Evalua on of the use of steel slag in micro surfacing

Recebido: 8 de dezembro de 2018 Aceito para publicação: 3 de fevereiro de 2020 Publicado: 15 de maio de 2020 Editor de área: Kamilla Vasconcelos ABSTRACT Micro surfacing (MS) is a type of bituminous coa ng frequently applied in the surface of pavement structures in order to prevent the occurrence of common distress and/or as a maintenance procedure. Micro surfacing has successfully been used in some states of Brazil as well as in many countries all over the world. Local aggregates used for micro surfacing composi on are in some cases scarce and/or expensive. Therefore, the main objec ve of the present research is to verify the technical, financial and environmental viability of using steel slag aggregate (SSA) in micro surfacing applica ons. To reach the main objec ve, a procedure was developed in four steps. In the first step it was made the characteriza on of conven onal (grani c) and alterna ve aggregates (steel slag) and polymer-modified emulsified asphalt. In the second step, mix design, surface abrasion resistance and se>ng me tests were performed. In the third step, the micro surfacing mechanical performance was analyzed using a laboratory traffic simulator to observe accumulated permanent deforma on; surface wear; micro texture and macro texture. In the fourth step, a compara ve study of costs concerning micro surfacing applica on using conven onal mineral aggregate (MS-MA) and alterna ve aggregates (MS-SSA) were performed. The laboratory procedure results indicated that the steel slag studied has the poten al to be used in micro surfacing applica ons with more ru>ng resistance and beAer durability. The cost analysis showed for medium transport distances bigger than 60 km, the MS-SSA will be cheaper than MS-MA solu on.


INTRODUTION
Micro surfacing (MS) is one of the techniques applied to pavement preventive maintenance, with good results when used in heavy traf ic highways. The MS is considered an evolution of the slurry seal (SS), but with differences in the speci ications of the asphalt emulsion type, of the aggregates, the use of sand is not allowed in MS applications for example, and the durability is better than SS (FHWA, 1994).
According to Hein et al. (2003) and from the association of asphalt emulsion manufacturers (AEMA), the MS can be employed as an alternative to hot asphalt mixture to correct pavement rutting. However, according to Ceratti and Reis (2011), it is possible to correct rutting problems with depths up to 40mm. It has also been mentioned that in this case the time of the traf ic release is about 3 hours.
The MS is composed of 90-95% of aggregates. In the way to propose sustainable alternative to pavement preventive maintenance, there is availability of several residues that can be better investigated to be used in the paving area in order to promote the environmental preservation. According to Loiola (2009), certain types of alternative aggregates have been applied to paving. One of the materials with potential to be employed in substitution of mineral aggregates is the steel slag aggregate (SSA).
Although the use of SSA has already been carried out in the pavement community since the 1990s as granular material for base and subbase and as aggregate for asphalt layer applications, studies that analyze the behavior of SSA in MS applications are still incipient in Brazil. SSA is a product that has been gaining prominence in the pavement area, since its use has been well researched for road purposes. However, as cited by Rohde (2002) SSA has in the presence of oxides such as CaO and MgO in its composition. These oxides have an expansive feature due to chemical reactions. Care should be taken to check whether the SSA is already cured before being applied to road works. According to Teixeira et al. (2019) the SSA expansive effects could be minimized too when combined with other asphalt mixture components.
Over the years, several studies were developed with the aim of making feasible the use of SSA in road works. In Brazil, the irst use of this by-product in the paving in 1986, in the state of Espıŕito Santo, where the SSA was used over a stretch of more than 100 km (Silva e Mendonça, 2001). Other authors that studied the use of SSA in granular layers were Rohde (2002), Parente et al. (2003), Santos Neto (2007), Cavalcante et al. (2010) and Pires et al. (2019). In addition, DNIT (2008) also studied SSA as an aggregate in base and sub-base layers. The use of SSA in hot-machined asphalt mixtures was also investigated and the results showed that it has the potential to replace the conventional aggregate in such blends as can be seen in Castelo Branco (2004), Pedrosa (2010), Silva (2010) Tavares (2012) and Teixeira et al. (2019).
In Brazil, the use of SSA for thinner coating type Surface Treatment (ST) is more recent, dating from 2007 and started by Loiola (2009) and Pereira (2013). The results showed that the SSA also had a satisfactory performance and an experimental pavement section was built as reported by Rocha (2011). It is important to study new materials such as SSA concerning bituminous wearing course applications since local aggregates used are in some cases scarce and/or expensive. In the case of ST, the required characteristics are sometimes dif icult to obtain in the local aggregates, for example the Los Angeles abrasion test resistance is sometimes over the limit (40 %). Vasconcelos (2013) evaluated the behavior of double surface treatment (DST) and cape seal, having performed a comparison using SSA and conventional aggregates. However, the use of SSA in MS has been insuf iciently researched, so studies should be more developed in this area. In this paper, the main objective is to verify the technical, inancial and environmental viability of using SSA in MS applications.

MATERIALS AND METHODS
For the experimental program execution, a mineral aggregate (MA) considered of excellent behavior for application in MS and a SSA were collected. The procedure was divided into the following steps: characterization of materials; MS mix design; evaluation of the coating behavior with the use of an LCPC (laboratory central des ponts et chaussées) type laboratory traf ic simulator and inancial analysis of the studied aggregates.
The characterization tests carried out, including those speci ic for the steel aggregates were: Particle size distribution (DNER-ME 083/98); Shape index (DNER-ME 086/94); Los Angeles Abrasion (DNER-ME 035/98); Methylene blue (NBR 14949/2003); Sand equivalent (DNER -ME 054/94). To complement the SSA characterization, the environmental tests of solubilization (NBR 10006/2004) and leaching (NBR 10005/2004), as well as the evaluation of its expansion (DNIT -ME 113/2009) were necessary. (NBR 6299/99); Penetration (DNER -ME 003/99); Elastic recovery (DNER -ME382/99) and Softening point (NBR 6560/2008). In order to perform the MS mix design, the Wet Track Abrasion Test (NBR 14746/2001) was performed with four different binder content for each type of aggregate studied. The binder contents, water, cement and additives vary according to the mix design, de ined in the laboratory. The mold used for the WTAT is a 279 mm diameter metal disc. The molding procedure followed the same methodology adopted by Castro (2011) and is shown in Figure 1.
In the WTAT, the specimen is measured and then the minimum binder content is de ined. The test starts by weighing the specimen and after that they are kept immersed in water for one hour. At the end of this immersion period, the specimen are subjected to WTAT test through the equipment for 5 minutes. At the end of WTAT test, the specimen is washed to remove all loose material and the specimen is again taken to the oven at a temperature of 60° C until reaching constant weight. The result is obtained by calculating the weight loss suffered by the specimen subjected to WTAT test. The DNIT (DNIT -ES 035/2005) establishes a maximum loss of 538 g/m². This process is illustrated in Figure 2. The other test to be performed to determine the binder optimum content for the MS is the sand adhesion by the LWT machine (NBR 14841/2002). The molding sequence is shown in Figure 3. The mixing process is also carried out in the same way as mentioned in WTAT procedure. The sand adhesion test by the Loaded Wheel Tester (LWT) allows the exudation measurement of the specimen and determines the maximum binder content to be employed in the MS. The procedure is divided into two parts. In the irst part the MS specimen is subjected to a load TRANSPORTES | ISSN: 2237-1346 33 of 56 kg during 1000 cycles on the LWT machine. The second part consists in measuring the sand adhesion to the specimen which varies according to the binder rate employed. The process is illustrated in Figure 4.

Figura 4.
Conducting the LWT test (Castro, 2011) The Mixture Adhesion Test (NBR 14757/2001) veri ies the binder-aggregate compatibility. This standard, based on the US ISSA TB-114/1990, determines the water resistance of residual asphalt adhered to the aggregate. The result is obtained by visual analysis, an area that remained covered by the asphalt residue.
The Minimum Mixing Time Determination Test (NBR 14758/2001) was performed to calculate the additive content to be employed in the MS and consists in measuring the emulsion set time by performing a laboratory mix. The standard establishes a minimum time of 120 seconds for the initiation of the process of emulsion set. If the measured time is lower, it is necessary to add an amount of additive to reach the time of 120 seconds. However, technicians and specialists in MS, as well as Ceratti and Reis (2011), indicate that the laboratory mixing time should be between 180 and 300 seconds, thus ensuring adequate time for emulsion disruption in ield applications. The time considered in this research was 240 seconds.
The MS laboratory behavior was measured by analyzing the surface wear (NBR 14746/2001) and the cure time for the traf ic release (NBR 14798/2002) of MS specimens molded at the optimum design level. The cure time test consists of measuring the pull-out resistance of aggregates on a surface of a MS specimen during its curing process. The samples were molded and tested at 30 min, 60 min, 90 min, 120 min and 150 min times. In each period, it was obtained the torque value. The molding sequence is shown in Figure 5.
According to speci ication concerning the time of 30 min, the minimum torque value accepted is 12 kg.cm, has indicated that the cure occurs in a satisfactory way. For the time of 60 min, it is expected to obtain a torque with a value greater than 20 kg.cm, minimum acceptable value to occur the release to traf ic. The procedure sequence is illustrated in Figure 6. This research also evaluated the MS behavior under the action of the small-scale laboratory traf ic simulator following the French standard guidelines NF P98-253-1 (AFNOR, 1991). In order to eliminate variables that could interfere in the analysis of results, it was chosen to use the methodology developed by Vasconcelos (2013) as shown in Figure 7. The aspects observed in the simulator test were: accumulated permanent deformation; surface wear; micro texture and macro texture. The macrotexture was measured by the Sand Patch Test (ASTM-E-965-96) and the micro texture was analyzed by the British Pendulum Test (ASTM-E-303-93).

Figura 7. Stages of the molding, simulation and measurement of the properties of the MS in the simulator LCPC
According to NF P98-253-1, the rutting measurement must occur in speci ic positions marked on the test plate and the recommended load application is 500 kg for hot mixes. Vasconcelos (2013) used the small traf ic simulator to evaluate cold mix asphalt pavements designed for low-traf ic. The author mentioned that for the loading proposed by the standard led the plates to rupture during the irst 100 cycles.
However, the mentioned author chose to use a load of 75 kg where it was possible to evaluate the surface wear and rutting along the cycles. For this research, the same load magnitude was adopted, and it was possible to make a comparison of the MS behavior with the DST and Cape Seal studied by Vasconcelos (2013). The rutting measurement readings were performed at 100, 500, 1000, 3000, 5000, 7000 and 10000 cycles.
The inancial feasibility of using alternative materials studied in this paper was veri ied through the application cost (U$$/m²) based on DNIT referential cost system (SICRO). The binder cost was de ined according to National Petroleum Agency of Brazil (ANP) referential table.

RESULTS PRESENTATION 3.1. Characteriza9on of materials
In order to select the grain size, it was decided to use the materials included in band II DNIT speci ications for MS (Figure 8), which is the most adopted particle size distribution range used by the Brazilian road agencies. The results are shown in Table 1 and met the requirements of current speci ications. Speci ically, for the SSA, the type used in this study was iron slag and it is important to mention that they were cured for 8 months with water injections and the percentage of slag on MS with SSA was 100%. However, it was not possible to use the slag size distribution from the steel industry, so it was necessary to build a grain size curve in laboratory.

Figura 8. Grain size curve of selected aggregates
According to the speci ication limits, both aggregates studied in this research attend the required standard limits to be used for MS applications. It is important to note that the Los Angeles abrasion test resistance of alternative aggregate was a little bit over the limit (40%) while the result obtained for SSA was only 17 %. DNER-ME 262/94 speci ies that the SSA should not have an expansion value greater than 3 %. For cured iron slags low expansion results are expected. Therefore, the SSA can be used without causing damage to the coating.

Mix Design and Analysis of MS Laboratory Behavior
The Mixture Adhesion Test (NBR 14757/2001) veri ied aggregate-emulsion compatibility. Then it was possible to verify that, through the MS sample visual inspection for both aggregates, over 90 % of the area was covered and it was considered satisfactory. For both aggregates, the binder content was 10.8 %.
The Minimum Mixing Time Determination was performed for the SSA as well as for the MA. The emulsion set time measured was greater than 300 seconds and higher than the minimum time limit. It was not necessary for the use of additives. It is emphasized that the incorporation of additives in the MS mixtures is only intended to delay the emulsion set time. Their use does not imply improvements in the wear surface and reduction concerning traf ic time.
The results obtained regarding the surface wear for the two aggregates studied reached the limits established in the current speci ications, and the SSA presented a lower wear when compared to the MA. This can be justi ied since MA Los Angeles abrasion test results (40.8 %) are higher than SSA (17.0 %). Regarding the traf ic time release, the torque values obtained for the time of 60 min are above the standard speci ications. This way, the traf ic can be released within one hour without causing any damage to the quality of the coating just applied. Table 3 summarizes the results obtained in the mix design and analysis of the laboratory behavior performed for the two groups studied. It should be (Table 3) noted that the binder content for AS, which has a higher speci ic weight was the same used for AM. However, these percentages are in relation to the weight of the aggregate used. As for costs, since there is a higher aggregate rate for AS one can af irm that this will result in a higher consumption of binder.

Analysis of the MS Submi?ed to the Laboratory Traffic Simulator
Before starting the rutting test, the application rate in kg/m² of MS was calculated. For the case of MS-MA, considering a thickness of 1.5 cm, the application rate was 29.83 kg/m². For the SSA, this rate was higher and equal to 38.33 kg/m². Following the same procedure adopted by Vasconcelos (2013), the MS plates were submitted to the traf ic simulator. Figure 9 shows the evolution of the deformation suffered by the MS with MA and SSA. By analyzing the plate deformation, it can be seen that the rutting when the SSA is used as an aggregate is about 50% lower for 10000 cycles when compared to MA, using the same criterion adopted by Vasconcelos (2013). Also, in this situation, it was decided to continue up to 30000 cycles only with the SSA, in order to analyze its behavior when submitted to a larger traf ic volume, certifying its better behavior to permanent deformation.
Based on this, it can be seen that with the MS-SSA is more rutting resistant. In relation to surface wear and aggregate detachment, none of these phenomena was observed during the simulation, indicating a suitable involvement of the aggregates by the binder and a good behavior of the coating when the two types of aggregate were used. Regarding macro and micro texture parameters, the results reached the current speci ications for the studied aggregates. Table 4 shows the results summary obtained for the granulometric combinations submitted to the simulation cycles. After conducting the simulator procedure, it was found that the SSA in the MS presented better behavior when compared to the MA.

Financial Evalua9on
The inancial evaluation was presented in US$/m². The cost compositions were based on DNIT referential cost system (SICRO) and the binder cost was based on National Petroleum Agency of Brazil (ANP) referential table. The value for each solution was calculated according to the mix design parameters and application rates.
Considering Medium Transport Distances (MTD) about 60 km for MA and analyzing the inal application costs in US$/m², the application cost of MS with SSA was about the same compared to the MA solution. For MTD bigger than 60 km, the MS-SSA will be cheaper than MS-MA solution. Table 5 presents the cost comparison between applying the MS with different aggregates studied. Therefore, considering regions with SSA suppliers located nearby roadway maintenance services, the MS-SSA application can be cheaper than MS-MA, because in this case it will be not necessary to purchase MA and spend more with transportation. Moreover, its behavior in wear and in the rutting test was better, concerning superior durability. This tendency is also observed in other studies that use SSA. It should be emphasized that this higher durability should be tested in real-scale experimental sections, so that other factors that may alter coating performance can also be analyzed.

CONCLUSIONS
The main objective of the present research was to verify the technical, inancial and environmental viability of using steel slag aggregate in Micro Surfacing applications. The aggregates and emulsion characterization, the mix design process and the MS behavior were evaluated through laboratory procedures and a comparison with the MA results was made. This paper also evaluated the results obtained in laboratory traf ic simulator. A cost analysis was made with the use of different types of aggregates. The results obtained showed that the SSA studied has the potential to be used in MS. The SSA was considered viable in the technical and environmental scope. Regarding the costs involved in the use of the SSA, it is important to mention that MTD can be decisive in choosing SSA instead of MA.
It should also be noted that the MS with SSA has better wear resistance and presented the smallest deformations in the laboratory traf ic simulator. In addition, road safety standards