Misconceptions about fuel-coolant interactions (FCI) & FAQ
For a scientist, nothing is more annoying (besides, arguably, being completely ignored) than their work being wrongly or misleadingly cited, or being misunderstood. This page provides a reply to the most common misconceptions in volcanology about induced fuel-coolant interactions:
Misconception 1: "FCI is a process driven by rapidly expanding steam"
It cannot be emphasized enough: FCI is a thermohydraulic process. It is the increase of hydraulic pressure that drives the fragmentation in FCI explosions. By definition, "hydraulic" refers to the pressure exerted by a liquid, and not by steam (or any other gas)! If it was steam, one would have to describe it as a "thermopneumatic process". In a volcanological setting, it is the water staying in its liquid phase that acts as hydraulic wedge and transmission medium to convert heat into mechanical stress.
(On a side note, for MFCI, steam expansion does play a certain role in the latest stage of the MFCI process, i.e. AFTER the fragments already have been generated. Then, the overheated water turns into steam and releases additional kinetic energy that accelerates the previously generated fragments. But it is vital to remember that the key for ash generation is the fragmentation phase, and for fragmentation, steam plays no supporting role. Quite in the contrary, steam is a good heat insulator that impedes FCI processes, see below.)
Understanding that FCI is a HYDRAULIC process (driven by rapidly expanding LIQUID water) is absolutely crucial to understand FCI. People who do not get this point, inevitably set their trains of thoughts on the wrong track.
Why is this so important to internalize? For two vital reasons:
1.) Hydraulic pressure is a much more effective way to exert forces on a target material (in our case the magmatic melt), compared to pneumatic processes. Your car most certainly has hydraulic breaks and no pneumatic ones, and this is for good reason!
2.) FCI cannot be explained without high heat-transfer rates. Heat exchange rates between a hot solid and a liquid are MUCH higher than between a solid and gas (by the order of magnitudes, in fact). Steam would therefore rather impede the heat transfer and thus interupt the FCI-loop.
We can therefore note for our first key conclusion:
Key point I: FCI is a thermohydraulic process, where melt fragmentation is driven by rapidly expanding liquid water.
Misconception 2: "Micro-film boiling would occur during the fragmentation phase, stopping any FCI process"
As pointed out above, any steam layer between melt and water would indeed be a factor that impedes an efficient heat exchange. This steam film plays an important role in the pre-mixing phase of MFCI (and in other cases could lead to the prevention of FCI processes altogether), but once when an FCI process has been kickstarted, any kind of micro-film boiling within the expanding cracks seems extremely unlikely from a physical point of view, because of two reasons:
1.) The establishment of a steam layer (Leidenfrost effect) and micro-film boiling is dependent on the water pressure. The larger the ambient pressure, the thinner and less frequent the established micro-films.
Now let's picture the situation of the water/melt interface within a crack during the phase of expanding water. According to our FCI model, the built-up of hydraulic pressure is sufficiently large to drive the cracks and create shock waves. It is therefore physically implausible how a micro-film can be established under such high pressures in this phase.
Key point II: During the phase of thermohydraulic expansion, the water inside the cracks exerts high shockwave-generating compressional as well as tensional stresses onto the melt.
2.) Equally important is to keep the time scale in mind: the expansion of the water is at least as fast as the propagation of the cracks. And cracks in a failing material are superfast! My experiments with hammer impacts on glass have shown that we talk here about time scales of the order of 10-5 s - 10-4 s (10µs - 100µs). Within this short time span the thermohydraulic feedback loop is leading to fragmentation of the melt. Micro-film boiling takes place in the order of 10-3 s - 10-1 s (1ms - 100ms). In other words, the built-up of a vapor film is slower than the total fragmentation process.
Key point III: FCI fragmentation is a very fast process, occuring on time scales between 10-5 s and 10-4 s. This process is faster than the process of establishing vapour films.
We can therefore conclude:
Key point IV: The pressure conditions and time scales, under which the expansion and fragmentation phase of FCI occurs, prevents micro-film boiling.
Misconception 3: "FCI ignores/neglects/underestimates the importance of quench fragmentation"
Establishing a (relatively) new type of fragmentation mechanism does not mean that the role of other fragmentation mechanisms is questioned. In fact, quench fragmentation (aka "thermal granulation") processes play a crucial role in the generation of submarine volcanic particles. There is no doubt about that.
Sometimes, however, volcanologists act a bit like football club supporters, and stop being fully objective. While these human traits are understandable, they might result in discussions that are less than fruitful. It is like supporters of Real Madrid and FC Barcelona arguing if Lionel Messi or Cristiano Ronaldo is the "better" player, and any acknowledgement of the quality of the one player would be seen as a kind of insult to the other player's club. In order to avoid falling into a pit of fruitless battles, it is best to simply look at the physics, and compare without emotions how FCI differs from the 'classic' quench fragmentation. Once the differences are understood, this should give us a sufficiently objective basis to to discuss the potential role of both processes under different ambient conditions, and will help us to avoid any blind spots due to a "club supporter's" mentality.
How quench fragmentation works
To start with, let us have a look at how quench fragmentation (also known as "thermal granulation") works.
In simple terms (and ignoring more complex subtypes of quench fragmentation mechanisms like, e.g., the ones leading to limu o'Pele particles) this mechanism runs the following steps: when a hot melt, like a lava flow, encounters water, the rapid heat loss of the melt body's outer layer will result in the immediate formation of a rigid crust. This crust is characterized by a very low heat conductivity, rendering it into an efficient heat insulator. Consequently, the parts of the lava body that are located further inside remain hot for a much longer time and cool down with a decreasing rate. As an effect of the cooling, these inner layers start to contract, and mechanical (tensional) stress is building up between the rigid crust and the contracting interior. Once the increasing tensional stresses reach a critical point, they begin to drive cracks that eventually lead to the fragmentation of the melt's body's outer layer.
Here we see already one important fact about quench fragmentation processes that we should keep in mind:
Key point V: In contrast to FCI, quench fragmentation is the result of cracks exclusively driven by tensional stresses.
From the description above, it should also have become clear that the quench fragmentation is a relatively slow process: the inner parts of the melt's body first need to cool down before they can start to contract, which then leads to the increase of tensional stresses. To sufficiently cool down, however, the heat has to be transferred from the melt body's interior through its own crust, which features a very low heat conductivity. Since the heat transfer curve therefore shows an exponentional decline. Due to this delay between the moment of water contact and the moment of tensional stresses reaching a critical fragmentation threshold, quench fragmentation occurs on time scales between 10-1 s and 106 s.
Key point VI: Quench fragmentation is a relatively slow process, occuring on time scales between 10-1 s and 106 s.
(But be aware that the "slow" time scales refer to the built-up times of stresses, and not necessarily to the velocity of cracks!)
Comparison between quench fragmentation and FCI mechanism
As discussed in this blog before, fragmentation is a process of energy conversion from one form of energy into (fracture) surfaces. What is the energy source for quench fragmentation? It is the same source as for FCI: the melt's heat.
Key point VII: Heat is the energy source for both quench fragmentation and FCI.
In contrast to quench fragmentation, during FCI the heat is 'directly' converted by means of the expanding (liquid) water, leading to a self-accelerated and rapidly increasing hydraulic pressure acting on the melt. Remember: it is these hydraulic pressures that drive the cracks, with the water keeping pace with (and trying to outrun) the propagating crack tips. The melt hence experiences a combination of compressional and tensional stresses, as well as shock waves.
It should therefore become clear which of the two mechanisms is faster: If we count the time between triggering (MFCI: vapor film collapse, IFCI: inducing fragmentation) and the end of fragmentation, it is obvious from key points III and VI that FCI are much faster than quench fragmentation processes.
Key point VIII: FCI is a fragmentation mechanismat that is at least a thousand times faster than quench fragmentation.
This difference in time scales is super important when it comes to discuss the efficiency of the different fragmentation mechanisms (see below).
In the following table, the key differences between the two mechanisms are summarized. In addition, for the matter of completeness, I added the (more or less hypothetical) mechanism of a "purely" thermopneumatic process, where entrapped water is simply transformed into steam before any kind of fragmentation happens.
| MFCI | IFCI | “pure steam” expansion | Quench fragmentation | |
| energy source | heat from melt | heat from melt | heat from melt | heat from melt |
| necessary precondition(s) | entrapped water; synchronized collapse of vapor film |
primary fragmentation causing synchronized opening of cracks | entrapped water | none |
| stress inducing element | entrapped liquid water | liquid water inside opening cracks | expanding steam | contracting interior of melt |
| physical principle | thermohydraulic | thermohydraulic | thermopneumatic | quenching |
| self-accelerating feedback loop? | yes | yes | no | No |
| types of stresses that drives cracks | compressional & tensional stresses | compressional & tensional stresses | compressional & tensional stresses | tensional stresses |
| generation of shock waves? | yes (high intensity) | yes (high intensity) | yes (low intensity) | no |
| time scales of stress build-up | 10-5 s – 10-4 s | 10-5 s – 10-4 s | 10-3 s – 10-1 s | 10-1 s – 106 s |
| possible at depth of > 1000m a.s.l.? | no | yes | no | yes |
FAQ
Based on the considerations listed above, we can now answer the following frequently asked questions.
1.) Quench fragmentation as primary vs secondary fragmentation process: which one is more efficient?
Up to now, we have not distinguished between two different types of quench fragmentation: primary (fragmentation of a melt body, such as a lava flow) vs secondary (fragmentation of volcanic particles that were formed by a different primary fragmentation mechanism). I think this distinction is crucial.
Let's have a closer look on the key parameters that control the extent of quench fragmentation: Besides the thermal gradient and the structural integrity (fracture mechanical properties) of the melt material, it is the melt's volume that is most influential on the entire process: the larger the contracting volume, the larger the tensional stresses that can be built up.
Imagine a very small particle (thus small volume) that is brought into contact with water. It will cool down "almost immediately" without having the time to form a large thermal gradient between crust and inner core, therefore there will be no noteworthy stresses between these two parts of the material. Consequently, no quench fragmentation will coccur. This example demonstrates that there must be a minimum object size, a critical threhold in volume, that determines if enough contraction can occur to generate sufficiently large stresses that lead to the fragmentation of the object.
Based on pouring experiments with heated volcanic ash particles, I learned that this "threshold" grain size appears to be rather large, and that secondary quench fragmentation for volcanic particles smaller than 500 µm might be an overrated mechanism. This is, however, a preliminary finding, and further systematic studies on this subject with different melt materials still have to be conducted, before making more definite statements.
Suffice to say that from a purely physical point of view, there are very strong indications that quench fragmentation as primary fragmentation mechanism is much more efficient in producing fragments than as a mechanism of secondary fragmentation.
Key point IX: Quench fragmentation is a considerably more efficient mechanism as a primary than as a secondary fragmentation process.
2.) Why do particles resulting from FCI look different to quench fragmentation products?
The fracture behavior of any material will depend on the way how and how fast it is overloaded. In my new line of work in the ballistics lab, I am often confronted with damaged car or train windows, and the question if it was a firearm bullet that has hit it, or just a rock. You can tell the difference by looking at the point of impact, because a glass will break differently when being hit by a slow rock, compared to a faster bullet -even if the total fragmentation energy would be the same. It is not alone the energy [in J], it is the energy rate [in J/s] that controls how the cracks (and therefore also the fragments) are formed. Simply written, the higher the impact energy rate, the rougher the fracture surfaces and the larger the amount of finer grains.
On top of that, shock waves are a game changer when it comes to fracture mechanical properties of a material. In contrast to quench fragmentation, intense shockwaves are observed in FCI processes, leading to characteristic fracture surfaces that are often described as "stepped surfaces".
A third factor that determines how fractures propagate is the type of stress that occur: FCI leads to a combination of compressional and tensional stresses, whereas quench fragmentation is driven exclusively by tensional stresses (see key point V).
All these factors contribute to the observation that different fragmentation mechanisms lead to very different fracture surfaces, and therefore to particles of different shapes.
3.) Quench fragmentation vs FCI: which one is the more efficient fragmentation mechanism?
When defining "efficiency" of a fragmentation mechanism by how much fracture surface is produced per input energy (i.e., heat) and time unit, there is little to no doubt that FCI is by far more "efficient" than quench fragmentation: FCI produces larger amounts of fine grains with larger specific surfaces when compared to the products of quench fragmentation, and FCI takes place on a much shorter time scale.
Due to the significantly larger time scales during quench fragmentation, large portions of heat are simply dissipated into the surrounding water and therefore "lost" for fragmentation.
All glory for team Messi, then?
Well, no. The considerations above only account for "perfectly" working FCI processes, i.e. FCI processes that neatly fulfill the pre-conditions. Even in lab experiments under very controlled conditions only ~60% of the FCI runs are "fully successful". Small perturbances of the pre-conditions will result in reduced efficiency or even failure of FCI. For example, if the vapor film collapses too early in the pre-mixing phase of an MFCI the resulting explosion is nothing more than a weak puff.
Quench fragmentation, on the other hand is a mechanism that is much more common. Whenever FCI does not happen or "fails", inevitably quench fragmentation will take over. If we would define "efficiency" as the mass of fragments produced by a fragmentation mechnism over the overall volume of available lava produced in submarine eruptions, then (primary) quench fragmentation would take the gold medal from all that we know. It is the much more frequently encountered process!
It is therefore a matter of perspective, how efficiency is defined, and both mechanisms have their important role in submarine volcanism.
4.) Can particles be synchroneously produced by FCI and quench fragmentation in the same eruptive episode?
On first glance, FCI and quench fragmentation seem indeed to be "competing" processes. Both require heat as fuel, and the heat reservoir might be limited.
In the following, let us discuss some scenarios:
a.) Quench fragmentation first, then FCI
A "first quench, then IFCI" scenario, where quench fragmentation occur as a first step, before FCI kicks in, is plausible as long as this happens before the interior of the melt has been cooled down below a certain threshold. Such a sequence could work because the cooling process during quench fragmentation is much slower than during an FCI event, so enough heat might be available for a subsequent FCI episode over a certain period.
b.) FCI first, then quench fragmentation
Since FCI is so much faster than quench fragmentation, and both processes take their energy from the melt's heat, it seems implausible that quench fragmentation plays a major role for the parts of the melt after they have been subjected to FCI. Primary quench fragmentation simply will come too late, and it is questionable if sufficient heat is left for further quench fragmentation in these regions.
This applies even more for secondary fragmentation processes,where the particles generated by FCI will have already spent their energy (heat) to the thermohydraulic expansion leading to their generation.
c.) "Simultaneous" FCI & quenching events
We should keep in mind that such a dichotomy (between either FCI or quench fragmentation) does not apply to scenarios, where only a part of the melt is subjected to FCI, while other regions are subjected to quench fragmentation. These settings might apply for the majority of submarine FCI eruptions.
Due to the very different time scales the two fragmentations processes would not run "simultaneously" in strict sense, but at least it is fair to call their occurence "synchroneous", and we would expect to find both types of fragments in the deposits.