Experimental Study on Bond Stress between Ultra High Performance Concrete and Steel Reinforcement

Due to axial deformations generally caused by flexure, shear stress will be generated across the interface between reinforcement and surrounding concrete. This longitudinal shear stress is called bond stress and coordinates deformation between concrete and reinforcement. With increasing a member's axial deformation, bond stress finally reaches its ultimate value, bond strength, after which deformation of reinforcement and surrounding concrete will be not coordinated any more. Studies have shown that addition of nanosilica into cement-based materials improves their mechanical properties. Considering the unique characteristics of nanosilica, it seems that this material can be used in ultra-high performance concrete. Therefore, further research is needed on how to use it in concrete mixes. Due to the importance of examining bond stress and the lack of exact equations for bond stress of ultra-high performance concrete and steel reinforcement, the present study aimed to assess the bond stress between concrete and steel reinforcement.


Introduction
High strength and ultra-high performance concrete has many advantages.Due to its better mechanical properties and low permeability, this type of concrete is gradually replacing conventional concrete.Because of its considerable properties, this type of concrete can either be used in structures to resist loads, or in large bridges and several constructions due to being affected by environmental conditions.Micro-silica is widely used as an additive to cement in producing high performance concrete.This matter is used to enhance the strength and efficiency of concrete.Several experiments have shown that replacing part of cement with micro-silica improves sulphate and acid resistance of concrete and reduces chlorine permeability.By addition of microsilica to concrete or cement mortar, due to being fine grained, it fills the space between cement particles, so the existing pores will become smaller.Moreover, due to the reaction between silica and calcium hydroxide remained from cement hydration process, more C-S-H gels are produced and, as a result, more capillary cracks will be covered [1].Recently, considering the unique characteristics of nanosilica, it seems that this material can be used in ultra-high performance concrete.Therefore, further research is needed on how to use it in concrete mixes.To this end, the present study used Pullout test to assess the effect of nanosilica on the bond stress between steel reinforcement and ultra-high performance concrete.Pullout test is the oldest, simplest, cheapest and less time-consuming way to measure local bond stress of concrete.In this test, a reinforcement is placed into a cylindrical or cube shaped concrete specimen, and then while the concrete is fixed in place, the reinforcement is pulled out.Since the reinforcement is under tension and concrete is under compression, the resultant relative strain will lead to relative slip.Many researchers have studied the bond between streel reinforcement and ultra-high performance concrete.Alkaysi, M., El-Tawil, S. ( 2016) conducted an experimental study on the bond stress between ultra-high performance concrete and streel reinforcement.They calculated the bond stress between 13, 16 and 19 mm reinforcements and ultra-high 1236 performance concrete in the pull out test.The average compressive strength of the ultra-high performance concrete made in these experiments was 190 MPa [2].Carbonell Munoz, M. A. et. al. (2014) examined the bond stress between conventional and ultra-high performance concrete and streel reinforcement.The conventional concrete with a compressive strength of about 50 MPa and the ultra-high performance concrete with a compressive strength of about 150 MPa were made for these experiments [3].Engstrom, B. et. al. (1998) presented the effects of concrete confinement and coating on bonding in high strength concretes.They showed that with reducing thickness of coating up to 16 mm (equal to the diameter of reinforcement) resulted in a 25% reduction in the maximum bond stress compared with the well-confined specimen (with sufficient coating).When using a 32 mm coating, the loading will be same as that of wellconfined concrete [4].Kim, S. et. al. (2016) conducted an experimental study on the bond stress between ultra-high performance concrete and 10, 13 and 19 mm steel reinforcements [5].Cake, K.H. et.al. ( 2010) conducted an experimental study on the bond stress between ultra-high performance concrete and high strength steel reinforcements.The pull out test based on RILEM standards was used in this study [6].Finally, these experiments showed that the bond stress of ultra-high performance concrete is 5-10 times higher than conventional concrete.Roy, M. et. al. (2017) used pull out test to determine the bond stress between ultra-high performance concrete and steel reinforcements.In this study, the strength of reinforcements was 415 MPa, and the compressive strength of the concrete was considered between 122.6 MPa to 176.1 MPa.Finally, sliding diagrams of reinforcement in concrete-force were plotted for all specimens [7].Xing G. et.al. (2015) performed pull out testing to determine the bond stress between ultra-high performance concrete and steel reinforcement [8].Guizani, L. et. al. (2017) conducted a Local bond stress-slip model for reinforced concrete joints and anchorages with moderate confinement.Guizani, L. et.al. in their paper presents a summary of an experimental investigation and the derivation of a bond-slip model for reinforcing steel embedded in moderately confined concrete under monotonic and cyclic loadings [9].Yan, C. and Mindes, S. (1994) conducted a Bond between epoxy-coated reinforcing bars and concrete under impact loading [10].Duchesneau, F., et. al. (2011) conducted a Monolithic and hybrid precast bridge parapets in high and ultra-high performance fibre reinforced concretes [11].

Local Bond Measurement
There are several bond tests used to examine the stress transfer from reinforcement to the surrounding concrete.These tests aim to find out how to model the actual interaction between reinforcement and concrete in real structures.However, achieving this goal is difficult because the bond between concrete and reinforcement is complicated by other structural parameters such as flexural bond, lateral pressure, riveting effect and cracking pattern.The suitable bond test should be carefully selected to reflect the actual conditions of the structure.Pull-out test is the oldest, simplest, most inexpensive and less time-consuming method to measure bond stress.In this test, a reinforcement is placed into a cylindrical or cube shaped concrete specimen, and then while concrete is fixed in place, reinforcement is pulled out.Because the reinforcement is subjected to tension and concrete is subjected to compression, the resultant relative strain will lead to relative slip.This test can provide a good comparison between bond strengths and corresponding development lengths.Pull-out test has been used by many researchers to study the effect of various parameters on the bond strength.In this test, short lap splice lengths are used to generate uniform bond stress along reinforcement, which is called local bond.

Steel Reinforcement Properties
Steel reinforcements having diameters 18  were used in determining the Ultra high performance concrete-steel bond strength.Some properties of these steel reinforcements, obtained through tensile test, are given in Table 1.

Mix Design of UHPC
Several mix designs have so far been offered for ultra-high performance concrete.After studying and testing several mix designs and assessing feasibility of producing them in laboratory, the mix design proposed by Schneider Jianxin was selected to be used in this study.In samples containing nanosilica, micro-silica equivalent to 2.5, 4.5 and 6.5 wt% cement was replaced by nanosilica.The mix designs used in the present study are given in Table 2 in kilograms per cubic meter.Type I cement with a strength class of 525 kg/cm was used in this study.The micro-silica used in this study was purchased from Zhikava company and its chemical composition is presented in Table 3.The superplasticizer was purchased from Silcrete company.This poly carboxylate-based superplasticizer is available with the brand Pema.The nanosilica used in this study were purchased from Lima Nano Pars company that its chemical composition is given in Table 3.The need for thermal treatment is one of the unique properties of ultra-high performance concrete.thermal treatment is a simple process and, in fact, it is an additional phase in the concrete production which in order to strengthen its structure microscopically.Thermal treatment is not required for all applications of ultra-high performance concrete, 1237 because without thermal treatment this type of concrete already has a considerable strength and flexibility compared with high performance concrete.According to the reports, thermal treatment improves mechanical properties of concrete by at least 15%.It has also been emphasized that after thermal treatment, durability of concrete is increased and shrinkage and creep are significantly reduced.thermal treatment improves the micro structure of concrete by accelerating the pozzolanic performance of micro-silica and modifying hydration structure.In this study, the specimens were demolded one day after concrete pouring.After demolding, two treatment techniques were applied to the specimens.Some specimens were placed in a vapor environment (90 ° C and 95% moisture) for 48 hours (Figure 1).After this step, the specimens were tested in a standard laboratory environment (22 ° C and moisture variations between 30-50%).The other specimens were placed in the laboratory environment since molding step until testing.Specimen production process was as follows: concrete pouring for all specimens was completed within 20 minutes after mixing.All specimens were placed on a vibrating table during concrete pouring and then were vibrated after pouring for 30 seconds.Then all specimens were covered by a plastic sheet to reduce the rate of loss of moisture.Before the main tests, compressive strength test was performed using standard cylindrical specimens at ages of 7, 28, 90 and 180 days, by breaking three specimens of each design per day.Compressive strength test results obtained for these ages are illustrated in Figure 2

Test Device
The loading method and the test device are shown in Figure 3. Loading was performed as load controlling and the load magnitude was recorded at any moment by an electronic load cell.The LVDT placed at the end of the reinforcement measures the slip of the reinforcement.The records were automatically saved by the data collection system.When the specimen is placed on the support and tensile loading is applied to the reinforcement, a compressive stress is formed at the contact of the specimen and the support.This compressive stress can lead to an increase in bond strength and thus an error in the test results.As a result, in the specimens made based on the RILEM standards, a spacing is considered between the bond zone and the contact point to the support to remove bearing pressure effect.Obviously, there is no contact between the concrete and the reinforcement at this spacing.

Ordinary treatment
Thermal treatment

Specimens
In this paper, 24 specimens were tested to assess the effect of nanosilica on local bond between ultra-high performance concrete and steel reinforcements.In this test, we measured the bond stress between ultra-high performance concrete and steel reinforcement of No. 18 with the bond lengths and concrete coatings of db,2db ,5db.Sample naming is so that e.g. in the sample R18C3L2N2.5-1,R18 means that tests were conducted based on RILEM standards and the reinforcement No.18 was tested.C3 represents concrete coating in cube specimens.According to considering three values of db,2db , and5db for concrete coating, C3 represents the third coating which is equal to 3 × 18 = 54  for this specimen.L2 shows the reinforcement-concrete bond length in cube specimens.L2 refers to the second reinforcement-concrete bond length which is equal to 2 × 18 = 36  for this specimen.N2.5 shows the nanosilica percentage in the mix design, and the number 1 or 2 at the end of name indicates type of treatment that can be ordinary or thermal treatment.Table 5 shows specifications of the specimens.In Figure 4, a series of specimens made in the laboratory after demolding are presented.

Result and Discussion
Summary of the results obtained from testing standard RILEM specimens [12] are presented in Tables 6 to 8. Where, u is bond stress and u/√  ′ is normalized bond stress for specimens.Bond length was considered twice the diameter of the reinforcement for all specimens.addition of nanosilica into concrete and thermal treatment increased the compressive strength of concrete.So that in the case of ordinary treatment, the 28-day compressive strength of concrete was increased by about 37 percent by replacing 6.5 percent by weight of cement nanosilica instead of microsilica, while this ratio was more than 40 percent in the case of thermal treatment.With increasing compressive strength of concrete, the bond stress between steel reinforcements and concrete was increased.Also, as can be seen in the tables, Increasing the concrete coating has increased the bond stress and normalized bond stress.Figures 5 and 6, respectively, reviews the results of the local tension stresses of steel reinforcement and ultra-high performance concrete in conventional mode.Figure 5 shows changes in bond stress of reinforcement No. 18 by increasing reinforcement coating.With increasing reinforcement coating, the bond stress was increased so that in the specimen containing 6.5% nanosilica, with increasing concrete coating by two times and three times increased bond stress 44% and up to 77%, respectively.Moreover, in the specimen containing 4.5% nanosilica, with increasing concrete coating by two times and three times increased bond stress 42% and up to 82%, respectively.Figure 6 shows changes in bond stress of reinforcement No. 18 by reducing nanosilica content in concrete under ordinary treatment, so that with increasing nanosilica content from zero to 6.5% by weight of cement, the bond stress in R18C1L2-1, R18C2L2-1 and R18C3L2-1 specimens was increased by 54%, 38% and 15%, respectively.Figure 6 shows the variation in the bond stress of reinforcement No. 18 by decreasing the amount of nanosilica in concrete in conventional curing mode.The bond stress in the R18C1L2-1 sample has increased by increasing the amount of nanosilica from zero to 2.5%, 4.5% and 6.5% of cement weight as much as 6.8%, 42% and 52%, respectively.The bond stress in the R18C2L2-1 sample has increased by increasing the amount of nanosilica from zero to 2.5%, 4.5% and 6.5% of cement weight as much as 19.8%, 27% and 38% respectively.The bond stress in the R18C3L2-1 sample has increased by increasing the amount of nanosilica from zero to 2.5%, 4.5% and 6.5% of cement weight as much as 1.4%, 10.6% and 15.4%, respectively.Similarly, Figures 7 and 8 illustrate changes in bond stress of reinforcement No. 18 by increasing reinforcement coating and reducing nanosilica content in concrete in the case of thermal treatment.The results indicate that with increasing reinforcement coating and increasing nanosilica content, the bond stress between steel reinforcements and concrete was increased.Figure 7 shows changes in bond stress of reinforcement No. 18 by increasing reinforcement coating.With increasing reinforcement coating, the bond stress was increased so that in the specimen containing 6.5% nanosilica, with increasing concrete coating by two times and three times increased bond stress 32% and up to 81%, respectively.Moreover, in the specimen containing 4.5% nanosilica, with increasing concrete coating by two times and three times increased bond stress 33% and up to 77%, respectively.Figure 8 shows the variation in the bond stress of reinforcement No. 18 by decreasing the amount of nanosilica in concrete in conventional curing mode.The bond stress in the R18C1L2-2 sample has increased by increasing the amount of nanosilica from zero to 2.5%, 4.5% and 6.5% of cement weight as much as 13%, 19%, and 28%, respectively.The bond stress in the R18C2L2-2 sample has increased by increasing the amount of nanosilica from zero to 2.5%, 4.5% and 6.5% of cement weight as much as 3%, 9%, and 16%, respectively.The bond stress in the R18C3L2-2 sample has increased by increasing the amount of nanosilica from zero to 2.5%, 4.5% and 6.5% of cement weight as much as 4.6%, 14%, and 25%, respectively.Table 8 shows the bond stress and normalized bond stress between steel reinforcement and concrete in the case of ordinary and thermal treatments.As can be seen, with increasing reinforcement coating and increasing the percentage of nanosilica, the bond stress was increased.According to the test observations, failure of the specimens can be divided into three main modes of pull-out, split, and bar yielding.In the mode of pulling the reinforcement out of the concrete by removing the concrete keys between reinforcement treads as much as concrete shear capacity, the keys are slipped off and the reinforcement is pulled out of concrete.In this case, the concrete specimen remains intact without any cracks or damage indicating destruction.This failure mode was observed in highly coated specimens.In the split mode, due to the reaching of hoop tensile stresses to the ultimate tensile strength of concrete, failure is done with wide radial cracking and splitting the specimen into two or more parts (Figure 9).The reinforcement bar yielding mode occurs due to the long bond length or high strength of concrete.In this case, before the bond zone reaches the ultimate capacity, the reinforcement yields.
Where, uc is the bond stress in MPa, c is the minimum concrete coating on the reinforcement, db is the diameter of the reinforcement,   = 0.55√  ′ , and   ′ is the compressive strength of the concrete.The initial value of fb is used to correct (modify) the equation 1. Where, c is the minimum concrete coating on the reinforcement, db is the diameter of the reinforcement ,   = 0.55√  ′ , and   ′ is the compressive strength of the concrete.By placing the equation 2 in the general form of Equation 1, we have: Where, c1 and c2 are constant coefficients.Equation 3 can be simplified to the following linear equation:

Conclusions
In this research, we investigated the local bonding stress of the ultra-high performance concrete and steel reinforcement.The results of the research show that the relationship of bond stress is reformed as follows.
Also Addition of nanosilica into concrete and thermal treatment increased the compressive strength of concrete.So that the 28-day compressive strength of concrete was increased about 37% by replacing 6.5 percent weight of cement microsilica with nanosilica.While this ratio was more than 40 percent in the case of thermal treatment.With increasing reinforcement coating, the bond stress was increased.so that in the specimen containing 6.5% nanosilica, with increasing concrete coating by two times and three times increased bond stress 44% and up to 77%, respectively.
When concrete in the case of normal treatment, with increasing reinforcement coating, the bond stress was increased so that in the specimen containing 6.5% nanosilica, with increasing concrete coating by two times and three times increased bond stress 44% and up to 77%, respectively.Moreover, in the specimen containing 4.5% nanosilica, with increasing concrete coating by two times and three times increased bond stress 42% and up to 82%, respectively.
In addition, the bond stress in R18C1L2-1 sample has increased by increasing the amount of nanosilica from zero to 2.5%, 4.5% and 6.5% of cement weight as much as 6.8%, 42%, and 52%, respectively.The bond stress in R18C2L2-1 sample has increased by increasing the amount of nanosilica from zero to 2.5%, 4.5% and 6.5% of cement weight as much as 19.8%, 27%, and 38%, respectively.The bond stress in R18C3L2-1 sample has increased by increasing the amount of nanosilica from zero to 2.5%, 4.5% and 6.5% of cement weight as much as 1.4%, 10.6%, and 15.4%, respectively.When concrete in the case of thermal treatment, with increasing reinforcement coating, the bond stress was increased so that in the specimen containing 6.5% nanosilica, with increasing concrete coating by two times and three times increased bond stress 32% and up to 81%, respectively.Moreover, in the specimen containing 4.5% nanosilica, with increasing concrete coating by two times and three times increased bond stress 33% and up to 77%, respectively.The bond stress in R18C1L2-2 sample has increased by increasing the amount of nanosilica from zero to 2.5%, 4.5% and 6.5% of cement weight as much as 13%, 19%, and 28%, respectively.The bond stress in R18C2L2-2 sample has increased by increasing the amount of nanosilica from zero to 2.5%, 4.5% and 6.5% of cement weight as much as 3%, 9%, and 16%, respectively.The bond stress in R18C3L2-2 sample has increased by increasing the amount of nanosilica from zero to 2.5%, 4.5% and 6.5% of cement weight as much as 4.6%, 14%, and 25%, respectively.

6 ,
Figure 10.    vs    plot obtained from tests for the ordinary treated specimens

Table 4 . Mix design 1 Mix design 2 Mix design 3 Mix design 4 Figure 2. Bar graph of the compressive strength tests for ages of 7, 28, 90 and 180 daysTable 4 . The compressive strength test results for ages of 7, 28, 90 and 180 days. No of Days Mix design 1 Mix design 2 Mix design 3 Mix design 4 Ordinary treatment Thermal treatment Ordinary treatment Thermal treatment Ordinary treatment Thermal treatment Ordinary treatment Thermal treatment
and