cond-mat0609505/COND.TEX
1: 
2: %\documentstyle[aps,multicol,pre]{revtex}
3: \documentstyle[preprint,prb,aps]{revtex}
4: %\documentstyle[twocolumn,prl,aps]{revtex}
5: %\usepackage{epsfig}
6: %\usepackage{graphicx}
7:  \textwidth 16cm
8:  \textheight 24cm
9:  \setlength{\topmargin}{-0.6cm}
10:  \setlength{\evensidemargin}{-0.6cm}
11:  \setlength{\oddsidemargin}{-0.3cm}
12: 
13: %\usepackage{multicol}
14: 
15:  \title{Constrained equilibrium as a tool for characterization of 
16:  deformable porous media.
17: }
18: 
19:  \author{E.V. Vakarin$^a$, Yurko Duda$^b$, J.~P.~Badiali$^a$
20: }
21:  \address{
22: $^a$ UMR 7575 LECA ENSCP-UPMC, 11 rue P. et M. Curie, 75231 Cedex 05, Paris, 
23: France\\ 
24: $^b$ Programa de Ingenier\'ia Molecular, Instituto Mexicano del Petr\'oleo,
25: 07730 D. F., M\'exico}
26: 
27: %\date{}
28:  
29: 
30: \begin{document}
31: \maketitle
32: \begin{abstract}
33: A new method for characterizing the deformable porous materials with
34: non-critical adsorption probes is proposed. The mechanism is based on a 
35: driving the adsorbate through a sequence of constrained equilibrium states
36: with the insertion isotherms forming a pseudo-critical point or a van der 
37: Waals-type loop. In the framework of a perturbation theory and Monte Carlo 
38: simulations we have found a link between the loop parameters and the host 
39: morphology. This allows one to characterize porous matrices through 
40: analyzing a shift of  the pseudo-critical point and a shape of 
41: the pseudo-spinodals.  
42: \end{abstract}
43: 
44: %\begin{multicols}{2}
45: 
46: \section{Introduction}
47: One of the methods for characterization of porous materials 
48: \cite{review,Jaroniec} is a retrieving of pore size 
49: distributions through adsorption probes.
50:  Capillary condensation in macro- or meso-pores 
51: permits to detect them rather reliably because the condensation 
52: pressure and temperature are very sensitive to the pore size. Nevertheless, 
53: it becomes more and more questionable\cite{avalanches} that the true 
54: equilibrium behavior is observed in experiments. One of the major problems 
55: is a well developed metastability leading to very long equilibration times. 
56: Therefore, the condensation is usually detected as an onset of hysteresis. 
57: Nevertheless, it has recently been argued\cite{Kierlik2} that the hysteretic 
58: behavior is not necessarily associated with an underlying equilibrium phase 
59: transition. In the case of micro-pores the situation is even more difficult 
60: because such pores are too small to support any sharp (bulk-like) phase 
61: transition. This makes one to search for alternative 
62: characterization mechanisms which are not based on the critical behavior. 
63: In the case of deformable matrices\cite{PRB}, such as 
64: aerogels\cite{Pablo,Herman} or amorphous intercalation 
65: compounds\cite{BisqL}, the problem is additionally complicated by variations 
66: of the host morphology with the guest density or pressure. Thus the guest 
67: response (e.g., adsorption isotherms) is a superposition of at least two 
68: contributions - the adsorbate thermodynamics under the host confinement and 
69: the adsorbate-induced host deformation.  
70: 
71: 
72: 
73: In this paper we propose a method for a  pore size 
74: determination through adsorption of non-critical fluids. The main idea
75: is to find an explicit relation between the density-dependent matrix 
76: morphology and some peculiarity in the fluid thermodynamic response.  
77: The mechanism is 
78: based on our finding\cite{JCPcomp} of negative compressibility 
79: states in confined dilatating systems. These states can appear as a 
80: consequence of a driving path - either a sequence of controlled lateral 
81: dilatations\cite{JCPcomp} or a sequential adsorption of prescribed 
82: portions\cite{Amarasekera}. This imposes an additional 
83: constraint\cite{Everett} on the fluid equilibrium. As a result, at certain 
84: conditions one can observe 
85:  a sequence of states which form a van der Waals-type loop in the isotherm. 
86: Then the loop parameters can be taken as the required specificities for a 
87: detection of the confining geometry. 
88: 
89: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
90: \section{Model}
91: 
92: The matrix is considered as a collection of slit-like pores with
93: different widths $h_i$ distributed according to some probability density.
94: For simplicity the pores are assumed to be uncorrelated. This allows us
95: to focus on a single pore of fluctuating width $h$ and surface area $S$. The 
96: width is known only statistically, that is, from a probability distribution
97: $f(h)$. The slit itself is a part of the hosting system, whose coupling
98: to the guest species changes the slit geometry (see below). 
99: The fluid is modeled as $N$ hard spheres of diameter $\sigma=1$,
100: injected into the pore. The hard core potential allows us to mimic
101: the excluded volume effects and ensures that the fluid is off-critical
102: (we do not consider freezing-type transitions).      
103: 
104: For any $h$ the total Hamiltonian is   
105: \begin{equation}
106: H=H_{ff}+H_{fw},
107: \end{equation}
108: where $H_{ff}$ is the hard sphere Hamiltonian, $H_{fw}$ is the slit
109: potential  
110: \begin{equation}
111: H_{fw}=A\sum_{i=1}^N 
112: \left[
113: \frac{1}{z_i^k}+\frac{1}{(h-z_i)^k}
114: \right];
115: \qquad
116: 1/2 \le z_i \le h-1/2.
117: \end{equation}
118: We do not take into account a short-ranged fluid-wall attraction,
119: responsible for the surface adsorption or layering effects. The 
120: inverse-power shape for $H_{fw}$ is chosen as a generic form of a soft 
121: repulsion, with $k$ controlling the softness. In some sense this mimics
122: the hydrophobic or non-wetting effects that, as we will see later, are
123: essential. On the other hand, this
124: simple model takes into account that the walls are impenetrable and
125: ignores possible localized adsorption effects, which could complicate
126: the overall picture.   
127: For technical purposes we are 
128: working with $k=3$. Moreover, our results are qualitatively insensitive to a 
129: particular choice of $k$. 
130: 
131: In principle, one may consider two adsorption mechanisms. One is 
132: an equilibrium between the pore and the bulk fluids. In 
133: that case $N$ would change according to the chemical potential $\mu=\mu_b$. 
134: In this paper we treat another mechanism. Namely, an adsorption in 
135: doses\cite{Amarasekera}, when $N$ is injected in prescribed portions.    
136: In that case we have a sequence of the fluid equilibrium
137: states corresponding to different $N$. The sequence is determined 
138: by an injection procedure, which imposes an additional constraint.   
139:  It should be emphasized that 
140: choosing the model we tried to retain only a 
141: minimal number of relevant ingredients: the absence of liquid-gas transition 
142: in the fluid bulk, excluded volume effects, and a 
143: density-dependent heterogeneous confinement. 
144: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
145: \section{Perturbation theory}
146: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
147: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
148: \subsection{Conditional  insertion isotherm}
149: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
150: At any pore width $h$ the free energy can be represented as
151: \begin{equation}
152: \beta F(h)=\beta F_0(h)-\ln \langle e^{-\beta H_{fw}}\rangle_0
153: \end{equation}
154: where $F_0(h)$ is the free energy of a reference system, and 
155: $\langle...\rangle_0$ is the average over the reference state.  
156: Taking a spatially confined hard sphere system (the one 
157: with $A=0$) as a reference, we consider a first order 
158: perturbation\cite{JCPcomp} for the conditional free energy
159: \begin{equation}
160: \beta F(h)=\beta F_0(h) +\beta S \rho (h) \int \prod_{i} d{z}_i H_{fw} 
161: \end{equation}
162: where $\beta=1/(kT)$ and $\rho(h)$ is the pore density in the "slab"
163: approximation 
164: \begin{equation}
165: \label{rh}
166: \rho(h)=\frac{N}{S(h-1)},
167: \end{equation}
168: ignoring a non-monotonic behavior of the  
169: density profile with increasing pore density. This is reasonable for wide 
170: pores and low fluid densities. 
171: The reference part is estimated in the excluded volume approximation, while
172: the perturbation contribution is $N\Psi(h)$, such that the total conditional
173: free energy is
174: \begin{equation}
175: \label{Fh}
176: \beta F (h)=-N\ln{\left[
177: \frac{1-b\rho(h)}{\Lambda_T^3\rho(h)}
178: \right] 
179: } -N +N\beta\Psi(h)
180: \end{equation}
181: where $b=2\pi\sigma^3/3$ is the excluded volume factor, $\Lambda_T$ is
182: the thermal de Broglie length and
183: \begin{equation}
184: \label{Psi}
185: \Psi(h)=16A\frac{h}{(1-2h)^2}
186: \end{equation}
187: 
188: The conditional insertion isotherm can be calculated in the standard way   
189: \begin{equation}
190: \label{muh}
191: \beta \mu(h)=\left(\frac{\partial \beta F(h)}{\partial N} 
192: \right)_{\beta,S,h}=
193: \ln{\left[
194: \frac{\Lambda_T^3\rho(h)}{1-b\rho(h)}
195: \right] 
196: }+\frac{b\rho(h)}{1-b\rho(h)} + \beta\Psi(h)
197: \end{equation}
198: which relates the fluid density $\rho(h)$ and the chemical potential 
199: $\mu(h)$. 
200: %%%%%%%%%%%%%%%%%%%%%%%%
201: \subsection{Matrix dilatation}
202: %%%%%%%%%%%%%%%%%%%%%%%%
203: As is discussed above, the pore dilatation can be taken into account,
204: assuming that the width $h$ is density-dependent. Real insertion materials
205: are usually rather complex (multicomponent and heterogeneous). For that 
206: reason one usually deals with a distribution of pore sizes or with an
207: average size. In this context we assume that $h$ is known only statistically
208: and the probability distribution $f(h|\rho)$ is 
209: conditional\cite{JCPcomp} to the guest density $\rho$, which should 
210: be selfconsistently found from
211: \begin{equation} 
212: \label{rp}
213: \rho=\int dh f(h|\rho) \rho(h). 
214: \end{equation}
215: This relation reflects an injection sequence $\rho(h)$ and the matrix 
216: dilatation. Note that the average density $\rho$ is not uniquely determined
217: by the particles insertion but is also conditional to the matrix response.
218: Namely this constraint makes the problem non-trivial, inducing a coupling
219: between the insertion and the dilatation modes.  
220: Taking our result for $\mu(h)$, we can focus on the 
221: insertion isotherm, averaged over the width fluctuations.  
222: \begin{equation}
223: \label{mu}
224: \mu(\rho)=\int dh f(h|\rho) \mu(h).
225: \end{equation}
226: This, however, requires a knowledge on the distribution $f(h|\rho)$.
227: Even without resorting to a concrete form for $f(h|\rho)$, it is
228: clear that the matrix reaction can be manifested as a change in
229: the distribution width or/and the mean value.  
230: One of the simplest forms
231: reflecting  at least one of these features is a $\delta$-like 
232: distribution, ignoring a non-zero width: 
233: \begin{equation}
234: f(h|\rho)=\delta[h-h(\rho)],
235: \end{equation}
236: where $\delta(x)$ is the Dirac $\delta$-function and the mean pore width
237: $h(\rho)$ is density-dependent. For concreteness we consider the swelling
238: behavior\cite{BisqL}. 
239: Therefore, the mean pore width $h(\rho)$ increases with $\rho$. 
240: For illustration purposes we take the $S$-shaped form, considered 
241: previously\cite{PRB,JCPcomp} 
242: \begin{equation}
243: \label{hr} 
244: h(\rho)=h_0(1+\tanh[\Delta (\rho-\rho_0)]+\tanh[\Delta \rho_0]).
245: \end{equation}
246: This form resembles a non-Vegard behavior,
247: typical for layered intercalation compounds\cite{PRB}. 
248: The dilatation is weak at low densities 
249: ($\rho << \rho_0$), the most intensive 
250: response is at $\rho \approx \rho_0$, and then the pore reaches a 
251: saturation, corresponding to its mechanical stability limit. Here  $\Delta$ 
252: is the matrix response constant or dilatation rate, controlling the 
253: slope near $\rho \approx \rho_0$, and $h_0$ is the mean pore width in the 
254: absence of insertion. This should be considered as a generic form that 
255: samples a non-linear increase of the average pore width, resolved at a given 
256: density $\rho$. The model mimics certain heterogeneity 
257: aspects which, as we believe, are essential. On the other hand, this choice 
258: allows us to map the problem in question onto the negative compressibility 
259: problem\cite{JCPcomp} in systems with fluctuating geometry.   
260:  
261: 
262: From eq.~(\ref{rp}) the average density is found 
263: to be 
264: \begin{equation}
265: \rho=\frac{N}{S}\frac{1}{h(\rho)-1}
266: \end{equation}
267: Changing the surface density $N/S$ (by increments in $N$) we vary the 
268: average pore density $\rho$. This allows us to eliminate $N/S$ in the favor 
269: of $\rho$ in all thermodynamic functions. Combining eqs~(\ref{muh}), 
270: (\ref{rp}) and (\ref{mu}) we obtain 
271: \begin{equation}
272: \label{mur}
273: \beta \mu (\rho)=
274: \ln{\left[
275: \frac{\Lambda_T^3\rho}{1-b\rho}
276: \right] 
277: }+\frac{b\rho}{1-b\rho} + 
278: 16\beta A\frac{h(\rho)}{(1-2h(\rho))^2}
279: \end{equation}
280: In the case of wide pores and a weak dilatation we expand $\beta \mu$ in 
281: terms of $1/h_0$ and $\Delta$, obtaining a generic van der Waals form  
282: \begin{equation}
283: \beta \mu (\rho)=
284: \ln{\left[
285: \frac{\Lambda_T^3\rho}{1-b\rho}
286: \right] 
287: }+\frac{b\rho}{1-b\rho} +
288: \frac{4\beta A}{h_0}  
289: -\frac{12\beta A}{h_0} \rho \Delta
290: \end{equation}
291: Therefore, as in the previous study\cite{JCPcomp}, we have an interplay of 
292: several effects -- the packing (first two terms), the fluid-matrix 
293: interaction  (density-independent), and the dilatation 
294: (linear in density). 
295:  It is seen that a coupling of the fluid-slit repulsion ($A$), spatial 
296:  confinement ($h_0$) and the dilatation ($\Delta$) acts 
297:  as an effective infinite-range fluid-fluid attraction. The latter induces
298:  an inflection point (or even a loop) in the isotherm $\mu(\rho)$. 
299: Introducing a dimensionless temperature $T^*=1/(4A^*)$, and solving
300: \begin{equation}
301: \frac{\partial (\beta\mu)}{\partial \rho}=0; \qquad
302: \frac{\partial^2 (\beta\mu)}{\partial \rho^2}=0
303: \end{equation} 
304: we find the pseudo-critical parameters at which the inflection point appears.
305: \begin{equation}
306: \label{parameters}
307: \rho_c=\frac{1}{3b}; \qquad
308: T^*_c=\frac{4}{27}\frac{\Delta}{bh_0}
309: \end{equation}
310: Plugging these back into eq.~(\ref{mur}) we get the third parameter
311: $\mu_c=\mu_c(b,h_0,\Delta)$. 
312: 
313: As has already been mentioned, the loop is not a signature of
314: a phase transition. It is a sequence of equilibrium states,
315: constrained by a coupling between the insertion and the dilatation modes. 
316: The physical mechanism is quite similar to the previously 
317: reported\cite{JCPcomp}. We deal with a competition of two effects:
318: the interparticle and wall-particle repulsions tend to increase the
319: insertion energetic cost $\mu$ with increasing $\rho$, while the pore 
320: dilatation diminishes it. As a result, at certain parameters the isotherm
321: becomes nonmonotonic. For a purely attractive pore ($A<0$) the inflection
322: point does not appear. Therefore, the wall-particle repulsion is an 
323: essential ingredient. This suggests that a suitable combination of the 
324: repulsive and attractive parts is a criterium for choosing the adsorption 
325: probe. As it  should be, the inflection (loop) feature 
326: disappears in the bulk limit $h_0 \to \infty$ or in the case of rigid 
327: matrices $\Delta \to 0$.  
328: 
329: In order to improve the accuracy of the theoretical results, the 
330: reference part is replaced by the Carnahan-Starling form 
331: \begin{equation}
332: \label{CS}
333: \beta \mu=\ln\left( \rho \Lambda_T^3\right)+
334: \frac{8\eta-9\eta^2+3\eta^3}{(1-\eta)^3}+
335: 16\beta A\frac{h(\rho)}{(1-2h(\rho))^2};
336: \qquad
337: \eta=\pi \rho/6.
338: \end{equation}
339: This allows us to avoid the unphysical behavior at high densities and 
340: escape from the flaws of the excluded volume approximation (\ref{mur}).
341: In particular, all the pseudo-critical parameters ($\rho_c, T_c^*, \mu_c$)
342: appear to be related to the parameters of the pore width distribution
343: ($h_0, \rho_0, \Delta$). Therefore, driving the system close to the 
344: pseudo-critical point permits to estimate the density dependent pore
345: size distribution. 
346: %%%%%%%%%%%%%%%%%%%%%%%%%%%
347: \section{Simulation}
348: %%%%%%%%%%%%%%%%%%%%%%%%%%
349: In order to verify the existence of the loop predicted by the perturbation 
350: theory as well as to test the accuracy of the theoretical estimation, the 
351: simulation of the model in question has been carried out appling the Widom 
352: test particle insertion method \cite{Widom}. As in our previous 
353: work\cite{JCPcomp} we applied the conventional canonical $NVT$ MC 
354: simulations of a confined hard sphere fluid. The simulation cell was 
355: parallelepiped in shape, with parallel walls at surface separation $h$, and 
356: constant surface area $S=L_x \times L_y = 12\times12$. The periodic boundary 
357: conditions were applied to the $X$ and $Y$ directions of the simulation box; 
358: the box length in the $Z$ direction is fixed by the pore width. For a given 
359: pore width the adsorbed fluid density is chosen according to the 
360: Eq.(\ref{rh}) varying the number of particles, $N$. At frequent intervals 
361: during this simulation, a coordinate ${\bf r}_{N+1}$ has been generated 
362: randomly and uniformly over the simulation cell. For this value of 
363: ${\bf r}_{N+1}$, we have computed 
364: $\langle \exp (-\beta H({\bf r}_{N+1})) \rangle_N 
365: $ averaged over all generated trial positions. In such a manner the chemical 
366: potential of the system with density $\rho$ has been calculated, 
367: \begin{equation}
368: \beta \mu = -\ln \left[ 
369: \frac{\rho}{\langle \exp (-\beta H({\bf r}_{N+1})) \rangle_N}
370: \right],
371: \end{equation}
372: where $H$ is the system Hamiltonian.
373:   
374: Each simulation runs $4\times 10^5$ MC cycles, whith the first 
375: half for the system to reach equilibrium 
376: whereas the second half for evaluating the ensemble averages.
377: 
378: %%%%%%%%%%%%%%%%%%%
379: \section{Results}
380: Adsorption isotherms are plotted in Figure~1. It is seen that the
381: inflection point around $\rho \approx 0.3$ transforms into the loop
382: with increasing repulsion $A$ (or decreasing temperature $T^*$).
383: As expected\cite{JCPcomp}, the perturbation theory underestimates the
384: magnitude of $A$ at which these effects appear. The loop gets sharper
385: with increasing dilatation rate $\Delta$ and becomes more pronounced
386: with decreasing pore width $h_0$. 
387: It is remarkable that the fitting on the pseudo-critical (inflection) point 
388: (part (a)) or on the loop (part (b)) allows us to find the set of the 
389: distribution parameters ($h_0$, $\Delta$, $\rho_0$) in a perfect agreement 
390: with the simulation. Attempts to fit under other criteria (e.g. 
391: the least-square deviations) give incorrect values of these parameters.
392: 
393: In agreement with our theoretical prediction (\ref{parameters}), the 
394: simulation results confirm that the pseudo-critical point (and the loop)
395: is reachable at any appropriate choice of the model parameters.  
396: This is in contrast to what was reported\cite{JCPcomp} for the negative 
397: compressibility states in the same model with lateral stretching 
398: (increasing $S$ while $N$ is fixed). 
399: There the loop has been found to appear 
400: only if the slit reaction $\Delta$ reaches some threshold value $\Delta^*$ 
401: which involves a combination of $h_0$, $\rho_0$ and $A$. In other words, 
402: the variation of the average density $\rho$ due to the transversal 
403: dilatation should dominate its variation, induced by the changes in the 
404: surface density $N/S$ due to the lateral stretch. This difference suggests 
405: that the driving path (namely, changing the fluid density by a lateral 
406: stretch or by a particle injection) is really important for this sort of 
407: systems. On the other hand, in the context of the present study, the fact 
408: that the pseudo-critical parameters do not involve any additional condition, 
409: is quite favorable for making a link between these parameters and the matrix 
410: morphology.  
411: 
412: 
413: 
414: In Figure~2 the pseudo-critical parameters are analyzed. It is seen
415: that $\rho_c$ grows with the dilatation rate $\Delta$ almost linearly
416: exhibiting a characteristic kink. The latter signals a crossover from a weak 
417: to a strong dilatation regime. The kink shifts to lower $\Delta$ and becomes 
418: less pronounced with decreasing $h_0$ (that is why it is almost 
419: undetectable for $h_0=5$). At low $\Delta$ the pseudo-critical 
420: density is practically independent of $h_0$, while at high $\Delta$ it 
421: approaches $\rho_0$. In agreement with our simplified estimation 
422: (\ref{parameters}), $T_c^*(\Delta)$ can be reasonably approximated by a 
423: straight line (the lower inset). Nevertheless, it exhibits a weak inflection 
424: point at the same $\Delta$ (around $\Delta=4$) as $\rho_c$ does. The 
425: pseudo-critical temperature incresases with decreasing $h_0$. This is also 
426: coherent with the estimation (\ref{parameters}) predicting a decrease in the 
427: slope inversly proportional to $h_0$. In contrast, the insertion energy 
428: $\mu_c$ decreases and becomes insensitive to $h_0$ with increasing $\Delta$ 
429: (the upper inset). This agrees well with the physical intuition that the 
430: pore dilatation diminishes an overlap between the repulisive wall fields, 
431: making the insertion less energetically consuming. 
432: 
433: 
434: In practice, an exact capture of the inflection point could be too 
435: demanding. Driving  at rather low temperatures ($T<T_c^*$), 
436: such that the system develops the loop, should be much easier. In that case 
437: a shape of the pseudo-spinodals could also be used as an indicator of the 
438: matrix morphology. These curves are plotted in Figure~3. It seen that
439: the pseudo-spinodals become broader with decreasing dilatation rate and   
440: shift to higher $\rho$ with increasing $\Delta$.
441: Comparing this with the previous figure, we can see again that $\rho_c$ 
442: converges to $\rho_0$ with increasing dilatation rate $\Delta$. 
443: If necessary, other quantitative characteristics, such as the curve 
444: asymmetry with respect to $\rho_c$ or $\mu$ values at the spinodal points, 
445: can also be easily extracted.
446: 
447: These results illustrate a correlation between the pseudo-critical point 
448: (or the loop) parameters and the density-dependent pore geometry.   
449: We did not try to obtain these parameters from the 
450: simulation because the procedure is time consuming, while new physical 
451: insights, compared to the anlytical estimations, are difficult to expect.  
452: The only promising point is to consider narrow ($1\le h_0 \le 2$) pores, 
453: for which the first-order perturbative scheme could fail even qualitatively. 
454: This point is left for a future analysis. 
455:  
456:  
457: %%%%%%%%%%%%%%%%%%%%%%%%%
458: %%%%%%%%%%%%%%%%%%%%%%%%%
459: \section{Discussion and Conclusion}
460: %%%%%%%%%%%%%%%%%%%%%%%%
461: A method for characterizing the deformable porous materials 
462: by adsorption of non-critical fluids is proposed. It is based
463: on a driving the host-guest system through a sequence of constrained
464: equilibrium states, forming a loop in the insertion isotherms. The 
465: constraint appears from a coupling of the insertion and the matrix
466: dilatation modes, inducing a competition between the repulsive
467: interactions and their relaxation by the host dilatation. As a result
468: the insertion isotherms develop pseudo-critical points or loops, whose
469: parameters are shown to be correlated with the parameters of the pore
470: size distribution. In particular, we have found that the magnitudes of the 
471: geometrical parameters ($h_0$, $\rho_0$, $\Delta$) which reproduce well the 
472: inflection point or the loop, perfectly agree with their values in the 
473: simulation. This suggests that the peculiar points (with $\partial 
474: \mu/\partial \rho=0$ or $\partial \rho$/$\partial \mu=0$) are extremely 
475: sensitive to the matrix morphology, while the overall isotherm shape is less 
476: important (unless on can achieve an exact isotherm coincidence). 
477: This is coherent with the recent finding\cite{Smit} that a matching
478: of the inflection points (with $\partial \rho$/$\partial \mu=0$) correctly 
479: reproduces the remaining part of the experimental isotherm. From the 
480: information-theoretical point of view\cite{power} such a criterium (the 
481: inflection matching) corresponds to a maximal amount of information one can 
482: get from a fitting of a model to a given experimental or simulation result.  
483: In our case we deal with the pseudo-critical point ($\partial \mu/\partial 
484: \rho=0$). Fitting on this peculiarity corresponds to a singular information 
485: rate\cite{power}, that allows us to update the geometrical parameters from 
486: their initial estimation in a limited range of $\rho$ (around the 
487: singularity). Namely this type of singularity is used to characterize
488: porous media\cite{review} through a shift of the fluid condensation
489: or freezing point.  
490:  
491:  
492: 
493: 
494: Although we have taken a simplified $\delta$-like 
495: distribution, our conclusion should hold in general because the 
496: characteristic loop feature is directly related to the dilatation mode.
497: The latter translates into a shift of the mean pore size, that has been
498: taken into account. Nevertheless, the role of a non-zero distribution width 
499: remains to be investigated. The only restriction is that the number of the 
500: detected loop parameters should be (at least) equal to the number of 
501: parameters determining the pore size distribution. This restriction is not 
502: too strong because in practice the distribution is usually characterized by 
503: only few parameters (e.g. leading mean pore sizes and dipersions around 
504: them). On the other hand, the proposed method is mainly 
505: intended for a complementary detection of the porosity range (e.g. 
506: micropores) where other techniques fail. For illustration and technical 
507: purposes we were working with rather large pores $10 \le h(\rho) \le 30$. 
508: This, however, does not affect our main concept.   
509: 
510: It should be emphasized that, since 
511: we deal with a constrained equilibrium, the loop does not violate the 
512: thermodynamic stability\cite{Everett}. Therefore, this feature can be 
513: readily found under suitable experimental conditions, similar to those
514: realized in the low-pressure adsorption into microporous 
515: matrices\cite{Amarasekera}. 
516: In this respect it is worth noting that quite similar loop
517: features have been detected\cite{Amarasekera} for nitrogen and argon
518: adsorption in microporous carbons and synthetic zeolites.    
519: As in our case, the fluids have been inserted in sequential doses.
520: Nevertheless, in contrast to our assumption, there is no evidence
521: for an insertion-induced dilatation of these slit-shaped materials. 
522: The loops have been interpreted as a signature of a monolayer-induced
523: micropore filling\cite{Amarasekera}. In this context we can propose
524: an alternative interpretation of our results, making a link to the 
525: experimental observations. Namely, it is well known\cite{Amarasekera} that 
526: the micropore filling starts from the narrowest pores ($h_0$ in our laguage).
527: As the fluid pressure (or density) increases, the larger micropores 
528: progressively fill. Microscopically this effect appears from an interplay
529: of repulsive and attractive parts of the wall-particle interaction. The
530: latter depends on the chemical nature of the probe and the matrix geometry
531: and composition. In order to extract general features, one may accept this
532: effect as an experimental fact, trying to see its consequences.  
533: 
534: Therefore, in a given density domain the fluid
535: resolves a quite narrow range of pore sizes because the narrow pores are 
536: already blocked by a preceding dose, but the filling of larger pores is not 
537: thermodynamically favorable at this density. This fact is captured by
538: our distribution $f(h|\rho)$ that can be re-interpreted as an effective
539: density-dependend matrix heterogeneity, with $\Delta$ being the 
540: heterogeneity index. In other words, an increase of the mean pore size
541: $h(\rho)$ may result from a pore blocking effect.   
542: This interpretation is coherent with the one proposed
543: by the authors of \cite{Amarasekera}, who claimed that the S-shaped feature
544: (the loop) signals a switching from the ultramicropore to micropore 
545: filling. Nevertheless, this qualitative analogy between the matrix swelling 
546: and heterogeneity requires further analysis, including a competition of the 
547: repulsive and attractive interactions. This issue is left for a future study.
548: 
549: %%%%%%%%%%%%%%%%%%%%%%%%%%%
550: \begin{thebibliography}{99}
551: \bibitem{review}
552: L.~D.~Gelb, K.~E.~Gubbins, R.~Radhakrishnan, M.~Sliwinska-Bartowiak,
553: Rep.~Prog.~Phys. {\bf 62}, 1573 (1999)
554: \bibitem{Jaroniec}
555: M.~Jaroniec, P.~Brauer, Surf.~Sci.~Rep. {\bf 6}, 65 (1986)
556: \bibitem{avalanches}
557: F.~Detcheverry, E.~Kierlik, M.~L.~Rosinberg, G.~Tarjus,
558: Phys.~Rev.~E {\bf 72}, 051506 (2005)
559: \bibitem{Kierlik2}
560: E.~Kierlik, P.~A.~Monson, M.~L.~Rosinberg, L.~Sarkisov, G.~Tarjus, Phys.
561: Rev. Lett. {\bf 87}, 055701 (2001)
562: \bibitem{PRB}
563:  E. V. Vakarin, J.P. Badiali, M. D. Levi, D. Aurbach, Phys. Rev. B, {\bf 
564:  63}, 014304 (2001)
565: \bibitem{Pablo}
566: F.~A.~Escobedo, J.~J.~de~Pablo, J.~Chem.~Phys. {\bf 106}, 793 (1997) 
567: \bibitem{Herman}
568: T.~Herman, J.~Day, J.~Beamish, 
569: Phys. Rev. B, {\bf 73}, 094127 (2006)
570: \bibitem{BisqL}
571: G.~Garcia-Belmonte, J.~Garcia-Canadas, J.~Bisquert
572: J.~Phys.~Chem.~B {\bf 110}, 4514 (2006) 
573: \bibitem{JCPcomp}
574: E.~V.~Vakarin, Yu.~Duda, J.~P.~Badiali,
575: J.~Chem.~Phys. {\bf 124}, 144515 (2006)
576: \bibitem{Amarasekera}
577: G.~Amarasekera, M.~J.~Scarlett, D.~E.~Mainwaring,
578: J.~Phys.~Chem. {\bf 100}, 7580 (1996)
579: \bibitem{Everett}
580: D.~Everett, Colloids and Surfaces A {\bf 141}, 279 (1998)
581: \bibitem{Widom}
582: B.Widom,  J.~Stat.~Phys. {\bf 19} 563 (1978)
583: \bibitem{Smit}
584: D.~Dubbeldam, S.~Calero, T.~J.~H.~Vlugt, R.~Krishna, T.~L.~M.~Maesen,
585: E.~Beerdsen, B.~Smit, Phys.~Rev.~Lett. {\bf 93}, 088302 (2004)
586: \bibitem{power}
587: E.~V.~Vakarin, J.~P.~Badiali, Phys.~Rev.~E {\bf 74} (2006) (in press)
588: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
589: %%%%%%%%%%%%%%%%%%%%%%%%%%%
590: \end{thebibliography}
591: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
592: \begin{figure}
593: \caption{Insertion isotherms for
594: a weakly (a) and strongly (b) repulsive slit. The other parameters are
595: $h_0=10$, $\Delta=15$, $\rho_0=0.3$ (both theory and simulation)} 
596: \end{figure}
597: %%%%%%%%%%%%%%%
598: \begin{figure}
599: \caption{Variation of the pseudo-critical parameters with the dilatation 
600: rate $\Delta$. $\rho_0=0.3$, $h_0=5$-squares, $h_0=10$-circles 
601:  } 
602: \end{figure}
603: %%%%%%%%%%%%%%%%%%%%
604: \begin{figure}
605: \caption{Pseudo-spinodal curves at different dilatation rates $\Delta$.
606: The other parameters are $h_0=7$, $\rho_0=0.3$} 
607: \end{figure}
608: %%%%%%%%%%%%%%%%%%%%%%%%%%%
609: %\begin{figure}
610: %\caption{}
611: %\end{figure}
612: 
613: %%%%%%%%%%%%%%%%%%%%%%%
614:  \end{document}
615: