The subsection we focus on here, “Adsorption processes in clays”,
contains very little descriptions of fundamental properties of
bentonite, and is instead almost exclusively devoted to detailed
discussions on various models. As an example, already in the
first paragraph the text digresses into dealing with the problem of
defining “surface species activity” in the “DDL”2 model…
TS15 discuss adsorption separately on “outer basal surfaces”, “interlayer basal surfaces”, and “edge surfaces”. Note that the distinction between “outer” and “interlayer” basal surfaces requires that we view the compacted bentonite as composed of stacks (referred to as “particles” in TS15). But this idea is just fantasy, as we have discussed in the previous part and in a separate blog post. Moreover, central to the description of adsorption processes in TS15 is the idea of a Stern layer. This concept was briefly introduced in the previous subsection (“Electrostatic properties, high surface area, and anion exclusion”)
The [electrical double layer] can be conceptually subdivided into a Stern layer containing inner- and outer-sphere surface complexes […] and a diffuse layer (DL) containing ions that interact with the surface through long-range electrostatics […].
The next time this concept is brought up is at the beginning of the
discussion on adsorption on “outer basal surfaces”
The high specific basal surface area and their electrostatic properties give rise to adsorption processes in the diffuse layer, but also in the Stern layer.
I have written a separate blog post arguing for that the idea of Stern layers on montmorillonite basal surfaces is unjustified. Note that the notion of Stern layers on montmorillonite basal surfaces in the contemporary bentonite literature de facto means that these surfaces are supposed to be full-fledged chemical systems. In particular, the basal surface is supposed to contain localized “sites” that interact generally with ions to form surface complexes and that can involve covalent bonding.
Note further that the Stern layer was originally introduced as a model (or a model component) that extends the Gouy-Chapman description of the electric double layer. TS15, on the other hand, use the term “Stern layer” to refer to an actual physical structural component. And just as in the case of several other “components” that has been introduced in the article (“particles”, “inter-particle water”, “free or bulk water”, “aggregates”…), the existence of a Stern layer is just declared rather than argued for. And just like with the other components, these are not universally adopted. I don’t think it is appropriate to include Stern layers in this way in a review article when established parts of the colloid science community refer to them as an “intellectual cul de sac”.
So in order to even begin to criticize what TS15 actually write about adsorption processes here, one has to accept both the flawed idea of stacks as fundamental structural units and the far from universally accepted idea of Stern layers on montmorillonite basal surfaces. I will therefore refrain from doing that, and simply proclaim that I don’t accept the premises. (I believe I will have reasons to return to the models presented here when reviewing later sections of TS15.)
Additional remarks
But I think it is worth reminding ourselves that at the end of the previous section (covered in part I) we were promised that this section should qualitatively link “fundamental properties of the clay minerals” to the diffusional behavior of compacted bentonite. A reader of TS15 will thus expect this section to contain, in particular, a reasonable description and discussion on how compacted montmorillonite works. Instead a very specific (and flawed) model is imposed on the reader: the first subsection (covered in part II), introduced the fictional stack concept, and gave a confused and irrelevant explanation of anion exclusion; the presently discussed subsection is centered around Stern layers.
If the authors truly did what they claimed, in this section they should have addressed the consequences of montmorillonite TOT-layers being charged — a universally accepted fact — without introducing further assumptions. This would naturally lead to a discussion on osmosis, swelling, swelling pressure and semi-permeable boundary conditions (all simple empirical facts). These topics, in turn, should lead to considerations of e.g. ionmobility and chemical interface equilibrium. Not a single one of these topics are, in any meaningful sense, actually addressed in this section.
Before ending this part of the review, I also would like to focus on what is being said about “interlayers”. We should keep in mind that TS15 — together with a large part of the contemporary bentonite research community — assume “interlayers” to be something different than simply the space between adjacent basal surfaces: these are supposed to be internal to the fantasy construct of a stack. When discussing adsorption in these presumed compartments they write
The interlayer space can be seen as an extreme case where the
diffuse layer vanishes leaving only the Stern layer of the adjacent
basal surfaces.
Of everything I’ve read in the bentonite literature, this is the closest I’ve come to see some actual description of what the fundamental difference between an “outer basal surface” and an “interlayer” is supposed to be. But let’s think this through. TS15 have claimed that an electric double layer is composed of a Stern layer and a diffuse layer, and we have vaugley been told that ions in the Stern layer are immobile. The above quotation thus implicitly says that that “interlayer” ions are not mobile, and that diffuse layers are only supposed to exist on “outer basal surfaces” (which, remember, is a fantasy component). But — disregarding that the stack-internal “interlayer” also is a fantasy concept — it is an indisputable experimental fact that has been known for a longtime that interlayers provide the only relevant transport mechanism in compacted bentonite.
Thus, either TS15 here provide us with yet another incorrect description of the behavior of compacted bentonite (that “interlayer” ions are immobile) or they are claiming, somewhat contradictorily, that Stern layer ions are mobile after all. But if Stern layer ions diffuse, such a structural component could reasonably not have been singled out in the first place! (The diffuse layer is supposed to have “vanished”.) As with many other issues in TS15, this question is left vague and unanswered.3 The continuation of the text does not make things clearer
For this reason, the interlayer space is often considered to be
completely free of anions (Tournassat and Appelo 2011), although
this hypothesis is still controversial (Rotenberg et al. 2007c;
Birgersson and Karnland 2009).
An interlayer completely devoid of anions certainly play by other rules than an “ordinary” electric double layer. Does this mean that TS15 assume “interlayer” ions to be immobile?4 Anyway, it is an indisputableexperimental fact that anions occupy interlayers, and I find it quite bizarre to find myself referenced in connection with the “controversial hypothesis”. The idea of compartments completely devoid of anions is widespread in the contemporary bentonite research community, but no one has ever suggested a mechanism for how such an exclusion is supposed to work; here, it apparently should be related to “Stern layers” in some (unexplained) manner. At the same time, the simplest application of Donnan equilibrium principally explains e.g. the behavior of the steady-state flux in anion tracer through-diffusion tests.
The agreement between [Poisson-Boltzmann] calculations and MD
simulation predictions was somewhat worse in the case of the
\(\mathrm{Cl^-}\) concentration profiles than in the case of the
\(\mathrm{Na^+}\) profiles (Figure 3), perhaps reflecting the poorer
statistics for interlayer Cl concentrations or the influence of
short-range ion-ion interactions (and possibly ion- water
interactions, as noted above) that are not accounted for in the
[Poisson-Boltzmann] equation. Nevertheless, reasonable quantitative
agreement was found (Table 2).
Here they acknowledge not only that anions do occupy interlayers, but also that the interlayer plays by the same rules as the “ordinary” electric double layer (“Poisson-Boltzmann calculations”). What happened to the “vanishing” diffuse layer, and to “considering” the interlayer to be “completely free of anions”? I find it quite outrageous that they fail to acknowledge these blatantly mixed messages with so much as a single word.
Update (251106): Part IV of this review is found here.
Footnotes
[1] As I have commented in the
earlier parts: TS15 are fond of using the very general terms
“clays” and “clay minerals”, while it is clear that the
publication mainly focus on systems with substantial ion exchange
capacity and swelling properties. Here we will continue to use the
term “bentonite” for these systems, and ignore the frequent
references in TS15 to more general terms.
[2] For some
reason, “DDL” is short for (the very generically sounding) “double
layer model”. Why not “DLM”?
[3] Spoiler: in later sections describing models, TS15 allow for the possibility of transport in “interlayers”.
[4] Questions like these can often not be answered because so many statements in TS15 are vague and ambiguous. In this discussion we have to refer to statements such as (my emphasis)
“The EDL can be subdivided into a Stern layer […] and a diffuse layer […].”
“The interlayer can be seen as an extreme case where the diffuse layer vanishes […]”
“The interlayer space is often considered to be completely free of anions […]”
I get annoyed by too much of such language in scientific
publications.
I argue that the only significant pore type in water saturated compacted bentonite is interlayers, by which I mean pores where the exchangeable cations reside (together with any other dissolved species). From this perspective it naturally follows that a homogeneous view is a suitable starting point for modeling compacted bentonite. I have presented, used, and discussed the homogeneous mixture model in many places on the blog, the main sources being
For reasons I can’t get my head around, a homogeneous view of
compacted bentonite is not the mainstream in contemporary
bentonite research. Instead we are stuck with
“the mainstream view”, which postulates several distinctly
different pore structures within the bentonite; in particular, the
mainstream view uses a bulk water phase as a starting point and also
distinguishes between “outer” and “inner” basal surfaces. Electric
double layers are assumed to only exist on “outer” surfaces, while
the function of the “inner” basal surfaces is mostly shrouded in
mystery.
On the blog I have also presented plenty of experimental support for a
homogeneous view. A main argument is that the conditions for swelling
pressure — the most profound feature of bentonite in equilibrium with an external solution — are essentially fulfilled automatically in the
homogeneous mixture model. The mainstream view, in contrast, requires
handling of the seemingly contradictory situation of having swelling
pressure while the water chemical potential is supposedly restored
without pressurization. Proponents of the mainstream view often deal
with this by simply ignoring swelling phenomena altogether.
I have also on the blog dissected several studies that argue for a
non-homogeneous view, but that actually provide evidence for the
opposite when examined more carefully. Consider in particular:
By systematically varying background concentration, material, and diffusing tracer, Glaus et al. (2007) clearly demonstrate, not only that the exchangeable cations are mobile, but that they dominate the flux in through-diffusion tests in highly compacted montmorillonite. While this certainly is an argument for that compacted bentonite is homogeneously structured, Glaus et al. (2007) still analyze their results from the perspective of the mainstream view, and do not — in my view — fully conclude what their results imply.
In particular they postulate the presence of an interlayer domain and a “free pore water” domain, and write for the “total” flux1 (their eq. 3)
where \(J_\mathrm{il}\) is a presumed diffusive flux in the interlayer domain and \(J_\mathrm{pw}\) is the presumed diffusive flux in the “free pore water” domain.
Their subsequent analysis shows that the measured flux in montmorillonite scales as
where \(C_\mathrm{bkg.}\) is the concentration of the background electrolyte (NaClO4), and \(Z\) is the charge number of the diffusing tracer (\(Z = 1\) for sodium and \(Z=2\) for strontium). Moreover, by considering ion exchange equilibrium, Glaus et al. (2007) show that also \(J_\mathrm{il}\) is expected to scale according to eq. 2. As they also confirm that this scaling behavior is not observed in systems without interlayer pores (kaolinite), they could have confidently concluded that their results imply that interlayers are the only significant pore structure in montmorillonite at these densities (as the title suggests).
Unfortunately, the discussion part of the article is considerably more tentative, focusing mainly on “interpretations” of the resulting flux
The present work shows that the interpretation of cation diffusion experiments in highly compacted swelling clays in terms of the concentration gradient in the aqueous phase may result in a nonsensical dependence of the effective diffusion coefficients on the salt concentration in the external aqueous phase. An alternative interpretation using an effective diffusion coefficient in the interlayer water (\(D_\mathrm{il}\)), being independent of the external salt concentration, with a corresponding concentration gradient in the interlayer water is more consistent with the experimental observations.
and the article ends on a quite apologetic note
The proposed interpretation should in turn not be blindly applied to
other experimental conditions. Diffusion of cations via the free
pore water may become increasingly important in swelling clays with
lower degrees of compaction or in clays in which the interlayer gel
pores are not that adjacent as they are in compacted
montmorillonite. In such cases, the assumption of
\(J_\mathrm{tot} \cong J_\mathrm{il}\) may no longer hold, and a
double-porous diffusion model would have to be applied in such
cases. The present concept may also reach its limits when dealing
with cations that rather sorb by surface complexation than by ion
exchange. Further work is therefore planned to extend the
investigations to such systems.
Given that the mainstream view to this day continues to be the default approach, one may think that this “further work” did show some convincing evidence for e.g. “diffusion of cations via the free pore water” at lower density. But what has actually been shown is that the “assumption of \(J_\mathrm{tot} \cong J_\mathrm{il}\)“ continues to be true for lower density!
Before we look at the additional results, we summarize the findings of
Glaus et al. (2007).
Findings in Glaus et al. (2007)
In the following we will consider the so-called “effective diffusion
coefficient”, here strictly defined as the experimental parameter
where \(j_\mathrm{ss}\) denotes the steady-state flux when an external
tracer concentration difference \(\Delta c^\mathrm{ext}\) is maintained
across a bentonite sample of length \(L\). We have
discussed through-diffusion and
the role of \(D_e\) in
many places on the blog, but in the present discussion we simply
view \(D_e\) as a normalized version the steady-state flux.
Note that we are required to compare diffusive fluxes in
different montmorillonite samples (an alternative test protocol
is suggested below). \(D_e\) varies both due to varying background
concentration (which is our object of study) and due to the variation
of different samples. It is thus crucial to minimize the latter type
of variation. This should be done (I suppose) by employing as
identical preparation protocols as possible. We will get back to this
complication of sorting out signal from noise as we comment the
results.
Glaus et al. (2007) present their results in diagrams where the logarithm of the evaluated quantities (diffusion parameters) is plotted against background concentration. This is of course convenient, as e.g. \(D_e\) can be expected to vary by two orders of magnitude as the background concentration is varied between 0.01 M and 1.0 M. But to remind ourselves what the actual dependency looks like between the normalized steady-state flux and background concentration, I will here insist on plotting the results in lin-lin diagrams.
The results for sodium in Glaus et al. (2007) plotted in lin-lin
diagrams, look like this (the data is the same in these three
diagrams)
We see that the data complies with the scaling law (eq. 2) and is quite well constrained (click on pictures to enlarge). \(D_e\) is evaluated in two ways in Glaus et al. (2007): by examining the breakthrough curve, and by examining the internal tracer profile at test termination. These methods of evaluation give more or less identical results, with the exception of the test performed at 0.01 M background concentration. In this low concentration limit, the confining filters increasingly restrict the flux, making it difficult to extract actual clay transport parameters. We have discussed this issue (and this particular study) at length in a previous blog post.
Even with the problem of accurately measuring \(D_e\) at the lowest background concentration, the results clearly demonstrate the behavior of a homogeneous system (eq. 2): e.g., \(D_e\) undoubtedly increases by a factor of approximately 10 when the background concentration is lowered from 1.0 M to 0.1 M.
The data for strontium in Glaus et al. (2007) only covers the
background concentration interval 0.5 M — 1.0 M, and is consequently
less constrained, as seen here
This data also has the peculiarity that the diffusivity of samples of length 5.4 mm is almost twice as large as for samples of length 10.4 mm. This clearly demonstrates how sample preparation becomes crucial when conducting these types of tests. In the plots above, I have allowed myself to treat samples of different length separately (Glaus et al. (2007) use average values). It is clear from the data, that also strontium is compatible with the scaling law of eq. 2. In particular, it can be distinguished that sodium and strontium have different dependencies.
The take away message from these results is clear: montmorillonite at this density (1950 kg/m3) behaves as a homogeneous system and shows no indication of containing additional pore structures.
Glaus et al. (2013) and NTB-17-12
After the publication of Glaus et al. (2007), corresponding results for lower densities has been presented. Glaus et al. (2013) — which is mostly recognized for demonstrating the seeming “uphill” diffusion effect — also contains measured \(D_e\) of sodium as a function of background concentration in conventional through-diffusion tests, both for density 1600 kg/m3 and 1300 kg/m3. These results are also published in more detail,2 together with new strontium results, in the NAGRA technical report NTB-17-12. We therefore look at these two publications together.
The additional data for sodium is here compared with the results from
Glaus et al. (2007)
For some of the additional tests, both through- and out-diffusion were performed. These points are labelled “TD” and “OD”, respectively, in the diagrams. We see that even for density as low as 1300 kg/m3, the data complies with the behavior of a homogeneous system (eq. 2) and is quite well constrained; in particular, there is nothing in the data for 1300 kg/m3 that suggests that these systems behave principally different than the 1950 kg/m3 samples.
For the system at 1300 kg/m3 and background concentration 0.1 M, two different values of \(D_e\) are presented in NTB-17-12. Only the lower of these values (\(7.0\cdot 10^{-10}\) m2/s) was published in Glaus et al. (2013), but NTB-17-12 presents a continued analysis that includes filter resistance, giving the value of \(D_e\) presented in the diagram. I think this is quite interesting, as the tests made at 0.1 M used “flushed” filters in order to minimize filter resistance. Apparently, filter resistance is still influential and it is not that easy to “design away” this problem.
NTB-17-12 also presents measured values of \(D_e\) for strontium under similar conditions (1300 — 1900 kg/m3, 0.1 — 1.0 M NaClO4 background), and are here compared with the earlier results
Although it naturally contains some scatter, we note that the additional data for ~1900 kg/m3 strengthens the earlier conclusion that also strontium scales in accordance with eq. 2. And just as for sodium, we see that the behavior does not qualitatively change, even for densities as low as 1300 kg/m3.
In the above diagrams are plotted single values for \(D_e\) for strontium at the lowest background concentration (0.1 M). It should be noted that these are burdened with large uncertainties as the transport restriction of the confining filters is severe; in NTB-17-12 are presented a whole set of simulations of the underlying flux evolution and concentration profiles with variations of the filter transport parameters. It is thus very clear that the problem of eliminating transport restrictions at the sample interfaces are not easy to completely eliminate. This is not surprising, as the theory suggests that \(D_e\) increases without limit with decreasing background concentration. Note that this behavior is strongly enhanced for divalent strontium; the measured values are many times larger than the corresponding diffusivity in bulk water (\(0.79\cdot 10^{-9}\) m2/s).
Even if the value of \(D_e\) is quite uncertain at the lowest background concentration, the mere observation that filter diffusivity strongly influence the process is, in a sense, itself a confirmation that the system still is governed by the behavior of interlayers.
The picture is quite clear from these findings: the combined results of Glaus et al. (2007), Glaus et al. (2013) and NTB-17-12 validates a homogeneous view of compacted bentonite, at essentially any relevant density!
The curious case of Bestel et al. (2018)
Bestel et al. (2018) further examine how \(D_e\) for sodium varies with background concentration. This publication shares some of the same authors with the previous studies, and presents additional measurements of \(D_e\) for sodium in essentially identical systems (similar preparation protocols, “Milos” montmorillonite, NaClO4 background electrolyte, flushed filters). Given the substantial evidence for homogeneous behavior collected in the publications discussed above, I find the conclusions of Bestel et al. (2018) rather odd.
Bestel et al. (2018) perform subsequent measurements of the steady-state flux in the same samples at different temperatures. The dependency of \(D_e\) on background concentration, however, looks essentially the same for each temperature, and — just as Bestel et al. (2018) — we here focus mainly on the results for 25 \(^\circ\mathrm{C}\). This data looks like this3
In their analysis, Bestel et al. (2018) include the results from Glaus et al. (2007) and Glaus et al. (2013), but treat them separately. They consequently conclude implicitly that, although the earlier studies found that \(D_e\) depends on background concentration in accordance with eq. 2, the new results show a different behavior. Specifically, they conclude that \(D_e\) scale with background concentration as \(C_\mathrm{bkg}^{-0.52}\) for density 1300 kg/m3 and as \(C_\mathrm{bkg}^{-0.76}\) for density 1600 kg/m3. Bestel et al. (2018) write
The results obtained in the present work for a broad variety of bulk
dry densities of Na-montmorillonite and concentrations of the
background electrolyte, give clear evidence that the equilibrium
distribution of cations between the clay phase and the external
aqueous phase is the main parameter influencing the observed overall
diffusive fluxes of cations. Whether the observed overall diffusive
fluxes are described by a physical subdivision of the pore space
into domains containing different species (e.g. the model proposed
in Appelo and Wersin (2007) or Bourg et al. (2007)), or whether
they are the result of the concentration gradients of such species
in a single type of pore (e.g. the model proposed by Birgersson and
Karnland (2009)), cannot be decided unambiguously from the available
data — notably because of the wide similarity of the model
predictions and because of some internal inconsistencies in the
experimental data. Both types of models would require some
adjustments in order to fully match the data. The diffusion data of
\(^{22}\mathrm{Na}^+\) can equally be described by a surface diffusion
model with a reduced, but non-zero mobility of sorbed cations,
similar to the median value determined in Gimmi and Kosakowski
(2011).
I think this is a problematic way of arguing and presenting data.
The data obviously has scatter
To begin with, why are the results from this study and the ones from Glaus et al. (2007) and Glaus et al. (2013) treated separately? When treated separately — according to Bestel et al. (2018) — these results are vaguely supposed to be incompatible: the dependence of \(D_e\) either comply with eq. 2 or it does not. I think that the appropriate thing to do is to discuss possible causes for why the new results supposedly differ from the earlier ones. As we have made clear above, all factors that determine \(D_e\) are not fully controlled in tests like these (e.g, what causes the difference in diffusivity for strontium in 5.4 mm and 10.4 mm samples, respectively, in Glaus et al. (2007)?). We have also seen that it is difficult to make accurate measurements at low enough background concentration, even with flushed filters.
Look e.g. at the specific values of \(D_e\) at background concentrations 1.0 and 0.1 M, respectively, in NTB-17-12 and Bestel et al. (2018) (unit is m2/s).
Under ideal conditions, these values would not differ for the same conditions in the two studies. The scatter of these values is moreover quite random, e.g. one study does not have values that are systematically larger than in the other. In Bestel et al. (2018) we also see that the mere disturbance of a sample in form of a temperature pulse may alter the diffusivity significantly (temperature is first increased in steps from 25 \(^\circ\mathrm{C}\) to 80 \(^\circ\mathrm{C}\), then decreased in steps to 0 \(^\circ\mathrm{C}\), and finally increased again to 25 \(^\circ\mathrm{C}\)). In e.g. one sample of density 1600 kg/m3 and background concentration 0.1 M is reported \(D_e = 3.4\cdot 10^{-10}\) m2/s at 25 \(^\circ\mathrm{C}\) before the conducted temperature changes, and \(2.3\cdot 10^{-10}\) m2/s after. One should also consider that the samples are not prepared equally, as they are saturated directly with the corresponding background solution. (This is also true for the previous studies.) Could this cause differences in diffusivity?
Bestel et al. (2018) should thus either argue for why the new results
are more accurate (or why the results of Glaus et al. (2007, 2013)
are less accurate) or treat the data from all studies in accumulation
and admit substantial experimental uncertainty. My impression is that
Bestel et al. (2018) make a little of both.
The data still complies with a homogeneous view
Looking at the aggregated sodium data, a somewhat different picture
emerges
Here is also included a model labelled “Full Donnan”, which takes into account the excess salt that is expected to enter the interlayers. For all other samples we have discussed, this contribution is only minor and can be neglected, and this assumption underlies eq. 2. For the sample of density 1300 kg/m3 with background concentration 5.0 M, however, the excess salt is not negligible and must be included in the analysis of the behavior of a homogeneous system (the deviation from eq. 2 is seen to become significant around 1.0 M background concentration). Bestel et al. (2018) actually present a full Donnan calculation for the excess salt, but, for unknown reasons, do not compare it directly with the experimental results (it is plotted in a separate diagram next to the data).
For 1300 kg/m3, I would claim that the “Full Donnan” model fits better to the accumulated data than the scaling law suggested in Bestel et al. (2018) (exponent \(-0.52\)). For 1600 kg/m3, the suggested scaling law (exponent \(-0.76\)) indeed fits better to the data than eq. 2, but the data is not that well constrained. To use this singular result to argue for a non-homogeneous bentonite structure basically boils down to claiming that the values measured at 0.1 M — a concentration range that is documented to be difficult to measure accurately — could not possibly be underestimated by, say, 50% (while also ignoring all other results).
If we also consider the results for strontium presented in NTB-17-12, I mean that the only reasonable conclusion that Bestel et al. (2018) can draw is that the results comply with a homogeneous bentonite structure.
Additional model components should not be motivated
solely by the ability of a model to be fitted to some arbitrary
data
A major motivation for measuring how \(D_e\) depends on background concentration at lower densities, according to Glaus (2007), is that “the assumption of \(J_\mathrm{tot} \cong J_\mathrm{il}\) may no longer hold”. What (I mean) has been demonstrated in the subsequent studies is that this assumption actually does hold. In particular, from the aggregated data it is not possible to claim that the behavior of \(D_e\) is qualitatively different at 1950 kg/m3 and 1300 kg/m3. Thus, there is no valid justification for introducing more complex model components. Moreover, introducing e.g. a bulk water phase causes fundamental conflicts with the description of other well-established properties of these systems, particularly swelling pressure. Adding such components merely to improve agreement with a specific dataset, while ignoring their broader implications, undermines the model’s overall coherence and validity. The data cannot“equally be described by a surface diffusion model”.
What does some alternative model actually predict?
Eq. 2 (or a full Donnan calculation) is a clean statement of the expected behavior of a homogeneous system (based on how interlayers function). If actual deviations from this behavior could be established we may conclude that a homogeneous description is not sufficient. However, any arbitrary deviation from eq. 2 does not automatically validate any specific alternative model. Validating a model requires that we can experimentally reproduce some of its non-trivial predictions. Bestel et al. (2018) don’t discuss what the exponents \(-0.52\) and \(-0.76\) are suppose to represent.
Note also that the arbitrary exponent \(-0.52\) is inferred by fitting to the data at 5.0 M background concentration. But we saw above that a full Donnan calculation within the homogeneous view actually explains the behavior in this concentration limit (Bestel et al. (2018) show this!). We have thus every reason to believe that the exponent \(-0.52\) is just spurious and do not represent some actual physical mechanisms.
A suggestion for how to preferably conduct these types of tests
The discussed studies are enough to convince me that cation tracer
diffusion behave in accordance with a homogeneous bentonite view at
any relevant density.
It is however also clear that the full variation of \(D_e\) in these
tests is caused by more factors than just background concentration and
density. To eliminate as much as possible of this scatter — and thus
to more accurately determine the dependence of background
concentration on \(D_e\) — I suggest the following test protocol.
Measure tracer flux at several background concentrations in the same sample.
This would eliminate both the unavoidable (small) variation in density between different samples as well as several unknown factors that determine the exact value of diffusivity (these may e.g. be related to variation in material or equipment and to sample handling)
Prepare samples by saturating them all with the same low concentration solution (e.g. 0.05 M).
To me it seems reasonable that the way samples saturate may influence the resulting detailed structure and thus the diffusivity. Saturating all samples in the same manner with the same solution will minimize variations from such effects.
Keep temperature constant.
I don’t think this is a crucial factor, but we see in Bestel et al. (2018) that larger temperature pulses may significantly alter the diffusivity.
Increase background concentration in steps and record the steady-state flux at each concentration.
I think a good range may be between 0.2 M and 1.0 M. For a homogeneous system, this corresponds to a variation in \(D_e\) by a factor 5 for monovalent and 25 for divalent cations.5 At the same time, the problem of filter transport resistance can hopefully be kept under control.
Decrease the background concentration (perhaps in steps) back to the first concentration where steady-state flux was measured.
Measure steady-state flux again and assert that no significant change in \(D_e\) has occurred as a consequence of the disturbance introduced by the background concentration pulse.
Final thoughts
The only reasonable conclusion to draw from the studies we have looked
at is that the behavior of cation tracer diffusion indicates a
relatively homogeneous structure, dominated by interlayers, in any
relevant bentonite system. Despite this, the contemporary scientific
bentonite literature is crammed with non-homogeneous descriptions of
compacted bentonite, centered around a bulk water phase (the
“mainstream view”). As we have seen here, this can even be the case
for studies that provide evidence for homogeneity.
What I find most frustrating is that interlayer effects often are viewed as some additional feature to be handled in specific cases. In reality, virtually all experimental findings (diffusion, swelling pressure, temperature response, Donnan effects, fluid flow, hyperfiltration, …) indicate that the behavior of compacted bentonite is fully governed by interlayers. The question is not if a presumed bulk water phase may dominate under certain conditions, but if such a phase is at all relevant. I want to emphasize this point: up until this day, no convincing evidence has ever been presented that compacted bentonite contains significant amounts of bulk water.
Even if the structure becomes more complex at lower densities, a
homogeneous model centered around interlayers guarantees to cover at
least some aspects of the system. On the contrary — if the goal is
process understanding — most experimental evidence rules out
bentonite models that assume a bulk water phase.
[2] As
far as I can see, these tests were done in duplicates for Na
diffusion with background concentrations 0.5 M and 1.0 M, and the
the numbers reported in Glaus et al. (2013) are averages.
[3] Bestel et
al. (2018) use a normalization scheme in their analysis that
involves corresponding measured water diffusivities and parameters
from “Archie’s law” (note, it is the
quotation marks version of the law). I think this handling makes
the presented results less transparent, and here we use the actual
reported values of \(D_e\).
[4] These are only values from the first phase at 25 \(^\circ\mathrm{C}\).
[5] I assume that measurements are being made in pure Na-montmorillonite.
The study consists of chloride and iodide though-diffusion tests in sets of samples of “Kunigel V1” bentonite, mixed with either 0%, 30%, or 50% silica sand. Here we mainly focus on the chloride tests. Also, we exclude the samples with 50% sand, as the montmorillonite content is judged to small. For each type of material, chloride diffusion tests were performed with NaNO3 background concentrations 0.01 M, 0.5 M, and 5.0 M. All samples are cylindrical with diameter 2 cm and height 1 cm (giving a volume of 3.14 cm3) and have dry density 1.6 g/cm3, which means that the effective montmorillonite density varies in the different test sets. To refer to a single test we use the notation “sand mixture percentage/background concentration”, e.g. “30/0.5” refers to the test made on the sample with 30% sand and with background concentration 0.5 M.
A single additional test was performed on purified “Kunipia F” material, at dry density 0.9 g/cm3 and a background concentration of 5.0 M NaNO3. This density was chosen in order to have a similar effective montmorillonite dry density as the “Kunigel V1” samples with 30% sand.
All tests were performed at elevated pH in the external solution of 12.5 (initially), and the Cl diffusion tests were performed in a N2 glove box, with vanishing CO2 and O2 pressures. In total we here investigate 7 tests (of the 22 tests in the full study, we exclude 12 that concern iodide diffusion, and 3 that have 50% sand). In addition to the published article, these tests are also reported in a technical report (in Japanese).
Materials
“Kunigel V1” and “Kunipia-F” are simply
brandnames rather than materials specifically aimed for scientific
studies. This is similar to e.g. “MX-80” and “KWK”, that we have
encountered in
previousassessments.
I have found it rather difficult to obtain official data on “Kunigel V1” and “Kunipia F”; data sheets or technical specifications do not seem readily available online. Moreover, the Japan Atomic Energy Agency seem to contain their data within a database, and restrict its usage (this site seems a bit deserted, though). Fortunately, the open scientific literature contains some entries. These sources, however, provide quite different values for e.g. montmorillonite content and exchangeable cations in “Kunigel V1”.
Montmorillonite content
Several studies of “Kunigel V1” — including Is08 — refer to a single source for e.g. mineral content and cation exchange capacity: Ito et al. (1994),1 which states that “Kunigel V1” contains 46% — 49% montmorillonite. Other sources, however, claim considerably different numbers; e.g. Cai et al. (2024) states a montmorillonite content of 54.3%, while Kikuchi and Tanai (2005) states 59.3%.
Here, I do not intend to critically assess these various sources, but simply conclude that the montmorillonite content stated in Is08 must be viewed with some skepticism. The study they reference (Ito et al. (1993)) is significantly older than their own, and they do not indicate that they have investigated the material actually used. In this assessment we adopt an uncertainty for the montmorillonite content in “Kunigel V1” of 45% — 60%.
Concerning “Kunipia F” most sources I have investigated state a montmorillonite content above 99%, although some — including Is08 — set a lower limit at 95%. Here we assume that the montmorillonite content of “Kunipia F” lies in the interval 95% — 100%.
Cation population
Reports on cation exchange capacity (CEC) and exchangeable cations in “Kunigel V1” are also quite scattered in the scientific literature, as demonstrated in the table below.
CEC values (roughly) in the range 0.55 — 0.80 eq/kg are reported. These numbers will not be further assessed here, and we will assume an uncertainty of this range for the CEC in “Kunigel V1”.
One observation to be made is that some of the sources reporting relatively high CEC also reports relatively high montmorillonite content. The data from e.g. Kikuchi and Tanai (2005) gives an estimate of the cation exchange capacity for the montmorillonite of 0.75/0.593 eq./kg = 1.26 eq./kg, while the data from Ito et al. (1994) gives roughly 0.556/0.475 eq./kg = 1.17 eq./kg. These numbers are quite consistent and suggest that the reported differences in CEC may partly be due to differences in montmorillonite content in different batches of “Kunigel V1”.
We can further conclude that the reported amount of exchangeable sodium in “Kunigel V1” is rather stable (with some exception), while the amount of exchangeable calcium and magnesium scatter significantly. This scatter is mainly due to interference of soluble accessory minerals (see below; entries in the above table where such interference is obvious are put within parentheses). Thus, the exchangeable cation population in “Kunigel V1” can be estimated to about 80% — 90% sodium and about 10% — 20% di-valent ions (calcium and magnesium).
Some cation data for “Kunipia F” found in the literature is listed in the table below (the table contains a few entries for the variants “Kunipia-G” and “Kunipia-P”; these are indicated).
The most commonly reported CEC value in this little survey is 1.19 eq./kg, and I suspect that this has been supplied by the manufacturer (although the value 1.15 eq./kg has also been reported as a given from the manufacturer). As “Kunipia F” is mainly pure montmorillonite, note that this value (1.19) is consistent with the montmorillonite CEC estimated from “Kunigel V1” above. That being said, the scatter in reported CEC for “Kunipia F” is in the range 1.0 — 1.22 eq./kg.
The few reported cation populations of “Kunipia F” (and the variant “Kunipia G”, which is supposed to be identical in composition) that I have found have a higher sodium content as compared with “Kunigel V1”, roughly in the range 85% — 95%.
Soluble accessory minerals
Basically all sources I have encountered — including Is08 — say that “Kunigel V1” contains smaller amounts of calcite and dolomite. This is also quite evident from some of the reported results on exchangeable cations, where the sum of these substantially exceeds the evaluated CEC. Obviously, the presence of additional calcium and magnesium contribute to the uncertainty and complexity when evaluating effects of ion equilibrium in this material (just as for the cases of “MX-80” and “KWK”).
Sample density
The samples in Is08 were ultimately sectioned and analyzed (for the final state concentration gradient). Is08 nowhere state that they measured density of these sections. We thus proceed with using the nominal density of 1.6 g/cm3. Using the above estimated uncertainty in montmorillonite content we get the following intervals for the effective montmorillonite density
Samples
EMMD interval (g/cm3)
0% sand
1.05 — 1.24
30% sand
0.83 — 1.01
Kunipia F
0.87 — 0.90
Note that these intervals do not include uncertainty due to variation in density of the actual samples.
Uncertainty of external solutions
The samples were prepared by first saturating them with deionized water for more than two weeks, and thereafter contacting them with NaNO3 solutions for more than five weeks.
We have no reason to doubt the accuracy of the initial concentration of the salt solutions, but contacting a bentonite containing di-valent ions with pure sodium solutions inevitably initiates an ion exchange process. We have made the same conclusion for studies using “MX-80” and “KWK” bentonite. Similar to the previous studies, Is08 do not keep track of the exact chemical evolution of the external solutions, but we can calculate an estimate of the extent of the sodium-for-di-valent exchange.
The above diagram shows the result of equilibrating the specified amount of bentonite (3.14 cm3) with the specified amount of external solution (100 ml) for different initial NaNO3 concentrations. The calculation assumes that the bentonite only contains sodium and calcium, with an initial calcium content of 15%, a selectivity coefficient of 5 M, and a cation exchange capacity of of 0.65 eq/kg. The diagram shows the amount of calcium left in the sample after equilibration, as a function of initial NaNO3 concentration for the cases of 0% and 30% mixed-in silica sand. The dashed vertical lines indicate the external concentrations in the performed tests. We note — as we have done for several other studies — that the equilibrium amount of di-valent ions still in the bentonite depends significantly on the initial NaNO3 concentration: tests performed at 0.5 M and 5.0 M gives essentially a pure sodium clay, while samples used at 0.01 M still contain the initial 15% di-valent ions in the clay.
Since the “Kunipia F” material only is used in a test with background concentration of 5.0 M, we can quite safely assume that the exchangeable cation population in this particular test is basically 100% sodium.
It should be noted that the calculations have not accounted for the
additional di-valent ions present in the bentonite in form of
accessory minerals (calcite, dolomite). They thus probably
underestimate the amount di-valent ions still left in the clay after
equilibration.
Evaluations from the diffusion tests
The diffusion tests were performed by sandwiching the clay samples between a source and target reservoir of equal volumes, 50 ml. The initial Cl tracer concentration was 0.05 mM in the source reservoir, and 0.0 mM in the target reservoir.
The tracer concentration in both the source and target reservoirs were
periodically measured, but as far as I understand, none of the
reservoir solutions were replaced during a test. This means that a
certain concentration build-up occurs in the target reservoir, and
a corresponding concentration drop occurs in the source reservoir.
The test set-up furthermore involves quite wide “filter” components at the interfaces between clay and reservoirs.2 Is08 mean that these components restrict diffusion to such an extent that they must be included in the test analyses. With a rather complex set-up that involves evolving reservoir concentrations and “filter” influence, the preferred way to evaluate them would be a full simulation of the whole process. This is however not the procedure followed in Is08 (below we make such simulations).
Instead, Is08 center most of their evaluation around the measured steady-state flux,3 taking filter diffusion into account. In the blog post on on filter influence on through-diffusion tests we derived an expression for the steady-state flux, which can be written
where \(D_e\) is the effective diffusivity, and \(L\) the length of the clay
component. \(\Delta c_\mathrm{res}\) is the difference in
concentration between the two reservoirs, and \(\omega\) is the relative
filter resistance, given by
\begin{equation} w = \frac{2D_eL_f}{D_fL} \tag{2} \end{equation}
where \(D_f\) and \(L_f\) denote effective diffusivity and length of the
two confining “filters” (assumed identical).
which is the same expression as found in Is08 (eqs. 2 and 3 in Is08).
\(D_e\) is thus evaluated in Is08 by measuring \(j^\mathrm{ss}\) and \(D_f\), estimating \(\Delta c_\mathrm{res}\), and knowing the lengths of the clay and filter components (\(L\) = 1 cm, \(L_f\) = 1.5 cm). Note that this is a quite involved procedure, necessitated by the test design: the source reservoir is small enough for the concentration to significantly drop during the course of a test; the target is not replaced during the course of a test, resulting in an increasing concentration significantly different from zero; the sample is sandwiched between wide “filter” components; and, as far as I can tell, the external solutions are not stirred or circulated. With a simpler test design, the reservoir concentration difference could have been kept effectively constant, and influence from confining filters could have been avoided (the only case, really, where filter influence is unavoidable is for cation through-diffusion at low ionic strength). With this being said, a re-evaluation of the results demonstrates that the “filter” influence, after all, is quite moderate. We will further discuss this below.
Is08 estimate \(\Delta c_\mathrm{res}\) by using the average source reservoir concentration during the course of a test (\(\bar{c}_\mathrm{source}\)), and by assuming zero target reservoir concentration, i.e. \(\Delta c_\mathrm{res} = 0-\bar{c}_\mathrm{source}\). I do not really understand this, because the target reservoir concentration is clearly not zero; since the two reservoirs have the same volume it seems more reasonable to assume that the concentration drop in the source reservoir corresponds to an equal concentration increase in the target reservoir.4
The “filter” diffusivities are claimed to be measured in separate tests without clay components, but the reported values does not make full sense to me. It is claimed that three different values for \(D_f\) were used for the three different background concentrations. But we do not expect any significant difference in diffusivity due to background concentration. Does this mean that tests performed at a specific background concetration all used the same test cell, while different test cells were used for different background concentrations? Furthermore, the specified values are \(D_f = 3\cdot 10^{-10}\) m2/s for background concentration 0.01 M, \(D_f = 2.6\cdot 10^{-9}\) m2/s for background concentration 0.5 M, and \(D_f = 1.8\cdot 10^{-9}\) m2/s for background concentration 5.0 M. The \(D_f\) values at high background concentration are thus not only almost an order of magnitude higher than that for 0.01 M background, these values also implies a diffusivity larger than for pure bulk water.5
If we anyway use these values for \(D_f\) to calculate the relative filter resistances (eq. 2) we get maximum values for \(\omega\) of 0.077, 0.037, and 0.055 for background concentrations 0.01 M, 0.5 M, and 5.0 M, respectively (anticipating the evaluated \(D_e\) values in table 1 in Is08). These values are tiny, showing that their own estimations indicate insignificant “filter” influence.
In the following we de-derive the values for \(j^\mathrm{ss}\) and \(\Delta c_\mathrm{b}\) (the final clay concentration difference) used for evaluating the reported values of \(D_e\), “\(D_a\)”, and \(\epsilon_\mathrm{eff}\),6 and compare them with the raw flux and concentration profile data (available for the tests performed with 30% sand mixture).
Steady-state fluxes
The steady-state fluxes are nowhere stated explicitly in Is08, but it is straightforward to read them off from the provided “breakthrough curves”. To check the consistency of the reported parameters we may use these values and the reported values for \(D_f\) and \(D_e\) to back-calculate \(\Delta c_\mathrm{res}\) using eq. 3.
In this table are also listed the “expected” values of of the reservoir concentration differences, \(\Delta c_\mathrm{res,ex}\), estimated from subtracting the average concentration increase in the target reservoir from from 0.05 mM. We see that the reported values of \(D_e\) “overestimates” \(\Delta c_\mathrm{res}\) by 8% — 40%.
We do not have more information to assess whether this mismatch is
due to some actual inconsistency in the reported values or if it indicates
that the concentration difference stated in the article was not
actually what was achieved in the experiment. In any case, this is low
quality scientific reporting.
Concentration profile gradients
We can, however, continue by also checking the consistency of the estimated pore diffusivity, \(D_p\),7 which was evaluated by measuring the concentration gradient in the clay at the termination of the tests (\(\Delta c_b/L\)).8
The concentration gradients are not explicitly stated in the article, but we can read them off from the published concentration plots. By using the tabulated values of \(D_p\) we can use eq. 4 to back-calculate what values for the steady-state flux was used for their evaluation.
Note that some of these values of \(j^\mathrm{ss}\) are smaller than
what can be read off from the “breakthrough curves”. In particular,
the value for the 30/5.0 test is reduced by more than 30%. If we use
these values of \(j^\mathrm{ss}\) to re-calculate the corresponding
reservoir concentration differences, we get
Although the calculated value for \(\Delta c_\mathrm{res}\) still is larger than 0.05 mM for the 30/0.5 test, these values are now generally in better agreement with the “expected” estimations.
I do not really know what to make of these results. For the 30/0.01 and 30/0.5 tests, the slightly different results perhaps reflect the uncertainty in the estimation of \(j^\mathrm{ss}\) and \(\Delta c_b\). But there is clearly something wrong with the evaluation of the 30/5.0 test. From the diagram (fig. 2 in Is08), it is, for example, clear that this test has the largest flux.
Chloride equilibrium concentrations
The chloride equilibrium concentration is evaluated in Is08 in terms of an “effective porosity,”6 \(\epsilon_\mathrm{eff} = D_e / D_p\). But from eq. 3 and eq. 4 we see that it is really evaluated from
Note that the factors \(j^\mathrm{ss}L\) cancel; the evaluation of \(\epsilon_\mathrm{eff}\) is therefore less sensitive to the estimation of \(j^\mathrm{ss}\) (the flux only appear in the correction term due to filter influence). Thus, even if the evaluation of \(j^\mathrm{ss}\) evidently has its flaws, the evaluation or \(\epsilon_\mathrm{eff}\) is more robust. This reflects the fact that the equilibrium concentration, as the name suggest, does not depend on transport quantities; as is clear from eq. 5, \(\epsilon_\mathrm{eff}\) is simply an interpretation of the clay concentration (\(c_b\)). We have discussed this issue severaltimesbefore.
Eq. 5 also shows that the uncertainty in estimating the equilibrium concentration (or \(\epsilon_\mathrm{eff}\)) mainly stem from uncertainties in \(\Delta c_\mathrm{res}\), and uncertainty stemming from filter resistance (\(2j^\mathrm{ss}L_f/D_f\)). Both of these uncertainties could have been avoided with a better test design — if filter resistance was avoided, and if the source and target reservoirs were kept at (virtually) constant concentrations, the equilibrium concentration would be given directly from the clay concentration profile.9
One way to estimate the effects of these uncertainties is to simply
compare the reported values for \(\epsilon_\mathrm{eff}\) with the ratio
\(\Delta c_b/\Delta c_\mathrm{res,init}\), where
\(\Delta c_\mathrm{res,init}\) = -0.05 mM is the initial reservoir
concentration difference.
The differences are not that great, demonstrating that reported values
of equilibrium concentrations (\(\epsilon_\mathrm{eff}\)) are quite
robust, even though we have found inconsistencies in the underlying
transport quantities.
Why not just simulate the whole thing?
A better way, in my view, to extract the equilibrium concentrations from this rather complex test set-up is to simulate the tests completely. This is done here, taking into account the external reservoir, the “filter” components and using the homogeneous mixture model for the bentonite component. Note that the homogeneous mixture and the effective porosity models are equivalent when it comes to modeling this type of diffusion: the effective porosity parameter can be calculated from \(\epsilon_\mathrm{eff} = \phi\cdot\Xi\), where \(\phi\) is the physical porosity and \(\Xi\) is the ion equilibrium coefficient. Similarly, the diffusion coefficient in the homogeneous mixture model (\(D_c\)) can in this case directly be identified with the pore diffusivity in the effective porosity model (\(D_p\)). In these simulations we used \(\Xi\) and \(D_c\) as fitting parameters.
The fitted parameters are listed in the table below and compared to
the reported values of \(D_p\) and \(\epsilon_\mathrm{eff}\).
Test
\(D_c\) (10-10 m2/s)
\(D_p\) (10-10 m2/s)
\(\Xi\)
\(\phi\cdot\Xi\)
\(\epsilon_\mathrm{eff}\)
30/0.01
2.14
1.8
0.103
0.043
0.043
30/0.5
2.50
2.4
0.457
0.19
0.13
30/5.0
2.39
1.6
0.625
0.26
0.21
Below is the simulated outflux curves and final state clay concentration profiles compared with experimental data.
0.01 M background concentration:
0.5 M background concentration:
5.0 M background concentration:
The simulations were performed both with (green lines) and without
(red lines) including filters. It may be noted that both models can be
fitted equally well, confirming that filter effects are after all
small in these tests. Also, although the diffusion parameters change
to some extent, the fitted ion equilibrium coefficients are
essentially the same regardless of whether filters are included or
not. We note that the spread in the values for the diffusion parameter
is smaller for the simulations as compared with the reported
values. As we expect similar diffusivity in these identically prepared
samples, I see this as a confirmation that a simulation better
captures the experimental parameters. Concerning the ion equilibrium,
or equivalently the “effective porosity”, we note that the
simulations provide somewhat higher values as compared with the
reported quantities, both for test 30/0.5 and test 30/5.0. The values
are however still comparable, again demonstrating that they have been
more robustly extracted.
Summary and verdict
We have seen that Is08 has several flaws and weaknesses: the test design is unnecessarily complex, and from the provided data on clay concentrations and fluxes, we have noted inconsistencies, e.g. in the values adopted for the steady state flux. It is also not completely clear if the actual initial concentration differences between the external reservoirs is 0.05 mM (as stated in the article) or if this is some measured but not reported quantity. We have also noted that the material used (primarily “Kunigel V1”) suffers from several uncertainties in its composition.
All of these factors lead to uncertainty in the quantities we are primarily interested in, i.e. chloride equilibrium concentrations. We have also seen, however, that we have reason to believe that these reported quantities are considerably more robust; most simplistically, the equilibrium concentration can be inferred by extrapolating the clay concentration profile to the interface on the target side and comparing that value to 0.05 mM.
My choice is therefore to keep these values to use for evaluating e.g. performance of models for salt exclusion. One reason that this data is interesting for this purpose is the measurement of equilibrium concentrations at an exceptionally high background concentration.
Below is a diagram that summarizes the findings of this assessment.
This figure includes gray stripes to indicate the estimated uncertainty in effective montmorillonite density (for these tests we have no means to estimate the uncertainty of the reported equilibrium concentration). For two of the tests that we have been able to look at in more detail (30/0.5 and 30/5.0) we have added an “area of uncertainty” that both include uncertainty in density and concentration. The estimation of the uncertainty in concentration is here simply done by including the values inferred from completely simulating these tests. These “areas” are no formal confidence intervals, but should be viewed as giving a hint of the uncertainties involved.
[1] Is08 actually refer to a corresponding technical report, from 1993.
[2] Apart from
“real” filters, the sample is also confined by two thick
perforated components; the total “filter” length is specified as
1.5 cm!
[3] As the concentration continually
changes in the reservoirs this is not a true steady-state, but what
we could call a “quasi”-steady state (it is still easily
distinguished from the initial part of a through diffusion test).
[4] With reservation for that the target is consumed
due to quite frequent sampling — but this would contribute to an
additional increase of the target concentration.
[5] Note that these are effective diffusivities, and includes “filter” porosity (we are not told their values, but they can of course not be larger than unity). The only value of \(D_f\) that seems reasonable is the one for 0.01 M, which corresponds to a geometric factor of 6.7. To make things even stranger, for iodide the same value is used for \(D_f\) regardless of background concentration (\(3.9\cdot 10^{-10}\) m2/s).
[6] Is08 refer to this quantity as \(\alpha\), “the capacity factor”. But it is clear from the text that it is interpreted as an effective porosity, and we will therefore use the notation \(\epsilon_\mathrm{eff}\), in accordance with earlier assessments. Is08 actually also relate the parameter \(\alpha\) to sorption via the relation \(\alpha = \phi + \rho K_d\) (their eq. 6). This is however a mix-up of two incompatible models, which I have commented on here. We also note that Is08 actually never use the distribution coefficient, \(K_d\), for anything in their analysis.
[7] Is08 call this quantity \(D_a\), but it is not an “apparent” diffusivity, and I do not accept using this bizarre nomenclature. I will call the corresponding parameter \(D_p\) in accordance with e.g. the effective porosity model.
[8] Is08 define the clay concentration, \(c_b\), in terms of total clay volume. Alternatively it can be defined in terms of total amount of water, the difference being a porosity factor.
[9] Or, rather, chloride equilibrium concentrations can then be inferred directly from the value of the clay concentration profile at the interface to the source reservoir.
This is the second part of the review of “Ionic Transport in Nano-Porous Clays with Consideration of Electrostatic Effects” (Tournassat and Steefel, 2015) (referred to as TS15 in the following). For background and context please check the first part. That part covered the introduction and the section “Classical Fickan Diffusion Theory”. The next section is titled “Clay mineral surfaces and related properties”, and is further partitioned into two subsections. Here we exclusively deal with the first one of these subsections: “Electrostatic properties, high surface area, and anion exclusion”. It only covers three and a half journal pages, but since the article here goes completely off the rails, there is much to comment on.
“Electrostatic properties, high surface area, and anion
exclusion”
As stated in the first part, I find it remarkable that the authors use
general terms such as “clay minerals” when the actual subject matter
is specifically systems with a significant cation exchange capacity,
and montmorillonite in particular. I will continue to refer to these
systems as “bentonite” in the following, disregarding the constant
references to “clay minerals” in TS15.
Stacks
After having established that montmorillonite and illite have
structural negative charge, it begins:
Clay mineral particles are made of layer stacks and the space
between two adjacent layers is named the interlayer space (Fig. 5).
This is the first mention of clay “particles” in the article, and they are introduced as if this is a most well-established concept in bentonite science (incredibly, it is also the first occurrence of the term “interlayer”). We will refer to “clay mineral particle” constructs as “stacks” in the following. I have written a detailed post on why stacks make little sense, where I demonstrate their geometrical impossibility and show that most references given to support the concept are studies on suspensions that often imply that montmorillonite do not form stacks. Sure enough, this is also the case in TS15
The number of layers per montmorillonite particle depends on the
water chemical potential and on the nature and external
concentration of the layer charge compensating cation (Banin and
Lahav 1968; Shainberg and Otoh 1968; Schramm and Kwak 1982a;
Saiyouri et al. 2000)
Banin and Lahav (1968), Shainberg and Otoh (1968), and Schramm and Kwak (1982) all report studies on montmorillonite suspensions. The abstract of Shainberg and Otoh (1968) even states “The breakdown of the tactoids occurred when the equivalent fraction of Na increased from 0.2 to 0.5. Montmorillonite clay saturated with 50% calcium (and less) exists as single platelets.”, and the abstract of Schramm and Kwak (1982) states “Upon exchange of Ca-counterions for Li-, Na-, or K-counterions, a sharp initial decrease in tactoid size was observed over approximately the first 30% of cation exchange.”. These are just different ways of saying that sodium dominated montmorillonite is sol forming.
I want to stress the absurdity of the description given in TS15. A pure fantasy is stated about how compacted bentonite is structured. As “support” for the claim are given references to studies on “dilute suspensions”. It should be clear that the way TOT-layers interact in such suspensions essentially says nothing about how they are organized at high density. But even if we pretend that these results are applicable, the given references say that most of the relevant systems (montmorillonite with about 30% sodium or more) do not form stacks.
Disregarding the references, note also how bizarre the above statement is that the number of layers in a “particle” depends on “the water chemical potential and on the nature and external concentration of the layer charge compensating cation”: stacks are supposed to be fundamental structural units, yet the number of layers in a stack is supposed to depend on the entire water chemistry?! (It makes sense, of course, for stacks in actual suspensions.) Also, for montmorillonite an actual number of layers is nowhere stated in TS15.
TS15 further complicate things by lumping together montmorillonite and
illite. In contrast to Na-montmorillonite, illite has by definition a
mechanism for keeping adjacent TOT-layers together: its layer charge
density is higher and compensated by potassium, which doesn’t hydrate
that well, leading to collapsed interlayers. As far as I understand,
one characterizing feature of illite is that the collapsed interlayers
are manifested as a
“10-angstrom peak” in X-ray diffraction measurements.
To treat montmorillonite and illite on equal footing (in a laid-back single sentence) again shows how nonsensical this description is. Stacking in montmorillonite suspensions occurs as a consequence of an increased ion-ion correlation effect when the fraction of e.g. calcium becomes large (> 70-80%). This process requires the ions to be diffusive and is distinctly different from the interlayer collapse in illite.
I actually have a hard time understanding what exactly is meant by the
term “illite” here. In clay science it is clear that what is
referred to by this name are systems that may have a quite
considerable cation exchange
capacity.1 Reasonably, such systems contain other types of cations
besides potassium2 (as they
are exchangeable), and must contain compartments where such ions can
diffuse (as they are exchangeable). To increase the complexity, there
are also “illite-smectite interstratified clay minerals”, which typically
are in “smectite-to-illite” transitional states. For these, it seems
reasonable to assume that the remaining smectite layers provide both
diffusable interlayer pores and the cation exchange capacity. I don’t
know if such “smectite layers” provides the cation exchange capacity
in general in systems that researchers call illite. Neither do I
understand how researchers can accept and use this, in my view, vague
definition of “illite”. Anyway, it is the task of TS15 to sort out
what they mean by the term. This is not done, and instead we get the
following sentence
Illite particles typically consist of 5 to 20 stacked TOT layers
(Sayed Hassan et al. 2006).
This study (Sayed Hassan et al., 2006) concerns one particular material (illite from “the Le Puy ore body”) that has been heavily processed as part of the study.3 I mean that such a specific study cannot be used as a single reference for the general nature of “illite particles”. Moreover, the stated stack size (5 — 20 layers) is nowhere stated in Sayed Hassan et al. (2006)!4
In their laid-back sentence, TS15 also implicitly define “interlayer space” as being internal to stacks. I criticized this way of redefining already established terms in the stack blog post, and TS15 serves as a good illustration of the problem: are we not supposed to be able to use the term “interlayer” without accepting the fantasy concept of stacks? To be clear, “interlayer spaces” in the context of montmorillonite simply means, and must continue to mean, spaces between adjacent TOT basal surfaces. It drives me half mad that the “stack-internal” definition is so common in contemporary bentonite scientific literature that this point seems almost impossible to communicate.
The provided illustration (“Fig 5”) explicitly shows how TS15 differ
between “interlayers” that are assumed internal, and “outer basal
surfaces” that are assumed external to the stack.
This illustration misrepresents the actual result of assembling a set of TOT-layers, just like any other “stack” picture found in the literature. The figure shows five identical TOT-layers that can be estimated to be smaller than 20 nm in lateral extension (while the text “conveniently” states that they should be 50 — 200 nm). Compared with “realistic” stacks, formed by randomly drawing TOT-layer sizes from an actual distribution, the depicted stack in TS15 looks like this5 (see here for details)
Besides the fact that “realistic” stack units cannot be used to form the structure of compacted bentonite, it should also be clear from this picture that “outer basal surfaces” and “interlayers” (in the sense of being internal to the stack) are not well defined. Note further that in actual compacted systems (above 1.2 g/cm3, say) such “realistic” stacks would be pushed together, something like this
In this picture, why should e.g. the interface between the green and the red stack be defined as an interface between two “outer surfaces” rather than an interlayer? Also, is this interface supposed to change nature and become an “interlayer”, as the water chemical potential or the external ion content changes? Like all other proponents of stack descriptions that I have encountered, TS15 do not in any way explain how “interlayers” and “outer surfaces” are supposed to function fundamentally differently. Similarly, they do not describe how the number of layers in a stack depends on water chemistry, nor do they provide a mechanism for why (sodium dominated) montmorillonite stacks of are supposed to keep together.
I want to emphasize that I do not favor any construction with
“realistic” stacks, but only use them to illustrate the absurd
consequences of taking a stack description seriously, and to
demonstrate that all such descriptions in the bentonite
literature are essentially pure fantasies, including the one given in
TS15. I’m also quite baffled as to why TS15 (and others)
provide such completely nonsensical descriptions, and how these can
end up in review articles. I believe a hint is given in this
formulation
[T]he number of stacked TOT layers in
montmorillonite particles dictates the distribution of water in two
distinct types of porosity: the interlayer porosity […]
and the inter-particle porosity.
The only way I can make sense of this whole description is as an
embarrassing attempt to motivate the introduction of models with
several “distinct types of porosity”: the outcome is simply a
macroscopic multi-porosity model (which will also be evident in later
sections).
I’ve written a detailed blog post on why multi-porosity models cannot be taken seriously. There I point out that basically all authors promoting multi-porosity for some reason attempt to dress it up in terms of microscopic concepts, while the models obviously are macroscopic. Moreover, no one has ever suggested a mechanism for how equilibrium is supposed to be maintained between the different types of “porosities”.
Anion exclusion
After hallucinating about the structure of compacted bentonite, TS15 change gear and begin an “explanation” of anion exclusion. Let’s go through the description in detail.
The negative charge of the clay layers is responsible for the
presence of a negative electrostatic potential field at the clay
mineral basal surface–water interface.
I cannot really make sense of the term “negative electrostatic
potential field”, although I think I understand what the authors are
trying to say here. What is true is that the electrostatic potential
near a montmorillonite basal surface is lowered compared to a
point farther away. But whether or not the value of the potential is
negative is irrelevant, as we are free to choose the reference
level. If the zero level is chosen at a point very far from the
surface, which often is done, it is true that the potential is
negative at the surface. But the key principle is that the potential
decreases towards the surface.6
A varying
electrostatic potential signifies an electric field, which in this
case is directed towards the surface (\(E = -d\phi/dx\)).
Furthermore, the electric field is not present merely because
of the presence of negative charge, but because this charge is
constrained to be positioned in the atomic structure of the
clay. Remember that the structural clay charge is compensated by
counter-ions, and that the system as a whole is charge neutral. The
reason for the presence of an electric field near the surface is due
to charge separation. And the reason for the potential
decreasing (i.e. the electric field pointing towards the surface) is
because it is the negative charge that is unable to be completely
freely distributed.7
The concentrations of ions in the vicinity of basal planar surfaces
of clay minerals depend on the distance from the surface
considered. In a region known as the electrical double layer (EDL),
concentrations of cations increase with proximity to the surface,
while concentrations of anions decrease.
Having established that the electrostatic potential varies in
the vicinity of the surface, it follows trivially that the ion
concentrations also vary. I also find it peculiar to label the regions
where the concentrations varies as the EDL. An electric double layer
is a structure that includes both the surface charge and the
counter-ions (hence the word “double”). What is described here
should preferably be called a diffuse layer. Note, moreover, that the
way an electric double layer here is introduced implies that TS15
consider a single interface, i.e. some variant of the Gouy-Chapman
model (this becomes clear below). But this model is not applicable to
compacted bentonite.
At infinite distance from the surface, the solution is neutral and
is commonly described as bulk or free solution (or water).
Here I think it becomes obvious that the authors try to motivate the presence of bulk water within the clay structure. As described in the blog post on “Anion accessible porosity”, it is only reasonable to assume that diffuse layers merge with a bulk solution in systems that are very sparse — i.e. in suspensions.8 This is how e.g. Schofield (1947) utilized the Gouy-Chapman model to estimate surface area. But how is the solution next to a basal surface in compacted bentonite supposed to merge with a bulk solution? Even if we use the authors’ own fantasy stack constructs, the typical structure of compacted bentonite must be envisioned something like this (I have color coded different stacks to be able to understand where they begin and end).
The regions where basal surfaces of different stacks face each other
(labelled A) are way too small in order to merge with a bulk solution
(and, as asked earlier, how are these regions even different from
“interlayers”?). Furthermore, regions adjacent to external
edge-surfaces of these imaginary stack units (B, C) are not at all
considered by applying a Gouy-Chapman model. The only way to make
“sense” out of the present description is to imagine larger voids in
the clay structure, something like this
But even if such voids would exist (in equilibrated water-saturated bentonite under reasonable conditions, they do not) they would only constitute an exotic exception to the typical pore structure. By focusing on this type of possible “anion” exclusion, TS15 completely miss the point.
This spatial distribution of anions and cations gives rise to the
anion exclusion process that is observed in diffusion experiments.
Now I’m lost. I don’t understand how ion distributions are supposed to cause a process. I think the authors here allude to Schofield’s approach to estimating surface area in montmorillonite suspensions. As discussed in detail in the blog post on anion-accessible porosity, if the suspension is so dilute that we can consider each clay layer independently, and if we equilibrate it with an external solution, we can measure its salt content, and use the Gouy-Chapman model to e.g. estimate surface area from the amount of excluded salt (as compared with the external solution).
But, as also discussed in the blog post on salt exclusion, the “Schofield type” of exclusion is not what we expect to be dominating in a dense system. Rather, in denser systems (and in Donnan systems generally — no surfaces need to be involved), salt exclusion occurs mainly because of charge separation at interfaces with the external solution. I find it revealing that TS15 so far in the article has not at all mentioned such interfaces.
Moreover, in the above sentence TS15 causally states that the anion exclusion process is “observed in diffusion experiments”, without further clarification. Given that the previous section treated diffusion, a reader would expect to have been introduced to the anion exclusion process and how it is observed in diffusion experiments. But this subsection is the first time in TS15 where the term “anion exclusion” is used! In the section on diffusion, “anion accessible porosity” was briefly mentioned, and I suppose a reader is here presumed to connect the dots. But the presence of an exclusion process certainly does not imply an “anion accessible porosity”! Furthermore, anion exclusion is not necessarily observed in diffusion experiments. A more correct statement is that we observe effects of salt exclusion in experiments where a bentonite sample is contacted with an external solution via a semi-permeable component (which typically is a filter that keeps the clay in place). The effect is most conveniently studied in equilibrium rather than diffusion tests, and salt exclusion is not present in e.g. closed-cell diffusion tests. Note that exclusion effects are always related to an external solution.
As the ionic strength increases, the EDL thickness decreases, with
the result that the anion accessible porosity increases as well.
Here it is fully clear that TS15 conflate “anion accessible porosity” and “anion exclusion”. If we consider the “Schofield type” of salt exclusion, it is true that the so-called “exclusion volume” changes with the ionic strength. However, an exclusion volume is not a physical space, but an effective, equivalent quantity. It is derived from the Gouy-Chapman model, which always has anions present everywhere.
Even more importantly, the “Schofield type” of exclusion is not really of interest in dense systems (nor is the Gouy-Chapman model valid in such systems). As discussed above, one must instead consider salt exclusion stemming from charge separation at interfaces with the external solution. For this case it does not even make sense to define an exclusion volume.
I can only interpret this entire paragraph as another fruitless attempt to motivate a multi-porous modeling approach. In this subsection we have so far been told that “two distinct types of porosity” can be defined (they cannot), and we have vaguely been hinted that “bulk or free solution” also is relevant for modelling compacted bentonite. And with the last quoted sentence it is relatively clear that TS15 try to establish that the relative sizes of various “porosities” are controlled by a simple parameter (ionic strength).
The final paragraph of this subsection contain several statements that
makes my jaw drop.
An equivalent anion accessible porosity can be estimated from the
integration of the anion concentration profile (Fig. 6) from the
surface to the bulk water (Sposito 2004)
Here the authors suddenly use the phrase “equivalent”! They are thus obviously aware of that “anion accessible porosity” is a spurious concept?! ?!?! I really don’t know what to say. Their own graph (“Fig. 6”) even show that the Gouy-Chapman model has anions (salt) everywhere! Note that this statement also implies that “the bulk water” is assumed to exist within the clay.
In compacted clay material, the pore sizes may be small as compared
to the EDL size. In that case, it is necessary to take into account
the EDLs overlap between two neighbouring surfaces.
I think this is a very revealing passage. The conditions of compacted bentonite are treated as an exception: pore sizes “may” be smaller than the EDL, and “in that case” it is necessary to account for overlapping diffuse layers. But for compacted bentonite, this is the only relevant situation to consider! Without “overlapping” diffuse layers there is no swelling and no sealing properties. An entire page has been devoted to discussing a model only relevant for suspensions (Gouy-Chapman), while “compacted clay material” here is commented in two sentences…
Clay mineral particles are, however, often segregated into
aggregates delimiting inter-aggregate spaces whose size is usually
larger than inter-particle spaces inside the aggregates.
All of a sudden — in the middle of a paragraph — we are introduced
to a new structural component! “Aggregates” have not been mentioned
earlier in the article and is here introduced without any references.
It is my strong opinion that this way of writing is not appropriate
for a scientific publication, especially not for a review article.
I’m not sure what type of system the authors have in mind here, but
“aggregates” are typically not present in actual water saturated
bentonite. I have commented more on this in
the blog post on stacks.
Conclusion
At the end of the previous section (on diffusion), we were promised that this section should qualitatively link “fundamental properties of the clay minerals” to the diffusional behavior of compacted bentonite. Instead, we are given a fictional description of the structure (conflated with structures of other “clay minerals”), along with a confused explanation of anion exclusion that is irrelevant for such systems. Not a single word is said about the equilibrium that must be considered, namely that at interfaces between bentonite and external solutions. Rather, the idea of “overlapping” diffuse layers — which is the ultimate cause for bentonite swelling — is treated as an exception and only commented on in passing (and nothing is said about how to handle such systems). Although nothing is fully spelled out, I can only interpret this entire part as a (failed) attempt to motivate a multi-porous approach to modeling bentonite. And multi-porosity models cannot be taken seriously.
I admit that scrutinizing studies and pointing out flaws can be fun. However, considering that the descriptions in TS15 are the rule rather than the exception in contemporary bentonite research, I mostly feel weary and resigned. I don’t mean that every clay researcher must agree with me that a homogeneous model is the only reasonable starting point for describing compacted bentonite, and I could only wish that this blog was more influential. But I feel almost dizzy thinking about how this research sector is so hermetically sealed that one can spend entire careers in it without ever having to worry about understanding the nature of swelling and swelling pressure.
Update (250901): Part III of this review is found here.
Footnotes
[1] The Wikipedia article on illite, for example, states that the cation exchange capacity is typically 0.2 — 0.3 eq/kg. Is a significant cation exchange capacity required for classifying something as illite?
[2] E.g. (Poinssot et al., 1999), that TS15 reference as a source on illite, work with sodium exchanged “illite du Puy”, i.e. “Na-illite”.
[3] The material was dispersed by diluting it in alkaline solution and sonicating it. It was thereafter dropped as a suspension on a glass slide and dried.
[4] We may note that the number 5 — 20 TOT-layers in a stack actually showed up when we investigated how this concept is (mis)used in descriptions of bentonite. There it turned out to be a complete misunderstanding of the behavior of suspensions of Ca-montmorillonite.
[5] I am not capable to produce anything reasonable in 3D,
but I think a 2D representation still conveys the message.
[6] Perhaps this criticism can be regarded as nitpicking. I have a nagging feeling, though, that electrostatics is quite poorly understood in certain parts of the bentonite research field. Take the phrase “negative electrostatic potential field”, for example. Although it can be understood at face value (a scalar field with negative values), it also appears to mix together stuff related to charges (“negative”), electric fields (“field”), and potentials. It certainly is important to separate these concepts. There are many examples in the clay literature when this is not done. E.g. Madsen and Müller-Vonmoos (1989) mean that two “potential fields” can repel each other (and also misunderstand swelling)
A high negative potential exists directly at the surface of the
clay layer. […] When two such negative potential fields overlap,
they repel each other, and cause the observed swelling in clay.
[…] the net negative electrical potential
between closely spaced clay particles repel anions attempting to
migrate through the narrow aqueous films of a compact clay […]
In this case, when the clay is compressed […] to the extent that
the electrostatic (diffuse double) layers surrounding the
particles overlap, the overlapping negative potentials repel
invading anions such that the pore becomes excluded to the anion.
[7] An isolated layer of negative charge of course also has an electric field directed towards it, but this is not the relevant system to consider here. (Such a system will actually have an electric field strength that is independent of the distance to the surface, as long as the layer can be regarded as infinitely extended.)
It should go without saying that modelers and model developers must justify every feature, mechanism, or component that they use. Failing to do so strongly increases the risk of being fooled by overparameterization rather than gaining insight. The bentonite scientific literature is nonetheless full of incorrect or unjustified model assumptions, several of which have been discussed previously on the blog. Examples include assuming the presence of bulk water, assuming “stack” structures, and assuming that diffusive fluxes from separate domains are additive. Here we discuss yet another unjustified common model component: Stern layers on montmorillonite basal surfaces.
In the bentonite literature, a Stern layer essentially means a layer of “specifically” sorbed ions on the basal surface, as e.g. illustrated here, in a a figure very similar to what is found in Leroy et al. (2006). Illustrations like this are ubiquitous in the literature.
If montmorillonite basal surfaces function roughly as uniform planes of charge we expect the counter-ions to form a diffuse layer, as e.g. described by the Gouy-Chapman model. By introducing a Stern layer, however, many bentonite researchers mean that exchangeable cations in general also interact with basal surfaces by forming immobile surface complexes. Such interactions necessarily involve mechanisms more “chemical” than the pure electrostatic interaction with uniform planes of charge, and the typical description postulates localized “sorption sites”, as illustrated above.
This blog post treats three different main arguments against Stern layers, presented in different sections
I want to make clear that this criticism concerns one particular type
of surface: the montmorillonite basal surface.
Stern layer models are found in many research fields dealing with
solid interfaces, and although they have been criticized more
generally, here we have no intention of doing so. Likewise, the
process of surface complexation is certainly important generally —
even in bentonite, e.g. on edge surfaces of montmorillonite
particles.1
To better be able to criticize the use of Stern layers on basal surfaces in bentonite modeling, we begin by discussing the origin of Stern’s model.
Origins of the Stern Layer model
The Stern layer concepts were introduced by Otto Stern2 as an extension of the Gouy-Chapman model. Stern’s main concern was metallic electrodes in electrochemical applications. In such systems, the surface (electrode) potential is externally controlled, and can typically be on the order of 1 volt. It is easily seen that the Gouy-Chapman model predicts nonsense for such surface potentials. For e.g. a 1:1 system, the counter-ion concentration at the surface is enhanced by a factor of the order of \(10^{17}\)(!), as seen directly from the Boltzmann distribution \(c^\mathrm{surf} = c^\mathrm{ext}\cdot e^{e\psi^0/kT}\), where \(\psi^0\) is the surface potential, \(e\) is the elementary charge, \(kT\) the thermal energy, and \(c^\mathrm{ext}\) is the concentration far away from the surface. The main problem is that the Gouy-Chapman model does not account for the finite size of ions, and therefore can accumulate an arbitrary amount of charge at the surface. To remedy this flaw, Stern suggested to divide the interface region into a “compact” layer and a “diffuse” part, with the division located an ionic radius from the electrode surface (sometimes referred to as the outer Helmholtz plane).
In the simplest version of Stern’s model the compact layer is free of charges but act as a plate capacitor with a prescribed capacitance (per unit area) \(K_0\). In the original paper Stern shows that, with electrode potential \(\psi^0 = 1\) V and external 1:1 solution concentration \(c^\mathrm{ext} = 1\) M, such a capacitive layer reduces the potential where the diffuse layer begins to \(\psi^1 = 0.08\) V; lowering the external concentration to \(c^\mathrm{ext} = 0.01\) M gives \(\psi^1 = 0.18\) V. For these calculations, Stern uses a value of \(K_0 = 0.29\;\mathrm{F/m^2}\), adopted from measurements on mercury electrodes. This version of the Stern model is essentially a way to take into account that ions cannot get arbitrarily close to the surface.
Stern also presented more elaborate versions of the model that include adsorption in the compact layer (as a Langmuir adsorption model). It is such mechanisms that is universally referred to as a Stern layer in the bentonite scientific literature. Clearly, such versions are substantially more conceptually complex; rather than to just account for a finite ion size at the first molecular layer, we must now also consider additional chemical interactions that typically are different for different types of ions. We also need to have an idea about the adsorption capacity.
Lack of a coherent description of “specific sorption” on
montmorillonite basal surfaces
When using Stern layers for describing montmorillonite basal surfaces, a first thing to note is that the surface potential is not independently controlled for these systems. In contrast to metallic electrodes, montmorillonite is characterized by a fixed surface charge and the problem of accumulating unrealistically large amounts of ions at the interface is significantly mitigated. As pointed by e.g. Norrish and Bolt already in the 1950s, even if we put all counter-ions within the first nanometer adjacent to the surface, the corresponding ion concentration is not larger than approximately 3 M. Here is an illustration of the montmorillonite basal surface on the nm scale, with a representative number of monovalent counter-ions (top layer oxygen atoms are red and the counter-ions blue).3
Clearly, there is room to accommodate all ions without running into the problems that was initially addressed by Stern’s model. Of course, solely accounting for the finite size of the ions — as is done in the simplest version of Stern’s model — is always well justified and will in principle improve the description. In particular, a pure diffuse layer model overestimates the capacitance of the surface. As shown in the table below, the introduction of an empty Stern layer “fixes” this problem.
Here \(c^\mathrm{ext}\) is the concentration of the 1:1 salt far away from the surface, \(\psi^1\) and \(c^1\) denote the electrostatic potential and the counter-ion concentration, respectively, at the point where the diffuse layer begins (i.e. at the interface to the compact layer), and \(\psi^0\) is the electrostatic potential at the surface. \(K_\mathrm{Stern}\) denotes the corresponding capacitance as calculated from the Stern model, with the choice \(K_0 = 0.29 \;\mathrm{F/m^2}\). \(K_\mathrm{DL}\) is instead the capacitance as calculated from a pure diffuse layer model (in which case the surface potential has the value of \(\psi^1\)). In the calculations are assumed a montmorillonite surface charge of 0.111 \(\mathrm{C}/\mathrm{m ^2}\).
But to simply account for the finite size of the ions by means of an empty compact layer is not how the term Stern layer is used in the bentonite scientific literature. As mentioned, most bentonite researchers mean that parts of the rather sparse collection of ions on the surface interacts chemically (“specific sorption”, “chemisorption”). The question of whether Stern layers on montmorillonite basal surfaces are well motivated thus reduces to what arguments there are for more elaborate chemical mechanisms being active on these surfaces. And descriptions of specific sorption on basal surfaces are really all over the place.
Deshpandeand Marshall (1959, 1961) claim that counter-ions are partitioned between (i) chemisorbed ions, which do not contribute to conductivity or activity, (ii) physisorbed ions in a Stern layer, which do not contribute to activity or D.C. conductivity, and (iii) diffuse layer ions, which contribute fully to activity and conductivity. If I interpret their numbers correctly, they state that about 75% — 80% of the counter-ions in pure K-montmorillonite are immobilized. Note that these authors mean that ions in the Stern layer are “physically” adsorbed, while the surface also has “chemically” adsorbed species. Thus, they use the term Stern layer for certain types of physisorption, while stating that ions also bond covalently to the surface.
ShainbergandKemper (1966, 1967), on the other hand, model ions as either mobile in a diffuse layer or immobile in a Stern layer. They argue that covalently bound ions are “extremely unlikely”, and mean that ions in the Stern layer form “ion pairs” with the surface, as suggested earlier by Heald et al. (1964). They use this idea as a starting point for analyzing differences in exchange selectivity for different monovalent cations in montmorillonite. They claim that “about 20 to 50% of sodium are specifically adsorbed”.
Note that Shainberg and Kemper, just like Deshpande and Marshall,
assume Stern layer ions to be “physically” bonded
to the surface (i.e. non-covalently), while having a completely
different opinion on the presence of chemisorbed
ions. Shainberg and Kemper (1966) provide a picture showing the conceptual
difference between an ion in the Stern layer (“unhydrated”) and ions
in the diffuse layer (“hydrated”) that looks very similar to this
In this context it may be worth to also mention the work of
Low
and co-workers. Low
argued consistently that swelling pressure is not primarily related to
the exchangeable ions — something that I strongly disagree with and
that I commented briefly on in a
previous blog post.4 Directly related to this view, these authors
claim that the exchangeable ions for the most part do not dissociate
from the surfaces, and in laterpapers they refer
to such ions as being part of a Stern layer.
To me, all the above descriptions seem like little more than speculation. None of these authors discuss how or why e.g. sodium ions (!) are supposed to from ion pairs with a charge center buried far inside the montmorillonite layer, nor how or why they bond covalently with the basal surface. Nevertheless, both Low as well as Shainberg and Kemper seem to have influenced the writings of Sposito, who, in turn, has had quite a huge impact on contemporary descriptions of bentonite. In e.g. Sposito (1992), which specifically discusses montmorillonite (“smectite”), he writes
Despite the long history of continual investigation of the surface
and colloid chemistry of smectites (van Olphen, 1977; Sposito,
1984), the structure of the electrical double layer at smectite
surfaces and its influence on the rheological properties of smectite
suspensions remain topics of lively controversy. One of the most
contentious issues is the partitioning of adsorbed monovalent
cations among the three possible surface species on the basal planes
of smectite particles, such as montmorillonite (see, e.g., Low,
1981, 1987). […] [A] monovalent cation can be
adsorbed on the basal planes by three different mechanisms:
inner-sphere surface complexes, in which the cation
desolvates and is captured by a ditrigonal cavity;
outer-sphere surface complexes, in which the cation remains
solvated but still is captured by a ditrigonal cavity and
immobilized; and the diffuse-ion swarm, in which the cation
is attracted to the basal plane, but remains fully dissociated from
the smectite surface (Sposito, 1989a, Chap. 7).
The view conveyed here is that exchangeable ions do not interact with montmorillonite basal surfaces as if these, to a first approximation, are planes of uniform charge. Such interaction is only supposed to govern an outer diffuse layer (called a “diffuse-ion swarm” for unclear reasons), and ions are also supposed to interact with the surface by no less than two other “mechanisms”, related to the “ditrigonal cavities”.
Note that while Sposito acknowledges an ongoing “lively controversy” regarding how to describe montmorillonite basal surfaces, he specifies that this debate is limited to how to distribute the exchangeable ions among “three possible surface species.” But, as we will explore below, there is certainly no consensus within colloid chemistry that exchangeable ions are involved in complexation chemistry on the basal surfaces! (I therefore find this way of formulating the “controversy” quite dishonest, to be honest.) For reasons I can’t get my head around, descriptions of “inner-” and “outer-sphere complexes” on montmorillonite basal surfaces are anyway ubiquitous in modern bentonite literature. Let’s therefore take a closer look at how these are introduced.
“Inner-” and “outer-sphere” surface complexes
A description that hardly enlightens me is given in Sposito (1989)5
[The inner- and outer-sphere complexes]
constitute the Stern layer on an adsorbent.
[…]
The diffuse-ion swarm and
the outer-sphere surface complex mechanisms of adsorption involve
almost exclusively electrostatic bonding, whereas inner-sphere
complex mechanisms are likely to involve ionic as well as covalent
bonding. Because covalent bonding depends significantly on the
particular electron configurations of both the surface group and the
complexed ion, it is appropriate to consider inner-sphere surface
complexation as the molecular basis of the term specific
adsorption. Correspondingly, diffuse-ion screening and outer-sphere
surface complexation are the molecular basis for the term
nonspecific adsorption. Nonspecific refers to the weak dependence on
the detailed electron configurations of the surface functional group
and adsorbed ion that is to be expected for the interactions of
solvated species.
Here, Sposito means that exchangeable ions bond covalently with the
montmorillonite basal surface,6 in agreement with Deshpende and Marschall, and in
disagreement with Shainberg and Kemper (we have one “extremely
unlikely” and one “likely” for covalent bonding…). In contrast to
Deshpende and Marschall, however, Sposito means that these
“inner-sphere” complexes are part of the Stern layer. To confuse
matters even more, Shainberg and Kemper assume their “unhydrated”
construct (which corresponds structurally to an “inner-sphere”
complex, see above figure) as being part of the Stern
layer, but not part of any covalent bonding. Moreover,
Shainberg and Kemper assume their “hydrated” construct
(corresponding structurally to an “outer-sphere” complex, see
above figure) to be part of the diffuse layer, while
Sposito wants his “outer-sphere” complexes to be immobile and part
of the Stern layer…
Given the above description (and others) it is hard to understand what
the difference is supposed to be between an “outer-sphere” complex
and an ion in the “diffuse-ion swarm”, other than that the former is
simply claimed to be immobilized; both ions are said to interact with
“exclusively electrostatic bonding”,7 both are classified as “nonspecific
adsorption”, and both are fully hydrated. In my head, this is simply
a recipe for achieving an overparameterized model description.
Sposito’s description also makes implicit statements about the montmorillonite basal surface: it contains “surface functional groups” whose specific electron configuration significantly influence covalent bonding, while being insensitive for the formation of “outer-sphere” complexes and the “diffuse ion-swarm”. In Sposito (1984) he suggests that the “functional groups” are groups of oxygen atoms on the surface of the montmorillonite particle (“ditrigonal cavities”) that qualifies as Lewis bases. The presence of atomic substitutions in the octahedral layer is supposed to enhance this Lewis base character
If isomorphic substitution of \(\mathrm{Al}^{3+}\) by
\(\mathrm{Fe}^{2+}\) or \(\mathrm{Mg}^{2+}\) occurs in the octahedral sheet,
the resulting excess negative charge can distribute itself
principally over the 10 surface oxygen atoms of the four silica
tetrahedra that are associated through their apexes with a single
octahedron in the layer. This distribution of negative charge
enhances the Lewis base character of the ditrigonal cavity and makes
it possible to form complexes with cations as well as with dipolar
molecules.
Note how completely different this whole description is compared to
the original Gouy-Chapman conceptual view. Here is implied that
montmorillonite basal planes8cannot be described as a passive layer of charge,
but that it is a fully reactive system, including covalent
bonding mechanisms.
Frankly, I dismiss the above description of the montmorillonite surface as a Lewis base as pure speculation. I will gladly admit that I am a physicist rather than a chemist, and perhaps I am missing something obvious, but I really don’t see any argumentation behind this description. I am also under the impression that montmorillonite basal planes are relatively chemically stable — that is why they form in the first place, and that is also one reason for why we are interested in using bentonite for e.g. long-term geological waste storage. Furthermore, a meta-argument for dismissing this description is that in later publications we find statements like this, in Sposito (2004):
The \(\mathrm{Na}^+\) that are counterions for the negative structural
charge developed as a result of isomorphic substitutions within the
clay mineral layer tend to adsorb as solvated species on the basal
plane (a plane of hexagonal rings of oxygen ions known as a siloxane
surface) near deficits of negative charge originating in the
octahedral sheet from substitution of a bivalent cation for
\(\mathrm{Al}^{3+}\). This mode of adsorption occurs as a result of
the strong solvating characteristics of Na and the physical
impediment to direct contact between Na and the site of negative
charge posed by the layer structure itself. The way in which this
negative charge is distributed on the siloxane surface is not well
known, but if the charge tends to be delocalized there, that would
also lend itself to outer-sphere surface complexation.
So, 20 years after the surface chemistry of montmorillonite was described as if it was completely understood (Sposito, 1984), the way the negative charge distributes is now described instead as “not well known”… Furthermore, in contrast to earlier statements, the formation of an “outer-sphere complex” is here associated mainly with the hydration properties of sodium. But if the idea of a “surface functional group” is discarded — or at least downplayed — why should a hydrated ion near the surface be described as a surface complex at all?
We note that Sposito (2004) still seem to imply that the “outer-sphere surface complex” is localized and immobile (“adsorbed near deficits of negative charge”) But the evidence is vast that sodium, and several other ions, are quite mobile even in monohydrates (see below).
Deviations from the Gouy-Chapman model do not imply surface complexation
Authors that promote Stern layers on montmorillonite basal surfaces
usually rely on the Gouy-Chapman model for describing the diffuse
layer part. Lyklema,
writing generally on colloid science, explicitly “defends” such an
approach
In the following our discussion will be based on the rather
pragmatic, though somewhat artificial, subdivision of the solution
side of the double layer into two parts: an inner part, or
Stern layer where all complications regarding finite ion
size, specific adsorption, discrete charges, surface heterogeneity,
etc., reside and an outer, Gouy or diffuse
layer, that is by definition ideal, i.e. it obeys
Poisson-Boltzmann statistics. This model is due to Stern following
older ideas of Helmholtz and has over the decades since its
inception rendered excellent services, especially in dealing with
experimental systems.
Bolt and co-workers […] investigated in detail the application of
the Gouy–Chapman diffuse-layer theory to ion-exchange
processes. Their work demonstrated that consideration of
electrostatic sorption alone is not sufficient to explain
ion-exchange data and that chemisorption (or “specific” sorption)
needs to be included in ion-exchange models.
It is not logically consistent to conclude that deviations from the
Gouy-Chapman model implies that specific sorption “needs to be
included”.9 On the contrary, introducing
specific sorption to compensate for a certain model rather than for
surface chemical reasons may, in my mind, be a recipe for an
overparameterized disaster. I don’t get reassured by statements like
this, also from Dzombak and Hudson (1995)
Surface complexation models can be extended to include diffuse-layer
sorption. This approach permits their application in modeling the
sorption of ions (such as monovalent electrolyte ions) that exhibit
weak specific sorption. The generality of such an extended surface
complexation approach together with the mathematical power of modern
chemical speciation models offers the potential for accurate
physicochemical modeling of ion exchange
Reasonably, a complex system may require complex models, but it is
certainly dangerous in a modeling context to rely too heavily on
“mathematical power” (I guess “numerical power” is the preferred
phrase).10
Note that very different attitudes towards the Stern layer concept is found in the colloid science literature, where e.g. Evans and Wennerström (1999) describe it as an “intellectual cul de sac”.11
One way of dealing with these difficulties is to say that the
solution layer closest to a charged surface has properties so
different from the bulk that it should be treated as a separate
entity. This device was introduced in the 1930s by the German
electrochemist Stern and the surface layer is commonly referred to
as the Stern layer, whose properties are specified by a number of
empirical parameters. It is the opinion of the authors of this book
that the Stern layer concept is an intellectual cul de sac for the
description of electrostatics in colloidal systems. One reason for
this point of view is that from modern spectroscopic measurements we
know molecular properties are not dramatically changed for a liquid
close to a charged surface.
I find it quite perplexing that so many authors in the bentonite scientific community attribute any deviation from the Gouy-Chapman model solely to surface-related mechanisms. The Gouy-Chapman model treats both ions (point particles) and water (a dielectric continuum) in a very simplified manner, and it is clear that “specific ion” effects are ubiquitous, also in systems that lack surfaces. Addressing differences in e.g. selectivity coefficients without considering ion polarizability and hydration, while postulating the existence of localized sorption “sites”, can, to my mind, only lead to incorrect descriptions.
The Poisson-Boltzmann equation is a mean field
approximation
Note also that the Poisson-Boltzmann equation — which underlies the Gouy-Chapman model — is only approximate. It is derived by assuming that the electrostatic potential experienced by any ion is the average potential from all other ions (and surfaces). More accurately, the ion distribution around a given ion deviates from the average, as a direct consequence of the presence of the central ion.
Including these ion-ion correlation effects makes the mathematical description considerably more complex. But with the advent of sufficiently powerful computers and algorithms, the electric double layer has been solved basically “exactly”. The “exact” solution may differ strongly from the Poisson-Boltzmann solution, with increasing concentrations towards the surfaces (and consequently a lowering of interlayer midpoint concentrations), and an explicit attractive electrostatic force between the two halves of an interlayer. Using Monte Carlo simulations, Guldbrand et al. (1984) demonstrated that with divalent counter-ions these effects are so large that the system becomes net attractive at a certain interlayer distance, in qualitative disagreement with the Poisson-Boltzmann solution. This effect, which has been thoroughlystudiedsincethe 80s, and which we have discussed in severalposts on this blog, is the prevailing explanation e.g. for the limited swelling of Ca-montmorillonite.
The lesson here is that observed deviations from predictions of the Poisson-Boltzmann equation not automatically can be taken as evidence for additional active system components, and certainly not as evidence for specific sorption. Note that limited swelling in divalent montmorillonite is explained by the ions being diffusive, not that they are sorbed and immobilized. I cannot overstate the importance of this insight.
It boggles my mind that the entire research area on ion-ion correlations in colloidal systems seems to have made no significant impact on parts of the bentonite scientific community; I seldom find references to works on ion-ion correlation, and when I do it’s quite confused. E.g. Sposito (1992) means that the formation of “quasicrystals”12 is due to “outer-sphere” complexes
The best known example of a montmorillonite quasicrystal is that
comprising stacks of four to seven layers. \(\mathrm{Ca}^{2+}\) ions,
solvated by six water molecules (outer-sphere surface complex),
serve as molecular “cross-links” to help bind the clay layers
together through electrostatic forces.
Sure, the ion-ion correlation effect that prevents Ca-montmorillonite from exfoliating is of electrostatic origin, but it is not related to “cross-links” or surface complexes. Sposito furthermore continues by claiming that “even […] Na-montmorillonite” forms “quasicrystals”. Such claim cannot be supported by ion-ion correlation — on the contrary, ion-ion correlation explains why Ca-montmorillonite forms “quasicrystals”, while Na-montmorillonite does not. It is thus relatively clear that Sposito do not refer to ion-ion correlation in the above statement. At the same time, later in the same publication he cites Kjellander et al. (1988) on going beyond the mean-field treatment of the Poisson-Boltzmann equation, and even claims that the Gouy-Chapman model is “completely inaccurate” for systems containing divalent ions. I can only conclude that this is a quite confused description.
Finite-size effects of water molecules
With focus on the first molecular layers at a solid interface, it is
clear that finite-size effects of water molecules — which are
not treated in the Gouy-Chapman model — reasonably influences
the resulting ion distribution. This influence is manifested both as a
steric effect — there can only be a discrete number of water
molecules between the ion and the surface — and as an effect of how
strongly a certain ion is hydrated.
Water molecules are treated explicitly e.g. in molecular dynamics (MD) simulations of montmorillonite/water interfaces, and here are results from simulating a three water-layer interlayer of Na-montmorillonite, from Hedström and Karnland (2012)13
Sodium is seen to accumulate “between” the water layers; in the above illustration we have also included schematic illustrations of the molecular configurations, as conceived by Shainberg and Kemper (1966) (shown earlier). As stated earlier, Shainberg and Kemper (1966) refer to these as “hydrated” and “unhydrated”, but they are clearly the same type of configurations that e.g. Sposito (1992) and Dzombak and Hudson (1995) call “outer-” and “inner-sphere” complexes.
While the above mentioned authors mean that these “complexes” involve specific interactions between ions and surface, the MD simulation suggests that such structures are mainly a consequence of the finite-size of the molecules and ions. In particular, the MD results do not support the idea that these structures depend critically on a specific, non-electrostatic, ion–surface interaction. Indeed, the simulations explicitly treat also the atoms of the montmorillonite layer, which could make it difficult to judge whether the appearance of “complexes” mainly is related to water–ion or ion–surface interactions. But note that Hedström and Karnland (2012) simulate two different systems: one where the montmorillonite charge is put on specific atoms in the octahedral sheet (Mg for Al substitutions), and one where it is distributed on all Al atoms (as a fraction of the elementary charge). Both systems have essentially an identical atomic configuration in the interlayer, which strongly suggest that no critical ion–surface interaction is involved in forming “outer-” and “inner-sphere complexes” (i.e. they really are not surface complexes). I am not aware of any published simulation where the basal surface is represented as a uniform sheet of charge while water molecules are treated explicitly, but I am convinced that “outer-” and “inner-sphere complexes” would appear also in such a simulation.
Regarding MD simulations of montmorillonite interlayers, you can also
simply observe them to convince yourself that the counter-ions
are not in any reasonable sense immobilized. These types of
simulations
are routinely used to calculate (quite significant) interlayer diffusion
coefficients, for crying out loud!
Experimental evidence of counter-ion mobility
A final argument for why Stern layers on montmorillonite basal
surfaces are unjustified is the vast amount of empirical evidence of
counter-ion mobility. We have discussed
severaldiffusion studies in
earlierblogposts that show that many ions (Na, Cl, K, Sr, I, Cs, Ca,…) have
a significant mobility even in very dense systems, dominated by bi- or
monohydrated interlayers. In
the previous post, we brought up the following result
This figure shows the resulting concentration profiles in two
diffusion experiments where sodium and chloride tracers, respectively, have diffused from an initial planar
source for the same amount of time (23.7 h), in samples of pure
Na-montmorillonite of dry density 1.8 \(\mathrm{g/cm^3}\), equilibrated
with deionized water. This result was used previously to dismiss the
ludicrous idea that these two ions are supposed to migrate in separate
parts of the pore volume, exposed to completely different
mechanisms. In the same vein, this result can be used to dismiss the
idea of a Stern layer on basal surfaces.
Sodium, which is universally acknowledged to reside in the interlayers,
is here demonstrated to diffuse just fine in bi- and monohydrated
interlayers. As chloride, which also resides in the interlayers
(despite all talk of
“anion-accessible porosity”), behave essentially identical, it is
quite far-fetched to assume any significant surface complexation
mechanism. And anyone who argues for that these tracers actually do
not diffuse in the interlayers should be reminded of the
seeming “uphill”diffusion experiment,14 which is performed at even higher density, and where
the “uphill” diffusion direction once and for all proves that the
transport occurs in interlayers.
Strangely, manyauthorsnowadays seem to promote both Stern layers and interlayer mobility in bentonite. Various simulation codes has been modified for this possibility, and there are several examples of researchers pointing out a possibility of “Stern layerdiffusion”. I think these authors should carefully examine their chain of assumptions: Surface complexation in a Stern layer (i.e. sorption) is initially suggested to explain e.g. why breakthrough times in cation through-diffusion tests are relatively long as compared with the steady-state flux (i.e. why “\(D_e\)” can be considerably larger than “\(D_a\)”). With evidence for that the “sorbed” ions actually dominate the mass transfer, the sorption mechanism is not reconsidered, but yet another mechanism is suggested: Stern layer mobility… Reasonably, such an approach is not adequate for developing models; researchers employing it should critically consider the intended purpose of a Stern layer component.
Counter-ion mobility is also related to swelling pressure. Bentonite swelling pressure is difficult to describe generally, and I have written a whole series of blog posts on the subject, but it is clear that measured swelling pressures in e.g. moderately dense Na- and Li-montmorillonites is quite well described by the Poisson-Boltzmann equation. As this set of conditions (not too dense clay, simple monovalent ions) are exactly those for which we expect the Poisson-Boltzmann equation to be adequate, this is a strong indication that all counter-ions contribute to the pressure.15 Also, the limited swelling in e.g. Ca-montmorillonite, as previously discussed, is explained by ion-ion correlation effects where all ions are included in the diffuse layer.
Finally, we can take a look at salt exclusion from compacted
bentonite. The magnitude of salt exclusion is
directly related to the amount of mobile counter-ions. Thus, if most
of the counter-ions were immobilized in a Stern layer, bentonite
should show small exclusion effects. In contrast, the empirical
results for e.g. chloride exclusion in sodium dominated bentonite
indicate, again, that all counter-ions are part of a diffuse
layer.
This diagram shows the relative amount of chloride in the bentonite as a function of \(c^\mathrm{ext}/c_\mathrm{IL}\), where \(c^\mathrm{ext}\) is the external salt concentration and \(c_\mathrm{IL}\) is the amount exchangeable cations, expressed as a monovalent interlayer concentration. The experimental data is from Van Loon et al. (2007), which we reevaluated and examined in detail in a previous blog post. The lines are the result from applying the “ideal” Donnan formula with various amounts of the counter-ions assumed diffusive. For details on Donnan theory, see this blog post.
Although the experimental data show considerable scatter, there is nothing in this plot that suggests that a fraction of the counter-ions are immobilized. And the quality of this data is certainly good enough to directly dismiss models that assume that the major part of ions are immobilized in a Stern layer.
[1] I find it quite frustrating that many descriptions
in the literature only refer abstractly to “mineral surfaces”
rather than specifically addressing montmorillonite. At the same
time it is often clear from the context that statements regarding
“mineral surfaces” should be understood as applicable to
montmorillonite basal surfaces. I would much appreciate if
researchers promoting Stern layers on basal surfaces would provide
descriptions for specific systems, e.g. pure Na-Ca-montmorillonite.
[2] Otto Stern is a fascinating character in the history of science, most famous for the Stern-Gerlasch experiment, that helped pave the way for quantum mechanics. I highly recommend this lecture by the late Sandip Pakvasa. An example of its contents:
A note on Stern’s style of working: He always had a cigar in one hand, and he left actual work with hands to others, as he did not trust his own manual dexterity! […] He described the beneficial effects of a large wooden hammer that he kept in his lab and used it to threaten the apparatus if it did not behave! (apparently it worked!)
[3] Note that di-valent counter-ions
would be even more sparsely distributed than this.
[4] It may be worth discussing my
objections against the work of Low and co-workers in more detail in
a future blog post.
[5] This is a general discussion on sorption on mineral surfaces, and is cited from the second edition of the book (2008). There is also a “Thired” edition.
[6] This particular description is
general (for “adsorbents”), but since Sposito, as well as a large
part of the contemporary bentonite scientific community, claim that
“inner-sphere” complexes are present on montmorillonite basal
surfaces, we can conclude that they mean that covalent bonding occur
on such surfaces.
[7] Sure, the full
quotation is “almost exclusively electrostatic bonding”, but what
is a reader supposed to do with that? Such vague and sloppy
scientific writing annoys me.
[8] Again, the discussion is on
general “mineral surfaces”, but from other writings it is clear
that this is supposed to apply to the montmorillonite basal
surface.
[9] I furthermore don’t believe that “Bolt and co-workers” concluded that specific sorption “needs” to be included, but that this rather is an interpretation made by Dzombak and Hudson themselves. Bolt considered and downplayed the Stern Layer already in the mid 50s, and although he indeed has expressed positive attitudes (seriously, these guys just write too much!), he continued to downplay its significance in e.g. Bolt (1979), writing “In conclusion it appears justified to assume that for homoionic clays saturated with common ions, if hydrated, the Stern layer will be an “empty” Stern layer according to the terminology of Grahame (1947).”
[10] Note also that the perspective in this quotation is that specific sorption models can be complemented with diffuse layer features — i.e. the existence of sorption “sites” is assumed a priori. But Dzombak and Hudson (1995) never really discuss the nature of such “sites” on montmorillonite basal surfaces, but rely on Sposito’s speculations about “inner-” and “outer-surface” complexes.
[11] Note that Stern’s original paper actually is from 1924. I also suspect that Stern would object to being labeled an “electrochemist”.
[12] The
terminology here is quite messy, and other authors may use other
terms such as “tactoids”, see
this post for a further discussion.
[14] Anyone making this argument should
also provide a plausible suggestion for where a significant
non-interlayer pore structure is located at these extreme
densities.
[15] The pressure in these types of calculations can be related to the interlayer midpoint concentration. But this does not mean that not all counter-ions are involved in the process.
Here’s an opinion: The compacted bentonite research field is currently
in a terrible state.
After a period away, I’ve recently begun catching up on newly published research in this field. With a fresh perspective, yet still influenced by writing over 30 long-reads over the past years, I can’t help but wonder: what is the problem? Why are a majority of researchers stuck with a view of bentonite1 that essentially makes no sense? And why has this view been the mainstream for decades now?
I get how this might come across: a solitary man ranting on a blog,
criticizing an entire research field in less-than-perfect English. I
probably smell bad and have some wild ideas about why General
Relativity is wrong as well. But what I’m aiming for with this blog
is simply a platform to present an alternative to the mainstream,
primarily because it annoys me as a science-minded person how absurd
this view is.2 I
understand that I will likely struggle to convince anyone who is
already invested in this view, but I’m trying to put myself in the
shoes of e.g. someone entering this field for the first time.
For these reasons, I will try something a little new here: reviewing already published papers. I have touched on this in various forms before, but then usually with a broader topic in mind. Now I intend to critically assess specific publications from the outset. As a first publication to review in this way, I have chosen “Ionic Transport in Nano-Porous Clays with Consideration of Electrostatic Effects” (Tournassat and Steefel, 2015), for the following reasons
It is published in “Reviews in Mineralogy & Geochemistry”, which claims that “The content of each volume consists of fully developed text which can be used for self-study, research, or as a text-book for graduate-level courses.” If anyone aims to learn about ion transport in bentonite from this publication, I would certainly recommend to also consider this review.
It is a quite comprehensive source for many of the claims of the contemporary mainstream view that I have described in earlier blog posts. I guess it makes sense for a publication in “Reviews in Mineralogy & Geochemistry” to reflect the typical view of a research field.
It is published as open access. The article is thus accessible to anyone who wants to check the details.
I will use the abbreviation TS15 in following to refer to this
publication.
Overview
The article covers 38 journal pages (+ references) and includes quite
a lot of topics. At the highest level of headings, the outline look
like this
Introduction (p. 1 — 2)
Classical Fickian Diffusion Theory (p. 2 — 9)
Clay mineral surfaces and related properties (p. 9 — 17)
Constitutive equations for diffusion in bulk, diffuse layer, and interlayer water (p. 17 — 23)
Relative contributions of concentration, activity coefficient and diffusion potential gradients to total flux (p. 24 — 28)
From diffusive flux to diffusive transport equations (p. 28 — 33)
Applications (p. 33 — 37)
Summary and Perspectives (p. 37 — 38)
Given the quite large scope of TS15, I will present this review in
parts, with this first part focusing on the introduction and the
section titled “Classical Fickian Diffusion Theory”.
“Introduction”
I find it remarkable that the authors use terms like “clays” and “clay minerals” when speaking of properties such as “low permeability”, “high adsorption capacity” and “swelling behavior”, and of applications such as nuclear waste storage. I mean that using such general terms here is too broad, as the article focuses solely on systems with swelling/sealing ability. Such an ability is generally connected with a significant cation exchange capacity. Here, I will refer to such systems as “bentonite”, although I am aware that I use the term quite sloppily. But I think this is better than to refer to the components as general “clay minerals” — I don’t think anyone consider it a good idea to e.g. use talc or kaolinite as buffer materials in nuclear waste repositories. Moreover, most of the examples considered in the article are systems that can be described as bentonite. Given the title of the article I also expect a definition of “nano-porous clays”. It is not given here, and the term is actually not used at all in the entire text! (Except one time at the very end.)
After providing a brief overview of the application of (sealing) clay materials, the introduction takes, in my opinion, a rather drastic turn (it happens without even changing paragraphs!).
Clay transport properties are however not simple to model, as they deviate in many cases from predictions made with models developed previously for “conventional” porous media such as permeable aquifers (e.g., sandstone). […] In this respect, a significant advantage of modern reactive transport models is their ability to handle complex geometries and chemistry, heterogeneities and transient conditions (Steefel et al. 2014). Indeed, numerical calculations have become one of the principal means by which the gaps between current process knowledge and defensible predictions in the environmental sciences can be bridged (Miller et al. 2010).
I think the first sentence is too subjective and general. Given the above discussion, here the term “clay transport properties” can cover a million things, if read at face value. Are all of them difficult to model? Also, something does not have to be more difficult just because it deviates from the “convention”. I would argue that several aspects of bentonite actually make it easier to model than, say, sandstone. Advective processes, for example, can often be neglected in compacted bentonite.
I find the statement regarding the advantage of reactive transport
models highly problematic. Not only does it read more like an
advertisement for the authors’ own tools than “fully developed text
for self-study”, but the authors also seem ignorant of issues like
the dangers of overparameterization (a theme that will recur).
“Classical Fickian Diffusion Theory”
As the title of the next section is “Classical Fickian Diffusion Theory,” a reader expects a discussion focused solely on diffusive process, especially when the immediate subtitle reads “Diffusion Basics.” I therefore find it peculiar that this section actually presents the traditional diffusion-sorption model, which describes a combination of diffusion and sorption processes. The model is summarized in eq. 10 in TS15
where \(c\) is the “pore water” concentration of the considered species, \(D_e\) its “effective diffusivity”, \(K_D\) the sorption partition coefficient, \(\rho_d\) dry density, and \(\phi\) porosity.3 For later considerations we also note that TS15 define the denominator on the right hand side as the “rock capacity factor”, \(\alpha = \phi + \rho_dK_D\).
I find it particularly odd that two of the fundamental assumptions of
this specific model are essentially left uncommented, namely that
sorbed ions are immobilized and that the pores contain bulk
water. Instead, the authors appear to question the assumption of
Fickian diffusion in the context of clay systems, i.e. that diffusive
fluxes are assumed proportional to corresponding aqueous concentration
gradients.
This section aims, as far as I can see, to point out shortcomings in the description of diffusion in bentonite, and to motivate further model development. But it should be clear from the outset that using the traditional diffusion-sorption model as the basis for such an endeavor is doomed to fail. The reason for this failure is not due to assuming Fickian diffusion, but due to the other two model assumptions; it has long been demonstrated that exchangeable ions are mobile, and the notion that compacted bentonite contains mainly bulk water is absurd.
After the traditional diffusion-sorption model has been presented, it is evaluated by investigating how it can be fitted to tracer through-diffusion data (this is restatement of original work of Tachi and Yotsjui (2014)). Not surprisingly, it turns out that fitted diffusion coefficients may be unrealistically large. This is of course a direct consequence of the incorrect assumption of immobility in the traditional diffusion-sorption model. TS15 also appear to dismiss the model, saying
This result […] is not physically correct and points out the
inconsistency of the classic Fickian diffusion theory for modeling
diffusion processes in clay media.
I am bothered, though, that they keep using the phrase “classic Fickian diffusion theory”, which inevitably focuses on the Fickian aspect rather than on the obviously incorrect assumptions of the chosen model. Also, rather than simply concluding that the model is incorrect, TS15 continues4
[T]he large changes of \(\mathrm{Cs}^+\) diffusion parameters as a function of chemical conditions (\(D_{e,\mathrm{Cs}^+}\) decreases when the ionic strength increases […]) highlight the need to couple the chemical reactivity of clay materials to their transport properties in order to build reliable and predictive diffusion models.
There is no rationale for such a conclusion. I don’t even completely
understand what “couple the chemical reactivity of clay materials to
their transport properties” mean. Isn’t that what the traditional
diffusion-sorption model attempts? What unrealistic \(D_e\) values
actually highlights is simply that one should not use a model that
assumes immobilization of “sorbed” ions.
However, the experimental observations were completely different: \(^{22}\mathrm{Na}^+\) accumulated in the high NaCl concentration reservoir as it was depleted in the low NaCl concentration reservoir, evidencing non-Fickian diffusion processes.
This is plain wrong. As explained in detail in an earlier post, the diffusion process in the “uphill” test is certainly Fickian. What the test demonstrates is, again, that “sorbed” ions are not immobile.
TS15 also comment on the results of fitting the model to anion tracer through-diffusion data. Here, as is well known, the fitted “rock capacity factor” \(\alpha = \phi + \rho_dK_D\) becomes significantly lower than the porosity \(\phi\). From the perspective of the traditional diffusion-sorption model, this is completely infeasible, as it implies a negative \(K_D\). But rather than simply dismissing the model, TS15 state
The lower \(\alpha\) values for anions than for water indicate that
anions do not have access to all of the porosity.
Also this is incorrect. The porosity5 is an input
parameter rather than a fitting parameter in the traditional
diffusion-sorption model. When claiming that a small value of \(\alpha\)
indicates a decreased porosity, TS15 reinterpret the parameter,
on the fly, in terms of a completely different model:
the effective porosity model. This model has not been mentioned at
all earlier in the article.6
As has been discussed earlier on the blog, the effective porosity model can be fitted to anion tracer through-diffusion data, but now we need to keep track of two different models in the evaluation (something that TS15 do not). Moreover, these two models (the traditional diffusion-sorption and the effective porosity models) are incompatible. But TS15 continue by saying
This result is a first direct evidence of the limitation of the
classic Fickian diffusion theory when applied to clay porous media:
it is not possible to model the diffusion of water and anions with
the same single porosity model. The observation of a lower \(\alpha\)
value for anions than for water led to the development of the
important concept of anion accessible porosity […]
This is a terrible passage. To begin with, the “Fickian” aspect is
also here implied as the problem. But the reason for why the
traditional diffusion-sorption model cannot be fitted to anion tracer
through-diffusion data is of course because this model assumes the
entire pore space to be filled with bulk water. Further, it’s hardly
comprehensible what the authors mean by “it is not possible to model
the diffusion of water and anions with the same single porosity
model”. I think they simply mean that for water you must choose
\(\alpha = \phi\), while for anion through-diffusion you instead must
“choose” \(\alpha < \phi\). But the result \(\alpha < \phi\) should only
lead to the conclusion that the traditional diffusion-sorption model
cannot in any reasonable sense be fitted. A favorable reading of this
passage is to assume that the authors actually mean that the
effective porosity model can only be fitted to anion and water
tracer through-diffusion data by using different values of the
(effective) porosity, and that any “rock capacity factor” should not
appear in this discussion.
Finally, the last sentence gives me headache. Rather than being an “important concept”, I mean that the idea of an “anion accessible porosity” has caused tremendous damage to the development of the bentonite research field for several decades now. We have earlier discussed on the blog that the whole idea of “anion accessible porosity” is based on misunderstandings. We have also demonstrated that the effective porosity model is not valid, even though it can be fitted to anion tracer through-diffusion data. A simple way to see this is to consider closed-cell diffusion data rather than through-diffusion data. Closed-cell tests are simpler than through-diffusion tests, as they don’t involve interfaces between clay and external solutions. We can e.g. take a look at the vast amount of diffusion coefficients for chloride in montmorillonite, presented in Kozaki et al. (1998).
There are in total 55(!) values, corresponding to 55(!) separate tests. These have been systematically varied with respect to density and temperature, but all of them were performed on montmorillonite equilibrated with distilled water. From the perspective of the effective porosity model, the effective porosity in such a system should be minute, perhaps even strictly zero; effective porosities evaluated from chloride through-diffusion tests are well below 1% even at a background concentration as large as 10 mM. Thus, if the idea of “anion accessible porosity” was reasonable, we’d expect extremely low values of the chloride diffusion coefficient in the above plot.7 We’d perhaps also expect a threshold behavior, where chloride diffusivity basically vanishes above a certain density. But this is not at all the behavior: chloride is seen to diffuse just fine in all 55(!) tests, with temperature- and density dependencies that seems reasonable for a homogeneous system. Moreover, chloride behaves very similarly to e.g. sodium, as seen here
Here the sodium data is from Kozaki et al. (1998),8 and it has also been measured in montmorillonite equilibrated with distilled water.
The effective porosity model and the notion of “anion accessible porosity” can consequently be dismissed directly, by comparing with simpler tests than what is done in TS15. The reason that the effective porosity model can be fitted to anion through-diffusion data must be attributed to a misinterpretation of such tests, as they involve also interfaces to external solutions. At least to me it is completely clear that what many researchers interpret as an effective porosity is actually effects of interface equilibrium.
If TS15 were serious about evaluating bentonite diffusion processes in
this section I think they should have done the following:
Discuss the assumptions of ion immobility of sorbed ions and bulk pore water when presenting the traditional diffusion-sorption model. Moreover, they should not call this “Classical Fickian Diffusion Theory”.
Also present and discuss the effective porosity model, as they obviously use it in their evaluations. They actually even seem to promote it! And it is as “Fickian” as the traditional diffusion-sorption model.
Evaluate the models using closed-cell data to avoid misinterpretations arising from complications at bentonite/external solution interfaces.
Conclude that the traditional diffusion-sorption model is not valid for bentonite, and that this is because of the assumptions of immobility of sorbed ions and bulk pore water.
Conclude that the effective porosity model is not valid for bentonite, and that the notion of “anion accessible porosity” is flawed.
Instead, we get a quite confused and incomplete description, mixed
with entirely inaccurate statements. In the end, it is difficult to
understand what the takeaway message of this section really is. A
reader is left with an impression that there is some problem with the
“Fickian” aspect of diffusion, but nothing is spelled out. We have
also been hinted that “anion accessible porosity” is important,
without really having been introduced to the concept/model.
The section ends with the following passage
The limitations of the classic Fickian diffusion theory must find
their origin in the fundamental properties of the clay minerals. In
the next section, these fundamental properties are linked
qualitatively to some of the observations described above.
If “classic Fickian diffusion theory” here is interpreted as “the traditional diffusion-sorption model” (which is literally what has been presented), the first sentence is both incorrect and trivial at the same time. The traditional diffusion-sorption model does not have “limitations” — it is fundamentally incorrect as a model for bentonite. The reason for this is that exchangeable ions are not immobile and that bentonite does not contain significant amounts of bulk water. Both of these reasons can be linked to “fundamental properties” of some specific clay minerals.
But it is clear that TS15 also have vaguely promoted the concept of “anion accessible porosity” and the effective porosity model. Are these not included in “the classic Fickian diffusion theory”? If not, why then is a model that assumes sorption and immobilization?
How can it not be immediately obvious to everyone that the
diffusion process is much simpler than the contemporary descriptions?
As we have brought up the data from Kozaki etal. (1998), I would like to end this blog post by further considering actual profiles of chloride and sodium diffusing in montmorillonite.
This figure shows the corresponding normalized concentration profiles after 23.7 hours in closed-cell tests performed at \(50\;^\circ\mathrm{C}\) in Na-montmorillonite at dry density \(1.8 \;\mathrm{g/cm^3}\) that has been equilibrated with distilled water. In the case of sodium, both the profile evaluated from Fick’s second law (orange line) and measured values (circles) are plotted. In the case of chloride, no measured values are available, but the value of the diffusion coefficient is the result of fitting Fick’s second law (green line) to such data.
From the perspective of the traditional diffusion-sorption model, the
sodium profile is supposed to represent the combined result of ions
diffusing in bulk water, at a rate many orders of magnitude larger
than in pure water, while being strongly retarded due to sorption onto
“the solid” (where the ions are immobile). This is clearly nonsense,
and something that I think TS15 actually tries to communicate.
From the perspective of the effective porosity model, on the other
hand, the chloride profile is supposed to be the result of the ion
diffusing in an essentially infinitesimal fraction of the pore volume,
which magically is perfectly interconnected in all samples on which such
tests are conducted. This is of course just as nonsensical as
the above interpretation of the sodium profile, but in this case TS15
appear to promote the model (the “important concept of anion
accessible porosity”).
Note that these two simple ions, at the end of the day, diffuse very similarly (please stop reading for a moment and contemplate the above plot). If sodium and chloride actually migrate in completely different domains and are subject to completely different physico-chemical processes, this “coincident” would be more than a little weird. Especially given that the two ions show similar diffusive behavior across a wide range of densities. To me, this simple observation makes it evident that ion diffusion in bentonite at the basic level is much simpler than what is suggested by the contemporary mainstream view. I mean that it is completely obvious that all ions in bentonite diffuse in the same type of quite homogeneous domain. And since it cannot be argued that the pore volume is dominated by anything other than interlayers at 1.8 g/cm³, this homogeneous domain is the interlayer domain at any relevant density. The evidence has been available for at least 25 years (in fact much longer than that). How can this be difficult to grasp?
Update (250213): Part II of this review is found here.
Footnotes
[1] By “bentonite” I here mean any type of smectite-rich system with a significant cation exchange capacity.
[2] The irony is that the “alternative” in a
broader perspective is more mainstream than the “mainstream”
view. I basically propose to obey the laws of thermodynamics.
[3] I have simplified the notation here somewhat compared with how it is written in TS15. As many others, TS15 call this equation “Fick’s second law” (via their eq. 4), which is not correct. Fick’s laws refer strictly to pure diffusion processes. However, the equation has the same form as Fick’s second law, if \(D_e/(\phi + \rho_d K_D)\) is treated as a single constant (often referred to as the apparent diffusivity).
[4] This behavior is of course not
unique for cesium; I don’t know why TS15 focus so hard on that ion
here.
[5] “Porosity” is a volume ratio. I’m not a fan of that the word has also begun to mean “pore space” in the bentonite scientific literature.
[6]
In fact, \(\alpha\) has earlier in the article been unambiguously
related to sorption:
If the species \(i\) is also adsorbed on or
incorporated into the solid phase, then it is possible to define a
rock capacity factor \(\alpha_i\) that relates the concentration in
the porous media to the concentration in solution
[7] That the diffusivity is much too large for an effective porosity interpretation to make sense can also be seen from invoking Archie’s law, which is quite popular in bentonite scientific papers.
Here \(D_0\) is the diffusivity in pure bulk water, which is about \(2\cdot 10^{-9} \;\mathrm{m^2/s}\) for chloride. Using the popular choice \(n \approx 2\) and choosing e.g. \(\epsilon_\mathrm{eff} = 0.001\) (most probably an overestimation when using distilled water), we get
This is more than twenty times the actual value for \(D_0\). (\(D_e = 5.1\cdot 10^{-14} \;\mathrm{m^2/s}\) is evaluated from Kozaki’s data at \(1.4 \;\mathrm{g/cm^3}\) and \(25\;^\circ\mathrm{C}\))
[8] Note! This publication is different from the
chloride study.
The paper follows a structure recognizable from several others that we have considered previously on the blog: It starts off with an introduction section containing several incorrect or unfounded statements1 regarding bentonite.2 It then presents some experimental results that makes it evident that no real progress has been made for a long time regarding e.g. experimental design.3 The major part of the paper is devoted to a “results and discussion” section with several incorrect statements and inferences, speculation, and irrelevant modeling.
[I]nfluence of a background electrolyte concentration gradient on the diffusion of anionic and cationic species at trace concentrations has […] been rarely investigated. Notable exceptions are the DR-A in situ diffusion experiment conducted at the Mont-Terri laboratory (Soler et al., 2019), and an “uphill” diffusion experiment of a \(^{22}\mathrm{Na}^+\) tracer in a compacted sodium montmorillonite (Glaus et al., 2013). These two studies demonstrated the marked influence of background electrolyte concentration gradient on tracer diffusion, and thus the necessity to understand the couplings between diffusion of several charged species present at contrasting concentrations and experiencing different concentration gradients. The experiment from Glaus et al. (2013) also demonstrated the importance of considering diffusion processes occurring in the porosity next to the charged surface of clay minerals (i.e., the porosity associated to the EDL of particles).
This quotation contains two statements relating to Glaus et al. (2013), both of which I think are problematic4
It basically claims that the “uphill” phenomenon is due to diffusive couplings between several types of ions. Of course, ion diffusion always involves couplings between different types of ions, due to the requirement of electroneutrality. But it is clear that Tertre et al. (2024) mean that the “uphill” effect is caused by additional couplings that are not present in chemically homogeneous systems.
It says that Glaus et al. (2013) demonstrates the importance to consider diffuse layers. I agree with this, but it is written in a way that implies that there also are other relevant “porosities”, and that there are other types of tests where ion diffusion in bentonite is not significantly influenced by the presence of diffuse layers.
As one of the authors of the “uphill” study, I would here like to argue for why I think the above statements are problematic and give some background context.
The “uphill” diffusion experiment
The “uphill” study actually originated from a prediction presented by me in a conference poster session. This poster discussed the role of the quantity \(D_e\), using the exact same theory that we had previously used to explain the diffusive behavior of tracer ions in compacted bentonite as an effect of Donnan equilibrium in a homogeneous system. In particular, it pointed out that \(D_e\) — although universally referred to as the (effective) “diffusion coefficient” — is not a diffusion coefficient in the context of compacted bentonite. I have continued this discussion in laterpapers, and in several postson this blog.
In the poster, we suggested the “uphill” experiment as a demonstration of the shortcoming of \(D_e\). If the two reservoirs in a through-diffusion test are maintained at different background concentrations, the theory predicts a non-zero tracer flux for a vanishing external tracer concentration difference, i.e. an “infinite” value of \(D_e\). The suggestion caught the interest of an experimental group, and after a successful collaboration we could present the results of an actual “uphill” experiment. Without making too much of an exaggeration, I would say that the results of this experiment were basically exactly as predicted.
Given this background, it should be clear that the tests in Glaus et al. (2013) follow exactly the same rules as tests in chemically homogeneous systems, rather than demonstrating “the necessity to understand the couplings between diffusion of several charged species present at contrasting concentrations”. Although it is quite clearly stated already in the abstract in Glaus et al. (2013), there is apparently still a need to communicate this explanation. Let me therefore try that here.
The “uphill” diffusion phenomenon explained
Consider an ordinary aqueous solution containing radioactive \(^{22}\mathrm{Na}\) and stable \(^{23}\mathrm{Na}\). The fraction of \(^{22}\mathrm{Na}\) ions can be written \(c_\mathrm{ext}/C_\mathrm{bkg}\), where \(c_\mathrm{ext}\) is the \(^{22}\mathrm{Na}\) concentration, and \(C_\mathrm{bkg}\) is the total sodium concentration (the “tracer” and “background” concentrations, respectively).
Since \(^{23}\mathrm{Na}\) and \(^{22}\mathrm{Na}\) are basically chemically indistinguishable, the same \(^{22}\mathrm{Na}\)-fraction will be maintained in any system with which this solution is in equilibrium. In particular, if the solution is in equilibrium with a montmorillonite interlayer solution, we can write
where \(c_\mathrm{int}\) and \(C_\mathrm{int}\) are the \(^{22}\mathrm{Na}\) and total interlayer concentrations, respectively. The total interlayer cation concentration (\(C_\mathrm{int}\)) can be handled in different ways, but it is important to note that this is a substantial number under all conditions, relating to the cation exchange capacity.5 Rearranging eq. 1 gives
Since the interlayer cation concentration is always larger than the corresponding background concentration, the above equation tells us that the corresponding interlayer tracer concentration becomes enhanced, by the factor \(C_\mathrm{int}/C_\mathrm{bkg}\).
Conventional through-diffusion
This enhancement mechanism causes the diffusional behavior of \(^{22}\mathrm{Na}\) in conventional through-diffusion experiments in bentonite. In such experiments, the tracer concentration in the target reservoir is usually kept near zero, and the actual steady-state concentration gradient in the interlayers is
where we have indexed the tracer concentration in the source reservoir with “\((1)\)”, labeled the sample length \(L\), and assumed that ions diffuse in the \(x\)-direction. The corresponding flux is thus (Fick’s law)
where \(D_c\) denotes the (macroscopic) diffusivity in the interlayers, and \(\phi\) is porosity. Keeping \(c_\mathrm{ext}^{(1)}\) constant, eq. 2 shows that the \(^{22}\mathrm{Na}\) steady-state flux increases indefinitely as the background concentration is made small, in full agreement with experimental observation.6
The picture below illustrates the concentration conditions in an conventional through-diffusion test.
Here we have chosen \(C_\mathrm{int}=\) 4.0 M, the background concentration in the two reservoirs (blue) is put equal to 0.1 M, and the tracer concentration (orange) is put to 0.1 mM in reservoir 1 (and zero i reservoir 2). The corresponding internal tracer gradient is plotted in the right side diagram, and the resulting diffusive flux is indicated by the arrow.
“Uphill” diffusion
To explain the “uphill” effect the only modifications needed in the above derivation is to allow for different background concentrations in the external reservoirs, and to recognize that the tracer concentration in the clay on the “target” side (indexed “\((2)\)”) no longer is zero. Considering the tracer concentration enhancement at both interfaces, the steady-state interlayer concentration gradient then reads
To be more concrete, let’s assume that \(C_\mathrm{bkg}^{(2)} = 5\cdot C_\mathrm{bkg}^{(1)}\), which is the same ratio as in Glaus et al. (2013). We then have
Note that we recover the conventional through-diffusion result (eq. 2) from this expression, if we put \(c_\mathrm{ext}^{(2)}= 0\). But if we e.g. set the tracer concentration equal in both reservoirs, we still have a flux from side \((1)\) to side \((2)\), of size \(j = 4/5 \cdot \phi D_c\cdot C_\mathrm{int}/C_\mathrm{bkg}^{(1)}\cdot c_\mathrm{ext}^{(1)}\). And even if we make \(c_\mathrm{ext}^{(2)}\) larger than \(c_\mathrm{ext}^{(1)}\) — as long as \(c_\mathrm{ext}^{(1)}< c_\mathrm{ext}^{(2)} < 5\cdot c_\mathrm{ext}^{(1)}\) — we still have a diffusive flux from side \((1)\) to side \((2)\), i.e seeming “uphill” diffusion.
Below is illustrated the concentration conditions in an “uphill” configuration.
In contrast to the above illustration for conventional through-diffusion, the background concentration in reservoir 2 is here raised to 0.5 M and the tracer concentration in reservoir 2 is put equal to 0.2 mM. We see that, although tracers are transported to the reservoir with higher concentration, the process is still ordinary Fickian diffusion, as the internal tracer gradient has the same direction as in the conventional case.
We can now conclude what was stated above: The “uphill” diffusion effect is caused by exactly the same mechanism that cause the behavior of cation diffusion in conventional bentonite through-diffusion tests. This mechanism is ion equilibrium between clay and external solutions at the two interfaces. In this particular case, with sodium tracers diffusing in a sodium background, we don’t need to invoke the full ion equilibrium framework in order to quantify the fluxes, but can rely on the very robust result that any two systems in equilibrium have the same tracer fraction (eq. 1).
Reexamining the Tertre et al. (2024) statements
With the explanation for the “uphill” effect established, let’s re-examine the problematic statements in Tertre et al. (2024) identified above
Glaus et al. (2013) cannot be used to support a claim of “marked influence” of additional diffusional couplings. The opposite is true: Glaus et al. (2013) found no significant influence from mechanisms beyond those in chemically homogeneous conditions.
The “uphill” effect was predicted from taking the idea seriously that diffusion in compacted bentonite is fully governed by interlayer properties. Singling out Glaus et al. (2013) as the study that demonstrates the importance of diffuse layers7 therefore gives the wrong impression. Rather, what Glaus et al. (2013) demonstrates, in conjunction with corresponding conventional through-diffusion results, is that compacted bentonite contains insignificant amounts of bulk water (what Tertre et al. (2024) call “interparticle water”).
A way forward (if anybody cares)
After the uphill study was published I was for a while under the illusion that things would begin to change within the compacted bentonite research field. Not only did the study, to my mind, deal a fatal blow to any bentonite model that relies on the presence of a bulk water phase in the clay. It also opened up a whole new area of interesting studies to conduct. Now, some 11 years later, I can disappointingly conclude that not a single additional study has been presented that explore the ideas here discussed.8 And, regarding bentonite models, bulk water is apparently alive and kicking, as has been discussed ad nauseum on this blog.
Experimentally, there are a number of interesting questions looking for answers. In particular, we actually do expect additional mechanisms to play a role in chemically inhomogeneous systems, e.g. osmosis, and other effects due to presence of salt concentration gradients and electrostatic potential differences. It may be argued for why such effects are not significant in Glaus et al. (2013), but it is of course both of fundamental and practical interest to understand under which conditions they are. The original “uphill” study is e.g. performed at quite extreme density (\(1900\;\mathrm{m^3/kg}\)). How would the result differ at \(1600\;\mathrm{m^3/kg}\) or \(1300\;\mathrm{m^3/kg}\)? Also, how would the results change with other choices of the reservoir concentrations, and how would the results differ if one of the cations is not at trace level (e.g. a system with comparable amounts of sodium and potassium)?
Even under the conditions of the original study, there are several predictions left to verify. If e.g. \(c^{(1)}_\mathrm{ext} = c^{(2)}_\mathrm{ext}/5\), the theory predicts zero flux (implying \(D_e = 0\)). The theory also implies that when performing “conventional” through-diffusion, the actual level of the background concentration in the target reservoir is irrelevant, as long as the tracer concentration is kept at zero.
In fact, one can imagine making a whole cycle of through-diffusion tests to explore the ideas here discussed, as illustrated in this animation
The resulting steady-state flux for various external conditions is indicated by the arrow. Here, the full ion equilibrium framework was used to calculate the internal concentrations (giving an internal gradient also in \(C_\mathrm{int}\)). Background concentrations and total interlayer concentration is chosen to be comparable with Glaus et al. (2013), while the choice for tracer concentration is arbitrary.9
With the risk of sounding hubristic, the number of experiments suggested in the above animation could have given enough material for several Ph.D. theses. But here we are, in the year 2024, without even a replication of the “uphill” effect. Instead, a basically entire research field has been stuck for decades with the ludicrous idea that models of compacted bentonite should be based on a bulk water description. I find this both hilarious and horrific.
Update (260803):A further treatment of the “uphill” test with focus on the electrostatic potential is found here. Update (261006):The “uphill” test is modeled with the full Nernst-Planck framework and the homogeneous mixture modelhere.
Footnotes
[1] For example (follow links to discussions on these issues):
Related to using the traditional diffusion-sorption model, it assumes \(D_e\) to be a real diffusion coefficient, which it is not. I find this particularly remarkable in a paper that deals with the presence of “saline gradients”. A motivation behind e.g. the “uphill” test is to point out the shortcomings of \(D_e\), as discussed in the rest of this blog post.
It claims that “anionic and cationic tracers do not experience the same overall accessible porosity”, which is unjustified.
It claims that “diffusion rates” of anions are decreased and “diffusion rates” of cations are increased, compared to “neutral species”, due to different interactions with diffuse layers. But this is not true generally.
[2] I use the word “bentonite” here quite loosely. Tertre et al. (2024) use wordings such as “clayey samples”, “argillaceous rocks” and “clayey formation”, but it is clear that the presented material is supposed to apply to actual bentonite.
[3] I’m specifically thinking about that cation tracer through-diffusion tests at low background concentration is not a good idea, and that it is completely clear from the results presented in Tertre et al. (2024) that some of these are mainly controlled by diffusion in the confining filters. Estimating a “rock capacity factor” larger than 750 for sodium tracers in a sodium-clay (at 20 mM background concentration) should have set off all alarm bells.
[4] Regarding Soler et al. (2019), I think that whole study is problematic, which I might argue for in a separate blog post.
[5] Glaus et al. (2013) invoke the “exchange site” activity \([\mathrm{NaX}]\) to discuss this quantity. I personally prefer relating it to the quantity \(c_\mathrm{IL}\) that is defined within the homogeneous mixture model.
[6] This agreement has been shown to be quantitative, see e.g. Glaus et al. (2007), Birgersson and Karnland (2009) and Birgersson (2017). Note that this result is quite independent of how many “porosities” you choose to include in a model; it’s merely a consequence of treating the dominating pores (interlayers) adequately. Further, note that measuring the diverging fluxes in the limit of low background concentration becomes increasingly difficult, as the confining filters becomes rate limiting.
[7] In the present context, I presume the terms “diffuse layer” and “interlayer” to be more or less equivalent. Other authors instead make an unjustified distinction, that I have addressed here.
[8] There are afewexamples of published studies where effects of the kind discussed here are present, but where the authors don’t seem to be aware of it.
[9] Tracer concentrations in Glaus et al. (2013) is much smaller, but this value does not affect any behavior, as long as it is small in comparison with total concentration.
Vl07 is centered around a set of through-diffusion tests in “KWK” bentonite samples of nominal dry densities 1.3 g/cm3, 1.6 g/cm3, and 1.9 g/cm3. For each density, chloride tracer diffusion tests were conducted with NaCl background concentrations 0.01 M, 0.05 M, 0.1 M, 0.4 M, and 1.0 M. In total, 15 samples were tested. The samples are cylindrical with diameter 2.54 cm and height 1 cm, giving an approximate volume of 5 cm3. We refer to a specific test or sample using the nomenclature “nominal density/external concentration”, e.g. the sample of density 1.6 g/cm3 contacted with 0.1 M is labeled “1.6/0.1”.
After maintaining steady-state, the external solutions were replaced
with tracer-free solutions (with the same background concentration),
and tracers in the samples were allowed to diffuse out. In this way,
the total tracer amount in the samples at steady-state was
estimated. For tests with background concentrations 0.01 M, 0.1 M, and
1.0 M, the outflux was monitored in some detail, giving more
information on the diffusion process. After finalizing the tests, the
samples were sectioned and analyzed for stable (non-tracer)
chloride. In summary, the tests were performed in the following
sequence
Saturation stage
Through-diffusion stage
Transient phase
Steady-state phase
Out-diffusion stage
Sectioning
Uncertainty of samples
The used bentonite material is referred to as “Volclay KWK”. Similar to “MX-80”, “KWK” is just a brand name (it seems to be used mainly in wine and juice production). In contrast to “MX-80”, “KWK” has been used in only a fewresearchstudies related to radioactive waste storage. Of the studies I’m aware, only Vejsada et al. (2006) provide some information relevant here.1
Vl07 state that “KWK” is similar to “MX-80” and present a table with chemical composition and exchangeable cation population of the bulk material. As the chemical composition in this table is identical to what is found in various “technical data sheets”, we conclude that it does not refer to independent measurements on the actual material used (but no references are provided). I have not been able to track down an exact origin of the stated exchangeable cation population, but the article gives no indication that these are original measurements (and gives no reference). I have found a specification of “Volclay bentonite” in this report from 1978(!) that states similar numbers (this document also confirms that “MX-80” and “KWK” are supposed to be the same type of material, the main difference being grain size distribution). We assume that exchangeable cations have not been determined explicitly for the material used in Vl07.
In a second table, Vl07 present a mineral composition of “KWK”, which I assume has been determined as part of the study. But this is not fully clear, as the only comment in the text is that the composition was “determined by XRD-analysis”. The impression I get from the short material description in Vl07 is that they rely on that the material is basically the same as “MX-80” (whatever that is).
Montmorillonite content
Vl07 state a smectite content of about 70%. Vejsada et al. (2006), on the other hand, state a smectite content of 90%, which is also stated in the 1978 specification of “Volclay bentonite”. Note that 70% is lower and 90% is higher than any reported montmorillonite content in “MX-80”. Regardless whether or not Vl07 themselves determined the mineral content, I’d say that the lack of information here must be considered when estimating an uncertainty on the amount of montmorillonite (“smectite”) in the used material. If we also consider the claim that “KWK” is similar to “MX-80”, which has a documented montmorillonite content in the range 75 — 85%, an uncertainty range for “KWK” of 70 — 90% is perhaps “reasonable”.
Cation population
Vl07 state that the amount exchangeable sodium is in the range 0.60 — 0.65 eq/kg, calcium is in the range 0.1 — 0.3 eq/kg, and magnesium is in the range 0.05 — 0.2 eq/kg. They also state a cation exchange capacity in the range 0.76 — 1.2 eq/kg, which seems to have been obtained from just summing the lower and upper limits, respectively, for each individual cation. If the material is supposed to be similar to “MX-80”, however, it should have a cation exchange capacity in the lower regions of this range. Also, Vejsada et al. (2006) state a cation exchange capacity of 0.81 eq/kg. We therefore assume a cation exchange capacity in the range 0.76 — 0.81, with at least 20% exchangeable divalent ions.
Soluble accessory minerals
According to Vl07, “KWK” contains substantial amounts of accessory carbonate minerals (mainly calcite), and Vejsada et al. (2006) also state that the material contains calcite. The large spread in calcium and magnesium content reported for exchangeable cations can furthermore be interpreted as an artifact due to dissolving calcium- and magnesium minerals during the measurement of exchangeable cations (but we have no information on this measurement). Vl07 and Vejsada et al. (2006) do not state any presence of gypsum, which otherwise is well documented in “MX-80”. I do not take this as evidence for “KWK” being gypsum free, but rather as an indication of the uncertainty of the composition (the 1978 specification mentions gypsum).
Sample density
Vl07 don’t report measured sample densities (the samples are ultimately sectioned into small pieces), but estimate density from the water uptake in the saturation stage. The reported average porosity intervals are 0.504 — 0.544 for the 1.3 g/cm3 samples, 0.380 — 0.426 for the 1.6 g/cm3 samples, and 0.281 — 0.321 for the 1.9 g/cm3 samples. Combining these values with the estimated interval for montmorillonite content, we can derive an interval for the effective montmorillonite dry density by combining extreme values. The result is (assuming grain density 2.8 g/cm3, adopted in Vl07).
Sample density (g/cm3)
EMDD interval (g/cm3)
1.3
1.04 — 1.32
1.6
1.36 — 1.67
1.9
1.67 — 1.95
These intervals must not be taken as quantitative estimates, but as giving an idea of the uncertainty.
Uncertainty of external solutions
Samples were water saturated by first contacting them from one side with the appropriate background solution (NaCl). From the picture in the article, we assume that this solution volume is 200 ml. After about one month, the samples were contacted with a second NaCl solution of the same concentration, and the saturation stage was continued for another month. The volume of this second solution is harder to guess: the figure shows a smaller container, while the text in the figure says “200 ml”. The figure shows the set-up during the through-diffusion stage, and it may be that the containers used in the saturation stage not at all correspond to this picture. Anyway, to make some sort of analysis we will assume the two cases that samples were contacted with solutions of either volume 200 ml, or 400 ml (200 ml + 200 ml) during saturation.
The through-diffusion tests were started by replacing the two saturating solutions: on the left side (the source) was placed a new 200 ml NaCl solution, this time spiked with an appropriate amount of 36Cl tracers, and on the right side (the target) was placed a fresh, tracer free NaCl solution of volume 20 ml. The through-diffusion tests appear to have been conducted for about 55 days. During this time, the target solution was frequently replaced in order to keep it at a low tracer concentration. The source solution was not replaced during the through-diffusion test.
As (initially) pure NaCl solutions are contacted with bentonite that contains significant amounts of calcium and magnesium, ion exchange processes are inevitably initiated. Thus, in similarity with some of the earlierassessed studies, we don’t have full information on the cation population during the diffusion stages. As before, we can simulate the process to get an idea of this ion population. In the simulation we assume a bentonite containing only sodium and calcium, with an initial equivalent fraction of calcium of 0.25 (i.e. sodium fraction 0.75). We assume sample volume 5 cm3, cation exchange capacity 0.785 eq/kg, and Ca/Na selectivity coefficient 5.
Below is shown the result of equilibrating an external
solution of either 200 or 400 ml with a sample of density 1.6 cm3/g,
and the corresponding result for density 1.3 cm3/g and external volume
400 ml. As a final case is also displayed the result of first
equilibrating the sample with a 400 ml solution, and then replacing it
with a fresh 200 ml solution (as is the procedure when the
through-diffusion test is started).
Although the results show some spread, these simulations make it relatively clear that the ion population in tests with the lowest background concentration (0.01 M) probably has not changed much from the initial state. In tests with the highest background concentration (1.0 M), on the other hand, significant exchange is expected, and the material is consequently transformed to a more pure sodium bentonite. In fact, the simulations suggest that the mono/divalent cation ratio is significantly different in all tests with different background concentrations.
Note that the simulations do not consider possible dissolution of accessory minerals and therefore may underestimate the amount divalent ions still left in the samples. We saw, for example, that the material used in Muurinen et al. (2004) still contained some calcium and magnesium although efforts were made to convert it to pure sodium form. Note also that the present analysis implies that the mono/divalent cation ratio probably varies somewhat in each individual sample during the course of the diffusion tests.
Direct measurement of clay concentrations
Chloride
clay concentration profiles were measured in all samples after
finishing the diffusion tests, by dispersing sample sections in
deionized water. Unfortunately, Vl07 only present this chloride
inventory in terms of “effective” or
“Cl-accessible porosity”, a concept often encountered in
evaluation of diffusivity. However, “effective porosity” is
not what is measured, but is rather an interpretation of
the evaluated amount of chloride in terms of a certain pore volume
fraction. Vl07 explicitly define effective porosity as
\(V_\mathrm{Cl}/V_\mathrm{1g}\), where \(V_\mathrm{1g}\) is the “volume
of a unit mass of wet bentonite”, and \(V_\mathrm{Cl}\) is the “volume
of the Cl-accessible pores of a unit mass of bentonite”. While
\(V_\mathrm{1g}\) is accessible experimentally, \(V_\mathrm{Cl}\) is
not. Vl07 further “derive” a formula for the effective porosity
(called \(\epsilon_\mathrm{eff}\) hereafter)
where \(n’_\mathrm{Cl}\) is the amount chloride per mass bentonite, \(\rho_\mathrm{Rf}\) is the density of the “wet” bentonite, and \(C_\mathrm{bkg}\) is the background NaCl concentration.2 In contrast to \(V_\mathrm{Cl},\) these three quantities are all accessible experimentally, and the concentration \(n’_\mathrm{Cl}\) is what has actually been measured. For a result independent of how chloride is assumed distributed within the bentonite, we thus multiply the reported values of \(\epsilon_\mathrm{eff}\) by \(C_\mathrm{bkg}\), which basically gives the (experimentally accessible) clay concentration
Here we also have divided by sample porosity, \(\phi\), to relate the clay concentration to water volume rather than total sample volume. Note that eq. 2 is not derived from more fundamental quantities, but allows for “de-deriving” a quantity more directly related to measurements. (I.e., what is reported as an accessible volume is actually a measure of the clay concentration.)
It is, however, impossible (as far as I see) to back-calculate the actual value of \(n’_ \mathrm{Cl}\) from provided formulas and values of \(\epsilon_\mathrm{eff}\), because masses and volumes of the sample sections are not provided. Therefore, we cannot independently assess the procedure used to evaluate \(\epsilon_\mathrm{eff}\), and simply have to assume that it is adequate.3 Here are the reported values of \(\epsilon_\mathrm{eff}\) for each test, and the corresponding evaluation of \(\bar{C}\) using eq. 2 (column 3)
*) The table in Vl07 says 0.076, but the concentration profile diagram says 0.090. **) The table in Vl07 says 0.16, but this must be a typo.
When using eq. 2 we have adopted porosities 0.536, 0.429, and 0.322,
respectively, for densities 1.3 g/cm3, 1.6 g/cm3, and 1.9 g/cm3.
The tabulated \(\epsilon_\mathrm{eff}\) values are evaluated as averages of the clay concentration profiles (presented as effective porosity profiles), which look like this for the samples exposed to background concentrations 0.01 M, 0.1 M and 1.0 M (profiles for 0.05 M and 0.4 M are not presented in Vl07)
The chloride concentration increases near the interfaces in all samples; we have discussed this interface excess effect in previousposts. Vl07 deal with this issue by evaluating the averages only for the inner parts of the samples. I performed a similar evaluation, also presented in the above figures (blue lines). In this evaluation I adopted the criterion to exclude all points situated less than 2 mm from the interfaces (Vl07 seem to have chosen points a bit differently). The clay concentration reevaluated in this way is also listed in the above table (last column). Given that I have only used nominal density for each sample (I don’t have information on the actual density of the sample sections), I’d say that the re-evaluated values agree well with those de-derived from reported \(\epsilon_\mathrm{eff}\). One exception is the sample 1.9/0.01, which is seen to have concentration points all over the place (or maybe detection limit is reached?). While Vl07 choose the lowest three points in their evaluation, here we choose to discard this result altogether. I mean that it is rather clear that this concentration profile cannot be considered to represent equilibrium.
As the reevaluation gives similar values as those reported, and since
we lack information for a full analysis, we will use the values
de-derived from reported \(\epsilon_\mathrm{eff}\) in the continued
assessment (except for sample 1.9/0.01).
Diffusion related estimations
Vl07 determine diffusion parameters by fitting various mathematical expressions to flux data.4 Parameters fitted in this way generally depend on the underlying adopted model, and we have discussed how equilibrium concentrations can be extracted from such parameters in an earlier blog post. In Vl07 it is clear that the adopted mathematical and conceptual model is the effective porosity diffusion model. When first presented in the article, however, it is done so in terms of a sorption distribution coefficient (\(R_d\)) that is claimed to take on negative values for anions. The presented mathematical expressions therefore contain a so-called rock capacity factor, \(\alpha\), which relates to \(R_d\) as \(\alpha = \phi + \rho_d\cdot R_d\). But such use of a rock capacity factor is a mix-up of incompatible models that I have criticized earlier. However, in Vl07 the description involving a sorption coefficient is in words only — \(R_d\) is never brought up again — and all results are reported, interpreted and discussed in terms of effective (or “chloride-accessible”) porosity, labeled \(\epsilon\) or \(\epsilon_\mathrm{Cl}\). We here exclusively use the label \(\epsilon_\mathrm{eff}\) when referring to formulas in Vl07. The mathematics is of course the same regardless if we call the parameter \(\alpha\), \(\epsilon\), \(\epsilon_\mathrm{Cl}\), or \(\epsilon_\mathrm{eff}\).
Mass balance in the out-diffusion stage
Vl07 measured the amount of tracers accumulated in the two reservoirs during the out-diffusion stage. The flux into the left side reservoir, which served as source reservoir during the preceding through-diffusion stage, was completely obscured by significant amounts of tracers present in the confining filter, and will not be considered further (also Vl07 abandon this flux in their analysis). But the total amount of tracers accumulated in the right side reservoir, \(N_\mathrm{right}\),5 can be used to directly estimate the chloride equilibrium concentration.
The initial concentration profile in the out-diffusion stage is linear (it is the steady-state profile), and the total amount of tracers, \(N_\mathrm{tot}\),6 can be expressed
where \(\bar{c}_0\) is the initial clay concentration at the left side interface, and \(V_\mathrm{sample}\) (\(\approx\) 5 cm3) is the sample volume.
A neat feature of the out-diffusion process is that two thirds of the tracers end up in the left side reservoir, and one third in the right side reservoir, as illustrated in this simulation
\(\bar{c}_0\) can thus be estimated by using
\(N_\mathrm{tot} = 3\cdot N_\mathrm{right}\) in eq. 3, giving
where \(c_\mathrm{source}\) is the tracer concentration in the left side reservoir in the through-diffusion stage.7 Although eq. 4 depends on a particular solution to the diffusion equation, it is independent of diffusivity (the diffusivity in the above simulation is \(1\cdot 10^{-10}\) m2/s). Eq. 4 can in this sense be said to be a direct estimation of \(\bar{c}_0\) (from measured \(N_\mathrm{right}\)), although maybe not as “direct” as the measurement of stable chloride, discussed previously.
Vl07 state eq. 4 in terms of a “Cl-accessible porosity”, but this is still just an interpretation of the clay concentration; \(\bar{c}_0\) is, in contrast to \(\epsilon_\mathrm{eff}\), directly accessible experimentally in principle. From the reported values of \(\epsilon_\mathrm{eff}\) we may back-calculate \(\bar{c}_0\), using the relation \(\bar{c}_0 / c_\mathrm{source} = \epsilon_\mathrm{eff}/\phi\). Alternatively, we may use eq. 4 directly to evaluate \(\bar{c}_0\) from the reported values of \(N_\mathrm{right}\). Curiously, these two approaches result in slightly different values for \(\bar{c}_0/c_\mathrm{source}\). I don’t understand the cause for this difference, but since \(N_\mathrm{right}\) is what has actually been measured, we use these values to estimate \(\bar{c}_0.\) The resulting equilibrium concentrations are
Test
\(N_\mathrm{right}\) (10-10 mol)
\(\bar{c}_0/c_\mathrm{source}\) (-)
1.3/0.01
4.10
0.038
1.3/0.05
10.2
0.097
1.3/0.1
17.8
0.168
1.3/0.4
41.4
0.395
1.3/1.0
52.4
0.445
1.6/0.01
1.21
0.014
1.6/0.05
3.64
0.043
1.6/0.1
6.15
0.072
1.6/0.4
13.0
0.154
1.6/1.0
21.6
0.225
1.9/0.01
0.41
0.006
1.9/0.05
1.14
0.018
1.9/0.1
1.64
0.025
1.9/0.4
3.19
0.051
1.9/1.0
8.19
0.113
We have now investigated two independent estimations of the chloride equilibrium concentrations: from mass balance of chloride tracers in the out-diffusion stage, and from measured stable chloride content. Here are plots comparing these two estimations
The similarity is quite extraordinary! With the exception of two
samples (1.3/0.4 and 1.9/0.1), the equilibrium chloride concentrations
evaluated in these two very different ways are essentially the
same. This result strongly confirms that the evaluations are adequate.
Steady-state fluxes
Vl07 present the flux evolution in the through-diffusion stage only for a single test (1.6/1.0), and it looks like this (left diagram)
The outflux reaches a relatively stable value after about 7 days,
after which it is meticulously monitored for a quite long time period.
The stable flux is not completely constant, but decreases slightly
during the course of the test. We anyway refer to this part as the
steady-state phase, and to the preceding part as the transient phase.
One reason that the steady-state is not completely stable is, reasonably, that the source reservoir concentration slowly decreases during the course of the test. The estimated drop from this effect, however, is only about one percent,8 while the recorded drop is substantially larger, about 7%. Vl07 do not comment on this perhaps unexpectedly large drop, but it may be caused e.g. by the ongoing conversion of the bentonite to a purer sodium state (see above).
Most of the analysis in Vl07 is based on anyway assigning a single
value to the steady-state flux. Judging from the above plot, Vl07 seem
to adopt the average value during the steady-state phase, and it is
clear that the assigned value is well constrained by the measurements
(the drop is a second order effect). The steady-state flux can
therefore be said to be directly measured in the through-diffusion
stage, rather than being obtained from fitting a certain model to
data.
Vl07 only implicitly consider the steady-state flux, in terms of a fitted “effective diffusivity” parameter, \(D_e\) (more on this in the next section). We can, however, “de-derive” the corresponding steady-state fluxes using \(j_\mathrm{ss} = D_e\cdot c_\mathrm{source}/L\), where \(L\) (= 0.01 m) is sample length. When comparing different tests it is convenient to use the normalized steady state flux \(\widetilde{j}_\mathrm{ss} = j_\mathrm{ss}/c_\mathrm{source}\), which then relates to \(D_e\) as \(\widetilde{j}_\mathrm{ss} = D_e/L\). Indeed, “effective diffusivity” is just a scaled version of the normalized steady-state flux, and it makes more sense to interpret it as such (\(D_e\) is not a diffusion coefficient). From the reported values of \(D_e\) we obtain the following normalized steady-state fluxes (my apologies for a really dull table)
Test
\(D_e\) (10-12 m2/s)
\(\widetilde{j}_\mathrm{ss}\) (10-10 m/s)
1.3/0.01
2.6
2.6
1.3/0.05
7.5
7.5
1.3/0.1
16
16
1.3/0.4
25
25
1.3/1.0
49
49
1.6/0.01
0.39
0.39
1.6/0.05
1.1
1.1
1.6/0.1
2.3
2.3
1.6/0.4
4.6
4.6
1.6/1.0
10
10
1.9/0.01
0.033
0.033
1.9/0.05
0.12
0.12
1.9/0.1
0.24
0.24
1.9/0.4
0.5
0.5
1.9/1.0
1.2
1.2
Plotting \(\widetilde{j}_\mathrm{ss}\) as a function of background concentration gives the following picture
The steady-state flux show a very consistent behavior: for all three
densities, \(\widetilde{j}_\mathrm{ss}\) increases with background
concentration, with a higher slope for the three lowest background
concentrations, and a smaller slope for the two highest background
concentrations. Although we have only been able to investigate the
1.6/1.0 test in detail, this consistency confirms that the
steady-state flux has been reliably determined in all tests.
Transient phase evaluations
So far, we have considered estimations based on more or less direct
measurements: stable chloride concentration profiles, tracer mass
balance in the out-diffusion stage, and steady-state fluxes. A major
part of the analysis in Vl07, however, is based on fitting solutions
of the diffusion equation to the recorded flux.
Vl07 state somewhat different descriptions for the through- and
out-diffusion stages. For out-diffusion they use an expression for the
flux into the right side reservoir (the sample is assumed located
between \(x=0\) and \(x=L\))
where \(j_\mathrm{ss}\) is the steady-state flux,9 \(D_e\) is “effective diffusivity”, and \(\epsilon_\mathrm{eff}\) is the effective porosity parameter (Vl07 also state a similar expression for the diffusion into the left side reservoir, but these results are discarded, as discussed earlier). For through-diffusion, Vl07 instead utilize the expression for the amount tracer accumulated in the right side reservoir
were \(S\) denotes the cross section area of the sample.
It is clear that Vl07 use \(D_e\) and \(\epsilon_\mathrm{eff}\) as fitting parameters, but not exactly how the fitting was conducted. \(D_e\) seems to have been determined solely from the the through-diffusion data, while separate values are evaluated for \(\epsilon_\mathrm{eff}\) from the through- and out-diffusion stages. As already discussed, Vl07 also provide a third estimation of \(\epsilon_\mathrm{eff}\), based on mass-balance in the out-diffusion stage. To me, the study thereby gives the incorrect impression of providing a whole set of independent estimations of \(\epsilon_\mathrm{eff}\). Although eqs. 5 and 6 are fitted to different data, they describe diffusion in one and the same sample, and an adequate fitting procedure should provide a consistent, single set of fitted parameters \((D_e, \epsilon_\mathrm{eff})\). Even more obvious is that the estimation of \(\epsilon_\mathrm{eff}\) from fitting eq. 5 should agree with the estimation from the mass-balance in the out diffusion stage — the accumulated amount in the right side reservoir is, after all, given by the integral of eq. 5. A significant variation of the reported fitting parameters for the same sample would thus signify internal inconsistency (experimental- or modelwise).
In the following reevaluation we streamline the description by solely using fluxes as model expressions,4 and by emphasizing steady-state flux as a parameter, which I think gives particularly neat expressions,10 (“TD” and “OD” denote through- and out-diffusion, respectively)
Here we use the pore diffusivity, \(D_p\), instead of the combination \(D_e/\epsilon_\mathrm{eff}\) in the exponential factors, and \(\widetilde{j} = j/c_\mathrm{source}\) denotes normalized flux. This formulation clearly shows that the time evolution is governed solely by \(D_p\), and that \(\widetilde{j}_\mathrm{ss}\) simply acts as a scaling factor.
In my opinion, using \(\widetilde{j}_\mathrm{ss}\) and \(D_p\) gives a formulation more directly related to measurable quantities; the steady-state flux is directly accessible experimentally, as we just examined, and \(D_p\) is an actual diffusion coefficient (in contrast to \(D_e\)) that can be directly evaluated from clay concentration profiles. Of course, eqs. 7 and 8 provide the same basic description as eqs. 5 and 6, and \(\widetilde{j}_\mathrm{ss}\) and \(D_p\) are related to the parameters reported in Vl07 as
When reevaluating the reported data we focus on the above discussed consistency aspect, i.e. whether or not a single model (a single pair of parameters) can be satisfactory fitted to all available data for the same sample. In this regard, we begin by noting that the fitting parameters are already constrained by the direct estimations. We have already concluded that the recorded steady-state flux basically determines \(\widetilde{j}_\mathrm{ss}\), and if we combine this with the estimated chloride clay concentration, \(D_p\) is determined from \(j_\mathrm{ss} = \phi\cdot D_p\cdot \bar{c}_0/L\), i.e.
Here are plotted values of \(D_p\) evaluated in this manner
Note that these values basically remain constant for samples of similar density (within a factor of 2) as the background concentration is varied by two orders of magnitude. This is the expected behavior of an actual diffusion coefficient,11 and confirms the adequacy of the evaluation; the numerical values also compares rather well with corresponding values for “MX-80” bentonite, measured in closed-cell tests (indicated by dashed lines in the figure).
Using eq. 10, we can also evaluate values of \(D_p\) corresponding to
the various reported fitted parameters \(\epsilon_\mathrm{eff}\). The
result looks like this (compared with the above evaluations from
direct estimations)
As pointed out above, a consistent evaluation requires that the
parameters fitted to the out-diffusion flux (red) are very similar
to those evaluated from considering the mass balance in the same process
(blue). We note that the resemblance is quite reasonable, although
some values — e.g. tests 1.3/1.0 and 1.6/1.0 — deviate in a perhaps
unacceptable way.
\(D_p\) evaluated from reported through-diffusion parameters, on the other hand, shows significant scattering (green). As the rest of the values are considerably more collected, and as the steady-state fluxes show no sign whatsoever that the diffusion coefficient varies in such erratic manner, it is quite clear that this scattering indicates problems with the fitting procedure for the through-diffusion data.
The 1.6/1.0 test
To further investigate the fitting procedures, we take a detailed look at the 1.6/1.0 test, for which flux data is provided. Vl07 report fitted parameters \(D_e = 1.0\cdot 10^{-11}\) m2/s and \(\epsilon_\mathrm{eff} = 0.063\) to the through-diffusion data, corresponding to \(\widetilde{j}_\mathrm{ss} = 1.0\cdot 10^{-9}\) m/s and \(D_p = 1.6\cdot 10^{-10}\) m2/s. We have already concluded that the steady-state flux is well captured by this data, but to see how well fitted \(\epsilon_\mathrm{eff}\) (or \(D_p\)) is, lets zoom in on the transient phase
This diagram also contains models (eq. 7) with different values of \(D_p\), and with a slightly different value of \(j_\mathrm{ss}\).12 It is clear that the model presented in the paper (black) completely misses the transient phase, and that a much better fit is achieved with \(D_p = 9.7\cdot10^{-11}\) m2/s (and \(\widetilde{j}_\mathrm{ss} = 1.06\cdot 10^{-9}\) m/s) (red). This difference cannot be attributed to uncertainty in the parameter \(D_p\) — the reported fit is simply of inferior quality. With that said, we note that all information on the transient phase is contained within the first three or four flux points; the reliability could probably have been improved by measuring more frequently in the initial stage.13
A reason for the inferior fit may be that Vl07 have focused only on the linear part of eq. 6; the paper spends half a paragraph discussing how the approximation of this expression for large \(t\) can be used to extract the fitting parameters using linear regression. Does this mean that only experimental data for large times where used to evaluate \(D_e\) and \(\epsilon_\mathrm{eff}\)? Since we are not told how fitting was performed, we cannot answer this question. Under any circumstance, the evidently low quality of the fit puts in question all the reported \(\epsilon_\mathrm{eff}\) values fitted to through-diffusion data. This is actually good news, as several of the corresponding \(D_p\) values were seen to be incompatible with constraints from direct estimations. We can thus conclude with some confidence that the inconsistency conveyed by the differently evaluated fitting parameters does not indicate experimental shortcomings, but stems from bad fitting of the through-diffusion model. Therefore, we simply dismiss the reported \(\epsilon_\mathrm{eff}\) values evaluated in this way. Note that the re-fitted value for \(D_p\) \((9.7\cdot10^{-11}\) m2/s) is consistent with those evaluated from direct estimations.
We note that when fitting the transient phase, it is appropriate to
use a value of \(\widetilde{j}_\mathrm{ss}\) slightly larger than the
average value adopted by Vl07 (as the model does not account for the
observed slight drop of the steady-state flux). This is only a minor
variation in the \(\widetilde{j}_\mathrm{ss}\) parameter itself (from
\(1.02\cdot10^{-9}\) to \(1.06\cdot10^{-9}\) m/s), but, since this value
sets the overall scale, it indirectly influences the fitted value of
\(D_p\) (model fitting is subtle!).
More questions arise regarding the fitting procedures when also examining the presented out-diffusion stage for the 1.6/1.0 sample. The tabulated fitted value for this stage is \(\epsilon_\mathrm{eff}\) = 0.075, while it is implied that the same value has been used for \(D_e\) as evaluated from the the through-diffusion stage (\(1.0\cdot 10^{-11}\) m2/s). The corresponding pore diffusivity is \(D_p = 1.33\cdot 10^{-10}\) m2/s. The provided plot, however, contains a different model than tabulated, and looks similar to this one (left diagram)
Here the presented model (black dashed line) instead corresponds to \(D_p = 8.5\cdot 10^{-11}\) m2/s (or \(\epsilon_\mathrm{eff}\) = 0.118). The model corresponding to the tabulated value (orange) does not fit the data! I guess this error may just be due to a typo in the table, but it nevertheless gives more reasons to not trust the reported \(\epsilon_\mathrm{eff}\) values fitted to diffusion data.
The above diagram also shows the model corresponding to the reported parameters from the through-diffusion stage (black solid line). Not surprisingly, this model does not fit the out-diffusion data, confirming that it does not appropriately describe the current sample. The model we re-fitted in the through-diffusion stage (red), on the other hand, captures the outflux data quite well. By also slightly adjusting \(\widetilde{j}_{ss}\), from from \(1.06\cdot10^{-9}\) to \(0.99\cdot10^{-9}\) m/s, to account for the drop in steady-state flux during the course of the through-diffusion test, and by plotting in a lin-lin rather than a log-log diagram, the picture looks even better! In a lin-lin plot (right diagram), it is easier to note that the model presented in the graph of Vl07 actually misses several of the data points. Could it be that Vl07 used visual inspection of the model in a log-log diagram to assess fitting quality? If so, data points corresponding to very low fluxes are given unreasonably high weight.14 This could be (another) reason for the noted difference between \(D_p\) evaluated from fitted parameters to the out-diffusion flux, and from the total accumulated amount of tracer (which should be equal).
From examining the reported results of sample 1.6/1.0 we have seen that the fitting procedures adopted in Vl07 appear inappropriate, but also that a consistent model can be successfully fitted to all available data (using a single \(D_p\)). Vl07 don’t provide flux data for any other sample, but we must conclude that the reported fitted \(\epsilon_\mathrm{eff}\) parameters cannot be trusted. Luckily, the preformed refitting exercise confirms the results obtained from analysis of stable chloride profiles and accumulated amount of tracers in out-diffusion, and we conclude that these results most probably are reliable. The corresponding value of \(\bar{c}_0/c_\mathrm{source}\) (using eq. 11) for the refitted model is here compared with the estimations from direct measurements
Summary and verdict
Chloride equilibrium concentrations evaluated from mass balance of the tracer in the out-diffusion stage and from stable chloride content show remarkable agreement. On the other hand, the scattering of estimated concentrations increases substantially if they are also evaluated from the reported fitted diffusion parameters. This could indicate underlying experimental problems, as a consistent evaluation should result in a single value for the equilibrium concentration; the various evaluations — stable chloride, out-diffusion mass balance, through-diffusion fitting and out-diffusion fitting — relate, after all, to a single sample.
By reexamining the evaluations we have found, however, that the problem is associated with how the fitting to diffusion data has been conducted (and presented), rather than indicating fundamental experimental issues. In the test that we have been able to examine in detail (1.6/1.0), we found that the reported models do not fit data, but also that it is possible to satisfactorily refit a single model that is also compatible with the direct methods for evaluating the equilibrium concentration. For the rest of the samples, we have also been able to discard the fitted diffusion parameters, as they are not compatible e.g. with how the steady-state flux (very consistently) vary with density and background concentration.
For these reasons, we discard the reported “effective porosity”
parameters evaluated from fitting solutions of the diffusion equation
to flux data, and keep the results from direct measurements of
chloride equilibrium concentrations (from stable chloride profile
analysis and mass-balance in the out-diffusion stage). I judge the
resulting chloride equilibrium concentrations as reliable and that
they can be used for increased qualitative process understanding. I
furthermore judge the directly measured steady-state fluxes as
reliable. This study thus provide adequate values for both chloride
equilibrium concentrations and diffusion coefficients.
However, a frustrating problem is that, although the equilibrium concentrations are well determined, we have little information on the exact state of the samples in which they have been measured. We basically have to rely on that the “KWK” material is “similar” to “MX-80”, keeping in mind that “MX-80” is not really a uniform material (from a scientific point of view). Also, the exchangeable mono/divalent cation ratio is most probably quite different in samples contacted with different background concentrations.
Yet, I judge the present study to provide the best information
available on chloride equilibrium in compacted bentonite, and will use
it e.g. for investigating the salt exclusion mechanism in these
systems (Ialreadyhave). That this information is the best available is, however, also
a strong argument for that more and better constrained data is
urgently needed.
The (reliable) results are presented in the diagram below, which includes “confidence areas”, that takes into account the spread in equilibrium concentrations, in samples where more than a single evaluation were performed, and the estimated uncertainty in effective montmorillonite dry density (the actual points are plotted at nominal density, assuming 80% montmorillonite content)
[1] Vejsada et al. (2006) call their material “KWK 20-80”. In other contexts, I have also found the versions “KWK food grade” and “KWK krystal klear”. I have given up my attempts at trying to understand the difference between these “KWK” variants.
[3] This should be relatively straightforward, but I get at bit nervous e.g. about the presence of a rather arbitrary factor 0.85 in the presented formula (eq. 19 in Van Loon et al. (2007)).
[4] As always for these types of diffusion tests, the raw data consists of simultaneously measured values of time (\(\{t_i\}\)) and reservoir concentrations (\(\{c_i\}\)). From these, flux can be evaluated as (\(A\) is sample cross sectional area, and \(V_\mathrm{res}\) is reservoir volume)
\(\bar{j}_i\) is the mean flux in the time interval between \(t_{i-1}\)
and \(t_i\), and should be associated with the average time of the
same interval: \(\bar{t}_i = (t_i + t_{i-1})/2\). The above formula
assumes no solution replacement after the \((i-1)\):th measurement (if
the solution is replaced, \(\left (c_i – c_{i-1} \right )\) should be
replaced with \(c_i\)).
Alternatively one can work with the accumulated amount of substance, which e.g. is \(N(t_i) = \sum_{j=1}^i c_j\cdot V_\mathrm{res}\), in case the solution is replaced after each measurement. I prefer using the flux because eq. * only depends on two consecutive measurements, while \(N(t_i)\) in principle depends on all measurements up to time \(t_i\). Also, I think it is easier to judge how well e.g. a certain model fits or is constrained by data when using fluxes; the steady-state, for example, then corresponds to a constant value.
Van Loon et al. (2007) seem to have utilized both fluxes and accumulated amount of substance in their evaluations, as discussed in later sections.
[8] From total test time, recorded flux, and sample cross sectional area, we estimate that about \(5.8\cdot 10^{-8}\) mol of tracer is transferred from the source reservoir during the course of the test (\(50\) days\(\cdot 2.7\cdot 10^{-11}\) mol/m2/s\(\cdot 0.0005\) m2). This is about 1% of the total amount tracer, \(c_\mathrm{source} \cdot V_\mathrm{source} = 2.65 \cdot 10^{-5}\) M \(\cdot 0.2\) L = \(5.3\cdot 10^{-6}\) mol.
[9] Van Loon et al. (2007) label this parameter \(J_L\), and don’t relate it explicitly to the steady-state flux. From the experimental set-up it is clear, however, that the initial value of the out-diffusion flux (into the right side reservoir) is the same as the previously maintained steady-state flux. Note that the expressions for the fluxes in the out-diffusion stage in Van Loon et al. (2007) has the wrong sign.
[10] The description provided by eqs. 5 and 6 not only mixes expressions for flux and accumulated amount tracer, but also contains three dependent parameters \(D_e\), \(\epsilon_\mathrm{eff}\), and \(j_\mathrm{ss}\) (e.g. \(j_\mathrm{ss} = D_e/(c_\mathrm{source}\cdot L)\)). In this reformulation, the model parameters are strictly only \(\widetilde{j}_\mathrm{ss}\) and \(D_p\). We have also divided out \(c_\mathrm{source}\) to obtain equations for normalized fluxes. Note that the expression for \(\widetilde{j}_{TD}(L,t)\) is essentially the same that we have used in previousassessments of through-diffusion tests. Note also that eqs. 7 and 8 imply the relation \(\widetilde{j}_{OD}(L,t) = \widetilde{j}_{ss} – \widetilde{j}_{TD}(L,t)\), reflecting that the out-diffusion process is essentially the through-diffusion process in reverse.
[11] Note the similarity with that diffusivity also is basically independent of background concentration for simple cations. Note also that there is no reason to expect completely constant \(D_p\) for a given density, because the samples are not identically prepared (being saturated with saline solutions of different concentration).
[12] As we here consider a single sample, we alternate a bit sloppily between steady-state flux (\(j_\mathrm{ss} \)) and normalized steady-state flux (\(\widetilde{j}_\mathrm{ss}\)), but these are simply related by a constant: \(\widetilde{j}_\mathrm{ss} = j_\mathrm{ss} / c_\mathrm{source}\). For the 1.6/1.0 test this constant is (as tabulated) \(c_\mathrm{source} = 2.65\cdot 10^{-2}\) mol/m3.
[13] I think it is a bit amusing that the pattern of data points suggests measurements being performed on Mondays, Wednesdays, and Fridays (with the test started on a Wednesday).
[14] I have warned about the dangers of log-log plots earlier.
“Multi-porosity” models1 — i.e
models that account for both a bulk water phase and one, or several,
other domains within the clay — have become increasingly
popular in bentonite research during the last couple of decades. These
are obviously macroscopic, as is clear e.g. from the benchmark
simulations described in
Alt-Epping et
al. (2015), which are specified to be discretized into 2 mm thick
cells; each cell is consequently assumed to contain billions and
billions individual montmorillonite particles. The macroscopic
character is also relatively clear in their description of two
numerical tools that have implemented multi-porosity
PHREEQC and CrunchFlowMC have implemented a Donnan approach to describe the electrical potential and species distribution in the EDL. This approach implies a uniform electrical potential \(\varphi^\mathrm{EDL}\) in the EDL and an instantaneous equilibrium distribution of species between the EDL and the free water (i.e., between the micro- and macroporosity, respectively). The assumption of instantaneous equilibrium implies that diffusion between micro- and macroporosity is not considered explicitly and that at all times the chemical potentials, \(\mu_i\), of the species are the same in the two porosities
On an abstract level, we may thus illustrate a multi-porosity approach
something like this (here involving two domains)
The model is represented by one
continuum for the “free water”/”macroporosity” and one for the
“diffuse layer”/”microporosity”,2 which are
postulated to be in equilibrium within each macroscopic cell.
But such an equilibrium (Donnan equilibrium)
requires a
semi-permeable component. I am not aware of any suggestion for such
a component in any publication on multi-porosity
models. Likewise, the co-existence of diffuse layer and free water
domains requires
a mechanism that prevents swelling and maintains the pressure
difference — also the water chemical potential should of course be
the equal in the two “porosities”.3
Note that the questions of what constitutes the semi-permeable
component and what prevents swelling have a clear answer in
the homogeneous mixture model. This answer also corresponds to an
easily identified real-world object: the metal filter (or similar
component) separating the sample from the external solution.
Multi-porosity models, on the other hand, attribute no particular
significance to interfaces between sample and external
solutions. Therefore, a candidate for the semi-permeable component has
to be — but isn’t — sought elsewhere. Donnan equilibrium
calculations are virtually meaningless without identifying this
component.
The partitioning between diffuse layer and free water in
multi-porosity models is, moreover, assumed to be controlled by water
chemistry, usually by means of the
Debye length. E.g. Alt-Epping et al. (2015) write
To determine the volume of the microporosity, the surface area of montmorillonite, and the Debye length, \(D_L\), which is the distance from the charged mineral surface to the point where electrical potential decays by a factor of e, needs to be known. The volume of the microporosity can then be calculated as \begin{equation*} \phi^\mathrm{EDL} = A_\mathrm{clay} D_L, \end{equation*} where \(A_\mathrm{clay}\) is the charged surface area of the clay mineral.
I cannot overstate how strange the multi-porosity description
is. Leaving the abstract representation, here is an attempt to
illustrate the implied clay structure, at the “macropore” scale
The view emerging from the above description is actually even more
peculiar, as the “micro” and “macro” volume fractions are supposed
to vary with the Debye length. A more general illustration of how the
pore structure is supposed to function is shown in this animation
(“I” denotes ionic strength)
What on earth could constitute such magic semi-permeable membranes?!
(Note that they are also supposed to withstand the inevitable pressure
difference.)
Here, the informed reader may object and point out that no researcher
promoting multi-porosity has this magic pore structure in
mind. Indeed, basically all multi-porosity publications instead
vaguely claim that the domain separation occurs on the nanometer scale
and present microscopic illustrations, like this (this is a
simplified version of what is found in
Alt-Epping et
al. (2015))
In the remainder of this post I will discuss how the idea of a domain separation on the microscopic scale is even more preposterous than the magic membranes suggested above. We focus on three aspects:
The implied structure of the free water domain
The arbitrary domain division
Donnan equilibrium on the microscopic scale is not really a valid concept
Implied structure of the free water domain
I’m astonished by how little figures of the microscopic scale are
explained in many publications. For instance, the illustration above
clearly suggests that “free water” is an interface region with
exactly the same surface area as the “double layer”. How can that
make sense? Also, if the above structure is to be taken seriously it
is crucial to specify the extensions of the various water layers. It
is clear that the figure shows a microscopic view, as it depicts an
actual diffuse layer.4 A diffuse layer width varies, say, in the
range 1 – 100 nm,5 but authors seldom reveal if we are
looking at a pore 1 nm wide or several hundred nm wide. Often we are
not even shown a pore — the water film just ends in a void, as in the
above figure.6
The vague nature of these descriptions indicates that they are merely “decorations”, providing a microscopic flavor to what in effect still is a macroscopic model formulation. In practice, most multi-porosity formulations provide some ad hoc mean to calculate the volume of the diffuse layer domain, while the free water porosity is either obtained by subtracting the diffuse layer porosity from total porosity, or by just specifying it. Alt-Epping et al. (2015), for example, simply specifies the “macroporosity”
The total porosity amounts to 47.6 % which is divided into 40.5 % microporosity (EDL) and 7.1 % macroporosity (free water). From the microporosity and the surface area of montmorillonite (Table 7), the Debye length of the EDL calculated from Eq. 11 is 4.97e-10 m.
Clearly, nothing in this description requires or suggests that the
“micro” and “macroporosities” are adjacent waterfilms on the
nm-scale. On the contrary, such an interpretation becomes quite
grotesque, with the “macroporosity” corresponding to half a
monolayer of water molecules! An illustration of an actual pore of
this kind would look something like this
This interpretation becomes even more bizarre, considering that
Alt-Epping et
al. (2015) assume advection to occur only in this half-a-monolayer
of water, and that the diffusivity is here a factor 1000 larger than
in the “microporosity”.
As another example, Appelo
and Wersin (2007) model a cylindrical sample of “Opalinus clay”
of height 0.5 m and radius 0.1 m, with porosity 0.16, by discretizing
the sample volume in 20 sections of width 0.025 m. The void volume of
each section is consequently
\(V_\mathrm{void} = 0.16\cdot\pi\cdot 0.1^2\cdot 0.025\;\mathrm{m^3} =
1.257\cdot10^{-4}\;\mathrm{m^3}\). Half of this volume (“0.062831853”
liter) is specified directly in the input file as the volume of the
free water;7 again, nothing suggests that this water
should be distributed in thin films on the nm-scale. Yet,
Appelo and Wersin (2007)
provide a figure, with no length scale, similar in spirit to that
above, that look very similar to this
They furthermore write about this figure (“Figure 2”)
It should be noted that the model can zoom in on the nm-scale suggested by Figure 2, but also uses it as the representative form for the cm-scale or larger.
I’m not sure I can make sense of this statement, but it seems that they imply that the illustration can serve both as an actual microscopic representation of two spatially separated domains and as a representation of two abstract continua on the macroscopic scale. But this is not true!
Interpreted macroscopically, the vertical dimension is fictitious, and
the two continua are in equilibrium in each paired cell. On a
microscopic scale, on the other hand, equilibrium between paired cells
cannot be assumed a priori, and it becomes crucial to specify
both the vertical and horizontal length scales. As
Appelo and Wersin (2007)
formulate their model assuming equilibrium between paired cells, it is
clear that the above figure must be interpreted macroscopically (the
only reference to a vertical length scale is that the “free
solution” is located “at infinite distance” from the surface).
We can again work out the implications of anyway interpreting the model microscopically. Each clay cell is specified to contain a surface area of \(A_\mathrm{surf}=10^5\;\mathrm{m^2}\).8 Assuming a planar geometry, the average pore width is given by (\(\phi\) denotes porosity and \(V_\mathrm{cell}\) total cell volume)
The double layer thickness is furthermore specified to be 0.628 nm.9 A microscopic interpretation of this particular model thus implies that the sample contains a single type of pore (2.51 nm wide) in which the free water is distributed in a thin film of width 1.25 nm — i.e. approximately four molecular layers of water!
Rather than affirming that multi-porosity model formulations are macroscopic at heart, parts of the bentonite research community have instead doubled down on the confusing idea of having free water distributed on the nm-scale. Tournassat and Steefel (2019) suggest dealing with the case of two parallel charged surfaces in terms of a “Dual Continuum” approach, providing a figure similar to this (surface charge is -0.11 C/m2 and external solution is 0.1 M of a 1:1 electrolyte)
Note that here the perpendicular length scale is specified,
and that it is clear from the start that the electrostatic potential
is non-zero everywhere. Yet,
Tournassat and Steefel
(2019) mean that it is a good idea to treat this system as if it
contained a 0.7 nm wide bulk water slice at the center of the
pore. They furthermore express an almost “postmodern” attitude
towards modeling, writing
It should be also noted here that this model refinement does not imply necessarily that an electroneutral bulk water is present at the center of the pore in reality. This can be appreciated in Figure 6, which shows that the Poisson–Boltzmann predicts an overlap of the diffuse layers bordering the two neighboring surfaces, while the dual continuum model divides the same system into a bulk and a diffuse layer water volume in order to obtain an average concentration in the pore that is consistent with the Poisson–Boltzmann model prediction. Consequently, the pore space subdivision into free and DL water must be seen as a convenient representation that makes it possible to calculate accurately the average concentrations of ions, but it must not be taken as evidence of the effective presence of bulk water in a nanoporous medium.
I can only interpret this way of writing (“…does not imply
necessarily that…”, “…must not be taken as evidence of…”)
that they mean that in some cases the bulk phase should be
interpreted literally, while in other cases the bulk phase
should be interpreted just as some auxiliary component. It is my
strong opinion that such an attitude towards modeling only contributes
negatively to process understanding (we may e.g. note that later in
the article, Tournassat
and Steefel (2019) assume this perhaps non-existent bulk water to
be solely responsible for advective flow…).
I say it again: no matter how much researchers discuss them in microscopic terms, these models are just macroscopic formulations. Using the terminology of Tournassat and Steefel (2019), they are, at the end of the day, represented as dual continua assumed to be in local equilibrium (in accordance with the first figure of this post). And while researchers put much effort in trying to give these models a microscopic appearance, I am not aware of anyone suggesting a reasonable candidate for what actually could constitute the semi-permeable component necessary for maintaining such an equilibrium.
Arbitrary division between diffuse layer and free water
Another peculiarity in the multi-porosity descriptions showing that they cannot be interpreted microscopically is the arbitrary positioning of the separation between diffuse layer and free water. We saw earlier that Alt-Epping et al. (2015) set this separation at one Debye length from the surface, where the electrostatic potential is claimed to have decayed by a factor of e. What motivates this choice?
Most publications on multi-porosity models define free water as a region where the solution is charge neutral, i.e. where the electrostatic potential is vanishingly small.10 At the point chosen by Alt-Epping et al. (2015), the potential is about 37% of its value at the surface. This cannot be considered vanishingly small under any circumstance, and the region considered as free water is consequently not charge neutral.
The diffuse layer thickness chosen by Appelo and Wersin (2007) instead corresponds to 1.27 Debye lengths. At this position the potential is about 28% of its value at the surface, which neither can be considered vanishingly small. At the mid point of the pore (1.25 nm), the potential is about 8%11 of the value at the surface (corresponding to about 2.5 Debye lengths). I find it hard to accept even this value as vanishingly small.
Note that if the boundary distance used by Appelo and Wersin (2007) (1.27 Debye lengths) was used in the benchmark of Alt-Epping et al. (2015), the diffuse layer volume becomes larger than the total pore volume! In fact, this occurs in all models of this kind for low enough ionic strength, as the Debye length diverges in this limit. Therefore, many multi-porosity model formulations include clunky “if-then-else” clauses,12 where the system is treated conceptually different depending on whether or not the (arbitrarily chosen) diffuse layer domain fills the entire pore volume.13
In the example from Tournassat and Steefel (2019) the extension of the diffuse layer is
1.6 nm, corresponding to about 1.69 Debye lengths. The potential is
here about 19% of the surface value (the value in the midpoint is
12% of the surface
value). Tournassat
and Appelo (2011) uses yet another separation distance — two Debye
lengths — based on
misusing the concept of exclusion volume in the Gouy-Chapman model.
With these examples, I am not trying to say that a better criterion is needed for the partitioning between diffuse layer and bulk. Rather, these examples show that such a partitioning is quite arbitrary on a microscopic scale. Of course, choosing points where the electrostatic potential is significant makes no sense, but even for points that could be considered having zero potential, what would be the criterion? Is two Debye lengths enough? Or perhaps four? Why?
These examples also demonstrate that researchers ultimately do not
have a microscopic view in mind. Rather, the “microscopic”
specifications are subject to the macroscopic constraints.
Alt-Epping et
al. (2015), for example, specifies a priori that the system
contains about 15% free water, from which it follows that the diffuse
layer thickness must be set to about one Debye length (given the
adopted surface area). Likewise,
Appelo and Wersin (2007)
assume from the start that Opalinus clay contains 50% free water, and
set up their model accordingly.14Tournassat and Steefel
(2019) acknowledge their approach to only be a “convenient
representation”, and don’t even relate the diffuse layer
extension to a specific value of the electrostatic
potential.15 Why
the free water domain anyway is considered to be positioned in the
center of the nanopore is a mystery to me (well, I guess because
sometimes this interpretation is supposed to be taken literally…).
Note that none of the free water domains in the considered models are actually charged, even though the electrostatic potential in the microscopic interpretations is implied to be non-zero. This just confirms that such interpretations are not valid, and that the actual model handling is the equilibration of two (or more) macroscopic, abstract, continua. The diffuse layer domain is defined by following some arbitrary procedure that involves microscopic concepts. But just because the diffuse layer domain is quantified by multiplying a surface area by some multiple of the Debye length does not make it a microscopic entity.4
Donnan effect on the microscopic scale?!
Although we have already seen that we cannot interpret multi-porosity models microscopically, we have not yet considered the weirdest description adopted by basically all proponents of these models: they claim to perform Donnan equilibrium calculations between diffuse layer and free water regions on the microscopic scale!
The underlying mechanism for a Donnan effect is the establishment of charge separation, which obviously occur on the scale of the ions, i.e. on the microscopic scale. Indeed, a diffuse layer is the manifestation of this charge separation. Donnan equilibrium can consequently not be established within a diffuse layer region, and discontinuous electrostatic potentials only have meaning in a macroscopic context.
Consider e.g. the interface between bentonite and an external solution
in
the
homogeneous mixture model. Although this model ignores the
microscopic scale, it implies charge separation and a continuously
varying potential on this scale, as illustrated here
The regions where the potential varies are exactly what we categorize
as diffuse layers (exemplified in two ideal microscopic geometries).
The discontinuous potentials encountered in multi-porosity model descriptions (see e.g. the above “Dual Continuum” potential that varies discontinuously on the angstrom scale) can be drawn on paper, but don’t convey any physical meaning.
Here I am not saying that Donnan equilibrium calculations cannot be performed in multi-porosity models. Rather, this is yet another aspect showing that such models only have meaning macroscopically, even though they are persistently presented as if they somehow consider the microscopic scale.
An example of this confusion of scales is found in
Alt-Epping et
al. (2018), who revisit the benchmark problem of
Alt-Epping et
al. (2015) using an alternative approach to Donnan equilibrium:
rather than directly calculating the equilibrium, they model the clay
charge as immobile mono-valent anions, and utilize the
Nernst-Planck
equations. They present “the conceptual model” in a figure very
similar to this one
This illustration simultaneously conveys both a micro- and macroscopic view. For example, a mineral surface is indicated at the bottom, suggesting that we supposedly are looking at an actual interface region, in similarity with the figures we have looked at earlier. Moreover, the figure contains entities that must be interpreted as individual ions, including the immobile “clay-anions”. As in several of the previous examples, no length scale is provided (neither perpendicular to, nor along the “surface”).
On the other hand, the region is divided into cells, similar to the
illustration in Appelo and Wersin (2007). These can hardly have any other meaning
than to indicate the macroscopic discretization in the adopted
transport code (FLOTRAN). Also, as the “Donnan porosity” region
contains the “clay-anions” it can certainly not represent a diffuse
layer extending from a clay surface; the only way to make sense of
such an “immobile-anion” solution is that it represents a
macroscopic homogenized clay domain (a homogeneous mixture!).
Furthermore, if the figure is supposed to show the microscopic scale
there is no Donnan effect, because there is no charge separation!
Taking the depiction of individual ions seriously, the interface
region should rather look something like this in equilibrium
This illustrates the fundamental problem with a Donnan effect between microscopic compartments: the effect requires a charge separation, whose extension is the same as the size of the compartments assumed to be in equilibrium.16
Despite the confusion of the illustration in Alt-Epping et al. (2018), it is clear that a macroscopic model is adopted, as in our previous examples. In this case, the model is explicitly 2-dimensional, and the authors utilize the “trick” to make diffusion much faster in the perpendicular direction compared to the direction along the “surface”. This is achieved either by making the perpendicular diffusivity very high, or by making the perpendicular extension small. In any case, a perpendicular length scale must have been specified in the model, even if it is nowhere stated in the article. The same “trick” for emulating Donnan equilibrium is also used by Jenni et al. (2017), who write
In the present model set-up, this approach was implemented as two connected domains in the z dimension: one containing all minerals plus the free porosity (z=1) and the other containing the Donnan porosity, including the immobile anions (CEC, z=2, Fig. 2). Reproducing instantaneous equilibrium between Donnan and free porosities requires a much faster diffusion between the porosity domains than along the porosity domains.
Note that although the perpendicular dimension (\(z\)) here is referred
to without unit(!), this representation only makes sense in a
macroscopic context.
Jenni et al. (2017) also provide a statement that I think fairly well sums up the multi-porosity modeling endeavor:17
In a Donnan porosity concept, cation exchange can be seen as resulting from Donnan equilibrium between the Donnan porosity and the free porosity, possibly moderated by additional specific sorption. In CrunchflowMC or PhreeqC (Appelo and Wersin, 2007; Steefel, 2009; Tournassat and Appelo, 2011; Alt-Epping et al., 2014; Tournassat and Steefel, 2015), this is implemented by an explicit partitioning function that distributes aqueous species between the two pore compartments. Alternatively, this ion partitioning can be modelled implicitly by diffusion and electrochemical migration (Fick’s first law and Nernst-Planck equations) between the free porosity and the Donnan porosity, the latter containing immobile anions representing the CEC. The resulting ion compositions of the two equilibrated porosities agree with the concentrations predicted by the Donnan equilibrium, which can be shown in case studies (unpublished results, Gimmi and Alt-Epping).
Ultimately, these are models that, using one approach or the other,
simply calculates Donnan equilibrium between two abstract,
macroscopically defined domains (“porosities”,
“continua”). Microscopic interpretations of these models lead — as
we have demonstrated — to multiple absurdities and errors. I am not
aware of any multi-porosity approach that has provided any kind of
suggestion for what constitutes the semi-permeable component required
for maintaining the equilibrium they are supposed to describe.
Alternatively expressed: what, in the previous figure,
prevents the “immobile anions” from occupying the entire clay
volume?
The most favorable interpretation I can make of multi-porosity approaches to bentonite modeling is a dynamically varying “macroporosity”, involving magical membranes (shown above). This, in itself, answers why I cannot take multi-porosity models seriously. And then we haven’t yet mentioned the flawed treatment of diffusive flux.
[1] This category has many other names,
e.g. “dual
porosity” and “dual continuum”, models. Here, I mostly use the term
“multi-porosity” to refer to any model of this kind.
[3] This lack of a full
description is very much related to the incomplete description of
so-called
“stacks” — I am not aware of any reasonable suggestion of a
mechanism for keeping stacks together.
[4] Note the difference between a diffuse layer and a diffuse layer domain. The former is a structure on the nm-scale; the latter is a macroscopic, abstract model component (a continuum).
[5] The scale of an electric double layer is
set by the Debye length, \(\kappa^{-1}\). From the formula for a 1:1
electrolyte, \(\kappa^{-1} = 0.3 \;\mathrm{nm}/\sqrt{I}\), the Debye
length is seen to vary between 0.3 nm and 30 nm when ionic strength
is varied between 1.0 M to 0.0001 M (\(I\) is the numerical value of
the ionic strength expressed in molar units). Independent of the
value of the factor used to multiply \(\kappa^{-1}\) in order to
estimate the double layer extension, I’d say that the estimation 1 –
100 nm is quite reasonable.
[6] Here, the informed reader may perhaps point out that authors don’t really mean that the free water film has exactly the same geometry as the diffuse layer, and that figures like the one above are more abstract representations of a more complex structure. Figures of more complex pore structures are actually found in manymulti-porositypapers. But if it is the case that the free water part is not supposed to be interpreted on the microscopic scale, we are basically back to a magic membrane picture of the structure! Moreover, if the free water is not supposed to be on the microscopic scale, the diffuse layer will always have a negligible volume, and these illustrations don’t provide a mean for calculating the partitioning between “micro” and “macroporosity”.
It seems to me that not specifying the extension of the free water is a way for authors to dodge the question of how it is actually distributed (and, as a consequence, to not state what constitutes the semi-permeable component).
[7] The PHREEQC input files are provided as
supplementary material to Appelo and Wersin (2007). Here I consider the input corresponding
to figure 3c in the article. The free water is specified with
keyword “SOLUTION”.
[8] Keyword
“SURFACE” in the PHREEQC input file for figure 3c in the paper.
[9] Using the identifier “-donnan” for the “SURFACE”
keyword.
[10] We assume a boundary
condition such that the potential is zero in the solution infinitely
far away from any clay component.
[11] Assuming exponential decay, which is only strictly true for a single clay layer of low charge.
[12] For example,
Tournassat and Steefel
(2019) write
(\(f_{DL}\) denotes the volume fraction of the diffuse layer):
In PHREEQC and CrunchClay, the volume of the diffuse layer (\(V_{DL}\) in m3), and hence the \(f_{DL}\) value, can be defined as a multiple of the Debye length in order to capture this effect of ionic strength on \(f_{DL}\): \begin{equation*} V_{DL} = \alpha_{DL}\kappa^{-1}S \tag{22} \end{equation*} \begin{equation*} f_{DL} = V_{DL}/V_{pore} \end{equation*} […] it is obvious that \(f_{DL}\) cannot exceed 1. Equation (22) must then be seen as an approximation, the validity of which may be limited to small variations of ionic strength compared to the conditions at which \(f_{DL}\) is determined experimentally. This can be appreciated by looking at the results obtained with a simple model where: \begin{equation*} \alpha_{DL} = 2\;\mathrm{if}\;4\kappa^{-1} \le V_{pore}/S\;\mathrm{and,} \end{equation*} \begin{equation*} f_{DL} = 1 \;\mathrm{otherwise.} \end{equation*}
[13] Some tools (e.g. PHREEQC) allow to put a maximum size limit on the diffuse layer domain, independent of chemical conditions. This is of course only a way for the code to “work” under all conditions.
[14] As icing on the cake, these estimations of free water in bentonite (15%) and Opalinus clay (50%) appear to be based on the incorrect assumption that “anions” only reside in such compartments. In the present context, this handling is particularly confusing, as a main point with multi-porosity models (I assume?) is to evaluate ion concentrations in other types of compartments.
[16] Donnan equilibrium between microscopic
compartments can be studied in
molecular dynamics simulations, but they require the considered
system to be large enough for the electrostatic potential to reach
zero. The semi-permeable component in such simulations is
implemented by simply imposing constraints on the atoms making up
the clay layer.
When discussing semi-permeability, we noted that a bentonite sample that is saturated with a saline solution probably contains more salt in the initial stages of the process than what is dictated by the final state Donnan equilibrium. This salt must consequently diffuse out of the sample before equilibrium is reached.
The reason for such a possible “overshoot” of the clay concentration is that an infiltrating solution is not subject to a Donnan effect (between sample and external solution) when it fills out the air-filled voids of an unsaturated sample. Also, even if the region near the interface to the external solution becomes saturated — so that a Donnan effect is active — a sample may still take up more salt than prescribed by the final state, due to hyperfiltration: with a net inflow of water and an active Donnan effect, salt will accumulate at the inlet interface (unless the interface is flushed). This increased concentration, in turn, alters the Donnan equilibrium at the interface, with the effect that more salt diffuses into the clay.
These effects are relevant for our ongoing assessment of studies of chloride equilibrium concentrations. If bentonite samples are saturated with saline solutions, without taking precautions against these effects, evaluated equilibrium concentrations may be overestimated. Note that, even if saturating a sample may be relatively fast, it may take a long time for salt to reach full equilibrium, depending on details of the experimental set-up. In particular, if the set-up is such that the external solution does not flow past the inlet, equilibration may take a very long time, being limited by diffusion in filters and tubing.
Interface excess salt
Another way for evaluated salt concentrations to overestimate the true equilibrium value — which is independent of whether or not the sample has been saturated with a saline solution — is due to excess salt at the sample interfaces.
Suppose that you determine the equilibrium salt concentration in a bentonite sample in the following way. First you prepare the sample in a test cell and contact it with an external salt solution via filters. When the system (bentonite + solution) has reached equilibrium (taking all the precautions against overestimation discussed above), the concentration profile may be conceptualized like this
The aim is to determine \(\bar{c}_\mathrm{clay}\), the
clay concentration of the species of interest
(e.g. chloride), and to relate it to the corresponding concentration in the
external solution (\(c_ \mathrm{ext}\)).
After ensuring the value of \(c_\mathrm{ext}\) (e.g. by sampling or controlling the external solution), you unload the test cell and isolate the bentonite sample. In doing so, we must keep in mind that the sample will begin to swell as soon as the force on it is released, if only water is available. In the present example it is difficult not to imagine that some water is available, e.g. in the filters.1
It is thus plausible that the actual concentration profile look
something like this directly after the sample has been isolated
We will refer to the elevated concentration at the interfaces as the
interface excess. The exact shape of the resulting
concentration profile depends reasonably on the detailed procedure for
isolating the sample.2 If the ion content of the sample is measured
as a whole, and/or if the sample is stored for an appreciable amount
of time before further analysis (so that the profile evens out due to
diffusion), it is clear that the evaluated ion content will be larger
than the actual clay concentration.
To quantify how much the clay concentration may be overestimated due
to the interface excess, we introduce an effective penetration
depth, \(\delta\)
\(\delta\) corresponds to a depth of the external concentration that
gives the same interface excess as the actual distribution. Using this
parameter, it is easy to see that the clay concentration evaluated as
the average over the entire sample is
This expression is quite interesting. We see that the relative
overestimation, reasonably, depends linearly on \(\delta\) and on the
inverse of sample length. But the expression also contains the ratio
\(r \equiv c_\mathrm{ext}/\bar{c}_\mathrm{clay}\), indicating that the effect may
be more severe for systems where the clay concentration is small in
comparison to the external concentration (high density, low
\(c_\mathrm{ext}\)).
An interface excess is more than a theoretical concept, and is frequently observed e.g. in anion through-diffusion studies. We have previously encountered them when assessing the diffusion studies of Muurinen et al. (1988) and Molera et al. (2003).3Van Loon et al. (2007) clearly demonstrate the phenomenon, as they evaluate the distribution of stable chloride (the background electrolyte) in the samples after performing the diffusion tests.4 Here is an example of the chloride distribution in a sample of density 1.6 g/cm3 and background concentration of 0.1 M5
The line labeled \(\bar{c}_\mathrm{clay}\) is evaluated from the average of only the interior sections (0.0066 M), while the line labeled \(\bar{c}_\mathrm{eval}\) is the average of all sections (0.0104 M). Using the full sample to evaluate the chloride clay concentration thus overestimates the value by a factor 1.6. From eq. 1, we see that this corresponds to \(\delta = 0.2\) mm. For a sample of length 5 mm with the same penetration depth, the corresponding overestimation is a factor of 2.1.
Here is plotted the relative overestimation (eq. 1) as a function of \(\delta\) for several systems of varying length and \(r\) (\(= c^\mathrm{ext}/\bar{c}_\mathrm{clay}\))
We see that systems with large \(r\) and/or small \(L\) become hypersensitive to this effect. Thus, even if it may be expected that \(\delta\) decreases with increasing \(r\),6 we may still expect an increased overestimation for such systems.
To avoid this potential overestimation of the clay concentration, I
guess the best practice is to quickly remove the first couple of
millimeters on both sides of a sample after it has been unloaded. In
many through-diffusion tests, this is done as part of the study, as
the concentration profile across the sample often is measured. In
studies where samples are merely equilibrated with an external
solution, however, removing the interface regions may not be
considered.
Summary
We have here discussed some plausible reasons for why an evaluated
equilibrium salt concentration in a clay sample may be overestimated:
If samples are saturated directly with a saline solution. Better practice is to first saturate the sample with pure water (or a dilute solution) and then to equilibrate with respect to salt in a second stage.
If the external solution is not circulated. Diffusion may then occur over very long distances (depending on test design). The reasonable practice is to always circulate external solutions.
If interface excess is not handled. This is an issue even if saturation is done with pure water. The most convenient way to deal with this is to section off the first millimeters on both sides of the samples as quickly as possible after they are unloaded.
Footnotes
[1] One way to minimize this possible effect could be to
empty the filter before unloading the test cell. This may, however,
be difficult unless the filter itself is flushable. Also, you may
run into the problem of beginning to dry the sample.
[2] The only study I’m aware of that has
systematically investigated these types of concentration profiles is
Glaus et
al. (2011). They claim, if I understand correctly, that the
interface excess is not caused by swelling during
dismantling. Rather, they mean that the profile is the result of an
intrinsic density decrease that occurs in interface regions. Still,
they don’t discuss how swelling are supposed to be inhibited,
neither during dismantling, nor in order for the density
inhomogeneity to remain. Under any circumstance, the conclusions in
this blog post are not dependent on the cause for the presence of a
salt interface excess.
[3] In through-diffusion tests, the problem of the
interface excess is usually not that the equilibrium clay
concentration is systematically overestimated, since the detailed
concentration profile often is sampled in the final state. Instead,
the problem becomes how to separate the linear and non-linear parts
of the profile.