Environ. Sci. Technol. 2008, 42, 403–409
Strontium Uptake by Cementitious Materials ERICH WIELAND,* JAN TITS, DOMINIK KUNZ, AND RAINER DÄHN Paul Scherrer Institut, Nuclear Energy and Safety Research Department, Laboratory for Waste Management, 5232 Villigen, Switzerland
Received May 24, 2007. Revised manuscript received September 18, 2007. Accepted September 24, 2007.
Wet chemistry experiments and X-ray absorption fine structure (XAFS) measurements were carried out to investigate the immobilization of nonradioactive Sr and 85Sr in calcite-free and calcite-containing Portland cement. The partitioning of pristine Sr between hardened cement paste (HCP) and pore solution, and the uptake of 85Sr and nonradioactive Sr were investigated in batch-type sorption/desorption experiments. Sr uptake by HCP was found to be fast and nearly linear for both cements, indicating that differences in the compositions of the two cements have no influence on Sr binding. The partitioning of pristine Sr bound in the cement matrix and 85Sr between HCP and pore solution could be modeled in terms of a reversible sorption process using similar Kd values. These findings allow 85Sr uptake to be interpreted in terms of an isotopic exchange process with pristine Sr. Sr K-edge EXAFS measurements on Sr doped HCP and calcium silicate hydrate (C-S-H) samples reveal no significant differences in the local coordination environments of pristine Sr and Sr bound to the cementmatrixuponsorption.Thefirstcoordinationsphereconsists of five to six oxygen atoms located at a distance of about 2.6 Å, which corresponds to Sr–O distances in the hydration sphere of Sr2+ in alkaline solution. Sr binds to the cement matrix via two bridging oxygen atoms located at a distance of about 3.6 Å. No further neighboring atoms could be detected, indicating that Sr is taken up as a partially hydrated species by HCP. Wet chemistry and spectroscopic data further indicate that Sr binding to C-S-H phases is likely to be the controlling uptake mechanism in the cement matrix, which allows Sr uptake by HCP to be predicted based on a Ca–Sr ion exchange model previously developed for Sr binding to C-S-H phases. The latter finding suggests that long-term predictions of Sr immobilization in the cementitious near field of repositories for radioactive waste can be based on a simplified sorption model with C-S-H phases.
of the engineered barrier systems in deep geological repositories in conjunction with disposal strategies for long-lived radioactive waste (2). Knowledge of the chemical mechanisms governing the interaction of metal cations and anionic species with the cement matrix, thus, is essential for predicting the long-term behavior of waste ions in these repositories. The cement matrix and radioactive wastes contain considerable strontium inventories. The concentration of pristine (nonradioactive) Sr in unhydrated cement was found to range between about one hundred and a few thousand parts per million (ppm) (3–5), depending on the initial Sr concentrations of the raw materials used for cement production. 90Sr is the most important radioactive Sr isotope found in operational waste from reactors and reprocessing plants. This radionuclide was considered to be safety relevant in the performance assessment for a proposed repository concept for long-lived intermediate level waste in Switzerland (6). In the past decades several studies have reported uptake measurements with radiostrontium on cementitious materials (e.g., 7–15). In all these studies the distribution coefficient (Kd) between crushed hardened cement paste (HCP) and highly alkaline cement pore waters was found to be lower on altered (Kd ∼10-3-10-2m3 kg-1) (7–11) than on unaltered cementitious materials (Kd ∼0.1 m3 kg-1) (13), suggesting that Ca may exert an influence on Sr uptake. Note that the Ca concentration is lower in the (Na,K)OH rich pore water ([Ca]t ∼2 mM, pH ∼13.3) of fresh cement than in a chemically altered system ([Ca]t ∼20 mM, pH ∼12.5). Sr binding in the complex cement matrix is still poorly understood. Sorption studies on single cement minerals indicate that Sr is taken up by calcium silicate hydrates (C-S-H) (12, 14, 15) and (Na+Al) substituted C-S-H (e.g., (16)). Strongest uptake, however, was reported under experimental conditions where AFt (Al2O3-Fe2O3-tri) phases are allowed to form (17) or in ettringite-containing systems (12, 18). Based on Sr leaching from degradation studies, it was further speculated that Sr uptake by cementitious systems might be controlled by a Sr-bearing solid of variable composition (solid solution) (19). This study reports results from wet chemistry experiments and X-ray absorption fine structure (XAFS) spectroscopy carried out to determine the binding mechanism of Sr in the cement matrix. In particular, isotopic exchange between 85Sr and pristine Sr of the cement matrix, reversibility of Sr binding and the nature of the uptake-controlling cement phases have been addressed. In earlier studies XAFS spectroscopy was used to determine the coordination environment of sorbed Sr species on mineral oxides (e.g., 20–23). Sr XAFS studies on cement systems have not been reported to date, although the technique can provide essential information on the cation and anion binding mechanisms in cementitious systems (e.g., 24–31).
Materials and Methods Introduction The safe disposal of cement-based hazardous and radioactive waste is of major importance in view of the need for sustainable waste treatment strategies. Stabilization and conditioning by cementitious materials has been recognized as technology well suited for the immobilization of cationic and anionic species present in these waste forms (1). Cementbased materials are foreseen to be used for the construction * Corresponding author phone: +41 56 310 2291; fax: + 41 56 310 3565; e-mail:
[email protected]. 10.1021/es071227y CCC: $40.75
Published on Web 12/07/2007
2008 American Chemical Society
Cementitious Materials and Pore Water Preparation. Two different sulfate resistant Portland cements (A, B), both denoted as HTS cement (CEM I 52.5 N HTS () Haute Teneur en Silice), Lafarge, France) were used for the preparation of the cement pastes according to a procedure described earlier (32). Determination of the chemical compositions of cements A and B as listed in Table S1 (Supporting Information) is described elsewhere (33). Cement A was manufactured 15 years ago and contains no calcite (see Figure S1, Supporting Information) (34). Cement B is a “modern” cement, which contains approximately 4–5 weight% (wt%) calcite. Thermodynamic modeling of the hydration process suggests that VOL. 42, NO. 2, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
403
the compositions of the hydrate assemblages of the two cements are slightly different (34). In HCP prepared from cement B monocarbonate is present instead of hydrogarnet, which is formed in cement A. Note, however, that the mineral compositions with respect to the dominant phases formed during hydration, i.e., portlandite, ettringite, and C-S-H, are similar (34). The Sr concentrations in the pastes of the two cements (denoted as pristine Sr) were determined to be 1290 and 1184 ppm, respectively. Crushed HCP materials were prepared from bulk samples by grinding them in a mortar under CO2-free conditions and sieving the crushed material (size fraction e63 µm). Determination of the ionic composition of the pore water in equilibrium with HCP and the basic preparation of the artificial cement water (ACW) used in this study is described elsewhere (33). ACW has the composition of cement pore water in the initial stage of the cement degradation (Table S2, Supporting Information). Calcium Silicate Hydrates (C-S-H). The Sr doped C-S-H phases used in the XAFS measurements were prepared using a procedure reported elsewhere (35). Preparation of the XAFS samples was carried out as described below. Wet Chemistry Experiments. Throughout this study, Fluka or Merck “pro analysis” chemicals and a high-purity deionized water (18.2 MΩ cm) generated by a Milli-Q Gradient A10 System (Millipore) were used. All experiments were carried out in duplicate samples in a glovebox under a N2 atmosphere (O2 and CO2 e 2 ppm) at ambient temperature (T ) 23 ( 3 °C). The experiments were carried out in 40 mL polyallomere centrifuge tubes, which were prewashed in 0.1 M HNO3 and thoroughly rinsed with deionized water before use. Labeled solutions for the uptake experiments were prepared using 85Sr (t1/2 ) 64.9 d) radiotracers (85Sr with carrier) purchased from Amersham International Plc. (UK) or Isotope Products Europe Blaseg GmbH (Germany). Radio assay was performed using a Canberra Packard Tricarb 2250 CA liquid scintillation analyzer (β counting). The 85Sr samples for radio assay were prepared by mixing 5 mL aliquots with 15 mL scintillator (Ultima Gold XR, Packard Bioscience S.A.) prior to counting. Measured counts were corrected for 85Sr decay in the time period between tracer addition to the samples and activity measurements. Batch-type uptake experiments with 85Sr on HCP were carried out as a function of time (kinetic tests) and at varying solid-to-liquid (S/L) ratios (2.5 × 10-3 to 0.13 kg L-1). Crushed HCP was weighed into the centrifuge tubes, or appropriate aliquots were withdrawn from a vigorously stirred stock suspension (0.05 kg L-1 crushed HCP in ACW) and pipetted into the tubes. The tubes were filled up to 40 mL by adding ACW. For both types of experiments the suspensions were pre-equilibrated on an end-over-end shaker for at least 3 days, and thereafter labeled with 85Sr ([85Sr]t ∼2 × 10-9 M). Samples were collected after regular time intervals in the kinetic tests. Sampling was carried out after 30 days equilibration in the experiment with varying S/L ratio. In both cases triplicate aliquots were withdrawn from the supernatant solution after centrifugation (60 min at 95000 g) and, together with standards, analyzed for 85Sr. Reversibility of 85Sr uptake was tested in desorption experiments using HCP prepared from cement A. Prior to desorption 85Sr was sorbed onto HCP (S/L ratio ) 2.5 × 10-2 kg L-1) for 1 and 30 days ([Sr]t )1.1 × 10-8 M). To start desorption the samples were centrifuged (60 min at 95000 g) and the supernatant solution replaced by fresh 85Sr free ACW. The activity of 85Sr released from HCP was determined after regular time intervals (kinetic tests) or after sequential replacement of ACW, i.e., centrifugation of the suspensions, replacement of the supernatant solution by fresh ACW free of 85Sr, and homogenization of the samples using a shear 404
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 2, 2008
mixer. After each replacement, the concentration of Sr (85Sr plus carrier) was determined using 85Sr radio assay. The kinetic tests showed that a time period of 7 days was sufficient to attain equilibrium in the desorption experiments (Figure S3, Supporting Information). The partitioning of nonradioactive (pristine) Sr between HCP and ACW was determined by varying the amount of HCP in contact with ACW (S/L ) 10-6 kg L-1 - 0.13 kg L-1) and from sorption isotherm measurements. HCP suspensions were prepared in manner similar to that described for the 85Sr uptake studies. For the sorption isotherm experiments appropriate volumes of 10-2 M and 5 × 10-2 M Sr stock solutions were added to the cement suspensions. The stock solutions were prepared by dissolving SrO in Milli-Q water. The cement suspensions were then equilibrated for 30 days and, in addition, 90 days in the case of sorption isotherm measurements. In both types of experiments aliquots were sampled from the supernatant solution after centrifugation and analyzed for Sr using an Applied Research Laboratory ARL 3410D inductively coupled plasma optical emission spectrometer (ICP-OES). Reversibility of the immobilization of pristine Sr bound in cement A was studied in leaching experiments. The procedure was similar to that employed in the desorption experiments with 85Sr, i.e., including sequential replacement of the Sr free ACW and re-equilibration for 7 days. The Sr concentrations in the supernatant solutions withdrawn after centrifugation were determined using ICP-OES. XAFS Sample Preparation. Sr XAFS measurements were performed on Sr doped C-S-H and HCP samples prepared from cement A. For the former, C-S-H phases with calciumto-silica (C/S) ratios ) 0.7 and 1.1 were synthesized in ACW as described elsewhere (35). After equilibrating the suspensions for 14 days, appropriate volumes of a Sr solution ([Sr]t ) 8 × 10-3 M prepared from SrO) were added to achieve Sr loadings of ∼1500 ppm. The Sr doped samples were then equilibrated for 30 days and centrifuged (20 min at 95000g). The cement samples were prepared from untreated (Cem_U) and treated HCP (Cem_HD, Cem_FD, Cem_LE). The sample denoted Cem_U consists of dry HCP material with ∼1300 ppm pristine Sr, which was not subjected to any treatment. The partially depleted HCP sample (Cem_HD) was prepared by carrying out one leaching cycle in ACW as described above to partially deplete the cement matrix of pristine Sr, and subsequently reloading with nonradioactive Sr. Taking into account the background concentration of Sr in ACW ([Sr]t ) 1.4 × 10-6 M) and by assuming Kd ) 0.1 m3 kg-1 it is estimated that the initial concentration of pristine Sr in HCP was reduced to about 600 ppm after a single leaching step, which corresponds to about half of the initial concentration. The partially depleted HCP sample was then reloaded by mixing the wet paste with a Sr/ACW solution (SrO dissolved in ACW, [Sr]t ) 2 × 10-3 M). The sample was equilibrated for 3 days, centrifuged, and the supernatant solution was replaced by a fresh Sr/ACW solution. The procedure was repeated three times to achieve a Sr surface concentration of ∼1200 ppm. The Cem_FD sample corresponds to HCP, which was completely depleted from pristine Sr (12 leaching cycles) and reloaded with Sr as described above to reach a Sr surface concentration of ∼1100 ppm. For the Cem_LE sample (leached cement), 1 g of HCP was mixed with 40 mL Milli-Q water, equilibrated for 3 days, centrifuged (20 min at 95000g), and the supernatant solution replaced by Milli-Q water. This procedure was repeated three times. The pH value of the supernatant solution after the last replacement was determined to be 12.5, indicating equilibrium with portlandite. The sample was then reloaded to reach a Sr surface concentration of ∼1000 ppm. All Sr-containing suspensions were centrifuged after equilibration for 7 days. The residual wet C-S-H and HCP pastes were packed into Plexiglas holders
FIGURE 1. Uptake kinetics of 85Sr by HCP in ACW (S/L ratio ) 2.5 × 10-2 kg L-1). HCP were prepared from the calcite-free cement A and the calcite-containing cement B. and sealed with Kapton tape. The holders were taken to the beamlines in a closed container filled with N2. SrO and Sr(OH)2 reference samples were prepared by diluting reagent grade chemicals with cellulose to achieve an X-ray absorption edge jump of 1. The Sr reference solution was prepared by dissolving SrO in Milli-Q water (pH 13, [Sr] ) 7 × 10-3 M). Sr K-edge (16.105 keV) XAFS Measurements. Absorption spectra were collected at cryogenic temperatures (77 K) to minimize the effect of thermal disorder (20). The samples were mounted on a coldfinger type sample holder and placed inside an airtight box flushed with N2 to avoid carbonation. The spectra of the Sr solution were collected at ambient (room temperature (RT) ∼300 K) and cryogenic temperature (CT). The monochromator energy was calibrated by assigning the first inflection point of the LIII-absorption edge of metallic lead foil to 13035 eV. Several scans were averaged to improve the signal-to-noise ratio (reference samples ∼3 scans; dilute samples ∼6 scans). The spectra of the reference compounds were recorded at BM1B (SNBL) at the European Synchrotron Radiation Facility (ESRF), Grenoble, France, using a channel cut crystal Si(111) monochromator in transmission mode (ionization chambers; I0: 90% N2/10% Ar; I1: 70% N2/30% Ar). XAFS spectra of the dilute Sr doped C-S-H and HCP samples ([Sr]sorbed ) 1000 - 1500 ppm) were collected at the X-11A beamline at the National Synchrotron Light Source (NSLS), Brookhaven National Laboratory, Uptown, USA, and at BM26A (DUBBLE) at ESRF both equipped with double crystal Si(111) monochromators. The measurements were carried out in fluorescence mode using a 13 element solid state detector (X-11A) or a nine channel monolithic Ge detector (BM26). Extended X-ray absorption fine structure (EXAFS) data reduction was performed by standard procedures (see the Supporting Information) using the WinXAS 97 2.3 software package (36).
Results and Discussion Uptake Studies with 85Sr on HCP. The results from the uptake studies with 85Sr are shown in Figure 1. The partitioning of 85Sr between HCP and ACW is expressed in terms of a distribution coefficient (Kd), which is defined as follows: Kd )
( ) ( cs,eq cl,eq
)
)( )
c0 - cl,eq Y × cl,eq mc
[m3 kg-1]
(1)
where Cs,eq is the equilibrium 85Sr (plus carrier) concentration on HCP [mol kg-1], Cl,eq is the equilibrium 85Sr (plus carrier) concentration in ACW [mol m-3], c0 is the initial 85Sr (plus carrier) concentration in ACW [mol m-3], mc, and V is the mass of HCP [kg] and volume of ACW [m3]. In this study Sr uptake is expressed in terms of Kd as a linear and reversible sorption-type process is involved (see
sections below). The difference between the initial 85Sr concentration (c0) and the concentration determined in the supernatant solution at equilibrium (c1,eq) accounts for 85Sr sorbed on the solid. Note that wall sorption was found to be negligible. Figure 1 shows that for both cements equilibrium is attained within about two weeks, indicating slow uptake of Sr. This finding suggest that diffusion into the cement matrix could be involved in 85Sr binding. In particular, the C-S-H gel precipitates as clusters of nanoscale colloidal particles with an associated internal pore system, which cause slow accessibility of sorption sites. At present, however, the observed slow uptake of Sr cannot be explained conclusively. The mean Kd values are similar for both cements: Kd ) (0.08 ( 0.01) m3 kg-1 (cement A) and (0.11 ( 0.01) m3 kg-1 (cement B), respectively, which were determined from averaging the data after two days equilibration. Determination of the 85Sr uptake at increasing S/L ratio show that the sorption values are constant in the selected concentration range, thus corroborating the concept of a sorption-type uptake process (Figure S2, Supporting Information). These findings show that differences in the mineral compositions of the two HCP materials have no significant influence on Sr uptake, and further suggest that neither calcium monocarbonate in calcite-containing cement nor hydrogarnet in calcite-free cement are the uptake-controlling phases for Sr in the cement matrix. Immobilization of Nonradioactive Sr in HCP. The interaction of nonradioactive Sr (pristine Sr and Sr added) with both cements was determined at increasing cement concentrations (Figure 2a) and from sorption isotherm measurements (Figure 2b), respectively. Sr partitioning between HCP and ACW can be quantified based on mass balance considerations (Supporting Information). This approach holds in case of reversible Sr binding, which is supported by the reversibility tests. Partitioning of pristine Sr was predicted by assuming Kd ) 0.10 m3 kg-1 in case of cement A and Kd ) 0.12 m3 kg-1 in case of cement B (Figure 2a). Note that these values are consistent with those determined from the radiotracer experiments (Kd ) (0.08 ( 0.01) m3 kg-1 (cement A) and (0.11 ( 0.01) m3 kg-1 (cement B). Figure 2b shows that the Sr sorption isotherms of both cements agree very well, and further, are nearly linear in the given concentration range (slope ) 0.9 ( 0.1). These findings indicate that the extent of sorption of 85Sr and pristine Sr to HCP are comparable and corroborate previous conclusions that differences in the HCP compositions do not have a significant influence on Sr uptake by HCP. For this reason the more detailed experimental studies on the reversibility of Sr binding and the EXAFS measurements were performed using solely the calcite-free cement A. Reversibility of Sr Binding to HCP. Reversibility of 85Sr and nonradioactive (pristine) Sr binding was tested for cement A by carrying out sequential replacement of the supernatant solutions by 85Sr free ACW solution or fresh ACW with [Sr]0 ∼1.4 × 10-6 M, respectively, as described above. In the latter case a sequential leaching of pristine Sr from the cement matrix was achieved. It should be noted that the Sr concentration in ACW, which is caused by impurities present in the chemicals used for ACW preparation, is negligibly small compared to the Sr inventory in HCP. Tests of the desorption kinetics using cement A showed that 85Sr release is a fast process (Figure S3, Supporting Information). Equilibrium was established within about two days. Figure 3a and b show that experimental data and predictions of the solution concentrations as a function of solution replacements agree very well. Predictions are based on mass balance considerations after each replacement step using Kd ) 0.08 m3 kg-1 for 85Sr and Kd ) 0.12 m3 kg-1 for pristine Sr (Supporting Information). These values agree VOL. 42, NO. 2, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
405
FIGURE 2. (a) Sr concentration in solution shown as a function of the S/L ratio of the HCP suspensions after 30 days equilibration. The broken and dashed lines represent predictions made for the concentration of dissolved Sr based on mass balance calculations and using Kd ) 0.10 m3 kg-1 (cement A) and Kd ) 0.12 m3 kg-1 (cement B). Solid lines outline the total Sr inventory associated with the HCP and the background concentration in ACW, respectively. (b) Sorption isotherms of nonradioactive Sr on the HCP (S/L ratio ) 2.5 × 10-2 kg L-1; equilibration time ) 30 and 90 days). The shaded areas indicate the initial concentration of sorbed Sr and Sr in ACW in an equilibrated ACW/HCP system at S/L ratio ) 2.5 × 10-2 kg L-1. within the estimated uncertainties ((15%) with those obtained from the 85Sr uptake tests (Kd ) 0.08 m3 kg-1) and the investigations of the partitioning of pristine Sr between HCP and ACW (Kd ) 0.10 m3 kg-1). The good agreement between experimental and predicted Sr concentrations shows that both 85Sr binding and the binding of pristine Sr to the cement matrix are reversible processes. Isotopic Exchange of 85Sr with Pristine Sr in HCP. Similar sorption values and reversibility of the sorption processes suggest that radiostrontium can replace pristine Sr in the cement matrix via an isotopic exchange process, which can be described in terms of a distribution coefficient, R, as follows (37): R)
(ns,M ⁄ ns,M) Kd,M* ) (nl,M ⁄ nl,M) Kd,M
(2)
ns,M* and ns,M denote the mole numbers of 85Sr or pristine Sr, respectively, in HCP, and nl,M* and nl,M are the respective mole numbers in ACW. Kd,M* and Kd,M denote the distribution ratios of 85Sr or pristine Sr, respectively, i.e. Kd,M* ) (0.08 ( 0.01) m3 kg-1 and Kd,M ) (0.10 ( 0.01) m3 kg-1. The distribution coefficient, R, is estimated to be 0.8 ( 0.1 based on the above Kd values, indicating that the large portion of pristine Sr is accessible to isotopic exchange with 85Sr. It is to be noted that R should equal 1 if the total inventory of pristine Sr was accessible to isotopic exchange. The above finding suggests that about 80% of the inventory of pristine Sr is subjected 406
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 2, 2008
FIGURE 3. Stepwise desorption of 85Sr (a) and pristine Sr (b) from the HCP of cement A. The aqueous 85Sr concentration was calculated based on the distribution coefficient determined from the uptake kinetic experiment (Kd ) (0.08 ( 0.01) m3 kg-1). The Kd value for pristine Sr was obtained from a best fit to the data (Kd ) 0.12 m3 kg-1). Experimental conditions: S/L ratio ) 2.5 × 10-2 kg L-1; [85Sr ]t ) 1.1 × 10-8 M. to reversible sorption and isotopic exchange processes. Further, only a small portion of Sr is bound in the structure of cement minerals (clinker or secondary minerals), which was not accessible to isotopic exchange over the time scale of the experiments. XAFS Measurements. HCP prepared from cement A was depleted from pristine Sr and reloaded with Sr to discern changes in the coordination environment in the treated HCP samples (Cem_LE, Cem_HD, Cem_FD) in comparison with Sr bound in the untreated sample (Cem_U). Furthermore, the EXAFS spectra of the Sr-containing HCP samples are compared to those of the Sr doped C-S-H samples and reference spectra (Supporting Information). C-S-H and HCP Samples. The EXAFS spectra collected at CT and the corresponding radial structure functions (RSFs) of the C-S-H and HCP samples are shown in Figure 4a and b. Comparison of the EXAFS spectra of the Sr doped C-S-H and HCP samples with the reference spectra of SrO(s) and Sr(OH)2(s) (Figure S4, Supporting Information) or SrCO3(s) and SrSO4(s) (reported in ref 20), respectively, show that the formation of crystalline solids or any solid solution in these samples can be excluded. All samples show a first peak located at ∼2.05 Å (not corrected for phase shift) corresponding to neighboring oxygen atoms (Figure 4b). Taking into account the given uncertainties in the bond distance ((0.02 Å), all Sr–O distances determined for the Sr doped C-S-H and HCP samples are similar (2.57–2.60 Å) (Table 1). Further, variation in the coordination number is small (N ) 6.1 – 7.6). The Sr-O bond distances of the first coordination shell are comparable to those determined for the aqueous Sr species and in Sr(OH)2(s) (Table 1). The number of first shell oxygen atoms is comparable to that in Sr(OH)2(s) (N ∼ 6), but lower than for the aqueous Sr species (N ∼ 9). These results suggest that the coordination number of the first oxygen shell is reduced when Sr is bound in solids or to C-S-H or HCP, respectively.
FIGURE 4. Sr EXAFS data (a) Experimental (solid line) and modeled (dashed line) EXAFS spectra of the Sr doped C-S-H and HCP samples. (b) Experimental (solid line) and theoretical Fourier transforms (open circles) obtained from the EXAFS spectra of the C-S-H (C-S-H 0.7, C-S-H 1.1) and HCP samples (Cem_LE, Cem_FD, Cem_HD, Cem_U). RSFs are not corrected for phase shift.
TABLE 1. Structural Information Obtained from Sr K-Edge Data Analysisa sample Sr2+
Sr–O Sr–O Sr–O Sr–Sr Sr–O Sr–Sr Sr–O Sr–O Sr–O Sr–Si Sr–O Sr–O Sr–Si Sr–O Sr–O Sr–Si Sr–O Sr–O Sr–O Sr–Si
aq
Sr2+aq RT Sr(OH)2 SrO C-S-H 0.7 C-S-H 1.1 Cem_U Cem_HD Cem_FD Cem_LE
N
R (Å)
σ2 (Å2)
∆E0 (eV)
res (%)
9.7 8.4 6.2 5.4 6.3 12.3 6.3 7.6 1.9 2.0 6.1 2.0 0.7 6.4 1.9 0.7 6.4 6.8 1.8 0.8
2.62 2.59 2.58 4.00 2.54 3.97 2.60 2.61 3.68 4.33 2.57 3.59 4.22 2.58 3.61 4.23 2.59 2.59 3.61 4.23
0.010 0.011 0.011 0.011 0.010 0.014 0.008 0.009 0.011 0.011 0.012 0.009 0.009 0.011 0.007 0.007 0.010 0.011 0.007 0.007
-6.7 -1.1 -2.7 -2.7 -4.1 -4.1 0.6 -0.3 -0.3 -0.3 -0.2 -0.2 -0.2 0.5 0.5 0.5 0.0 0.8 0.8 0.8
7.7 2.7 8.7 8.7 7.2 7.2 6.4 6.6 6.6 6.6 10.7 10.7 10.7 8.4 8.4 8.4 7.8 8.3 8.3 8.3
a All data are at cryogenic temperature unless indicated (RT: room temperature). N is the coordination number, R denotes bond distance (Å), σ is the Debye–Waller factor (Å), and ∆E0 the energy shift (eV). The residual factor ((%res )
n
n
(( Σ |yexp(i) - ytheo(i)|)⁄( Σ yexp(i)))) represents the quality of the fit (%) with yexp and ytheo as the experimental and theoretical i)1 i)1 data points.
Binding Mechanism. Backscattering contributions of further shells above the noise level appear in the RSFs of the C-S-H 1.1, Cem_LE, Cem_HD, and Cem-U samples. The quality of the EXAFS spectrum for the Cem_FD samples is insufficient to allow further shells to be considered in the fitting. No significant backscattering contributions from more distant shells are detected in the C-S-H 0.7 sample. In case of the C-S-H 1.1, Cem_LE, Cem_HD, and Cem-U samples good agreement between experimental and modeled oscillations was achieved by taking into account neighboring oxygen (∼2 O at 3.59–3.68 Å) and silica atoms (1 - 2 Si at 4.22–4.33 Å) (Table 1). The fact that the same structural models could be applied to mimic the EXAFS data of Sr doped C-S-H and HCP samples and that very similar structural parameters resulted for Sr taken up by the two materials
implies that C-S-H is the solid phase controlling Sr uptake in the cement matrix. The EXAFS data indicate that Sr species taken up by C-S-H and the cement matrix are bound via bridging oxygen atoms to surface sites of C-S-H. The presence of two second shell oxygen atoms at ∼3.6–3.7 Å suggests that the Sr species immobilized in the cement matrix is not only surrounded by H2O molecules in the first coordination sphere. The reduction in the latter compared to the aqueous species seems to be partially compensated by oxygen atoms associated with the second coordination sphere. Modeling Sr Uptake by HCP in Terms of Sr–Ca Exchange on C-S-H. Wet chemistry and EXAFS data support the idea that C-S-H phases are the uptake-controlling cement phase for Sr in the complex cement matrix. In a recent study it was proposed that Sr binding to C-S-H phases in alkali-free VOL. 42, NO. 2, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
407
solutions can be described in terms of a Sr–Ca ion exchange model (35). In the framework of the present study the existing model was extended to account for Na/K–Sr exchange processes in the presence of the high concentrations of alkalis in ACW (Supporting Information). The following exchange reactions can be written in line with the earlier study (35): Ca–C-S-H(s) + Sr2+(aq) ) Sr–C-S-H(s) + Ca2+(aq) (3) Na2–C-S-H(s) + Sr2+(aq) ) Sr–C-S-H(s) + 2Na+(aq) (4) K2–C-S-H(s) + Sr2+(aq) ) Sr–C-S-H(s) + 2K+(aq)
(5)
The Kd value for Sr uptake by HCP in ACW can be estimated based on the extended ion exchange model (Supporting Information). The underlying assumptions are (i) Sr binding to HCP is due to ion exchange reactions with Ca (and Na, K in fresh HCP) on the edge and planar silanol groups of the C-S-H phases (35), (ii) the cation exchange capacity (CEC) of C-S-H phases is independent of the C/S ratio (35), (iii) all silanol sites of C-S-H are deprotonated under the alkaline conditions prevailing in cement systems (35), (iv) the negative charge carried by deprotonated silanol groups is neutralized by Ca, Sr or in addition by Na and K in ACW, (v) Na and K have the same affinity for silanol groups. The Kd value for Sr uptake by HCP is estimated to be 0.14 m3 kg-1 using the proposed C-S-H-based ion exchange model (Supporting Information). The predicted value is in good agreement with the experimental data (Kd ∼ 0.08 – 0.12 m3 kg-1) in view of the simplifying assumptions that have been made. This finding suggests that long-term predictions of Sr immobilization in the cementitious near field of repositories for radioactive waste can be based on a simplified sorption model with C-S-H.
Acknowledgments Experimental assistance from the staff of the Swiss-Norwegian beamline (SNBL), the Dutch-Belgium beamline (DUBBLE) at the ESRF and X-11A at the NSLS is gratefully acknowledged. We thank J.-P. Dobler (PSI) for his contribution to the radioanalytical work and S. Köchli (PSI) for the ICP-OES measurements. Thanks are extended to Dr. I. Bonhoure for assistance during the EXAFS measurements. Dr. M. Bradbury (PSI) is thanked for the many valuable discussions and helpful comments on this work. Gratitude is expressed to I. Hagenlocher (Nagra) for her continuous interest in the research project and for reviewing the manuscript. Partial financial support was provided by the National Cooperative for the Disposal of Radioactive Waste (Nagra), Switzerland.
Supporting Information Available Tables S1 and S2 show the compositions of the cements and porewater used in the experiments. Figures S1–S5 show the results from thermogravimetric analysis of the cements, uptake of radiostrontium at increasing S/L ratio, desorption kinetics of radiostrontium, EXAFS data of Sr reference compounds and Sr uptake by C-S-H. In addition, mass balance considerations of Sr partitioning and extension of the ion exchange model for predicting Sr uptake by C-S-H in HCP are included. The material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Spence, R. Chemistry and Microstructure of Solidified Waste Forms; Lewis Publishers: Boca Raton, 1992. (2) Chapman, N.; McCombie, C. Principles and Standards for the Disposal of Long-Lived Radioactive Wastes; Pergamon: Amsterdam, 2003. (3) Duchesne, J.; Reardon, E. J. Determining controls on element concentrations in cement kiln dust leachate. Waste Manage. 1998, 18, 339–350. 408
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 2, 2008
(4) Tamàs, F. D.; Abonyi, J. Trace elements in clinker. I. A graphical representation. Cem. Concr. Res. 2002, 32, 1319–1323. (5) Graham, I. J.; Goguel, R. L.; St John, D. A. Use of strontium isotopes to determine the origin of cement in concretes. Case examples from New Zealand. Cem. Concr. Res. 2000, 30, 1105– 1111. (6) Nagra. Project Opalinus Clay––Safety Report, Nagra Technical Report NTB 02–05; Nagra: Wettingen, Switzerland, 2002. (7) Allard, B.; Persson, G.; Torstenfeldt, B. Radionuclide Sorption on Concrete, Nagra Technical Report NTB 85–21; Nagra: Wettingen, Switzerland, 1985. (8) Hietanen, R.; Alaluusua, M.; Jaakola, T. Sorption of Cesium, Strontium, Iodine, Nickel and Carbon in Mixtures of Concrete, Crushed Rock and Bitumen, Report YJT-84–04; Nuclear Waste Commission of Finnish Nuclear Power Companies: Helsinki, Finland, 1985. (9) Ewart, F. T.; Terry, S.; Williams, S. J. Near Field Sorption Data for Cesium and Strontium, AERE report M 345; Harwell Laboratory: Oxfordshire, UK, 1985. (10) Atkinson, A.; Nickerson, A. K. Diffusion and sorption of cesium, strontium, and iodine in water-saturated cement. Nucl. Technol. 1988, 81, 100–113. (11) Sugiyama, D.; Fujita, T. Sorption of radionuclides onto cement materials altered by hydrothermal reactions. Mater. Res. Soc. Symp. Proc. 1999, 556, 1123–1130. (12) Atkins, M.; Glasser, F. P. Application of Portland cement-based materials to radioactive waste immobilization. Waste Manage. 1992, 12, 105–131. (13) Wieland, E.; Tits, J.; Spieler, P.; Dobler, J. P.; Scheidegger, A. M. Uptake of nickel and strontium by a sulphate-resisting portland cement. In Applied Mineralogy; Rammlmair, D., Mederer, J., Oberthür, T., Heimann, R. B., Pentinghaus, H., Eds.; Balkema: Rotterdam 2000; Vol. 2. (14) Iwaida, T.; Nagasaki, S.; Tanaka, S. Sorption study of strontium onto hydrated cement phases using a sequential desorption method. Radiochim. Acta 2000, 88, 483–486. (15) Tits, J.; Wieland, E.; Dobler, J. P.; Kunz, D. The uptake of strontium by calcium silicate hydrates under high pH conditions: An experimental approach to distinguish adsorption from coprecipitation processes. Mater. Res. Soc. Symp. Proc. 2004, 807, 689–694. (16) Shrivastava, O. P.; Shrivastava, R. Sr2+ uptake and leachability study on cured aluminium-substituted tobermorite-OPC admixtures. Cem. Concr. Res. 2001, 31, 1251–1255. (17) Lameille, J. M.; Goutiere, G.; Petit, J. C.; Regourd, M. Retention of cobalt, cesium, and strontium in the hydrates of C3S, C3A, and gypsum. J. Am. Ceram. Soc. 1987, 70, 604–614. (18) Gougar, M. L. D.; Sheetz, B. E.; Roy, D. M. Ettringite and C-S-H Portland cement phases for waste ion immobilization: A review. Waste Manage. 1996, 16, 295–303. (19) Berner, U. Geochemical modelling of repository systems: Limitations of the thermodynamic approach. Radiochim. Acta 1998, 82, 423–428. (20) O’Day, P.; Newville, M.; Neuhoff, P. S.; Sahai, N.; Caroll, S. A. X-ray absorption spectroscopy of strontium(II) coordination. I. Static and thermal disorder in crystalline, hydrated, and precipitated solids in aqueous solution. J. Colloid Interface Sci. 2000, 222, 184–197. (21) Sahai, N.; Caroll, S. A.; Roberts, S.; O’Day, P. A. X-ray absorption spectroscopy of strontium(II) coordination. II. Sorption and precipitation at kaolinite, amorphous silica, and goethite surfaces. J. Colloid Interface Sci. 2000, 222, 198–212. (22) Duff, M. C.; Hunter, D. B.; Hobbs, D. T.; Fink, S. D.; Dai, Z.; Bradley, J. P. Mechanism of strontium and uranium removal from high-level radioactive waste simulant solutions by the sorbent monosodium titanate. Environ. Sci. Technol. 2004, 38, 5201–5207. (23) Choi, S.; O’Day, P.; Rivera, N. A.; Mueller, K.; Vairavamurthy, M. A.; Seraphin, S.; Chorover, J. Strontium speciation during reaction of kaolinite with simulated tank-waste leachate: Bulk and microfocused EXAFS analysis. Environ. Sci. Technol. 2006, 40, 2608–2614. (24) Allen, P. G.; Siemering, G. S.; Shuh, D. K.; Bucher, J. J.; Edelstein, N. M.; Langton, C. A.; Clark, S. B.; Reich, T.; Denecke, M. A. Technetium speciation in cement waste forms determined by X-ray absorption fine structure spectroscopy. Radiochim. Acta 1997, 76, 77–86. (25) Rose, J.; Moulin, I.; Hazemann, J.-L.; Masion, A.; Bertsch, P. M.; Bottero, J.-Y.; Mosnier, F.; Haehnel, C. X-ray absorption spectroscopy study of immobilization processes for heavy metals in calcium silicate hydrates. I. Case of lead. Langmuir 2000, 16, 9900–9906.
(26) Scheidegger, A. M.; Wieland, E.; Scheinost, A. C.; Dähn, R.; Spieler, P. Spectroscopic evidence for the formation of layered Ni-Al double hydroxides in cement. Environ. Sci. Technol. 2000, 34, 4545–4548. (27) Pomiés, M.-P.; Lequeux, N.; Boch, P. Speciation of cadmium in cement. I. Cd2+ uptake by C-S-H. Cem. Concr. Res. 2001, 31, 563–569. (28) Ziegler, F.; Scheidegger, A. M.; Johnson, A. C.; Dähn, R.; Wieland, E. Sorption mechanism of zinc to calcium silicate hydrate: X-ray absorption fine structure (XAFS) investigation. Environ. Sci. Technol. 2001, 35, 1550–1555. (29) Bonhoure, I.; Scheidegger, A. M.; Wieland, E.; Dähn, R. Iodine species uptake by cement and CSH studied by I K-edge X-ray absorption spectroscopy. Radiochim. Acta 2002, 90, 647–651. (30) Bonhoure, I.; Wieland, E.; Scheidegger, A. M.; Ochs, M.; Kunz, D. EXAFS study of Sn(IV) immobilization by hardened cement paste and calcium silicate hydrates. Environ. Sci. Technol. 2003, 37, 2184–2191. (31) Schlegel, M.; Pointeau, I.; Coreau, N.; Reiller, P. Mechanism of europium retention by calcium silicate hydrates: An EXAFS study. Environ. Sci. Technol. 2004, 38, 4423–4431.
(32) Sarrot, F.-A.; Bradbury, M. H.; Pandolfo, P.; Spieler, P. Diffusion and adsorption studies on hardened cement paste and the effect of carbonation on diffusion rates. Cem. Concr. Res. 1992, 22, 439–445. (33) Wieland, E.; Tits, J.; Ulrich, A.; Bradbury, M. H. Experimental evidence for the solubility limitation of the aqueous Ni(II) concentration and isotopic exchange of63Ni in cementitious systems. Radiochim. Acta 2006, 94, 29–36. (34) Lothenbach, B.; Wieland, E. A thermodynamic approach to the hydration of sulphate-resisting Portland cement. Waste Manage. 2006, 26, 706–719. (35) Tits, J.; Wieland, E.; Müller, C. J.; Landesman, C.; Bradbury, M. H. Strontium binding by calcium silicate hydrates. J. Colloid Interface Sci. 2006, 300, 78–87. (36) Ressler, T. WinXAS: A program for X-ray absorption spectroscopy data analysis under MS-Windows. J. Synchrotron Rad. 1998, 5, 118–122. (37) Lieser, K. H. Nuclear and Radiochemistry, 2nd ed.; Wiley-VCH: Weinheim, Germany, 2001.
ES071227Y
VOL. 42, NO. 2, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
409
