Introduction
Cole and others (1985) described a system for mounting end caps on ice-core specimens that typically achieves end-plane parallelism within 0.5 μm/mm of specimen diameter. The importance of nearly parallel end planes for successful tensile testing has been discussed by Reference Lee, Lee, Currier, Lim and SchulsonLee and others (1984). (A successful test results in fracture normal to the tensile axis within the central third of a right-circular cylinder.) Using the procedure described here, approximately 45 right-circular cylinders of both Arctic sea ice and laboratory-grown saline ice have been prepared and successfully tested under uniaxial tension.
Procedure
The system and procedure described by Cole and others (1985) for attaching end caps to cored ice specimens was followed in the development of this procedure with one important exception. Instead of using end caps whose bonding surfaces had been machined with concentric grooves, end caps were used which had jute-backed carpeting bonded to their ice-bonding surfaces with epoxy resin. As discussed by Reference Cole, Cole, Gould and BurchLee and others (1984), the addition of a layer of jute-backed carpeting provides sufficient compliance at the ice/end-cap interface to prevent the development of large shear stresses in the end zones during testing. Minimizing shear stresses at the end zones prevents the need for reducing the cross-sectional area within the gage section. Additionally, the carpeting enhances the strength of the ice/end-cap bond.
An end cap with jute-backed carpeting is shown on the system in Figure 1. The procedural differences from those outlined by Cole and others are: (1) the carpeting is soaked in distilled water at 0°C until saturated; (2) a latex membrane is placed around the carpeted end cap and distilled water at 0°C is added to re-saturate the carpeting; and (3) once the ice is brought into contact with the carpeting, the water is allowed to freeze for approximately 1 h before removal of the specimen. The latex membrane keeps the water in place during freezing and accommodates freezing strains.
To measure accurately and control sample deformation during testing, extensometers are mounted on holders (Fig. 2) which attach directly to the ice. The holders were originally designed for use on fresh-water ice specimens and contain spring-loaded hinges for easy mounting (see Reference LeeLee, unpublished). The holders are frozen to the specimen using a small amount of cooled water. They were machined so that the springs are unstressed when the mounts are in place, thus preventing disturbance of the uniaxial stress state at their points of contact. Four spacer rods provide a consistent gage length and are removed prior to testing. The extensometers are mounted on rods extending towards the central region of the sample and measure the relative displacement of the holders during testing. While using these mounts, initial failure has never been observed to initiate near their contact points.
Results of the Procedure
The procedure has been used to test 60 specimens under tension at –10° C and at constant strain-rates of 10-3 s-1 and 10-7 S -1. The average end-plane parallelism of these specimens was 0.45 μ m/mm of specimen diameter with a standard deviation of 0.17 μ m/mm. The success rate was 75%. Figure 3 illustrates a typical case for a laboratory-grown saline ice specimen strained parallel to its growth direction.
Acknowledgements
We wish to thank Reference Cole, Cole, Gould and BurchD. Cole of the US Army Cold Regions Research and Engineering Laboratory, Hanover, New Hampshire, for use of the end-capping system. We also thank R. Roberts, of US Army CRREL, and E. Schulson, of the Thayer School of Engineering, for valuable Figure 3 assistance. The work was funded through a grant from the Ice Research Laboratory, Dartmouth College, which is funded by a consortium of the following firms and federal agencies: Amoco, Exxon, Mobil, Shell, Sohio, Minerals Mangement Service, Maritime Administration, and the National Science Foundation.