- Email: [email protected]
Evaluation of buffer materials by associating engineering and sorption properties
ARTICLE IN PRESS
Applied Radiation and Isotopes 61 (2004) 1163–1172
Evaluation of buffer materials by associating engineering and sorption properties Yi-Lin Jana,b,*, Shih-Chin Tsaic, Hwai-Ping Chengd, Chun-Nan Hsua a
Department of Nuclear Science, National Tsing Hua University, Hsinchu 300, Taiwan b Department of Civil Engineering, Ching Yun University, Jungli 320, Taiwan c Department of Industrial Safety and Hygiene, Fooyin University, Kaohsiung 831, Taiwan d US Army Engineer Research and Development Center, Vicksburg, MS 39180-6199, USA Received 20 October 2003; received in revised form 9 January 2004; accepted 30 March 2004
Abstract To provide an overall functional evaluation of buffer materials, this study attempted to investigate the relationships among the engineering properties, plastic index (PI), compaction efficiency, sorption properties, and distribution ratio (Rd) for some buffer materials composed of quartz sand and bentonite. Th and U were nuclides of interest, and both synthetic groundwater (GW) and seawater (SW) were used for batch sorption experiments, while the deionized water (DIW) was used for engineering property tests. SW and GW were also used to evaluate the effects on PI. The results show that the maximum dry density was reached when bentonite content was 30% with the same compaction energy by the ASTM D698 method. PI and bentonite content of tested buffer materials consisting of bentonite and quartz sand demonstrated a linearly proportional relationship regardless of the solution used. The following sequence of PIDIW > PIGW > PISW is due to coagulation and flocculation effects. The buffer materials of lower PI value could decrease swelling potential and increase permeability. The Rd observed in GW and SW of U increased linearly with PI measured in DIW, although the Rd of Th remained relatively constant above a PI of 88. From the viewpoints of associated engineering and sorption properties, the buffer materials containing 30–50% bentonite are probably the most favorable choice. Another result shows that U has a better additivity with respect to Rd than Th in both synthetic GW and synthetic SW. These results will allow a determination of more effective buffer material composition, and improved estimates of the overall Rd of the buffer material mixture from the Rd of each mineral component. r 2004 Elsevier Ltd. All rights reserved. Keywords: Buffer material; Plastic index; Compaction efficiency; Sorption; Additivity
1. Introduction Bentonite clay and quartz sand have been selected as potential buffer and backfill materials for high-level radwaste disposal (Jedinakova-Krizova, 1998). The functional requirements of buffer materials consist of *Corresponding author. Department of Nuclear Science, National Tsing Hua University, Hsinchu 300, Taiwan. Fax: +886-3-459-0208. E-mail address: [email protected] (Y.-L. Jan).
both engineering and chemical properties (JNC, 2000; Yong, 1996). The engineering properties include: (a) preventing the waste container from settling and moving, (b) conducting the heat of decay away from the container, (c) providing a low permeability, (d) maintaining a high swelling potential, and (e) providing a high compaction efficiency. Chemical properties include: (a) retardation of the transport of nuclides, and (b) maintaining any intruding water in a reduced condition. Some recent studies of buffer-backfill materials have focused on physical properties (e.g., hydraulic
0969-8043/$ - see front matter r 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.apradiso.2004.03.116
ARTICLE IN PRESS 1164
Y.-L. Jan et al. / Applied Radiation and Isotopes 61 (2004) 1163–1172
conductivity, thermal conductivity, mechanics, swelling properties, etc.; Komine and Ogata, 1994; Radhakrishna and Chan, 1989; Borgesson et al., 2001), and others studies have focused on chemical properties (e.g., sorption, diffusion, migration and transport, etc.; Bout et al.,1998; Muurinen and Lehikoinen, 1999; Tsai et al., 2001). Bentonite was associated with high cation exchange capacity, high swelling potential and low hydraulic conductivity, while quartz sand was associated with high thermal conductivity and better mechanical properties. The researches that were aimed primarily at understanding the physical or the chemical properties of buffer/backfill materials composed of quartz sand and bentonite could not achieve an optimal mixture. Establishing the relationship between physical and chemical properties in composite materials should, however, be beneficial to selecting an optional composite material mixture. In addition, other research also pointed out that the sorption properties (related to Rd value) of radionuclides (RN) with respect to particular mineral components of the mixture could be used to evaluate the Rd value of the mixture, thus providing additional benefits in choosing effective composite materials (Jacquier et al., 2001; NAGRA, 1993). The purpose of this study was to measure the Rd value of U and Th with respect to various composite ratios of bentonite and quartz sand using batch sorption tests, and to acquire the PI values and maximum dry density with respect to various composite ratios of bentonite and quartz sand, using an Atterberg limit test and standard proctor compaction tests. In addition, this study also attempted to establish the relationship between the physical and chemical properties of composite materials. The plastic index (PI) was calculated by the Atterberg liquid limit (LL) and also by the plastic limit (PL) of soil as follows: PI ¼ LL PL:
ð1Þ
PI values are usually applied to the classification of soil and the evaluation of various engineering properties such as compressive and swelling characteristics. Standard proctor compaction tests provide the maximum dry density and optimum moisture content with the same compaction energy. The compaction efficiency is closely related with construction cost (Das, 1990). The equilibrium distribution coefficient (Kd) and the distribution ratio (Rd) for non-equilibrium partitioning of RN can be calculated by the following formula: Kd or Rd ðml=gÞ ¼ S=C;
ð2Þ
where S is the concentration of RN sorbed on solid phase (mol/g), and C is the concentration of RN in solution (mol/ml). If the sorption is a surface process, a proportional relationship should exist between sorption
sites and RN. Furthermore, Rd or Kd of a mixture can be written as Kd or Rd ðml=gÞ ¼ S fi Rdi :
ð3Þ
Here fi is the mass fraction of mineral i in the mixture, and Rdi is the distribution ratio of RN on mineral i (Jacquier et al.,2001; NAGRA, 1993). In this work, mixtures of bentonite and quartz sand were used as the solid phase. Batch experiments were used to measure the Rd values of Th and U. Atterberg tests and standard proctor compaction tests were used to measure LL and PL by deionized water during the construction phase to establish the relationship between PI and Rd. Synthetic seawater (SW) and groundwater (GW) were used as liquid phases to simulate the possible conditions that may be encountered with geologic disposal within an island, including the effects for PI and sorption properties.
2. Experimental 2.1. Solid phase The solid phase material used for both the Atterberg limit tests and batch sorption tests were composed of mixtures of varying proportions of MX-80 bentonite with particle size less than 0.074 mm and quartz sand with particle size from 0.297 to 0.84 mm. The composite ratios of bentonite to quartz sand used for the Atterberg limit tests and standard proctor compaction tests were as follows: 9:1, 7:3, 5:5, 3:7 and 1:9, respectively (by dry weight). The ratios of bentonite to quartz sand used for the batch sorption tests were 10:0, 7:3, 5:5, 3:7 and 0:10. 2.2. Liquid phase To simulate the geochemical conditions that a deep geological repository might encounter, the solutions used for the batch sorption tests were synthetic compositions of GW and SW (Table 1) in which UO2(NO3)2 5H2O and Th(NO3)4 6H2O were added to yield a 104 M concentration of U and Th, respectively. To simulate the conditions for the in situ buffer materials during the construction phase, deionized water (DIW) was used as the solution in the Atterberg limit test and standard proctor compaction tests. SW and GW were also used to evaluate the effects on PI. 2.3. Batch sorption tests Batch sorption tests were conducted with a solid/ liquid ratio of 1 g/30 ml. The liquid phase solutions (i.e., SW and GW) containing 104 M U and 104 M Th were individually added to the solid phase materials of
ARTICLE IN PRESS Y.-L. Jan et al. / Applied Radiation and Isotopes 61 (2004) 1163–1172
various composition ratios; they were placed into the centrifugal tubes in triplicate and were shaken at 120 rpm. Samples collected at the 7th, 14th, 28th, and 56th day after centrifugalization by 10 380 g, and the residual concentrations of U and Th in the liquid phase were measured by ICP-MS. Values of pH and Eh were also measured by glass electrode and platinum redox electrode.
1165
plastic limit test using DIW, SW and GW, respectively. These tests were performed following the method of ASTM D4318, (ASTM, 1994) and the PI value was calculated by (Eq. (1)). Standard Proctor Compaction Test was performed by the method of ASTM D698 (ASTM, 1994); here only DIW was used. The energy of compaction per unit volume was 593.5 kJ/m3.
2.4. Atterberg limit tests and standard proctor compaction tests 3. Results and discussion Using the solid-phase mixtures described above, we performed the Atterberg liquid limit test and Table 1 The composition of synthetic GW and SW Composition
GW
SW
CaCl2 2H2O (g) MgCl2 6H2O (g) Na2SO4 (g) NaHCO3 (g) KCl (g) NaBr (g) SrCl2 6H2O (g) LiCl (g) NaF (g) NaCl (g) H3BO3 (g) CsCl (g) 0.1 N NaOH (ml) Total solution volume (l) pH
138.6634 7.0260 16.5704 0.2756 0.3086 1.0311 2.1330 0.1221 0.0672 92.2581 — — — 20 B7.5
30.3008 215.8030 79.8975 — 15.2157 1.7267 0.4639 0.0212 0.0571 482.2667 0.5145 — 16 20 B 8.2
Remarks: GW: synthetic groundwater; SW: synthetic sea water.
3.1. Results of Atterberg tests and standard proctor compaction tests of mixture DIW was used to simulate the liquid phase during construction, while SW and GW were used to simulate the water intruding after construction was completed. These three kinds of solutions and quartz sand/ bentonite mixtures were independently tested for liquid limit and plastic limit, and the PI values were calculated. The PI values using the three liquid phases were linearly proportional to the bentonite content (Fig. 1). As Das (1990) pointed out, when the PI value was larger than 35, this was a result of the high swelling potentials. The higher the PI value, the lower the water permeability is. The experimental results indicated that the PI for GW was lower than for DIW, and the PI for SW lower than for GW (PIDIW>PIGW>PISW). In the mixture, bentonite was the material that most affected the magnitude of the PI value. Because of the relatively high cation concentrations in the GW and SW, and the coagulation and flocculation that might have resulted, the surface electrical double layer of the bentonite
Fig. 1. The relationship between PI and bentonite content in the three liquid phases.
ARTICLE IN PRESS Y.-L. Jan et al. / Applied Radiation and Isotopes 61 (2004) 1163–1172
1166
Fig. 2. The compaction results and void ratio (e) for varying PI of mixture in deionic water. Table 2 pH ( Eh ) values of GW and SW in Th sorptiona Time (d)
7 14 28 56
Solution
c
SW GWc SW GW SW GW SW GW
pH ( Eh ) B100+Q0
B70+Q30b
B50+Q50
B30+Q70
B0+Q100
7.80 7.95 7.59 7.92 7.57 7.58 7.66 7.50
7.87 7.84 7.71 7.59 7.69 7.44 7.67 7.50
7.89 7.57 7.71 7.65 7.74 7.35 7.71 7.38
7.87 7.50 7.84 7.64 7.59 7.28 7.59 7.49
5.02 4.96 4.97 4.88 4.72 5.05 4.58 4.80
(126) (202) (231) (217) (209) (163) (195) (215)
(132) (172) (217) (171) (130) (155) (204) (216)
(120) (165) (160) (163) (143) (156) (214) (217)
(120) (159) (145) (168) (146) (152) (226) (218)
(300) (268) (286) (231) (222) (215) (244) (219)
a
Initial pH = 4.85, Eh = 192 mV in SW and initial pH = 4.78, Eh =192 mV in GW. Bentonite mix with quartz sand. B70+Q30 means bentonite 70% mixed with quartz 30% by dry weight. c GW: synthetic groundwater; SW: sea water. b
Table 3 pH (Eh) values of GW and SW in U sorptiona Time (d)
Solution
c
7 14 28 56 a
SW GWc SW GW SW GW SW GW
pH (Eh ) B100+Q0
B70+Q30b
B50+Q50
B30+Q70
B0+Q100
7.85 7.72 7.78 7.96 7.58 7.46 7.56 7.57
7.81 7.28 7.74 7.78 7.57 7.27 7.57 7.48
7.88 7.49 7.76 7.57 7.81 7.26 7.63 7.38
7.97 7.23 7.94 7.63 7.66 7.38 7.67 7.06
5.58 6.38 5.85 6.87 5.80 6.92 6.04 6.88
(104) ( 72) (205) (211) (124) (220) (239) (172)
(111) (121) (224) (204) (119) (205) (241) (167)
(99.2) (126) (204) (195) (114) (189) (238) (196)
(103) (126) (189) (184) (116) (179) (249) (203)
Initial pH =6.21, Eh = 108 mV in SW and initial pH = 7.55, Eh =117 mV in GW. Bentonite mix with quartz sand. B70 + Q30 means bentonite 70% mixed with quartz 30% by dry weight. c GW: synthetic groundwater; SW: sea water. b
(118) (125) (257) (207) (166) (173) (238) (219)
ARTICLE IN PRESS Y.-L. Jan et al. / Applied Radiation and Isotopes 61 (2004) 1163–1172
1167
Fig. 3. Sorption percentage of U on different bentonite/quartz sand ratio in SW and GW.
particle was probably compressed. This also demonstrates that the swelling potential would decrease when using GW or SW, and water permeability would increase. The results of standard proctor compaction tests show that the maximum dry density (gd;max =1.72 g/cm3) was obtained when bentonite content is 30% (Fig. 2). That is, when the PI=88, there was better compaction efficiency as well as lower void ratio (e). In addition, these results demonstrated a better gradation when bentonite content was 30%.
3.2. Sorption of U and Th on the mixtures The pH and Eh of the samples collected from the sorption batch tests after 7, 14, 28, and 56 days are shown in Tables 2 and 3, respectively. The initial pH and Eh of SW and GW with 104 M U and that with 104 M Th are also shown in Tables 2 and 3. The results showed that, because bentonite has a good pH buffer capacity, the final pH was in the range of 7–8 when the bentonite content exceeded 30%, regardless of whether U or Th was present in SW and GW. The final pH and initial pH for pure quartz sand
ARTICLE IN PRESS 1168
Y.-L. Jan et al. / Applied Radiation and Isotopes 61 (2004) 1163–1172
Fig. 4. Sorption percentage of Th on different bentonite/quartz sand ratio in SW and GW.
were similar. The measured Eh values ranged between 100 and 300 mV. With respect to the sorption batch tests, the results showed that as the bentonite increased, the sorption of U proportionally increased, regardless of whether SW or GW was used in the liquid phase. Because bentonite has higher ion exchange capacity than quartz sand, bentonite dominated the sorption with respect to U (Fig. 3). Furthermore, U in aqueous solution may exist in four oxidation states: 3+, 4+, 5+, and 6+. This could easily result in a complex reaction with the anions in
GW and SW. The solubility values of these species were relatively high (Hsu, 2001; Joseph and Seaborg, 1957). In comparison, the only oxidation state of Th in aqueous solution is 4+. When bentonite exceeded 30% and the pH range was between 7 and 8, Th generally formed Th(OH)4 and precipitated, having a dramatic affect on the apparent Rd. The Rd value of the sorption on quartz sand was lower, partly because the sorption of the quartz sand with respect to Th was relatively low when pH = 5–6. In addition, in these pH ranges, Th tended to form Th(OH)3+, Th(OH)2+ and 2
ARTICLE IN PRESS Y.-L. Jan et al. / Applied Radiation and Isotopes 61 (2004) 1163–1172
1169
Fig. 5. The relationship between Rd of U in SW and GW and PI in DIW.
Th(OH)+ 3 , while the percentage of Th(OH)4 was relatively small (given in Fig. 4) (Joseph and Seaborg, 1957; Cromieres et al., 1998). 3.3. Relationship between PI and Rd The results of the Rd measurements as described in Section 3.2 could be used to evaluate the sorption characteristics of mixture materials. Likewise, the results of the PI measurements by DIW as described in Section 3.1 could be used to evaluate engineering properties (e.g., water permeability, swelling potential and compression properties, etc.) of mixture materials. Therefore, if one could establish the relationship between Rd
and PI, it should provide incremental benefits for the choice and evaluation of the optional composition of mixture materials. The relationship between Rd and PI for U in SW and GW and that for Th in SW and GW are shown in Figs. 5 and 6, respectively. As shown in Fig. 5, at each sampling time, Rd values for U varied linearly with PI in both GW and SW. As shown in Fig. 6, Rd values for Th remained relatively constant for PI values higher than 88 (corresponding to a bentonite content of 30%). This behavior occurred at all sampling times for both GW and SW. To summarize, when using a quartz sand/bentonite mixture as a buffer-backfill material for the purpose of retarding nuclides transport, increasing the ratio of
ARTICLE IN PRESS 1170
Y.-L. Jan et al. / Applied Radiation and Isotopes 61 (2004) 1163–1172
Fig. 6. The relationship between Rd of Th in SW and GW and PI in DIW.
bentonite to sand up to 100% (PI = 398) will result in higher retardation of U, but no additional retardation of Th is obtained if the ratio of bentonite to sand is increased above 30% (PI=88). From the engineering point of view, to obtain a PI value of approximately 400 may be impractical during construction, because of high swelling potential and plasticity. The heat conductivity would also be poor without the addition of quartz sand. To balance the engineering needs against desirable chemical properties, a material with a PI of approximately 100 (bentonite content 30–50%) would be a better choice for the composition. As described in Section 3.1, when a mixture with bentonite content of 30–50% was intruded by SW or GW, it would still exhibit proper swelling potential and
have the ability to self-seal and prevent cracks, thereby minimizing the intrusion of water and the release of nuclides, in addition to having a higher compaction efficiency. 3.4. Additivity of Rd Another concern was whether one could use the Rd value of the individual materials to predict the Rd value of various composition ratios. This could prove useful when choosing and evaluating buffer-backfill materials. Figs. 7 and 8 show a comparison of experimental Rd and calculated Rd values (Eq. (3)) for U and Th in GW and SW for the various tested mixtures. The results of this study demonstrated that the Rd value of U showed good additivity in the mixture materials for both SW and GW
ARTICLE IN PRESS Y.-L. Jan et al. / Applied Radiation and Isotopes 61 (2004) 1163–1172
1171
Fig. 7. The additivity of Rd of Th on quartz sand/bentonite mixtures.
Fig. 8. The additivity of Rd of U on quartz sand/bentonite mixtures.
(r2 = 0.9368 and 0.9553), while Th did not show similar results (r2 = 0.6563 and 0.5592). Although the calculated Rd value and experimental Rd values of Th did not show good additivity, those values all tended to be conservative.
4. Conclusion Under the experimental conditions of this study, PI and bentonite content of tested buffer material consisting of bentonite and quartz sand demonstrated a linearly
proportional relationship regardless of the solution used (SW, GW, or DIW). The results also showed that PIDIW > PIGW > PISW due to coagulation and flocculation effects. Under the same compaction energy condition, the PI of 88 (bentonite content 30%) has the maximum dry density and minimum void ratio for various ratios of the mixture. In both SW and GW, the sorption percentage of U increased linearly with bentonite content, although the sorption percentage of Th remained relatively constant above a PI of 88 (bentonite content 30%). If one increased the bentonite content of a buffer material above 30%, the Rd of U was increased,
ARTICLE IN PRESS 1172
Y.-L. Jan et al. / Applied Radiation and Isotopes 61 (2004) 1163–1172
and the ability to conduct heat due to proportionately smaller quartz sand content was decreased. The difficulty of working at lower compaction efficiency with a material containing a high proportion of bentonite (from an engineering standpoint) indicates that buffer materials containing 30–50% bentonite are probably the most favorable choice. This investigation further indicates that U has better additivity with respect to Rd than Th in both synthetic groundwater and synthetic seawater. These results provide useful information in the choice of buffer material composition, and in evaluating the overall Rd of the mixture from the Rd of each buffer component.
References ASTM, 1994. Soil and rock. Construction. Annual Book of ASTM Standard, Section 4, Vol. 04.08. ASTM, Philadelphia. Borgesson, L., Chijimatsu, M., Fujita, T., Nguyen, T.S., Rutqvist, J., Jing, L., 2001. Thermal-hydro-mechanical characterization of a betonies-based buffer material by laboratory tests and numerical back analysis. Int. J. Rock Mech. Mining Sci. 38, 95–104. Bout, K.A., Cowper, M.M., Heath, T.G., Sato, H., Shibutani, T., Yui, M., 1998. Toward an understanding of the sorption of U(VI) and Se(IV) on sodium bentonite. J. Contam. Hydrol. 35, 141–150. Cromieres, L., Moulin, V., Fourest, B., Guillaumont, R., Giffaut, E., 1998. Sorption of thorium onto hematite colloids. Radiochim. Acta 82, 249–255.
Das, B.M., 1990. Principles of Geotechnical Engineering. PWSKent Publishing Company, Boston. Hsu, C.N., 2001. An examination and analysis of the safe evaluation for LLW disposal site—to set up migration parameters of nuclides. Technical Report, Fuel Cycle and Materials Administration. Taipei, Taiwan. Jacquier, P., Meier, P., Ly, J., 2001. Asorption of radioelements on mixtures of minerals—experimental study. Appl. Geochem. 16, 85–93. Jedinakova-Krizova, V., 1998. Migration of radionuclides in environment. J. Radioanal. Nucl. Chem. 229 (1–2), 3–18. JNC, 2000. Repository design and engineering technology. Project to Establish the Scientific and Technical Basis for HLW Disposal in Japan, H12. Japan Nuclear Cycle Development Institute. Joseph, J.K., Seaborg, G.T., 1957. The Chemistry of Actinide Elements, 1st Edition. Wiley, New York. Komine, H., Ogata, N., 1994. Experimental study on swelling characteristics of compacted bentonite. Can. Geotechnol. J. 31, 478–490. Muurinen, A., Lehikoinen, J., 1999. Porewater chemistry in compact bentonite. Eng. Geol. 54, 207–214. NAGRA, 1993. Sorption database for bentonite. Technical Report 93-06. National Cooperation for the Disposal of Radioactive Waste, Switzerland. Radhakrishna, H.S., Chan, H.T., 1989. Thermal and physical properties of candidate buffer-backfill materials for nuclear fuel waste disposal vault. Can. Geotechnol. J. 26, 629–639. Tsai, S.C., Ouyang, S., Hsu, C.N., 2001. Sorption and diffusion behavior of Cs and Sr on Jih-Hsing bentonite. Appl. Radiat. Isot. 54, 209–215. Yong, R.N., Boonsinsuk, P., Wong, G., 1996. Formulation of backfill material for a nuclear fuel waste disposal vault. Can. Geotechnol. J. 23, 216–228.
Applied Radiation and Isotopes 61 (2004) 1163–1172
Evaluation of buffer materials by associating engineering and sorption properties Yi-Lin Jana,b,*, Shih-Chin Tsaic, Hwai-Ping Chengd, Chun-Nan Hsua a
Department of Nuclear Science, National Tsing Hua University, Hsinchu 300, Taiwan b Department of Civil Engineering, Ching Yun University, Jungli 320, Taiwan c Department of Industrial Safety and Hygiene, Fooyin University, Kaohsiung 831, Taiwan d US Army Engineer Research and Development Center, Vicksburg, MS 39180-6199, USA Received 20 October 2003; received in revised form 9 January 2004; accepted 30 March 2004
Abstract To provide an overall functional evaluation of buffer materials, this study attempted to investigate the relationships among the engineering properties, plastic index (PI), compaction efficiency, sorption properties, and distribution ratio (Rd) for some buffer materials composed of quartz sand and bentonite. Th and U were nuclides of interest, and both synthetic groundwater (GW) and seawater (SW) were used for batch sorption experiments, while the deionized water (DIW) was used for engineering property tests. SW and GW were also used to evaluate the effects on PI. The results show that the maximum dry density was reached when bentonite content was 30% with the same compaction energy by the ASTM D698 method. PI and bentonite content of tested buffer materials consisting of bentonite and quartz sand demonstrated a linearly proportional relationship regardless of the solution used. The following sequence of PIDIW > PIGW > PISW is due to coagulation and flocculation effects. The buffer materials of lower PI value could decrease swelling potential and increase permeability. The Rd observed in GW and SW of U increased linearly with PI measured in DIW, although the Rd of Th remained relatively constant above a PI of 88. From the viewpoints of associated engineering and sorption properties, the buffer materials containing 30–50% bentonite are probably the most favorable choice. Another result shows that U has a better additivity with respect to Rd than Th in both synthetic GW and synthetic SW. These results will allow a determination of more effective buffer material composition, and improved estimates of the overall Rd of the buffer material mixture from the Rd of each mineral component. r 2004 Elsevier Ltd. All rights reserved. Keywords: Buffer material; Plastic index; Compaction efficiency; Sorption; Additivity
1. Introduction Bentonite clay and quartz sand have been selected as potential buffer and backfill materials for high-level radwaste disposal (Jedinakova-Krizova, 1998). The functional requirements of buffer materials consist of *Corresponding author. Department of Nuclear Science, National Tsing Hua University, Hsinchu 300, Taiwan. Fax: +886-3-459-0208. E-mail address: [email protected] (Y.-L. Jan).
both engineering and chemical properties (JNC, 2000; Yong, 1996). The engineering properties include: (a) preventing the waste container from settling and moving, (b) conducting the heat of decay away from the container, (c) providing a low permeability, (d) maintaining a high swelling potential, and (e) providing a high compaction efficiency. Chemical properties include: (a) retardation of the transport of nuclides, and (b) maintaining any intruding water in a reduced condition. Some recent studies of buffer-backfill materials have focused on physical properties (e.g., hydraulic
0969-8043/$ - see front matter r 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.apradiso.2004.03.116
ARTICLE IN PRESS 1164
Y.-L. Jan et al. / Applied Radiation and Isotopes 61 (2004) 1163–1172
conductivity, thermal conductivity, mechanics, swelling properties, etc.; Komine and Ogata, 1994; Radhakrishna and Chan, 1989; Borgesson et al., 2001), and others studies have focused on chemical properties (e.g., sorption, diffusion, migration and transport, etc.; Bout et al.,1998; Muurinen and Lehikoinen, 1999; Tsai et al., 2001). Bentonite was associated with high cation exchange capacity, high swelling potential and low hydraulic conductivity, while quartz sand was associated with high thermal conductivity and better mechanical properties. The researches that were aimed primarily at understanding the physical or the chemical properties of buffer/backfill materials composed of quartz sand and bentonite could not achieve an optimal mixture. Establishing the relationship between physical and chemical properties in composite materials should, however, be beneficial to selecting an optional composite material mixture. In addition, other research also pointed out that the sorption properties (related to Rd value) of radionuclides (RN) with respect to particular mineral components of the mixture could be used to evaluate the Rd value of the mixture, thus providing additional benefits in choosing effective composite materials (Jacquier et al., 2001; NAGRA, 1993). The purpose of this study was to measure the Rd value of U and Th with respect to various composite ratios of bentonite and quartz sand using batch sorption tests, and to acquire the PI values and maximum dry density with respect to various composite ratios of bentonite and quartz sand, using an Atterberg limit test and standard proctor compaction tests. In addition, this study also attempted to establish the relationship between the physical and chemical properties of composite materials. The plastic index (PI) was calculated by the Atterberg liquid limit (LL) and also by the plastic limit (PL) of soil as follows: PI ¼ LL PL:
ð1Þ
PI values are usually applied to the classification of soil and the evaluation of various engineering properties such as compressive and swelling characteristics. Standard proctor compaction tests provide the maximum dry density and optimum moisture content with the same compaction energy. The compaction efficiency is closely related with construction cost (Das, 1990). The equilibrium distribution coefficient (Kd) and the distribution ratio (Rd) for non-equilibrium partitioning of RN can be calculated by the following formula: Kd or Rd ðml=gÞ ¼ S=C;
ð2Þ
where S is the concentration of RN sorbed on solid phase (mol/g), and C is the concentration of RN in solution (mol/ml). If the sorption is a surface process, a proportional relationship should exist between sorption
sites and RN. Furthermore, Rd or Kd of a mixture can be written as Kd or Rd ðml=gÞ ¼ S fi Rdi :
ð3Þ
Here fi is the mass fraction of mineral i in the mixture, and Rdi is the distribution ratio of RN on mineral i (Jacquier et al.,2001; NAGRA, 1993). In this work, mixtures of bentonite and quartz sand were used as the solid phase. Batch experiments were used to measure the Rd values of Th and U. Atterberg tests and standard proctor compaction tests were used to measure LL and PL by deionized water during the construction phase to establish the relationship between PI and Rd. Synthetic seawater (SW) and groundwater (GW) were used as liquid phases to simulate the possible conditions that may be encountered with geologic disposal within an island, including the effects for PI and sorption properties.
2. Experimental 2.1. Solid phase The solid phase material used for both the Atterberg limit tests and batch sorption tests were composed of mixtures of varying proportions of MX-80 bentonite with particle size less than 0.074 mm and quartz sand with particle size from 0.297 to 0.84 mm. The composite ratios of bentonite to quartz sand used for the Atterberg limit tests and standard proctor compaction tests were as follows: 9:1, 7:3, 5:5, 3:7 and 1:9, respectively (by dry weight). The ratios of bentonite to quartz sand used for the batch sorption tests were 10:0, 7:3, 5:5, 3:7 and 0:10. 2.2. Liquid phase To simulate the geochemical conditions that a deep geological repository might encounter, the solutions used for the batch sorption tests were synthetic compositions of GW and SW (Table 1) in which UO2(NO3)2 5H2O and Th(NO3)4 6H2O were added to yield a 104 M concentration of U and Th, respectively. To simulate the conditions for the in situ buffer materials during the construction phase, deionized water (DIW) was used as the solution in the Atterberg limit test and standard proctor compaction tests. SW and GW were also used to evaluate the effects on PI. 2.3. Batch sorption tests Batch sorption tests were conducted with a solid/ liquid ratio of 1 g/30 ml. The liquid phase solutions (i.e., SW and GW) containing 104 M U and 104 M Th were individually added to the solid phase materials of
ARTICLE IN PRESS Y.-L. Jan et al. / Applied Radiation and Isotopes 61 (2004) 1163–1172
various composition ratios; they were placed into the centrifugal tubes in triplicate and were shaken at 120 rpm. Samples collected at the 7th, 14th, 28th, and 56th day after centrifugalization by 10 380 g, and the residual concentrations of U and Th in the liquid phase were measured by ICP-MS. Values of pH and Eh were also measured by glass electrode and platinum redox electrode.
1165
plastic limit test using DIW, SW and GW, respectively. These tests were performed following the method of ASTM D4318, (ASTM, 1994) and the PI value was calculated by (Eq. (1)). Standard Proctor Compaction Test was performed by the method of ASTM D698 (ASTM, 1994); here only DIW was used. The energy of compaction per unit volume was 593.5 kJ/m3.
2.4. Atterberg limit tests and standard proctor compaction tests 3. Results and discussion Using the solid-phase mixtures described above, we performed the Atterberg liquid limit test and Table 1 The composition of synthetic GW and SW Composition
GW
SW
CaCl2 2H2O (g) MgCl2 6H2O (g) Na2SO4 (g) NaHCO3 (g) KCl (g) NaBr (g) SrCl2 6H2O (g) LiCl (g) NaF (g) NaCl (g) H3BO3 (g) CsCl (g) 0.1 N NaOH (ml) Total solution volume (l) pH
138.6634 7.0260 16.5704 0.2756 0.3086 1.0311 2.1330 0.1221 0.0672 92.2581 — — — 20 B7.5
30.3008 215.8030 79.8975 — 15.2157 1.7267 0.4639 0.0212 0.0571 482.2667 0.5145 — 16 20 B 8.2
Remarks: GW: synthetic groundwater; SW: synthetic sea water.
3.1. Results of Atterberg tests and standard proctor compaction tests of mixture DIW was used to simulate the liquid phase during construction, while SW and GW were used to simulate the water intruding after construction was completed. These three kinds of solutions and quartz sand/ bentonite mixtures were independently tested for liquid limit and plastic limit, and the PI values were calculated. The PI values using the three liquid phases were linearly proportional to the bentonite content (Fig. 1). As Das (1990) pointed out, when the PI value was larger than 35, this was a result of the high swelling potentials. The higher the PI value, the lower the water permeability is. The experimental results indicated that the PI for GW was lower than for DIW, and the PI for SW lower than for GW (PIDIW>PIGW>PISW). In the mixture, bentonite was the material that most affected the magnitude of the PI value. Because of the relatively high cation concentrations in the GW and SW, and the coagulation and flocculation that might have resulted, the surface electrical double layer of the bentonite
Fig. 1. The relationship between PI and bentonite content in the three liquid phases.
ARTICLE IN PRESS Y.-L. Jan et al. / Applied Radiation and Isotopes 61 (2004) 1163–1172
1166
Fig. 2. The compaction results and void ratio (e) for varying PI of mixture in deionic water. Table 2 pH ( Eh ) values of GW and SW in Th sorptiona Time (d)
7 14 28 56
Solution
c
SW GWc SW GW SW GW SW GW
pH ( Eh ) B100+Q0
B70+Q30b
B50+Q50
B30+Q70
B0+Q100
7.80 7.95 7.59 7.92 7.57 7.58 7.66 7.50
7.87 7.84 7.71 7.59 7.69 7.44 7.67 7.50
7.89 7.57 7.71 7.65 7.74 7.35 7.71 7.38
7.87 7.50 7.84 7.64 7.59 7.28 7.59 7.49
5.02 4.96 4.97 4.88 4.72 5.05 4.58 4.80
(126) (202) (231) (217) (209) (163) (195) (215)
(132) (172) (217) (171) (130) (155) (204) (216)
(120) (165) (160) (163) (143) (156) (214) (217)
(120) (159) (145) (168) (146) (152) (226) (218)
(300) (268) (286) (231) (222) (215) (244) (219)
a
Initial pH = 4.85, Eh = 192 mV in SW and initial pH = 4.78, Eh =192 mV in GW. Bentonite mix with quartz sand. B70+Q30 means bentonite 70% mixed with quartz 30% by dry weight. c GW: synthetic groundwater; SW: sea water. b
Table 3 pH (Eh) values of GW and SW in U sorptiona Time (d)
Solution
c
7 14 28 56 a
SW GWc SW GW SW GW SW GW
pH (Eh ) B100+Q0
B70+Q30b
B50+Q50
B30+Q70
B0+Q100
7.85 7.72 7.78 7.96 7.58 7.46 7.56 7.57
7.81 7.28 7.74 7.78 7.57 7.27 7.57 7.48
7.88 7.49 7.76 7.57 7.81 7.26 7.63 7.38
7.97 7.23 7.94 7.63 7.66 7.38 7.67 7.06
5.58 6.38 5.85 6.87 5.80 6.92 6.04 6.88
(104) ( 72) (205) (211) (124) (220) (239) (172)
(111) (121) (224) (204) (119) (205) (241) (167)
(99.2) (126) (204) (195) (114) (189) (238) (196)
(103) (126) (189) (184) (116) (179) (249) (203)
Initial pH =6.21, Eh = 108 mV in SW and initial pH = 7.55, Eh =117 mV in GW. Bentonite mix with quartz sand. B70 + Q30 means bentonite 70% mixed with quartz 30% by dry weight. c GW: synthetic groundwater; SW: sea water. b
(118) (125) (257) (207) (166) (173) (238) (219)
ARTICLE IN PRESS Y.-L. Jan et al. / Applied Radiation and Isotopes 61 (2004) 1163–1172
1167
Fig. 3. Sorption percentage of U on different bentonite/quartz sand ratio in SW and GW.
particle was probably compressed. This also demonstrates that the swelling potential would decrease when using GW or SW, and water permeability would increase. The results of standard proctor compaction tests show that the maximum dry density (gd;max =1.72 g/cm3) was obtained when bentonite content is 30% (Fig. 2). That is, when the PI=88, there was better compaction efficiency as well as lower void ratio (e). In addition, these results demonstrated a better gradation when bentonite content was 30%.
3.2. Sorption of U and Th on the mixtures The pH and Eh of the samples collected from the sorption batch tests after 7, 14, 28, and 56 days are shown in Tables 2 and 3, respectively. The initial pH and Eh of SW and GW with 104 M U and that with 104 M Th are also shown in Tables 2 and 3. The results showed that, because bentonite has a good pH buffer capacity, the final pH was in the range of 7–8 when the bentonite content exceeded 30%, regardless of whether U or Th was present in SW and GW. The final pH and initial pH for pure quartz sand
ARTICLE IN PRESS 1168
Y.-L. Jan et al. / Applied Radiation and Isotopes 61 (2004) 1163–1172
Fig. 4. Sorption percentage of Th on different bentonite/quartz sand ratio in SW and GW.
were similar. The measured Eh values ranged between 100 and 300 mV. With respect to the sorption batch tests, the results showed that as the bentonite increased, the sorption of U proportionally increased, regardless of whether SW or GW was used in the liquid phase. Because bentonite has higher ion exchange capacity than quartz sand, bentonite dominated the sorption with respect to U (Fig. 3). Furthermore, U in aqueous solution may exist in four oxidation states: 3+, 4+, 5+, and 6+. This could easily result in a complex reaction with the anions in
GW and SW. The solubility values of these species were relatively high (Hsu, 2001; Joseph and Seaborg, 1957). In comparison, the only oxidation state of Th in aqueous solution is 4+. When bentonite exceeded 30% and the pH range was between 7 and 8, Th generally formed Th(OH)4 and precipitated, having a dramatic affect on the apparent Rd. The Rd value of the sorption on quartz sand was lower, partly because the sorption of the quartz sand with respect to Th was relatively low when pH = 5–6. In addition, in these pH ranges, Th tended to form Th(OH)3+, Th(OH)2+ and 2
ARTICLE IN PRESS Y.-L. Jan et al. / Applied Radiation and Isotopes 61 (2004) 1163–1172
1169
Fig. 5. The relationship between Rd of U in SW and GW and PI in DIW.
Th(OH)+ 3 , while the percentage of Th(OH)4 was relatively small (given in Fig. 4) (Joseph and Seaborg, 1957; Cromieres et al., 1998). 3.3. Relationship between PI and Rd The results of the Rd measurements as described in Section 3.2 could be used to evaluate the sorption characteristics of mixture materials. Likewise, the results of the PI measurements by DIW as described in Section 3.1 could be used to evaluate engineering properties (e.g., water permeability, swelling potential and compression properties, etc.) of mixture materials. Therefore, if one could establish the relationship between Rd
and PI, it should provide incremental benefits for the choice and evaluation of the optional composition of mixture materials. The relationship between Rd and PI for U in SW and GW and that for Th in SW and GW are shown in Figs. 5 and 6, respectively. As shown in Fig. 5, at each sampling time, Rd values for U varied linearly with PI in both GW and SW. As shown in Fig. 6, Rd values for Th remained relatively constant for PI values higher than 88 (corresponding to a bentonite content of 30%). This behavior occurred at all sampling times for both GW and SW. To summarize, when using a quartz sand/bentonite mixture as a buffer-backfill material for the purpose of retarding nuclides transport, increasing the ratio of
ARTICLE IN PRESS 1170
Y.-L. Jan et al. / Applied Radiation and Isotopes 61 (2004) 1163–1172
Fig. 6. The relationship between Rd of Th in SW and GW and PI in DIW.
bentonite to sand up to 100% (PI = 398) will result in higher retardation of U, but no additional retardation of Th is obtained if the ratio of bentonite to sand is increased above 30% (PI=88). From the engineering point of view, to obtain a PI value of approximately 400 may be impractical during construction, because of high swelling potential and plasticity. The heat conductivity would also be poor without the addition of quartz sand. To balance the engineering needs against desirable chemical properties, a material with a PI of approximately 100 (bentonite content 30–50%) would be a better choice for the composition. As described in Section 3.1, when a mixture with bentonite content of 30–50% was intruded by SW or GW, it would still exhibit proper swelling potential and
have the ability to self-seal and prevent cracks, thereby minimizing the intrusion of water and the release of nuclides, in addition to having a higher compaction efficiency. 3.4. Additivity of Rd Another concern was whether one could use the Rd value of the individual materials to predict the Rd value of various composition ratios. This could prove useful when choosing and evaluating buffer-backfill materials. Figs. 7 and 8 show a comparison of experimental Rd and calculated Rd values (Eq. (3)) for U and Th in GW and SW for the various tested mixtures. The results of this study demonstrated that the Rd value of U showed good additivity in the mixture materials for both SW and GW
ARTICLE IN PRESS Y.-L. Jan et al. / Applied Radiation and Isotopes 61 (2004) 1163–1172
1171
Fig. 7. The additivity of Rd of Th on quartz sand/bentonite mixtures.
Fig. 8. The additivity of Rd of U on quartz sand/bentonite mixtures.
(r2 = 0.9368 and 0.9553), while Th did not show similar results (r2 = 0.6563 and 0.5592). Although the calculated Rd value and experimental Rd values of Th did not show good additivity, those values all tended to be conservative.
4. Conclusion Under the experimental conditions of this study, PI and bentonite content of tested buffer material consisting of bentonite and quartz sand demonstrated a linearly
proportional relationship regardless of the solution used (SW, GW, or DIW). The results also showed that PIDIW > PIGW > PISW due to coagulation and flocculation effects. Under the same compaction energy condition, the PI of 88 (bentonite content 30%) has the maximum dry density and minimum void ratio for various ratios of the mixture. In both SW and GW, the sorption percentage of U increased linearly with bentonite content, although the sorption percentage of Th remained relatively constant above a PI of 88 (bentonite content 30%). If one increased the bentonite content of a buffer material above 30%, the Rd of U was increased,
ARTICLE IN PRESS 1172
Y.-L. Jan et al. / Applied Radiation and Isotopes 61 (2004) 1163–1172
and the ability to conduct heat due to proportionately smaller quartz sand content was decreased. The difficulty of working at lower compaction efficiency with a material containing a high proportion of bentonite (from an engineering standpoint) indicates that buffer materials containing 30–50% bentonite are probably the most favorable choice. This investigation further indicates that U has better additivity with respect to Rd than Th in both synthetic groundwater and synthetic seawater. These results provide useful information in the choice of buffer material composition, and in evaluating the overall Rd of the mixture from the Rd of each buffer component.
References ASTM, 1994. Soil and rock. Construction. Annual Book of ASTM Standard, Section 4, Vol. 04.08. ASTM, Philadelphia. Borgesson, L., Chijimatsu, M., Fujita, T., Nguyen, T.S., Rutqvist, J., Jing, L., 2001. Thermal-hydro-mechanical characterization of a betonies-based buffer material by laboratory tests and numerical back analysis. Int. J. Rock Mech. Mining Sci. 38, 95–104. Bout, K.A., Cowper, M.M., Heath, T.G., Sato, H., Shibutani, T., Yui, M., 1998. Toward an understanding of the sorption of U(VI) and Se(IV) on sodium bentonite. J. Contam. Hydrol. 35, 141–150. Cromieres, L., Moulin, V., Fourest, B., Guillaumont, R., Giffaut, E., 1998. Sorption of thorium onto hematite colloids. Radiochim. Acta 82, 249–255.
Das, B.M., 1990. Principles of Geotechnical Engineering. PWSKent Publishing Company, Boston. Hsu, C.N., 2001. An examination and analysis of the safe evaluation for LLW disposal site—to set up migration parameters of nuclides. Technical Report, Fuel Cycle and Materials Administration. Taipei, Taiwan. Jacquier, P., Meier, P., Ly, J., 2001. Asorption of radioelements on mixtures of minerals—experimental study. Appl. Geochem. 16, 85–93. Jedinakova-Krizova, V., 1998. Migration of radionuclides in environment. J. Radioanal. Nucl. Chem. 229 (1–2), 3–18. JNC, 2000. Repository design and engineering technology. Project to Establish the Scientific and Technical Basis for HLW Disposal in Japan, H12. Japan Nuclear Cycle Development Institute. Joseph, J.K., Seaborg, G.T., 1957. The Chemistry of Actinide Elements, 1st Edition. Wiley, New York. Komine, H., Ogata, N., 1994. Experimental study on swelling characteristics of compacted bentonite. Can. Geotechnol. J. 31, 478–490. Muurinen, A., Lehikoinen, J., 1999. Porewater chemistry in compact bentonite. Eng. Geol. 54, 207–214. NAGRA, 1993. Sorption database for bentonite. Technical Report 93-06. National Cooperation for the Disposal of Radioactive Waste, Switzerland. Radhakrishna, H.S., Chan, H.T., 1989. Thermal and physical properties of candidate buffer-backfill materials for nuclear fuel waste disposal vault. Can. Geotechnol. J. 26, 629–639. Tsai, S.C., Ouyang, S., Hsu, C.N., 2001. Sorption and diffusion behavior of Cs and Sr on Jih-Hsing bentonite. Appl. Radiat. Isot. 54, 209–215. Yong, R.N., Boonsinsuk, P., Wong, G., 1996. Formulation of backfill material for a nuclear fuel waste disposal vault. Can. Geotechnol. J. 23, 216–228.