3643

Russian Journal of Building Construction and Architecture

The aim of this paper is to validate the use as a concrete filler of the waste from the drying process of the aggregate used in the manufacture of hot-mix asphalt (HMA). According to the standard NF P 98-728-1 [7], this material is called recovery filler (RF). After the drying process of the aggregate in a rotating drum with the temperature generally in the range of 150 °C to 180 °C, RF is retained by bag house filters to control its emission [8, 9]. It should be noted that a part (3––4 %) of this RF is stored in a silo for later use in the manufacturing of HMA [7, 10] because the asphalt mixture is a combination of aggregate, asphalt binder, and filler [11]. The majority of the RF, however, is introduced into water to prevent its dispersal into the air because the official journal of the Republic of Algeria No. 24 limits dust values in the manufacture of HMA to 100 mg/m3. In addition, if the production plant has their own quarry, RF may be deposited as illegal filling, resulting in environmental and health problems [10]. In 2017, 3 million tonnes of asphalt mix was produced in Algeria, with an estimated generation of RF of over 120.000 tonnes. These facts have guided us to reflect on the use of RF in the manufacture of concrete.

1. Existing studies. Several research programs have been carried out to study the properties of concrete and HMA mixtures using the mineral filler. The mineral filler is comprised of particles with a physical size that passes through a sieve with a pore diameter of "75µm" [12]. Researchers [3, 11, 13] have demonstrated that the use of filler in foam concrete improves the compressive strength and increases the stiffness modulus of HMA mixtures. Several studies [3, 4, 14] have demonstrated that the type of filler material used can influence properties of both fresh and hardened concrete. Joudi-Bahri et al. [4] have demonstrated that limestone filler gives composite a more homogeneous bond. Bederina et al.’s [15] research has indicated that adding limestone fillers improves the workability and mechanical strength of concrete and reduces dimensional variations. It has also been reported [16] that a higher strength is obtained with the use of filler, as the resulting mix has smaller air voids. Researchers [17] have found that the use of 7––10 % of mineral filler improves the properties of concrete. Al Shamaa et al. [18] have concluded that the use of limestone filler increases the swelling of concrete.

On the other hand, researchers [19] have shown that when the sand grading is kept nearly invariable, the filler properties are important factors for concrete workability. The study of [8] investigated the effect of RF from HMA plants on the mechanical behavior of selfcompacting concrete. The studies [19] reported that the use of RF decreases the compressive strength, flexural strength, splitting tensile strength, and static modulus of elasticity of selfcompacting concrete due to the large particle size of the RF.

However, in [10] it was found that it was possible to obtain a high durability of selfcompacting concrete with RF in terms of resistance to attack of aggressive agents such as car-

60

bonate ions, sulfate, and chloride. Other authors [20, 21] have reported that the use of superplasticizers in concrete reduced water demand and improved slump, thus increasing the mechanical strength of the concrete. This could be explained by the stabilities of ettringite varying in the presence or absence of superplasticizers [22].

Several studies [23, 24] have examined the freeze-thaw resistance of concrete produced with fine recycled aggregates. Their findings show that freeze-thaw resistance is affected more by the water/cement (W/C) ratio than by the type of aggregate used and that air entrainment has a positive effect on improving concrete resistance.

2. Characterization of Materials

2.1. Cement. This study used commercial Portland (CEM II) class 42.5 MPa cement from the Hamma Bouziane factory (Constantine, Algeria). The chemical and physical characteristics of this cement are presented in Table 1. The potential mineralogical composition of the cement was calculated according to the empirical formula of Bogue [25].

2.2. Water.This study used tap water from a civil engineering research laboratory at the University of Constantine 1. Its quality conformed to the requirements of standard NFP 18-404. The chemical compositions of the water are presented in Table 2.

Table 2

61

Russian Journal of Building Construction and Architecture

2.3. Crushed Sand. The crushed sand used (0/5 mm) in this study was from Constantine (ENG de Khroub). The chemical and physical properties of the sand are presented in Table 3 (they were measured using the standards of NF P18-553, NF P18-555, NF P18-560, NF P18-597, and NF P18-598). The grading curves of the different sands are given in Fig. 1.

2.4. Gravel. We used fractions of crushed stone (8/16 and 16/25 mm) from Constantine (National Company of Aggregates: ENG) with the apparent density of 1558.5 kg/m3, specific density of 2550 kg/m3, and coefficient of Los Angeles of 26.84% (hard). The properties were measured using standards NF P18-560, NF P18-554, and NF P18-573. The grading curves of the gravel used are given in Fig. 1.

Fig. 1. Grading curves of gravel and crushed sand compared with the normalized curve

2.5. Superplasticizer. The superplasticizer used in this study belonged to the polycarboxylate group and was supplied in liquid form. Superplasticizer was added at the dosage of 1.6 % of the cement mass. This proportion was determined to be the optimal dosage after comparison of concrete mixes with others dosages (0.8 %, 1.2 %).

62

2.6. Recovery Filler. We used fine particles which are available at HMA plants. In our case the fines were obtained from the HMA plants of El-Djzira and El-Arabiya. This RF is obtained from drying of sand and stone in a rotating drum at the temperatures between 150 °C and 180 °C. The chemical composition, physical properties and grading curve of the RF are presented in Table 4 and in Fig. 2.The significant findings to note are the presence of high percentages of SiO2 and CaO, which exceeds 20 % and 44 %, respectively.

Fig. 2. Comparison of the grading curves of the crushed sand (CS) and recovery filler (RF) and comparison with the normalized curve

3. Characterization of RF with Aphelion Lab 4.4 and Astra 1.6 Software Packages. The present section assesses the output values that characterize the RF (compactness, elongation, fill ratio and average diameter) obtained with the Aphelion Lab 4.4 and Astra 1.6 software packages. The microscopic morphology of the RF particles from a scanning electron microscopy (SEM) image is displayed in Fig. 3. After opening the SEM image with Aphelion Lab 4.4, output values characterizing the RF were obtained. These are presented in Fig. 4 and 5 as well as in Table 5. We also used the Astra 1.6 software package for segmentation and measurement of average diameters. These results are given in Fig. 6.

63

Russian Journal of Building Construction and Architecture

Fig. 3. Scanning electron microscopy image displaying the morphology of different recovery fillers

Analysis of the measurement file obtained with Astra 1.6 allowed us to determine the average diameters to be 1.22 μm. The results also showed that the RF had excellent compactness and a good fill ratio despite the poor state of the particle shape. This was attributed to the small diameter of the grain.

64

Fig.6. Segmentation obtained with the Astra 1.6 software package

4. Mix proportions. Three samples were then subjected to 50 freezing and thawing cycles using a Controls Group 10-D1429/A climatic chamber conforming to Russian National State Standard (GOST) 10060-2012 [26]. The temperature in the climatic test chamber varied from −15 °C to 15 °C, as shown in Fig. 7. The curve is a stair-step shape [27] ascent, part of which is the time when the temperature dropped from 15 °C to −15 °C and then remained at −15 °C, and the horizontal part is when the temperature rose from −15 °C to 15 °Cand then remained at 15 °C.

Fig. 7.Climatic test chamber (Controls Group 10 D1429/A)

The method of B. Scramtaiv [1, 28] was used for all concrete mixtures. Three aspects must be considered: water/cement ratio, incorporation ratio of the RF, and the effect of superplasticizer. The concrete compositions are reported in Table 6.

65

Russian Journal of Building Construction and Architecture

C1: Concrete without a recovery filler (RF) and a superplasticizer: 0.5 (350) C2: Concrete without a recovery filler (RF) and a superplasticizer: 0.6 (300) C3: Concrete without a recovery filler (RF) and a superplasticizer: 0.6 (350) C4: Concrete without a recovery filler (RF) and a superplasticizer: 0.6 (400)

C1 + 10 % RF: Concrete with a 10% recovery filler (RF) and without a superplasticizer : 0.5 (315) C2 + 10 % RF: Concrete with a 10% Recovery filler (RF) and without a superplasticizer : 0.6 (270) C3 + 10 % RF: Concrete with a 10% recovery filler (RF) and without a superplasticizer : 0.6 (315) C4 + 10 % RF: Concrete with a 10% recovery filler (RF) and without a superplasticizer : 0.6 (360) C5 + 15 % RF: Concrete with a 15% recovery filler (RF) and without a superplasticizer : 0.5 (297.5) C6 + 20 % RF: Concrete with a 20% recovery filler (RF) and without a superplasticizer : 0.5 (280) C7 + 15 % RF: Concrete with a 15% recovery filler (RF) and without a superplasticizer : 0.6 (297.5) C8 + 20 % RF: Concrete with a 20% recovery filler (RF) and without a superplasticizer :0.6 (380)

C1 + SP1.6 % : Concrete without recovery filler (RF) and with a superplasticizer : 0.5 (350)

C1 + 10 % RF + SP1.6 %: Concrete with 10% recovery filler (RF) and with a 1.6% superplasticizer : 0.5 (315) C3 + SP1.6 %: Concrete without recovery filler (RF) and with a superplasticizer : 0.6 (350)

C3 + 10 % RF + SP1.6 %: Concrete with a 10% recovery filler (RF) and with a 1.6% superplasticizer : 0.6 (315)

66

5. Results of concrete slump. A concrete slump test was performed according to the standard NF P 18-451 using an Abrams cone. The results obtained are shown in Fig. 8.

Fig. 8.Concrete slump of mixtures without modification and modified with a recovery filler (RF) and a superplasticizer

The results shown in Fig. 8 indicate that better values of concrete slump were obtained with superplasticizers due to the stabilities of ettringite, which confirms the studies in [20––22]. We note that in the case of a W/C ratio of 0.6, the workability increases. However, when a RF is used as a cement replacement in concrete, workability decreases. This decrease was 15 % for concrete C3 + 10 % RF + 1.6 % SP. This can be explained by the fact that RF increases water absorption and decreases concrete slump. Indeed, the morphological aspect of RF, with a low coefficient of elongation of 0.463, does not allow for a high workability.

6. Results of compressive strength. The compressive strength test was performed in accordance with standard NF P 18-406. The compressive strengths were estimated on the 10 × 20 mm samples using a universal press (Controls Group Digimax Plus 70-C0019/B) over 28 days. The obtained results are illustrated in Fig. 9.

The experimental results in Fig. 9 indicate that after 28 days the greatest compressive strength was obtained for a concrete mixture with 10 % RF, 1.6 % superplasticizer, and a W/C ratio of 0.6. Despite the RF having a good fill ratio (0.670) and a small diameter (1.22 μm), its use beyond 10 % in concrete mixtures causes a decrease in the compressive strength. This can be explained due to the percentage of zinc in RF of 52.64 mg/l. Indeed, these results confirm those of [29] regarding the increase in the percentage of zinc in concrete which decreased the compressive strength.

67

Russian Journal of Building Construction and Architecture

Fig. 9. Compressive strength of all mixtures without modification and modified with a recovery filler (RF) and a superplasticizer

We can, therefore, infer that 10 % of RF should be considered optimal. Indeed, a large amount of RF (up to 67 %) reduces the stiffness of the granular skeleton and hence the compressive strength of concrete.

On the other hand, we can see that the compressive strength of concrete decreases at the lower W/C ratio of 0.5. But when we use a W/C ratio of 0.6, an increase in the compressive strength is noted (until 32 %). The higher is the quantity of cement in concrete, the greater is the compressive strength. These results were attributed to the small particle size of cement compared to RF filler (1.22 μm), which confirms the results of [30] regarding the relationship between particle size and compressive strength. Finally, we can say that the compressive strength of concrete with RF is significantly dependent on the W/C ratio.

7. Results of compressive Strength after Freeze-Thaw Cycling. After demolding at 24 hours, the samples were water cured for 28 days. Then three samples were subjected to 50 freeze-thaw cycles using a Controls Group 10-D1429/A climatic chamber conforming to the Russian National State Standard (GOST) 10060-2012 [26].The obtained results are illustrated in Fig. 10.

The results from Fig. 10 clearly demonstrate that 50 cycles of freezing and thawing decreased the compressive strength when a superplasticizer was not used. The decrease for concrete C1 + 10 % RF was 57 %.This can be explained by the fact that the RF increased water absorption, and therefore the freeze-thaw cycles generated microcracks and deterioration of the concrete. As a superplasticizer decreases the water consumption, the compressive strength is improved. The increase for concrete without a RF and with a superplasticizer was 16.59 % (C1 + SP 1.6 %) and 7.54 % (C3 + SP 1.6 %). On the other hand, the increase for concrete with both a RF and a superplasticizer was 0.3 % (C1 + 10 % RF + SP 1.6 %) and 2.08 %

68

(C3 + 10 % RF + SP 1.6 %).These results demonstrate that the freeze-thaw resistance of concrete with RF is affected by several parameters, including the percentage of RF, water content, cement content, and the use of a superplasticizer.

Fig. 10. Compressive strength of concrete mixtures without modification and modified with recovery filler (RF) and superplasticizer after freeze-thaw cycles

8. Conclusions. The main conclusions of this study can be summarized as follows.

––An increase in the content of RF increases water absorption and decreases concrete slump.

––The morphological aspect of RF with its low coefficient of elongation does not allow for good concrete workability.

––Despite the RF’s good fill ratio a threshold of 10 %RF is considered optimal. Beyond this percentage, the amount of zinc in the concrete reaches a point where the compressive strength decreases.

––The use of RF provides a positive effect on the compressive strength if the W/C ratio is optimized. In this study, we found concrete C4 with 10 % RF and without SP (W/C ratio = 0.6).

––The experimental results confirm that superplasticizer reduces the water consumption while improving the workability and the compressive strength of concrete.

––The freeze-thaw resistance of concrete with RF is affected by several parameters such as the percentage of RF, W/C ratio, and the use of a superplasticizer.

––Therefore it is possible to use RF from HMA plants, especially when limited to 10 % and accompanied by the use of a superplasticizer.

69