Water Dynamics in Nanoporous Alteration Layer Coming from Glass

Article pubs.acs.org/JPCC

Water Dynamics in Nanoporous Alteration Layer Coming from Glass Alteration: An Experimental Approach D. Rébiscoul,*,† J. Cambedouzou,†,‡ I. Matar Briman,‡ M. Cabié,§ H.-P. Brau,† and O. Diat† †

CEA, ICSM − UMR 5257 CEA-CNRS-UM-ENSCM, 30207 Bagnols-sur-Cèze Cedex, France CEA, DEN, DTCD/SECM/LCLT - Marcoule, F-30207 Bagnols-sur-Cèze, France § Aix-Marseille Université, CP2M, F-13397 Marseille, France ‡

S Supporting Information *

ABSTRACT: Protons dynamics in borosilicate glasses of various compositions (SiO2/B2O3/NaO/CaO/X, with X = Al2O3 or ZrO2) at various stages of alteration have been characterized at a picosecond scale using quasielastic neutron scattering (QENS). The results obtained were compared to the one from porous silica-based material (MCM41) having pore wall surfaces with Si−OH, Al−OH, or Zr− OH terminal groups and pore sizes around 2.3 nm. The composition and the nanoporosity of the alteration layers were characterized using solution analysis, small angle X-ray scattering and transmission electron microscopy. The strength hydrogen bonds of confined water were studied using thermogravimetric analysis and differential thermal analysis. The results showed that the free water bonding and the mobility of protons depend on the altered glass composition and that the residence time of water obtained from QENS analysis is apparently correlated to the alteration rate of the glass. Moreover, whereas the composition of the alteration layer can partially explain this result, the presence of ions in the leachate filling the gel nanoporosity seems to be the main factor affecting the water/protons mobility. The latter result is really important since the ion solvation and the ion adsorption on the surface can strongly impact the hydrolysis rate of the pore wall of the gel and can also modify the kinetics of dissolved elements recondensation into the nanopore.

■

INTRODUCTION

water chemical reactivity. This is what we propose to investigate in this paper. The long-term evolution assessment of materials is of great interest, particularly for the nuclear borosilicate glass in geological repository for which the long-term behavior regarding its alteration by water has to be determined typically for a period of 1 My.12 When a borosilicate glass is in contact with water, its alteration leads to the formation of a nanoporous silica layer named “gel”. This layer is hydrated, and its porosity is filled of leachate (water and ionic species coming from glass alteration). The gel is formed by hydrolysis and condensation of the species coming from the glass and dissolved in solution such as Si and Ca around the insoluble network former as Zr and Al. Moreover, the gel limits the water transport toward the glass and the release of the mobile glass components.13−17 The transport through the gel has been already characterized at large time and space scales.16−19 Meanwhile, the underlying mechanisms at the pore interfaces, i.e. at a shorter scale, such as the water motions and water reactivity, are still not understood. In this paper, we have characterized the water

Water properties in porous silica-based materials such as ordered mesoporous silica, minerals, cement, and glasses have been studied for several years in the field of membrane science, catalysis, biology, nanofluidics, building, and earth sciences.1−3 The numerous investigations related to this subject also highlight the interest to study water reactivity in order to explain the long term evolution of minerals and silicates glasses. Indeed, rate laws developed and used to explain or to predict mineral and glass dissolution kinetics generally take into account some thermodynamic and kinetic data arising from measurements in diluted media (concentration of dissolved ions, pH, solubility, adsorption constants, etc.). However, nanoporous minerals such as diatomaceous,4 clay materials,5 secondary minerals formed from rocks6 and amorphous nanoporous alteration layers of silicate glasses consist in sets of confined media,7 where the validity of thermodynamic and kinetic data are not proved. Indeed, strong interactions of water molecules with pore surfaces8,9 and ions slowing down water motions are expected.10,11 Then, the availability of water to react with pore walls and dissolved species present into the porosity can be limited. Thus, probing confined water dynamics in such nanoporous materials can provide new insight into © 2015 American Chemical Society

Received: March 31, 2015 Revised: June 25, 2015 Published: June 26, 2015 15982

DOI: 10.1021/acs.jpcc.5b03073 J. Phys. Chem. C 2015, 119, 15982−15993

Article

The Journal of Physical Chemistry C dynamics at the picosecond scale in gels in order to clarify the water behavior in the porosity. Gels coming from glass alteration are complex materials. They can be inhomogeneous in term of porosity and composition, and the leachate composition into the pores is not attainable. Thus, the dynamics of water confined in gels have been studied in two steps. In the first step, we have used model materials, trying to determine the predominant factor impacting the water dynamics: the composition of the pore surface and/or the pore size and/or the presence of ions into the porosity.21,22 Many studies have shown that the decrease of the confinement size,22 the chemical nature of the interfaces23,24 as well as the presence of ions25−28 in solution lead to a slowdown of water motions as observed in Vycor porous glass,29 ordered mesoporous silica such as MCM-41 of various pores size,22,30,31 and mesoporous alumina.24 However, none of these studies has been performed on the same materials. To reach this goal, we have characterized the water dynamics at a time scale between 0.1 and 10 ps by quasi-elastic neutron scattering (QENS) in model materials such as silica-based materials (MCM-41) having various pores sizes (from 2 to 2.7 nm) and presenting Si−OH, Al−OH, and Zr−OH groups at the pore surface tuned by a grafting process.20 We have shown that one part of water molecules has the same mobility as bulk water, and the other part is fixed at the pore surface. These results were also confirmed by molecular dynamics simulations in silica nanopores by Bourg et al.32 Moreover, the fixed water molecules at the pore interface were surface sensitive. Pores having Al−OH and Zr−OH surfaces lead to the highest water immobilization, creating coordination bonds with water molecules. For pores smaller than 2.3 nm, the pore surface composition was the predominant parameter affecting the fixed proton content. A similar study performed with model materials filled of leachate have shown that the presence of ions does not modify the water dynamics in the pores. However, the leachate used to fill the model materials came from glass alteration, which is probably different from the more concentrated leachate existing in the gel porosity. The purpose of this paper, which consists of the second step of our approach, is to determine, for the first time, the water dynamics at a picosecond scale by QENS in the gels formed during the alteration of borosilicate glasses having various compositions (SiO2/B2O3/NaO/CaO/X, with X = Al2O3 or ZrO2). The composition and the porosity of the alteration layers were characterized using solution analysis, small angle Xray scattering (SAXS), and transmission electron microscopy (TEM). The hydrogen bonds of confined water were studied using thermogravimetric analysis (TGA) and differential thermal analysis (DTA). We compared the water behavior in the gel coming from glass alteration with the one in model materials and we intended to relate the water mobility in the gel with the kinetics of glass alteration.

Table 1. References, Properties of the Structure and the Porosity of Model Materials

reference

precursors

MCMSi-2 MCM-Si-3 MCMAl-3 MCMZr-3

Si(OCH3)4

a

number of grafting cycle

da (nm)

pore diameter (nm)

mean temperature of free water removal (°C) (20)

2 3 3 3

3.7 3.7 3.7 3.7

2.3 2.0 2−2.4 2.1−2.5

106 120 117 116

Al(OC2H5)3 Zr(OC2H5)4

d: the inter-reticular distance.

Suitable amounts of analyticalgrade SiO2 (Millisil), H3BO3, Na2CO3, Al2O3, and ZrO2 and CaO (Prolabo) were mixed (Supporting Information Table 1). The powder mixture was decarbonated for 1 h at 800 °C and melted between 1300 and 1400 °C for 3 h in a platinum−gold crucible. The glasses were quenched in air and crushed before refining (second melt) for 2 h at 1300 or 1400 °C to ensure their homogeneity. The glass references, the molar compositions, and the mass densities are presented in Table 2. Glass specimens were ground, sieved to Table 2. References, Molar Compositions, and Mass Density ρm of Soda-Lime Borosilicate Glasses G Reference

SiO2

B2O3

Na2O

CaO

Zr G Zr/Ca G (21) Al/Ca G (21) Al/Zr G (21)

63.9 56.5 61.2 63.8

15.3 17.5 16.3 17.0

16.9 18.3 12.9 13.4

3.8 5.8 -

ZrO2 Al2O3 3.8 3.9 1.8

3.9 4.1

ρm (g.cm−3) 2.532 2.569 2.471 2.427

recover the 20−40 μm powder size fractions and washed in acetone in order to remove the fine particles. The specific surface area of the glass powders were measured by krypton adsorption−desorption using the Brunauer−Emmett−Teller (BET) method. The specific surface area obtained for the Zr, Zr/Ca, Al/Ca, and Al/Zr glasses were respectively 0.1765, 0.1655, 0.1665, and 0.1800 ± 0.0200 m2·g−1. The four glasses were altered at a glass surface area to solution volume ratio (S/V) of 80 cm−1 in ultrapure water in a polytetrafluoroethylene (PTFE) reactor for 7, 22, and 58 days at 90 °C in an oven. After each alteration, the altered glasses were removed from the reactor, filtered, and dried at 60 °C for 15 h in order to remove only the water sorbed on the grain surface and avoiding removal of water from pores. One part of the altered glass powders was freeze-dried for 48 h to avoid a porosity modification35,36 and analyzed by SAXS and TEM. The other part of the altered glass powders was placed in a desiccator in the presence of a beaker containing a saturated solution of KCl, imposing a relative humidity (RH) of 86% at 25 °C, allowing the complete filling of porosity and avoiding the water removal from porosity for several weeks. Then samples were characterized by TGA/DTA and QENS. Materials Characterization. Solution Analysis. At the end of the experiments, pH were measured at 90 °C at the laboratory atmosphere, and leachates were ultrafiltered to 10 000 Da. Then, 3 mL of filtered leachate were acidified with 50 μL of HNO3 15 N (HNO3 65 w% Suprapur Merck) and analyzed by ICP-AES to determine the atomic concentrations of Si, B, Na, Ca, Al, and Zr from acidified solutions. The analytical uncertainty was ±3%.

■

EXPERIMENTAL METHODS Materials. Model Materials. Model materials, i.e., porous silica-based materials MCM41,33 were synthesized by a hydrothermal route and grafted via a hydrolytic surface sol− gel method in order to tune the pore size and have pore wall surfaces with Si−OH, Al−OH, or Zr−OH terminations. Material elaboration and properties were already published in.20 Their mean properties are summarized in Table 1. Altered glasses. Four borosilicate glasses34 were elaborated by adding oxides to the SiO2−B2O3−Na2O ternary system. 15983


Article

The Journal of Physical Chemistry C The normalized mass losses NL(i) were calculated from the concentrations of the element i [i] (g·m−3) (eq 1). NL(i) =

[i] xi·(S /V )

(1)

with xi being the mass fraction of the element i in solution and S being the glass surface area (m2) in a volume of solution V (m3). Due to its solubility in these experimental conditions, boron is the tracer of glass alteration. Thus, the altered glass thicknesses AGth(B) were determined from the normalized mass losses of B (eq 2). AGth(B) =

NL(B) ρv

(2)

with ρv being the mass density of the glass indicated in Table 2. The retention factors RF(i) of the element i were calculated from eq 3: RF(i) = 1 −

NL(i) NL(B)

(3)

Materials Characterization. The morphology of the materials was characterized using SAXS in transmission geometry with a Xenocs setup equipped with a Mo anode (λ = 0.71 Å) over a Q range from 0.2 nm−1 to 10 nm−1 using a MAR345 2D imaging plate detector. Samples were analyzed in glass capillaries of 2 mm diameter. TEM observations of ultramicrotomic thin sections (preparation of thin sections is described in ref 37) were carried out on a Tecnai ̈ G2 (FEI), equipped with a LaB6 source operating at 200 kV and in scanning mode. The detectors were a Gatan CCD camera and a STEM BF-DF detector. The dehydration of altered glass was carried out using a SETARAM TGA 92−16.18 thermogravimetric apparatus. TGA and DTA data were collected under an air stream (16 cm3· min−1) running from room temperature up to 700 °C with an heating rate of 10 °C·min−1. Buoyancy effects were corrected by a blank run with an empty alumina crucible. Water Dynamics Characterization. Before analysis, samples stored in the desiccator were placed into aluminum cells, and then cells were sealed with an indium wire to avoid water loss and ensure a constant hydration level during neutron scattering experiments. The time-of-flight spectrometer FOCUS for cold neutrons at SINQ at the Paul Scherrer Institute (Villigen, Switzerland) was used for water dynamics studies. The QENS measurements of samples were carried out at 300 K with an incident wavelength of 4.32 Å and a resolution of 90 μeV in order to cover a correlation time ranging from 0.1 to 10 ps. The QENS signal was treated with the DAVE software and analyzed using the QENSH program provided by LLB and designed to treat such data.

Figure 1. Evolutions of the altered glass thickness AGth(B) (a) and of the alteration rate (b) as a function of time.

Table 3. Altered Glass Thickness AGth(B) and Retention Factor RF of the Si, Na, and Ca of the Glasses Altered for 7, 22, and 58 days reference

alteration time (days)

AGth(B)a (nm)

RF(Si)

RF(Na)

RF(Ca)

7 22 58 7 22 58 7 22 58 7 22 58

842 1032 1298 658 748 794 135 167 209 155 208 250

94 94 95 88 88 90 83 84 87 82 87 89

27 28 28 9 9 10 2 2 3 2 5 5

99 99 100 87 94 95 -

Zr AG

Zr/Ca AG Al/Ca AG

Al/Zr AG

a

AG: altered glass.

These results highlight that glasses containing Al (Al/Ca G and Al/Zr G) are 4−6 times less altered than the glass without Al (Zr G and Zr/Ca G). For these alteration durations, this difference is mainly due to the various initial dissolution rates of the glasses lasting for a few hours in these experimental conditions,38 as well as the observed altered glass thicknesses obtained after 7 days (Table 3). Indeed, the initial rate, corresponding to the maximum dissolution rate of the glass matrix, depends on the glass composition and structure through the distribution of network-forming cations and exchange sites.42 Al and Zr have a low solubility, and their presence in the glass lead to a more polymerized network. However, the presence of Zr in the glass does not lead to an initial rate as low

■

RESULTS Glass Alteration Kinetics and Composition of Alteration Layers. Figure 1 presents the evolutions of the altered glass thicknesses AGth(B) calculated from the boron concentrations as a function of time and the Table 3 and the retention factors of Si, Na and Ca in the altered glass zones. Zr and Al concentrations in solution were under the detection limits, thus, we have considered a maximal retention of these elements into the alteration layers. 15984


Article

The Journal of Physical Chemistry C

The first one concerns the glasses without Al (Zr and Zr/Ca glasses). Their evolution is similar to that observed in ref 7 on simple borosilicate glasses containing ZrO2. The SAXS pattern of Zr and Zr/Ca altered glasses displays the presence of a second regime located between 0.5 and 3 nm−1. Using the Guinier approximation, the characteristic size is around 1.6 nm−1 in samples altered for 7 days. It can be related to structural inhomogeneities presenting typical dimensions around 4 nm. After 58 days of aging, this broad feature shift toward lower q values, and reaches 0.8 nm−1. Such an evolution indicates that the typical length of structural unhomogeneities has increased to a mean distance of ∼8 nm. These so-called unhomogenities could be attributed to the formation of a porous network emerging from the partial dissolution of the glass in the alteration layer according to a mechanism of hydrolysis of the siloxane bonds and the condensation of the silanol fractions.41,42 As no secondary structure peaks are visible in the SAXS patterns of these altered glasses, this porosity seems not organized. These results are confirmed by the TEM images of these glasses altered during 58 days (Figure 3), showing porous structures. In order to precise the mean pore size, Watershed image processing43 was used, and the Feret diameters, i.e., the longest distance from two contour points of the object, were determined. The mean pore sizes obtained are about 3.7 and 3.4 nm for the Zr and Zr/Ca AG, respectively, and porosity was estimated around 50%. These results are in line with the values calculated from SAXS patterns, since in the latter case the typical distances involve both the pore mean diameter and the wall diameter. The second type of evolution concerns the two glasses containing Al, namely, the Al/Ca and Al/Zr glasses. The scattering intensities after 7 days of alteration still varies as q−3 indicating the presence of rough interfaces but on a larger domain than for the pristine glasses (3 nm < d < 30 nm vs 6 < d < 30 nm). The TEM images of the alteration layers from Al/Ca and Al/Zr glasses altered for 58 days show two types of microstructure (Figure 3c,d). For both samples, the alteration layer consists in a gel presenting a gradient of pore size from 1 to 40 nm. This result can be correlated with the larger domain observed on SAXS data than for the pristine glasses characteristic of a highly polydisperse porosity.7 Such porosity gradient has been already characterized on similar glass.14,15,44 In the case of Al/Zr AG, some secondary phases having a filamentary structure are also observed at the top of the gel (Figure 3d). The gel thicknesses of the ultrathin cross sections varies between 300 and 450 nm and between 320 and 380 nm for the Al/Ca AG and Al/Zr AG 58 days, respectively. We have to conclude that the gel thickness obtained by TEM can be different from the glass altered thickness AGth(B) determined from solution analysis, i.e., 209 nm for Al/Ca AG and 250 nm for Al/Zr AG, due to the grain arrangement creating various local S/V and leading to slight variations of glass grain alteration. The temperatures of water removal in altered glasses were determined by TGA and DTA analyses. The results are given in Figure 4 and Table 4. The TGA analyses present continuous mass losses as a function of the temperature, which can be attributed to the desorption of physically adsorbed water and to the removal of the hydroxyl groups. These continuous mass losses do not allow dissociating the amount of physically adsorbed water from the amount of hydroxyl groups. As shown in Table 4, the weight percentage of physically adsorbed water and hydroxyl groups at 700 °C increases with the altered glass

as that for glass with Al. This phenomenon has already been observed in ref 38. As presented in Figure 1, the dissolution rates rapidly drop in the first 20 days. The different magnitudes and durations of these rate drops are related to the gel formation and its ability to restructure. This restructuration depends on the experimental conditions and on the elements existing in the pristine glass.7 As presented in Table 3, the alteration layers mainly consist of silica, with a low amount of insoluble element such as Zr or Al and Ca2+ or Na+ acting as charge compensators. For the Zr G, the presence of 4 mol % of ZrO2 in the glass leads to the formation of a gel containing Zr, limiting its restructuration.7,34 This is traduced by a constant alteration rate between 7 and 58 days (Figure 1b). When Ca is contained into the glass (Al/Ca G and Zr/Ca G) at the pH of the experiment, Ca is integrated into the gel playing the role of charge compensator of [AlO4]2− and [ZrO6/2]2− units as highlighted by the RF(Ca) (Table 3). This Ca integration into the gel is known to increase the protective properties of the gel.39,40 The consequence is a continuous decrease of the alteration rate for the Zr/Ca G and to a lesser extent for the Al/Ca G between 7 to 58 days (Figure 1). For the glass without Ca2+, it is Na+ that plays the role of charge compensator in the gel. Altered Glass Characterization. In order to determine the structural evolution of the porosity of the alteration layers and their morphologies, SAXS and TEM characterizations were performed. Figure 2 shows the SAXS patterns of the glasses and the altered glasses. While all pristine glasses present scattering intensities varying as a q−3 power law characteristic of rough interfaces at typical length scales ranging between 3 and 30 nm, the altered glasses present two types of evolution depending on the glass composition.

Figure 2. SAXS patterns of glasses and altered glasses. 15985


Article


Figure 3. STEM BF of ultramicrotomic thin sections of the alteration layers of (a) Zr AG, (b) Zr/Ca AG, (c) Al/Ca AG, and (d) Al/Zr AG 58 days. The images of Zr and Zr/Ca AG 58 days after a Watershed processing used to determine the mean pore size are also presented.

the amount of proton increase and not to the altered glass matrix. Thus, we have plotted the maximal intensity of the sum over Q of the scattering intensity as a function of the altered glass thicknesses AGth(B) (Figure 7). The results show that the scattering intensity has a linear evolution with AGth(B). This traduces the fact that the intensity of the signal is mainly related to the protons amount in the alteration layer which increases with time. The QENS spectra of the altered glasses depend strongly on the Al presence into the alteration layers. For the Zr and Zr/Ca AG, the QENS spectra show a much broader signal as compared with pristine glasses. Inversely, the Al/Ca and Al/Zr QENS spectra do not present any broadening compared to pristine glasses, meaning that protons are not mobile in these altered glass samples at a picosecond time scale. In order to clarify the behavior of protons in the gel formed from Zr and Zr/Ca glass alteration, we have determined their translational motions. To reach this goal, QENS data S(Q,ω) were fitted with a Dirac function modulated by the elastic contribution and one Lorentzian function having a half width at half-maximum of Γt(Q), which was overall convoluted by the instrumental resolution R(ω) (eq 4). This function was the most appropriate fitting function.

thickness. Comparing the samples, it means that there are around 4 times more physically adsorbed water and hydroxyl groups in samples Zr and Zr/Ca AG than in samples Al/Ca and Al/Zr AG. Moreover, the DTA curves show only one endothermic peak for all samples with a minimum comprised between 108 and 130 °C as presented in Table 4. The temperatures corresponding to this minimum depend on the altered glass since the temperatures are higher for the altered glass with Al (Al/Ca AG and Al/Zr AG) than for the altered glass without Al (Zr AG and Zr/Ca AG). These results mean that free water is more bonded in Al/Ca AG and Al/Zr AG than in Zr AG and Zr/Ca AG. Moreover, except for the Zr/Ca AG, the temperature of the minima increases with the alteration thickness, as presented on the Figure 5. Water Dynamics. In order to determine and compare the protons dynamics in the alteration layer of glasses with the one in the model materials, samples were analyzed by QENS.45 QENS spectra were collected for model materials, pristine glasses and altered glasses, and are presented in Figure 6a−e showing the sum over Q of QENS spectra. A vanadium foil was measured and used as a reference to determine the instrumental resolution. The QENS spectra of the model materials show a much broader signal as compared with vanadium and a similar evolution to that obtained on Mibemol (LLB, Saclay, France) and published in ref 20. This signal broadening comes from the small energy transfer, of typically a few millielectronvolts, between incoming neutrons and the moving protons of confined water and hydroxyl groups. In the case of the altered glasses, which present a slight amount of protons from water molecules and hydroxyl groups, compared to model materials filled of water, we had to control that the increase of the scattering signal was related mainly to

⎧ ⎡I ⎤ Γt(Q ) ⎥· S(Q , ω) = ⎨IEl . δ(ω − ω0) + ⎢ B 2 2 ⎣ 2π Γt (Q ) + (ω − ω0) ⎦ ⎩ ⎛ ω − ω0 ⎞⎫ exp⎜ ⎟⎬ ⊗ R(ω) + Background ⎝ kBT ⎠⎭ (4) ⎪

⎪

⎪

⎪

with kB the Boltzmann constant and T the temperature. 15986


Article


Figure 4. TGA and TDA curves of samples under air flow (a) Zr AG, (b) Zr/Ca AG, (c) Al/Ca AG, and (d) Al/Zr AG.

Table 4. Mean Temperature of Free Water Removal and H2O/OH Loss at 700 °C reference Zr AG

Zr/Ca AG Al/Ca AG Al/Zr AG


mean temperature of free water removal (°C)

H2O/OH loss at 700 °C (wt %)

7 22 58 7 22 58 7 22 58 7 22 58

114 117 120 111 111 108 118 128 130 118 128 130

3.3 4.0 4.3 3.6 3.9 4.2 0.5 0.7 0.9 0.8 1.1 13

Figure 5. Evolution of the temperature of the minima of the free water desorption as a function of the altered glass thickness.

that protons are strongly confined for distances longer than 6 Å. At higher Q, Γt asymptotically reaches a constant value for all samples characteristic of a jump-diffusion mechanism, which can be fitted according the Singwi and Sjolander model (SS)46 based on an exponential distribution of jump lengths given by the eq 5:

An example of fits of the QENS data S(Q,ω) for the sample Zr/Ca AG 58 days using eq 4 is presented in Figure 8. The evolutions of the Γt(Q2) obtained from the fits using one Lorentzian for model materials, Zr and Zr/Ca AG are presented in Figure 9. For bulk water, Γt presents a linear evolution as a function of Q2 characteristic of a Brownian diffusion and can be fitted using Fick’s law. In that case, a value of Dt = 2.5.10−9 m2·s−1 is obtained. Γt of confined protons in model materials is Q2dependent from 0.5 A−2. For Zr and Zr/Ca AG glasses, Γt is constant for Q2 lower than 1, meaning that the proton dynamics does not follow a Brownian diffusion and, therefore,

Γt(Q ) =

ℏDt Q 2 1 + Dt Q 2τt

(5)

with ℏ being the reduced Planck constant, τt the average residence time between two consecutive jumps, and Dt the translational diffusion coefficients of confined water molecules defined by Dt = ⟨l2⟩/6τt, with ⟨l2⟩ being the mean square jump length. For model materials, the best fits using eq 5 (Figure 9a) is obtained with a translational diffusion coefficient of bulk 15987


Article


Figure 6. (a−e). Sum of QENS spectra of dried and hydrated samples and vanadium obtained at room temperature and λ = 5.2 Å.

water (Dt(20 °C) = 2.5.10−9 m2·s−1) and residence times presented in Table 5. The same tendencies as in ref 20 are obtained. In the case of Zr and Zr/Ca altered glasses, Γt being constant for Q2 lower than 1, we have fitted Γt(Q2) considering the translational diffusion coefficient of bulk water ((Figure 9b,c)). The average residence times obtained for the altered glasses are presented in Table 5. First, the τt values obtained for the Zr and Zr/Ca altered glasses are lower than the τt obtained for the model materials and are from 2 to 5 times lower than the τt of the model materials having a lower pore size and Si−OH (MCMSi-2 and MCMSi-3) and Zr−OH pore wall surface (MCMZr-3). Second, τt varies as a function of time, and these variations depend on altered glass composition. The τt of proton in the gel of Zr AG decreases with the alteration duration, whereas the τt of proton in the gel of Zr/Ca AG increases. This result seems to be correlated to the alteration rate. Indeed, as presented on

Figure 7. Evolution of the maximum scattering intensities of the SUM of QENS signal of altered glasses as a function of the altered glass thickness.

15988


Article


Table 5. Fitting Parameters Dt and τt Obtained by Fitting of Γt(Q) Using the Jump Diffusion Model of the Singwi and Sjolander Model46,a reference


bulk focus MCMSi-2 MCM-Si-3 MCMAl-3 MCMZr-3 Zr AG

-

Zr/Ca AG

a

Figure 8. Examples of simulation of experimental S(ω) of Zr/Ca AG 58 days by one Lorentzian function from Q = 0.7 Å−1 to Q = 2.1 Å−1.

Dt (m2·s−1) 2.5 0.10−9 m2·s−1

7 22 58 7 22 58

2.5 0.10−9 m2·s‑1

2.5 0.10‑9 m2·s‑1

τt (ps) 1.7 1.6 2.9 2.9 1.8 1.4 0.9 0.6 1.0 1.8

± ± ± ± ± ± ± ± ± ±

0.3 0.1 0.3 0.2 0.2 0.1 0.1 0.2 0.3 0.1

The underlined values are fixed during the fitting procedure.

glass alteration rate, i.e., the tracers and water transport through the gel at a higher observation time. In order to determine whether the nature of the gel has an impact on the proportion of fixed proton in the time-window corresponding to the experimental resolution, we have calculated the elastic incoherent structure factor (EISF). EISF can be calculated experimentally as the ratio of the elastic intensity over the sum of the elastic intensity and quasi-elastic

Figure 10, the decrease of alteration rate of the Zr/Ca G goes with an increase of the τt, i.e., a longer immobilization of protons into the pore and/or at the pore surface with the alteration duration. The Zr glass does not present the same behavior. Indeed, while the alteration rate remains constant, the τt decreases with time. These results could suggest that the proton dynamics at a picosecond scale is correlated with the

Figure 9. Half width at half-maximum Γt(Q) obtained from the fitting of QENS spectra for bulk water, water confined in models materials, and Zr and Zr/Ca altered glasses at 300 K. 15989


Article


phenomena: the porosity of the gel and/or the wall and surface composition of the pore and/or the presence of ions in the leachate filling the porosity. First, the alteration layers of Al/Ca AG and Al/Zr AG present similar gradients of pore sizes ranging from 1 to 40 nm with secondary phases at the top of the gel for Al/Zr AG. In that case, pore sizes of Al/Ca AG and Al/Zr AG are higher than the pore size of the homogeneous gels of Zr and Zr/Ca AG, around 3.7 to 3.4 nm, respectively. Nevertheless, protons are not mobile in alteration layers of Al/Ca AG and Al/Zr AG, while protons motions can be characterized in the gels of Zr and Zr/Ca AG. Based on the results of refs 30 and 32, confinement has a strong impact on water motions beyond 2 nm of pore size. Thus, the strong immobilization of protons in Al/Ca AG and Al/Zr AG cannot be explained by the pore size. Second, in this study and in ref 20 we have shown that the pore surface could have a strong impact on the proton mobility at a picosecond scale. Indeed, protons are strongly immobilized in model materials having Al−OH and Zr−OH pore surface, which behave as Lewis acids with stronger H-bonds. Another interpretation of these results could be that alumina or zirconiabase surfaces strongly interact with the water dipoles and texture the water layers more strongly because they are more ionic than silica. However, as shown in Tables 2 and 3, the composition of the alteration layers mainly consist in SiO2/ SiOH with a low fraction of Al2O3/AlOH and ZrO2/Zr−OH, which could not be sufficient to explain our results. The charge compensator could also play a role as illustrated with the fraction of fixed protons in gels of Zr/Ca AG being higher than in the gel of Zr AG. Their porosities being the same, the only difference between these gels is the charge compensator, Ca2+ versus Na+, compensating the charge of [SiO4]4− and [ZrO6]2− units. In hydrous aluminosilicate glasses, a slower H2O molecule diffusion in nanoscale time was characterized when Ca2+ compensates [SiO4]4− and [AlO4]2− rather than when Na+ plays the same role. It was suggested that H2O molecules can form strong complexes with Ca2+, decreasing the protons mobility.25 Thus, the composition of the wall and the surface of the gels could partially explain our results. Third, the presence of ions in the leachate filling the porosity could also strongly modify the water properties.30,50 Indeed, Mamontov and al,30 have shown that water dynamics is slowed down in CaCl2 aqueous solution confined in mesoporous silica at a nanosecond scale. The cation charge and the ion hydration environment have been suggested to explain the dynamics of water molecules. Confined water molecules are expected to be strongly influenced by the presence of ions, which depends otherwise on the ions ability to be solvated, modifying the electrostatic interactions between the different elements of the system.51 The sorption of ions on the pore surface can also modify the interactions of water with the pore surface and then its dynamics and activity. Thus, the water dynamics at a picosecond scale could be mainly related to the composition of the leachate filling the porosity of the alteration layer, modifying the structuration of water through the ion solvation and ion sorption at the pore surface. This latter result is also really important since the ion adsorption on the surface and the ion solvation strongly impact the hydrolysis rate of materials52 and can modify the kinetics of dissolved element recondensation into the pore, and therefore the kinetics of glass alteration. This could explain the apparent relation of proton dynamics at a picosecond scale with the glass alteration rate, i.e., the tracers (boron) and water transport

Figure 10. Evolutions of the alteration rate and of the residence time τt as a function of time.

intensity obtained from the quasi-elastic spectra analysis. EISF(Q) values were fitted using the model EISFM of Volino and Dianoux,47,48 which considers the motion of the mobile hydrogen atoms as a restricted diffusion within a sphere of radius R and the fraction of fixed protons f at the time scale of the experiment:49 ⎡ 3j (QR ) ⎤2 ⎥ +f EISFM (Q ) = (1 − f )⎢ 1 ⎣ QR ⎦

(6)

where j1(QR) is the first-order spherical Bessel function. The best fitting parameters f and R are presented in Table 6 and Figure 11. Table 6. Physical Parameters Obtained by Fitting the EISF Using the Model of Volino and Dianoux47 reference


bulk focus MCMSi-2 MCM-Si-3 MCMAl-3 MCMZr-3 Zr AG

-

Zr/Ca AG

7 22 58 7 22 58

R (Å)

f

± ± ± ± ± ± ± ± ± ± ±

0.02 ± 0.02 0.69 ± 0.06 0.79 ± 0.04 0.89 ± 0.05 0.74 ± 0.05 0.79 ± 0.2 0.80 ± 0.2 0.84 ± 0.2 0.93 ± 0.2 0.92 ± 0.2 0.79 ± 0.2

6.4 2.7 2.0 2.1 1.9 2.5 2.9 2.2 2.7 2.1 2.5

0.2 0.1 0.2 0.1 0.2 0.2 0.2 0.2 0.2 0.2 0.2

The results obtained in this study on the model materials are close to the one obtained in ref 20. Moreover, the fractions of fixed protons in the gel of Zr and Zr/Ca AG are equal to or higher than those in MCMSi-2, MCMSi-3, and MCMZr-3 and increase with the mean temperature of the free water removal.

■

DISCUSSION This work has shown that water properties into the pores of the alteration layer such as the temperature of free water removal and the proton dynamics (type of mechanisms for distances higher than 6 Å, residence time, and fixed proton fraction) are modified compared to bulk water. These modifications are more or less important depending on the altered glass and seem related to the kinetics of glass alteration. A combination of several properties of the alteration layers can explain these 15990


Article


Figure 11. Experimental EISF and fitted EISFM curves as a function of Q for models materials and Zr and Zr/Ca altered glasses at T = 300 K.

through the gel at a longer observation time. This is the case of glasses containing Al presenting alteration layers containing immobile protons at a picosecond time scale, which are less altered and present long-term alteration rates lower than glasses containing Zr. This phenomenon is also observable for the glass Zr/Ca G, where the increase of residence time is associated with a decrease of glass alteration rate and vice versa for the Zr G. The variation of leachate composition, depending on the glass and when the alteration layer is formed, of the gel properties (composition and porosity) could explain the water mobility at a picosecond scale through ion solvation and ion sorption at a gel pore surface structuring water molecules.

experiments with ions such as those performed on mesoporous silica filled of water as Gouze et al.,53 are necessary to determine the validity of such laws in this peculiar environment.

CONCLUSION For the first time, we have characterized the proton dynamics in simplified glasses at various stages of alteration and shown that the mobility of protons at a picosecond time scale depends on the altered glass composition and can be related to the alteration rate of the glass. Even if the composition of the alteration layer can partially explain this result, the presence of ions in the leachate filling the gel porosity seems to be the main factor affecting the water/protons mobility. This latter result is really important since the ion solvation and the ion adsorption on the surface can strongly impact the hydrolysis rate of the pore wall of the gel and can also modify the kinetics of dissolved elements recondensation into the pore. This is also the case at the alteration layer−glass interface where hydrolysis drives the progress of glass alteration. Glass alteration is generally predicted using kinetics and thermodynamics laws coming from data obtained in bulk solution. However, two main issues limit the use of these laws: their validity in confined media and the unreachable composition of the leachate in confined media. Some

Corresponding Author

■

ASSOCIATED CONTENT

S Supporting Information *

Table listing the mass of raw products used for the elaboration of the soda-lime borosilicate glasses G. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcc.5b03073.

■

■

AUTHOR INFORMATION

*E-mail: [email protected]; Tel: +33 4 66 39 79 40; Fax: +33 4 66 79 66 20. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Patrick Jollivet for the glass elaboration and JeanMarc Zanotti for the fruitful discussion of the QENS data. We are grateful to the PSI-SINQ in providing the neutron facilities and to Fanny Juranyi, the local contact.

■

REFERENCES

(1) Beau, J.; Webber, W. Studies of nano-structured liquids in confined geometries and at surfaces. Prog. Nucl. Magn. Reson. Spectrosc. 2010, 56, 78−93. (2) Kalinichev, A. G. Molecular Structure and Dynamics of NanoConfined Water: Computer Simulations of Aqueous Species in Clay, Cement, and Polymer Membranes. In Transport and Reactivity of Solutions in Confined Hydrosystems; Mercury, L., Tas, N., Zilberbrand,

15991


Article

The Journal of Physical Chemistry C M., Eds.; NATO Science for Peace and Security Series C: Environmental Security; Springer: New York, 2014; pp 103−115. (3) Wang, Y. Nanogeochemistry: nanostructures, emergent properties and their control on geochemical reactions and mass transfers. Chem. Geol. 2014, 378−379, 1−23. (4) Oliver, S.; Kuperman, A.; Coombs, N.; Lough, A.; Ozin, G. A. Lamellar aluminophosphates with surface patterns that mimic diatom and radiolarian microskeletons. Nature 1995, 378, 47−50. (5) Murray, H. H. Traditional and new applications for kaolin, smectite, and palygorskite: a general overview. Appl. Clay Sci. 2000, 17, 207−221. (6) Simonyan, A. V.; Dultz, S.; Behrens, H. Diffusive transport of water in porous fresh to altered mid-ocean ridge basalts. Chem. Geol. 2012, 306−307, 63−77. (7) Cailleteau, C.; Angeli, F.; Devreux, F.; Gin, S.; Jestin, J.; Jollivet, P.; Spalla, O. Insight into silicate-glass corrosion mechanisms. Nat. Mater. 2008, 7, 978−983. (8) Buch, V.; Devlin, J. P. Water in Confined Media; Springer Series in Cluster Physics; Springer: Berlin, 2003; p 473. (9) Teixeira, J.; Zanotti, J.-M; Bellissent-Funel, M. C.; Chen, S.-H. Water in confined geometries. Phys. B 1997, 234−236, 370−374. (10) Hewish, N. A.; Enderby, J. E.; Howells, W. S. The dynamics of water molecules in ionic solution. J. Phys. C: Solid State Phys. 1983, 16, 1777−1791. (11) Herdman, G. J.; Salmon, S. Dynamics of water protons in concentrated gallium(3+), aluminum(3+), iron(3+) and dysprosium(3+) aqueous solutions: a study using incoherent quasi-elastic neutron scattering. J. Am. Chem. Soc. 1991, 113, 2930−2939. (12) Poinssot, C.; Gin, S. Long-term behavior science: the cornerstone approach for reliably assessing the long-term performance of nuclear waste. J. Nucl. Mater. 2012, 420, 182−192. (13) Gin, S.; Ribet, I.; Couillard, M. Role and properties of the gel formed during nuclear glass alteration: importance of gel formation conditions. J. Nucl. Mater. 2001, 298, 1−2 1−10. . (14) Rébiscoul, D.; Van der Lee, A.; Rieutord, F.; Né, F.; Spalla, O.; El-Mansouri, A.; Frugier, P.; Ayral, A.; Gin, S. Morphological evolution of alteration layers formed during nuclear glass alteration: new evidence of a gel as a diffusive barrier. J. Nucl. Mater. 2004, 326 (1), 9− 18. (15) Rébiscoul, D.; Frugier, P.; Gin, S.; Ayral, A. Protective properties and dissolution ability of the gel formed during nuclear glass alteration. J. Nucl. Mater. 2005, 342 (1−3), 26−34. (16) Gin, S.; Guittonneau, C.; Godon, N.; Neff, D.; Rebiscoul, D.; Cabie, M.; Mostefaoui, S. Nuclear glass durability: new insight into alteration layer properties. J. Phys. Chem. C 2011, 115, 18696−19706. (17) Valle, N.; Verney-Carron, A.; Sterpenich, J.; Libourel, G.; Deloule, E.; Jollivet, P. Elemental and isotopic (29Si and 18O) tracing of glass alteration mechanisms. Geochim. Cosmochim. Acta 2010, 74 (12), 3412−3431. (18) Gin, S.; Ryan, J. V.; Schreiber, D. K.; Neeway, J.; Cabié, M. Contribution of atom-probe tomography to a better understanding of glass alteration mechanisms: Application to a nuclear glass specimen altered 25 years in a granitic environment. Chem. Geol. 2013, 349− 350, 99−109. (19) Parruzot, B.; Jollivet, P.; Rébiscoul, D.; Gin, S. Long-term alteration of basaltic glass: Mechanisms and rates. 2015. Geochim. Cosmochim. Acta 2015, 154, 28−48. (20) Briman, I. M.; Rébiscoul, D.; Diat, O.; Zanotti, J. M.; Jollivet, P.; Barboux, P.; Gin, S. Impact of pore size and pore surface composition on the dynamics of confined water in highly ordered porous silica. J. Phys. Chem. C 2012, 116 (12), 7021−7028. (21) Briman, I. M.; Rébiscoul, D.; Diat, O.; Jollivet, P.; Zanotti, J. M.; Gin, S. Dynamics of water confined in gel formed during glass alteration at a picosecond scale. Procedia Earth Planet. Sci. 2013, 7, 733−737. (22) Takahara, S.; Kittaka, S.; Mori, T.; Kuroda, Y.; Yamaguchi, T.; Bellissent-Funel, M.-C. Neutron scattering study on dynamics of water molecules confined in MCM-41. Adsorption 2005, 11, 479−483.

(23) Sklari, S.; Rahiala, H.; Stathopoulos, V.; Rosenholm, J.; Pomonis, P. The influence of surface acid density on the freezing behavior of water confined in mesoporous MCM-41 solids. Microporous Mesoporous Mater. 2001, 49, 1−3. (24) Mitra, S.; Mukhopadhyay, R.; Tsukushi, I.; Ikeda, S. Dynamics of water in confined space (porous alumina): QENS study. J. Phys.: Condens. Matter 2001, 13, 8455−8465. (25) Indris, S.; Heitjans, P.; Behrens, H.; Zorn, R.; Frick, B. Fast dynamics of H2O in hydrous aluminosilicate glasses studied with quasielastic neutron scattering. Phys. Rev. B: Condens. Matter Mater. Phys. 2005, 71, 064205. (26) Swenson, J.; Bergman, R.; Longeville, S.; Howells, W. S. Dynamics of 2D-water as studied by quasi-elastic neutron scattering and neutron resonance spin-echo. Phys. B 2001, 301, 28−34. (27) Michot, L. J.; Delville, A.; Humbert, B.; Plazanet, M.; Levitz, P. Diffusion of water in a synthetic clay with tetrahedral charges by combined neutron Time-of-Flight measurements and molecular dynamics simulations. J. Phys. Chem. C 2007, 111, 9818−9831. (28) Marry, V.; Dubois, E.; Malikova, N.; Durand-Vidal, S.; Longeville, S.; Breu, J. Water dynamics in hectorite clays: Infuence of temperature studied by coupling neutron spin echo and molecular dynamics. Environ. Sci. Technol. 2011, 45, 2850−2855. (29) Bellissent-Funel, M. C.; Chen, S. H.; Zanotti, J. M. Singleparticle dynamics of water molecules in confined space. Phys. Rev. E: Stat. Phys., Plasmas, Fluids, Relat. Interdiscip. Top. 1995, 51, 4558. (30) Mamontov, E.; Cole, D. R.; Dai, S.; Pawel, M. D.; Liang, C. D.; Jenkins, T.; Gasparovic, G.; Kintzel, E. Dynamics of water in LiCl and CaCl2 aqueous solutions confined in silica matrices: A backscattering neutron spectroscopy study. Chem. Phys. 2008, 352, 117−124. (31) Floquet, N.; Coulomb, J. P.; Dufau, G.; André, G.; Kahn, R. Structural and dynamic properties of confined water in nanometric model porous materials (8 Å⩽⌀⩽40 Å). Phys. B 2004, 350, 265−269. (32) Bourg, I. C.; Steefel, C. I. Molecular dynamics simulations of water structure and diffusion in silica nanopores. J. Phys. Chem. C 2012, 116, 11556−11564. (33) Chen, H.; Wang, Y. Preparation of MCM-41 with high thermal stability and complementary textural porosity. Ceram. Int. 2002, 28, 541−547. (34) Bergeron, B.; Galoisy, L.; Jollivet, P.; Angeli, F.; Charpentier, T.; Calas, G.; Gin, S. First investigations of the influence of IVB elements (Ti, Zr, and Hf) on the chemical durability of soda-lime borosilicate glasses. J. Non-Cryst. Solids 2010, 356, 2315−2322. (35) Korpa, A.; Trettin, R. The influence of different drying methods on cement paste microstructures as reflected by gas adsorption: Comparison between freeze-drying (F-drying), D-drying, P-drying and oven-drying methods. Cem. Concr. Res. 2006, 36, 634−649. (36) Galle, C. Cement and Effect of drying on cement-based materials pore structure as identified by mercury intrusion porosimetry: A comparative study between oven-, vacuum-, and freeze-drying. Cem. Concr. Res. 2001, 31, 1467−1477. (37) Ehret, G.; Crovisier, J. L.; Eberhart, J. P. A new method for studying leached glasses: analytical electron microscopy on ultramicrotomic thin sections. J. Non-Cryst. Solids 1986, 86, 72−79. (38) Gin, S.; Beaudoux, X.; Angéli, F.; Jegou, C.; Godon, N. Effect of composition on the short-term and long-term dissolution rates of ten borosilicate glasses of increasing complexity from 3 to 30 oxides. J. Non-Cryst. Solids 2012, 358, 2559−2570. (39) Mercado-Depierre, S.; Angeli, F.; Frizon, F.; Gin, S. Antagonist effects of calcium on borosilicate glass alteration. J. Nucl. Mater. 2013, 441, 402−410. (40) Chave, T.; Frugier, P.; Gin, S.; Ayral, A. Glass−water interphase reactivity with calcium rich solutions. Geochim. Cosmochim. Acta 2011, 75, 4125−4139. (41) Casey, W. H.; Bunker, B. C. Leaching of mineral and glass surfaces during dissolution. Rev. Mineral. Geochem. 1990, 23, 397−486. (42) Bunker, B. C. Molecular mechanisms for corrosion of silica and silicate glasses. J. Non-Cryst. Solids 1994, 179, 300−308. (43) Zhu, Y.; Zuo, T.; Jiang, L.; Cai, Z. T.; Yan, C.; Liu, X.; Malcolm, X.; Wu, Q. L. Pore structure analysis on activated carbon fibers-By 15992


Article

The Journal of Physical Chemistry C cluster and watershed transform method. Appl. Surf. Sci. 2009, 256 (3), 587−592. (44) Rébiscoul, D.; Van der Lee, A.; Frugier, P.; Gin, S.; Ayral, A. Xray reflectometry characterization of SON 68 glass alteration films. J. Non-Cryst. Solids 2003, 325, 113−123. (45) Bée, M.; Amoureux, J. P. Incoherent quasielastic neutron scattering from plastic crystals: An approximate method for multiple scattering evaluation. Mol. Phys. 1980, 41, 287. (46) Singwi, K. S.; Sjolander, A. Diffusive motions in water and cold neutron scattering. Phys. Rev. 1960, 119, 863−71. (47) Volino, F.; Dianoux, A. J. Neutron incoherent scattering law for diffusion in a potential of spherical symmetry: general formalism and application to diffusion inside a sphere. Mol. Phys. 1980, 41, 271−279. (48) Zanotti, J.-M. Vibrations et relaxations dans les molécules biologiques. Apports de la diffusion incohérente inélastique de neutrons. J. Phys. IV 2005, 130, 87−113. (49) Wanderlingh, U. N.; Giordano, R.; Dianoux, A. J.; Wanderlingh, F. IQENS dynamics in hydrated Crambin. Phys. A 2002, 304, 276− 282. (50) Marry, V.; Dubois, E.; Malikova, N.; Breu, J.; Haussler, W. Anisotropy of water dynamics in clays: insights from molecular simulations for experimental QENS analysis. J. Phys. Chem. C 2013, 117, 15106−15115. (51) Parsegian, V. A.; Zemb, T. Hydration forces: Observations, explanations, expectations, questions. Curr. Opin. Colloid Interface Sci. 2011, 16, 618−624. (52) Jollivet, P.; Gin, S.; Schumacher, S. Forward dissolution rate of silicate glasses of nuclear interest in clay-equilibrated groundwater. Chem. Geol. 2012, 330-331, 207−217. (53) Gouze, B.; Cambedouzou, J.; Parres-Maynadié, S.; Rébiscoul, D. How hexagonal mesoporous silica evolves in water on short and long term: Role of pore size and silica wall porosity. Microporous Mesoporous Mater. 2014, 183, 168−176.

15993


Water Dynamics in Nanoporous Alteration Layer Coming from Glass

Recommend Documents