www.elsevier.com/locate/wasman

Waste Management 26 (2006) 758–768

The effect of sodium chloride on the dissolution ofcalcium silicate hydrate gels

J. Hill a,*, A.W. Harris a, M. Manning b, A. Chambers c, S.W. Swanton c

a UK Nirex Limited, Curie Avenue, Harwell, Didcot Oxfordshire, OX11 0RH, United Kingdomb Formerly AEA Technology, Building 220, Harwell, Didcot Oxfordshire, OX11 0RA, United Kingdom

c Serco Assurance, Building 150, Harwell International Business Centre, Didcot, Oxfordshire, OX11 0RA, United Kingdom

Accepted 31 January 2006Available online 13 March 2006

Abstract

The use of cement based materials will be widespread in the long-term management of radioactive materials in the United Kingdom.One of the applications could be the Nirex reference vault backfill (NRVB) as an engineered barrier within a deep geological repository.NRVB confers alkaline conditions, which would provide a robust chemical barrier through the control of the solubility of some keyradionuclides, enhanced sorption and minimised corrosion of steel containers. An understanding of the dissolution of C–S–H gels incement under the appropriate conditions (e.g., saline groundwaters) is necessary to demonstrate the expected evolution of the chemistryover time and to provide sufficient cement to buffer the porewater conditions for the required time.

A programme of experimental work has been undertaken to investigate C–S–H gel dissolution behaviour in sodium chloride solutionsand the effect of calcium/silicon ratio (C/S), temperature and cation type on this behaviour. Reductions in calcium concentration and pHvalues were observed with samples equilibrated at 45 �C compared to those prepared at 25 �C. The effect of salt cation type on salt-con-centration dependence of the dissolution of C–S–H gels was investigated by the addition of lithium or potassium chloride in place ofsodium chloride for gels with a C/S of 1.0 and 1.8. With a C/S of 1.0, similar increases in dissolved calcium concentration with increasingionic strength were recorded for the different salts. However, at a C/S of 1.8, anomalously high calcium concentrations were observed inthe presence of lithium.� 2006 Elsevier Ltd. All rights reserved.

1. Introduction

UK Nirex Ltd. has been established to provide the Uni-ted Kingdom with safe, environmentally sound and pub-licly acceptable options for the long-term management ofradioactive materials. These materials include operationaland decommissioning wastes from nuclear power plantsand other wastes from industrial, defense and medicalapplications involving radioactivity. In order to deliver thismission Nirex undertakes various activities including devel-oping concepts for the long-term management of wastes.The Nirex Phased Geological Repository Concept is basedon a deep geological repository and provides multi-barrier

0956-053X/$ - see front matter � 2006 Elsevier Ltd. All rights reserved.

doi:10.1016/j.wasman.2006.01.022

* Corresponding author. Tel.: +44 0 1235 825 203; fax: +44 0 1235 825289.

E-mail address: joanne.hill@nirex.co.uk (J. Hill).

containment using a combination of engineered and natu-ral barriers including:

� physical containment in stainless steel and concretecontainers;� geological isolation in vaults excavated deep

underground;� a chemical barrier provided by backfilling the vaults with

a specially designed cement-based material (Nirex refer-ence vault backfill, NRVB) based on Portland cementcombined with fine limestone and hydrated lime aggre-gate (Francis et al., 1995);� geological containment through eventual sealing of the

repository (Nirex, 2003).

The chemical barrier is based on the control of the chem-ical conditions within the repository through the slow

J. Hill et al. / Waste Management 26 (2006) 758–768 759

dissolution of the NRVB into slowly-flowing groundwater.The demonstration of the persistence of the necessary condi-tions for an extended time period and the provision of a suit-able quantity of NRVB depends on an understanding of thedissolution of material and the minerals that comprise it.

The major constituents of Portland cement are calciumsilicates and aluminates, which react with water to producea firm, hard mass. The main products of this reaction arecalcium hydroxide and calcium silicate hydrate (C–S–H)and the behaviour of the NRVB is, therefore, expected tobe dominated by the dissolution of these materials (Atkin-son, 1985; Atkinson et al., 1987; Glasser et al., 1985; Har-ris, 1998). Calcium silicate hydrates possess a remarkablelevel of structural complexity with more than 30 crystallinephases known. Preparations from room temperature havestructures that range from semi-crystalline to nearly amor-phous (Taylor, 1997). The term ‘C–S–H gel’ is used in thispaper for the range of materials prepared. The range ofpossible compositions may encompass phases with cal-cium/silicon molar ratios (C/S) in the range 0.8–1.8although spatial variations are commonly observed on avery fine scale (Richardson, 1999). Although C–S–H gelsare believed to be metastable with respect to crystallineC–S–H materials such as tobermorite and afwillite, theyare known to persist for extremely long times under geolog-ical conditions (McConnell, 1955; Milodowski et al., 1989).

The dissolution of C–S–H gels of varying C/S into purede-ionised water has been the subject of a number of exper-imental studies (Brown et al., 1984; Flint and Wells, 1934;Fujii and Kondo, 1981; Greenberg and Chang, 1965; Gru-tzeck et al., 1989; Kalousek, 1952; Roller and Ervin, 1940;Taylor, 1950). The results of these studies were similar inthat they showed the equilibrated calcium concentrationand the pH decrease with decreasing C/S, whereas the sili-con concentration increased. It has become evident, how-ever, that some groundwater solutes, in particular alkalimetals and perhaps counter-ions such as chloride, interactwith C–S–H gels during dissolution (Glasser et al., 1999).Such interactions apparently affect calcium solubility andtherefore the net solubility and chemical conditioning per-formance of materials such as NRVB. It is now apparentthat experimental data for dissolution in pure water donot adequately describe the behaviour of C–S–H gels in sal-ine solutions, and therefore additional data are required tounderpin models of the long-term chemical conditioningperformance of NRVB.

For this reason, a series of experiments were performed(Chambers et al., 2005) to investigate the dissolutionbehaviour of a series of synthetic C–S–H gels in sodiumchloride (NaCl) solutions. The aim of the study was to pro-vide experimental data to allow the interpretation of theeffect of NaCl on C–S–H dissolution and the developmentof a mathematical model to describe the effect. This paperconcentrates on the experimental studies.

C–S–H gels of varying C/S (in the range 0.8–2.0) wereprepared and the amount of water associated with the gelsdetermined by two methods of drying to a constant mass.

This allowed the calculation of the ‘bound’ water associ-ated with each gel. The gels were then studied in a seriesof dissolution experiments to investigate the effects of NaClconcentration and water:solid (w:s) ratio on the aqueouselemental concentrations of calcium and silicon and onthe pH. Uncertainties highlighted from this study led to asecond set of experiments being carried out to investigatethe effect of temperature and cation type (Swanton et al.,2005).

2. Experimental

2.1. Preparation of gels

A silica stock (�5% by mass) was prepared by diluting aLudox TM-50 suspension (50% by mass) using nitrogen-sparged de-ionised water. Approximately 35 g of the solstock was weighed into glass beakers and oven-dried to aconstant weight. The results showed that a 5.4% sol bymass had been prepared.

A number of C–S–H gels were prepared in 1 l bottlesinside a nitrogen atmosphere glove box with different C/Sratios: 0.8, 1.0, 1.1, 1.2, 1.5, 1.8 and 2.0. The requiredamount of CaO was dispersed in a known amount of nitro-gen-sparged de-ionised water to which the appropriateamount of stock silica gel was added to give the correctC/S. The bottles were shaken vigorously as the gels beganto form and left for at least 1 month to cure.

2.2. Drying experiments

The gels were prepared in a significant amount of excesswater (a water:solid ratio (w:s) of greater than 10) as it hadbeen found this was necessary to prevent them from settingsolid. To determine the bound water associated with thegel, it was necessary to remove this excess. The weightsof the gels were determined after three different treatments:

� centrifugation and decanting;� drying over silica gel at ambient temperature;� oven-drying at 105 �C.

Tightly-bound water was defined as that removed byoven-drying but not by drying over silica gel. Loosely-bound water was defined as that removed by drying oversilica gel but not by centrifugation. In this relatively simplemanner, the necessary information was obtained to therequired accuracy.

Portions of 17–23 g (including excess free water) of eachC–S–H mixture were assessed to be equivalent to �2 g ofC–S–H. These amounts were placed in polypropylene cen-trifuge tubes and placed in an oven at 105 �C inside a nitro-gen atmosphere glove box. All of the samples had dried toconstant weight within 53 days. Portions of the gel mixturewere also placed in polypropylene tubes, centrifuged, theexcess liquid decanted and the mass determined. The tubeswere then left to dry over periodically-replenished silica gel

Table 2Matrix of C–S–H gel dissolution experiments (second series)

Equilibration temperature Ambientc 25 �C 45 �C

C–S–H gel C/S 1.0 1.0 1.5 1.8 1.0 1.5Equilibration time (days)a

1 X

7 X

6 X

90 X

760 J. Hill et al. / Waste Management 26 (2006) 758–768

and, within a period of 320 days, all of the samples haddried to a constant weight.

With the masses of the tightly- and loosely-bound waterassociated with each C/S ratio of C–S–H gel determined,the correct amount of solid NaCl necessary to give therequired aqueous concentration could be calculated.

2.3. Dissolution experiments

pH equilibration time trials were performed on each gel.Samples of each gel mixture were weighed in duplicate intocentrifuge tubes; one sample was equilibrated in a0.1 mol dm�3 NaCl solution and the second in a1.0 mol dm�3 solution. The evolution of the pH was deter-mined over 1 month. The results of the evolution of the pHshowed that only slight changes occurred, with the largestchange taking place at very short times. It was concludedfrom this that a 1 month equilibration time was sufficientfor all experiments but further experiments were carriedout to investigate the effect of equilibration time (Swantonet al., 2005).

At least one batch of gels at each C/S ratio was used toprepare a C–S–H gel equilibrated water for use in the dis-solution experiments. After curing for at least 1 month, theequilibrated water was removed from the gel by vacuum fil-tration through a 0.45 lm Nalgene PES membrane filterunit that had been pre-conditioned.

Two sets of experiments were performed initially(Chambers et al., 2005), one to investigate the dissolutionof gels in varying sodium chloride concentration and thesecond to investigate the influence of varying w:s ratiosat a constant NaCl concentration. The parameters are sum-marised in Table 1. All experiments were left to equilibratefor 1 month in a nitrogen atmosphere glove box.

From the analysis of the results of the initial study, sev-eral further potentially useful experiments were identifiedand were conducted for the second part of this investigation.

Table 1Matrix of C–S–H gel dissolution experiments (first series)

C–S–H gel C/S ratio

0.8 0.9 1.0 1.1 1.2 1.5 1.8 2.0

Concentration of sodium chloride (mol dm�3)a

0.1 X X X X X X X X

0.3 X X X X X X X X

0.5 X X X X X X X X

0.7 X X X X X X X X

1.0 X X X X X X X X

Water/solid ratio (cm3 g�1)b

13 X X X X

20 X X X X

30 X X X X

50 X X X X

80 X X X X

100 X X X X

a Experiments performed at w/s ratio of 20:1.b Experiments performed at a sodium chloride concentration of

0.5 mol dm�3.

These experiments are summarised in Table 2. The C–S–Hgels were prepared to the same recipe as that used in the firstinvestigation outlined above. The experiments examined theeffect of equilibration time on the dissolution of a C/S = 1.0gel at a constant NaCl concentration at 25 �C, the effect ofequilibration temperature as a function of NaCl concentra-tion and the influence of cation type (lithium or potassiumchloride in place of NaCl) on dissolution of C–S–H gel. Fur-ther details of experimental procedure can be found inChambers et al. (2005) and Swanton et al. (2005).

3. Results

3.1. Preparation and characterisation of C–S–H gels

The results of the initial C–S–H gel drying experimentsare summarised in Table 3. The gel masses after oven dry-ing and drying over silica gel are expressed as a percentageof the mass of gel mixture (including excess water), and the‘tightly-bound’ water was calculated as the difference.

3.2. C–S–H gel dissolution experiments

3.2.1. Effect of NaCl concentrations

3.2.1.1. pH. The final pH values are presented in Fig. 1. Aswould be expected from experiments in non-saline systems,the main effect on the pH was the C/S ratio. The valuesranged from pH 12.6 at C/S = 2.0 to 10.9 at C/S = 0.8.

Concentration of NaCl

(mol dm�3)b

0 X X X X

0.1 X X X X

0.3 X X X X

0.5 X X X X X

0.7 X X X X

1.0 X X X X

Concentration of LiCl

(mol dm�3)b

0.1 X X

0.5 X X

1.0 X X

Concentration of KCl

(mol dm�3)b

0.1 X X

0.5 X X

1.0 X X

a Experiments performed at a NaCl concentration of 0.5 mol dm�3.b Experiments equilibrated for 1 month.c Equilibrated at ambient temperature in a nitrogen atmosphere glove

box.

10.5

11.0

11.5

12.0

12.5

13.0

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6

NaCl Concentration/mol dm-3

pH

C/S 0.8C/S 0.9C/S 1.0C/S 1.1C/S 1.2C/S 1.5C/S 1.8C/S 2.0C/S 0.85 Glasser dataC/S 1.1 Glasser dataC/S 1.4 Glasser dataC/S 1.8 Glasser dataCHGlasser data

Nirex Data at w/s 20 cm3 g-1

Glasser data at w/s 25 to 40 cm3 g-1

CH denotes calcium hydroxide solid

Fig. 1. Variation of equilibrated pH with NaCl concentration for various C–S–H gels.

Table 3Mass fractions of solids and tightly-bound water in C–S–H gel mixtures and derived mole ratios

C/S ratio 0.8 0.9 1.0 1.1 1.2 1.5 1.8 2.0

Mean % by mass ofsolid in gel mixtureafter centrifugation

64.07 73.87 65.97 77.04 74.18 62.72 64.07 61.39

Mean % by mass ofsolid in gel mixtureafter drying oversilica gela

7.864 8.290 8.294 8.094 8.003 8.955 9.391 9.364

Mean % by mass ofsolid in gel mixtureafter oven dryingat 105 �Cb

7.403 7.965 7.813 7.899 7.626 8.695 9.318 9.453

Mean % by mass ofloosely-bound waterc

56.20 65.58 57.67 68.94 66.18 53.77 54.68 52.03

Mean % by mass oftightly-bound waterd

0.461 0.325 0.481 0.195 0.377 0.260 0.073 �0.089

Amount per unit massof gel mixture(mol kg�1)

CaO 0.52 0.58 0.62 0.64 0.65 0.80 0.90 0.95SiO2 0.65 0.65 0.61 0.58 0.54 0.53 0.50 0.48H2Oe 0.59 0.63 0.63 0.59 0.64 0.70 0.76 0.64

Mole ratiose

Ca 0.80 0.90 1.00 1.10 1.20 1.51 1.81 2.01Si 1 1 1 1 1 1 1 1H2O 0.90 0.98 1.03 1.03 1.20 1.31 1.54 1.35

a Mean of four determinations.b Mean of two determinations.c ‘Loosely-bound’ water is defined as that removed by drying over silica gel but not by centrifugation/ decanting.d ‘Tightly-bound’ water is defined as that removed by oven drying at 105 �C but not by drying over silica gel.e Based on the amount of water determined from the mass of gel after silica gel drying minus the mass of CaO + SiO2 in the gel.

J. Hill et al. / Waste Management 26 (2006) 758–768 761

A small decrease in the pH of about 0.2 was observed foreach gel as the NaCl concentration was increased from0.1 to 1.0 mol dm�3. The results are consistent with thedata determined by Glasser et al. (1999), which are alsoshown in Fig. 1.

3.2.1.2. Ca concentration. The equilibrated aqueous Caconcentrations are shown in Fig. 2. Both the C/S ratio ofthe gels and the NaCl concentration were found to havea strong effect on the Ca concentration. For all of the gels,the Ca concentration increased with NaCl concentration,

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6

NaCl Concentration/ mol dm-3

Eq

uili

bra

ted

Ca

con

cen

trat

ion

/ mm

ol d

m-3

C/S 0.8C/S 0.9C/S 1.0C/S 1.1C/S 1.2C/S 1.5C/S 1.8C/S 2.0C/S 0.85 Glasser dataC/S 1.1 Glasser dataC/S 1.4 Glasser dataC/S 1.8 Glasser dataCH Glasser data

Nirex Data at w/s 20 cm3 g-1

Glasser data at w/s 25 to 40 cm3 g-1

CH denotes calcium hydroxide solidIndicative error bars are shown for C/S 0.9 gel data

see Section 3.2.2 for explanation

Fig. 2. Variation of equilibrated calcium concentration with sodium chloride concentration for various C–S–H gels.

762 J. Hill et al. / Waste Management 26 (2006) 758–768

with the most rapid increase occurring at lower NaCl con-centrations. These results are again compared and consis-tent with the data produced by Glasser et al. (1999).

The overall trend shown by varying the C/S ratio of thegels was, as expected, an increase in the aqueous Ca con-centration with increasing C/S ratio. Whilst this trendwas reasonably consistent at the lowest NaCl concentra-tion, some of the Ca concentration curves were found tocross over at higher NaCl concentrations. In particular,the curves for the C–S–H gels with C/S of 0.8 and 1.0 werefound to increase more rapidly and give higher Ca concen-trations at higher NaCl levels, than was the case for somegels of higher C/S. However, no corresponding cross-overwas observed in the pH or Si concentration data.

3.2.1.3. Silicon concentration. The major effect on the aque-ous Si concentration was due to the C/S ratio of the gels,with the measured concentrations decreasing at higher C/S ratios. The NaCl concentration had little effect on theSi concentration, with the exception of the data for the

0.0

5.0

10.0

15.0

20.0

25.0

30.0

0 20 40 60

Water/Solid Ra

Eq

uili

bra

ted

Ca

con

cen

trat

ion

/ mm

ol d

m-3

Fig. 3. Variation of equilibrated calcium concentratio

lowest C/S ratio C–S–H gel. In this case (C/S = 0.8), adecrease in the Si concentration at higher NaCl levelswas observed.

3.2.2. Effect of water/solid ratio

3.2.2.1. pH. Little dependence of the final pH values on thew/s ratio was observed, with the exception of the lowest C/S ratio C–S–H gel experiment. In this case (C/S = 0.8), thepH was found to increase from around 10.8 to 11.2between w/s ratios of 13 and 100 cm3 g�1. As was the casefor the experiments at varying NaCl concentrations, thefinal pH was found to increase with increasing C/S ratio.

3.2.2.2. Ca concentration. For the higher C/S ratio gel (C/S = 1.8), the highest aqueous Ca concentrations wereobserved at around 2.8 · 10�2 mol dm�3 (see Fig. 3). Nodependence on C/S ratio was observed in this case. How-ever, for the lower C/S ratio gels (0.80–1.2), a decrease inCa concentration with increasing w/s ratio was found athigher w/s ratios. A cross-over in the Ca concentration

80 100 120

tio/ cm3 g-1

C/S 0.8C/S 1.0C/S 1.2C/S 1.8

Nirex data at 0.5 mol dm-3 NaCl

n with water/solid ratio for various C–S–H gels.

J. Hill et al. / Waste Management 26 (2006) 758–768 763

curves was observed for C/S ratios of 0.8, 1.0 and 1.2. At13 cm3 g�1, the highest Ca concentration was measuredfor the C/S = 0.8 gel and the lowest for the C/S = 1.2gel. At w/s ratios of 30 cm3 g�1 and above, the trend hadreversed with the highest Ca concentration being measuredfor the C/S = 1.2 gel and the lowest for the C/S = 0.8 gel.The cross-over point was very close to a w/s ratio of20 cm3 g�1, the value at which the experiments of varyingNaCl concentrations were carried out. The similarity ofthe results at a w/s ratio of 20 cm3 g�1 might be relatedto the observed trends in the experiments performed atvarying sodium chloride concentrations described above.In the latter experiments, also performed at a w/s ratio of20 cm3 g�1, the calcium concentrations did not follow asimple trend with C/S for gels with C/S from 0.8 to 1.2.

3.2.2.3. Silicon concentration. Little dependence of the Siconcentration on w/s ratio was found for C/S ratios of1.8 and 1.2 (see Fig. 4). For the C/S ratio = 1.0 gel, a smallincrease in Si concentration at higher w/s ratios wasobserved, whereas in the C/S = 0.8 case a significantincrease of around 50% was observed across the range ofw/s ratios investigated.

3.3. Effect of temperature and cation experiments

3.3.1. Effect of equilibration time

The solution pH and calcium concentration were variedas a function of equilibration time. For samples measuredat different temperatures (water bath at 25 �C and glovebox at ambient), the pH measurements were in good agree-ment. However, for samples measured at different times,there was a wider variation in pH. These differences werethought to be due to a combination of measurement errorsand temperature variations between the samples whenmeasurements were taken. The experimental uncertaintiesof associated with the dissolved calcium concentrationswere ±10%, but the measured calcium concentrations werein agreement with experimental error. Thus, for C/S = 1.0

0.0

0.1

0.2

0.3

0.4

0 20 40 60

Water/Solid R

Eq

uili

bra

ted

Si c

on

cen

trat

ion

/ mm

ol d

m-3

Nirex data at 0.5 mol dm-3 NaCl

Fig. 4. Variation of equilibrated silicon concentratio

gel, there appeared to be no significant variation in C–S–Hgel equilibration on salt addition with equilibration timebetween 1 and 90 days. It was, therefore, assumed thatthe gels equilibrated rapidly on salt addition.

3.3.2. Effect of temperatureThe results of the effect of temperature on pH, Ca con-

centration, and Si concentration are shown in Figs. 5–7,respectively. Each data point is the average of duplicateexperiments and is compared to results from samples equil-ibrated at lower temperatures.

It can be seen from Fig. 5 that for the C/S = 1.5 samplesthere was a significant decrease in pH with an increase inequilibration temperature when it was measured at theequilibration temperature (around 12.5 at 25 �C to 11.8at 45 �C). The pH determined at 25 �C was about 0.1 Ulower than that determined previously under ambient con-ditions and sampled at a temperature of 14 �C. However,when the sample was allowed to cool from 45 �C to ambi-ent temperature before measurement occurred, there wasno significant difference in pH from that determined at25 �C.

A slightly larger decrease in measured pH of about 0.2–0.3 U was found between the C/S = 1.0 gels equilibrated at25 �C in this study and those equilibrated under ambientconditions previously. The measured pH values for theC/S = 1.0 gels that were equilibrated for 1 month at45 �C and then allowed to cool to ambient temperaturebefore sampling were almost identical to those equilibratedat 25 �C.

Results of the preliminary study showed an increase incalcium concentration with an increase in C/S ratio(Fig. 1), and data in the second study exhibited the sametrend (Fig. 6). Also, as expected, there was a generaldecrease in calcium concentration with an increase in equil-ibration temperature and the dependence on NaCl concen-tration appeared to be reduced.

The effect of temperature on silicon concentrationappeared to be negligible (Fig. 7) although the data for

80 100 120

atio/ cm3 g-1

C/S 0.8C/S 1.0C/S 1.2C/S 1.8

n with water/solid ratio for various C–S–H gels.

0.00

0.02

0.04

0.06

0.08

0.10

0 0.2 0.4 0.6 0.8 1 1.2

Salt concentration / mol dm-3

Sili

con

co

nce

ntr

atio

n /

mm

ol d

m-3

C/S 1.0 NaCl ambient C/S 1.0 NaCl 25 oCC/S 1.0 NaCl 45 oCC/S 1.5 NaCl ambient C/S 1.5 NaCl 25 oCC/S 1.5 NaCl 45 oC

Fig. 7. Equilibrium silicon concentration for C–S–H Ca/Si = 1.0 and 1.5 gels with increasing salt concentration as a function of equilibration temperature.

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

0 0.2 0.4 0.6 0.8 1 1.2

Salt concentration / mol dm-3

Cal

ciu

m c

on

cen

trat

ion

/ m

mo

l dm

-3

C/S 1.0 ambient C/S 1.0 25 oCC/S 1.0 45 oCC/S 1.5 ambient C/S 1.5 25 oCC/S 1.5 45 oC

Fig. 6. Equilibrium calcium concentrations for C–S–H Ca/Si = 1.0 and 1.5 gels with increasing sodium chloride concentration at different equilibrationtemperatures.

11.6

11.8

12.0

12.2

12.4

12.6

12.8

0 0.2 0.4 0.6 0.8 1 1.2

Salt concentration / mol dm-3

pH

C/S 1.0 ambient

C/S 1.0 25 oC

C/S 1.0 45 oC cooled to 23 oC

C/S 1.5 ambient

C/S 1.5 25 oC

C/S 1.5 45 oC

C/S 1.5 45 oC cooled to 23 oC

Fig. 5. Equilibrium pH for C–S–H Ca/Si 1.0 and 1.5 gels with Increasing salt concentration as a function of equilibration temperature and measurementtemperature.

764 J. Hill et al. / Waste Management 26 (2006) 758–768

J. Hill et al. / Waste Management 26 (2006) 758–768 765

C/S = 1.5 samples equilibrated at 25 and 45 �C seemed tobe slightly higher than that measured at ambient tempera-ture previously. There also appeared to be a slight decreasein silicon concentration with increasing NaCl.

3.3.3. Effect of cation typeFigs. 8 and 9 show the effects of cation type on calcium

concentration and pH, respectively. The calcium concen-tration (Fig. 8) was found to increase with salt concentra-tion for all salts. The most significant difference betweenthe samples using different cation types occurred for theC/S = 1.8 gel. In this case, the LiCl caused a greaterincrease in the calcium concentration than that observedfor either sodium or potassium. Also, for the C/S = 1.0gel the dissolved calcium concentrations were slightly lowerin the presence of lithium than for either sodium or potas-sium at the two lower salt concentrations studied.

11.4

11.6

11.8

12.0

12.2

12.4

12.6

12.8

0 0.2 0.4 0.6

Salt concentration

pH

Fig. 9. Variation of equilibrated pH for various C–S–H gels with

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

0 0.2 0.4 0.6

Salt concentration

Cal

ciu

m c

on

cen

trat

ion

/ m

mo

l dm

-3

Fig. 8. Variation of equilibrated calcium concentration for various C–S–H g

Fig. 9 illustrates the general trend of increasing pH withincreasing salt concentration and increasing C/S. Alsothere were differences caused by the type of cationemployed in that, regardless of C/S, the trend of decreasingpH was KCl > NaCl > LiCl.

4. Discussion

4.1. Effect of NaCl concentration

A series of C–S–H gels dissolution experiments wereundertaken at various NaCl concentrations and w:s ratios.The results indicated that NaCl had a significant effect onthe dissolution behaviour. In order to investigate furtherand isolate the data relating to the effect of the NaCl, addi-tional experiments were performed to examine the effects ofequilibration time, temperature and salt cation type.

0.8 1 1.2

/ mol dm-3

C/S 1.0 LiCl

C/S 1.0 NaCl

C/S 1.0 KCl

C/S 1.5 NaCl

C/S 1.8 LiCl

C/S 1.8 KCl

increasing concentration of chloride salt with different cations.

0.8 1 1.2

/ mol dm-3

C/S 1.0 NaCl ambient

C/S 1.5 NaCl ambient

C/S 1.8 NaCl ambient

C/S 1.0 LiCl 25 oC

C/S 1.0 NaCl 25 oC

C/S 1.0 KCl 25 oC

C/S 1.5 NaCl 25 oC

C/S 1.8 LiCl 25 oC

C/S 1.8 KCl 25 oC

els with increasing concentration of chloride salts with different cations.

766 J. Hill et al. / Waste Management 26 (2006) 758–768

Previous studies (Brown et al., 1984; Flint and Wells,1934; Fujii and Kondo, 1981; Greenberg and Chang,1965; Grutzeck et al., 1989; Kalousek, 1952; Roller andErvin, 1940; Taylor, 1950) investigating the dissolution ofC–S–H in pure water observed that the equilibrated cal-cium concentration and the pH decreased with decreasingC/S, whereas the silicon concentration increased. The maindifference observed in this study was the enhancement ofthe dissolved calcium concentration at higher NaCl con-centrations, which is consistent with the results reportedby Glasser et al. (1999). It was also shown that increasingthe w:s resulted in a decrease in calcium concentration.

This observed increase in calcium concentration withincreasing concentration of NaCl means the assumptionthat a Ca–Na ion exchange process is occurring is reason-able. It is thought that, for high C/S gels, the aqueouschemistry would be determined by equilibrium with cal-cium hydroxide. The good agreement of the data for thecalcium concentration and pH of the high C/S gels withthat for calcium hydroxide reported by Glasser et al.(1999) bears this out. No ion exchange process would beexpected for calcium hydroxide so for systems at a higherC/S ratio in which both calcium hydroxide and C–S–Hgel are present, it would be expected that equilibration withcalcium hydroxide would constrain the calcium concentra-tion and pH to a narrow range of values.

4.1.1. Effect of equilibration time

The effect of the equilibration time may affect theobserved leaching behaviour of the gels depending uponthe kinetics of the dissolution process. Also, it is recognisedthat C–S–H gels formed by the direct reaction of calciumoxide and silica undergo changes in crystallinity with time(Chen et al., 2004).

A recent study by Walker (2003) has shown that the gelsprepared by direct reaction methods similar to thoseemployed in this study, require over 1 year to reach asteady state with the formation of more crystalline C–S–H phases, including 14 A tobermorite and a new 12 A C–S–H phase, in C/S = 1.0–2.0 gels. Younger gels are identi-fied as primarily containing the 14 A tobermorite-like gelreferred to as C–S–H(I) after Taylor (1986) consistent withresults of other workers (Chen et al., 2004).

The samples used in these studies were equilibrated forapproximately 1 month for most experiments. No signifi-cant differences in solution phase calcium concentrationswere found for C/S = 1.0 gels equilibrated for periods of1 day to 3 months. It can be concluded, therefore, thatequilibration of gels used in the current study was fastand that the 30-day equilibration times were appropriatefor all of the experiments with the young C–S–H gels used.The finely divided nature of the gels prepared is consideredlikely to underlie the fast equilibration of the gels tochanges in salinity. Coarser, more crystalline gels areexpected to respond more slowly to changes in solutionphase composition. The pH values of water equilibratedfor 51 days with coarse samples of natural C–S–H minerals

were found to be about 2 pH units lower than that equili-brated with finely-ground specimens of the same material(Atkinson et al., 1995).

It is considered unlikely that true thermodynamic equi-librium had been reached in the dissolution experimentspresented here, rather a metastable equilibrium wasachieved with a poorly crystalline C–S–H(I) phase. Chenet al. (2004) also pointed out that the poorly crystallisedC–S–H phases are examples of metastable phases and, assuch, their equilibria with aqueous solutions are also meta-stable equilibria, even if an apparent steady state has beenreached. It is possible that the dissolution equilibria maychange with the age of the gel but this has yet to beinvestigated.

4.1.2. Effect of temperature

The starting point for interpreting the temperaturedependence of gel dissolution is by comparison to calciumhydroxide for which solubility data as a function of tem-perature are available (e.g., Glasser et al., 1999).

There are three variables controlling the temperaturedependence of the solution chemistry in the calciumhydroxide system:

� the solubility of the calcium hydroxide solid phase;� the dissociation constant for water, Kw;� the degree of dissociation of Ca(OH)+ ions.

The solubility of calcium hydroxide is known todecrease with increasing temperature. Combined with anincrease in Kw (e.g., from 13.995 at 25 �C to 13.405 at45 �C) (Lide and Frederikse, 1994) and a slight reductionin the Ca(OH)+ ion dissociation constant, this leads to asignificant reduction in solution pH when measured atthe equilibration temperature of 45 �C.

In the case of the gels, the underlying solubility will bedependent on the nature of the C–S–H phase or phasespresent. The gel phases will be determined by the C/Sand the extent of ageing (crystallinity) of the gel. In thepresence of added salt, the process, or processes, that areoperating to increase the dissolved calcium concentrationsmay also be temperature dependent. The dissociation ofthe silicate species, for example:

H3SiO4� ¼ H2SiO2�4 þHþ

may also affect the solution phase pH. However, given thelow concentration of silicate species in solution (except atlow C/S), these are not expected to affect the solution pHsignificantly.

Qualitatively, both sets of experiments undertaken withC–S–H-gels equilibrated at 45 �C show reductions in cal-cium concentration compared to those at 25 �C, althoughit should also be noted that the magnitude of the changesat a given sodium chloride concentration are comparablewith the measurement errors. In the case of C/S = 1.5 thepH values also reduced with increasing temperature. Thisis similar to the behaviour observed with calcium hydrox-

J. Hill et al. / Waste Management 26 (2006) 758–768 767

ide. The result of pH measurements for the samples equil-ibrated at 45 �C and cooled to 25 �C before measuring weresimilar to those equilibrated and measured at 25 �C. In theabsence of a solid phase, and given that the solubility of theC–S–H gel increases at lower temperature (so that thereshould be no loss of calcium from solution), the observedincrease in pH on cooling may be accounted for by the tem-perature dependence of Kw and re-equilibration of the iondissociation reactions.

For the experiments undertaken with C/S = 1.0 gels, thesampling and pH measurements were conducted at roomtemperature. The values were almost identical for thosemeasured in the experiments conducted and measured at25 �C as found for the cooled C/S = 1.5 samples. However,given that the gels appear to equilibrate rapidly to changesin salt concentration, there is an uncertainty concerning theextent to which the experiments may have re-equilibrated(i.e., more calcium dissolved) at ambient temperature priorto sampling.

4.1.3. Effect of cation type

The three cations studied appear to have different effectson the extent to which they cause a decrease in pH. Forboth C/S = 1.0 and 1.8 the trend is: K > Na > Li (Fig. 5).This may have been due to the so-called sodium or salterror which occurs with measurements using glass elec-trodes at high pH values. In strongly alkaline solutionsthe electrode will respond to sodium and lithium ions aswell as to hydrogen ions with the pH reading being biasedtowards lower values (Midgley, undated).

To confirm that the measurements had been affected inthis way, a series of pH measurements were made on satu-rated calcium hydroxide solutions containing three differ-ent concentrations (0.1, 0.5 and 1.0 mol dm�3) of lithium,sodium or potassium chloride. The pH value of the calciumhydroxide solution was unaffected by the presence of up to1.0 mol dm�3 potassium chloride or at salt concentrationsof 0.1 mol dm�3. However, at higher concentrations ofeither lithium or sodium ions, the measured pH decreasedwith increasing salt concentrations with trends similar tothose observed for the C–S–H gels (Fig. 9). It should benoted, however, that the effect is pH-dependent and thebias would be reduced at lower pH.

The experiments with different salts were undertaken toinvestigate the hypothesis that a cation-exchange type ofprocess occurred between the calcium ions in the gels andalkali metal ions in solution. This process was invoked toexplain the observed increases in calcium dissolution in sal-ine waters for low C/S gels observed in the initial study. If acation-type exchange process was operating, some sensitiv-ity to cation type may be expected, which may result in dif-ferences in the dissolution behaviour in the presence ofdifferent cations. In contrast, if the observed trends resultedfrom ionic strength effects alone, the cation would beexpected to exert only a negligible effect of the gel dissolu-tion behaviour.

Equilibration of the C/S = 1.0 gel with different chlo-ride salts led to similar increases in dissolved calcium con-centrations with increasing ionic strength. There appearto be some slight differences between the data for the dif-ferent cations. While greater than the combined experi-mental uncertainties, these differences are not sufficientlylarge to be definitive. In general, they are less than twotimes the combined uncertainties Therefore, the resultsappear to be consistent with the operation of an ionexchange type process although an ionic strength effectcannot be ruled out due to some slight differences in datafor different cations even though these do not appear tobe significant.

The inclusion of an ion-exchange process was found notto be necessary to explain the experimental data for gels ofC/S greater than 1.5 in the presence of NaCl, where thesolution chemistry is expected to be dominated by equili-bration with a calcium hydroxide-type solid. The anoma-lous results for lithium observed with the gel of C/S = 1.8suggest that the situation may be more complicated andthat there may be additional processes contributing to theincreased calcium dissolution from C–S–H gels withincreased ionic strength. Additional data would be requiredto gain further insight into the effect of cation type on geldissolution behaviour.

In a recent paper Chen et al. (2004) summarised the cur-rent understanding of the structure of the CaO–SiO2–H2Osystem. As noted previously, the reaction of CaO and SiO2

at room temperature tends to lead to the formation of animperfect version of 14 A tobermorite (C–S–H(I)). TheC/S of 14 A tobermorite (Ca5Si6O16(OH)2 Æ 8H2O) is 0.83,although C–S–H(I) can accommodate substantial concen-trations of defects that allow variations of C/S rangingfrom 0.67 to 1.5. Spectroscopy studies indicate that Si–OH groups are present in C–S–H gels and that theydecrease with increasing C/S (Cong and Kirkpatrick,1996; Yu et al., 1999), reaching zero at C/S � 1.2–1.3(Yu et al., 1999). The results suggest that with increasingC/S, calcium ions replace protons in the Si–OH groups.In fact, this is one of the mechanisms by which the C/Sof the gel may vary at low ratios. Substitution of the cal-cium ions attached to the Si–O� groups by sodium ionsmay provide an explanation for the ion-exchange processpostulated for low C/S gels in saline waters.

5. Conclusions

� Gels with a C/S of 1.0 equilibrate rapidly in response tovariations in added salt concentration and, therefore, a30-day equilibration period was appropriate for theimmature C–S–H gels used in this study.� NaCl has been shown to have a significant effect on dis-

solution behaviour.� Equilibration of the gels in this study was fast a 30-day

equilibration time was shown to be sufficient.

768 J. Hill et al. / Waste Management 26 (2006) 758–768

� The effect of increasing temperature was to decrease dis-solved calcium concentration and pH, for C–S–H gels ofC/S = 1.0 and 1.5.� Equilibration of C/S = 1.0 gel with different salts led to

similar increases in dissolved calcium concentrationswith increasing ionic strength although there appearedto be some small differences between the effects of thedifferent cations. This may, however, not be significant.As a result, even though the results appear to be consis-tent with the operation of an ion exchange-type processbetween alkali metal ions in solution and calcium ions inlow C/S gels, alternative processes cannot be ruled out.� For C/S = 1.8 gels equilibrated with sodium or potassium

salts, dissolved calcium concentrations were similar butanomalously high calcium concentrations were observedfor the gel in the presence of lithium at higher ionicstrengths. This suggests that additional processes mightbe occurring in the presence of lithium at higher C/S gels.� Further work is necessary to explain the unexpected dif-

ferences observed.

Acknowledgements

Janice Haines, Charlotte Byrne and Andrew Jones ofHarwell Scientifics Ltd. are thanked for undertaking thesolution analyses. The authors thank Steve Williams andMark Cowper of Serco Assurance for bringing the salt ef-fect on pH measurements to our attention and for pH mea-surements of calcium hydroxide solutions in the presence ofdifferent cations, respectively.

References

Atkinson, A., 1985. The time dependence of pH within a repository forradioactive waste disposal. UKAEA Report AERE-R11777, UKAEA,UK.

Atkinson, A., Everitt, N.M., Guppy, R., 1987. Evolution of pH in aRadwaste repository: experimental simulation of cement leaching.UKAEA Report AERE-12594, UKAEA, UK.

Atkinson, A., Harris, A.W., Hearne, J.A., 1995. Hydrothermal alterationand ageing of synthetic calcium silicate hydrate gels. Nirex ReportNSS/R374. The Nirex report can be obtained by emailing info@nirex.co.uk.

Brown, P.W., Franz, E., Frohnsdorff, G., Taylor, H.F.W., 1984. Analysesof the aqueous phase during early C3S hydration. Cem. Concr. Res. 14,257–262.

Chambers, A.V., Heath, T.G., Hunter, F.M.I., Manning, M.C., Williams,S.J., 2005. The effect of sodium chloride on the dissolution of calciumsilicate hydrate gels. Report SA/ENV-063. The Nirex report can beobtained by emailing info@nirex.co.uk.

Chen, J.J., Thomas, J.J., Taylor, H.F.W., Jennings, H.M., 2004. Solubilityand structure of calcium silicate hydrate. Cem. Concr. Res. 34, 1499–1519.

Cong, X., Kirkpatrick, R.J., 1996. 17O MAS NMR investigations of thestructure of calcium silicate hydrate gel. J. Am. Chem. Soc. 79, 1585–1592.

Flint, E.P., Wells, L.S., 1934. Study of the system CaO–SiO2–H2O at30 �C and the reaction of water on the anhydrous calcium silicates. J.Natl. Bur. Stand. 12, 751–783.

Francis, A.J., Cather, R., Crossland, I., 1995. Development of the Nirexreference vault backfill. Report on Current Status 1994; Nirex ReportS/95/011. The Nirex report can be obtained by emailinginfo@nirex.co.uk.

Fujii, J., Kondo, W., 1981. J. Chem. Soc. Trans. 2, 645–651.Glasser, F.P., Angus, M.J., McCulloch, C.E., MacPhee, D., Rahman,

A.A., 1985. The chemical environment in cements. In: Jantzen, C.M.,Ewing, J.A., Ewing, R.C. (Eds.), Scientific Basis for Nuclear WasteManagement VIII, MRS Symposia Proceedings, vol. 44. MaterialsResearch Society, Pittsburgh, PA, pp. 849–858.

Glasser, F.P., Tyrer, M., Quillin, K., Ross, D., Pedersen, J., Goldthorpe,K., Bennett, D., Atkins, M., 1999. The chemistry of blended cementsand backfills intended for use in radioactive waste disposal. Environ-ment Agency Technical Report, P98.

Greenberg, S.A., Chang, T.N., 1965. Investigation of the colloidalhydrated calcium silicates. J. Phys. Chem. 69, 182–188.

Grutzeck, M., Benesi, A., Fanning, B., 1989. 29Si magic angle spinningNMR study of calcium silicate hydrates. J. Am. Ceram. Soc. 72, 665–668.

Harris, A.W., 1998. An assessment of the pH buffering provided by theNirex reference vault backfill within a radioactive waste repository.Nirex Report NSS/R323. The Nirex report can be obtained byemailing info@nirex.co.uk.

Kalousek, G.L., 1952. In: Proceedings of the Third InternationalSymposium on Cement Chemistry. Cement and Concrete Association,London, pp. 296–311.

Lide, E.D.R., Frederikse, H.P.R., 1994. CRC Handbook of Chemistryand Physics, 75th ed. CRC Press, London, pp. 8–57.

McConnell, J.D.C., 1955. The hydration of larnite (b-Ca3SiO4) andbredigite (a1-Ca2SiO4) and the properties of the resulting gelatinousmineral plombierite. Mineral. Mag. 30, 672–680.

Midgley, D.H., undated. Systematic fault-finding in the measurement ofpH. Lab Practice 37, pp. 65–71.

Milodowski, A.E., Nancarrow, P.H.A., Spiro, B., 1989. A mineralogicaland stable isotope study of natural analogues of ordinary Portlandcement (OPC) and CaO–SiO2–H2O (C–S–H) compounds. NirexReport NSS/R240. The Nirex report can be obtained by emailinginfo@nirex.co.uk.

Nirex, 2003. Generic repository studies – answers to public questions aboutthe phased disposal concept for radioactive waste. Nirex Report N/069.The Nirex report can be obtained by emailing info@nirex.co.uk.

Richardson, I.G., 1999. The nature of C–S–H in hardened cements. Cem.Concr. Res. 29, 1131–1147.

Roller, P.S., Ervin, G., 1940. The system calcium oxide–silica–water at30�. The association of silicate ion in dilute alkaline solution. J. Am.Chem. Soc. 62, 461–471.

Swanton, S.W., Fairbrother, H.J., Turner, N.A., 2005. The effects ofsodium chloride on the dissolution of calcium silicate hydrate gels. II.Effects of temperature and cation effects. Report S/ENV-0725. TheNirex report can be obtained by emailing info@nirex.co.uk.

Taylor, H.F.W., 1950. Hydrated calcium silicates, Part 1, compoundformation at ordinary temperatures. J. Chem. Soc., 3682–3690.

Taylor, H.F.W., 1986. Proposed structure for calcium silicate hydrate gel.J. Am. Chem. Soc. 9, 464–467.

Taylor, H.F.W., 1997. Cement Chemistry, second ed. Thomas Telford,London.

Walker, C.S., 2003. Characterisation and solubility behaviour of syntheticcalcium silicate hydrates. PhD Thesis, University of Bristol.

Yu, P., Kirkpatrick, R.J., Poe, B., McMillan, P.F., Cong, X., 1999.Structure of calcium silicate hydrate (C–S–H) near-, mid- and far-infra-red spectroscopy. J. Am. Chem. Soc. 82, 742–748.