Magazine of Concrete Research, 2005, 57, No.

8, October, 463468

Concrete-to-concrete bond strength: influence of an epoxy-based bonding agent on a roughened substrate surface
E. N. B. S. Ju lio*, F. A. B. Branco and V. D. Silva*
University of Coimbra, Portugal; Technical University of Lisbon, Portugal

An experimental study was performed to evaluate the bond strength between two concrete layers, using different techniques for increasing the roughness of the substrate surface and a commercial epoxy-based bonding agent. A total of 40 slant shear half specimens and 40 pull-off half specimens first had the substrate surface prepared by wire-brushing, sand-blasting, chipping with a light jackhammer, or were left as-cast against steel formwork. Three months later, the bonding agent was applied and the new concrete was added. Pull-off tests and slant shear tests were performed to evaluate the bond strength in tension and in shear. Analysis of the results indicates that the application of an epoxy-based bonding agent does not improve the bond strength since the adopted method for surface preparation adequately increases its roughness.

Introduction
Adding new concrete to an existing concrete substrate is a usual technique for repairing and/or strengthening structures. The usual practice consists of first increasing the roughness of the substrate surface and frequently, the application of a bonding agent is also adopted. However, this procedure is empirically based and the efficiency of the second step has not been proved. The authors performed an experimental study with the objective of quantifying the influence of the application of a bonding agent on the bond strength between two concrete layers with different ages. The selected bonding agent was a commercial, widely used, twocomponent epoxy resin. The methods adopted to increase the roughness of the substrate surface, before the application of the bonding agent, were those most commonly used in practice. From the conclusions of this study, designers can choose the best preparation technique for their sub-

strate surface in order to obtain higher bond strengths at lower costs.

Previous research
There are some published works on adhesion of repairing materials to a concrete substrate where bonding agents are used.17 Nevertheless, the results obtained by different researchers are not always in agreement.36 Furthermore, due to the variability of the parameters that influence the bond strength, it is not possible either to generalise or to extrapolate the conclusions drawn. According to Garbacz et al.7 the adhesion in the repair system depends on the surface roughness of the concrete substrate, the presence of micro-cracks and the properties of the materials to be used for the repair. The authors state the increasing necessity of using a bond coat as the violence of surface treatment increases.7 Cleland and Long8 concluded that the principal function of a bonding agent is to develop a bonding bridge between the repairing material and the concrete substrate. Based on published works, Talbot et al.9 have stated that the use of a bonding agent reduces the variability of results. In terms of the characteristics of the bonding agent, Emmons4 states that it should be easily absorbed by the 463
0024-9831 # 2005 Thomas Telford Ltd

* Department of Civil Engineering, Faculty of Sciences and Technology, University of Coimbra, Portugal. Department of Civil Engineering, Instituto Superior Te cnico, Technical University of Lisbon, Portugal. (MCR 41246) Paper received 20 February 2004; last revised 21 March 2005; accepted 29 March 2005

 lio et al. Ju pore structure of the substrate and must be compatible with both the substrate and the repairing material. This author indicates three main types of bonding agents that are frequently used: cement-based slurries, epoxies, and latex emulsions.4 With regard to the bond strength, the perceived advantage of considering a bonding agent is not unanimous. Austin et al.5 have reported that bond coats can significantly increase the adhesion between the concrete substrate and the repairing material, remarking, however, that a misuse can lead to much lower strengths than without a bond coat. Cleland and Long8 also indicate that, for some repairing materials, the values of bond strength in tension are greatly reduced if no bonding coats are used. Emmons4 points out that adequate bonding may be developed by placing the repairing material directly against the prepared concrete substrate and that the use of an epoxy bonding agent may produce a vapour barrier, resulting in the failure of the bonding. Saucier and Pigeon3 concluded that casting new concrete directly on an old concrete base was better than using a cement slurry with a high water/ cement ratio as a bonding agent. test provides a sensitive means of assessing the effect of variations in resin formulation. The pull-off test (Fig. 2) was chosen to evaluate the bond strength in tension because it can be carried out in situ, which is contrary to the slant shear test.16,17 The adopted geometry for the slant shear specimens was a 20 cm 3 20 cm 3 40 cm prism with the interface line at 308 to the vertical. The specimens were tested under compression using the standard procedure for the testing of cubes or cylinders for compressive strength. The adopted geometry for the pull-off specimens was a 20 cm cube with the interface line at the middle. A core of 75 mm diameter was drilled into the added concrete and extending 15 mm beyond the interface into the substrate. A circular steel disc was bonded with an epoxy resin to the surface of the core. A tension force was applied to the disc with a commercial device at a steady rate of 0.05 MPa until failure occurred. For each situation considered, five slant shear specimens and five pull-off specimens were built as well as six standard specimens to characterise the compressive strength of the concrete substrate and of the added concrete (three cubes for each). All parameters that could influence the bond strength were kept constant, apart from the roughness of the substrate surface and the method to promote it: (a) con-

Experimental investigation
The following objectives were defined for this experimental research: (a) to quantify the influence of applying a bonding agent on the bond strength of the interface, considering different methods for increasing the surface roughness; and (b) to examine the correlation between the bond strength in shear and the bond strength in tension. The slant shear test (Fig. 1) was chosen to measure the bond strength in shear because, according to different researchers, this test is roughness sensitive.2,4,1013 In 1976, Kriegh14 published that the Arizona slant shear test was probably the most meaningful and discriminating test method available to select an epoxy resin. Furthermore, in 1978, Tabor15 declared that this

Fig. 1. Slant shear test

Fig. 2. Pull-off test Magazine of Concrete Research, 2005, 57, No. 8

464

Concrete-to-concrete bond strength crete substrate mix; (b) added concrete mix; (c) age of both concretes; (d ) temperature and relative humidity of the environment; (e) method used for cleaning the interface surface; ( f ) moisture condition of the interface surface; (g) bonding agent, including type and thickness; (h) method adopted for applying the bonding agent, including the time gaps between preparation/application and application/casting of new concrete; and ( i ) tests used to measure the bond strength. As a consequence of keeping the first four parameters constant, also the strength of both concretes and the differential shrinkage between them were kept constant. Preliminary tests were conducted to define the following parameters: concrete substrate mix; added concrete mix; and their ages.10 To avoid monolithic rupture modes, it was decided to adopt the same concrete mix for both the substrate and the added concrete, although this would never be the situation in the field. In order to obtain shrinkage deformations between the substrate and the added concrete, their ages were set, respectively, at 112 days and 28 days, at the time of the test. A concrete mix with an estimated compressive strength of 50 MPa was used. The constituents of this concrete were (per m3 ) 360 kg of type I:32.5 Portland cement, 1.6 l of a modified lignosulphonate admixture, 168 l of water, 813 kg of siliceous sand with 2.84 fineness modulus, 469 kg of limestone crushed aggregates with 6.16 fineness modulus and 567 kg of limestone crushed aggregates with 6.93 fineness modulus. The adopted roughening techniques were the most commonly used in practice. The following situations have been considered: (1r) and (1) surface cast against steel formwork; (2r) and (2) surface prepared with a steel brush (Fig. 3); (3r) and (3) surface partially chipped (Fig. 4); and (4r) and (4) surface treated with sand blasting (Fig. 5). For situations (1) to (4), the substrate surface preparation was followed by the application of a bonding agent (Fig. 6). A commercial epoxy resin that is widely used in practice was chosen. The chemical composition of this product is: bisphenol A-epichlorohydrin resin, with 4,4-methylene bis (cyclohexylamide), as curing component. According to the manufacturer, this product has 90 MPa of compressive strength and 45 MPa of bending strength. In all cases the bonding agent was applied immediately before casting the strengthening layer of concrete.

Fig. 3. Surface prepared with steel brush

Results and discussion

Table 1 presents the average values of the compressive strength of the original concrete and of the added concrete, used in all eight situations. These results were obtained with tests performed on standard cubic specimens. The average value of the bond strength in shear, determined with the slant shear test is also given in
Magazine of Concrete Research, 2005, 57, No. 8

Fig. 4. Surface partially chipped

465

 lio et al. Ju
Variation coefficient: % Bond strength in shear: MPa Variation coefficient: % Bond strength in tension: MPa Added concrete: MPa Added concrete: MPa Original concrete: MPa

Slant shear test

Pull-off test

Tensile strength

Fig. 5. Surface treated with sand blasting

Compressive strength

Epoxy resin application

Fig. 6. Application of the epoxy resin on the substrate surface

Table 1 for each of these situations. The corresponding average value of the bond strength in tension, measured with the pull-off test, and the theoretical value of the tension strength of the added concrete are also indicated. It should be mentioned that the rupture mode, observed in all specimens, tested with both methods, was always an adhesive failure at the interface. For all five specimens of situation (1r), de-bonding of the core occurred when executing the hole and so the corresponding average value of the bond strength is not presented. As mentioned previously, with the exception of the 466

Substrate surface treatment

As-cast against

Wire-brushing

Partially chipped

Table 1. Test results

Situations considered

Magazine of Concrete Research, 2005, 57, No. 8

1r 1 2r 2 3r 3 4r 4

Sand-blasting

No Yes No Yes No Yes No Yes

50.95 50.08 49.66 50.88 51.40 49.69 50.60 50.80

47.42 45.22 46.11 49.38 45.46 46.61 45.14 45.30

3.81 3.75 3.78 3.97 3.69 3.86 3.58 3.64

2.40 1.92 2.24 1.47 1.93 2.65 2.08

14.17 13.54 12.05 7.48 19.69 6.42 21.63

1.30 11.20 10.67 12.63 6.24 11.16 14.13 11.57

33.85 7.95 8.90 15.44 20.67 11.47 8.56 2.59

Concrete-to-concrete bond strength surface roughness, all variables that could influence test results were kept constant, including the concrete mix, the concrete constituents, their ages, and all the laboratory procedures involved. However, it should be noted that the average value of the compressive strength of the concrete specimens was 50.51 MPa, for the substrate, and 46.33 MPa, for the added concrete (see Table 1). This difference was probably due to the fact that the first halves and the second halves of the composite specimens were cast during different times of the year and differences in temperature and relative humidity were registered in the laboratory. When analysing the results, it must be remembered that the compressive strength of the added concrete of situation 2 presented an average value (49.38 MPa) that was significantly higher than the average value used for reference. Examining the results of the slant shear tests (Table 1), it can be seen that the bond strength in shear was approximately constant, and slightly superior to 11 MPa, for all situations with epoxy resin application. The value of 12.63 MPa, corresponding to situation 2, may probably be justified by the fact cited above. Analysing the results of the pull-off tests (Table 1), it can be seen that values varied within a small range, between 1.93 and 2.51 MPa, for all situations considered with epoxy resin application. Noting that the upper limit corresponds to the situation with no preparation of the substrate surface, it is probable that this difference is inherent to the test itself and not to the roughening method used. The results indicate that there is no influence of surface roughness on the bond strength in shear, when an epoxy resin is used as bonding agent. The comparison between results of the slant shear tests (Table 1), with and without application of the epoxy-based bonding agent, demonstrates that the application of the latter on the substrate surface does not improve its bond strength if a surface preparation method is adopted that adequately increases its roughness. In fact, using sand blasting to prepare the substrate surface, the bond strength in shear, registered with the slant shear specimens, was 14.13 MPa without epoxy resin application, and 11.57 MPa with epoxy resin application. When the epoxy resin was applied to the substrate surface cast against steel formwork, partially chipped, wire brushed, and sand blasted, the visual aspect became identical. It can be said that the epoxy resin application eliminated the roughness differences between the substrate surfaces in these situations. Therefore, it is not surprising that the bond strength in shear of these situations was also identical. This may also explain the higher value obtained with the slant shear specimens treated with sand blasting without epoxy resin application relative to the value registered with the slant shear specimens treated with sand blasting followed by epoxy resin application.
Magazine of Concrete Research, 2005, 57, No. 8

The second objective defined was to investigate the possibility of correlation between the slant shear test results and the pull-off test results. In Figs 7 and 8, the average values obtained with the slant shear test are plotted versus the corresponding average values measured with the pull-off test, without and with epoxy resin application, respectively. Taking into account the linear trend line that correlates the values of the graph in Fig. 7, one concludes that there is a good correlation between the slant shear and the pull-off tests, with a correlation coefficient of 0.948, for the situations without epoxy resin application. Taking into account the corresponding values of the situations with epoxy resin application shown in Fig. 8, one concludes that, due to the reduced variation of test results and in spite of the occurrence previously referred to, that correlation is approximately constant.

Conclusions and future research

The major conclusion drawn from the experimental study described in this paper is that the application of
600 Bond strength in tension: MPa 500 400 300 200 100 000 000 y 5 01886x R2 5 09614

200 400 600 800 1000 1200 1400 1600 1800 Bond strength in shear: MPa

Fig. 7. Correlation between slant shear and pull-off tests results of situations considered without epoxy resin application

600

Bond strength in tension: MPa

500 400 300 200 100 000 000 200 400 600 800 1000 1200 1400 1600 1800 Bond strength in shear: MPa

Fig. 8. Correlation between slant shear and pull-off tests results of situations considered with epoxy resin application

467

 lio et al. Ju the adopted bonding agent on the substrate surface does not improve the bond strength of the interface if a surface preparation method is chosen that adequately increases the roughness of the substrate surface. It can also be stated that surface roughening by sandblasting provides a better method than applying the adopted bonding agent as the achieved bond strength of the interface was higher and the operation cost was lower. As the adopted bonding agent was a commercial epoxy resin, which is widely used, attention should be given to this conclusion. Further experimental research is planned to investigate the influence on the bond strength of the method of application of this bonding agent and other epoxy-based bonding agents will also be considered in this study. It was also concluded that the results of the slant shear test and of the pull-off test may be correlated. This is of special interest since the slant shear test can only be performed in the laboratory whereas the pulloff test can be performed in the field and so if the correlations have been previously defined, the bond strength in shear can be assessed in the field.
4. Emmons P. H. Concrete Repair and Maintenance, Part Three: Surface Repair, Section 6: Bonding Repair Materials to Existing Concrete. R. S. Means Company, MA, 1994, pp. 154163. 5. Austin S., Robins P. and Pan Y. Tensile bond testing of concrete repairs. Materials and Structures, 1995, RILEM, 28, No. 179, 249259. lio E. S., Branco F. and Silva V. D. Structural rehabilitation 6. Ju of columns using reinforced concrete jacketing. Progress in Structural Engineering and Materials, 2003, 5, No. 1, 2937. rka M. and Courard L. On the effect of 7. Garbacz A., Go concrete surface treatment on adhesion in repair systems. Magazine of Concrete Research, 2005, 57, No. 1, 4960. 8. Cleland D. J. and Long A. E. The pull-off test for concrete patch repairs. Proceedings of the Institution of Civil Engineers, Structures & Buildings, 1997, 122, No. 4, 451460. D. and Morgan D. R. Influ9. Talbot C., Pigeon M., Beaupre ence of surface preparation on long-term bonding of shotcrete. ACI Materials Journal, 1994, 91, No. 6, 560566. lio E. N. B. S. The Influence of the Interface on the Behav10. Ju iour of Columns Strengthened by RC Jacketing. PhD Thesis, University of Coimbra, 2001. (only available in Portuguese). 11. Abu-Tair A. I., Rigden S. R. and Burley E. Testing the bond between repair materials and concrete substrate. ACI Materials Journal, 1996, 93, No. 6, 553558. 12. Wall J. S., Shrive N. G. and Gamble B. R. Testing of bond between fresh and hardened concrete. Proceeding of the Rilem International Symposium Adhesion between polymers and concrete, Aix-en-Provence, France, 1619 September 1986, pp. 335343. 13. Momayez A., Ehsani M. R., Ramezanianpour A. A. and Rajaie H. Comparison of methods for evaluating bond strength between concrete substrate and repair materials. Cement and Concrete Research, 2005, 35, No. 4, 748757. 14. Kriegh J. D. Arizona slant shear test: a method to determine epoxy bond strength. ACI Journal, 1976, 73, No. 7, 372373. 15. Tabor L. J. The evaluation of resin systems for concrete repair. Magazine of Concrete Research, 1978, 30, No. 105, 221225. 16. Long A. E. and Murray A. McC. The pull-off partially destructive test for concrete. In In Situ /Nondestructive Testing of Concrete, (Malhotra V. M.(ed)), SP-82. American Concrete Institute, Detriot, MI, 1984, pp. 327350 17. Hindo K. R. In-place bond testing and surface preparation of concrete. Concrete International, 1990, 12, No. 4, 127129.

Acknowledgements
O LIZ, We are grateful to SIKA, HILTI, BETA FIVINTE, DYWIDAG, PREGAIA, CIMPOR and SECIL for their collaboration in this research project.

References
1. Murray M. A. Surface preparation for adhesives. Concrete International, 1989, 11, No. 9, 130132. 2. Rizzo E. M. and Sobelman M. B. Selection criteria for concrete repair materials. Concrete International, 1989, 11, No. 9, 4649. 3. Saucier F. and Pigeon M. Durability of new-to-old concrete bondings. Proceedings of the ACI International Conference Evaluation and Rehabilitation of Concrete Structures and Innovations in Design, Hong Kong, 1991, Vol. 1, pp. 689707.

Discussion contributions on this paper should reach the editor by 1 April 2006

468

Magazine of Concrete Research, 2005, 57, No. 8