Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Freezing point depression and freeze-thaw damage by nano-fuidic salt trapping

Freezing point depression and freeze-thaw damage by nano-fuidic salt trapping Freezing point depression and freeze-thaw damage by nano- uidic salt trapping Tingtao Zhou Massachusetts Institute of Technology, Department of Physics Mohammad Mirzadeh Massachusetts Institute of Technology, Department of Chemical Engineering Roland J.-M. Pellenq The MIT / CNRS / Aix-Marseille University Joint Laboratory, "Multi-Scale Materials Science for Energy and Environment" and Massachusetts Institute of Technology, Department of Civil and Environmental Engineering Martin Z. Bazant Massachusetts Institute of Technology, Department of Chemical Engineering and Massachusetts Institute of Technology, Department of Mathematics A remarkable variety of organisms and wet materials are able to endure temperatures far below the freezing point of bulk water. Cryo-tolerance in biology is usually attributed to \anti-freeze" proteins, and yet massive supercooling (< 40 C) is also possible in porous media containing only simple aqueous electrolytes. For concrete pavements, the common wisdom is that freeze-thaw (FT) damage results from the expansion of water upon freezing, but this cannot explain the large pressures (> 10 MPa) required to damage concrete, the observed correlation between pavement damage and de-icing salts, or the FT damage of cement paste loaded with benzene (which contracts upon freezing). In this work, we propose a di erent mechanism { nano uidic salt trapping { which can explain the observations, using simple mathematical models of dissolved ions con ned between growing ice and charged pore surfaces. When the transport timescale for ions through charged pore space is prolonged, ice formation in con ned pores causes enormous disjoining pressures via the ions rejected from the ice core, until their removal by precipitation or surface adsorption at a lower temperatures releases the pressure and allows complete freezing. The theory is able to predict the non-monotonic salt-concentration dependence of FT damage in concrete and provides some hint to better understand the origins of cryo-tolerance from a physical chemistry perspective. I. INTRODUCTION rectly caused by solid phase transformations, not only ice 18,19 22,23 formation , but also salt crystallization . Interest- ingly, there is strong correlation between FT damage and The durability of wet porous materials against freeze- the use of de-icing salts on concrete pavements , which thaw (FT) damage is critical in many areas of science and are often less durable than concrete structures without engineering. In biology, it is a matter of life and death. salt exposure in the same cold climates. Moreover, FT Living cells must somehow maintain a liquid state within damage only occurs when the water saturation level ex- 1{3 the cellular membrane during winter , while avoiding ceeds a critical value . Previous models of \frost heave" anoxia due to external ice encasement . Various anti- (see e.g. ref ) achieved some successes in explaining the freeze proteins have been identi ed in cryo-tolerant ani- deformation of saturated soils due to the dynamics of pre- mals, and cryo-protectant chemicals have been used for melted liquid and its coupling with the solid. However, 2,5{7 cryo-preservation and in vitro fertilization . In addi- the applicability of these theories to hardened cement is tion, the complex thermodynamics of supercooled water questionable, due to its much higher sti ness compared could play a role. Even in bulk water, deep supercool- 27,28 to capillary stresses . In summary, despite the soci- 8{10 ing can lead to multiple metastable disordered states . etal importance of FT damage in cement, a physics-based 11,12 Phase transitions under nano-con nement can lead theory has not yet been developed that can predict the to exotic new phases, as well as modi ed ice nucleation, enormous pressures required (> 10 MPa), as well as all 13{15 16,17 in both experiments and molecular simulations puzzling observations above. of water in nanopores. In engineering, the most familiar example of FT dam- In this article, we develop a predictive theory of age is the fracture of concrete pavements during the freezing-point depression and FT damage in charged winter , commonly attributed to the expansion of water porous media, based on a simple new mechanism 19,20 transforming to ice within the pores . However, this sketched in Figure 1: nano uidic salt trapping. It is contradicts the observation that FT damage occurs when well known in colloid science that, when two charged sur- cement is loaded with benzene , a normal liquid that faces are separated by a liquid electrolyte, the crowding shrinks upon freezing. Recent experiments have chal- of ions in solution results in large repulsive forces when- lenged the prevailing hypothesis that FT damage is di- ever the electric double layers overlap, at the scale of arXiv:1905.07036v2 [cond-mat.soft] 8 Dec 2019 2 Figure 1: Physical picture of nano uidic salt trapping. (a) In an open pore, where ions and water molecules can easily exchange with a nearby reservoir, no signi cant pressure or freezing point depression is predicted. (b) Nanoscale bottlenecks, especially in poorly connected porous networks, with charged surfaces can signi cantly hinder this exchange by size or charge exclusion of the co-ions, and counter-ions are forced to stay to maintain charge neutrality. (c) Once ice nucleates, even an initially open pore will eventually trap a nanoscale thin lm of supercooled, concentrated electrolyte near the charged surface, until surface ion condensation or solid salt precipitation occurs. (d) In biological cells, the charged cytoskeleton (indicated by bers) could enable such passive nano uidic salt trapping, while further active control of water and ion ux across the cell membrane is performed by ion channels and pumps. In all cases, nano uidic salt trapping can lead to dramatic supercooling and, once ice nucleates, severe damage to the solid matrix. the Debye screening length (1-100nm in water). This the pore surface charges and preserve overall electroneu- \disjoining pressure" is responsible for the stabilization trality. Ions in solution mediate surface forces , which of colloidal dispersions in aqueous electrolytes , surface play a crucial role in the mechanical properties of con- 30 32,33,42{45 forces in clays and other porous media , and the elec- crete and the function of biological systems. In trostatic properties of membranes . Disjoining pressure most cases, the ions are assumed to have negligible solu- has been successfully modeled by the Poisson-Boltzmann bility in the frozen solid, as is the case with pure ice. mean- eld theory for solutions of monovalent ions, and Freezing begins in the larger \macropores" (> 100 extensions are available to describe correlation e ects in- nm), where bulk water easily transforms to ice, slightly 32,33 volving multivalent ions . Here we treat the disjoin- below the thermodynamic melting point of the solution, ing pressure between ice core and the charge pore surface which may be depressed from that of the pure solvent with a mean- eld approximation. Although the physics by the dissolved salt and any anti-freeze solutes. This of electrolyte freezing under con nement has been consid- bulk ice can form by homogeneous nucleation, spinodal ered for nanoporous materials , we propose that nano- decomposition, or (most likely) by heterogeneous nucle- uidic salt trapping is the key mechanism for large su- ation on impurities. Regardless of its origin, the advanc- percooling and FT damage in cement and other charged ing ice rejects ions, causing the salt concentration to rise nano-porous materials. This physical picture is consis- in the nearby, increasingly con ned liquid electrolyte. tent with all the available experimental evidence for con- What happens next depends on the degree of super- crete. cooling, the surface charge and importantly the pore connectivity. As shown in Figure 1(a), even after par- tial freezing, an individual pore may remain open, allow- A. Physical Picture ing ions and water molecules to exchange freely with a reservoir of bulk solution via a percolating liquid path 46{50 to neighboring unfrozen pores or an external bath . Consider a heterogeneous porous material saturated In this scenario, the liquid electrolyte and any solid ice with liquid and subjected to continuously decreasing within the pore remain in quasi-equilibrium with the bulk temperatures. As in most organisms and construction reservoir at constant chemical potential. The connected materials, suppose that the pore surfaces and suspended 35{41 path to the reservoir may pass through liquid-saturated materials are hydrophilic and charged , e.g. by pores, or partially frozen pores with suciently thick liq- the dissociation of surface functional groups or the ad- uid lms to allow unhindered transport. sorption of charged species. The large capillary pores (>5 nm) are typically also lled with water, but can be As freezing proceeds, many ions and water molecules replaced with other uids such as benzene. However, the will inevitably be trapped out of global equilibrium, small \gel" pores (1-5 nm) are always lled with liquid although still in local quasi-equilibrium within each water due to the strong surface charge and hydrophilic- nanoscale pore. The simplest case is that of a pore ity even in benzene-loaded cement samples. Importantly, connected to external reservoir only via a bottleneck, the liquid must contain dissolved salts, possibly at low sketched in Fig. 1(b). Water molecules that are not concentration, as well as excess counter-ions to screen closely associated with ions can still go through the bot- 3 tleneck, with a possibly di erent viscosity. The bottle- thus allowing complete freezing of the pores. Ions may neck may block solvated ions (with their solvation shells) also be cleared by adsorption reactions on the pore sur- from passing by steric hindrance or charge exclusion. face, which regulate and neutralize the surface charge. Even if some solvated ions can di use through a given bottleneck, their electrokinetic transport rate may be II. THEORY too slow to allow many to escape prior to more com- 51{53 plete freezing . Such slow ion transport may be en- hanced by long, tortuous pathways through a series of As mentioned above, under the assumption of sepa- 54{57 bottlenecks and compounded by a large volume of rated timescales for ion transport and freezing, we ap- micropores, e ectively cut o from the macropores with proximate the dynamic problem as a quasi-equilibrium insucient time for salt release, in materials of low pore- problem: in the limit of free ions, ion and water trans- space accessivity . Even in relatively well connected port is much faster than freezing; in the other limit of porous structures, nano uidic salt trapping can also re- trapped ions, ion transport is much slower than freez- sult from bottlenecks created by the advancing ice, as ing. The solutions of both limits can be uni ed in the shown in Figure 1(c), where the larger open pore on same quasi-equilibrium mean- eld framework. Below we the right side freezes almost completely rst. Due to the present details of these solutions. surface hydrophilicity, a supercooled liquid lm often re- The mean- eld free energy for a liquid electrolyte and mains between the pore surface and the ice core prior its frozen solid inside a charged pore can be described by 11,16,58 to complete freezing , which is now the only path- F = F + F + F way for water and ions in the smaller pore shown on the tot liquid solid interface left side. As temperature decreases further, ice forma- s 2 = dV   krk s l tion starts in the left pore, but transport of solvated ions through the thin liquid lm is now slow, and the entropy h i l 2 (1) of these con ned ions builds up a pressure. In biological + dV g(fc g) +  krk cells, as shown in Figure 1(d), electrolytes are contained l within the cell walls, and nano uidic salt trapping is facil- + dS ( + q ) j j itated by the charged cytoskeleton and abundant charged j=s;l;sl macromolecules (including cryo-resistant proteins). In- ternal salt concentrations are also actively maintained where the integrations are performed over volumes of by ion channels and pumps in the cell membrane . solid (V ) and liquid (V ) with permittivities  and  , re- s l s l To quantitatively calculate the timescales of freezing spectively, and over surfaces of the solid-liquid interface and ion transport, one needs to solve a proper electroki- (S ), the liquid-pore interface (S ) and the solid-pore sl l netic model of the 3D charged pore structure, with in- interface (S ), with corresponding surface charge densi- formation of the tortuosity and connectivity in addition ties, q , q and q and interfacial tensions, , and ; sl l s sl l s to the pore sizes. Here we focus on the asymptotic be- is the bulk chemical potential di erence between s l havior of very long ion transport timescale v.s. freez- solid and liquid phases; r is the electric eld; g(fc g) ing timescale, which hereafter referred to as the limit the non-electric part of homogeneous liquid electrolyte of trapped ions. This approximation of timescale sep- free energy; c the concentration of ion species i having aration is similar in essence to the adiabatic or Born- charge z e; and  = z ec the net charge density, as- i i i Oppenheimer approximation , where the short time sumed to be negligible in the solid phase. We focus on quasi-equilibrium is solved|as we show in the next situations of complete wetting by the liquid,  , s l sl sections|neglecting the slowly changing physics, in our in which case we can neglect S and assume S covers the s l case the ion transport. The phenomenon of ion trap- entire pore surface. ping in charged nanochannels, while water remains free Setting F = = 0 for bulk and surface variations, tot to di use and ow to a nearby reservoir or larger we obtain Poisson's equation pore, is well established in the eld of nano uidics and forms the basis for various devices, such as electro-  r  =  in V l l 61 (2) osmotic micropumps , nano- uidic diodes and bipolar r  = 0 in V s s 54,62,63 64 transistors , and nano uidic ion separators . The supercooling of con ned liquids can be greatly en- and electrostatic boundary conditions hanced by the salt rejected by freezing, as the remain- ~ ~ q = ( E  E ) n ^ on S ing solution becomes more concentrated inside a trapped sl s s l l ls sl (3) freezing pore. Large disjoining pressures are then pro- q =  r n ^ on S l l l l duced in the very concentrated liquid solution and trans- mitted to the solid matrix, potentially causing damage. The equilibrium state of liquid-solid coexistence is At suciently low temperatures, salt-enhanced super- found by minimizing the total free energy with respect cooling and freeze-thaw pressure are relieved by the sud- to the position and shape of the solid-liquid interface, den precipitation of ions from the concentrated liquid, S . Here, we consider two cases: (1) an open pore sl 4 where ions of species i exchange freely with a reservoir by minimizing the total free energy with respect to r, of concentration c , or (2) a pore with trapped ions, F =r = 0: tot whose total number is xed by screening the pore sur- r = argmin F (r) (5) tot face charge in the liquid, prior to freezing, by the mech- anisms shown in Fig. 1. Importantly, we neglect the ef- which yields the equilibrium ice volume fraction,  = fects of volume changes due to the water/ice transforma- tion, under the assumption that liquid water molecules (r =R) . Once r is found, mechanical equilibrium at the solid-liquid interface gives pressure of both phases, (of size  3A) are mobile and small enough to escape the pore as freezing progresses, regardless of whether sol- which is transmitted to the pore boundary vated ions are trapped. In contrast to the common wis- @F @F solid liquid dom about freeze-thaw damage in pavements, this pic- P = = (6) @r @r ture must also hold for well-connected hierarchical porous r=r r=r materials such as concrete. The rst equality describes the tendency to form more The preceding thermodynamic framework for con ned ice and hence expand its volume, while the second equal- electrolyte phase transformations can be extended in var- ity shows the free energy cost to squeeze the electrolyte, ious ways, e.g. to account for ion-ion correlations (es- resisting the growth of ice. pecially involving multivalent ions), nite ion sizes and For a symmetric pore, after freezing starts, the free 67,68 hydration surface forces , but here we focus on the energy of ice is given by simplest Poisson-Boltzmann mean- eld theory , which suces to predict the basic physics of freezing-point de- F = (  ) V (d)r ; (7) ice s l pression and material damage. The homogeneous free en- ergy is then given by the ideal gas entropy for point-like where (  ) is the Gibbs free energy change per vol- s l ume for bulk water freezing, which can be calculated ions, g = c [ln(v c ) 1], with v the molecular volume, i i i i i and the electrostatic potential in the liquid electrolyte is using the Gibbs-Helmholtz relation, as shown in ref. . In principle, the electric eld energy of the ice core (x < r) then given by the Poisson-Boltzmann (PB) equation: depends on its shape and the electrostatic boundary con- 2 1 z e ditions, but vanishes here by symmetry. The interfacial r  =  = z ec ; c = c e (4) l l i i i d1 energy is F = S(d)r , which gives rise to the i surface sl Gibbs-Thomson e ect of freezing point depression for Since we focus on highly con ned electrolyte liquid lms, con ned pure water. The free energy of the electrolyte we set the relative permittivity,  = 10 , to that of wa- l 0 shell is given by 69,70 ter near dielectric saturation at high charge density . elec To assess the prevalence of nano uidic salt trapping d1 = q (R)R + within Poisson-Boltzmann theory, the state of a bottle- S(d) (8) R h i neck shown in Fig. 1 can be estimated by comparing l 2 d1 x dx g(fc g) +  krk the double layer thickness  (or hydrated ion size a) inside with its radius R: if   R (or a & R) then the double layer(s) span across and the bottleneck is ap- where the rst term is the electrostatic energy of surface proximated as \closed" to ions, since freezing rate may charges, and the integrand takes the form given above for exceed ion transport rate, given a high tortuousity of the mean- eld theory of point-like ions. To summarize, we pore network. If   R (or a . R), then the channel D are solving a free boundary problem where the liquid-ice may be viewed as open to ion exchange. For an initial boundary position r is unknown beforehand. We adopt salt concentration of 0:1 mol=L in a binary monovalent a numerical algorithm to search for the r that minimizes electrolyte, (with relative permittivity   10), we nd r total free energy at a given temperature T , surface charge 4 k T l B density q and initial salt concentration c : 0:5 nm. D 2 2c e 1. starting from r = 0, compute the total free energy F (0). A. Symmetric pores 2. increment r by a small amount dr, compute the to- tal free energy F (r). In order to obtain analytical results, we consider when computing the total free energy at a given isotropic electrolyte freezing in d dimensions, where ice r value, we always solve the Poisson-Boltzmann nucleates to form a plate (d = 1), cylinder (d = 2) or Eqn. 4 to obtain the electric potential pro le , sphere (d = 3) of radius r within a pore of the same and insert into the integration of Eqn. 8. symmetry, whose surface is located at x = R. The total d d1 pore volume is V (d)r , and S(d)r the surface area of 3. after sweep r from 0 to the pore size R, nd the the ice core (x < r), surrounded by a liquid electrolyte minimum of F and the corresponding r gives the shell (r < x < R). At thermodynamic equilibrium, position of the quasi-equilibrium ice front. the location r of the solid-liquid interface is determined 5 Figure 2: Electrolyte freezing and pressure generation in a parallel slit pore (d = 1) with free ions exchanging with a reservoir. There is no e ects of interfacial tension. The freezing point depression, T  0:1 K, Figure 3: In contrast to Fig.2, for a binary electrolyte and disjoining pressure, P  0:1 MPa, are quite small, with trapped ions, freezing-point depression as large as in the limit of one-component plasma of only -40 K can occur. The quasi-equilibrium approximation counter-ions. In this case the total number of gives a continuous freezing temperature range marked counter-ions is determined by the surface charge density by two T values: the temperature to start freezing, T , only and does not depend on pore size. And the T f and that of complete freezing of the pore T , when ff and P only depends on the distance between ice front ions are removed by precipitation. and the pore surface, which denoted by L = 5 nm here. B. Free ions As freezing proceeds in an open pore, where all ions can escape to a reservoir, the surface charge is eventually screened in a thin liquid lm containing only counter-ions, which corresponds to one component 73,74 plasma (OCP) . The Poisson-Boltzmann equation for the OCP can be integrated for symmetric pore shapes to obtain the mean electrostatic potential. For a slit pore (d = 1), we obtain to the rst order 4 k T Z T l B 0 jPj  Q 2 2 q e T s 0s 1 Figure 4: Large disjoining pressures up to  10 MPa (9) ~ ~ jPj jPj Rqe occur during the freezing process, below the @ A tan = 2 2 2 4 k T temperature to start freezing, T , and above that of l B f complete freezing of the pore T . The range of ff where Q is the latent heat of bulk water freezing, T pressure is marked by P and P , correspondingly. f ff Blue and red dots are numerical results, while the solid the bulk freezing point, and T = T T the freezing point depression. Notice that for OCP limit, the total lines connecting them are guiding the eyes. amount of counter-ions does not depend on the pore size, but is simply determined by the surface charge density. Hence, the quasi-equilibrium solution only depends on the distance between the ice front and the pore surface, which we here denote as L. spherical (d = 3) pores are even smaller than in a slit Inserting typical values, we can estimate the freezing pore (d = 1) and may often be neglected compared to point depression in the slit pore as T  0:1 K and the Gibbs-Thompson e ect of interfacial tension in such the pressure as P  0:1MPa. In this case, the freez- curved geometries. In general, if excess salt ions (and wa- elec ing point is only depressed by . 1 K, and no signi cant ter molecules) are free to escape the pore during freezing, pressure is generated, as shown in Fig.2. As shown in then we expect very little freeze-thaw damage in a wet ref. , the e ects of ions in open cylindrical (d = 2) or porous material. 6 C. Trapped ions The situation is completely di erent in the opposite limit, where all ions in the original liquid binary elec- trolyte remain trapped within the pore during freezing. Total ion number conservation is then imposed on the d1 PB equations, c S(d)x dx = N , and signi cant i i freezing-point depression can be achieved. The math- ematical details can be found in a companion paper , and here we focus on explaining the physical predictions of the theory. To separate the e ect of curvature, here we focus on the slit symmetry (d = 1). First we consider a binary 1:1 liquid electrolyte freezing in a parallel slit pore (d = 1). In this case, there is no ef- fect of solid-liquid interfacial tension, as the interface area does not change as ice front advances (zero curvature). As shown in Fig.3, the freezing point is substantially de- creased by increasing the initial salt concentration c in the con ned liquid. After freezing starts at temperature T , due to the resistance of the electrolyte, the equilib- rium ice volume fraction  monotonically increases as temperature decreases. The freezing process continues until the trapped ions are suddenly removed from the thin liquid lm at the temperature of freezing nished T , when the salt solubility limit is reached, and  sud- ff denly jumps to 1. The pore is completely frozen now. Complete freezing may also occur if the trapped ions are adsorbed on the pore surface, thereby neutralizing the Figure 5: (a) typcial experiment protocol reproduced surface charge (as shown below). from . T and T correspond to the temperature As shown in Figure 4, signi cant disjoining pressures f ff when ice formation initiates and solubility limit ( 10 MPa for R = 5 nm) can be generated by con- reached, as indicated in Fig.3. (b) shaded area shows ned ions during the freezing process. The pressures the pressure range after freezing starts in a trapped at the freezing start temperature T and the complete pore. P and P correspond to the pressure when ice freezing temperature T are labeled as P and P , re- f ff ff f ff formation initiates and solubility limit reached, as spectively. The disjoining pressure varies approximately indicated in Fig.4. Dash line shows open pores with free linearly with temperature between these values during ions, which is close to the horizontal line of 0. Data the freezing process in a slit pore. points show measured damage in cement paste FT experiment , a non-monotonic function of salt concentration. Tensile strength of hardened cement D. Salt solubility limit and surface charge paste is  3 MPa. regulation As ice volume fraction increases, salt concentration goes up. At some point the concentrated electrolyte will 22,23 become saturated and salt will crystallize. The volume of opposed to the concept of \crystallization pressure" salt crystal precipitate is neglected. The solubility equi- that has been proposed to account for pressure and dam- librium for 1:1 electrolyte (M + B ) at saturation is age (under room temperature) in construction materials, + sat [M ][B ] c here the pressure of the freezing pore is determined by K = = . Here c is the concen- eq solid [MB] c solid thermodynamc equilibrium between freezing and precip- tration in solid crystal phase, which is typically regarded itation, thus salt crystallization is merely a consequence, sat as constant 1. Once c , the saturated concentration of instead of the cause for pressure. salt ions, is reached the equilibrium position of ice front becomes thermodynamically unstable and all the liquid When the concentration of trapped ions is high enough, turned into solid phases of ice and salt crystal. In Fig.3, counter-ion recombination with the surface charge be- all the curves at some point reach the solubility limit and comes important. This e ect can be included by a modi- undergo sudden crystallization, when ice volume fraction ed boundary condition for the PB equations, where the discontinuously jumps from  < 1 to  = 1. The pressure surface charge is computed self-consistently based on a 75 71 at this point is denoted as P in both Fig.4 and Fig.5. As charge regulation model (see more details in ). ff 7 III. APPLICATION TO CONCRETE and complete freezing. Freezing point depression, ice vol- ume fraction and pressure are calculated using a simple mean- eld theory. The predictions of this theory are semi-quantitatively consistent with experimental observations of freeze-thaw Many extensions of the theory could be considered damage in cement. Below critical degree of water satu- in future work. Ion correlations, including the strong- 76{78 ration, plenty of large pores remain open transport path- coupling limit , can be introduced via higher order ways for ions during freezing, hence no signi cant dam- terms in Eqn.8, resulting in modi ed PB equations . age observed . The volume expansion of water during At larger length scales, models of interfacial instabilities 79{81 freezing is irrelevant in this theory, so it can also explain leading to dendritic growth could be extended to ac- 82,83 qualitatively similar results observed in freeze-thaw ex- count for electrokinetic phenomena in charged pores . perimetns on cement samples loaded with benzene, which Here we always assume bulk phase of ice (the Ih phase) shrinks upon freezing . The non-monotonic dependence is formed, since the freezing conditions discussed here of damage on NaCl concentration can be explained by (T > 200 K , P < 100 MPa, d < 100 nm) are not very crossover from salt trapping to channel opening though extreme. Exotic phases of ice (non-Ih phases) are known 9,84{89 charge regulation, as shown in Fig.5. A fully quantitative to dominate under more extreme conditions . Salt comparison requires the plasticity and fracture mechan- ions can also a ect the surface tension of ice-electrolyte ics of the solid matrix due to these local high pressures, interface, as well as other aspects of nucleation under and the connectivity of the pores, which is currently a con nement, described in a companion paper . missing link. Also to quantify the transport timescales As a rst application to material durability, our theory for ions as freezing proceeds, pore connectivity is key in- is consistent with complex trends of freeze-thaw damage formation. Nevertheless, to our knowledge for the rst in hardened cement paste. These predictions could in u- time this mechanism shows potential to encompass all ence industrial practices in road de-icing and pavement these observations. design. The theory may also provide some perspective on the physics of cryo-tolerance and cryo-preservation in biological materials, which abound in electrolyte-soaked IV. CONCLUSION AND DISCUSSIONS macromolecules, nanopores and membranes. V. ACKNOWLEDGEMENT In this article, we present a theory of the freezing of electrolytes in charged porous media. The key insight is that, if ions become trapped by the advancing ice This work is carried out with the support of the Con- front, large disjoining pressures can cause material dam- crete Sustainability Hub at MIT. The authors thank S. age, until further supercooling triggers salt precipitation Yip, C. Qiao, J. Weiss, M. Pinson for useful discussions. 1 13 C. L. Guy, Annual review of plant biology 41, 187 (1990). K. Morishige and H. Iwasaki, Langmuir 19, 2808 (2003). 2 14 J. Saragusty and A. Arav, Reproduction 141, 1 (2011). M. Nosonovsky and B. Bhushan, Physical Chemistry J. M. Storey and K. B. Storey, in Life in the frozen state Chemical Physics 10, 2137 (2008). (CRC press, 2004) pp. 269{300. K. V. Agrawal, S. Shimizu, L. W. Drahushuk, D. Kilcoyne, C. Andrews, Annals of Botany 78, 529 (1996). and M. S. Strano, Nature nanotechnology 12, 267 (2017). 5 16 F. F. AbdelHafez, N. Desai, A. M. Abou-Setta, T. Falcone, P. A. Bonnaud, B. Coasne, and R. J. Pellenq, Journal of and J. Goldfarb, Reproductive biomedicine online 20, 209 Physics: Condensed Matter 22, 284110 (2010). (2010). S. Cerveny, F. Mallamace, J. Swenson, M. Vogel, and R. Wakchaure, S. Ganguly, S. Sharma, P. K. Praveen, L. Xu, Chemical reviews 116, 7608 (2016). M. Sharma, and T. Mahajan, Int. J. Phar. & Biomedi. Y. Farnam, D. Bentz, A. Sakulich, D. Flynn, and J. Weiss, Rese 2, 11 (2015). Advances in Civil Engineering Materials 3, 316 (2014). 7 19 L. Rienzi, C. Gracia, R. Maggiulli, A. R. LaBarbera, D. J. H. Cai and X. Liu, Cement and Concrete Research 28, Kaser, F. M. Ubaldi, S. Vanderpoel, and C. Racowsky, 1281 (1998). Human reproduction update 23, 139 (2017). P. C. Association, \Freeze-thaw resistance," (2018). 8 21 O. Mishima and H. E. Stanley, Nature 396, 329 (1998). J. J. Beaudoin and C. MacInnis, Cement and Concrete C. Tulk, C. Benmore, J. Urquidi, D. Klug, J. Neuefeind, Research 4, 139 (1974). B. Tomberli, and P. Egelsta , Science 297, 1320 (2002). E. M. Winkler and P. C. Singer, Geological society of P. G. Debenedetti, Journal of Physics: Condensed Matter America bulletin 83, 3509 (1972). 15, R1669 (2003). M. Steiger, Journal of crystal growth 282, 455 (2005). 11 24 L. D. Gelb, K. Gubbins, R. Radhakrishnan, and Y. Farnam, D. Bentz, A. Hampton, and W. Weiss, Trans- M. Sliwinska-Bartkowiak, Reports on Progress in Physics portation Research Record: Journal of the Transportation 62, 1573 (1999). Research Board , 81 (2014). 12 25 D. T. Limmer and D. Chandler, The Journal of chemical W. Li, M. Pour-Ghaz, J. Castro, and J. Weiss, Journal of physics 137, 044509 (2012). Materials in Civil Engineering 24, 299 (2011). 8 26 56 J. Wettlaufer and M. G. Worster, Annu. Rev. Fluid Mech. H.-C. Chang, G. Yossifon, and E. A. Demekhin, Annual 38, 427 (2006). review of uid mechanics 44, 401 (2012). 27 57 T. Zhou, K. Ioannidou, E. Masoero, M. Mirzadeh, R. J.-M. G. Yossifon, P. Mushenheim, Y.-C. Chang, and H.-C. Pellenq, and M. Z. Bazant, Langmuir 35, 4397 (2019). Chang, Physical Review E 79, 046305 (2009). 28 58 T. Zhou, K. Ioannidou, F.-J. Ulm, M. Z. Bazant, and R. Denoyel and R. Pellenq, Langmuir 18, 2710 (2002). R.-M. Pellenq, Proceedings of the National Academy of H. Bertil, Ionic channels of excitable membranes (Sinauer Sciences 116, 10652 (2019). Sunderland, MA, 1992). 29 60 J. Lyklema, Fundamentals of Interface and Colloid Sci- M. Born and R. Oppenheimer, Annalen der physik 389, ence. Volume II: Solid-Liquid Interfaces (Academic Press 457 (1927). Limited, San Diego, CA, 1995). S. Zeng, C.-H. Chen, J. C. Mikkelsen Jr, and J. G. Santi- J. N. Israelachvili, Intermolecular and Surface Forces (Aca- ago, Sensors and Actuators B: Chemical 79, 107 (2001). demic Press, New York, NY, 2011). H. Daiguji, Y. Oka, and K. Shirono, Nano Letters 5, 2274 D. Andelman, in Handbook of Biological Physics (Elsevier, (2005). 1995) pp. 603{641. G. Yossifon, Y.-C. Chang, and H.-C. Chang, Physical re- R. J.-M. Pellenq and H. Van Damme, Mrs Bulletin 29, 319 view letters 103, 154502 (2009). (2004). D. Gillespie and S. Pennathur, Analytical chemistry 85, R.-M. Pellenq, J. Caillol, and A. Delville, The Journal of 2991 (2013). Physical Chemistry B 101, 8584 (1997). M. Z. Bazant, B. D. Storey, and A. A. Kornyshev, Physical C. Alba-Simionesco, B. Coasne, G. Dosseh, G. Dudziak, Review Letters 106, 046102 (2011). K. Gubbins, R. Radhakrishnan, and M. Sliwinska- M. Z. Bazant, M. S. Kilic, B. Storey, and A. Ajdari, Ad- Bartkowiak, Journal of Physics: Condensed Matter 18, vances in Colloid and Interface Science 152, 48 (2009). R15 (2006). K. Bohinc, A. Shrestha, M. Brumen, and S. May, Physical K.-M. Jan and S. Chien, The Journal of general physiology Review E 85, 031130 (2012). 61, 638 (1973). M. A. Brown, G. V. Bossa, and S. May, Langmuir 31, J. Visser, (1973). 11477 (2015). 37 69 R. Giese, Clays and Clay Minerals 21, 145 (1973). F. Booth, The Journal of Chemical Physics 19, 391 (1951). 38 70 S. McLaughlin, in Current topics in membranes and trans- M. Aguilella-Arzo, A. Andrio, V. M. Aguilella, and port, Vol. 9 (Elsevier, 1977) pp. 71{144. A. Alcaraz, Physical Chemistry Chemical Physics 11, 358 S. L. Swartzen-Allen and E. Matijevic, Chemical Reviews (2009). 74, 385 (1974). T. Zhou, M. Mirzadeh, R. J.-M. P. Dimitrios Fraggedakis, A. J. Allen, J. J. Thomas, and H. M. Jennings, Nature and M. Z. Bazant, \Theory of freezing point depression in materials 6, 311 (2007). charged porous media," In preparation. 41 72 R. J.-M. Pellenq, A. Kushima, R. Shahsavari, K. J. J. W. Gibbs, (1879). Van Vliet, M. J. Buehler, S. Yip, and F.-J. Ulm, Pro- S. Brush, H. Sahlin, and E. Teller, The Journal of Chem- ceedings of the National Academy of Sciences 106, 16102 ical Physics 45, 2102 (1966). (2009). M. Baus and J.-P. Hansen, Physics Reports 59, 1 (1980). 42 75 A. Delville, N. Gasmi, R. J. Pellenq, J. M. Caillol, and T. Markovich, D. Andelman, and R. Podgornik, EPL (Eu- H. Van Damme, Langmuir 14, 5077 (1998). rophysics Letters) 113, 26004 (2016). 43 76 R. Jellander, S. Mar celja, and J. Quirk, Journal of Colloid A. G. Moreira and R. R. Netz, EPL (Europhysics Letters) and Interface Science 126, 194 (1988). 52, 705 (2000). 44 77 J. Stankovich and S. L. Carnie, Langmuir 12, 1453 (1996). R. R. Netz, The European Physical Journal E 5, 557 J. P. Valleau and L. K. Cohen, The Journal of chemical (2001). physics 72, 5935 (1980). L. Samaj and E. Trizac, Physical Review E 84, 041401 S. Torquato, Random heterogeneous materials: microstruc- (2011). ture and macroscopic properties, Vol. 16 (Springer Science M. B. Hastings and L. S. Levitov, arXiv preprint cond- & Business Media, 2013). mat/9607021 (1996). 47 80 J. Van Brakel, Powder technology 11, 205 (1975). W. W. Mullins and R. F. Sekerka, Journal of applied J. A. Quiblier, Journal of Colloid and Interface Science 98, physics 34, 323 (1963). 84 (1984). W. W. Mullins and R. Sekerka, in Dynamics of Curved M. B. Pinson, T. Zhou, H. M. Jennings, and M. Z. Bazant, Fronts (Elsevier, 1988) pp. 345{352. Journal of colloid and interface science (2018). M. Z. Bazant, J. Choi, and B. Davidovitch, Physical re- Z. Gu and M. Z. Bazant, arXiv preprint arXiv:1808.09804 view letters 91, 045503 (2003). (2018). M. Mirzadeh and M. Z. Bazant, Physical review letters P. Biesheuvel and M. Bazant, Physical Review E 94, 119, 174501 (2017). 050601 (2016). L. G. Dowell and A. P. Rinfret, Nature 188, 1144 (1960). 52 85 P. Peters, R. Van Roij, M. Z. Bazant, and P. Biesheuvel, D. Steytler, J. Dore, and C. Wright, The Journal of Phys- Physical review E 93, 053108 (2016). ical Chemistry 87, 2458 (1983). 53 86 J. Catalano, R. Lammertink, and P. Biesheuvel, arXiv E. Mayer and A. Hallbrucker, Nature 325, 601 (1987). preprint arXiv:1603.09293 (2016). B. J. Murray, D. A. Knopf, and A. K. Bertram, Nature G. Yossifon, I. Frankel, and T. Miloh, Physics of Fluids 434, 202 (2005). 18, 117108 (2006). E. B. Moore and V. Molinero, Physical Chemistry Chem- S. Pennathur and J. G. Santiago, Analytical chemistry 77, ical Physics 13, 20008 (2011). 6772 (2005). J. Finney, A. Hallbrucker, I. Kohl, A. Soper, and D. Bowron, Physical review letters 88, 225503 (2002). http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Condensed Matter arXiv (Cornell University)

Freezing point depression and freeze-thaw damage by nano-fuidic salt trapping

Loading next page...
 
/lp/arxiv-cornell-university/freezing-point-depression-and-freeze-thaw-damage-by-nano-fuidic-salt-X4qV4eHDPu
ISSN
2469-990X
eISSN
ARCH-3331
DOI
10.1103/PhysRevFluids.5.124201
Publisher site
See Article on Publisher Site

Abstract

Freezing point depression and freeze-thaw damage by nano- uidic salt trapping Tingtao Zhou Massachusetts Institute of Technology, Department of Physics Mohammad Mirzadeh Massachusetts Institute of Technology, Department of Chemical Engineering Roland J.-M. Pellenq The MIT / CNRS / Aix-Marseille University Joint Laboratory, "Multi-Scale Materials Science for Energy and Environment" and Massachusetts Institute of Technology, Department of Civil and Environmental Engineering Martin Z. Bazant Massachusetts Institute of Technology, Department of Chemical Engineering and Massachusetts Institute of Technology, Department of Mathematics A remarkable variety of organisms and wet materials are able to endure temperatures far below the freezing point of bulk water. Cryo-tolerance in biology is usually attributed to \anti-freeze" proteins, and yet massive supercooling (< 40 C) is also possible in porous media containing only simple aqueous electrolytes. For concrete pavements, the common wisdom is that freeze-thaw (FT) damage results from the expansion of water upon freezing, but this cannot explain the large pressures (> 10 MPa) required to damage concrete, the observed correlation between pavement damage and de-icing salts, or the FT damage of cement paste loaded with benzene (which contracts upon freezing). In this work, we propose a di erent mechanism { nano uidic salt trapping { which can explain the observations, using simple mathematical models of dissolved ions con ned between growing ice and charged pore surfaces. When the transport timescale for ions through charged pore space is prolonged, ice formation in con ned pores causes enormous disjoining pressures via the ions rejected from the ice core, until their removal by precipitation or surface adsorption at a lower temperatures releases the pressure and allows complete freezing. The theory is able to predict the non-monotonic salt-concentration dependence of FT damage in concrete and provides some hint to better understand the origins of cryo-tolerance from a physical chemistry perspective. I. INTRODUCTION rectly caused by solid phase transformations, not only ice 18,19 22,23 formation , but also salt crystallization . Interest- ingly, there is strong correlation between FT damage and The durability of wet porous materials against freeze- the use of de-icing salts on concrete pavements , which thaw (FT) damage is critical in many areas of science and are often less durable than concrete structures without engineering. In biology, it is a matter of life and death. salt exposure in the same cold climates. Moreover, FT Living cells must somehow maintain a liquid state within damage only occurs when the water saturation level ex- 1{3 the cellular membrane during winter , while avoiding ceeds a critical value . Previous models of \frost heave" anoxia due to external ice encasement . Various anti- (see e.g. ref ) achieved some successes in explaining the freeze proteins have been identi ed in cryo-tolerant ani- deformation of saturated soils due to the dynamics of pre- mals, and cryo-protectant chemicals have been used for melted liquid and its coupling with the solid. However, 2,5{7 cryo-preservation and in vitro fertilization . In addi- the applicability of these theories to hardened cement is tion, the complex thermodynamics of supercooled water questionable, due to its much higher sti ness compared could play a role. Even in bulk water, deep supercool- 27,28 to capillary stresses . In summary, despite the soci- 8{10 ing can lead to multiple metastable disordered states . etal importance of FT damage in cement, a physics-based 11,12 Phase transitions under nano-con nement can lead theory has not yet been developed that can predict the to exotic new phases, as well as modi ed ice nucleation, enormous pressures required (> 10 MPa), as well as all 13{15 16,17 in both experiments and molecular simulations puzzling observations above. of water in nanopores. In engineering, the most familiar example of FT dam- In this article, we develop a predictive theory of age is the fracture of concrete pavements during the freezing-point depression and FT damage in charged winter , commonly attributed to the expansion of water porous media, based on a simple new mechanism 19,20 transforming to ice within the pores . However, this sketched in Figure 1: nano uidic salt trapping. It is contradicts the observation that FT damage occurs when well known in colloid science that, when two charged sur- cement is loaded with benzene , a normal liquid that faces are separated by a liquid electrolyte, the crowding shrinks upon freezing. Recent experiments have chal- of ions in solution results in large repulsive forces when- lenged the prevailing hypothesis that FT damage is di- ever the electric double layers overlap, at the scale of arXiv:1905.07036v2 [cond-mat.soft] 8 Dec 2019 2 Figure 1: Physical picture of nano uidic salt trapping. (a) In an open pore, where ions and water molecules can easily exchange with a nearby reservoir, no signi cant pressure or freezing point depression is predicted. (b) Nanoscale bottlenecks, especially in poorly connected porous networks, with charged surfaces can signi cantly hinder this exchange by size or charge exclusion of the co-ions, and counter-ions are forced to stay to maintain charge neutrality. (c) Once ice nucleates, even an initially open pore will eventually trap a nanoscale thin lm of supercooled, concentrated electrolyte near the charged surface, until surface ion condensation or solid salt precipitation occurs. (d) In biological cells, the charged cytoskeleton (indicated by bers) could enable such passive nano uidic salt trapping, while further active control of water and ion ux across the cell membrane is performed by ion channels and pumps. In all cases, nano uidic salt trapping can lead to dramatic supercooling and, once ice nucleates, severe damage to the solid matrix. the Debye screening length (1-100nm in water). This the pore surface charges and preserve overall electroneu- \disjoining pressure" is responsible for the stabilization trality. Ions in solution mediate surface forces , which of colloidal dispersions in aqueous electrolytes , surface play a crucial role in the mechanical properties of con- 30 32,33,42{45 forces in clays and other porous media , and the elec- crete and the function of biological systems. In trostatic properties of membranes . Disjoining pressure most cases, the ions are assumed to have negligible solu- has been successfully modeled by the Poisson-Boltzmann bility in the frozen solid, as is the case with pure ice. mean- eld theory for solutions of monovalent ions, and Freezing begins in the larger \macropores" (> 100 extensions are available to describe correlation e ects in- nm), where bulk water easily transforms to ice, slightly 32,33 volving multivalent ions . Here we treat the disjoin- below the thermodynamic melting point of the solution, ing pressure between ice core and the charge pore surface which may be depressed from that of the pure solvent with a mean- eld approximation. Although the physics by the dissolved salt and any anti-freeze solutes. This of electrolyte freezing under con nement has been consid- bulk ice can form by homogeneous nucleation, spinodal ered for nanoporous materials , we propose that nano- decomposition, or (most likely) by heterogeneous nucle- uidic salt trapping is the key mechanism for large su- ation on impurities. Regardless of its origin, the advanc- percooling and FT damage in cement and other charged ing ice rejects ions, causing the salt concentration to rise nano-porous materials. This physical picture is consis- in the nearby, increasingly con ned liquid electrolyte. tent with all the available experimental evidence for con- What happens next depends on the degree of super- crete. cooling, the surface charge and importantly the pore connectivity. As shown in Figure 1(a), even after par- tial freezing, an individual pore may remain open, allow- A. Physical Picture ing ions and water molecules to exchange freely with a reservoir of bulk solution via a percolating liquid path 46{50 to neighboring unfrozen pores or an external bath . Consider a heterogeneous porous material saturated In this scenario, the liquid electrolyte and any solid ice with liquid and subjected to continuously decreasing within the pore remain in quasi-equilibrium with the bulk temperatures. As in most organisms and construction reservoir at constant chemical potential. The connected materials, suppose that the pore surfaces and suspended 35{41 path to the reservoir may pass through liquid-saturated materials are hydrophilic and charged , e.g. by pores, or partially frozen pores with suciently thick liq- the dissociation of surface functional groups or the ad- uid lms to allow unhindered transport. sorption of charged species. The large capillary pores (>5 nm) are typically also lled with water, but can be As freezing proceeds, many ions and water molecules replaced with other uids such as benzene. However, the will inevitably be trapped out of global equilibrium, small \gel" pores (1-5 nm) are always lled with liquid although still in local quasi-equilibrium within each water due to the strong surface charge and hydrophilic- nanoscale pore. The simplest case is that of a pore ity even in benzene-loaded cement samples. Importantly, connected to external reservoir only via a bottleneck, the liquid must contain dissolved salts, possibly at low sketched in Fig. 1(b). Water molecules that are not concentration, as well as excess counter-ions to screen closely associated with ions can still go through the bot- 3 tleneck, with a possibly di erent viscosity. The bottle- thus allowing complete freezing of the pores. Ions may neck may block solvated ions (with their solvation shells) also be cleared by adsorption reactions on the pore sur- from passing by steric hindrance or charge exclusion. face, which regulate and neutralize the surface charge. Even if some solvated ions can di use through a given bottleneck, their electrokinetic transport rate may be II. THEORY too slow to allow many to escape prior to more com- 51{53 plete freezing . Such slow ion transport may be en- hanced by long, tortuous pathways through a series of As mentioned above, under the assumption of sepa- 54{57 bottlenecks and compounded by a large volume of rated timescales for ion transport and freezing, we ap- micropores, e ectively cut o from the macropores with proximate the dynamic problem as a quasi-equilibrium insucient time for salt release, in materials of low pore- problem: in the limit of free ions, ion and water trans- space accessivity . Even in relatively well connected port is much faster than freezing; in the other limit of porous structures, nano uidic salt trapping can also re- trapped ions, ion transport is much slower than freez- sult from bottlenecks created by the advancing ice, as ing. The solutions of both limits can be uni ed in the shown in Figure 1(c), where the larger open pore on same quasi-equilibrium mean- eld framework. Below we the right side freezes almost completely rst. Due to the present details of these solutions. surface hydrophilicity, a supercooled liquid lm often re- The mean- eld free energy for a liquid electrolyte and mains between the pore surface and the ice core prior its frozen solid inside a charged pore can be described by 11,16,58 to complete freezing , which is now the only path- F = F + F + F way for water and ions in the smaller pore shown on the tot liquid solid interface left side. As temperature decreases further, ice forma- s 2 = dV   krk s l tion starts in the left pore, but transport of solvated ions through the thin liquid lm is now slow, and the entropy h i l 2 (1) of these con ned ions builds up a pressure. In biological + dV g(fc g) +  krk cells, as shown in Figure 1(d), electrolytes are contained l within the cell walls, and nano uidic salt trapping is facil- + dS ( + q ) j j itated by the charged cytoskeleton and abundant charged j=s;l;sl macromolecules (including cryo-resistant proteins). In- ternal salt concentrations are also actively maintained where the integrations are performed over volumes of by ion channels and pumps in the cell membrane . solid (V ) and liquid (V ) with permittivities  and  , re- s l s l To quantitatively calculate the timescales of freezing spectively, and over surfaces of the solid-liquid interface and ion transport, one needs to solve a proper electroki- (S ), the liquid-pore interface (S ) and the solid-pore sl l netic model of the 3D charged pore structure, with in- interface (S ), with corresponding surface charge densi- formation of the tortuosity and connectivity in addition ties, q , q and q and interfacial tensions, , and ; sl l s sl l s to the pore sizes. Here we focus on the asymptotic be- is the bulk chemical potential di erence between s l havior of very long ion transport timescale v.s. freez- solid and liquid phases; r is the electric eld; g(fc g) ing timescale, which hereafter referred to as the limit the non-electric part of homogeneous liquid electrolyte of trapped ions. This approximation of timescale sep- free energy; c the concentration of ion species i having aration is similar in essence to the adiabatic or Born- charge z e; and  = z ec the net charge density, as- i i i Oppenheimer approximation , where the short time sumed to be negligible in the solid phase. We focus on quasi-equilibrium is solved|as we show in the next situations of complete wetting by the liquid,  , s l sl sections|neglecting the slowly changing physics, in our in which case we can neglect S and assume S covers the s l case the ion transport. The phenomenon of ion trap- entire pore surface. ping in charged nanochannels, while water remains free Setting F = = 0 for bulk and surface variations, tot to di use and ow to a nearby reservoir or larger we obtain Poisson's equation pore, is well established in the eld of nano uidics and forms the basis for various devices, such as electro-  r  =  in V l l 61 (2) osmotic micropumps , nano- uidic diodes and bipolar r  = 0 in V s s 54,62,63 64 transistors , and nano uidic ion separators . The supercooling of con ned liquids can be greatly en- and electrostatic boundary conditions hanced by the salt rejected by freezing, as the remain- ~ ~ q = ( E  E ) n ^ on S ing solution becomes more concentrated inside a trapped sl s s l l ls sl (3) freezing pore. Large disjoining pressures are then pro- q =  r n ^ on S l l l l duced in the very concentrated liquid solution and trans- mitted to the solid matrix, potentially causing damage. The equilibrium state of liquid-solid coexistence is At suciently low temperatures, salt-enhanced super- found by minimizing the total free energy with respect cooling and freeze-thaw pressure are relieved by the sud- to the position and shape of the solid-liquid interface, den precipitation of ions from the concentrated liquid, S . Here, we consider two cases: (1) an open pore sl 4 where ions of species i exchange freely with a reservoir by minimizing the total free energy with respect to r, of concentration c , or (2) a pore with trapped ions, F =r = 0: tot whose total number is xed by screening the pore sur- r = argmin F (r) (5) tot face charge in the liquid, prior to freezing, by the mech- anisms shown in Fig. 1. Importantly, we neglect the ef- which yields the equilibrium ice volume fraction,  = fects of volume changes due to the water/ice transforma- tion, under the assumption that liquid water molecules (r =R) . Once r is found, mechanical equilibrium at the solid-liquid interface gives pressure of both phases, (of size  3A) are mobile and small enough to escape the pore as freezing progresses, regardless of whether sol- which is transmitted to the pore boundary vated ions are trapped. In contrast to the common wis- @F @F solid liquid dom about freeze-thaw damage in pavements, this pic- P = = (6) @r @r ture must also hold for well-connected hierarchical porous r=r r=r materials such as concrete. The rst equality describes the tendency to form more The preceding thermodynamic framework for con ned ice and hence expand its volume, while the second equal- electrolyte phase transformations can be extended in var- ity shows the free energy cost to squeeze the electrolyte, ious ways, e.g. to account for ion-ion correlations (es- resisting the growth of ice. pecially involving multivalent ions), nite ion sizes and For a symmetric pore, after freezing starts, the free 67,68 hydration surface forces , but here we focus on the energy of ice is given by simplest Poisson-Boltzmann mean- eld theory , which suces to predict the basic physics of freezing-point de- F = (  ) V (d)r ; (7) ice s l pression and material damage. The homogeneous free en- ergy is then given by the ideal gas entropy for point-like where (  ) is the Gibbs free energy change per vol- s l ume for bulk water freezing, which can be calculated ions, g = c [ln(v c ) 1], with v the molecular volume, i i i i i and the electrostatic potential in the liquid electrolyte is using the Gibbs-Helmholtz relation, as shown in ref. . In principle, the electric eld energy of the ice core (x < r) then given by the Poisson-Boltzmann (PB) equation: depends on its shape and the electrostatic boundary con- 2 1 z e ditions, but vanishes here by symmetry. The interfacial r  =  = z ec ; c = c e (4) l l i i i d1 energy is F = S(d)r , which gives rise to the i surface sl Gibbs-Thomson e ect of freezing point depression for Since we focus on highly con ned electrolyte liquid lms, con ned pure water. The free energy of the electrolyte we set the relative permittivity,  = 10 , to that of wa- l 0 shell is given by 69,70 ter near dielectric saturation at high charge density . elec To assess the prevalence of nano uidic salt trapping d1 = q (R)R + within Poisson-Boltzmann theory, the state of a bottle- S(d) (8) R h i neck shown in Fig. 1 can be estimated by comparing l 2 d1 x dx g(fc g) +  krk the double layer thickness  (or hydrated ion size a) inside with its radius R: if   R (or a & R) then the double layer(s) span across and the bottleneck is ap- where the rst term is the electrostatic energy of surface proximated as \closed" to ions, since freezing rate may charges, and the integrand takes the form given above for exceed ion transport rate, given a high tortuousity of the mean- eld theory of point-like ions. To summarize, we pore network. If   R (or a . R), then the channel D are solving a free boundary problem where the liquid-ice may be viewed as open to ion exchange. For an initial boundary position r is unknown beforehand. We adopt salt concentration of 0:1 mol=L in a binary monovalent a numerical algorithm to search for the r that minimizes electrolyte, (with relative permittivity   10), we nd r total free energy at a given temperature T , surface charge 4 k T l B density q and initial salt concentration c : 0:5 nm. D 2 2c e 1. starting from r = 0, compute the total free energy F (0). A. Symmetric pores 2. increment r by a small amount dr, compute the to- tal free energy F (r). In order to obtain analytical results, we consider when computing the total free energy at a given isotropic electrolyte freezing in d dimensions, where ice r value, we always solve the Poisson-Boltzmann nucleates to form a plate (d = 1), cylinder (d = 2) or Eqn. 4 to obtain the electric potential pro le , sphere (d = 3) of radius r within a pore of the same and insert into the integration of Eqn. 8. symmetry, whose surface is located at x = R. The total d d1 pore volume is V (d)r , and S(d)r the surface area of 3. after sweep r from 0 to the pore size R, nd the the ice core (x < r), surrounded by a liquid electrolyte minimum of F and the corresponding r gives the shell (r < x < R). At thermodynamic equilibrium, position of the quasi-equilibrium ice front. the location r of the solid-liquid interface is determined 5 Figure 2: Electrolyte freezing and pressure generation in a parallel slit pore (d = 1) with free ions exchanging with a reservoir. There is no e ects of interfacial tension. The freezing point depression, T  0:1 K, Figure 3: In contrast to Fig.2, for a binary electrolyte and disjoining pressure, P  0:1 MPa, are quite small, with trapped ions, freezing-point depression as large as in the limit of one-component plasma of only -40 K can occur. The quasi-equilibrium approximation counter-ions. In this case the total number of gives a continuous freezing temperature range marked counter-ions is determined by the surface charge density by two T values: the temperature to start freezing, T , only and does not depend on pore size. And the T f and that of complete freezing of the pore T , when ff and P only depends on the distance between ice front ions are removed by precipitation. and the pore surface, which denoted by L = 5 nm here. B. Free ions As freezing proceeds in an open pore, where all ions can escape to a reservoir, the surface charge is eventually screened in a thin liquid lm containing only counter-ions, which corresponds to one component 73,74 plasma (OCP) . The Poisson-Boltzmann equation for the OCP can be integrated for symmetric pore shapes to obtain the mean electrostatic potential. For a slit pore (d = 1), we obtain to the rst order 4 k T Z T l B 0 jPj  Q 2 2 q e T s 0s 1 Figure 4: Large disjoining pressures up to  10 MPa (9) ~ ~ jPj jPj Rqe occur during the freezing process, below the @ A tan = 2 2 2 4 k T temperature to start freezing, T , and above that of l B f complete freezing of the pore T . The range of ff where Q is the latent heat of bulk water freezing, T pressure is marked by P and P , correspondingly. f ff Blue and red dots are numerical results, while the solid the bulk freezing point, and T = T T the freezing point depression. Notice that for OCP limit, the total lines connecting them are guiding the eyes. amount of counter-ions does not depend on the pore size, but is simply determined by the surface charge density. Hence, the quasi-equilibrium solution only depends on the distance between the ice front and the pore surface, which we here denote as L. spherical (d = 3) pores are even smaller than in a slit Inserting typical values, we can estimate the freezing pore (d = 1) and may often be neglected compared to point depression in the slit pore as T  0:1 K and the Gibbs-Thompson e ect of interfacial tension in such the pressure as P  0:1MPa. In this case, the freez- curved geometries. In general, if excess salt ions (and wa- elec ing point is only depressed by . 1 K, and no signi cant ter molecules) are free to escape the pore during freezing, pressure is generated, as shown in Fig.2. As shown in then we expect very little freeze-thaw damage in a wet ref. , the e ects of ions in open cylindrical (d = 2) or porous material. 6 C. Trapped ions The situation is completely di erent in the opposite limit, where all ions in the original liquid binary elec- trolyte remain trapped within the pore during freezing. Total ion number conservation is then imposed on the d1 PB equations, c S(d)x dx = N , and signi cant i i freezing-point depression can be achieved. The math- ematical details can be found in a companion paper , and here we focus on explaining the physical predictions of the theory. To separate the e ect of curvature, here we focus on the slit symmetry (d = 1). First we consider a binary 1:1 liquid electrolyte freezing in a parallel slit pore (d = 1). In this case, there is no ef- fect of solid-liquid interfacial tension, as the interface area does not change as ice front advances (zero curvature). As shown in Fig.3, the freezing point is substantially de- creased by increasing the initial salt concentration c in the con ned liquid. After freezing starts at temperature T , due to the resistance of the electrolyte, the equilib- rium ice volume fraction  monotonically increases as temperature decreases. The freezing process continues until the trapped ions are suddenly removed from the thin liquid lm at the temperature of freezing nished T , when the salt solubility limit is reached, and  sud- ff denly jumps to 1. The pore is completely frozen now. Complete freezing may also occur if the trapped ions are adsorbed on the pore surface, thereby neutralizing the Figure 5: (a) typcial experiment protocol reproduced surface charge (as shown below). from . T and T correspond to the temperature As shown in Figure 4, signi cant disjoining pressures f ff when ice formation initiates and solubility limit ( 10 MPa for R = 5 nm) can be generated by con- reached, as indicated in Fig.3. (b) shaded area shows ned ions during the freezing process. The pressures the pressure range after freezing starts in a trapped at the freezing start temperature T and the complete pore. P and P correspond to the pressure when ice freezing temperature T are labeled as P and P , re- f ff ff f ff formation initiates and solubility limit reached, as spectively. The disjoining pressure varies approximately indicated in Fig.4. Dash line shows open pores with free linearly with temperature between these values during ions, which is close to the horizontal line of 0. Data the freezing process in a slit pore. points show measured damage in cement paste FT experiment , a non-monotonic function of salt concentration. Tensile strength of hardened cement D. Salt solubility limit and surface charge paste is  3 MPa. regulation As ice volume fraction increases, salt concentration goes up. At some point the concentrated electrolyte will 22,23 become saturated and salt will crystallize. The volume of opposed to the concept of \crystallization pressure" salt crystal precipitate is neglected. The solubility equi- that has been proposed to account for pressure and dam- librium for 1:1 electrolyte (M + B ) at saturation is age (under room temperature) in construction materials, + sat [M ][B ] c here the pressure of the freezing pore is determined by K = = . Here c is the concen- eq solid [MB] c solid thermodynamc equilibrium between freezing and precip- tration in solid crystal phase, which is typically regarded itation, thus salt crystallization is merely a consequence, sat as constant 1. Once c , the saturated concentration of instead of the cause for pressure. salt ions, is reached the equilibrium position of ice front becomes thermodynamically unstable and all the liquid When the concentration of trapped ions is high enough, turned into solid phases of ice and salt crystal. In Fig.3, counter-ion recombination with the surface charge be- all the curves at some point reach the solubility limit and comes important. This e ect can be included by a modi- undergo sudden crystallization, when ice volume fraction ed boundary condition for the PB equations, where the discontinuously jumps from  < 1 to  = 1. The pressure surface charge is computed self-consistently based on a 75 71 at this point is denoted as P in both Fig.4 and Fig.5. As charge regulation model (see more details in ). ff 7 III. APPLICATION TO CONCRETE and complete freezing. Freezing point depression, ice vol- ume fraction and pressure are calculated using a simple mean- eld theory. The predictions of this theory are semi-quantitatively consistent with experimental observations of freeze-thaw Many extensions of the theory could be considered damage in cement. Below critical degree of water satu- in future work. Ion correlations, including the strong- 76{78 ration, plenty of large pores remain open transport path- coupling limit , can be introduced via higher order ways for ions during freezing, hence no signi cant dam- terms in Eqn.8, resulting in modi ed PB equations . age observed . The volume expansion of water during At larger length scales, models of interfacial instabilities 79{81 freezing is irrelevant in this theory, so it can also explain leading to dendritic growth could be extended to ac- 82,83 qualitatively similar results observed in freeze-thaw ex- count for electrokinetic phenomena in charged pores . perimetns on cement samples loaded with benzene, which Here we always assume bulk phase of ice (the Ih phase) shrinks upon freezing . The non-monotonic dependence is formed, since the freezing conditions discussed here of damage on NaCl concentration can be explained by (T > 200 K , P < 100 MPa, d < 100 nm) are not very crossover from salt trapping to channel opening though extreme. Exotic phases of ice (non-Ih phases) are known 9,84{89 charge regulation, as shown in Fig.5. A fully quantitative to dominate under more extreme conditions . Salt comparison requires the plasticity and fracture mechan- ions can also a ect the surface tension of ice-electrolyte ics of the solid matrix due to these local high pressures, interface, as well as other aspects of nucleation under and the connectivity of the pores, which is currently a con nement, described in a companion paper . missing link. Also to quantify the transport timescales As a rst application to material durability, our theory for ions as freezing proceeds, pore connectivity is key in- is consistent with complex trends of freeze-thaw damage formation. Nevertheless, to our knowledge for the rst in hardened cement paste. These predictions could in u- time this mechanism shows potential to encompass all ence industrial practices in road de-icing and pavement these observations. design. The theory may also provide some perspective on the physics of cryo-tolerance and cryo-preservation in biological materials, which abound in electrolyte-soaked IV. CONCLUSION AND DISCUSSIONS macromolecules, nanopores and membranes. V. ACKNOWLEDGEMENT In this article, we present a theory of the freezing of electrolytes in charged porous media. The key insight is that, if ions become trapped by the advancing ice This work is carried out with the support of the Con- front, large disjoining pressures can cause material dam- crete Sustainability Hub at MIT. The authors thank S. age, until further supercooling triggers salt precipitation Yip, C. Qiao, J. Weiss, M. Pinson for useful discussions. 1 13 C. L. Guy, Annual review of plant biology 41, 187 (1990). K. Morishige and H. Iwasaki, Langmuir 19, 2808 (2003). 2 14 J. Saragusty and A. Arav, Reproduction 141, 1 (2011). M. Nosonovsky and B. Bhushan, Physical Chemistry J. M. Storey and K. B. Storey, in Life in the frozen state Chemical Physics 10, 2137 (2008). (CRC press, 2004) pp. 269{300. K. V. Agrawal, S. Shimizu, L. W. Drahushuk, D. Kilcoyne, C. Andrews, Annals of Botany 78, 529 (1996). and M. S. Strano, Nature nanotechnology 12, 267 (2017). 5 16 F. F. AbdelHafez, N. Desai, A. M. Abou-Setta, T. Falcone, P. A. Bonnaud, B. Coasne, and R. J. Pellenq, Journal of and J. Goldfarb, Reproductive biomedicine online 20, 209 Physics: Condensed Matter 22, 284110 (2010). (2010). S. Cerveny, F. Mallamace, J. Swenson, M. Vogel, and R. Wakchaure, S. Ganguly, S. Sharma, P. K. Praveen, L. Xu, Chemical reviews 116, 7608 (2016). M. Sharma, and T. Mahajan, Int. J. Phar. & Biomedi. Y. Farnam, D. Bentz, A. Sakulich, D. Flynn, and J. Weiss, Rese 2, 11 (2015). Advances in Civil Engineering Materials 3, 316 (2014). 7 19 L. Rienzi, C. Gracia, R. Maggiulli, A. R. LaBarbera, D. J. H. Cai and X. Liu, Cement and Concrete Research 28, Kaser, F. M. Ubaldi, S. Vanderpoel, and C. Racowsky, 1281 (1998). Human reproduction update 23, 139 (2017). P. C. Association, \Freeze-thaw resistance," (2018). 8 21 O. Mishima and H. E. Stanley, Nature 396, 329 (1998). J. J. Beaudoin and C. MacInnis, Cement and Concrete C. Tulk, C. Benmore, J. Urquidi, D. Klug, J. Neuefeind, Research 4, 139 (1974). B. Tomberli, and P. Egelsta , Science 297, 1320 (2002). E. M. Winkler and P. C. Singer, Geological society of P. G. Debenedetti, Journal of Physics: Condensed Matter America bulletin 83, 3509 (1972). 15, R1669 (2003). M. Steiger, Journal of crystal growth 282, 455 (2005). 11 24 L. D. Gelb, K. Gubbins, R. Radhakrishnan, and Y. Farnam, D. Bentz, A. Hampton, and W. Weiss, Trans- M. Sliwinska-Bartkowiak, Reports on Progress in Physics portation Research Record: Journal of the Transportation 62, 1573 (1999). Research Board , 81 (2014). 12 25 D. T. Limmer and D. Chandler, The Journal of chemical W. Li, M. Pour-Ghaz, J. Castro, and J. Weiss, Journal of physics 137, 044509 (2012). Materials in Civil Engineering 24, 299 (2011). 8 26 56 J. Wettlaufer and M. G. Worster, Annu. Rev. Fluid Mech. H.-C. Chang, G. Yossifon, and E. A. Demekhin, Annual 38, 427 (2006). review of uid mechanics 44, 401 (2012). 27 57 T. Zhou, K. Ioannidou, E. Masoero, M. Mirzadeh, R. J.-M. G. Yossifon, P. Mushenheim, Y.-C. Chang, and H.-C. Pellenq, and M. Z. Bazant, Langmuir 35, 4397 (2019). Chang, Physical Review E 79, 046305 (2009). 28 58 T. Zhou, K. Ioannidou, F.-J. Ulm, M. Z. Bazant, and R. Denoyel and R. Pellenq, Langmuir 18, 2710 (2002). R.-M. Pellenq, Proceedings of the National Academy of H. Bertil, Ionic channels of excitable membranes (Sinauer Sciences 116, 10652 (2019). Sunderland, MA, 1992). 29 60 J. Lyklema, Fundamentals of Interface and Colloid Sci- M. Born and R. Oppenheimer, Annalen der physik 389, ence. Volume II: Solid-Liquid Interfaces (Academic Press 457 (1927). Limited, San Diego, CA, 1995). S. Zeng, C.-H. Chen, J. C. Mikkelsen Jr, and J. G. Santi- J. N. Israelachvili, Intermolecular and Surface Forces (Aca- ago, Sensors and Actuators B: Chemical 79, 107 (2001). demic Press, New York, NY, 2011). H. Daiguji, Y. Oka, and K. Shirono, Nano Letters 5, 2274 D. Andelman, in Handbook of Biological Physics (Elsevier, (2005). 1995) pp. 603{641. G. Yossifon, Y.-C. Chang, and H.-C. Chang, Physical re- R. J.-M. Pellenq and H. Van Damme, Mrs Bulletin 29, 319 view letters 103, 154502 (2009). (2004). D. Gillespie and S. Pennathur, Analytical chemistry 85, R.-M. Pellenq, J. Caillol, and A. Delville, The Journal of 2991 (2013). Physical Chemistry B 101, 8584 (1997). M. Z. Bazant, B. D. Storey, and A. A. Kornyshev, Physical C. Alba-Simionesco, B. Coasne, G. Dosseh, G. Dudziak, Review Letters 106, 046102 (2011). K. Gubbins, R. Radhakrishnan, and M. Sliwinska- M. Z. Bazant, M. S. Kilic, B. Storey, and A. Ajdari, Ad- Bartkowiak, Journal of Physics: Condensed Matter 18, vances in Colloid and Interface Science 152, 48 (2009). R15 (2006). K. Bohinc, A. Shrestha, M. Brumen, and S. May, Physical K.-M. Jan and S. Chien, The Journal of general physiology Review E 85, 031130 (2012). 61, 638 (1973). M. A. Brown, G. V. Bossa, and S. May, Langmuir 31, J. Visser, (1973). 11477 (2015). 37 69 R. Giese, Clays and Clay Minerals 21, 145 (1973). F. Booth, The Journal of Chemical Physics 19, 391 (1951). 38 70 S. McLaughlin, in Current topics in membranes and trans- M. Aguilella-Arzo, A. Andrio, V. M. Aguilella, and port, Vol. 9 (Elsevier, 1977) pp. 71{144. A. Alcaraz, Physical Chemistry Chemical Physics 11, 358 S. L. Swartzen-Allen and E. Matijevic, Chemical Reviews (2009). 74, 385 (1974). T. Zhou, M. Mirzadeh, R. J.-M. P. Dimitrios Fraggedakis, A. J. Allen, J. J. Thomas, and H. M. Jennings, Nature and M. Z. Bazant, \Theory of freezing point depression in materials 6, 311 (2007). charged porous media," In preparation. 41 72 R. J.-M. Pellenq, A. Kushima, R. Shahsavari, K. J. J. W. Gibbs, (1879). Van Vliet, M. J. Buehler, S. Yip, and F.-J. Ulm, Pro- S. Brush, H. Sahlin, and E. Teller, The Journal of Chem- ceedings of the National Academy of Sciences 106, 16102 ical Physics 45, 2102 (1966). (2009). M. Baus and J.-P. Hansen, Physics Reports 59, 1 (1980). 42 75 A. Delville, N. Gasmi, R. J. Pellenq, J. M. Caillol, and T. Markovich, D. Andelman, and R. Podgornik, EPL (Eu- H. Van Damme, Langmuir 14, 5077 (1998). rophysics Letters) 113, 26004 (2016). 43 76 R. Jellander, S. Mar celja, and J. Quirk, Journal of Colloid A. G. Moreira and R. R. Netz, EPL (Europhysics Letters) and Interface Science 126, 194 (1988). 52, 705 (2000). 44 77 J. Stankovich and S. L. Carnie, Langmuir 12, 1453 (1996). R. R. Netz, The European Physical Journal E 5, 557 J. P. Valleau and L. K. Cohen, The Journal of chemical (2001). physics 72, 5935 (1980). L. Samaj and E. Trizac, Physical Review E 84, 041401 S. Torquato, Random heterogeneous materials: microstruc- (2011). ture and macroscopic properties, Vol. 16 (Springer Science M. B. Hastings and L. S. Levitov, arXiv preprint cond- & Business Media, 2013). mat/9607021 (1996). 47 80 J. Van Brakel, Powder technology 11, 205 (1975). W. W. Mullins and R. F. Sekerka, Journal of applied J. A. Quiblier, Journal of Colloid and Interface Science 98, physics 34, 323 (1963). 84 (1984). W. W. Mullins and R. Sekerka, in Dynamics of Curved M. B. Pinson, T. Zhou, H. M. Jennings, and M. Z. Bazant, Fronts (Elsevier, 1988) pp. 345{352. Journal of colloid and interface science (2018). M. Z. Bazant, J. Choi, and B. Davidovitch, Physical re- Z. Gu and M. Z. Bazant, arXiv preprint arXiv:1808.09804 view letters 91, 045503 (2003). (2018). M. Mirzadeh and M. Z. Bazant, Physical review letters P. Biesheuvel and M. Bazant, Physical Review E 94, 119, 174501 (2017). 050601 (2016). L. G. Dowell and A. P. Rinfret, Nature 188, 1144 (1960). 52 85 P. Peters, R. Van Roij, M. Z. Bazant, and P. Biesheuvel, D. Steytler, J. Dore, and C. Wright, The Journal of Phys- Physical review E 93, 053108 (2016). ical Chemistry 87, 2458 (1983). 53 86 J. Catalano, R. Lammertink, and P. Biesheuvel, arXiv E. Mayer and A. Hallbrucker, Nature 325, 601 (1987). preprint arXiv:1603.09293 (2016). B. J. Murray, D. A. Knopf, and A. K. Bertram, Nature G. Yossifon, I. Frankel, and T. Miloh, Physics of Fluids 434, 202 (2005). 18, 117108 (2006). E. B. Moore and V. Molinero, Physical Chemistry Chem- S. Pennathur and J. G. Santiago, Analytical chemistry 77, ical Physics 13, 20008 (2011). 6772 (2005). J. Finney, A. Hallbrucker, I. Kohl, A. Soper, and D. Bowron, Physical review letters 88, 225503 (2002).

Journal

Condensed MatterarXiv (Cornell University)

Published: May 16, 2019

There are no references for this article.