Project Pacer And the Unbuilt Pacer Economy Part 2: Ralph Moir’s 2 Kiloton Variant and the Future of Pacer

Project Pacer And the Unbuilt Pacer Economy Part 1:

This is the second of two articles. This gives some history of the possible need for near term D-D fusion neutrons to maximize fission fuel supplies. This  second part covers more modern redesigns of the concept, notably by Ralph Moir in the 1990s and the fact that the vast industries Pacer enables were never built.
 the Unbuilt Pacer Economy
We also discuss some ideas about the future of Pacer.

In the first post of this series  we discussed the general question of the length of time earthly fissionables might last and what an aid D-D fusion is to stretching those supplies into geological time.

 The 1975 Los Alamos Progress Report on Project Pacer was covered in the first part of this article and can be found here:
There have been various incarnations of Project Pacer, including before it was formally born, from 1957 on; a lot of people (especially those working in the nuclear weapons complex in the USA) began thinking on how to get power out of nuclear explosions, because at that time, only 15 years after the construction of the first artificial uranium reactor in 1942, we knew how to make uncontrolled fusion reactions, and in fact had been doing so for 5 years already: The hydrogen bomb.

 In a Project Pacer containment facility underground, you would not see mushroom clouds but this gives you a feel for the yields under consideration (as well as a reminder why you want them underground):

Remember that the first considerations were 1 megaton devices, Project Pacer in 1975 was thinking about 800 individual 50 Kiloton devices yearly  (40 megatons explosive equivalent yearly or around 2 gigawatts electrical at 30% efficiency,and Ralph Moir’s 2 kiloton device based variant is considered below.

Were larger explosions ever considered? I’m sure they must have been, I considered them in this article

 1 megaton device– $10 million–4184000 gigajoules per megaton
$2.39 per giajoule– about the same price as coal
29.3 gJ/ton coal at $80 ton gives $2.73 a gigajoule coal

10 megaton device– $10 million–41840000 gigajoules per megaton
0.239 per giajoule 1/10th the cost of coal

So, 100 megaton device– $10 million–418400000 gigajoules per megaton = ~1/100th the cost of coal

The takeaway is anything much below a megaton device is very
expensive relative to the larger sizes.

If the devices were free, the deuterium being the only cost, the price would be 1/1891 that of coal at $80 ton (given Deuterium at $500/kilogram)

which basically comes to the conclusion that assuming ‘weapons complex like costs per device’, nothing much under a megaton makes economic sense.

I should explain that phrase– a lot of government operations ramp up costs continually and every few years allocate the overhead expansion to costs to each of their subprograms. It is a huge problem in government and Western civilization in general, and it comes down to a kind of organizational hardening of the arteries where more and more financial and time input is required for less and less result.

 Thus today we spend more time and money on studying and building the latest government large launch vehicle then a much newer and leaner NASA working with incredibly less equipment and worse conditions had to spend actually building the real Saturn V and the new rocket has not yet flown and may not ever.  We killed the Saturn V to save what wiki characterizes as  Cost per launch, $494 million in 1964–73 dollars ($3.2 billion present day) 
 And today pay for a development budget as big as that which launched it without actually having anything to launch.
Same for fusion reactors and same for the nuclear weapons complex.  The US Air Force in 1960 could not hit targets more accurately than we can today with better equipment but they could generate more sorties on less notice for less real money, and so on. In 1959 the US government could build 20 nuclear weapons a day at probably a tenth the real cost each that merely rebuilding a single weapon would take today. Similarly there is a fusion complex  of overhead to pay for even though there are no working fusion reactors.

All this is a roundabout way of saying that if you can build explode and recycle the isotopes of  nuclear devices for $1000 each you can make money running a 2KT Pacer. If it takes $25000 to build explode and recycle the isotopes you can make money with a 50KT Pacer (assuming you can keep the cavity engineering cost down which tends to be easier with the smaller sizes than larger)  If it takes $25 million per device to build explode and recycle the isotopes  you will lose money unless you are blowing off 50 megaton Tsar Bombas twice a day.
On the raw power of a 50 megaton bomb–

  (There is a concept-800 Tsar Bombas per year, output 2 terawatts, not gigawatts electrical-– Wow. If 97% fusion or better like the original was that is enough isotopes to produce 9300 kg of  U-233 per shot. Wow.).  I am not sure how you would leakproof piping in a Tsar Bomba like Pacer but what’s a little local earthquake between friends?  Unless you have AB-Matter available

 If you had AB-Matter to build a super-Pacer –yes you could make huge money on the 50 megaton units– but if you had AB-Matter you wouldn’t NEED to do that because there are enormously more profitable things you could do with AB-Matter— sigh. Back to the article

It’s not just  that the 50 megaton versions shake the ground for 10 times the radius; the real reason they would be very hard to build is the wider the unsupported arch of the roof the easier it is to collapse. THAT (tightness of the chamber against 800 thermonuclear blasts a year for decades) is why salt domes (highly plastic) were considered for the 50 Kiloton version of Pacer in 1975 and why the Ralph Moir version is only 2 kilotons (smaller, tighter, easier to engineer).

EVERY Pacer unit involves local earthquakes. (a number of kilometers away they are not noticable to anyone without a seismograph)

In the last article the 1975 report stated “Seismic effects, which are peculiar to PACER are kept small: the baseline 2000 Mwe station produces “thumps” so small as to go unnoticed a few miles from the site”
I believe I was saying that if you can keep your costs down you can make money building smaller versions of Pacer.

Remember it is the TOTAL complex cost including the bombs that adds up. You got overhead? It goes on the bill. Bombs go on the bill, piping problems go on the bill. That visitor center? On the bill.  To keep it economical keep expenses down.

This is of course the exact opposite approach to fusion as ITER which is basically free organizational money forever with no binding deliverables and no thought of  actual financial payback ever except as a theoretical far-future exercise whenever a complaining government asks about it. (This is not to diss their work or say that they are not trying serious approaches to breakeven, just that I don’t believe that future practical working fusion reactors will claim direct descent from ITER)

 Iter is Latin for “The Way”.  ITER claims it is the way to new energy. But one premise of this pair of articles is that perhaps PACER was ‘the way’ that was never seriously considered because to be anti-nuclear and anti Project Plowshare was cool in the 70s even before Three Mile Island.  President Carter was trained as a nuclear engineer but cut fusion research and part of that was ending the Plowshare Project under which PACER began. (Google Carter cut fusion research and see what you get)  However those who hate Republicans can read this,  The key thing is I am confident some version of PACER would work and Carter’s budget cut that; by the time the Reagan Administration cut magnetic fusion research PACER was long gone.

Another key thing about Pacer is you have to conserve the surface of the blast chamber so you keep the temperature lower than you’d like and the cavern bigger than optimum from a heat point of view because you don’t want to maximize radiation and wall contact with the fireball but rather minimize it. (AB-Matter of course would be immune to both. )

The chamber will move but needs to recover preshot position, the surrounding matrix (rock, salt, whatever) will take shocks but needs to stabilize and get sealed, you need to crank out many many bombs a year and they frankly cannot have much at all in common with regular nuclear bombs– no expensive detonators inside, ideally primaries fired externally say through gas gun electromagnetic implosion of a metal collar or other exotic means, all expensive parts of which are never in the blast chamber.

No bomb can be independently triggerable, the complex itself is the trigger so no stolen bomb risk. Etc.(note that the fissionables are by definition usable in weapons and they can NEVER leave the complex except downblended as U233 from Thorium 232 and mixed with U-238 down to reactor grade.

 If we produce Pu-239 it is supergrade, the best bomb material there is and so for the U-233 because of the clean D-D generation of neutrons without a lot of reactor dwell time in the messy fission environment that irradiates daughter isotopes over time.

 Here mother isotopes are irradiated directly quickly and finally.

The neutrons, the blasts, everything will eventually weaken the chamber. That blast chamber needs to be either regeneratable or end of life disposable all underground. You can’t have leaks that get through all layers of defenses, whatever they be, not in the pipes, not in the blast chamber.

 Yet you can’t spend too much on mega-engineering your way out of impossible situations or it becomes too expensive.

 I am pretty sure that Ralph Moir thought about it a bit and realized that solvable small problems  held more promise than huge profits divided by huge expenses.

But there are still big problems because of chemistry physics radiochemistry and all stops in between. All pieces of the former bombs will go into the chamber. Various working fluids are possible, some of which are optimized for tritium extraction (water on the other hand is designed for tritium entrainment as tritiated water) All kinds of isotopes will be produced although if well insulated from the chamber these can be a greatly reduced subset of the otherwise possible. The interaction of whatever working fluids are used with the pieces and the radiation will all generate complexity.

The huge and singular advantage of Pacer, of course, is that OFF means OFF. Not that all radiation is gone but that no runaway nuclear reaction is possible once you refuse to load another bomb.

There is no huge flammable inventory of graphite or hydrogen gas as in other reactor designs, or liquid sodium (unless you design it that way– never read of it in any of the PACER literature)

 From the point of last detonation all radioactivity begins to decay and however horrible a possible emergent situation it is (if you have sited correctly) already buried underground and under the water table. Rare indeed are the situations involving a Pacer unit that you could not simply turn your back on (physically, not necessarily financially).

 Here is Wiki data and pictures

Molten FLiBe flowing; this sample’s green tint is from dissolved uranium tetrafluoride.

The 2:1 mixture forms a stoichiometric compound, Li2BeF4, which has a melting point of 459 °C, a boiling point of 1430 °C, and a density of 1.94 g/cm3. Its heat capacity is 4540 kJ/m3, which is similar to that of water, more than four times that of sodium, and more than 200 times that of helium at typical reactor conditions

The low atomic weight of lithiumberyllium and to a lesser extent fluorine make FLiBe an effective neutron moderator. As natural lithium contains ~7.5% lithium-6, which tends to absorb neutrons producing alpha particles and tritium, nearly pure lithium-7is used to give the FLiBe a small cross section;[8] e.g. the MSRE secondary coolant was 99.993% lithium-7 FLiBe.[9]
Beryllium will occasionally disintegrate into two alpha particles and two neutrons when hit by a fast neutron.
In the liquid fluoride thorium reactor (LFTR) it serves as solvent for thefissile and fertile material fluoride salts, as well as moderator and coolant.
Some other designs (sometimes called molten-salt cooled reactors) use it as coolant, but have conventional solid nuclear fuel instead of dissolving it in the molten salt.
The liquid FLiBe salt was also proposed as a liquid blanket for tritium production and cooling in the compact tokamak reactor design by MIT
Ampoules of FLiBe with uranium-233 tetrafluoride: solidified chunks contrasted with the molten liquid.

Purified FLiBe. Originally ran in the secondary loop of the MSRE.
 there’s an opening for a new lithium-7 extraction process. However, any company attempting such a development will have to work under the watchful eye of DOE’s Y-12 group. WHAT DO WE KNOW ABOUT FLIBE? molten salt for reactor use

Ralph Moir Pacer Links —

Peaceful nuclear explosives to make electrical energy:
PACER Revisited

A study and project called Pacer suggested exploding a 20 kton nuclear explosive in a steam filled, earth walled cavity once every three hours
to produce 1000 MWe of power. In a series of papers this idea was revisited replacing the steam filled, earth walled cavity with a steel lined
underground cavity using molten salt droplets to cushion the effects of the explosive and absorb its energy. The yield chosen was typically
2 ktons once every 20 minutes to produce the same 1000 MWe of power. If such explosives could be initiated with 20 tons [84 GJ] of fission yield for a total of 2 ktons then the resulting power system would be 1% fission and 99% fusion.
• R.W. Moir. “PACER Revisited” Fusion Technology 15 (March 1989) 5 pages
• Call, Charles and R.W. Moir. “A Novel Fusion Power Concept Based on Molten-Salt Technology,” Nuclear Science & Engineering 104 (1990)
10 pages
• Szoke, Abraham and R.W. Moir. “A Practical Route to Fusion Power,” Technology Review (July 1991) 8 pages
• Szoke, Abraham and R.W. Moir. “A Realistic Gradual and Economical Approach to Fusion Power,” Fusion Technology 20 (December 1991)
10 pages
• Sahin, Sumer, R.W. Moir, Unalan, S. “Neutronic Investigation of a Power Plant Using Peaceful Nuclear Explosives,” Fusion Technology 26

Ralph Moir data on PACER from Neutronic analysis of a PACER reactor

The original PACER concept called for 20 kiloton charges in a 200 m diameter cavity under 200 atmospheres pressure of  500 C steam  generating 1000
MW electric with one bomb every 7 hours.

The 50 kiloton concept of 1975 was detailed in the previous article

Modified by Ralph Moir the concept became 2 kiloton charges each 40 minutes in a 20 meter radius cylindrical cavity engineered with a 1 cm thick stainless steel liner rock bolted or otherwise joined to the natural earth interface.

FLiBe is 2 Lithium 1 Beryllium 4 Fluorine with a density of .495 and jets of it, in the modified PACER architecture of Ralph Moir, shield the wall of the blast chamber. 

At the moment of blast, 25% of the volume in the ‘liquid zone’ is molten FLiBe jets and 75% void.
 They shield the 1 cm thick stainless steel blast chamber which is 30 meters in radius 60 in diameter and that chamber is rockbolted to the excavated wall.

  To achieve a tritium breeding ratio of 1.15 (a shot!) requires FLiBe thickness of 2.0 resulting in energy density of 19085 joules/gram.  

If over 2.5 meters thick FLiBe thickness after 30 years of 800 blasts a year the steel wall will be low enough activation (and the rock behind it) that USNRC rules of shallow burial apply.

This change was instrumental in reducing cavity volume by factor of 50 and peak pressure by a factor of 9.

Steam working fluid was replaced by molten salt jets absorbing energy and shocks, reducing pressure right after the explosion to 3 megapascals or so (the heat evaporates the liquid)

 Because tritium is nearly insoluble in the salt it reduces the tritium inventory by a factor of 10 e 5 from 10 million curies to 100 curies of tritium.

The FLiBe working temperature would be 500 C or 773 K

Fusion reaction products: 3.5 mev alpha and 14.1 mev neutron

40 cm thickness of rock behind the 1 cm metal experiences the heaviest irradiation

As viewed from the side 5 meters away from the bomb the pouring FLiBe zone begins and 30 meters away it ends so 25 meters thick of 2mm diameter shower sprays amid 75% void.

Each 2 kiloton shot is  neutron source with strength 2.95 x 10e24 neutrons per shot.

Lithium 6 burnup of 37 gram shot of 6 li in FLiBe natural lithium makeup of ~170 tons 12 tons lithium 6 content
315000 shots in plant lifetime 630000 kilotons (all nuclear tests ever equivalent)
Amazingly close– I wrote an article on that. The actual number of detonations is under 3000 so the first PACER team rapidly becomes world champions in about 4 years.
Apparently around 629 megatons. 452 megatons Soviet, 140 USA 7 megatons UK 10 megatons France and 20 megatons China. Plus under a megaton for India, Pakistan and North Korea and everyone else.
Berylllium burnup from neutrons 10 gram of Be per shot 3 tons total over a 30 yr USA beryllium resource ~150 kt (If you take total crustal abundance it is about the same as Uranium so say 40 trillion tons in the world)
One cool thing about this paper is the Pacer PNE reactor vessel at end of lifetime can be a disposal site for other nuclear waste so theoretically you have an end of cycle profit center

Ralph Moir data on PACER from

A PNE (Peaceful Nuclear Explosion–another name for PACER)  plant would produce a tenth as much waste as a conventional fission plant even if the explosions were all fission. over time 90% fusion should be possible.

DT is implied in the beginning at least. Lithium is only in supply for some centuries in a D-T powered world– deuterium is needed.  You really need to conserve lithium for the molten salt and not waste it on breeding tritium that you can get from the harder to fuse deuterium.  On the other hand everything has a learning curve.

Engineers might wrap the fissile material in a  cylindrical  jacket through which they would pass a large electric current squeezing the cylinder–rail guns and gas guns might also be used to keep bomb cost down and the charges more secure

In the early 1960s physicist Albert Latter, then of the Rand Corp, devised a scheme called PACER (but did not plan to reprocess the unburned fuel) and assumed Hiroshima sized explosions.

PNE power using 1  kiloton  explosives would be economically competitive at $1000 per explosive including both the nuclear explosive and processing the nuclear materials.

Oak Ridge NL estimate processing nuclear leftovers with molten salt system could cost as little as $10 per kg of recovered U which could translate to $100-500 per explosion.

Previous Project Plowshare estimates relied on custom rather than mass produced technology and were orders of magnitude more expensive than needed here.

To be economical the bombs have to be under $1000 each– that means radically different detonation and arming procedures–you can’t have expensive electronics aboard the bomb–
On the other hand 25,000 identical units a year would amortize costs and bring mass production economies and the explosives would not have to be packaged or guarded external to the site and Plowshare also assumed no recycling of fuel– none of that true here.

Test facilities could be made to withstand explosions of 30 to 300 tons as we engineer the actual unit.

If a leak did occur it would be well underground, small (the entire facility is small) detectable and contained and the worst conceivable accident is no more than 1 percent that from today’s fission plants since cleaning is done once a week vs once in 3 yrs.

  Not lowering the next explosive would shut the plant down; no runaway dynamic exists even in potential (the article mentions Chernobyl’s stock of hot graphite burned for days)

Some waste would accumulate on the walls of the cavity. Decommissioning might mean filling it.

Over the experience curve the amount of fission in each explosive should decline.

Any existing nuclear weapons state could operate a PACER without gaining any knowledge of new bomb tech. Non weapons states might have them run by contract by weapons states

Inspection is needed because of the potential of plutonium production even though Thorium U-233 is the logical cycle to choose because of the safety and down-blending potentials.

No ideal energy source for base-load electric power on Earth as opposed to space is in sight.  Conservation is not an energy source though it can moderate oncoming demand. Space based approaches are not build-able on demand without some new tech. PACER is one answer.

The explosives would not be self contained and transportable, thus not immediately usable as other than raw bomb material offsite.

Many of the technical ideas in this article come from the  High Energy Density Facility to study matter at high density pressure and temperature

20 minute intervals between explosion would produce 3000 megawatts thermal and 1000 megawatts electrical

Ralph Moir data on PACER from

The cost of reprocessing has been estimated to be $600 kg in aqueous solutions but $10 kg in molten salt
 Ralph Moir mentions here that one proposed way to store nuclear energy is to use the pulse to pump water uphill.
J. Pettibone A Novel Scheme For Making Cheap Electricity With Nuclear Energy UCID 18153 1979
Modified search name UCRL-JC–107068 PDF Idea is from 1979 PDF from 1991 UCRL-JC–107068
. In
1979 Joseph Pettibone conceived of a large water piston external engine driven by a nuclear
weapon-like release of nuclear energy. [29, 30] Although this engine had obvious proliferation
problems, it highlights the scientific feasibility and the economic advantages of eliminating
the nuclear steam supply
More on PACER by Ralph Moir–
If a fraction p is fissioned in the explosion the he amount of uranium used to run a 4 gw thermal reactor with a 1 day processing cycle is about 5/p kg  This has to be compared with the 4800 kg plutonium inventory of the Super Phenix breeder reactor (translation–with a breeder reactor you process infrequently with PACER you reprocess literally between shots so you only have a few kg in inventory instead of tons– amazing improvement in nuclear site safety)

Ralph Moir data on PACER from
Steam entrains tritium becomes radioactive
hard to separate tritium from steam

political and health hazards if vented tritium

Key differences to Ralph Moir version of Pacer
The Moir reboot of PACER changes chamber size from an unlined salt cavity, spherical shape 100 m radius to a steel lined cavity cylindrical shape 20 to 50 m radius 60 to 150 m height
The charge yield went down from 20 KT to 1 to 10 kilotons
ambient pressure 20 mpa (200 atmospheres) to .1 pa  (1 atmosphere)
equilibrium pressure after explosion from 26 mpa to 3 mpa
tritium inventory in cavity from over 10 million curies to 100 curies
fluid inventory from 330000 tons of steam to 1000 to 10000 tons of FliBe salt.

Line cavity with steel for predictable seal and properties under hundreds of thousands of explosions
use molten salt not steam for working fluid 70 pct energy then absorbed pressure contained reduced factor 3 or more
tritium produced insoluble in molten salt so can be pumped away purified
tritium inventory one hundred thousandth what was in case of accidental venting much reduced risk
smaller yield means reduced cost to cavity
Sterling  reentry test in the same cavity as Salmon proved decoupling worked–only nuclear tests in Mississippi –proved salt cavity practicality

This picture is from Wikipedia

the only nuclear weapons test detonations known to have been performed in the eastern United States.
Two underground detonations, a joint effort of the US Atomic Energy Commission and the US Department of Defense, took place under the designation of Project Dribble, part of a larger program known as Vela Uniform (aimed at assessing remote detonation detection capabilities). The first test, known as the Salmon Event, took place on October 22, 1964. It involved detonation of a 5.3 kiloton device at a depth of 2,700 feet (820 m). The second test, known as the Sterling Event, took place on December 3, 1966 and involved detonation of a 380 ton device suspended in the cavity left by the previous test. Further non-nuclear explosive tests were later conducted in the remaining cavity as part of the related Project Miracle Play.

no experience base of hundreds of thousands of explosions in salt caverns (but fatigue experience with steel)

steel liner stops impurities from earth contaminating molten salt
Euctectic FLiBe is 67.1 pct wt bef2 32.9 pct liF melting point 363 degrees Centigrade small amount ThF4 added to this
fuel charge surrounded beryllium thorium flouride FLiBe vaporizes all energy released U233 bred from Thorium and tritium
arbitrary height cylinder 4.67 radius
Vertical jet streams molten FLiBe around charge ~2mm diameter
in cylindrical cavity 20 m radius 4tj fuel charge exploded no evaporation pressure 10.4 mpa 90pct energy absolrbed vaporizationp 2.9 mpa 100pct absorbed vaporization p – 2.2 mpaAlthough D-D is assumed in the 50 KT pacer, tritium may be used up to 50 50 mix with D T in this model
D-T surrounded solid beryllium up to 20 cm thick absolrb muliplies neutrons from DT
breeding configvuration based on fission suppressed concept to maximize number of neutrons per unit energy

If DD used beryllium not necessary  D D reactions create even more excess neutrons per unit energy than DT reactions among beryllium.

 Beryllium metal is not soluble in FLiBe
Keeping flourine ratio high gives chance for beryllium to convert to BeF2 so can dissolve in FLIBE
Tritium has low solubility in FLiBe so pumpable out with cryopump few cubic meters after each explosion
at 37 kg FLiBe cost 19 million
amont FLiBeneeded absorb 4 tj without evaporation for temp rise form 400 to 1300 c
1 atmos vapor pressure is 2000 tons of FLiBe  salt
direct cost whole pacer unit must not exceed 2 billion

Ralph Moir data on PACER from Pacer Revisited– 2 kiloton charges in small engineered chamber
UCRL 98468 Rev 1
Older version of this at

 In the PACER concept, a 20-kT
peaceful nuclear explosion is contained in a cavity about 200
m in diameter, filled with 200 atm of 500C steam.

Energy from the explosion is used to produce power, and the neutrons
are used to produce materials such as
 U233, Pu, Co-60, and T.

The present idea is to modify the PACER concept in 3
ways to improve the practicality, predictability, and safety
of power production from this technologyst; improvements
are (1) line the cavity with steel; (2) replace the steam with
molten salt: and (3) reduce the explosive yield to about 2

… the only fusion power concept where the
underlying technology is proven and in hand today.

… Lining the cavity with steel makes it engineer able and
predictable, and prevents contamination of the working fluid…

The steam working fluid is replaced with molten  (FLiBe) in the form of droplets, to absorb energy and
suppress shocks

 This change results in an ambient pressure
below 1 atm soon after the explosion and allows much
of the energy to go into evaporation, thus reducing the pressure
in the cavity right after the explosion to about 3 MPa.
Also, because tritium is insoluble in the molten salt, it can
he removed almost completely, thus reducing the tritium inventory
by a factor of 100000 to 100 Ci using FLiBe. Then when the explosive yield is
reduced to 2 kT, the cavity volume is reduced by a factor of
50. which reduces the peak pressure in the cavity by a factor
of 9

In the modified concept, the cavity is tall and
cylindrical—rather than spherical-with a smaller radius of
curvature and a hemispherical roof (Fig. 1). As such, the
cavity should be much more durable.

The steel skin also must be corrosion resistant For example,
alloys high in nickel would be good. Haslelloy-n would be
excellent; and type 316 stainless steel may be adequate.

Pipes used to carry the molten salt to the droplet spray system  and to spray the walls will be made of the same material as the skin.

The system or pipes to carry the molten salt
from the cavity to the pumps and primary heat exchanger is
conventional, except the pipes must withstand pressures to
about 3 MPa at a pulse rale of about 1/hr

The cavity, its
liner, and the piping system must withstand about 200,000
shots over a 30-yr period

The walls of the lower part of the cavity will be cooled
actively so that a reasonable temperature ~650c is maintained
during the intershot lime of ~1 hour while the heat is
removed by circulating the molten salt

To keep the temperature
rise of the salt pool reasonable, a bed of balls in the salt
pool is assumed to store much of the beat. The balls could
be made of nickel-coaled iron.

One limiting
process is heat conduction from the surface of a droplet
to its  interior. 
this time is characteristic of the thermal diffusivity
k/pc, which for FLiBe is 1.7 times ten to the minus 7 meters squared per second where k the thermal conductivity is .8watt/m kelvin, p the density is 2050 kg m3 and c the heat capacity is 2350 J kg-1K-1 
thermal diffusivity time for a droplet of  1
mm in diameter, this time is 140 milliseconds 
The vapor rushing
past the droplets from the expanding fireball will distort
and break up droplets and cause internal circulation or  vortex motion so can enhance heat transfer
by  a factor of 2.7 and oscillations by a similar factor.
 The time to extinguish the fireball appears to be limited by conduction into the droplet rather than heat transfer within the gas or from the gas to the droplets or by condensation onto the droplets
After the heat is distributed over
2 kT of molten salt for each Kt of nuclear energy yield the pressure in the cavity will drop below 1 atm which corresponds to a temperature of below 1200 C an additional 5 KT of molten salt will bring the temperature down to 700 C. 
Before the next shot the tritium helium and other noncondensible gases must be pumped out and molten salt pumped through heat exchanger to lower its temperature and recharge the upper reservoir (for the shower spray system) shown in Figure 1

 The next charge is then lowered on a tether or dropped and the droplet spray system turned on to fill the cavity with the appropriate distribution of molten salt droplets.
The salt can be kept in a reduced state by continuously reacting it with metallic beryllium then the tritium will exist as T2 gas and can be removed by pumping. The uranium can also be removed by reacting with beryllium by fluorination or the salt can be fluorinated directly in a separate tank. The small amount of fissile material left would be very dilute in the huge amount of salt so criticality is prevented.
Since the excavated cavity will not confirm to the desired skin shape grout must be chosen to that it has low vapor pressure above 500 C. 
Ideally vapor pressure should be low in oxygen nitrogen silicon and other materials that would contaminate the molten salt if an inward leak occured (this would cause difficulties in reprocessing the molten salt)
Holes drilled for the rock bolts will form a collection system for pumping gaseous material in the region behind the metal skin to maintain the low gas phase pressure under 1 atmosphere.
Maintaining a low gas pressure in the cavity makes pumping tritium gas easy. The thermal design of the cavity skin is important. If salt is sprayed on the walls before each shot the salt-carrying debris from the nuclear device and its surrounding material will not be frozen directly onto the wall instead it will flow to the pool at the bottom of the cavity or freeze onto the existing frozen salt layer. Flowing fresh molten salt can clear or remelt this layer. 
Therefore the steel skin will remain at an ambient temperature below the melting point of the molten salt (363 C) except for a short time (under an hour) between each shot typically 1/hour. The frozen salt layer can then reduce the thermal stress on the wall. 

Contrast that with the 1975 design of PACER

A key question– how realistic  would it be to expect $1000 per unit small thermonuclear charges? Pure fusion of D-D would be nice, Deuterium-Tritium boosting fissionables would work, there might be a booting path of many architectures because you are debugging the charges as well as the first PACER but in the end the expectation is cheap Deuterium-Deuterium devices, as close to pure fusion as you can get.  This might in practice mean 200 tons of fission, 1800 tons of D-D- fusion. Moir himself has expressed the hope that  “If such explosives could be initiated with 20 tons [84 GJ] of fission yield for a total of 2 ktons then the resulting power system would be 1% fission and 99% fusion.”

Some references on small fission/fusion charge engineering–
discussion of 4th generation weapons FGNW…  …. nuclear shaped charges 15 kg of tritium in an arsenal equivalent to one million 1-ton-FGNWs,..

The physical principles of
thermonuclear explosives, inertial
confinement fusion, and the quest for
fourth generation nuclear weapons

Andre Gsponer and Jean-Pierre Hurni  

Note from Friedlander– This is a good as something Winterburg would write. Enjoy. 

“any country with access to tritium and high-power x-ray imaging technology could
easily develop and weaponize simple boosted fission explosives without nuclear
testing….with boosting — the problem of the preinitiation of the chain reaction, which creates difficulties in making a non-boosted fission bomb [66, 69], is no longer a serious problem….Boosting can also be used to make efficient and reliable fission weapons in which reactor grade plutonium is used instead of weapons grade plutonium…..It is therefore clear that ICF experiments will contribute very significantly to progress in weapons physics…A modern, sophisticated proliferator with access to ICF computer codes and today’s computer workstations would have far more tools for designing a secondary than the U.S., U.K. or USSR had in the 1950s or France and China in the 1960s…, in subcritical burn, the quality of the fissile material is of little importance: reactor-grade plutonium is just as good as weapons-grade plutonium….many technologically sophisticated countries (and, in particular, Germany, India, Israel, Japan, and Pakistan which have highly developed nuclear
infrastructures) are today in a good position to make not only atomic bombs but
also hydrogen bombs…currently preferred technique is to use magnetic compression to increase the
density of the fissile material (which may consist of low-quality, reactor-grade
plutonium) and a very small amount of antimatter to initiate the subcritical burn…..Fourth generation nuclear weapons based on such processes, and with yields of 1 to 10 tons equivalents of TNT, may weigh less than a few kilograms.”

Friedlander here to finish up: So we may end up with a underground antimatter triggered boosted fission burning D-D fusion molten salt reactor. That kind of crosses over many boundaries of what is a traditional reactor.
Now a general discussion on the future of PACER. 
Moir’s variant is far more engineerable and deployable. Not in every place but politics allowing (a big if) within say 5 to 10 kilometers of every place.  The 1975 scheme in the previous article was limited to a hundred or more sites near the US Gulf coast (or other locations in the world where salt domes are plentiful such as Iran).

But the salt domes were only chosen because they would allegedly be cheap. If engineering a big cavity got far cheaper anywhere PACER like units could be located anywhere. They would also be ideal waste disposal units at end of lifetime. But if we have the kind of free neutron factory that PACER promised to provide huge numbers of portable fission reactors could operate anywhere it was safe to operate one. It would enable a whole new energy economy, including say high temperature reactors for chemical processing and vast associated industries.Trading 252 kilos of  deuterium  at $500 a  kg worth $126,000 getting a million plus in power would pay if the processing was say only a couple of hundred thousand dollars more.   But with a self-booting new nuclear chemical economy the deuterium itself might get cheaper, just from economies of scale and lower power costs. Assured supplies for a lifetime of cheap power can settle peoples’ minds and get them producing wealth.
Contrast that with people fearful of peak everything and with constantly ramping up taxes because of a shrinking real economy.  (By which I mean the portion of the economy not created by finance games and government contracts of any kind but rather by industry and engineering selling honest goods at market prices)  Had the late 1970s seen a prototype PACER unit we might well now be living in a different world.  

Forty years ago, Project Pacer was– and remains– the most practical hope for short term D-D- fusion reactors that could be commercially deployed within a decade or even less.

Because of the decisions not made in 1975 the economy of 2015 is far smaller than it might be.  Notably China negotiated in an exemption for peaceful nuclear explosions in the test ban treaty. Perhaps there was a reason for that:  Someday PACER may return.

If you liked this article, please give it a quick review on ycombinator or StumbleUpon. Thanks