Laboratory preparation of clay specimens with low-pressure mixing

Shan H., Kerenyi K., Guo J, Xie Z., and Shen J.

ICSE6-302

Laboratory preparation of clay specimens with low-pressure mixing Haoyin SHAN1, Kornel KERENYI2, Junke GUO3, Zhaoding XIE4, J. Jerry SHEN4 1

Hydraulics Research Engineer, Turner-Fairbank Highway Research Center GENEX Systems, 6300 Georgetown Pike, McLean, VA 22101 [email protected] 2

Hydraulic Research Program Manager, FHWA Hydraulic R&D Program Department of Transportation, 6300 Georgetown Pike McLean, VA 22101 [email protected] 3

Associate Professor, Department of Civil Engineering, University of Nebraska-Lincoln, Omaha, NE 68182 [email protected]

4

Research Engineer, Turner-Fairbank Highway Research Center GENEX Systems, 6300 Georgetown Pike, McLean, VA 22101 [email protected], [email protected]

Abstract: Cohesive soils form the foundation of vital infrastructures like bridge piers, levees and embankment dams. Prototype soils vary widely in composite and gradation, thus their properties are difficult to control. While laboratory prepared artificial clay specimens can separate parameters that control the erosion of cohesive soils and facilitate judgment on effect of each parameter using ex-situ erosion testing devices. Porcelain Kaolin specimen is prepared for an ex-situ scour testing device (ESTD) developed in the Turner Fairbank Highway Research Center (TFHRC). This device simulates the process of underwater erosion, and aims to quantify the principle of erosion in cohesive soils around bridge piers. The device needs cohesive soils that have similar characteristics to represent the prototype soils. Standard proctor method was tested unsatisfied since the prepared specimen quickly slaked even in stilled water. Consolidation method in air improves the resistance to collapse to a certain extent. But a regular ESTD test runs about one hour, and the boundary of a specimen would crack in the soaking condition. That incurred a difficulty to specify how much of the soil loss is due to eroding not slaking. Thus, a low-pressure tank was designed and fabricated to reproduce clay specimens as those located underwater. It is of 30 cm long, 10 cm wide and 30 cm high. It is capable to execute mixing under a 7000-Pa-pressure condition. The slurry was mixed inside a stainless steel tube sitting in the tank. The dry clay powder was pre-mixed with water, poured into the tube. Then the slurry was re-mixed under lowpressure condition and pre-compacted to eliminate air. The pre-compacted specimen was then transferred into a consolidation cell. Dead weights applied consolidation to the specimen. Crumb tests of prepared clay specimens show satisfactory improvement on resisting slaking.

Keywords: Cohesive soils, low-pressure mixing, erosion, crumb test, slaking. I

INTRODUCTION

Cohesive soils, i.e. clay and silts, and rocks can be defined as chemical gels compared to non-cohesive sands which are physical gels (Annandale 2005). They both compose materials cover the top few hundred meters of the Earth. In chemical gels, particles are connected by fixed bonds, i.e. hydrogen bond, van der Waals bonds, and cation bonds (Partheniades 2009). These bonds provide primary erosion resistance to water flow instead of submerged particle weight in physical gels. Since these bonds develop in different situations in the formation of a specific cohesive soil, it is more challenging to quantify them than to quantify erosion resistance of sand by directly calculating the submerged weight of an individual particle. As the erosion process in cohesive soil is complex, currently engineers adopt scour equations developed in noncohesive soils (Briaud et al. 2004), which potentially results in very conservative bridge pier design since the erosion rate in cohesive soils are much lower than non-cohesive sands and it takes a long time to reach equilibrium scour with cohesive bed material. This warrants more research on erosion of cohesive soils. Arulanandan (1980) summarized the erodibility of cohesive sediments are affected by the following characteristics: (1) placement conditions, rheology and stress history of the soil; (2) particle size distribution of non-cohesive portion; (3) clay content and type; (4) organic matter and cementing agents content; (5) soil 1

Corresponding author

57

Laboratory preparation of clay specimens with low-pressure mixing

Shan H., Kerenyi K., Guo J, Xie Z., and Shen J.

temperature and water content; and (6) chemical composition of the pore fluid. Given that prototype soils vary widely in composite and gradation, thus their properties are difficult to control. It might be much feasible to use remolded cohesive specimens to separate these parameters. Traditional geotechnical preparation of soil specimens follows procedures from the compaction test (ASTM D698 - 07e1). This method keeps the moisturized clayey mixture stand (maturing) 16 hours before compacted in standard 10.16-cm-diameter mold. Then it uses a 24.4 N rammer blowing three layers of mixture into a specimen with 25 blows for each layer. It works well at low water content. The French Institute of Sciences and Technology for Transportation, Development and Networks (IFSTTAR, Haghighi et al. 2012) proposed using only the maturing step of compaction test to prepare specimens at higher water content when mixture is too plastic to be compacted. It was found that prepared specimens slake with releasing of air bubbles in still water at water content up to 42%. Sharif (2003) used natural consolidation method to prepare 8.9-mm-diameter cylindrical Kaolin clay specimens on top of a piston inside a soil holder for flume test. He achieved pure Kaolin clay specimens with a surface bulk density of 1200 kg/m3 naturally consolidated for 48 hours from slurry with a soil mass concentration of 350 g/L. Apparently, the density increases with the soil depth. At a depth of 5 cm, the soil density increases to 1420 kg/m3. This is very realistic and close to the natural clay. This specimen worked quite well for flume test. However, it takes several months or years for the specimen to develop proper strength to maintain its shape. Shan et al. (2011) prepared Kaolin clay specimens in a consolidometer cell in air for an ex-situ scour testing device (ESTD). The ESTD is designed and developed at the Hydraulics Laboratory in the Turner Fairbank Highway Research Center of the Federal Highway Administration. It simulates underwater erosion by generating a horizontal near-bed flow. The flow has a log-law profile similar to that of open channel flow. The moisturized mixture was steadily compacted in a 63.5-mm-diameter consolidometer cell with a compacting pressure of 65 KPa for three hours. The regular consolidometer method (RCS) shortens the preparation. However, if the extruded clay specimen was immediately immersed into water, it again slakes with releasing of air bubbles. Berghager and Ladd (1964) explained this phenomenon was because the pressurized air bubble formed in the consolidation would swell and burst. This destroyed the connection among soil particles, thus the soil collapsed. Shan et al. (2011) leaves such specimens in a closed zip bag for two hours, the soil demonstrated a stronger resistance to still water to some extent. The inner structure of the soil revealed voids with different sizes. These voids were occupied by either water or air. Since clay has a low permeability, trapped air bubbles inside wet clay is not able to escape within three hours. If pressurized air bubbles accounts for slaking of specimens, it is vital to eliminate them from the preparation. Thus, the authors propose a new fast method to prepare clay specimens to reduce air bubbles inside soil specimens. The new method mixes the clay powder and water under low-pressure condition. After pre-compaction to reduce the air bubble, the mixture is consolidated in a consolidometer cell in air. The transfer of mixture from the vacuum condition to the consolidometer cell is assumed not absorbing much air into the mixture. The air can be mostly eliminated from the final specimen. II

LOW-PRESSURE MIXING

Consolidation for several hours in air cannot eliminate air bubbles trapped in the slurry during the mixing process. It is straightforward to finish the mixing under vacuum condition to prevent the air from filling the void. However, the boiling temperature of water is below 0°C at extreme low pressure. A moderately lowpressure can be achieved in laboratory condition to keep the water in liquid state. The regular room temperature of 25°C corresponds to a low pressure around 3200 Pa. The proposed low pressure tank was built with plexiglass panels with a thickness of 17 mm. Figure 1 illustrates the tank and other components for soil preparation. The tank is 30 cm long, 10 cm wide and 30 cm high. The slurry is mixed inside a stainless steel tube sitting in the tank. The tube has an inner diameter of 63.5 mm, which is the size required by the ex-situ scour testing device (Shan et al. 2011). The tube has a wall thickness of 3 mm, and it is nested to the tank bottom through a 3-mm groove. A laboratory electric mixer mixes the dry Kaolin powder with distilled water into slurry. The mixer shaft is 30 cm long so it shakes and hits the steel tube when rotating. A needle is screwed on the tank bottom to supply a constraint to the mixer shaft at the propeller end. These propellers have a diameter of 58 mm. More propellers can be mounted to the shaft to prepare higher soil specimen.

58

Laboratory preparation of clay specimens with low-pressure mixing

Shan H., Kerenyi K., Guo J, Xie Z., and Shen J.

Figure 1: Sketch of the low pressure tank and soil preparation components.

The slurry mixing in the low pressure tank is the most challenging part of the whole preparation process. A vacuum feedthrough is used to solve the dilemma of rotating and sealing. It uses hydrocarbon-based ferrofluid, specifically optimized for introducing rotary motion with a magnetic liquid hermetic seal in most standard environments. Its inner shaft is rotary, while the outer solid part keeps it stationary. The feedthrough has an O-ring to help the sealing of the tank cover. The feedthrough is a rigid unit, which does not allow reciprocating motion of the mixer shaft. The vacuum condition in the tank is obtained by a vacuum pump. The pressure meter on the pump is not sufficiently precise to show the low pressure. The pressure inside the tank was determined by the method of water boiling points under different low pressures. Tests with water of different temperatures reveal that 3839 °C water was boiling while the vacuum pump worked stable. This corresponds to a low pressure of 66407010 Pa, 7% of the atmosphere pressure. III

SOIL PREPARATION PROCEDURE

Vacuum the dry clay powder before adding water to it will perfectly eliminate air bubbles that might be trapped into soil specimens in later preparation. However, it is difficult to get slurry uniformly mixed inside the stainless steel tube since the mixer shaft and propellers are fixed. Different propellers with various shapes had been tested, and the specific propeller with certain shape needs more research to make vacuum mixing realistic. Currently, the uniform mixing of the slurry is finished in air. The water content of the field clay is unavailable. The authors defined an initial water content of 67% to provide sufficient water for saturated specimens. Desired amounts of dry Kaolin powder and distilled water were uniformly mixed by an electric mixer in a standard 6-inch-diameter proctor for one minute. The mixture was quickly transferred into the stainless steel tube, and a porous stone was placed at the bottom. The bottom of the lowest propeller was attached with the needle on the bottom. Then all screws were tightened on the tank cover, and it was ready to be vacuumed. The slurry was re-mixed for five minutes. Then the mixing shaft was replaced with a loading shaft to pre-compact the slurry. The loading shaft had a pad on the end touching the slurry. A saturated filter paper was placed between the loading pad and the slurry. The paper helps draining during the pre-compaction. The pre-compaction took about 5 minutes in low-pressure condition to eliminate air in the slurry as much as possible. The slurry was taken out of the tank and put into a consolidation cell. Depending on the compaction pressure, different dead weights (10 kg each) can be placed on its loading pad to compact the slurry in air for three hours. The compaction time was also affected by the final water content and density of the specimen. After the compaction, the specimen was extracted and trimmed to the desired height (15 mm) with a wire saw. The specimen was fixed onto a porous stone and mounted in the ESTD for erosion testing. With an initial water content of 67%, the authors prepared several specimens following the low-pressure mixing method (LPS, Figure 2.a) and the regular consolidation method (RCS, Figure 2.b). Natural mud (NYS, Figure 2.d) from New York port was collected, deposited, and dried naturally. After erosion, the

59

Laboratory preparation of clay specimens with low-pressure mixing

Shan H., Kerenyi K., Guo J, Xie Z., and Shen J.

remaining soil specimen made of Armoricaine Kaolin by naturally maturing method was collected (AAE, Figure 2.c). IV

RESULTS AND DISCUSSION

EPK Kaolin has a median size of 1 micron, while the Armoricaine Kaolin is coarser, having a median size of 2.5 microns. This makes different specific surface area for both materials. The EPK Kaolin is 22 m2/g, while the Armoricaine Kaolin is 9 m2/g. Nevertheless, they have the same chemical composition (Chevalier and Haghighi 2011). All six aforementioned specimens were broken by hand. Figure 2 shows their fracture planes. As can be seen, the specimen by the low-pressure mixing method (Figure 2.a) has the least voids. The naturally deposited New York mud and regular consolidated specimen have considerable voids, and two specimens by the naturally mature method have the most voids. These voids explain the slaking phenomenon of specimens.

Figure 2: Regular images of 6 specimens prepared in the laboratory.

Table 1 gives properties of three specimens. The wet and dry densities of New York mud (NYS), regular consolidation method specimen (RCS), and Armoricaine Kaolin specimen after erosion were not measured. Soil LPS EPK ARM

Wet density (kg/m3) 1694 1678 1716

Dry density (kg/m3) 1136 1150 1210

Water content (%) 49.1 46 41.5

Porosity ϕ (%) 57 57 54

Saturation degree (%) 98 93 93

Table 1: Properties of soil specimens prepared with different methods.

The saturation degree and the porosity in Table 1 give the volume percentage of air in the whole soil specimen. It was 1.14% for the LPS and 3.78% for the ARM. The LPS has only 1/3 of the air bubbles inside than that of the ARM. The larger air volume might increase the exchange of water between the outer flow and the inner soil structure by infiltration (Shan 2010). The wet densities are much larger than that of specimens prepared by Sharif (1420 kg/m3). IV.1

Crumb test

Crumb tests were done for specimens ARM, RCS and LPS. These specimens were immersed in still water right after they were prepared. Figure 3 illustrates the beginning and last images of these specimens. Two upper left corner images are those of ARM. Two lower left corner images are those of RCS, and two right images are those of LPS. The ARM specimen was broken in 144 minutes. The RCS was a little stronger than the ARM. It did not collapse like the ARM. Nevertheless, it showed cracks around 950 minutes. The LPS is the strongest among these three specimens. It stayed unbroken in water for 4 days. Only little cracks showed on the surface. These cracks were much narrower than that on the RCS. Both RCS and LPS had dissolution

60

Laboratory preparation of clay specimens with low-pressure mixing

Shan H., Kerenyi K., Guo J, Xie Z., and Shen J.

around them in the test. That might because the water content on the surface was quite low after the consolidation. After a certain time, the dissolution stopped.

Figure 3: Crumb test of three soil specimens prepared with different methods.

Figure 3 demonstrates that the low-pressure mixing method can prepare strong specimens resisting the infiltration of water due to the low air content inside the specimen. IV.2

Scanning electron microscope test

Figure 4 shows images of six soil specimens using a scanning electron microscope. They all have a magnifier of 800 times. The AAE specimen is the remaining soil after eroded of the ARM specimen. Comparison between the images of AAE and ARM shows the AAE has a continuous structure while the ARM has much more individual particles or flocs connected to each other by both face-to-face and edge-toface connection. The water flow eroded those individual particles or flocs with low bonding forces. The EPK and NYS have much finer particles compared to the ARM. Unlike the ARM, the EPK has a structure of steep hills with finer particles or flocs scattered on them. The NYS has a finer structure resembling vegetation in microscopic scale. In the upper right corner, there seems to be a large particle or flocs. Undulate shape exists in both RCS and LPS. Individual particles or flocs are seldom seen in the images. Images with larger magnifier show the flocculent structure inside the RCS and LPS specimens. This structure might indicate the fracture plane was developed by tensile failure. It means the connection between the original interfaces was much stronger than that of other three specimens.

Figure 4: SEM images of specimens prepared with different methods.

61

Laboratory preparation of clay specimens with low-pressure mixing

V

Shan H., Kerenyi K., Guo J, Xie Z., and Shen J.

CONCLUSION AND FUTURE RESEARCH

A low-pressure mixing method to prepare cohesive soil specimens for erosion testing devices in the laboratory was proposed. It was compared to other different preparation methods. Both crumb test and SEM images prove the low-pressure mixing method significantly improved the specimen capability of resisting slaking. This reproduces a realistic soil condition of submerged stream bed material. Future research includes finding the specific propeller to uniformly mix soil slurry under vacuum condition. Cohesive soil specimens prepared using the new method needs to be compared to field soils. The slaking mechanism needs to be clarified for specimens prepared by other methods. VI

ACKNOWLEDGEMENT

The authors want to thank Michael Gruber and Oscar Suaznabar for their help on the drawing. Special thanks go to Ms. Yuan Yao and the Chemistry Laboratory in the TFHRC for support of the SEM tests. VII

REFERENCE

Annandale, G. W. (2005). Scour technology – mechanics and engineering practice. McGraw-Hill. Arulanandan,K., Gillogley, E., and Tully, R. (1980). Development of a Quantitative Method to Predict Critical Shear Stress and Rate of Erosion of Natural Undisturbed Cohesive Soils. Technical report – US Army Engineering Waterways Experiment Station. ASTM D698 - 07e1. (2007). Standard Test Methods for Laboratory Compaction Characteristics of Soil Using Standard Effort. ASTM International, West Conshohocken, PA, 2003, DOI: 10.1520/D0698-07E01, www.astm.org. Briaud, J.-L., Chen, H. C., Nurtjahyo, P., and Wang, J. (2004). Pier and contraction scour in cohesive soils. NCHRP report 516. Texas A & M University, College Station, TX. Chevalier, C. and Haghighi, I. (2011). Progress report on TFHRC-LCPC collaboration in highway infrastructure. The French Institute of Sciences and Technology for Transportation, Development and Networks (IFSTTAR). Paris, France. Haghighi I., Chevalier C., Duc M., Reiffsteck P., Guédon S (2012). Improvement of hole erosion test and results on reference soils, submitted to Journal of Geotechnical and Geoenvironmental Engineering. Partheniades, E. (2009). Cohesive sediments in open channels – properties, transport, and applications. Elsevier. Shan, H. (2010), Experimental Study on Incipient motion of Non-cohesive and Cohesive Sediments, Ph.D. Dissertation, University of Nebraska-Lincoln, Lincoln, NE. Shan, H., Wagner, A., Kerenyi, K., Guo, J., and Xie, Z. (2011). An Ex-situ Scour Testing Device for erosion research of cohesive soils. Proceedings of the engineering mechanics institute 2011 conference, Wang, M., Bernal, D., Hajjar, J. and Cao, Y. Ed., Boston, MA, 1020-1027. Sharif, Ali R. (2003). Critical shear stress and erosion of cohesive soils. PhD dissertation. Department of Civil, Structural, and Environmental Engineering, University of New York at Buffalo.

62