The purpose of this paper is to clarify some of the mechanisms and phenomena relevant to description of the debonding process between ribbed reinforcement and concrete.

Before discussing the rationale of the study, it would be useful to introduce the basic phenomena of debonding. A typical case of a debonding problem is the pull-out of a reinforcing bar from a plain concrete specimen (Figure 1.a). While this illustration may seem to oversimplify the debonding process, it is actually a basic problem of debonding, and in many respects, close to practical application.

As shown in Figure 1.a, at the concrete surface of the pull-out specimen, where the reinforcement bar is leaving the concrete, the abrupt change of elastic conditions results in significant stress concentrations. Due to these accumulated stresses, the debonding process will start at the surface and extend inside the specimen along the reinforcing bar. The shear stresses will be largest close to the reinforcing bar, thus the debonding will start on the interface between the concrete and the reinforcing bar. The crack patterns resulting from this process depend on the interface geometry and the fracture mechanical properties of the interface and the surrounding concrete; furthermore, these different crack patterns do not form independently from one another, but interact through complicated non-linear mechanisms.

The simple pull-out problem depicted in Figure 1a can be generalized to the crack in a concrete beam in bending shown in Figure 1.b. Once the concrete is fully cracked, i.e., when no stresses are transferred between the two crack faces, the opening of the crack is controlled solely by the pull-out of the reinforcement. If no debonding is taking place, the crack faces close to a reinforcing bar are prevented from moving apart, i.e., the crack cannot open; it is only able to open in case of a local pull-out, and even then only around a reinforcing bar. Hence, the pull-out of a single reinforcing bar from a concrete specimen is essential to an understanding of the opening of bending cracks in concrete beams.

Since simple pull-out seems to reflect some of the basic phenomena in formation of cracks in real concrete structures, it merits further investigation. The purpose of the present study is to shed light on this phenomenon through investigation of a pull-out problem.

Experiments performed by other researchers, have clarified some of the significant features of the debonding process. It is generally accepted that two types of cracks appear: cone-shaped cracks and longitudinal splitting cracks, both of which start at the interface or close to the interface between concrete and reinforcement (Figure 2.)

Figure 2. Internal cone-shaped cracks and longitudinal splitting cracks

One of the aims of this paper is to present the adaptation of an existing technique for exact identification of cracks under load for the pull-out-induced cracks. The use of The technique, known as Molten Metal Injection method (Nemati 1994), has until now, been limited to plain concrete specimens in compression. However, the technique has been successful in accurate identification of cracks and pores down to microscale, and it is considered to be the best possible solution for identification of the complicated pull-out induced crack patterns.

Using this experimental technique, it is possible to investigate crack patterns and crack widths, and to study the process of creation of different types of cracks by investigating the crack patterns at different stages of pull-out. For instance, it is possible to study in detail the interaction between cone-shaped and longitudinal splitting cracks.

Goto (1971) studied experimentally the crack formation in axially loaded specimens. The specimens where loaded axially in tension, and cracks where observed at the surface (primary cracks). The formation of internal cracks related to the debonding of the reinforcement was investigated by injection of ink in narrow holes parallel to the reinforcing bar. Later, the specimens were cut in half along planes including the bar axes.

Goto (1971) found that numerous internal cracks formed around the deformed bars. At steel stresses less than 100 N/mm², these internal cone-shaped cracks are initiated around the ribs close to the primary cracks. When steel stresses are increased, the internal cracks develop further from the primary cracks at almost every rib (observed for deformed bars with lateral ribs). The internal cone-shaped cracks are formed with their apexes near the bar lugs and with their bases generally directed towards the nearest primary cracks. The angles of the internal cracks are generally within the range of 45 to 80 degrees to the bar axis, the angle being larger the longer the distance is to the primary crack. Through formation of these cracks, the concrete around the reinforcing bars form a mechanism like a comb with slanted teeth's as shown in Figure 3.

Goto (1971) also observed longitudinal splitting cracks, which were initiated near the bar at the faces of the primary cracks and extending towards the outside of the specimen. He suggests that both the action on the front faces of deformed bar lugs and deformation of the comb-like mechanism formed by the cone-cracked material can be the major causes for formation of longitudinal cracks. The latter cause would indicate that cone-shaped cracks develop before longitudinal splitting cracks, and that the splitting cracks are controlled by geometrical non-linearities caused by deformation of the comb-like mechanism formed by the cone-cracked material.

Figure 4. 4a (left) shows how the radial components of the bond forces are balanced against tensile stress rings. Figure 4b shows the distribution of tangential stresses around a deformed reinforcing bar (Tepfers 1973).

Based on the ideas from Tepfers (1973), Reinhardt and van der Veen (1991) a model has been established for concrete as a strain-softening material (Reinhardt and van der Veen 1991). They used both a power function and an exponential function as softening functions. The results obtained by using these models were in the area between a plastic material and a partly cracked elastic material. Furthermore, the models taking strain-softening into account lead to results which agree well with the experimental findings.

The test equipment to be used for the pull-out experiments was originally created to preserve the cracks under applied compressive load. The experiments to be carried out involve four procedures: (i) reinforced concrete cylinder casting and preparation, (ii) pull-out specimen fabrication, (iii) pull-out crack induction, and (iv) molten metal injection and solidification. The third and fourth procedures will be carried out simultaneously. In order to preserve the microstructure and cracks due to pull-out in reinforced concrete specimens under applied pull-out load, the cracks will be filled with a liquid metal alloy called Wood's metal during pulling of the reinforcing bar, and the alloy will be solidified in place before releasing the load. Wood’s metal is a fusible alloy which in the liquid phase is nonwetting, with an effective surface tension of about 400 mN/m. It consists of 42.5% bismuth, 37.7% lead, 11.3% tin, and 8.5% cadmium. It has a melting point range from 71°C to 88°C, and is solid at room temperature. Wood’s metal has a Young’s modulus of 9.7 GPa and a density of 9.4 g/cm³. The advantage of such an alloy is that it can be injected into pull-out-induced cracks at the desired stress level, then solidified at any stage of the experiment to preserve, in three-dimensional form, the geometry of the cracks induced at any given stage of the experiment.

The equipment used for this research consists of five pieces: pedestal, vessel, piston, top cap, and heater. Figure 5 shows the schematic diagram of the test assembly. The reinforced concrete specimen to be tested for the pull-out test will be assembled inside a special pull-out device. It consists of two equal size brackets, one of which is stationary and resting on the pedestal with the reinforced concrete specimen attached to it, and the other one is connected to the reinforcing bar and transfers the applied load in compression to the reinforcing bar, causing it to be pulled out of the concrete. The dimensions of the pull-out device is 203 mm height and 102 mm diameter. Figure 6 shows a schematic diagram of the reinforced concrete specimen pull-out device.

After fabrication of the reinforced concrete cylinder inside the pull-out device, the specimen will be dried in an oven at a temperature of 40°C. Removal of the moisture and preheating the specimen, ensuring that the molten metal alloy could penetrate into pores and cracks deep within its core without solidifying prematurely. The pull-out device will then be placed on the pedestal inside the vessel and the piston will be placed on top of it. The top cap will be left open resting on three wedges a short distance from the top of the vessel so that the molten metal could be poured through the gap using a funnel.

At this point, a minimum load will be applied to the piston to prevent the pull-out device from floating after the Wood’s metal is poured in. Once the pull-out device is fully submerged in the molten metal the top cap will be dropped by removing the wedges and then tightening the bolt to the vessel. To monitor the temperature, a thermocouple will be inserted into a pre-drilled hole on the top cap. The heater will then be placed around the assembled system with a special noncombustible board placed on top to prevent heat convection and thus uniform heating of the test assembly. The molten metal will be driven into voids and fractures by pore pressure provided using nitrogen gas. The alloy can be solidified at any stage of the experiment to preserve the geometry of pull-out induced cracks as they exist under load. With a surface tension of 400 mN/m, the alloy could penetrate into flat cracks with apertures as fine as 0.08 microns under a pore pressure of 10 MPa. Such a technique allows one to observe the geometry of the pull-out cracks induced at any given stage of the experiment by injecting the alloy into pull-out-induced cracks at the desired stress level, then solidify it at any stage of the experiment to preserve, and facilitate observing the cracks in three dimensions. After the metal is thus solidified, each of the cylinders will be sectioned into smaller specimens. The specimens will be observed using optical microscope and Scanning Electron Microscopy (SEM).

The heat will be supplied in three stages. Starting at room temperature, the heat will be ramped up to 50°C and held at that temperature for 10 minutes. Then the temperature will be ramped up to 75°C and held there for an additional 10 minutes. The final stage involves ramping the temperature up to a target of 96°C for a period of 15 minutes and holding it at that temperature until the heat is no longer needed. The ceramic heater will be placed around the vessel to liquefy the Wood’s metal inside and to maintain a constant temperature throughout the experiment. This temperature will, in turn, be monitored by a thermocouple that is attached to the side of the top cap. Figure 7 shows a schematic setup for the testing.

Figure 7. Schematic diagram of the test assembly

With the internal temperature thus established and maintained at 96°C, a vacuum Will be applied to the vessel and will be kept constant for at least 30 minutes. The vacuum removes any air that is trapped in the reinforced concrete cylinder when it is assembled inside the vessel. An the desired pull-out stress, the vacuum will be removed.

Finally, in order to saturate the induced cracks with the molten metal, nitrogen pressure will be applied to the top of the vessel. It will be controlled by a high-pressure regulator on a bottle of nitrogen. A nitrogen pressure of 10 MPa should applied to the molten metal as the pore pressure, which will be kept constant throughout the tests and which will not alter the effective stresses on the reinforced concrete cylinder.

The pull-out stress of interest shall be kept constant for 2 hours to allow the liquid metal to penetrate into the fractures. Afterwards, fans will be used to cool the vessel down to room temperature and to expedite solidification. Approximately 3 hours elapses between the time pore pressure is applied and the period during which the metal is allowed to solidify.

After solidification of the metal, the reinforced concrete specimens will be extruded and each of the cylinders will be sectioned along its long axis, using oil as a coolant. Smaller specimens are then extruded along the reinforcing bar for examination using scanning electron microscopy (SEM) in conjunction with an image analyzer.

The samples will be examined in the SEM with backscattered electrons (BSE). The images thus acquired by the SEM are transferred to an image analyzer and digitized into an array of 512 ´ 512 pixels with 255 gray levels. a typical gray level BSE image is shown in Figure 8 (this image is obtained from the compression tests on plain concrete).

From the histogram of the distribution of gray levels in a BSE image, the threshold value for discriminating the areas of Wood’s metal from other components in the image can be selected (Figure 9). Once the other components in the BSE image, i.e., cementitous phases, the reinforcing bar, and aggregates are removed, and after careful image manipulations, a skeletonized binary image of the cracks is obtained, which represents the Wood’s metal, as shown in Figure 10. From the binary image of Figure 10 various crack identification and measurements can be performed.

A special experimental technique has been developed which makes possible the preservation of the pull-out-induced cracks in reinforced concrete as they exist under applied pull-out stresses, such as cone-shaped cracks and longitudinal splitting cracks. This technique involved injecting a molten-metal alloy into the induced pull-out cracks and solidifying it before unloading. By analysis of BSE images, various measurements of the crack orientation, density, and length can be performed. This method could clarify the formation and interaction of cone-shaped cracks and longitudinal splitting cracks and the order of their formation.

Goto Y. (1971): Cracks Formed in Concrete Around Deformed Tension Bars. Journal of the American Concrete Institute Vol. 68, No. 4, April 1971, pp 244-251.

Nemati, K. M. (1994): "Generation and Interaction of Compressive Stress-Induced Microcracks in Concrete." Ph.D. Thesis, University of California at Berkeley.

Reinhardt H.W. and van der Veen C. (1991): Splitting Failure of a Strain-Softening Material due to Bond Stresses. In Applications of fracture Mechanics to Reinforced Concrete, Ed. A. Carpinteri, Elsevier Sci. Publ. 1991.

Tepfers R. (1973): A theory of bond applied to overlapped tensile reinforcement splices for deformed bar. Dissertation, Publication 73:2, Chalmers University of Technology, Division of Concrete Structures, Göteborg, Sweden, 1973, 328 pp.

Figure 5. Transition zone cracks