmas during merging, while two multi-
turn coils provided for a small mirror field (~0.2 T) on either end of the chamber. An array of
external flux and Bz field loops were installed under each coil as well as an axial array of Bz
probes inside the tungsten shell. From this array the excluded flux due to the presence of the FRC
was obtained, and the FRC velocity, radius, length, and energy was determined. A HeNe laser
based interferometer measured the cross tube line density at the axial midplane. From this
diagnostic and the magnetic measurements, the plasma density and pressure balance temperature
were obtained. Deuterium plasmas were employed and a calibrated neutron detector was
positioned radially outside the magnets at the chamber center to measure the D-D fusion neutron
flux.
The ratio of FRC directed velocity to thermal velocity was found to be greater than three. The
merging and conversion of the FRC kinetic energy was observed to take place on the Alfvenic

4
 Supersonic Merging and Compression of FRCs

timescale. The distinct difference between merging and the simple passage of the FRC unopposed
is clearly demonstrated in the array of excluded flux signals in the interaction chamber as
illustrated in figure 4. The two FRCs merged as they collided to form a single FRC indicated by
the peak diamagnetism appearing and remaining at the axial midplane. It is worth noting that with
merging the excluded flux increased by a factor of at least four greater than observed for the
transiting FRC.

Figure 4. Excluded flux from diagnostic array in merging chamber for the
experimental configuration shown in Fig. 1 for (top) single FRC passage, and
(bottom) colliding FRCs.

In a constant flux cylindrical coil the thermal conversion of kinetic energy would be
manifested by a large increase in plasma length as the equilibrium axial magnetic pressure
constraint determines the FRC average beta [3],

1
  1  xs2 (3)
2

An increase in plasma radius indicated by the greatly increased excluded flux signal would
normally only occur from a large increase in poloidal flux. This large increase in excluded flux
has also been observed in some translation experiments where a single FRC expands and comes
to rest to fill a large chamber, and has been attributed to the possible conversion of toroidal flux
into poloidal flux [9,10]. Based on the results here however, this is not likely to be the correct
explanation for several reasons. Any formation induced toroidal flux is oppositely directed for

5
 Supersonic Merging and Compression of FRCs

each of the merging FRCs and would be annihilated with merging as observed in spheromak
counter helicity merging [11].
There is a more plausible explanation for the increased plasma radius than an increase in flux.
If the equilibrium length of the FRC were to expand beyond the region over which the flux is held
constant, the changing end conditions will significantly affect the average beta. In equilibrium the
plasma pressure along a field line is constant, and must also be in radial pressure balance with the
local external field, or Bend = Be = Bvac/(1-xs2). At this critical value an average beta condition can
be derived from the Maxwell stress tensor in a manner similar to that obtained for constant axial
flux [3] with the result:

1 2 4
 0
1 xs 0 R (4)
2

where R is the ratio of the FRC separatrix in the end region to that at the midplane. If the region
at the end of the FRC is maintained with a magnetic field that is greater than the external field at
the FRC midplane, (Bend > Be), the FRC is forced to contract axially and expand radially. With a
large increase in plasma energy from thermal conversion, the FRC must expand in radius
compressing the axial field between the FRC and flux conserving coil until Be ~ Bend. At this
point the FRC can expand axially and equation (4) will hold. In this case, the FRC is
characterized by a higher  than given by the constant axial flux relation in equation 3.
The interpretation of the IPA experimental results was greatly aided by the use of 2D resistive
Magnetohydrodynamic (MHD) calculations performed with the Moqui code [12]. By initializing
the code with the experimental magnetic waveforms from the experiments, the dynamic behavior
of the FRC formation, acceleration and velocity was matched by adjusting the initial flux and
plasma inventory. The same large increase in excluded flux is observed in the code on
thermalization as it was in the experiments with no change in the total poloidal flux, thus
confirming the interpretation given here.
There was one significant difference between the code and the observed behavior of the
colliding FRCs as can be seen in the results of the calculations shown in figure 2. Complete
merging was not produced in the code whereas it was always observed in the experiments. In the
code the FRC kinetic energy is almost completely thermalized upon collision (>99%), but the
small residual rebound velocity of several km/sec causes immediate separation of the FRCs
prohibiting further reconnection, and undoing whatever reconnection that had occurred. Magnetic
reconnection is one of the central processes in plasma physics. One of the key issues in
understanding magnetic reconnection is what determines the magnitude of the reconnection rate.
In simple two-dimensional models as employed here, the reconnection site is characterized by a
local null point of the magnetic field with field lines extending from the null forming an x-type
separatrix. In resistive MHD models of reconnection this rate is controlled by the magnitude of
the resistivity. However modifying the resistivity in the Moqui code, even with density weighted
resistivities, it was not possible to obtain complete reconnection of the merging FRCs. This is
counter to the resistively driven reconnection model found to explain the reconnection rates
observed in merging spheromak experiments [13]. The observed rate in spheromak reconnection
is far too slow to explain the rate observed in the experiments here where the FRC is found to
reconnect at Alfvénic (MA ~ 1) speed.
In contrast to typical resistive models, numerical simulations of collisionless reconnection
have demonstrated that magnetic reconnection in sufficiently thin current sheets, in which Hall
effects and whistler dynamics become important, may occur at a fast rate that does not depend on
the dissipation mechanism. The reconnection rates predicted from codes with the Hall term, as
well as pure particle codes, predict reconnection on timescales of 30-50 ci-1 [14]. The ion
cyclotron time, ci-1, at peak axial compression was 1 and 0.1 s for the expansion merging and

6
 Supersonic Merging and Compression of FRCs

compression merging experiments respectively. Again, the reconnection timescale predicted is

much too long. The likely explanation is that the reconnection is inherently three dimensional,
and that the local resistivity may be highly anomalous. Work is currently underway at the
University of Washington employing a 3D MHD code referred to as NIMROD [15,16] which
includes both the Hall and resistive terms. Hopefully this code can shed more light on the exact
mechanism at play here.
The basic FRC equilibrium parameters observed during compression are displayed in figure
5.

Figure 5. FRC experimental data at the axial midplane during compression for
the IPA configuration shown in figure 2. Plotted from top to bottom: the radially
averaged electron density, the axial magnetic field external to the FRC, the
excluded flux (plasma) radius, the radially averaged total plasma temperature,
and the D-D fusion neutron signal detected by a scintillator based neutron
detector.

Total temperature was calculated based on radial pressure balance. During compression it can
be seen that this resulted in total temperatures of 1 keV or more. The overwhelmingly larger ion
mass dictates that the ions receive virtually all the heating from kinetic conversion of the FRC
motion. The MHD calculations reflect this with ion temperatures Ti ~ 850 eV after collision and
compression. A strong neutron signal was detected during magnetic compression from two
calibrated scintillator-based detectors, manufactured by Eljen Technology. These detectors were
employed at the axial midplane to monitor the time history of the neutron flux. The EJ-410
detector employed is specifically designed for detecting fast neutrons while being nearly
insensitive to gamma radiation. The detector consists of zinc sulfide phosphor embedded in a
hydrogenous polymer matrix structured in a series of concentric cylinders of clear plastic to
facilitate light collection. The fast neutron detector is directly mounted to a photomultiplier with a
time response of 0.2 sec (scintillator decay constant). A Monte Carlo N Particle (MCNP)
transport code was employed to correct for neutron passage through the quartz tube, Aluminum
magnets as well as the lead shielding placed around the neutron detector itself. Scattering and
slowing from surrounding structures accounted for about a factor of two reduction in neutrons
detected on IPA according to the MCNP calculations. Even with correction for FRC geometry,
attenuation and scattering in intervening material a much higher ion temperature, T i ~ 2 keV was
inferred. The anomalously large signal was well beyond what could be attributed to measurement

7
 Supersonic Merging and Compression of FRCs

error of the plasma density and volume. Most likely it is the result of a non-thermal ion
population, but the mechanism for maintaining this over the ~30 s of the FRC compression time
is not clear.
The confinement observed for the merged and compressed FRC was better than the scaling
inferred from equation 2 by roughly a factor of 2 for the conditions shown in Fig. 5. This scaling
was based on FRCs confined in constant flux coils. Enhanced confinement has also been
observed in translation experiments where the FRC is expanded into, and completely fills a larger
chamber [9,10]. It would be tempting to attribute the improved confinement to flux generation,
but as discussed, this interpretation is not warranted when the FRC is mirror confined. Previous
compression experiments of a translated FRC into a constant flux coil did not exhibit enhanced
confinement [17]. Mirror confinement provides several possible mechanisms that could lead to
improved confinement. It produces a stronger magnetic nozzle at the ends of the FRC slowing the
open field line plasma loss, which in turn relaxes the large radial pressure gradient across the
separatrix that would suppress instabilities such as lower hybrid drift that has been calculated to
produce the transport scaling observed for FRCs. The increased plasma radius and beta reduce the
magnetic gradient at the null and could thus reduce the flux loss for the same plasma resistivity.
The rotational instability that is always observed later in the equilibrium with FRCs formed in
situ was observed in the merged FRC experiments with compression, but not expansion. The
rotational instability occurs when , the ratio of the plasma rotation compared to the ion
diamagnetic drift frequency, approaches one. For the rigid rotor profile it can be shown that  
L/xs2 [3]. The experimental results are in accord with this scaling if it is assumed that the angular
momentum L acquired by the FRC is similar for both cases. It has been shown that a significant
toroidal field can stabilize this mode [18], so that the observed spin up under compression would
indicate that little toroidal field is present after merging.

3. FRC fusion scaling

The fusion reaction rate, dNfus/dt (n/sec), for D-T fusion systems is given by

d N fus
dt
  nD nT v dVol , (5)
vol

where n is the number density of species , and v is the reaction rate parameter for the fusion
reaction of interest for a plasma with ion temperature Ti. It will be assumed that nD = nT = ½ n
where n is the total ion density. For pulsed systems such as the IPA based FRC device considered
here, the relevant quantity is the total neutron yield per pulse. Equation 5 should thus be
integrated over the lifetime of the plasmoid and then multiplied by the repetition rate Rp. to obtain
the neutron flux. For the purposes here it will be sufficient to multiply the left hand side of
equation 5 by the energy confinement time. The energy confinement is typically 2/3 of the
particle confinement time [19], and the particle confinement for merged FRCs was found to be
roughly a factor of two better than the expression in equation 2. During the decay of the FRC the
plasma density and temperature are maintained as the product must be in radial pressure balance
with the external magnetic field (see figure 1). The total fusion energy provided by the D-T
reaction is Efus = Nfus(22.4 MeV)(1.6x10-13 J/MeV). This includes an additional 4.8 MeV from
the lithium fusion in the blanket required to obtain a new triton for continued operation of the
reactor. With this one can write for the FRC:

Pfus  1.2  1012 n2 v N Vol FRC R p . (6)

8
 Supersonic Merging and Compression of FRCs

For the highly elongated FRC, the volume can be approximated VolFRC = rs2ls. For temperatures
between 10 and 20 keV, the usual approximation for the D-T fusion cross section is:  
1.1x10-31 Ti2(eV). It is clearly desirable to operate at the upper limit in magnetic field pressure as
this maximizes the fusion yield. Since increasing the ion temperature comes at the expense of
decreasing n for a given magnetic pressure (see equation 1) there is thus no advantage in seeking
higher ion temperatures. In fact there is a disincentive in that the FRC lifetime also decreases with
decreasing n. It turns out that for ion temperatures in the range of 3 to 12 keV - the range
anticipated for the optimal operation of the IPA device, a fit to the cross section is better
approximated by   4x10-33 Ti2.6(eV) as is indicated in Figure 6.

Figure 6. Nuclear cross section for D-T fusion and approximate fits over range of
interest for pulsed FRC fusion.

Given the FRC confinement scaling, and the ion temperature within this range, the neutron yield
is essentially independent of the ion temperature. More precisely, the increase in fusion cross
section at higher temperature is offset by decreased density and FRC lifetime at a given
compression magnetic field.
While the physical dependencies for all terms in equation 6 have been identified, it is worth
introducing quantities that are primarily determined by the device parameters that can be changed
and optimized. The fundamental parameters in this regard are the compression magnetic field, Be,
the compression coil radius, rc, and the magnitude of the FRC poloidal flux p. Assuming that the
current profile is similar to the rigid rotor profile commonly inferred by observed pressure
profiles [20], the closed poloidal flux, along with Be and rc in turn determine the plasma radius, rs,
as:

1/ 3
r3  rc  p 
 p  s Be  rs   
 . (7)
rc  Be 

With equation 7 for the FRC radius, and solving equation 1 for the density, along with the
appropriate approximation for the D-T cross section as stated above, the FRC fusion energy per
pulse (from equation 6) can be rewritten in terms of the experimentally controllable variables
as:

9
 Supersonic Merging and Compression of FRCs

E fus  1.14  105 Be3.73 1p.47 l1s .5 rc0.67 (8)

The FRC poloidal flux and final FRC length are primarily a function of the scale of the formation
sections positioned at each end of the device (see figures 2 and 3). To obtain the fusion gain,
equation 8 must be normalized to the FRC energy acquired each pulse. The FRC energy can be
stated as:

B2
N k Te  Ti   Vol FRC  e  rs2 ls ,
3
EFRC  (9)
2 2 0
Employing equation 7 again for the FRC radius, one has:

E fus
G  0.093Be2.4  0p.82 ls0.5 . (10)
EFRC

It is notable that the gain has no dependence on the compression coil radius, and only a fairly
weak dependence on the FRC length. Gain contours as a function of the FRC poloidal flux and
compression field are shown in figure 7.

Figure 7. Gain contours as a function of the FRC poloidal flux and compression
magnetic field. It was assumed that the FRC length ls = 1 m.

This is not to say that one is free to pick any radius for the burn chamber compression coil. As
was pointed out earlier, there is the kinetic stability criterion, which was expressed as S*/ < 5
Stating this criterion in terms of the control variables:

S* 2 rs rs  pi Be1/ 3  2p / 3 rc2 / 3
  8.75  103  5, (11)
 ls c ls Ti1/ 2

10
 Supersonic Merging and Compression of FRCs

where the ion atomic number was assumed to be 2.5 for the D-T plasma. For kinetic and/or Hall
effects to stabilize the FRC, the ion gyro orbit must remain a significant fraction of the plasma
radius (e.g. either a higher Ti or lower p) With a reasonable assumption for the fusion ion
temperature (9 keV), equation 11 expresses the stability limit on the amount of poloidal flux for a
given coil radius and confining field:

ls3 / 2
 p  0.0126 . (12)
Be1 / 2 rc

The plot of the regions of the p – Be space of figure 7 that are excluded for various coil radii are
shown in figure 8.

Figure 8. Gain plot as in figure 7 with regions of exclusion for different

compression coil radii rc, where an ion temperature was 9 keV and a FRC length
ls = 1 m were assumed.

It is clear that to achieve fusion gain at the lowest possible compression field a coil radius of 0.2
m or less is desired, with a coil radius of less than 0.1 m assuring kinetic stability for the largest
range of both flux and compression field.
There are several reasons to keep the compression coil small. Certainly the amount of
compression field energy required per pulse is thereby minimized, but perhaps more importantly,
the wall loading per pulse is reduced as the ratio of wall surface area to reacting volume scales as
1/rc.

4. Pulsed FRC Fusion Reactor Prototype

One is now in a position to determine the basic parameters of a fusion reactor based on the
pulsed compression of the FRC. The initial design should aim to validate the concept with a
device that is capable of attaining G = 1. The goal would be to demonstrate this with an
experiment that would operate essentially in a single pulse mode (R p >> 1 Hz) to reduce the cost
and complexity of the pulse power systems, but to do it in a manner that could evolve into a
commercially viable device. This is deemed possible due to the several unique attributes of the

11
 Supersonic Merging and Compression of FRCs

concept that eliminate or greatly mitigate the major cost drivers inherent in the other approaches.
For instance, with cyclical plasma generation the need for sustainment, fueling and auxiliary
heating systems, including current drive, are eliminated which could potentially simplify reactor
operation. Having a transient burn, the vacuum boundary should be much easier to maintain
thereby significantly reducing recycling, tritium retention and wall interaction issues. A
conceptual design is illustrated in figure 9.

Figure 9. Conceptual depiction of a prototype fusion reactor based on the pulsed

compression of the FRC

In figure 9 the radial scale has been exaggerated and the length of the burn region inside the
compression coil has been foreshortened for purposes of illustration. The breeding blanket, which
would immediately surround the high field chamber, is also not shown for clarity. The cross
section of the burn region and blanket is illustrated in figure 10. There are several alternative
possibilities for the configuration of compression coil and fusion blanket. The use of a beryllium
inner wall as part of a flux concentrator and neutron multiplier is one possibility as indicated in
the figure. Other arrangements include the use of a conventional multi-turn solenoid at the inner
boundary. While the neutron fluence would no doubt limit the lifetime of such a coil, the linear,
simply connected nature of the confinement system could make periodic replacement an
acceptable arrangement. A structural analysis of these various options has yet to be done and
would need to be performed to determine the best method for field creation in the burn chamber.

Figure 10. Midplane cross-section of compression section with fissile-fussile

breeding blanket.

12
 Supersonic Merging and Compression of FRCs

In order to make sure that the assumptions made in the prototype design are consistent with
the scaling process of the FRC formation and merging demonstrated with the IPA results, several
runs of the 2D MHD Moqui code were conducted using typical fields, voltages, and fill
inventories used in past FRC experiments for the FRC formation. The radial scale of the
formation coils was set the same as the LSX FRC experiment [21]. The coil fields and timings
were adjusted to produce the same supersonic FRCs as achieved in the IPA experiments with
merging occurring in a compression coil with the same radius (rc = 0.1). The major difference
being that the compression field for the calculations was increased to 10 T. The results are
displayed in figures 11 and 12.

Figure 11. Field line and pressure contours from 2D MHD calculation of the
formation merging and compression of the FRC in the prototype.

The unique ability of the FRC plasmoid to be simultaneously translated and compressed over
distances of several meters allows for the FRC formation and kinetic energy input to be added
incrementally outside of the burn chamber and breeding blanket. This is clear from the time
history of the ion temperature shown in figure 12. The ion temperature increases from a
combination of pre-compression and kinetic conversion prior to the merging and final conversion
in the compression chamber at 24 sec. In the process the ions receive the bulk of the heating.
After the FRC merges and heats, any residual axial motion is only weakly damped as the FRC
velocity is subsonic. There was sufficient reconnection in this calculation that the FRCs do not
become separated even though the residual velocity is not small.

13
 Supersonic Merging and Compression of FRCs

Figure 12. (top) the time history of the average electron and ion temperature
inside the FRC. (bottom) The time history of the FRC axial velocity and external
magnetic field at the midplane of the FRC.

The compression of the FRC to high energy density can be performed inside a simple, small
cylindrical coil set that is constructed for pulsed, high magnetic compression fields. This kind of
magneto-kinetic heating is efficient and avoids the numerous challenges confronting steady state
approaches where large ports and sophisticated neutral beam devices are needed to create, heat,
and possibly sustain the plasma currents. These ancillary devices must also be co-located with the
fusion reactor blanket and power processing systems making for a more complex and inefficient
reactor.
Operation with tritium will necessitate recovery of unused fuel. A large diverter region is thus
located at each end of the formation sections (see figure 9). The plasma exhaust (diverter) regions
can be as large as desired and well removed from the reactor, eliminating critical power loading
issues. As can be seen in figure 11, the entire high field reactor vacuum flux is external to FRC
plasmoid flux and is thus essentially diverter flux. In a transient burn, the particle loss from the
FRC will be overwhelmingly directed to the diverter regions as the axial flow time is many orders
of magnitude smaller than the perpendicular particle diffusion time across the open flux region.
The strong axial gradient magnetic field outside the burn chamber assures that the plasma lost
from the FRC remains contained in the divertor regions essentially eliminating tritium co-
deposition and making the tasks of full tritium recovery and diverter maintenance much easier to
carry out.

5. Discussion and Conclusion

With the successful merging of supersonic FRCs in the IPA experiments, a method with the
potential for the production of fusion energy in a simple and technologically appealing manner
has been demonstrated. FRC plasmoids were brought to high temperature and density via a linear,
sequenced array of low field compression coils establishing a means for obtaining fusion
conditions remote from the plasma formation system. To obtain the requisite Lawson product a
larger FRC must be formed, accelerated and compressed to higher fields. From the results of the

14
 Supersonic Merging and Compression of FRCs

IPA experiments together with equation 10, breakeven conditions will require an increase in
poloidal flux which can be achieved by a radial scale increase of roughly a factor of three to a
formation size similar to the LSX experiment [21]. The merging chamber would remain at 0.1 m
radius but at much higher field (10 T). The same magneto-kinetic compression scheme as used on
IPA would be extended in order to impel the FRC into the higher fields of the burn chamber.
The suitability of this approach to fusion or fusion-fission hybrid applications [22] is the
consequence of the numerous critical advantages that significantly reduce the technological
challenges, as well as the cost of development and operation. The more significant begin with the
ability of the FRC plasmoid to be translated and merged where the FRC formation and kinetic
energy input is added outside the burn chamber and realized only within the burn chamber as
demonstrated in the IPA experiments. The plasma divertor region is well removed from the
reactor, and can be scaled to eliminate any power loading issues. With the diverter located
remotely in an essentially neutron-free environment, tasks such as full tritium recovery and
diverter maintenance are made much easier to perform.
In order to rapidly spawn additional fusion systems, maintaining a low tritium inventory and
creating more tritium than consumed will be essential. Due to the pulsed FRC reactor’s linear,
cylindrical geometry and the small scale of the fusion plasma compared to the neutron absorption
region, the breeding blanket coverage can be optimized. It is critical that a sufficient number of
energetic neutrons interact with lithium in the blanket to breed new tritium fuel. A high
conversion efficiency of the fast fusion neutrons in the blanket surrounding the device makes
significantly more fusion neutrons available for both fussile and fissile fuel generation as well as
waste burning applications. The simply connected, linear geometry of the reactor vessel is also
amenable to rapid and frequent first wall replacement if necessary. A fluid metal wall interface in
this geometry is also feasible [23]. This would allow for operation at the highest power density,
and resolve several plasma-material wall issues.
There are also several critical attributes related to the unique nature of the FRC as a fusion
plasma. Reactor costs will generally scale inversely with the fusion power density which in turn
scales as 2B4. The FRC has the highest  of all magnetically confined fusion plasmas, and the
simple cylindrical nature of the confining field coils provides for the highest magnetic fields.
Utilizing the FRC plasmoid for magnetic fusion energy thus represents the highest fusion power
density that can be attained in a non-destructive, repetitive manner. As indicated in figure 10, the
small ratio of the FRC volume to the neutron absorbing blanket volume provides for the
maximum fusion power density obtainable without exceeding thermal limitations in the blanket.
For hybrid operation, the small footprint allows for easy integration into existing fission power
plant infrastructure. The reduced reactor scale also means much faster and far less expensive
iterations during the development phase which will be essential to the integration of
improvements and new technologies.
At first thought, it would appear that a pulsed system would be inferior to a steady system,
primarily due to the advantages of self-sustainment (i.e. ignition). There are several limitations
imposed by a steady fusion burn, not the least being constrained to an inherently low power
density. As is now being realized, the steady burn does not easily allow for maintenance of the
plasma-wall interface. The ability to maintain the critical balance between alpha heating and
plasma losses in an ignited reactor is an open question; as is tritium retention, fueling, burn
diagnostics, plasma current sustainment, profile control, as well as several other device specific
issues such as disruptions. By virtue of the cyclic nature of the burn, virtually all these issues are
eliminated or significantly mitigated. Both fusion reactor gain and diverter wall loading can be
readily regulated by the compression strength and rep rate Rp. All of the D-T fuel can be
introduced during the initial formation of the FRC plasmoid, eliminating the need for refueling.
The major challenge for pulsed systems is operation at the highest efficiency possible in order
to minimize what would otherwise be a large recirculating power fraction. Using pulsed magnetic
fields to form, accelerate and compress the FRC to fusion conditions is efficient with Ohmic

15
 Supersonic Merging and Compression of FRCs

losses in the drive circuits as the primary loss channel. Employing a pulsed magnetic field also
affords a unique opportunity for direct energy recovery. It has been appreciated in the past that
pulsed systems employing flux compression can realize direct recovery of the fusion alpha energy
[24]. For the pulsed compression of the FRC this takes the form of the back emf experienced by
the compression circuit due to the increase in FRC plasmoid pressure from the presence of the
fusion alphas. This push-back during the natural oscillatory behavior of the compression circuit
causes the energy storage capacitors to be recharged to greater energy from the increased
expansion energy of the FRC. A key technology for such a system is the use of an opening switch
to allow this energy to then be stored for the next pulse. This appears feasible as the voltages and
energies envisioned for the prototype are within the range of what can be achieved with current
solid state switches and power delivery systems [25]. Additional recuperative operation of all the
magnetic field circuits would provide for a way to significantly reduce the energy losses to
primarily that of the FRC itself. Even that energy may be recoverable in some measure by direct
conversion in the divertor as the plasma energy loss is primarily in the form of a high energy,
directed plasma jet at the divertor entrance [23,26]. The capability to make use of the plasma,
fusion and electrical energy in an efficient manner is unique to the concept described here, and is
critical for the commercial application of fusion to be realized without resorting to a larger scale,
higher fusion gain system.
Possibly the most underappreciated advantage of a pulsed system comes from the ability to
evaluate key performance milestones with single pulses. This minimizes neutron activation
during development and further reduces cost as much simper and less robust pulse power systems
can be employed. Given the scaling and the desire to optimize at smaller radius and higher power
density, it is worth determining the parameters of a device capable of a gain of five. From
equation 10 the following values will yield such a gain:

ls  2 m  p  25 mWb Be  16 T  Q5. (12)

With a compression coil radius rc = 0.2 m, the FRC should remain well below the kinetic
stability limit given in equation 12 even prior to full compression. The fusion energy per pulse is
given by equation 8 and for the assumed parameters yields Efus = 14 MJ. The rep rate Rp is
determined by the wall loading limits. Taking advantage of the translatability of the burning FRC,
the compression section can be made much longer than the FRC so that the drifting FRC deposits
the fusion energy along a series of compression coil sections. Alternatively, the FRCs can be
merged and brought to rest in a different part of the elongated compression chamber with each
pulse. In either case, the rep rate will likely never exceed 2 – 4 Hz for average power of 40 to 60
MW in order to maintain the average neutron fluence below 4 MW/m2. The low frequency
minimizes the pulse power issues as well as any pump out concerns.
A low total fusion yield would at first seem restrictive. It does however have several
advantages. With eventual fusion plant power output in the 200 to 500 MW range, the pulsed
FRC reactor can be modularized. The investment-intensive energy input and fuel recovery
systems can be used to power multiple, relatively inexpensive, easily replaceable burn chambers.
This will be particularly important in the early stages of deployment when reliability and
maintenance schedules will be less well understood.
In summary, with the repetitive and efficient generation of FRC plasmoids, brought to high
temperature and density as they are injected into the burn chamber, a compact, low-cost fusion
neutron source can be achieved. At a minimum it would be an attractive source for fast neutron
based applications such as isotope production, fissile fuel breeding and fission waste
transmutation. The expectation is that with the successful demonstration of a prototype scale
device, a viable alternative for fusion energy production can be rapidly developed at an affordable

16
 Supersonic Merging and Compression of FRCs

cost, and will spur a renewed urgency to perform the required reactor development on a timescale
that will allow fusion to be a major player in the world’s energy future.

Acknowledgements
The authors would like to thank Dr. David Kirtley for his assistance in calibrating the neutron
detector response, and Dr. Samuel Andreason for invaluable contributions to the design and
construction of several electronic components. The work was partially supported by grants from
the U.S. Dept. of Energy.

17
 Supersonic Merging and Compression of FRCs

References
[1] M. Tuszewski, Nuclear Fusion 28 2033 (1988).
[2] M. Tuszewski, D.C. Barnes, D.J. Rej, and D.P. Taggart, Phys. Fluids B 3, 2205 (1991).
[3] W.T. Armstrong, et al., Phys. Fluids 24, 2068 (1981).
[4] J.T. Slough, et al., Physics of Plasmas, 2, 2286 (1995).
[5] John Slough, et al., J. Fusion Energy 26, 199 (2006).
[6] G. Votroubek, J. Slough, S. Andreason, and C. Pihl, J. Fusion Energy 27, 123 (2008).
[7] John Slough and Richard Milroy, J. Fusion Energy 29, 567 (2010).
[8] Hirokazu Tahara, Yoichi Kagaya, and Takao Yoshikawa, Journal of Propulsion and Power,
13, 651 (1997).
[9] H.Y. Guo, A.L. Hoffman, K.E. Miller, and L.C. Steinhauer, Phys. Rev. Lett. 92, 245001
(2004).
[10]. M.W. Binderbauer, et. al. Phys. Rev. Lett., 105, 045003 (2010)
[11] Y. Ono, M. Yamada, T. Akao, T. Tajima, and R. Matsumoto, Phys. Rev. Lett. 76, 3328
(1996).
[12] R.D. Milroy and J.U. Brackbill, Phys. Fluids 25, 775 (1982).
[13] M. Yamada, et al., Phys. Rev. Lett. 65, 721 (1990).
[14] J. Birn et al., Geophys. Res. Lett. 32, L06105 (2005).
[15] C.R. Sovinec, et al., J. Comp. Phys. 195, 355 (2004).
[16] R.D. Milroy, A.I.D. Macnab, C.C. Kim and C.R. Sovinec, J. Fusion Energy 27, 73 (2008).
[17] D. J. Rej, et al., Phys. Fluids B 4, 1909 (1992).
[18] R.D. Milroy and L.C. Steinhauer, Phys. Plasmas 15, 022508 (2008)
[19] A.L. Hoffman, J.T. Slough, Nuclear Fusion 33, 23 (1993).
[20] R.E. Chrien and S. Okada, Phys. Fluids 30, 3574 (1987)
[21] J.T. Slough, et al, Phys. Rev. Lett., 69 2212 (1992).
[22] John Slough, David Kirtley, Chris Pihl, George Votroubek, and Philip Wallace, “The Pulsed
High-Density FRC as a Neutron Source for Fusion-Fission Hybrid Applications”,
http://www.science.doe.gov/ofes/FF_Hybrid_Report_Final.pdf, pg. 70, (2010).
[23] A.A. Harms, K.F. Schoepf, G.H. Miley, D.R. Kingdom, 2002 Principles of Fusion Energy,
World Scientific Publishing, Singapore 912805, pgs. 267-277.
[24] R.W. Moir, Nuclear Fusion, 37, 557 (1997)
[25] A. Pokryvailo, C. Carp, C. Scapellati, IEEE Transactions on Plasma Physics,38, 2734
(2010).
[26] G.L. Kulcinski, et. al., Fusion Technology 19, 791 (1991).

18

View publication stats