The RCA and RFA were obtained by crushing waste concrete, which was taken from an old demolished reinforced concrete frame structure, as seen in Fig. (1). The average compressive strength of concrete cores (100mm cube) extracted from the original waste concrete was 38.7MPa. The natural coarse aggregate (NCA) used in the mix was crushed gravel with a maximum size of 25 mm. The natural fine aggregate (NFA) was local river sand with a fineness modulus of 2.40. The basic physical properties of aggregates are listed in Table (1).

Fig. (1). Recycled aggregates (a) RCA, (b) RFA.

In order to study the influence of the substitution of RCA and RFA with different replacement ratios (0%, 50% and 100%) on the freeze-thaw behaviors, nine kinds of concrete were prepared with different mixture proportion listed in Table (2). High efficiency water reducing agent (WRA), mineral powder (MP), and fly ash (FA) were added to enhance the workability of the RAC. The mix proportions and properties of fresh concrete were summarized in Table (2). The concrete type designations starting with RC refer to recycled concrete, and the following two numbers (0, 0.5, 1) indicates the replacement ratio of RCA and RFA respectively. For example, the label ‘RC-1-0.5’ indicates that the recycled concrete with 100% substitution of NCA by RCA, and 50% replacement of NFA by RFA.

As can be seen in Table (2), the slump value decreased with the increasing replacement rate of recycled aggregate. When the NCA and NFA were both totally replaced by recycled aggregates, a significant reduction of slump value was observed for the higher water absorption of RCA and RFA.

A total of 90 specimens, including 54 cubes and 36 prismatic reinforced concrete specimens were cast in a standard laboratory at 25±2 ̊C, as seen in Fig. (2).

The specimens were divided into two groups according to the different shape feature and research content:

Group 1-Fifty-four 100×100×100mm³ cubes with different replacement ratio of RCA and RFA were under two conditions: 27 specimens were exposed to freeze-thaw test, the rest were not. Each condition involved in nine kinds of RAC and three samples were cast with the same concrete mix.

Fig. (2). All specimens.

Fig. (3). Steel bar details.

Group 2-Thirty-six 100×100×400mm³ reinforced concrete prisms with different replacement ratio of RCA and RFA were under two conditions: 18 specimens were exposed to 100 freezing and thawing cycles, the rest were not. Each condition involved in nine kinds of RAC and two specimens were cast with the same concrete mix. The steel bar details are shown in Fig. (3). Three coupons were taken from the longitudinal rebar to conduct the standard tensile coupon tests, as seen in Fig. (4). From these tests, the measured average yielding strength, tensile strength and modulus of elasticity of the rebar was 327MPa, 438MPa and 2.04×10⁵MPa respectively.

Fig. (4). Steel reinforcement test.

After 28 days curing with condition of 20±3°C and 95% RH (relative humidity), the specimens were subjected to freeze-thaw cycles according to “Standard for test methods of long-term performance and durability of ordinary concrete” GB/T50082-2009 [16].

The accelerated freeze-thaw test as recommended by GB/T50082-2009 is water freezing and water thawing method, that is, all specimens are soaked in water for rapid freezing and thawing. Based on the standards, 2~4h for each freeze-thaw cycle is required, and the test time in this study is controlled around 3 hours. Simultaneously, the central temperature of specimens should be controlled within (-18±2)̊C at the lowest temperature and (5±2)̊C at the highest temperature. The freeze-thaw behaviors of RAC is conducted with the KDR-V5 testing machine, as shown in Fig. (5). The specimens were first inserted in rubber specimen boxes with the size of 115×115×500mm³. The rubber specimen box and its cross section dimensions are presented in Figs. (6 and 7). Then, clean water was poured into the rubber specimen box, and the water surface is 5mm above the upper face of specimens. Finally, the rubber specimen boxes were put into the test machine Fig. (8) and the freeze-thaw test started. After 100 freeze-thaw cycles, the compressive strength of cubes and prisms were measured and compared with the responding specimens without freeze-thaw experiment.

Fig. (5). Freeze-thaw testing machine.

Fig. (6). Rubber specimen box.

Fig. (7). Cross section of rubber specimen box.

Fig. (8). Specimens in Freeze-thaw testing machine.

After the completion of 100 freeze-thaw cycles, the failure mode of cube specimens with different replacement ratios of RCA and RFA was similar to that of specimens without freeze-thaw action. During the process of loading, vertical cracks first appeared at the corners and extended upward and downward quickly, this increased the damage of the angular range. As the load increasing, vertical cracks began to appear in succession at the middle position of each side. Subsequently, the middle part of concrete was gradually crushed and peeled, which results in the loss of load-carrying capacity and the final failure, as seen in Fig. (9).

Fig. (9). Failure mode of cubes.

After 100 freeze-thaw cycles, the ultimate strength of cubes decreased when compared to the same mix proportion concrete cubes without freeze-thaw test. The main reason for this phenomenon is that, under the action of freezing and thawing, the ice pressure leads to the accumulation and development of the initial micro cracks inside the concrete. With the increasing load, the initial cracks developed fast. Consequently, the compressive strength decreased, and the decline was associated with the substitution ratio of RCA and RFA.

The measured compressive strength of RAC is given in Table (3). f_c0 and f_c100 is the average strength of three samples before and after 100 freeze-thaw cycles respectively. Δf_c is variation ratio of compression strength after freeze-thaw test.

As shown in Table (3), the compressive strength of RAC with different replacement ratio of recycled aggregate, on the whole, is lower than that of NC, except specimen RC-0-1 and RC-1-0.5. It seems that the compressive strength is not affected by the incorporation of recycled aggregates. This is completely contradictory to the other previous researches [5-7]. The reason lies in the same water consumption, as seen in Table (2). The effective water/cement ratio of RAC is lower than that of NCA, due to the higher water absorption of recycled aggregates. Therefore, although there are some defects in recycled aggregates, the compressive strength of RAC is very close to that of NAC. Considering the higher water absorption of recycled aggregates, many researchers took different measures to compensate the water that recycled aggregates absorb. Some recommended pre-saturating the aggregates [17, 18], others recommended immersion in water for 30 minutes [19], the rest advised adding more water during the concrete production, in order to obtain the same effective w/c ratio. However, not only the lower strength but the higher water absorption, they both are the characteristics of recycled aggregates. If the additional water were added, it is hard to judge who caused the change of mechanical properties of RAC. Therefore, no correction of water was done in this research to study the effects of different substitution rate on the compression performance of RAC.It can also be seen in Table (3) that, after exposure to 100 freeze-thaw cycles, the loss rate of compressive strength (LRCS) of NC is minimum(10.4%), and the LRCS of the remaining specimens in an order from small to large is RC-0.5-0, RC-0-0.5, RC-1-0, RC-0.5-0.5, RC-0-1, RC-1-0.5, RC-0.5-1, and RC-1-1. When the coarse and fine aggregates were totally replaced by recycled aggregates, the LRCS of RC-1-1 was up to the maximum as 41.8%. It can also be found that the interface damage is more complex and serious than that of other specimens.

This may be explained by the higher water absorption (WA) of recycled aggregate. A lot of research results show that the WA of RCA and RFA is generally 3.6%-8.0% and 8.3%-12.1% respectively, while natural rock is always lower than 3% [20]. Therefore, with the increasing replacement rate of RCA and RFA, recycled concrete is easier to damage under the inside pressure during the process of freezing and thawing, so the compressive strength decreased with the increasing contents of recycled aggregate.