The influence of warm mix asphalt on binders in mixes that contain recycled asphalt materials

Construction and Building Materials 77 (2015) 50–58

Contents lists available at ScienceDirect

Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat

The influence of warm mix asphalt on binders in mixes that contain recycled asphalt materials Ashley Buss a,⇑, R. Christopher Williams b, Scott Schram c a

Department of Civil, Construction and Environmental Engineering, Iowa State University, 394 Town Engineering Building, Ames, IA 50011, United States Department of Civil, Construction and Environmental Engineering, Iowa State University, 490 Town Engineering Building, Ames, IA 50011, United States c Iowa Department of Transportation, Office of Materials, 800 Lincoln Way, Ames, IA 50010, United States b

h i g h l i g h t s  Field production and evaluation of warm mix asphalt (WMA) using three technologies.  Similar performance for HMA and WMA, with WMA having lower air voids on average.  Evaluation of WMA/HMA using recycled asphalt materials via extraction and recovery.  After recovery, no measurable long-term influences from WMA additives in binders.  Reduced production temperatures do not show long term impacts on binder properties.

a r t i c l e

i n f o

Article history: Received 5 August 2014 Received in revised form 12 December 2014 Accepted 13 December 2014

Keywords: Warm mix asphalt Recycled asphalt pavement Recycled asphalt shingles

a b s t r a c t Rheological effects of warm mix asphalt (WMA) additives have been carefully studied in the laboratory since their introduction. Research has shown that reduced plant temperatures decrease the aging of the asphalt binder. It is important to determine if the reduction in the asphalt binder grade is still detectable after in-service aging. Each pavement represents one of the three predominate types of WMA technologies: chemical modifiers, wax modifiers and a foaming process. All mixes in the study included RAP and some with recycled asphalt shingles (RAS). The performance grade (PG) was determined for virgin binders. Field-cores were measured for density. Binder from cores was extracted and recovered to compare between the HMA and WMA. Field performance surveys compared HMA and WMA test sections. Findings show little evidence to suggest WMA facilitates the incorporation of higher amounts of recycled asphalt materials. The recycled binder had a larger influence on binder properties compared to WMA additives. Performance surveys for HMA and WMA mixtures were comparable. Recovered binder from field cores showed similar performance in both HMA and WMA binders. All three mixtures comparing WMA and HMA mixes show evidence that the reduced production temperatures do not have long term impacts on the binder properties. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction Warm mix asphalt (WMA) technologies are being increasingly used to reduce mixing and compaction temperatures of hot mix asphalt (HMA). The WMA technologies are available in the form of additives or asphalt plant modifications. The use of warm mix asphalt with recycled asphalt materials has added another layer of complexity to quality control and quality assurance practices in the asphalt industry. During the introduction of WMA to the

⇑ Corresponding author. E-mail addresses: [email protected] (A. Buss), [email protected] (R.C. Williams), [email protected] (S. Schram). http://dx.doi.org/10.1016/j.conbuildmat.2014.12.023 0950-0618/Ó 2014 Elsevier Ltd. All rights reserved.

industry, multiple benefits were exhibited and discussed including the combined use of WMA and recycled asphalt material but little research has been done to investigate the influence of WMA binder on mixes containing recycled asphalt materials that have undergone some in situ aging. This research focuses on measuring the effect of using a warm mix asphalt additive in conjunction with recycled asphalt materials that include recycled asphalt pavement (RAP) and/or recycled asphalt shingles (RAS). WMA works in two ways that are synergistic to using recycled asphalt materials. First, the WMA reduces the mixing/compaction temperatures which will reduce aging of the virgin binder, creating an overall softer binder compared to a mix that is produced at a higher temperature. This effect is measured in this research by collecting virgin binder and

A. Buss et al. / Construction and Building Materials 77 (2015) 50–58

field cores. The binder from the field cores is then extracted and recovered to compare WMA/HMA differences for both virgin and recovered binders. The second reason WMA works well in mixes using recycled asphalt materials is due to the improved compactability. Compaction improvements allow for a stiffer mixture, which is characteristic of mixes with recycled asphalt materials, to reach target density. In this research, field cores were collected after 2 years of traffic for WMA and HMA sections and comparing the densities of the field cores will determine if WMA and HMA show differences in the compactability. The purpose of this paper is to measure the impact of WMA technologies on mixes with various amounts and types of recycled asphalt materials and measure the influence of WMA on the binder, compaction and pavement performance after 1 or 2 years of in situ aging in the field. This will evaluate the collective use of WMA with recycled asphalt materials. The detailed objectives are:  to determine if any reduction in the continuous asphalt grade is still detectable after some in-service aging,  determine the influence shingles have on WMA binder compared with RAP and show how the differences change with additional shingles added to the mixture,  determine if there is evidence to support the premise that WMA may allow for additional RAP,  observe pavement performance for two consecutive years, and  compare field core density to determine if WMA binders play a role in the densification process of WMA. 2. Background Within the last 15 years, a great deal of asphalt research has focused on sustainable practices and developing warm mix asphalt materials. The research of warm mix asphalt began in the 1950’s with foamed asphalt [1]. Further studies on foamed asphalt were conducted in the 1980’s and found that curing temperature, length and moisture conditions dramatically affect the strength of foamed asphalt mixtures that contain sand and RAP [2]. In addition to foaming asphalt with water via plant modifications, synthetic zeolite additives were also developed for the purpose of foaming asphalt. Chemical and wax modifiers were also developed as warm mix asphalt additives. Driving the development of the WMA technologies are the production benefits such as reduction in fuel cost, reduced emissions, improved compactability, a longer paving season, longer haul distances and the use of WMA with recycled asphalt materials [3–5]. The commercially available additives used in this study included a wax additive, a forestry products chemical additive, and a foaming plant modification. The wax additive is a Fischer– Tropsch paraffin wax. The Fischer–Tropsch (F–T) process produces the fine crystalline, long-chain aliphatic hydrocarbon that makes up the wax. The wax allows for a reduction in production temperatures of 10–30 °C and is added at 3% by weight of the mix to gain the desired temperature reduction but should not exceed 4% due to potential impact to the binder’s low temperature properties [6]. The chemical additive used in this study is derived from the forest products industry and is commercially available. This additive contains surfactant properties that emulsify the asphalt [7,8]. A plant modification was used to produce foamed asphalt at reduced plant temperatures. The foaming of the asphalt is controlled by injecting water to the asphalt through nozzles that create small water droplets as the binder is being mixed with hot aggregates. Studying differences in the HMA and WMA field binders after in-service aging will help to detect long-term differences between HMA and WMA in the various WMA technologies. Laboratory studies have shown that measureable differences do occur in mixture properties between WMA and HMA mixes [9,10].

51

A study which used laboratory aged binders demonstrated how certain WMA modifiers will significantly change the rheological and failure properties of the binders. Investigation of artificially aged binders helps to identify where potential changes in the field may occur as a result of using WMA technologies. Measuring the differences in viscosity between WMA and HMA in laboratory studies provides evidence for what will likely occur in the field. Another WMA study suggests some of the laboratory differences between binders may be due to variability in oxidative stability or due to viscosity effects [11]. A laboratory study focused on quantifying the long-term effects of WMA additives on binders aged in the laboratory at set time intervals. This study found that the WMA binder had the lower shear modulus which is likely due to the 20 °C reduction in rolling thing film oven (RTFO) aging [12]. As other variables in the field are added, such as recycled asphalt materials, it is important to investigate the impacts of WMA additives and processes on rheological properties, pavement density and performance. In general, studies have shown good performance of WMA pavements and WMA technologies do achieve reduced mixing and compaction temperatures [13]. The incorporation of moderate to high amounts of RAP in WMA is hypothesized to work well because the reduction in temperature allows for less stiffening of the binder compared to a conventional mixture and a national study showed that using RAP with WMA exhibited adequate binder blending [5]. A study looking at high RAP mixes in Europe found no differences between a WMA and conventional mixture with the exception of a significant reduction in production temperature. The mixture performance test indicated that the WMA additive did not affect the stiffness nor the fatigue life [14]. In a WMAhigh RAP field trial in Florida, mixture testing indicated a softer material response in WMA pavement with a high RAP content of 45%. Dynamic modulus values were similar but the Hirsh and Witczak models under estimated the E⁄ values. An explanation for the differences between the HMA and WMA test results presented in the study is the incomplete blending of virgin and recycled binder in the WMA mixture, due to the lower production temperatures [15]. Since differences were detectable, further studies are needed for mixes with both high and typical amounts of RAP. A synthesis review of WMA suggested selecting slightly higher temperatures when WMA is used with high RAP amounts [16]. Identifying differences in WMA and HMA mixes is complicated by the multiple technologies available. Some changes in mixture properties may exist with a particular additive where another additive may indicate no detectable difference between WMA and HMA properties. For example, studies have shown that a Fischer–Tropsch wax increases the viscosity of binders at 60 °C [17]. The variety of technologies can influence mixture properties in different ways. The focus of this research is to examine the influence of multiple WMA additive types in field projects that contain various amounts of recycled asphalt materials over time.

3. Mixture and material information Three pavement projects were selected to be constructed with both hot mix asphalt and warm mix asphalt test sections. Each project used a different type of WMA technology and all three were constructed during the same construction season. The following construction season, three additional WMA projects were constructed. At the time of production, virgin binder was collected from the tank at the asphalt plant for each project. The virgin binder was tested to verify the performance grade (PG) and compare HMA and WMA binders. Testing of the virgin binder also established a baseline from which to compare recovered binder samples.

52

A. Buss et al. / Construction and Building Materials 77 (2015) 50–58

The projects were selected in different regions within the State of Iowa. Fig. 1 displays the location for each pavement included in this research and shows each of the major types of WMA, chemical/wax/foaming, which were added to the mixture by the contractor or supplier. Table 1 shows the project name, location, HMA/ WMA technology and important mixture details. The mixture designs range from 300 thousand (K) to 10 million equivalent single axle loads (ESAL) and the binder grades include PG 64-22, PG 64-28, and PG 58-28. Most of the pavements contain approximately 20% RAP with the exception of Hwy 13 in Clayton County which contained 5% RAP. The limited amount of recycled asphalt material in this pavement helps to compare the difference between using only 5% and 20% RAP in a WMA mixture. The US 61 (Muscatine) project studies the use of shingles with WMA and includes 0%, 5% and 7% recycled asphalt shingles (RAS). 4. Experimental plan and test methodology Fig. 2 shows the experimental plan that was developed to compare the asphalt binders and field cores. Original binder properties were evaluated to measure initial differences between the WMA binder and the HMA binder. Table 2 shows the full testing plan

for the binders. The testing plan allows for comparison of WMA/ HMA binder in the field with the WMA/HMA virgin binder properties. Rheology testing for virgin and recovered binders were performed in the same way except binders were not RTFO aged since the aging that occurs during construction is assumed to have already taken place. It is also expected that recovered binder will be stiffer than the virgin binder because of the recycled asphalt materials used in the pavements and the natural aging that occurs in the top layer of asphalt pavements due to oxidation and ultraviolet light exposure in the field. 4.1. Performance grade binder tests The binder testing for original and recovered binders followed Superpave standard specification for Performance Graded Asphalt Binders, American Association of State Highway and Transportation Officials (AASHTO) M320. The AASHTO M320 test provides each binder’s performance grade and gives detailed information about the rheological differences between the binder groups shown in Table 2. First, virgin binders were tested in a dynamic shear rheometer (DSR) manufactured by TA Instruments, model AR1500ex, shown in Fig. 3(a). Testing was performed according

Fig. 1. Project locations for mixes included in study.

Table 1 Mixture summary. Code

Year

Road name

Project location

Project number

WMA technology

ESAL level

Binder grade

RAP (%)

RAS (%)

Floyd Marcus

2009 2009

U.S. Route 218 Iowa Hwy 143

Charles City, IA Bypass North of Marcus, IA

NHSX-218-9(129)–3H-34 STP-143-1(4)–2C-18

HMA 10M HMA 3M

64-28 64-22

17 20

– –

Warren

2009

U.S. Route 65

STP-065-3(57)–2C-91

HMA 3M

64-22

20

–

Tama

2010

STP-S-C064(110)–5E-64

Chemical additive

20

–

2010

MP-013-2(704)59–76-22

Chemical additive

64-22

5

–

Muscatine-0 Muscatine-5 Muscatine-7

2010 2010 2010

U.S. Route 61 U.S. Route 61 U.S. Route 61

South of Strawberry Point, IA Northbound lanes between Muscatine, IA and Blue Grass, IA

HMA 300K HMA 1M

64-22

Clayton

County Hwy E67 Iowa Hwy 13

SB Lanes of US 65 North of Indianola, IA East of Laurel, IA

Chemical additive Fischer–Tropsch wax Foaming

HSIPX-061-4(107)–3L-70 HSIPX-061-4(107)–3L-70 HSIPX-061-4(107)–3L-70

Chemical additive Chemical additive Chemical additive

HMA 1M HMA 1M HMA 1M

58-28 58-28 58-28

20 13 6

– 5 7

53

A. Buss et al. / Construction and Building Materials 77 (2015) 50–58

Fig. 2. Experimental plan overview.

to AASHTO T-315 where the high temperature failure parameter of G⁄/sin(d) is equal to 1.0 kPa. Virgin binder was then short-term aged in an RTFO, model CS325-B, manufactured by James Cox & Sons, shown in Fig. 3(b). RTFO simulates the aging that occurs during the construction process. The RTFO aged binder is tested in the DSR with a failure parameter of G⁄/sin(d) equal to 2.2 kPa. The remaining RTFO binder is placed in the PAV for long term aging at 100 °C at 2.1 MPa for 20 h according to AASHTO R28. This simulates aging in the field that occurs over 7–10 years in service. The PAV used in this testing was manufactured by Applied Systems, Inc. After PAV aging, the binder is degassed and bending beam rheometer (BBR) beams are prepared. The BBR, manufactured by Cannon Instrument Company and shown in Fig. 3(d), measures low temperature properties according to AASHTO T-313. The binder properties of stiffness and the rate of change in stiffness (m-value) are measured in the BBR. The binder stiffness properties impact the amount of recycled materials that can be added to a mix. WMA may allow for higher incorporations of recycled asphalt materials if a binder stiffness reduction is detectable in WMA pavements. 4.2. Extraction and recovery Extraction and recovery was performed according to ASTM D2172 and ASTM D5404, respectively. A toluene–ethanol blend was used as the solvent to avoid using harsher chemicals. Two centrifuges were used in the extraction process. The first centrifuge uses an aluminum bowl and filter paper to filter out the aggregates from the asphalt–toluene solution. The very fine particles were then removed using the second high speed centrifuge where the mineral filler is separated from the solvent–binder solution. Once the binder is fully separated from the aggregate, the solution is placed in the rotary evaporator system for distillation.

The rotary evaporator system uses nitrogen gas as a blanket over the solution so no oxidation of the asphalt will occur during recovery. Once the solvent is fully distilled, the binder will be ready for subsequent performance grade binder tests. Prior to performing extraction and recovery on field cores, the recovery process was calibrated by using a binder that had already been tested. The binder was dissolved in solvent and then recovered using the same process that was used for the cores. The recovered binder was tested in the DSR to ensure that the rheological properties were similar to the original binder properties. Binder properties were matched only when glass marbles were used in the rotary evaporator to ensure all of the solvent was distilled off. When marbles were not used, the binder displayed a significantly reduced stiffness. At least three marbles were used for each recovery. This testing study series found that the toluene–ethanol solvent does not work with the wax additive. The recovered wax-modified binder was extremely soft and was not able to be tested at typical DSR temperatures. This indicated that the wax additive had trapped solvent in the binder. The first sample was discarded and a second sample of the wax-modified mixture was prepared for extraction and recovery using normal propyl bromide. The recovery was successful with the normal propyl bromide solvent. 4.3. Collection of field cores and field performance surveys Studying binder properties allows for material characterization at the micro-level but WMA additive impacts should be studied at a system performance level as well. Each of the mixes studied have physical pavement locations in Iowa which allows for annual pavement condition surveys. Field cores were collected allowing for 1 or 2 years of field aging and traffic densification, depending on the project date of construction. Density of field cores was measured using the CoreLok system according to AASHTO T331. Binder from pavement cores was extracted and recovered to evaluate the impact of WMA additives after at least 1 year of aging in the field. The extracted binder properties will also help to show how RAP influences the performance grade and whether there are detectible benefits from using warm mix asphalt additives. The binder recovery was performed on only the surface layer which consisted of the mixtures used in this study. The pavement condition survey information is used to compare overall performance of each pavement section and to investigate any differences between the HMA and WMA sections. The projects were too large to survey the entire pavement so three 500 ft (152.4 m) sections were selected randomly within the stationing for each mixture. The survey occurred on those sections. The surveyed areas were marked with roadway marking paint and were surveyed the following year. Primary measurements include the length and severity of transverse, longitudinal, edge cracking, rutting and popouts. Studying this evidence will show if WMA and HMA have similar performance in the field.

Table 2 Testing plan for original and recovered binders. Binder

Virgin binder DSR

RTFO DSR

Recovered binder DSR

PAV DSR

Recovered PAV DSR

BBR

Recovered BBR

Floyd WMA Floyd HMA Marcus WMA Marcus HMA Warren WMA Warren HMA Tama WMA Clayton WMA Muscatine-0% Muscatine-5% Muscatine-7%

XXX XXX XXX XXX XXX XXX XXX XXX XXX – –

XXX XXX XXX XXX XXX XXX XXX XXX XXX – –

XXX XXX XXX XXX XXX XXX XXX XXX XXX XXX XXX

XXX XXX XXX XXX XXX XXX XXX XXX XXX – –

XXX XXX XXX XXX XXX XXX XXX XXX XXX XXX XXX

XXX XXX XXX XXX XXX XXX XXX XXX XXX – –

XXX XXX XXX XXX XXX XXX XXX XXX XXX XXX XXX

54

A. Buss et al. / Construction and Building Materials 77 (2015) 50–58

Fig. 3. Binder testing equipment. (a) Dynamic shear rheometer. (b) Rolling thin film oven. (c) Pressure aging vessel. (d) Bending beam rheometer (Photos by Ka Lai Ng Puga).

5. Results 5.1. Binder results The binder data for each mixture is graphed in the same chart for easy comparison. All tests were run in triplicate and the graphs show the average of the three test results. The error bars in each figure show the 95% confidence interval. The important comparisons are the observed differences between HMA and WMA binder properties in the original binders and in the cores. Comparing the original binder measures the influence WMA has on the binder at the beginning of a pavement’s life and the cores measure the differences in the binder after the pavement has undergone aging in the field. The binders will also show different low temperature properties. The lower mixing and compaction temperatures reduce aging of the asphalt binder [4] and it is important to see if there are any measureable low-temperature benefits in the WMA binders extracted from the field cores. Similarly, the high temperature comparison will ensure that no negative effects are occurring due to WMA technologies. Figs. 4–7 display the failure temperatures of the binders tested and serve as a visual representation of the performance grade range for binders at both high and low temperatures. Fig. 4 displays the binder test results for the Floyd mixture. The binder for the Floyd mixture is a PG 64-28 and the WMA technology is a forest products chemical modifier. Seventeen percent RAP was added to the mixture and the increased stiffness is reflected in the recovered binder properties. The virgin HMA binder has a slightly higher failure temperature compared with the WMA, indicating reduced stiffness in the WMA binder. The WMA has a slightly lower failure temperature for the RTFO aged binder. Field cores were obtained after 2 years of in-service aging and the recovered binder

Fig. 4. Binder results for Floyd mixture containing 17% RAP.

for HMA and WMA shows almost identical results indicating that there is little difference between the two binders after 2 years in the field and binder recovery. The PAV test results showed slightly higher failure temperatures for the WMA binders at intermediate temperatures for both original and recovered binders. The similarities in the intermediate failure temperatures suggest no evidence of long term benefits for adding the chemical-based WMA additive for fatigue cracking. The BBR data in Fig. 4 shows both binders meeting the 28 °C minimum with the WMA original binder having the lowest failure temperature followed by the HMA failure temperature with just under a 0.5 °C difference. Comparison of the recovered binders for HMA and WMA show the same low fail-

A. Buss et al. / Construction and Building Materials 77 (2015) 50–58

Fig. 5. Binder results for Marcus mixture containing 20% RAP.

Fig. 6. Binder results for Warren mixture containing 20% RAP.

ure temperature with only a 0.2 °C difference. The BBR binder tests indicate that the inclusion of RAP and in-field aging increased the low temperature grade of the binder and reduced the influence of the WMA additive. The RAP benefits the high temperature by stiffening the binder to exceed the high temperature requirements. There is 17% RAP in this mix and the recovered binder grade reflects the changes due to RAP and 2 years of in-service aging. Fig. 5 shows the Marcus mixture binder testing. The HMA and WMA binders have similar original binder properties when compared for virgin, RTFO and PAV aged material. The binder grade for the Marcus mixture is a PG 64-22 and the WMA additive is a wax additive. Twenty percent RAP was used in this mixture and the influence on binder stiffness is evident in Fig. 5, showing an increase in the recovered binder stiffness. The WMA and HMA binders recovered from field cores show similar increases with HMA being slightly stiffer. The stiffness differences are approximately 3 °C. The wax-modified WMA binder was initially extracted with a toluene–ethanol blend which left a soft sticky binder that failed immediately in the DSR. The toluene was an adequate solvent for other WMA mixes, but did not perform well as a solvent for the wax-modified binder. One hypothesis is that the wax and binder structure trapped the solvent within the molecular structure. The binder was subjected to a long period of time in the rotary evaporator without any success in removing additional solvent. Extraction was repeated on additional cores using a normal

55

propyl bromide based solvent. The n-propyl-bromide appeared to adequately dissolve and be extracted from the wax modified binder. The PAV data for Marcus, shown in Fig. 5, displays the same trends that were observed in the chemical WMA additive in the Floyd mixture. No apparent differences are shown at intermediate and low failure temperatures for the recovered WMA binders. The final HMA/WMA comparison is the Warren mixture, Fig. 6, which uses a foaming additive to achieve WMA characteristics. The virgin binder is a PG 64-22 and the mixture contained 20% RAP. The original binder results actually show the same binder because the ‘‘foaming’’ occurred on a plant modification but testing the virgin binder collected on the date of construction was important because there was a 9 day time lapse between the HMA and WMA mixture production due to inclement weather. The additional testing ensures that each tank binder had the same properties. The binder results show the same properties for the WMA and HMA production days. The recovered binder also shows similar high failure temperatures grades for WMA and HMA. This is expected because it would be unlikely as the foaming process should leave no longterm impacts on the binder. The intermediate and low failure temperatures are the same for the recovered HMA and WMA. Overall, the binder grade will be influenced by the addition of RAP and aging in the field and the binder testing results show no long-term influences from the WMA foaming production process for high, intermediate and low temperatures. All three mixtures use different WMA technologies and all show evidence that the reduced production temperatures do not have long term impacts on the binder properties. The remaining mixes in the binder study incorporate WMA binders modified using a chemical additive. The parameter of interest is the comparison of high and low temperature binder properties of these WMA mixes as higher amounts of binder are replaced by different amounts and types of recycled asphalt materials. The Tama mix contains 20% RAP and the Clayton mixture contains only 5% RAP and results are shown in Fig. 7. Comparatively, the recovered binders show a higher increase in binder stiffness for the Tama mixture due to the higher RAP content but the benefits of using recycled material may be worth the slight increase in stiffness. The difference between the virgin and recovered high temperatures was an increase of 7.6 and 10.4 °C for Clayton (5% RAP) and Tama (20% RAP), respectively. The difference between the virgin and recovered low temperatures was an increase of 1.6 and 6.31 °C for Clayton (5% RAP) and Tama (20% RAP), respectively. The limitations of this comparison is that these are two different mixes produced at different plants; however, both mixes use a PG 64-22 binder and in situ aging conditions are similar as the pavements are located within 125 miles of each other. Comparing the two mixes shows that the extra 15% of RAP will likely increase the low temperature grade but provides a more sustainable mixture. Continuation of pavement surveys will show if measureable differences in performance is evident. The Muscatine WMA mixture used a combination of RAP and RAS to investigate the effects of using RAP and RAS on WMA binder properties. The Muscatine mix used a PG 58-28 virgin binder and was produced in three variants with different amounts of recycled asphalt material. ‘‘Muscatine 0% RAS’’ contained 20% RAP, ‘‘Muscatine 5% RAS’’ contained 13% RAP and ‘‘Muscatine 7% RAS’’ contained 6% RAP. Binder replacement was approximately 20%, 30% and 30%, respectively. The overall binder replacement of 30% is the same for the 5% and 7% shingle mixes but the binder in the RAS is stiffer than the RAP binder. Since the 30% binder replacement uses different amounts of RAP and RAS, the mixes will have different resultant PG grades which will influence the mixture performance. The Muscatine binder results are shown in Fig. 8. The virgin binder grade met PG 58-28 requirements. The binder grade increased as more

56

A. Buss et al. / Construction and Building Materials 77 (2015) 50–58

Fig. 7. Binder results for Clayton and Tama mixtures.

binder replacement occurred with the recycled binder and as higher amounts of shingles were added. This trend is also reflected at intermediate and low temperature testing with the stiffness increasing from 0% RAS to 7% RAS. The 5% RAS mixture increased the low temperature by approximately 6.5 °C and the 7% RAS increased the low temperature by 13 °C compared to the Muscatine 0% RAS mixture. The low temperature grade increases are expected due to the relatively high stiffness of binders in RAS. The primary limitation of RAS is the increase in stiffness documented in the binder testing. RAS should be used with caution in cold regions that are prone to thermal cracking. The low temperature grade increase was 6.5 °C as a result of increasing shingles from 5% to 7%. The recommended best practice is to verify the resultant PG of a mixture that uses high amounts of RAS.

5.2. Field core density and pavement performance surveys Studying control and experimental WMA mixes in the field provide an opportunity to compare densities of constructed pavements. Density and air voids are dependent on the aggregate gradation curve, the amount of asphalt and construction practices. Traffic on an asphalt pavement over time will also increase the density. Core air voids obtained by AASHTO T331 are shown in

Fig. 8. Binder results for Muscatine mixture. ⁄High temperature estimated by extrapolation due to temperature being too high for current DSR measurements.

Fig. 9 and each bar represents the average of seven core air voids and the error bars represent the 95% confidence interval. The cores for Floyd, Marcus and Warren were collected after 2 years of densification from traffic and the cores from Tama, Clayton and Muscatine were collected after 1 year of service. Floyd, Marcus and Warren (the mixes where HMA and WMA were studied) show increased average density for all three WMA mixtures. The Muscatine mixture, which studied the influence of RAP and RAS, shows that increased levels of recycled asphalt materials did not negatively influence the field air voids. The largest difference in HMA and WMA air voids is the Warren mixture where the foamed asphalt had an approximate average of 4% air voids and the HMA mix had an average of 5.8% air voids. Overall, the average WMA air voids were lower for WMA mixes compared to HMA mixes but statistical differences were not observed at the 95% confidence interval. The pavement surveys were performed for 2 years. Pavement surveys were conducted according to the long-term pavement performance (LTPP) program guidelines [18]. The full project length was too long to survey so three 500 ft (152.4 m) sections were selected at random. Floyd, Marcus and Warren pavements were surveyed 2 and 3 years following construction. Tama, Clayton and Muscatine were surveyed 1 and 2 years following construction. No fatigue cracking was present on the pavement sections surveyed. The most prevalent pavement distresses were transverse cracking and rutting. Transverse cracking will sometimes be caused by low-temperature cracking but may also be caused by reflective cracking. Rutting depth was measured at three locations in the wheel path and an average is reported in Fig. 10. Rutting measurements are not available for Clayton and the rutting for the Muscatine mixtures is not reported because it is located on a shoulder and does not see traffic loads that would cause rutting. Overall, rutting measurements were comparable between HMA and WMA mixes with the Marcus WMA showing an average lower rutting for WMA compared to the HMA control in year 3. The transverse cracking was measured using a measuring wheel transverse within the 500 ft (152.4 m) survey sections. The transverse crack spacing is presented in Fig. 11 and the pavements included in this graph are overlays on concrete pavement. All mixtures except Floyd and Tama were HMA overlays on PCC. Floyd and Tama mixes were an HMA on HMA. The crack spacing measured will be dependent on overlay properties as well as the underlying condition of the PCC. Underlying pavement condition can greatly influence the performance of an overlay. The Floyd, Tama and Muscatine pavements exhibited no transverse cracking and were not included in Fig. 11. The lower spacing of the transverse cracks indicates more cracking of the test section. The WMA section of the Warren mixture has the highest amount of transverse cracking the first year of pavement surveys but the Warren HMA/WMA sections showed similar transverse crack spacing the following year. The transverse cracking is likely due to reflection cracking but cracks appeared faster in the WMA section. There was no transverse cracking for the Floyd, Tama and Muscatine mixes. When comparing the Tama (20% RAP) and Clayton (5% RAP) pavement performance, it is important to look at transverse cracking because this is the failure mode that is most likely influenced by the additional of RAP. The Tama mixture exhibited no transverse cracking while the Clayton mixture developed transverse cracking after the first year but did not show additional cracking the second year the pavements were surveyed. The Tama mixture was an asphalt overlay over an asphalt pavement and the Clayton mixture was an asphalt overlay over a jointed reinforced concrete pavement. The performance data shows that the additional 15% RAP in the Tama mixture did not negatively influence the mixture performance because no transverse cracking was observed.

A. Buss et al. / Construction and Building Materials 77 (2015) 50–58

Fig. 9. Field core air voids comparing HMA and WMA.

Fig. 10. Rutting measurements from pavement surveys.

Fig. 11. Spacing between transverse cracks in pavement survey sections.

57

58

A. Buss et al. / Construction and Building Materials 77 (2015) 50–58

6. Conclusions and recommendations Binder research of WMA technologies helps to predict the influence of WMA on mixture properties but outside factors such as recycled asphalt materials, pavement structure and pavement density also have system-wide implications. The purpose of this research is to investigate the in situ performance/influence WMA technologies have on binder properties, pavement density/air voids and pavement performance. Studying control and experimental mixes help to better evaluate the role WMA plays in the recovered binder properties. Three WMA additives, a chemical modifier, wax additive and foaming process were investigated in multiple pavement projects located in different regions throughout the State of Iowa. These various projects used different amounts of recycled asphalt materials which allowed for the measurement WMA additives/processes have on mixes using recycled asphalt pavement and recycled asphalt shingles. Testing of virgin and recovered binders helped to compare and contrast the ways in which WMA influences binder properties in the presence of recycled asphalt materials in the mixture. The wax modified binder could not be adequately separated from the toluene–ethanol blend solvent. The use of toluene as a solvent is a less toxic alternative to normal propyl bromide and trichloroethylene. The effect of solvent type when extracting binder containing waxes should be further evaluated because of the quality control and quality assurance implications. Field cores were collected and pavement density was compared for HMA and WMA mixtures. Field surveys were conducted in consecutive years following construction and primary distresses were rutting and transverse cracking. The pavement performance surveys for HMA and WMA mixtures were comparable. Pavement air voids was slightly lower for the WMA mixes on average and the foamed asphalt mixture showed the largest difference. The Tama mixture had high air voids and was the only mix that appeared to contain too high of air voids (above the average acceptance value of 8%) but since other WMA mixtures did not exhibit the same problems, it is likely there were other causes for the high air voids measured in the core samples. The WMA additives allow producers to reduce the mixing and compaction temperatures of asphalt production. There was little evidence to suggest WMA facilitates the incorporation of higher amounts of recycled asphalt materials because the recycled binder had a much larger influence on measured binder properties compared to the lesser influence from the WMA additives. As would be expected, the recovered foamed asphalt gave the most similar binder results to the HMA control mixture due to the release of the entrapped water. Interestingly, the foamed asphalt pavement sections exhibited the highest pavement density indicating improved compactability. Recovered binder from field cores showed similar performance in both HMA and WMA binders. Field densities for HMA and WMA pavements were similar in most cases and the field performance results were similar as well. WMA pavement densities were higher on average indicating that WMA technology does complement the use of recycled asphalt materials. The recovered binders showed an increase in the low temperature performance grade due to oxidative aging and the addition of recycled asphalt materi-

als. The binder stiffness was particularly sensitive to the amount of RAS in a mixture. To ensure low temperature performance, even when using WMA, the final mixture’s performance grade should be evaluated through extraction and recovery. Future studies should investigate the use of WMA technologies combined with recycled asphalt materials and rejuvenators. The rejuvenators may help to reduce the increased stiffness in recycled binders and if used in conjunction with WMA at reduced temperatures, may work toward creating superior-performing sustainable mixtures. Acknowledgments The authors would like to thank the Iowa Department of Transportation for sponsoring the demonstration projects. The authors would like to thank the Iowa Highway Research Board for funding this research. Special thanks are due to Paul Ledtje at Iowa State University and to Ka Lai Ng for the pictures of binder testing equipment. The authors would also like to thank Bill Rosener and the members of the Asphalt Paving Association of Iowa. References [1] Csanyi L. Foamed asphalt. Technical Bulletin No. 240. Washington (DC): American Road Builders Association; 1959. [2] Roberts FL, Engelbrecht JC, Kennedy TW. Evaluation of recycled mixtures using foamed asphalt. Transp Res Board 1984:78–85. [3] D’Angelo J, Harm E, Bartoszek J, Baumgardner B, Corrigan M, Cowsert J, et al. Warm mix asphalt: European practice publication. Washington (DC): US Department of Transportation; 2008. [4] Rubio MC, Martinez G, Baena L, Moreno F. Warm mix asphalt: an overview. J Cleaner Prod 2012;24:9. [5] Bonaquist R. Mix design practices for warm mix asphalt. Washington (DC): National Cooperative Highway Research Program; 2011. [6] Jamshidi A, Hamzah MO, You ZP. Performance of warm mix asphalt containing SasobitÒ: state-of-the-art. Constr Build Mater 2013;38:530–53. [7] Hurley G. Evaluation of new technologies for use in WMA. Auburn (AL): Auburn University; 2006. [8] Hurley GC, Prowell BD. Evaluation of EvothermÒ for use in warm mix asphalt. Auburn (AL): Auburn University; 2006. [9] Kvasnak AW, West R, Moore J, Nelson J, Turner P, Tran N. Case study of warm mix asphalt moisture susceptibility in Birmingham. In: Transportation Research Board annual meeting. Washington (DC): Transportation Research Board; 2009. [10] Zelelew HP. Laboratory evaluation of the mechanical properties of plantproduced warm-mix asphalt mixtures. Road Mater Pavement Des 2012. [11] Ahmed EI, Hesp SAM, Senthil KPS, Rubab SD, Warburton G. Effect of warm mix additives and dispersants on asphalt rheological, aging and failure properties. Constr Build Mater 2012;37:6. [12] Banerjee A, Smit AdF, Prozzi JA. The effect of long-term aging on the rheology of warm mix asphalt binders. Fuel 2012;97:9. [13] Oliveira JRM, Silva HMRD, Abreu LPF, Fernandes SRM. Use of a warm mix asphalt additive to reduce the production temperatures and to improve the performance of asphalt rubber mixes. J Cleaner Prod 2013;41:8. [14] Oliveira JRM, Silva HMRD, Abreu LPF, Fernandes SRM. The role of a surfactant based additive on the production of recycled warm mix asphalts – less is more. Constr Build Mater 2012;35:8. [15] Copeland A, D’Angelo J, Dongre R, Belagutti S, Scholar G. Field evaluation of high reclaimed asphalt pavement-warm-mix asphalt project in Florida: case study. Transp Res Board 2010;2179:9. [16] Capitao SD, Picado-Santos LG, Martinho F. Pavement engineering materials: review on the use of warm-mix asphalt. Constr Build Mater 2012;36:9. [17] Kim H, Lee S-J, Amirkhanian SN. Rheology of warm mix asphalt binders with aged binders. Constr Build Mater 2011;25(1):7. [18] Miller JS, Bellinger WY. Distress identification manual for the long-term pavement performance program. 4th Revised ed. McLean (VA): Federal Highway Administration; 2003. p. 164.