The propaganda does not work and neither do the explanations that degradation in support of fusion is related to external reasons, like excessive oil supply, budget deficit, etc., as we heard from our leaders during the previous Snowmass meeting. Soon after those explanations during the energy "crisis" in California and when the gas prices doubled in Illinois, we watched as no one recalled fusion as a potential energy source. Also the recent budget surplus has bypassed fusion.
New expectations are emerging inside the US fusion community in conjunction with a chance of rejoining the "small" ITER. The hopes are expressed that ignition achieved either with ITER or with a smaller ignition experiment will magically reposition the fusion in the eyes of the society and then give a new momentum to the program.
Still our hopes of a future success in fusion will remain futile until the reasons why fusion is not accepted as a potential energy source (and, in fact, has lost interest of its former supporters) are recognized and until the appropriate strategy is determined. ITER, a result of 15 year work by international team, serves as a good reference basis for the analysis.
TFTR, JET and JT-60 have culminated the plasma physics stage in fusion. Designed in late 70s - early 80s they provided the access to the fusion grade plasma and calibrated the plasma theories. Although a little bit short in the reactor physics achievements (the breakeven was not reached with a good margin) they (together with smaller experiments) gave a background physics for the future reactor development.
The next step in fusion is the first step into reactor physics. It has a different logic than plasma physics experiments. Costly and not flexible for experimentation, with activation of critical structural elements, the next step should be made within a transparent strategy leading to a competitive fusion reactor. We have no such strategy. All the proposed next step machines are based on ideas formulated in the early 80s and essentially mimic the big plasma physics experiments.
The focus of the reactor strategy is different. This strategy should take from plasma physics what is compatible with the reactor needs, with emphasis on efficiency, simplicity, low activation, and reliability in reactor control. The same strategy should have a transparent research path, when the problems are gradually solved along the way instead of piling them up at the bigger and bigger scales.
There are several layers of the reactor physics which should determine the strategy. Besides them there are deeper layers which specify the necessary physics and technology R&D path.
n T tE [m-3 keV sec] > 5 1021,
which should be fulfilled in both ignition and operation stages of the reactor, was always considered as a major milestone in the fusion research. In the conventional form the criterion looks 3 dimensional (in the parameter space n, T, tE), thus, suggesting many routes toward the fusion reactor (even in a specific area of the quasi-stationary magnetic fusion, i.e., tokamaks and stellarators). In particular, the former ITER was expected to pass the ignition margin in a long lasting discharge (1000 sec), while the present (smaller) version meets its 60 % (and with some luck 100 %) margin on a shorter time scale.
Everything looks OK at this level of consideration for both variations of this, still non-reactor machine. So, what was wrong 4 years ago when big ITER was rejected. Many explanations are circulating (most referring to an unfortunate political will), but none are statisfactory. In fact, the same ignition condition
B2 ß tE [Tesla2 sec] > 4, ß=(2/3) Eplasma/Emagnetic
written in terms of a dimensionless parameter ß (ratio of the thermal Eplasma to magnetic Emagnetic energy inside the plasma), the magnetic field B, and the energy confinement time tE explains the problem much better than the fusion science leaders do.
The factor B2 in it is limited at some realistic level of B = 5-6 T (5.3 T in ITER). The dimensionless ß factor, controlling the plasma stability, was studied intensively by theory and experiment and is exhausted at a very low level ß =0.02-0.04 when projected realistically to the reactor (0.03 in ITER). As a result, magnetic fusion, which already has a good confinement in tokamaks (2-3 orders of magnitude better than in other systems), was left relying solely on further increase in tE (about 4 sec in the small ITER). Thus, instead of 3-D parameter space, conventional magnetic fusion (regardless of the choice of tokamaks or stellarators) is being forced to follow the one lane road where everything depends essentially on one parameter.
This causes the problem even at the most fundamental level of the reactor physics. The total power P of the fusion reactor is related to the plasma energy content Eplasma as
P [GW] = 5Eplasma/tE [GJ/sec]
At a given power (which should be of the order of 2-3 GW in the reactor taking into account its complexity), an increase in tE leads to blowing up the energy content together with the size and the cost of the machine.
Unfavorable for the reactor strategy, these are unavoidable and fundamental tendencies in magnetic fusion. Being dependent on a single parameter, they require just one point for calibration. In 1998, ITER, the first self-consistent reactor oriented project, made the real numerical calibration of the cost of the next step in fusion. The result is well known. The $10 B big ITER (which was not a power reactor at all) died, at the same time, fulfilling its mission by proving that the dominant doctrine of magnetic fusion based on enhanced confinement is incompatible with real life. At such a high price tag, other difficult technologies are providing real, tangible service to society. Thus, spacecraft and jet bombers are flying, fission/non-fission reactors are producing the electricity, the submarines are hiding deeply in the ocean waters and supercomputers are crunching terabytes of information per second. Something, at least, resembling what was advertised by the propaganda would be expected from fusion by society.
The post ITER step on this road of low ß would be even more costly. Without being a ``rocket'' scientist, it is possible to understand that one of the reason of failure of the next step in magnetic fusion is its scientific strategy, focused on enhancement of the plasma core confinement and inconsistent with the most basic reactor physics.
Note, that the chasers of an enhanced tE for the reactor have been, in fact, doubly punished by plasma physics. First, the plasma temperature at the plasma edge in their "reactors" is low (relative to fusion scales) and, thus, only 1/4 of the plasma volume participates in fusion. Second, the same profiles lead to imminent turbulence and reduction of so-anticipated ``good'' confinement. So, even in pursuing tE the dominant doctrine in fusion uses not the best approach.
As a result, following the dominant line, the "small" ITER (with the size of what has to be a good fusion reactor) has acquired very modest "reactor" parameters, such as the total fusion power of 0.4 GW (instead of 2-3 GW) and the average wall load less than 1 MW/m2 (instead of 10 MW/m2 reasonable for a reactor).
It is not a failure of 15 year efforts by an experienced team. It is an intrinsic fundamental inconsistency in the present fusion strategy which targets essentially the fusion analog of the "steam engine". The "steam engine" may work (at the sub MW/m2 level of the power flux) but it will never "fly" (as a power reactor).
The major lesson of the big ITER was that it has shown the "dead end" sign across the road for the "steam engine" version of the magnetic (quasi-stationary) fusion. Being a conventional tokamak, and, thus, the simplest and cheapest device for its capacity, ITER has left no expectations to bypass the same calibration, not only for its "small" brother but also for other concepts (e.g., stellarators) if they would follow the low ß-high tE road. In this regard, it is necessary to think twice whether approval of the "small" ITER indeed leads to a contribution to the fusion reactor physics and technology.
Reliance on the increase in the energy confinement time has very important negative reactor control implications. The resulting increase in the total plasma energy, which is the major driver of plasma instabilities, leaves no realistic chance to control the reactor at the same level of ß as in experimental machines. Any abnormal event, which is tolerable in experimental devices (including ITER), in a reactor will trigger the expulsion of really big chunks of the plasma energy (portions of GJ) toward the plasma facing material surfaces. The following erosion, plasma contamination by impurities and plasma disruption would have unpredictable consequences for the operation of the reactor itself.
This scenario is not scalable even from a big ITER (were it built) to the power reactor, thus, requiring experimentation (simply unthinkable) with the plasma control and stability in the power fusion reactor. Such a steam engine "strategy" is evidently piling the problems at bigger scales instead of resolving them.
In contrast, the fusion reactor strategy, first of all, should target opening the high-ß dimension for the magnetic fusion reactor rather than ironing once more time (as it was reasonable at the plasma physics stage of research) a piece of the "dead end" road of good confinement.
Within the tokamak approach for reactor, a straightforward way of increasing ß is to involve the walls into plasma stabilization. In this regard, experiments on T-11M with lithium covered walls (inspired by TFTR results) as well as data on lithium surface properties of ALPS technology program show compatibility of the lithium covered walls with the high temperature plasma edge. Based on lithium potential in controlling the plasma edge, the LiWall concept has been recently developed. It suggests a radical way of increasing ß limits to the level of 15-20% (without sacrificing tE) even in tokamaks with a circular cross-section.
In a general sense, the ``first wall'' includes the plasma facing surfaces and the very next layer of material. In the reactor the first wall surface should withstand a high energy flux from the plasma and absorb 1/5 of the reactor power, while the next 10-15 cm's, absorbing most of remaining 4/5 of the power, suffer from losing mechanical properties under intense neutron radiation and produce radioactivity. At the same time this layer should breed tritium, transmit the neutron energy to the high-temperature cooler of the reactor, be transparent for the plasma control systems and should withstand possible abnormal thermal or electro-magnetic events due to plasma instability.
There is no strategy in magnetic fusion regarding the "first wall". The problem remains unresolved even at the conceptual level. The conventional approach just copies the first wall from the big physics experiments and relies on not yet determined high-Z materials, which are supposed to be studied in future (low fluence) prototypes of the reactor. Clearly, this approach works against fusion. In the neutron flux high-Z elements unavoidably produce the radioactive junk in what is advertised as a ``clean'' fusion reactor.
The ``divertor'' plasma geometry, adopted as the party line in magnetic fusion, contributed severely to the problems of the "first wall". Highly non-uniform power deposition on the plasma facing components has created unresolvable power extraction problem from the plasma. As a result, even if a miracle happens with the free boundary plasma stability limits, the divertor geometry with its low power extraction capabilities prohibits an increase in the ß-value. Ideas of cooling the divertor plates by the liquid metal or "waterfalls" of gallium, LiSn along the solid walls cannot correct the deficiency of the approach. The dominant doctrine, low efficiency and the high cost are really unseparatable.
Also, the high plasma edge temperature regime (obtained recently on DIII-D), excellent for confinement and stability (and consistent with the LiWall concept), was condemned as "reactor" irrelevant (in the ITER sense, which requires high density/low temperature at the plasma edge). A divertor in the reactor cannot tolerate the high temperature of the plasma edge.
As the result of a wrong physics approach to the "first wall", magnetic fusion is trapped in the low average wall loading. Its low neutron wall load expands the activation over a larger volume of high-Z materials, thus, making the fusion reactor even less ``clean'' than it would be with high-Z materials at high-ß.
Based on solid structural elements of the first wall and sensitive to thermal deformatons, conventional magnetic fusion requires a stationary plasma regime. Relying of high tech current drive and profile control, it is not only very expensive for tokamaks, but is just unrealistic for the plasma surrounded by solid walls in other systems.
It is impossible to make a finite list when there is so much wrong regarding the key problem of magnetic fusion. While there is nothing wrong in the tokamak plasma physics, it is evident that the simple mimicking of existing large tokamak designs does not work for reactor.
The reactor strategy should find a proper way to reconcile the tokamak physics with the first wall requirements. It clearly should go back to distributed power deposition over wall surface in order to enhance efficiency of the reactor. It also should rely preferably on low-Z materials and liquid elements in the first wall structure.
The LiWall concept did the reconciliation for the first time in fusion when the propulsion of intense lithium streams has been invented for tokamaks 3 years ago. Later, with invention of the dual Li streams/FLiBe blanket first wall structure by S.Zinkle and B.Nelson (ORNL) and of the "Yacht sail" approach, the first wall concept became compatible with the reactor physics. Based on low-Z materials and liquid first wall structure, LiWalls are the real step toward the clean and efficient fusion reactor, which is not sensitive to intensity of the high neutron flux (and, thus, does not require the stationary regime).
In fact, after 50 years of development, having no idea what kind of materials will be allowed in the reactor, this problem is postponed for an indefinite future when the necessary neutron fluence will be accumulated using of the next generation of the steam engine "fusion reactors".
The fusion reactor strategy should address the efficient power extraction from the neutron zone in clear way. This would be a real test of competitiveness of the fusion reactor and maturity of the strategy which determines the acceptance of fusion by the society.
While only vague speculations on the "cost" of electricity can be found in the conventional approach, the LiWall concept has found FLiBe as a perfect, high-temperature coolant for the fusion reactors, not activatable, not damagable and compatible with the active cooling of the plasma facing surface by intense lithium streams.
Not scalable from one power level to another, the stability control in the dominant doctrine (together with the divertor physics) suggests that even the power reactor has to be an experimental machine for plasma physics studies.
In fact, in the fusion reactor strategy the plasma physics concept should certainly resolve the plasma control problems at the sub $$ B level of experiments and, thus, make the reactor development step with sufficient stability margins which would guarantee the reactor oriented focus of the next step experiment.
In the LiWall concept the plasma is controlled by the lithium walls with excessive power and particle extraction capabilities. Also, LiWalls arrange the best environment for controlling plasma stability by combination of a conducting wall (situated right at the plasma boundary) and, if necessary, feedback system with executive elements (protected by the FLiBe layer) in the very proximity of the plasma edge.
Trapped in numerous limitations, full of illusions and hidden problems, this dominant doctrine, at the same time, is very aggressive in suppressing deviations from the party line. Its most prominent recent "achievement" is demolishing the $0.5 B (or so) TFTR facility, the major US investment into magnetic fusion. This happened on the eyes of the same community which were "seeking" new ways for fusion at the previous Snowmass meeting and now is talking about the next step fusion device.
Development of the LiWall concept during the 3 post ITER years made it clear that while the divertor machines are trapped on the same 1-D trail, TFTR was very consistent with the high-ß path in the reactor parameter space. In fact, its lithium conditioning experiments have contributed heavily into the LiWall concept which was completed in a very short time a year ago. Consistent with the reactor physics and basic plasma physics experiments and theory, it now demands the experimental testing and development of what appears as "unsubstantiated" claims to its opponents. In fact, there is a number of well identified issues which cannot be resolved without specially designed experiments.
Although many aspects of LiWalls are orthogonal to conventional fusion, the LiWall research path (up to ignition and the reactor regime) is very clear. It relies on the copper shell conventional tokamaks with the circular cross-section and lithium coating of the plasma facing surface. Contrary to a widespread misinterpretation, the flowing liquid lithium was never proposed for the research plasma physics machines. The flowing Li technology could be developed separately of tokamaks and then should merge with the tokamak plasma at the power reactor stage of development.
An appropriate experiment in the US, which would be able to go behind the T-11M experiments with lithium walls in Russia (already 4 years old), can demonstrate the potential of the new plasma regimes with high edge temperature and high-ß. This would start implementation of the new vision for the fusion reactor. The scientific fusion establishment has blocked any lithium wall experimental development which would require redirection of the DoE provided resources. This establishment still consider its false main stream doctrine as eligible for further consumption of funds, while leaving only residual funding for so-called "innovative" concepts. In particular, this time, it is difficult to see a reactor relevant logic be present in recommendations on joining the downsized ITER.
It is a matter of decisiveness of DoE administration to choose the road
and strategy. The present, the low ß
road, regardless of what vehicle is used (tokamaks or stellarators), leads
to the same "dead end" revealed by the big ITER. On the other hand, the
LiWalls concept is able to open an additional dimension in the reactor
parameter space in a way consistent with other aspects of the fusion reactor
physics. Starting with a modest, but appropriate scale experiments on TEXT
tokamak (presently shut down, but saved as a machine by University of Texas
at Austin) the LiWall program can generate a momentum for fusion. Then,
gradually learning how to control the plasma by lithium coated walls at
higher ß and on bigger size scales,
the LiWalls have the chance of advancing fusion to the ignition level
and to make it finally happen in a power reactor.