Current Research in to Nanoparticles
As liquefaction can cause significant damage during earthquakes, and the current techniques and practises are not always suitable especially for existing developed sites, there is a need for new methods to be studied regarding liquefaction mitigation.
One potential new method of liquefaction mitigation is the use of nanoparticles as a passive site remediation technique. This technique involves slowly injecting the stabilisation material into a liquefiable subsoil and allowing the natural groundwater and hydraulic gradient to disperse the stabiliser uniformly throughout the soil (Gallagher and Mitchel, 2002). This technique allows the entire sites to be permeated instead of grout columns, which tend to form when using more traditional cement or silicate grouting.
During the review of literature information with a connection to liquefaction mitigation was found for the following nanoparticles:
a) Silica
b) Bentonite
c) Laponite
a) Various researchers have studied the effects of using colloidal silica as a passive site remediation technique (Díaz-Rodríguez et al., 2008; Gallagher and Mitchel, 2002; Gallagher, Conlee and Rollins, 2007; Gallagher, Pamuk and Abdoun, 2007; Mollamahmutoglu and Yilmaz, 2010). Colloidal silica is an aqueous dispersion of silica nanoparticles. Initially a low density and low viscosity solution forms which is similar to water. After a long induction period, where the viscosity stays low, a high viscous gel will form. Gallagher and Mitchel (2002) suggest that 5-10% by weight colloidal silica is expected to reduce liquefaction potential of saturated loose sands and still be economical (further results shown below).
b) The addition of the nanoparticle bentonite to a sand sample resistance to cyclic loading has also been studied (Gratchev et al., 2007). Gratchev et al. (2007) suggest that for bentonite to be successful in increasing liquefaction resistance over 7% by weight bentonite is necessary to be added to a sand sample.
c) Research has shown that aqueous dispersions of laponite also form a similar low viscous solution that transforms in to a high viscous gel over time (Bonn et al., 1999; Mongondry, Tassin and Nicolai, 2004; Mourchid et al., 1994). Bonn et al. (1999, p 7534) state that “[l]ow concentrations (typically a few percent by weight) of Laponite in water form suspensions with viscosity several orders of magnitude higher than of the solvent (water)”.
Following this research and discussion with the thesis supervisor, it has become clear that there is a gap in researching the affects of laponite on cyclic resistance of saturated sands. With laponite’s high viscosity at low concentrations it is hoped that laponite will be able to provide greater liquefaction resistance at low concentrations than other nanoparticles that have previously been studied.
Due to the large amount of research available for silica nanoparticles and its similarity to laponite an in-depth review of literature of the effects of its addition has been presented below. Research into colloidal silica has proven to be very successful and could provide a good framework for laponite testing. It is hoped that laponite will prove to be equally successful but with lower concentrations making it more economically viable.
Collodial Silica – In-depth Review
Gallagher and Mitchel (2002) performed cyclic triaxial tests on saturated loose sands grouted with 5, 10, 15 and 20% colloidal silica, and compared deformations against what was observed during cyclic triaxial tests for ungrouted samples. It was observed that both grouted and ungrouted samples had distinctly different deformation characteristics, ungrouted samples experienced very little strain prior to liquefaction. Though it was found that once liquefaction occurred that large strains quickly arose and the sample would collapse within a few additional loading cycles. Also observed was that as the cyclic stress ratio (CSR, maximum cyclic shear stress to initial effective confining stress) increased, the number of cycles that occurred before failure decreased. Comparing these results to grouted samples Gallagher and Mitchel (2002) noticed that for the same relative density and CSR, grouted samples experienced much less strain and none of the samples were found to have failed. Samples treated with lower concentrations of collodial silica experienced greater strains than those with higher concentrations. All the grouted samples returned to the original shape after loading had ceased.
Gallagher and Mitchel (2002) also tested the effect of curing time on cyclic resistance. It was noted that as the curing time increased, the strains reached during cyclic loading decreased. It is thought that this happens due to the fact that even after gelation occurs bonds in the collodial silica are still forming and these bonds are what increases the deformation resistance. These authors do hypothesize that this improvement to cyclic resistance will taper off with time. Unconfined compression tests were also undertaken in this study and showed that as the silica concentration increased within a sample the unconfined compression strength also increased. After cyclic loading had occurred unconfined compression tests where also undertaken to establish whether there had been a loss in compressive strength due to cyclic loading. It found that the loss in compressive strength varied with strains the sample had experienced during cyclic loading. As a general trend, the larger the strains experienced during cyclic loading, the greater the reduction in strength. It is suggested that this occurs as bonds will be broken during cyclic loading, greater levels of strains will therefore create larger numbers of broken bonds.
Testing done by Díaz-Rodríguez et al. (2008) and Mollamahmutoglu and Yilmaz (2010) support the findings of Gallagher and Mitchel (2002) in that these researchers also found that the addition of colloidal silica can significantly increase cyclic resistance. Diaz-Rodrigues et al. (2008) state that for a given initial relative density and initial vertical pressure, liquefaction resistance is significantly greater for colloidal silica grouted samples in comparison with ungrouted samples. In general, sand with higher initial relative densities has a greater cyclic resistance for both treated and untreated samples (Díaz-Rodríguez et al., 2008). Mollamahmutoglu and Yilmaz (2010) studied the effect of cyclic loading on the unconfined strength of colloidal silica grouted sand. It was observed that performing cyclic triaxial tests, with CSR varying from 0.13 to 0.52 for 1000 cycles, that no failures occurred. Comparing the unconfined strength before loading to the unconfined strength after loading showed an insignificant decrease in strength of around 10%. Previous studies have shown that ungrouted saturated sand samples lose their strength and liquefy with a CSR of 0.26 in around 13 cycles (Gallagher, 2000 cited in Mollamahmutoglu and Yilmaz, 2010 p.346), therefore the findings shown by Mollamahmutoglu and Yilmaz (2010) compare favourably with the results by Díaz-Rodríguez et al. (2008) and Gallagher and Mitchel (2002).
In a study by Gallagher, Pamuk and Abdoun (2007) a centrifuge model was undertake and subjected a loose density sand sample grouted with 6% collodial silica to two in-flight shaking events with 0.2g and 0.25g maximum acceleration respectively. During both shaking events no liquefaction was found to take place. The deformations were recorded throughout the shaking events and a maximum shear strain of 0.5 and 1% were recorded for the shaking events respectively. Vertical strains were measured to be 0.3% and 0.1%. On previous centrifuge model studies using untreated sand samples, vertical strains of 2-2.5% were recorded for shaking events with 0.23-0.26g. These authors also studied a box model to test the delivery system of 5% by weight colloidal silica for feasibility in field use. The use of low-head injection wells and extraction wells created a hydraulic gradient for the collodial silica to flow. Taking chlorine concentrations at sampling wells showed that the colloidal silica had been uniformly permeated. The compressive strength samples taken from the box model support the unconfined compressive strength of 5% by weight colloidal silica strengths documented by Gallagher and Mitchel (2002).
Full scale field testing was undertaken by Gallagher, Conlee and Rollins (2007), who successful demonstrated, that with the use of hydraulic gradient created by an extraction well and low pressure injections well 8% by weight colloidal silica can permeate a layer of liquefiable sand. When subjected by a dynamic load there was a decrease in the settlement that occurred in the colloidal silica grouted area compared with a nearby ungrouted area thus increased liquefaction resistance was shown.
As liquefaction can cause significant damage during earthquakes, and the current techniques and practises are not always suitable especially for existing developed sites, there is a need for new methods to be studied regarding liquefaction mitigation.
One potential new method of liquefaction mitigation is the use of nanoparticles as a passive site remediation technique. This technique involves slowly injecting the stabilisation material into a liquefiable subsoil and allowing the natural groundwater and hydraulic gradient to disperse the stabiliser uniformly throughout the soil (Gallagher and Mitchel, 2002). This technique allows the entire sites to be permeated instead of grout columns, which tend to form when using more traditional cement or silicate grouting.
During the review of literature information with a connection to liquefaction mitigation was found for the following nanoparticles:
a) Silica
b) Bentonite
c) Laponite
a) Various researchers have studied the effects of using colloidal silica as a passive site remediation technique (Díaz-Rodríguez et al., 2008; Gallagher and Mitchel, 2002; Gallagher, Conlee and Rollins, 2007; Gallagher, Pamuk and Abdoun, 2007; Mollamahmutoglu and Yilmaz, 2010). Colloidal silica is an aqueous dispersion of silica nanoparticles. Initially a low density and low viscosity solution forms which is similar to water. After a long induction period, where the viscosity stays low, a high viscous gel will form. Gallagher and Mitchel (2002) suggest that 5-10% by weight colloidal silica is expected to reduce liquefaction potential of saturated loose sands and still be economical (further results shown below).
b) The addition of the nanoparticle bentonite to a sand sample resistance to cyclic loading has also been studied (Gratchev et al., 2007). Gratchev et al. (2007) suggest that for bentonite to be successful in increasing liquefaction resistance over 7% by weight bentonite is necessary to be added to a sand sample.
c) Research has shown that aqueous dispersions of laponite also form a similar low viscous solution that transforms in to a high viscous gel over time (Bonn et al., 1999; Mongondry, Tassin and Nicolai, 2004; Mourchid et al., 1994). Bonn et al. (1999, p 7534) state that “[l]ow concentrations (typically a few percent by weight) of Laponite in water form suspensions with viscosity several orders of magnitude higher than of the solvent (water)”.
Following this research and discussion with the thesis supervisor, it has become clear that there is a gap in researching the affects of laponite on cyclic resistance of saturated sands. With laponite’s high viscosity at low concentrations it is hoped that laponite will be able to provide greater liquefaction resistance at low concentrations than other nanoparticles that have previously been studied.
Due to the large amount of research available for silica nanoparticles and its similarity to laponite an in-depth review of literature of the effects of its addition has been presented below. Research into colloidal silica has proven to be very successful and could provide a good framework for laponite testing. It is hoped that laponite will prove to be equally successful but with lower concentrations making it more economically viable.
Collodial Silica – In-depth Review
Gallagher and Mitchel (2002) performed cyclic triaxial tests on saturated loose sands grouted with 5, 10, 15 and 20% colloidal silica, and compared deformations against what was observed during cyclic triaxial tests for ungrouted samples. It was observed that both grouted and ungrouted samples had distinctly different deformation characteristics, ungrouted samples experienced very little strain prior to liquefaction. Though it was found that once liquefaction occurred that large strains quickly arose and the sample would collapse within a few additional loading cycles. Also observed was that as the cyclic stress ratio (CSR, maximum cyclic shear stress to initial effective confining stress) increased, the number of cycles that occurred before failure decreased. Comparing these results to grouted samples Gallagher and Mitchel (2002) noticed that for the same relative density and CSR, grouted samples experienced much less strain and none of the samples were found to have failed. Samples treated with lower concentrations of collodial silica experienced greater strains than those with higher concentrations. All the grouted samples returned to the original shape after loading had ceased.
Gallagher and Mitchel (2002) also tested the effect of curing time on cyclic resistance. It was noted that as the curing time increased, the strains reached during cyclic loading decreased. It is thought that this happens due to the fact that even after gelation occurs bonds in the collodial silica are still forming and these bonds are what increases the deformation resistance. These authors do hypothesize that this improvement to cyclic resistance will taper off with time. Unconfined compression tests were also undertaken in this study and showed that as the silica concentration increased within a sample the unconfined compression strength also increased. After cyclic loading had occurred unconfined compression tests where also undertaken to establish whether there had been a loss in compressive strength due to cyclic loading. It found that the loss in compressive strength varied with strains the sample had experienced during cyclic loading. As a general trend, the larger the strains experienced during cyclic loading, the greater the reduction in strength. It is suggested that this occurs as bonds will be broken during cyclic loading, greater levels of strains will therefore create larger numbers of broken bonds.
Testing done by Díaz-Rodríguez et al. (2008) and Mollamahmutoglu and Yilmaz (2010) support the findings of Gallagher and Mitchel (2002) in that these researchers also found that the addition of colloidal silica can significantly increase cyclic resistance. Diaz-Rodrigues et al. (2008) state that for a given initial relative density and initial vertical pressure, liquefaction resistance is significantly greater for colloidal silica grouted samples in comparison with ungrouted samples. In general, sand with higher initial relative densities has a greater cyclic resistance for both treated and untreated samples (Díaz-Rodríguez et al., 2008). Mollamahmutoglu and Yilmaz (2010) studied the effect of cyclic loading on the unconfined strength of colloidal silica grouted sand. It was observed that performing cyclic triaxial tests, with CSR varying from 0.13 to 0.52 for 1000 cycles, that no failures occurred. Comparing the unconfined strength before loading to the unconfined strength after loading showed an insignificant decrease in strength of around 10%. Previous studies have shown that ungrouted saturated sand samples lose their strength and liquefy with a CSR of 0.26 in around 13 cycles (Gallagher, 2000 cited in Mollamahmutoglu and Yilmaz, 2010 p.346), therefore the findings shown by Mollamahmutoglu and Yilmaz (2010) compare favourably with the results by Díaz-Rodríguez et al. (2008) and Gallagher and Mitchel (2002).
In a study by Gallagher, Pamuk and Abdoun (2007) a centrifuge model was undertake and subjected a loose density sand sample grouted with 6% collodial silica to two in-flight shaking events with 0.2g and 0.25g maximum acceleration respectively. During both shaking events no liquefaction was found to take place. The deformations were recorded throughout the shaking events and a maximum shear strain of 0.5 and 1% were recorded for the shaking events respectively. Vertical strains were measured to be 0.3% and 0.1%. On previous centrifuge model studies using untreated sand samples, vertical strains of 2-2.5% were recorded for shaking events with 0.23-0.26g. These authors also studied a box model to test the delivery system of 5% by weight colloidal silica for feasibility in field use. The use of low-head injection wells and extraction wells created a hydraulic gradient for the collodial silica to flow. Taking chlorine concentrations at sampling wells showed that the colloidal silica had been uniformly permeated. The compressive strength samples taken from the box model support the unconfined compressive strength of 5% by weight colloidal silica strengths documented by Gallagher and Mitchel (2002).
Full scale field testing was undertaken by Gallagher, Conlee and Rollins (2007), who successful demonstrated, that with the use of hydraulic gradient created by an extraction well and low pressure injections well 8% by weight colloidal silica can permeate a layer of liquefiable sand. When subjected by a dynamic load there was a decrease in the settlement that occurred in the colloidal silica grouted area compared with a nearby ungrouted area thus increased liquefaction resistance was shown.
References
Bonn, D., Kellay, H., Tanaka, H., Wegdam., and Meunier, J., 1999. Laponite: What Is the Difference between a Gel and a Glass?. American Chemical Society, Issue 15, pp. 7534-7536.
Díaz-Rodrígues, J. A., Antonio-Izarraras, W. M., Bandini, P. and López-Molina, J. A., 2008. Cyclic strength of natural liquefiable sand stabilized with colloidal silica grout. Canadian Geotechnical Journal, Issue 45, pp. 1345-1355.
Gallagher, P.M., 2000. Passive Site Remediation for Mitigation of Liquefaction Risk. PhD thesis, Virginia Polytechnic Insitute and State University, Blacksburg, VA
Gallagher, P. M., Conlee, C. T. and Rollins, K. M., 2007. Full-Scale Field Testing of Colloidal Silica Grouting for Mitigation of Liquefaction Risk. Journal of Geotechnical and Geoenvironmental Engineering, 133(2), pp. 186-196.
Gallagher, P. M. and Mitchel, J. K., 2002. Influence of colloidal silica grout on liquefaction potential and cyclic undrained behaviour of loose sane. Soil Dynamics and Earthquake Engineering, Issue 22, pp. 1017-1026.
Gallagher, P. M., Pamuk, A. & Abdoun, T., 2007. Stabilization of Liquefiable Soils Using Colloidal Silica Grout. Journal of Materials in Civil Engineering, 19(1), pp. 33-40.
Gratchev, I. B., Sassa, K., Osipov, V.I., Fukuoka, H. and Wang, G., 2007. Undrained cyclic behavior of bentonite–sand mixtures. Geotechnical and Geological Engineering, Issue 25, pp. 349-367.
Mollamahmutoglu, M. & Yilmaz, Y., 2010. Pre- and post-cyclic loading strengths of silica grouted sand. Geotechnical Engineering, Issue GE6, pp. 343-348.
Mongondry, P., Tassin, J. F. and Nicolai, T., 2004. Revised state diagram of laponite solutions. Journal of Colloid and Interface Science, pp. 397-405.
Mourchid, A., Delville, A., Lambard, J., Lécolier, E. and Levitz, P., 1995. Phase Diagram of Colloidal Dispersions of Anistropic Charged Particles: Equilibrium Properties, Structure, and Rheology of Laponite Suspensions. American Chemical Society, Issue 11, pp. 1942-1950.