One critical area for many of the 4th generation reactor types we will analyse is the issue of our choice of materials. The decision on which material used in there construction could make or break many of these proposed reactors concepts. Several of these proposed reactors have operating temperatures in excess of +800 °C. Some, such as the LFTR would need critical parts to go even higher as much as +1,600 °C while exposed to corrosive liquids.
3.1 Stainless Steel and its alloys
Current nuclear reactor pressure vessels are typically forged from low alloy (ferritic) steel ingots, then lined with a (low cobalt) Stainless steel coating (for corrosion resistance purposes). Alternatively the pressure vessels are made entirely out of Cr-Ni-Mn Stainless Steel (that being Chromium for hardness and corrosion resistance, Nickel raises the operating temperature and Molybdenum improves high temperature strength). Such an arrangement typical yields a maximum operating temperature range of 450 – 600 °C. http://www.areva-np.com/common/liblocal/docs/Brochure/BROCHURE_EPR_US_2.pdf
The forged pressure vessel of a French made N4 reactor (1,500 MWe output)
Our Generation IV reactors almost all require much higher operating temperatures than this. To achieve this we’d likely need to use alternative materials. Top of the list would be “high temperature” stainless steels, notably blends of Cr-Ni-Mn-W (with “W” standing for Tungsten, which has the highest melting point of any metal) which would give us an operating temperature range of about 650 – 800 °C. http://www.smithmetal.com/downloads/stainless_datasheets.htm
3.2 Nickel Alloys
Alternatively we could opt for various high temperature Nickel alloys, in particular Nimonic, Inconel and Hastelloy so called “super” alloys. These are a series of high temperature, corrosion resistant Nickel alloys, typically containing various blends of elements such as Chromium, Silicon, Molybdenum, Tungsten and Titanium (amongst other ingredients). They’re principle disadvantage over stainless steels is higher costs, but they would allow us to push the reactors up to a 1,050 °C working temperature or possibly even a bit higher (my textbooks suggest 1,350 °C if we accept a reduced level of creep and stress resistance however the following link suggest we can push Hastalloy up to only 1,200 °C, neither of these two sources considers neutron fluxes though!).
However, one problem is that these “high temperature” alloys tend to be expensive to use, certainly much more expensive than the more “conventional” stainless steel alloys currently in use. And this is not just a factor of higher material costs, but the extra difficulty involved in making things out of them. For example, Inconel and Hastalloy have a reputation for being difficult to machine and are thus typically cast as whole components, then finished by (diamond tipped) grinding. Now, given that they are typically used these days to make relatively small high temperature parts (such as turbine blades) this isn’t really a big deal. Indeed engineering “best practice” for such parts is that they be cast.
However, casting an entire reactor Pressure Vessel (or Caladaria) out of such materials, while not impossible (quite large parts have been cast out of Nickel Alloys before, see the photo here) but certainly not easy, nor cheap. And just to complicate things, current engineering “best practice” would be that such parts (pressure vessels or heat exchangers) should be forged or machined, thought I would note that those high temperature Stainless steel alloy’s I mentioned earlier aren’t exactly easy to machine or forge either (as compared to low carbon steel anyway).
This inevitably means that any reactor made out of these materials will be much more expensive to build. Furthermore a good deal of that expense is down to the fact that so few facilities in the world can make such components (a sellers market) creating a potential production bottle neck. It is thus, difficult to see how you could therefore build such reactors quicker than existing reactors are constructed. Bearing in mind that one of the problems for nuclear is the slow rate of reactor construction, as compared to renewables or fossil fuels.
Another potential problem is that some nickel alloys (including Hastalloy) contain trace amounts of Cobalt, which under a neutron flux can produce the isotope Cobalt-60, which is a potent gamma emitter. This problem already means that efforts have been focused on eliminating “Stellite” (another nickel superalloy that for awhile was frequently used by the nuclear industry for high temperature parts) not increasing its usage. The EPR for example boasts of relatively little Stellite used, particularly in any parts exposed to a heavy neutron flux or likely to have high flow rates of coolant passing over them (again, see the EPR design brochure). Obviously, by using Alloys with lots of different ingredients, we risk hitting on similar problems, and our selection of such alloys will have to be very carefully undertaken, and thoroughly researched beforehand.
3.3 Materials, Temperature and Corrosion
I would note that while yes, if you fish out a text book it will tell you that the maximum working temperature of “conventional” stainless steel spec is up to 900 °C for Cr-Ni-Mn, or up to 1,300 °C for special high temperature (Cr-Ni-Mn-W) blends and higher still for the various nickel alloys mentioned. It would not, however, be recommended that you push a reactor’s working temperature that high. Really “bad things” will happen if you do that on a regular basis (once in a blue moon during an emergency, not a worry, but not 24/7 for 50 years!).
3.3.1 – Thermal Creep & Thermal Fatigue
There are several factors that limit the operating temperature of a power plant. Firstly the phenomenon known as thermal creep. This is a process under which materials can undergo deformation and eventually fracture upon exposure to relatively mild structure loads, at elevated temperatures over prolonged periods. To make matters worse, the neutron flux issue (see below) can accelerate thermal creep, making this a particular problem for high temperature reactors.
The failure via thermal creep of a steam pipe in a geothermal power plant
Another problem, is fatigue, which occurs where repeated cyclic loads over time causing crack propagation leading ultimately to failure of the component, often due to loading well below its maximum yield strength. While typically associated with structural loads (pressurisation of an aircraft fuselage during climb to cruising altitude for example), “thermal fatigue” can occur when a component is exposed to large cyclic temperature fluctuations.
It is important to note that the failure mechanisms for fatigue and creep are very different, but as far as the current discussion is concerned, many of the solutions to the two problems are relatively similar. Firstly, keep the operating temperature of the reactor below some minimum threshold level (again how low depends on the material, and for our reactors the neutron flux intensity). The occasional drift of operating temperature above this threshold (as would likely happen in an emergency situation) is not cause for alarm, so long as it doesn’t happen on a regular basis. Secondly, avoid exposing the material to a high rate of thermal cycles , i.e not having it constantly heating up rapidly and then cooling down again, try to keep it at more or less the same temperature constantly (save of course periods when the reactors being shut down for maintenance). Finally, any parts vulnerable to creep or fatigue should be subject to regular inspection (not easy if a part is buried at the bottom of a nuclear reactor!) to identify the tell-tale signs of impending failure.
3.3.2 – Neutron Fluxes
The third problem is the radiation that the components of the reactor are exposed to. This radiation exposure can gradually over time damage the material (by radioactive particles essentially “bumping” into atoms within the alloy and gradually changing its composition) this is a particular problem in a situation where there is a high neutron flux, as would be the case in any “fast” reactor or a fusion reactor. The cumulative effect of this radiation exposure is to degrade the physical properties of the material, effectively lowering the safe working temperature of our material to varying degrees. Also, as noted earlier, prolonged neutron exposure can cause some materials to mutate into another radioactive form, increasing the amount of high level waste the reactor ultimately generates when it is decommissioned.
3.3.3 – Phase changes
A fourth problem we need to worry about is the fact that alloy metals often have very different structural properties at different temperatures. For example, the ferrite (i.e iron with low carbon solution) in steel assumes a body centre cubic atomic formation up to a temperature of around 700-800 °C upon which it swaps over to a face centre cubic molecular arrangement. This naturally produces very different physical properties. Worse still, cooling the steel back down again below this temperature can often lead to a very different atomic structure being adopted (rapid cooling, after exposure to the critical temperature above, tends to result in a hard brittle steel, while slow cooling produces a soft ductile steel). It is therefore important not to exceed certain temperature limits with many alloys or you may find that you’ve basically “written off” the reactor.
Phase diagram for Stainless Steel
3.3.4 – Inter-granular corrosion and cracking
A related problem to the phase chance issue raised above is the danger of inter-granular corrosion and cracking. For example Chromium (a critical alloy component in both stainless steels as well as those Nickel alloys mentioned earlier) has a problem where at elevated temperatures it can precipitate out of solution, forming into regions rich in chromium, but leaving the surrounding areas devoid of chromium and thus vulnerable to corrosion (as usually the point of the Chromium being added to our alloy is to prevent corrosion).
This phenomenon is usually solved by adding in a small quantity of Molybdenum which serves to “lock-in” the Chromium, although with nuclear reactors the problem becomes that this neutron flux can gradually erode the Mn additives ability to correct this problem.
The point I’m making here is that many materials have very different properties at different temperatures. Thus, it’s important not to exceed said operating parameters on a regular basis, else you’ll quickly invalidate you’re reactor’s warranty. It also means material selection is a critical stage in the design of any plant.
3.4 Turbine issues
I would also note that it’s not just the reactor pressure vessel and all the components inside this that we need to worry about, but also the turbogenerator set as well (for anyone who I’ve lost, that’s the big spiny thing that makes the electricity!). Making a reactor pressure vessel, particularly one designed such as to have as few moving parts as possible, out of Nimonic HR4 or even Titanium (if we really want to push the boat out!), expensive granted, but certainly doable. However, a turbine set, something that will inevitably involve lots of moving parts, different kettle of fish! It could well cost as much money to build such a system, especially a Brayton cycle turbine, out of these exotic alloys, as it costs to build the reactor itself!
The turbogenerator set at a typical modern nuclear station, in this example, steam driven with hydrogen cooled bearings
To make matters worse there is the problem of Hydrogen embrittlement to worry about, something that a steam turbine cycle running at elevated temperatures is potentially prone too. There are two possible solutions, firstly prevent exposure to hydrogen. This can be done by using an inert gas such as helium as our working fluid (and by using nitrogen or helium cooled bearings). However helium has both a lower density and lower specific heat capacity than water. These means either increasing the size of your turbine set (bigger turbine = higher costs), or by raising the working pressure (thicker material needed, which complicates manufacturing and thus raises costs). The other option is to use other alloys, such as austenitic stainless steels or vanadium/chromium alloy steels, which are less vulnerable to hydrogen embrittlement, and construct the turbine out of this. But these materials have lower working temperatures, typically less than 600 °C, and this would result in a lower thermal efficiency (if used), amongst other issues.
Also, a key element of the proposal for many so called generation IV reactors is to use the Brayton cycle rather than the Rankine cycle. There is good reason to do this, as the Brayton cycle has a much higher level of thermal efficiency (typically 45-60%, but up to 65% possible) compared to the Rankine cycle (typically 33-40%, thought multi stage units have recently pushed it up to 47%). However, given the materials we are talking about using, a Brayton cycle turbine would in certain situations work out to be more expensive to build than a Rankine cycle set. While my preference as an engineer would be for a Brayton cycle, my suspicion is that when presented with the options of a Rankine cycle turbine set costing X amount and a Brayton cycle turbine that is 10-20% more efficient, but costs 1.5X, the suits will plumb for the Rankine cycle turbines. Of course, this won’t always be the case, but it’s important not to get fixated on the Brayton cycle.
Brayton cycle against Rankine cycle
A similar problem exists for the idea of making hydrogen directly out of high temperature reactors using the Sulfur-Iodine process. The high costs of such hardware (all of which again will have to be made out of high temperature alloys, where again Hydrogen embrittlement will bedevil us) may result in the whole idea being dropped in favour of more conventional electricity generation (and if we really want hydrogen that badly then use electrolysis and accept a lower rate of thermal efficiency) or co-generation (using waste heat for industrial heating purposes) once hard headed economics set in. So again, its important not to get fixated with this idea.
3.5 Alternatives to Steel or Nickel Alloys
Of course given the problems mentioned above with regard to Nickel and stainless steel alloys, one might question whether we’d be better off using some alternative materials. Indeed for some of the reactor designs we’ll be reviewing, such as the Gas-cooled fast Reactor (with its high neutron flux), the LFTR (high operating temperatures) and the Fusion reactor (both of these problems!) this could well be a necessity as it might be frankly impossible to build such reactors out of the alloys discussed, at least a reactor with a decent operating life and a good margin of safety.
3.5.1 – Titanium and its Alloys
Titanium alloys are one option. Titanium has a very high strength to weight ratio, as well as good temperature, creep and corrosion resistance. However, it has two main draw backs. Firstly the maximum operating temperature of titanium alloys isn’t that much higher than those for the Nickel alloys already discussed (indeed in many cases its lower!). The key reason why Titanium is frequently used in preference to Nickel in the aviation industry is its superior strength to weight ratio. So long as were not planning to fly our reactors around anywhere then this property doesn’t really make the extra expensive and difficulty of Titanium use worthwhile.
Secondly, Titanium in its liquid form reacts with pretty much everything it touches, even air! Thus it’s necessary to process it completely in a vacuum (see Kroll process). To be blunt, this sort of makes working with Titanium a complete pain in the as$ and not something you want to go through unless you have to. One point to note though is that certain Titanium alloys allow one to get around the issue of “active ingredients” discussed as regards Nickel alloys turning into their “evil twin” variants after exposure to radiation, although that entirely depends what you’re using them for!
3.5.2 – Tungsten and Refractory metals
Other options include so called Refractory metals, chief among these Tungsten and its alloys. Indeed tungsten is a key ingredient of the high temperature stainless steels and Nickel “super-alloys” mentioned earlier. By opting for pure tungsten we take advantage of its incredible heat resistance – an ultimate melting point of 3,410 °C (give or take) against a measly 1725°C for Titanium and a puny 1458°C for Nickel. Tungsten has excellent creep resistance and, better still, good resistance to high neutron fluxes.
But, as always there are drawbacks. Firstly, tungsten is fairly brittle (at least as far as metals go) and prone to shock load failure. Using it on its own would thus be problematic and require careful design. With the exception of a few relatively small parts, nose cones of reentry vehicles and the exhaust nozzles of rocket engines, we have very little experience at building large components (such as a reactor pressure vessel) out of pure Tungsten (or its alloys). Such manufacturing techniques will have to be invented and the costs of said research charged against the reactor development costs.
Indeed it is this not so small matter of cost that is the problem. Tungsten typically trades at a price twice that of Titanium, 4 times higher than Nickel and a good 10-20 times higher than steel. Building an entire reactor out of it would likely be very expensive, probably to the point that it simply wouldn’t be economically viable. This is why we typically use Tungsten as an additive to other cheaper materials, reducing costs. It is probable thus, as a work around, you could use pure Tungsten (or a tungsten alloy) to make key critical components in your reactor and other “cheaper” (relatively speaking!) materials for everything else. However, if lower costs and ease of construction are among you’re goals, I’d stay away from Tungsten.
Another two refractory metals worth mentioning are Tantalum and Niobium (often called Columbium in the US). Both materials have high melting points, good creep resistance and in the case of Tantalum, excellent resistance to corrosion. Unusually for Refractories they are also both reasonably ductile. However, high costs can be an issue with both materials (Niobium in particular), hence they are frequently used as alloying elements, or applied as a coating onto a base of some other material. While there is some experience with using Niobium in nuclear reactors (it was used as a fuel cladding material in the Dounreay reactor) there is little such experience with Tantalum.
3.5.3 –Refractory materials and ceramics
Another option is to use non metallic materials, notably so-called “refractory” materials, which is a catch all phrase referring to a variety of materials ranging from ceramics, graphite and high temperature concretes (plus metals as already discussed).
In General terms, refractory materials have the advantage of extremely high strengths (particularly in compression), good wear, corrosion and high temperature resistance (as in working temperatures in the order of 1,200 – 1,800 °C). Some refectories, such as Silicon Carbide, Boron Nitride or Silicon Nitride are practically indestructible, i.e tensile strengths in the order of 500 MN/m2 (with many times greater than this possible in compression) even at working temperatures in the order of 1,200 °C.
So you might say, well there’s our solution, build our reactor out of an “indestructible” material. If only life was that easy! Unfortunately Refectories come with a number of severe drawbacks. To be blunt, they’re just a total ba$tard to work with!
Two quick case studies, firstly the UK AGR program which used a high temperature blend of reinforced concrete to make the pressure vessels. It was the difficulty of making the pressure vessels out of this material (and concrete is pretty much the “easiest” of “refractory” materials to work with) that led to the AGR’s being late and massively over budget. Dungeness famously being 13 years late and costing 4 times the original (inflation adjusted) cost estimate (10 times more if we neglect inflation). Worse still, problems with cracking in the graphite core (most likely as a consequence of neutron bombardment) has resulted in several AGR’s having to operate at lower working temperatures and some may be forced to close down early as a result.
The NASA space shuttle also used Ceramics, principally Boron Silicon coated tiles underneath and RCC panels (Reinforced Carbon-Carbon) on the leading edges. It was the escalating costs of maintaining this TPS (thermal protection system) after each flight that prevented the shuttle from becoming the cheap space truck NASA wanted. Indeed it would end up costing two to four times more to launch (per kg to orbit) than expendable alternatives. The Colombia shuttle disaster also highlighted another problem with ceramic materials (and refectories in general) – they are prone to brittle fracture. That is, they can unexpectedly fail and crack as a result of relatively mild shock loads (either thermal or structural).
I remember doing some experiments once on ceramics (notably Silicon Nitrite) to test the tensile strength of some small samples of them. The action of merely pushing the samples into the die for testing could sometimes break them (and I’m not that strong a person!). Refractories, I concluded, can sometime break for no reason other than the fact that you looked at them the wrong way! I was also glad that I was only doing the testing and not the machining of them!
A NASA technician performs maintenance on the Space Shuttle’s extensive network of 8,000 ceramic tiles
I would note however, that Ceramic manufacturing technology is coming along in leaps and bounds and such materials are working there way into an increasing number of uses. In the future making things out of Ceramic may not be as bleak an option as I propose.
But for the moment, I won’t recomend the use of ceramics of any kind to build a nuclear reactor, not unless there was no other alterative. Anyone advocating we do otherwise, would be best advised to start an extensive (and expensive!) program of research into them before you even commence design of the plant, let alone construction. Else, like the AGR’s, you’ll likely wind up with a reactor that’s drastically over budget and late. This is especially true if, as noted, we’re forced to build the turbine kit out of ceramic materials (or Tungsten) as well, which would naturally be a lot more complex than building it out of the alloys mentioned earlier.
So before we even begin our evaluation, we have to conclude that a big stumbling block to several of the proposed reactor designs is this issue of materials choice. It’s certainly not a “show stopper” but it does “complicate things”, and that potentially means higher costs and slower build rates. There are, however, various work-a-round’s for some of the proposed reactors and I’ll discuss those as we go along.