Research Communications Sorption of Isosaccharinic Acid, a Cellulose Degradation Product, on Cement L. R. VAN LOON,* M. A. GLAUS, S. STALLONE, AND A. LAUBE Laboratory for Waste Management, Paul Scherrer Institute, Wu ¨ renlingen and Villigen, CH-5232 Villigen, PSI, Switzerland
Introduction Low- and intermediate-level radioactive waste (L/ILW) contains substantial amounts of cellulosic materials. In Switzerland ∼50% of the organic waste is cellulosic (1). In a cementitious L/ILW repository, the large amounts of cement produce alkaline environments in which the pH of the cement pore water remains above 12.5 for periods of the order of 105 years (2). It is well known in the literature that cellulose is unstable under alkaline conditions and will degrade to watersoluble, low molecular weight compounds (3). The type of degradation products formed depends strongly on the composition of the solution in contact with the cellulose. In the presence of Ca2+, as is the case for pore waters of cementstabilized waste forms, isosaccharinic acid (ISA, see Figure 1) is the main degradation product (4, 5). The importance of ISA in the context of radwaste disposal is that it can formsin analogy with gluconic acid (see Figure 1)sstable complexes with tri- and tetravalent radionuclides such as Am3+ and Pu4+ (6-8), which adversely influences their sorption on the cement phase (9). This could lead to an enhanced release of such radionuclides to the geo- and biosphere. A key factor for complexation is the concentration of ISA in the cement pore water. The concentration of ISA in the pore water is essentially governed by the quantity of the cellulose present (cellulose loading), its degradation rate, and the cement porosity. Conservative estimates indicate that concentrations up to 0.1 M can be achieved, which would have severe consequences for the sorption of radionuclides (9). A few processes, however, can reduce the concentration of ISA in the pore water, i.e., the degradation of ISA, the formation of sparingly soluble ISA compounds, and/or the sorption of ISA on the cement. Stability experiments showed that ISA is stable under alkaline conditions, and consequently, a decrease of the ISA concentration by a degradation process is unlikely (10). The only sparingly soluble salt that can be formed in the repository and can potentially bound the upper limit of the ISA concentration is Ca(ISA)2. On the basis of the solubility product of Ca(ISA)2, a maximum concentration of ISA of ∼5 × 10-2 M could be calculated (10). The only process capable of significantly reducing the ISA concentration in the cement pore water is its sorption on the cement. No studies on the sorption of ISA are known in the literature. Consequently, we started an investigation on the sorption of ISA on cement under repository conditions, i.e., a (Na,K)OH solution saturated with respect to Ca(OH)2 (Portlandite) and having a pH value of 13.3. In this study, we focus on measuring * Corresponding author fax: 41-56-310 44 38; e-mail: Luc.
[email protected].
S0013-936X(96)00505-6 CCC: $14.00
1997 American Chemical Society
FIGURE 1. Chemical structure of r-isosaccharinic acid and gluconic acid. sorption isotherms for ISA at 25 °C and on determining the sorption capacity and the ISA-cement interaction constant.
Experimental Section Hardened Cement Paste. The sorbent used in the sorption experiments was a hardened cement paste (CPA 55 HTS cement, Lafarge, France). An overview of the chemical and mineral composition of the HTS cement is given in Table 1. After complete hydration (6 years), the hardened cement paste was crushed and sieved. The particle fraction between 100 and 400 µm was used in the sorption experiments. Preparation of r-ISA. Ca(ISA)2 was synthesized by contacting lactose with limewater as decribed in ref 11 and transformed into NaISA by treating the Ca(ISA)2 with an ion exchange resin in the Na+ form: 8 g of Ca(ISA)2 was mixed with 200 g of Chelex-100 (BioRad) in 1000 mL of demineralized water. After mixing the suspension for 3 h, the resin was filtered off with a membrane filter (Millipore 0.2 µm), and the filtrate was evaporated to a volume of about 100 mL. The solution was further evaporated in an oven at 80 °C till a thick syrup was obtained. The syrup was cooled and left standing at room temperature (20 °C) to crystallize. The obtained crystals were triturated in the presence of water-free diethyl ether to remove remaining water, filtered off, washed again with water-free diethyl ether, and finally dried under reduced pressure in a vacuum oven at 55 °C. Analysis showed that the Ca level in the product obtained was ∼0.01%. HPLC analysis on a Carbopac PA-100 column (Dionex, amperometric detection using a gold working electrode) showed the presence of two peaks in the NaISA salt. The peak area of ISA contributed to 98% of the total peak area. Artificial Cement Pore Water (APW). The composition of the artificial cement pore water (denoted hereafter as APW) was taken from ref 2 and had the following composition: 114 mmol L-1 Na, 180 mmol L-1 K, 2 mmol L-1 Ca, and a pH of 13.4. One liter of demineralized water was flushed with argon for 0.5 h. Samples of 4.56 g of NaOH (Merck, 100%), 11.61 g of KOH (Merck, 87%), and 10 g of Ca(OH)2 (Merck, 100%) were added. All chemicals used were of analytical grade quality. The mixture was left standing for 24 h, after which time the liquid was filtered off through a 0.45 µm membrane filter (cellulose nitrate, Schleicher & Schuell ME 25, Germany).
VOL. 31, NO. 4, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1243
TABLE 1. Chemical and Mineralogical Composition of CPA 55 HTS Cement Used in Sorption Experiments oxide
wt %
oxide
wt %
CaO SiO2 Al2O3 Fe2O3 SO3
66.4 23.8 2.7 2.8 1.8
K2O Na2O MgO TiO2
0.26 0.09 0.88 0.12
C3S (alite) C2S (belite)
Clinker Compositiona 64.1 C3A (aluminate) 24.1 C4AF (ferrite)
0.5 7.6
a
C, CaO; S, SiO2; A, Al2O3; F, Fe2O3, e.g. C3S ) 3CaO‚SiO2.
All manipulations were carried out in a glovebox under a controlled N2 atmosphere (CO2, O2 < 5 ppm). After preparation, the composition of APW was checked by ICP-AES analysis and showed the following composition: 110.5 mmol L-1 Na, 176 mmol L-1 K, and 1.75 mmol L-1 Ca. The pH of the solution was 13.3. Sorption of ISA. The sorption of ISA on cement was studied by the batch technique. To cover a broad equilibrium concentration range of ISA, the sorption of ISA was carried out with different solid to liquid (S:L) ratios ranging from 1:2 (10 g of cement:20 mL of APW) to 1:40 (1 g of cement:40 mL of APW). The cement was placed in 50 mL polysulfonate centrifuge tubes (ISA was found not to sorb on these tubes), and the artificial cement pore water, containing ISA between 10-5 and 3 × 10-1 M, was added. The mixtures were shaken end-over-end for 1 day, 1 week, and a few systems for 1 month, 3 months, and 9 months. All manipulations were performed in a glovebox under a controlled N2 atmosphere. The suspensions were centrifuged at 27000g for 15 min (Beckman L7-35 Ultracentrifuge). Since a thin film of cementitious material remained on the top of the supernatant, solutions were additionaly filtered through a 0.45 µm membrane filter (Acrodisc, Gelman, U.S.) before analysis. The ISA in the equilibrium solution was measured by HPLC (Dionex DX 500) using a Carbopac PA-100 analytical column and an electrochemical detector (ED 40) in the supressed conductivity detection mode. The eluent used contained 0.08 M NaOH. Standards were made up from pure Ca(ISA)2 in demineralized water. The relative error on the HPLC measurements was at most 10%. The amount of ISA sorbed on the cement ([ISA]sorbed, mol kg-1) was calculated from the difference in concentration before and after sorption using the following equation:
((ISA)in - (ISA)eq)V [ISA]sorbed ) m
(mol kg ) (1) -1
where (ISA)in is the initial concentration of ISA (mol L-1), (ISA)eq is the equilibrium concentration of ISA (mol L-1), V is the volume of the liquid phase (L), and m is the amount of cement used (kg). The maximum relative error on the measured sorption values was ∼20%.
Results and Discussion Figure 2 shows the results of a preliminary experiment using a S:L ratio of 1:40. The decrease of ISA carbon [6∆(ISA) ) six times the difference between initial and equilibrium concentration of ISA] is plotted against the decrease of DOC [∆(DOC) ) difference between the initial and equilibrium concentration of total dissolved organic carbon]. As can be seen, the equilibrium concentration of ISA is smaller than the initial concentration [∆(ISA) > 0]. This can be interpreted either as a sorption process or as a chemical transformation reaction. The latter process, however, can be excluded because the concentration of DOC in the equilibrium solution decreases [∆(DOC) > 0] as well and the ratio between the
1244
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 4, 1997
FIGURE 2. Relationship between the decrease of ISA carbon in solution and the decrease of the total dissolved organic carbon (DOC). The symbols represent the experimental data. The solid line is a least squares fit with a slope of 0.94 ( 0.08 and a correlation coefficient of 0.96. The dashed line represents a plot with a slope of unity.
FIGURE 3. Sorption isotherm of r-ISA on Portland cement at pH 13.3. The symbols are the experimental data. The dashed line represents the best fit by a Langmuir isotherm with one adsorption site (q ) 0.17 ( 0.01 mol kg-1; K ) 286 ( 41 L mol-1). The solid line represents the best fit using a two-site model with Langmuir adsorption behavior (q1 ) 0.1 ( 0.01 mol kg-1; K1 ) 1730 ( 385 L mol-1; q2 ) 0.17 ( 0.02 mol kg-1; K2 ) 12 ( 4 L mol-1). decrease in ISA carbon and the decrease in DOC [6∆(ISA)/ ∆(DOC)] is close to unity (slope of the regression line is 0.94). Degradation of ISA would result in values of this ratio being much smaller than 1. A separate study on the stability of ISA in APW in the absence of cement showed that ISA is stable for months under such conditions and supports the exclusion of transformation of ISA in our sorption studies. The sorption of ISA on cement is a fast process almost reaching equilibrium after 1 day. Adsorption of ISA on the cement is not consistent with electrostatic concepts because of the repulsion between the negatively charged cement surface and the negatively charged ISA molecules (at pH 13.3 the carboxylic group of isosaccharinic acid is completely deprotonated). Strong specific interactions between ISA and the surface sites, however, can overcome this repulsion and can lead to sorption. Examples of the adsorption of anions by ligand exchange () specific interaction) on iron and aluminum oxides beyond the zero point of charge (at high pH) are known in the literature (12-14). The ISA adsorption data were plotted as a sorption isotherm (Figure 3). The simplest assumption that can be made in adsorption phenomena is that only one adsorption site is involved and that this site becomes reversibly saturated
with increasing concentration of adsorbate in the equilibrium solution. In such a case, the adsorption isotherm can be described by a Langmuir isotherm. For the ISA/cement system considered in this work, one can write
[ISA]sorbed )
Kq(ISA)eq 1 + K(ISA)eq
(2)
where q is the adsorption capacity of cement for ISA (mol kg-1), K is the adsorption affinity constant (L mol-1), and (ISA)eq is the equilibrium concentration of ISA (mol L-1). Close inspection of the data in Figure 3, however, showed that the experimental results cannot be described by a Langmuir type of isotherm with only one adsorption site. The introduction of a second site with a Langmuir adsorption behavior gives a much better fit of the experimental results as is shown in Figure 3:
[ISA]sorbed )
K1q1(ISA)eq 1 + K1(ISA)eq
K2q2(ISA)eq +
1 + K2(ISA)eq
(3)
where q1 and q2 represent the sorption capacities and K1 and K2 are the adsorption affinity constants for the two sites. The introduction of additional sorption sites would undoubtably give an even better fit to the experimental data. However, without more information on the solid phases responsible for the sorption and the underlying adsorption mechanism(s), we see no benefit from such an exercise. The total sorption capacity of the cement for ISA (q1 + q2) is ∼0.3 mol kg-1. This value is in good agreement with the value of 0.32 mol kg-1 proposed by Bradbury and Sarott (9). Attempts to fit the complete data set with a Freundlich type of isotherm failed. Cement has a quite complicated composition (see Table 1). The most important component within a hardened cement is the CSH gel (calcium silicate hydrate). A general composition of xCaO‚SiO2‚yH2O is assumed for these phases, where x lies in the range 1.4-1.8 for common Portland cements and y g x (2). It is well known that many organic products such as lignosulfonate, naphthalenesulfonate, etc. sorb on cement (15-21). This phenomenon found an application in the cement industry because the adsorption of organics changes the surface characteristics of the cement and consequently also its workability. It was shown that the sorption of lignosulfonate and naphthalenesulfonate on cement made the cement surface more negatively charged (17), resulting in modified rheological properties (fluidification). The adsorption mechanism of such organic molecules is still poorly understood. However, it was observed that especially the C3A and C4AF (see Table 1) phases and its hydration products showed a strong affinity for organic molecules. The adsorption was also found to be highly irreversible. It can be concluded that the observed sorption behavior of ISA on cement is very important for radwaste disposal since this sorption process drastically reduces the concentration of ISA in the pore water and consequently the adverse effect of ISA on radionuclide sorption. Although the adsorption behavior of ISA on cement could be satisfactorily described by a Langmuir isotherm considering two sorption sites, the underlying mechanism(s) is(are) still unknown, and further investigations in this field are necessary. From the
many possible adsorption mechanisms, specific interaction (chemical interaction) with the surface is the most likely one. A combination of surface analytical techniques such as reflection-absorption infrared spectroscopy (22, 23) and calorimetric measurements might give some valuable information on the mechanisms involved. In addition, adsorption of ISA on pure cement components will be undertaken in order to identify the different sorption sites involved.
Acknowledgments This work was partly financed by the Swiss National Cooperative for the Disposal of Radioactive Waste (Nagra). The authors would like to thank M. Bradbury for helpful comments during preparation of the manuscript.
Literature Cited (1) Endlager fu ¨ r schwach- und mittelaktive Abfa¨lle (Endlager SMA). Bericht zur Langzeitsicherheit des Endlagers SMA am Standort Wellenberg (Gemeinde Wolfenschiessen, NW); Nagra Technical Report NTB 94-06; Nagra: Wettingen, Switzerland, 1994. (2) Berner, U. A thermodynamic description of the evolution of pore water chemistry and uranium speciation during the degradation of cement; PSI Bericht 62; Paul Scherrer Institute: Villigen, Switzerland, 1990. (3) Whistler, R. L.; BeMiller, J. N. Adv. Carbohydr. Chem. Biochem. 1958, 13, 289. (4) Blears, M. J.; Machell, G.; Richards, G. N. Chem. Ind. 1957, August 24, 1150. (5) Machell, G.; Richards, G. N. J. Chem. Soc. A 1960, 1932. (6) Sawyer, D. T. Chem. Rev. 1964, 64, 633. (7) Moreton, A. D. Mater. Res. Soc. Symp. Proc. 1993, 294, 753. (8) Greenfield, B. F.; Holtom, G. F.; Hurdus, M. H.; O’Kelly, N.; Pilkington, N. J.; Rosevaer, A.; Spindler, M. W.; Williams, S. J. Mater. Res. Soc. Symp. Proc. 1995, 353, 1151. (9) Bradbury, M. H.; Sarott, F. A. Sorption Databases for the Cementitious Near-field of a L/ILW Repository for Performance Assessment; PSI-Bericht 95-06; Paul Scherrer Institute: Villigen, Switzerland, 1995. (10) Bradbury, M. H.; Van Loon, L. R. Cementitious Near-Field Sorption Databases for Performance Assessment of a L/ILW Repository in a Palfris Marl Host Rock (in preparation); Paul Scherrer Institute: Villigen, Switzerland, 1997. (11) Whistler, R. L.; BeMiller, J. N. In Methods in Carbohydrate Chemistry, Vol. 2: Reactions of Carbohydrates; Wolfrom, M. L., BeMiller, J. N., Eds.; Academic Press: New York, 1963; pp 477479. (12) Dzombak, D. A.; Morel, F. M. M. Surface Complexation Modelling, Hydrous Ferric Oxide; John Wiley & Sons, Inc.: New York, 1990. (13) Stumm, W.; Morgan, J. J. Aquatic Chemistry; John Wiley & Sons, Inc.: New York, 1996. (14) Sigg, L. M. Ph.D. Dissertation, Swiss Federal Institute of Technology Zu ¨ rich, Diss ETHZ No. 6417, Zu ¨ rich, Switzerland, 1979. (15) Blank, B.; Rossington, D. R.; Weinland, L. A. J. Am. Ceram. Soc. 1963, 46, 395. (16) Rossington, D. R.; Runk, E. J. J. Am. Ceram. Soc. 1968, 51, 46. (17) Costa, U.; Massazza, F. Il Cemento 1984, 127. (18) Andersen, P. J.; Roy, D. M. Cem. Concr. Res. 1987, 17, 805. (19) Uchikawa, H.; Hanehara, S.; Shirasaka, T.; Sawaki, D. Cem. Concr. Res. 1992, 22, 1115. (20) Singh, N. B.; Sarvahi, R.; Singh, N. P. Cem. Concr. Res. 1992, 22, 725. (21) Spanka, G.; Thielen, G. Beton 1995, 320. (22) Tejedor-Tejedor, M. I.; Anderson, M. A. Langmuir 1986, 2, 203. (23) Hug, S.; Sulzberger, B. Langmuir 1994, 10, 3587.
Received for review June 13, 1996. Revised manuscript received January 2, 1997. Accepted January 3, 1997. ES960505I
VOL. 31, NO. 4, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1245
