materials Article

Towards Understanding the Polymerization Process in Bitumen Bio-Fluxes Jan B. Król 1 1 2

*

ID

, Łukasz Niczke 2 and Karol J. Kowalski 1, *

ID

Faculty of Civil Engineering, Warsaw University of Technology, 00-637 Warsaw, Poland; [email protected] General Directorate for National Roads and Motorways, 66-004 Zielona Góra, Poland; [email protected] Correspondence: [email protected]; Tel.: +48-22-234-6487

Received: 9 August 2017; Accepted: 5 September 2017; Published: 9 September 2017

Abstract: Bitumen is a commonly used material for road construction. According to environmental regulations, vegetable-based materials are applied for binder modification. Fluxed road bitumen containing a bio-flux oxidation product increases the consistency over time. The efficiency of crosslinking depends on the number of double bonds and their position in the aliphatic chain of fatty acid. The main goal of this paper was to examine the structural changes taking place during hardening bitumen with bio-flux additives. Two types of road bitumens fluxed with two different oxidized methyl esters of rapeseed oil were used in this study. Various chemical and rheological tests were applied for the fluxed-bitumen at different stages of oxygen exposure. The oxidation of rapeseed oil methyl ester reduced the iodine amount by about 10%–30%. Hardening of the fluxed bitumen generally results in an increase of the resins content and a reduction of the aromatics and asphaltenes. In the temperature range of 0 ◦ C to 40 ◦ C, bio-flux results with a much higher increase in the phase angle than in temperatures above 40 ◦ C in the bitumen binder. The increase in the proportion of the viscous component in the low and medium binder temperature is favorable due to the potential improvement of the fatigue resistance of the asphalt mixture with such binders. Keywords: bio-oil; bio-flux; oxypolymerization; reclaim asphalt pavement (RAP); warm mix asphalt (WMA)

1. Introduction Bitumen is a commonly used material for construction of flexible and semi-rigid pavements. Crude oil, from which bitumen is refined, is one of the natural resources. According to the sustainability policy, it is important to replace limited raw material with the one from renewable resources, e.g., vegetable-based material. Vegetable-based materials can be applied independently, fully or partially replacing traditional materials or supporting recycling of currently used materials. Plants are valuable source of materials that can be used to produce binders, liqueurs, or binder solvents. Among the plants that can be used for this are the trees and shrubs from which natural resins are derived or oilseeds to obtain oils. One of the bio-oil applications can be replacement of organic origin fluxes containing volatile substances. Fluxed asphalts are used in the technology of asphalt mixtures for the production of cold-mixed asphalt and for the repair and maintenance of pavements [1], or they can be used as asphalt substitutes or rejuvenators [2,3]. The experience of using rapeseed oils and animal-origin oils shows that it is possible to apply them to pavement repairs as bitumen fluxes, but bleeding and binder drain down in the pavement can potentially occur [4]. As the aforementioned implies, it is important to select the right amount of bio-oils for the bitumen and to investigate their properties in particular with regard to the climate in which this technology is used. Another application of vegetable additives is the rejuvenation of reclaimed asphalt pavement (RAP), which is constantly striving to increase RAP’s content while maintaining pavement durability. Materials 2017, 10, 1058; doi:10.3390/ma10091058

www.mdpi.com/journal/materials

Materials 2017, 10, 1058

2 of 12

When using high RAP content, it is necessary to renew the properties of the RAP-contained-bitumen by using refreshing additives. As previously shown [5–12], bio-oil can successfully play the role of such rejuvenators. Vegetable oils added to aged bitumen allow the reduction of the asphaltene content in the binder, but do not change its colloidal structure. It has been found that the addition of too high an amount of vegetable oil can lead to a reduction in the thermostability of the bitumen binder and to the increase of the rutting potential of the mixture [13]. However, by optimizing the composition of binders containing bio-oils it is possible to improve the fatigue life of bitumen and asphalt mixtures, as well as to reduce the bitumen brittle fracture temperature (in single-edged notched bending (SENB) test) and thermal stress temperature (in the thermal stress restrained specimen test (TSRST)) [6,13,14]; such rheological characterization is important in bitumen modification with bioadditives, polymers, and rubber from recycling or acids [15]. Using vegetable origin additives, in particular oils, it should be remembered that most of the current applications are intended to soften or flux the bitumen. This is a convenient solution due to the good mixability of vegetable oils with bitumen, a homogeneous product with reduced stiffness is obtained. Especially in the works [4,13,16,17] there are indications of restrictions in too high a content of these fluxes due to the potential problem of pavement properties destruction in a form of a permanent deformation. The problem of too high liquefaction of the bitumen binder can be reduced by polymerizing oils in the bitumen binder. It has been found that methyl esters of fatty acids, obtained by transesterification from vegetable oils, may be used as bitumen flux for this purpose [18]. In this case, the hardening of the binder is obtained not by evaporation but by crosslinking of the flux in the presence of oxygen. Oxidative polymerization of vegetable oils is a free radical chain reaction, includes initiation, propagation, and termination steps [19]: Initiation RH → R• + H•; Propagation R• + O2 → ROO•, ROO• + RH → ROOH + R•; Termination ROO• + R• → ROOR, R• + R• → RR Oxypolymerization leads to the crosslinking of their structural units [20–24]. The efficiency of crosslinking depends on the number of double bonds and their position in the aliphatic chain of fatty acid. Vegetable oils comprise a variable number of double bonds depending on the composition. The composition of the oil and, hence, its reactivity to oxypolymerization is genotype-specific and, therefore, variable and dependent on the climatic conditions [25,26]. Previous studies have shown that the liquid products obtained from oxidation of rapeseed oil methyl esters can be used as a bitumen flux [27,28]. Fluxed road bitumen containing this oxidation product increased its consistency over time, had a high ignition temperature, and displayed satisfactory adhesion to the aggregate. The limited amount of double bonds in rapeseed oil esters does not over-stiffen the asphalt mixture. It allows for a two-stage production process and full use of bio-oil potential. In the first production stage the bitumen is liquefied and refreshed, and in the second stage (service life) polymerization of the oil in the bitumen occurs. This type of two-stage technology can be used in sustainable road surfaces without compromising its performance [29,30]. The aim of the work presented in this paper was to examine the structural changes occurring during hardening bitumen with a bio-flux additive. 2. Materials and Methods 2.1. Materials Two types of road bitumens (35/50 and 70/100 grades) produced from Uralic crude oil and vegetable origin flux were used for the study. This bio-flux was prepared in a special oxidation process. The material for oxidation was rapeseed oil methyl ester (RME) obtained by transesterification of rapeseed oil. Some of the physicochemical properties are summarized in Table 1. Cobalt acetate (Co(CH3 COO)2 •4H2 O) was applied as a catalyst of oxidation while cumene hydroperoxide (C9 H12 O2 ) was used as a promoter of the reaction.

Materials 2017, 10, 1058 Materials 2017, 10, 1058

3 of 12 3 of 12

Table 1. Physicochemical properties of rapeseed oil methyl ester.

Properties Rapeseed Oil Methyl Ester Tableat1.25 Physicochemical properties of rapeseed oil methyl ester. Density °C, g/cm3 0.8787 Flash point, °C 197 Properties Rapeseed Oil Methyl Dynamic viscosity at 25 °C, Pas 0.006 Ester 0.8787116 at 25 g/cm IodineDensity number, g ◦I2C, /100 g 3 ◦ Flash point, C 197 0.29 Acid value, mg KOH/g◦

Dynamic viscosity at 25 C, Pa·s 0.006 Iodine number, g I2 /100 g 116 Cobalt acetate (Co(CH 3COO) •4H2O) was applied as a catalyst Acid value, mg2KOH/g 0.29 of oxidation while cumene

hydroperoxide (C9H12O2) was used as a promoter of the reaction. 2.2. Production Procedure 2.2. Production Procedure Methyl esters of rapeseed oil were oxidized at 20 ◦ C or 90 ◦ C in a 500 mL reactor, depending on Methyl esters of rapeseed oil were oxidized at 20 °C or 90 °C in a 500 mL reactor, depending on the composition (Version A or Version B, respectively). The composition and production parameters the composition (Version A or Version B, respectively). The composition and production parameters are shown in Table 2 and the reactor view is shown in Figure 1. are shown in Table 2 and the reactor view is shown in Figure 1. Table 2. Bio-agent types. Table 2. Bio-agent types. Compounds Compounds Rapeseed methyl Rapeseed methylesters esters(RME) (RME) Cobalt(II) acetate tetrahydrate catalyst Cobalt(II) acetate tetrahydrate catalyst Cumene hydroperoxide Cumene hydroperoxide Process parameters

Process parameters

Oxidation temperature Oxidation temperature Process duration

Process duration

Version A

Version A 98.9% 98.9% 0.1% 0.1% 1.0% 1.0% 20 ◦ C 20 °C 2h

2h

Version B

Version B

99.9% 99.9% 0.1% 0.1% -

-

90 ◦ C 90 °C 1h

1h

The ca.ca. 300 g of rapeseed oiloil methyl ester. TheThe air The glass glass column columnreactor reactor(1000 (1000mL) mL)was wasfilled filledwith with 300 g of rapeseed methyl ester. inlet waswas at the bottom of the and and air was dispersed through sintered glass. The The air flow rate air inlet at the bottom of reactor the reactor air was dispersed through sintered glass. air flow −1 h−1 ) increased the air-oil phase volume by about 50%. The catalyst used in the experiment (500 L kg rate used in the experiment (500 L kg−1 h−1) increased the air-oil phase volume by about 50%. The was added the raw in 0.1 wt in % 0.1 amounts (in terms (in of cobalt it was subjected catalyst wastoadded tomaterial the raw material wt % amounts terms content) of cobaltbefore content) before it was to oxidation. Catalytic oxidation was also performed in the presence of cumene hydroperoxide subjected to oxidation. Catalytic oxidation was also performed in the presence of (1%/wt) cumene used as a promoter of oxidation Theofliquid oxidation products forproducts analysis hydroperoxide (1%/wt) used as reactions. a promoter oxidation reactions. The were liquidcollected oxidation every h. The for volatile products were a scrubber with trichloroethylene was placed were 0.5 collected analysis every 0.5 trapped h. The in volatile products were trapped inwhich a scrubber with in a dewar vessel filled with a cooling mixture. trichloroethylene which was placed in a dewar vessel filled with a cooling mixture.

Figure 1. Laboratory setup for bio-agent production: reactor for oxidation of rapeseed oil esters. Figure 1. Laboratory setup for bio-agent production: reactor for oxidation of rapeseed oil esters.

Materials 2017, 10, 1058

4 of 12

Oxidized RME in the presence of reaction promoters was stored at 5 ◦ C in a sealed container without light and oxygen. Modifications of bitumen were carried out by heating the binder to 150 ◦ C in a laboratory oven. Then a bio-flux (at room temperature) was added to the hot bitumen and stirred with a glass baguette for about half a minute—methyl esters of rapeseed oil were well homogenized with the bitumen. No component separation was observed during the modification process. Test samples were taken immediately after modification, and the remaining binder was poured onto a flat tray in a thickness of approximately 1 mm for further conditioning over time. Samples were conditioned under room conditions. After 28 days, the binder was collected cold with a heated spatula and test specimens were formed. 2.3. Analysis Methods Concentration of organic peroxide in liquid oxidation products was assessed on the basis of the peroxide value determined according to the ISO 3960 standard. The acid value was tested according to the ISO 660 standard. Iodine amount was determined according to the ISO 3961 standard. Viscosity of the flux was measured using a Brookfield DV-II+ Pro apparatus (Brookfield Engineering Laboratories, Middleborough, MA, USA). Flash point of the flux was determined using a Marcusson flash-point apparatus (Labor Muszeripari Muvek, Esztergom, Hungary). Road bitumen modified with bio-additive was subjected to a dynamic viscosity test at 60 ◦ C using a Brookfield apparatus (according to EN 13302:2010) and tests with a dynamic shear rheometer MCR 101 (Anton Paar GmbH, Graz, Austria) were conducted according to AASHTO T 315 (Standard Method of Test for Determining the Rheological Properties of Asphalt Binder Using a Dynamic Shear Rheometer (DSR)). Tests in DSR apparatus were conducted in a set of two parallel 25-mm diameter plates, gap 1 mm, in a temperature range between 0 ◦ C and 100 ◦ C, with strain between 0.1% and 12%, depending on the test temperature. The main test method to evaluate to chemical changes during bitumen hardening was Fourier transform infrared spectroscopy (FTIR), combined with attenuated total reflection (ATR). ALPHA FTIR device was used with the diamond ATR crystal (Bruker Corporation, Billerica, MA, USA). Infrared spectroscopy was applied to the analysis of original bitumen, as well as the bitumen with bio-flux. The fractional composition of bitumens (asphaltenes, resins, aromatics, and saturates) was determined by the thin-layer chromatography with a flame-ionization detection (TLC-FID). For the analysis an IATROSCAN MK-5 (IATRON Laboratories, INC., Tokyo, Japan) apparatus equipped with Chromarod was used. 3. Results 3.1. Physicochemical Changes The changes in the peroxide value of liquid products with oxidation time are plotted in Figure 2a. The highest concentration of peroxides is observed after approximately 2 h of oxidation in the presence of a cobalt catalyst. The addition of cumene hydroperoxide into the raw material to be oxidized in the presence of a cobalt catalyst enhances the generation of peroxides at the initial stage of oxidation. The rise in their concentration observed during the first hour of oxidation is an indication of the prevalence of the rate of peroxide generation over their decomposition at this stage. The oxidation of rapeseed oil methyl esters without additives suggests a minimal increase in the peroxide value. The changes in acid value confirm the scission of fatty acid methyl esters with the formation of carboxylic groups. The oxidation time resulted in approximately linear changes in the acid value which approached 0.7 mg KOH/g after 2 h of oxidation (Figure 2b). Final physicochemical properties of oxidation liquid products are summarized in Table 3.

Materials 2017, 10, 1058

5 of 12 1

a)

Acid KHO/gmg KHO/g mgvalue, Acid value,

RME + 0.1% Co

1500 2000 Peroxide value, meq O2/kg

Peroxide value, meq O2/kg

Materials 2017, 10, 1058 2000 Materials 2017, 10, 1058 RME

a)

RME + 0.1% Co + 1.0% Cumem RME Hydroperoxide RME + 0.1% Co

1000 1500

RME + 0.1% Co + 1.0% Cumem Hydroperoxide

500 1000

0500 0

0.5

1 time, h

0

1.5

2

RME

0.5

1 peroxide time, (a) h

1.5

5 of 12

RME + 0.1% Co

0.8 1

RME + 0.1% Co + 1.0% Cumem RME

b)

Hydroperoxide

0.6 0.8

RME + 0.1% Co RME + 0.1% Co + 1.0% Cumem Hydroperoxide

0.4 0.6 0.2

0.4

0.20

0

0.5

1 time, h

0 0

5 of 12

b)

2

0

0.5

1.5

1

2

1.5

2

Figure 2. Changes in and acid (b) value of rapeseed oil methyl ester vs. time of time, h oxidation. Figure 2. Changes in peroxide (a) and (b) value of rapeseed oil methyl time of Figure 2. Changes in peroxide (a) and acid acid (b) value of rapeseed oil methyl ester vs.ester timevs. of oxidation. oxidation. Table 3. Physicochemical properties of the oxidation of liquid products. Table 3. Physicochemical properties of the oxidation of liquid products. Product of Oxidation Table 3. Physicochemical properties ofLiquid the oxidation of liquid products.

Properties

RME + 0.1% Co +Liquid 1% Cumene Product of Oxidation + 0.1% Co (Version B) Liquid Product ofRME Oxidation Hydroperoxide (Version RME + 0.1% Co + 1% Cumene A) RME + 0.1% Co (Version B) Properties RME + 0.1% Co(Version + 1% Cumene Hydroperoxide A) Dynamic viscosity at 25 °C, Pas 0.011 0.008 RME + 0.1% Co (Version B) Hydroperoxide (Version A) ◦ C, Pa·s Dynamic viscositygatI225 0.011 0.008102 Iodine number, /100 g 82 Dynamic 25 °C, 0.011 0.008 Iodineviscosity number, gat I2 /100 g Pas 82 102 AcidAcid value, mg KOH/g 0.68 0.67 value, mg KOH/g 0.68 0.67 Iodine number, g I2/100 g 82 102 ◦ Flash point, 136 Flash point,°CC 136 191 191 Acid value, mg KOH/g 0.68 0.67 Flash point, °C 136 191 Data in Table Table33suggest suggestan anincrease increaseininthe theviscosity viscosityofofliquid liquid products. The oxidation Data presented presented in products. The oxidation of of rapeseed oil methyl ester reduced the iodine number of the liquid products by about 10% and 30%. Data presented in Table 3 suggest an increase in the viscosity of liquid products. The oxidation rapeseed oil methyl ester reduced the iodine number of the liquid products by about 10% and 30%. The additionof of a promoter (hydroperoxide) results both in higher consistency and faster ofaddition rapeseed oil ester reduced the iodine number liquid products by about 10% and 30%. The amethyl promoter (hydroperoxide) results both of inthe higher consistency and faster consumption consumption of the double bonds in fatty acid methyl esters. addition of ain promoter (hydroperoxide) results both in higher consistency and faster of The the double bonds fatty acid methyl esters. consumption of the double bonds in fatty acid methyl esters. 3.2. 3.2. Structural Structural Changes Changes and and Fractional Fractional Composition Composition 3.2. Structural Changes and Fractional Composition Structural changesthat thathave have occurred upon hardening fluxed bitumen are shown in the Structural changes occurred upon hardening fluxed bitumen are shown in the infrared Structural changes that have occurred upon hardening fluxed bitumen are shown in the infrared spectra of original (neat) bitumen and binder fluxed by the oxidation of the liquid products spectra of original (neat) bitumen and binder fluxed by the oxidation of the liquid products of rapeseed infrared spectra of original (neat) bitumen and binder fluxed by the oxidation of the liquid products of rapeseed oil methyl esters in the presence of a cobalt catalyst (Figure 3). oil methyl esters in the presence of a cobalt catalyst (Figure 3). of rapeseed oil methyl esters in the presence of a cobalt catalyst (Figure 3). Properties

0.0

a)

0.0

a)

0.1

3440 3440 4000 4000

2940 2940 3500 3500 neat 70/100 neat 70/100

3000 3000

559

746

559

869 812869 746812

1026

1315

1160 1160 1105 1052 1105 1026 1052

1315

1458

0.06 0.06

1458

0.2 0.2

0.6 0.6 4500 4500

0.05 0.05

0.1

2924 2924

0.5 0.5

1377

0.04 0.04

2953

0.4 0.1

1377

0.03 0.03

0.0

1601 1601

0.0

2953

0.4

0.3

3314

0.3

3314

Absorbance

Absorbance

0.2

0.00 0.00 0.01 0.01 0.02 0.02

1744 1744

0.2

1978 1978

2855 2855

0.1

0.07 0.07 2000 2000

1500 1500

2500 2000 2500 2000 Wavenumber, cm-1-1 Wavenumber, cm

70/100++5.0% 5.0%bio-fluxe bio-fluxe 70/100

1500 1500

1000 1000 1000 1000

70/100 days) 70/100++5.0% 5.0%bio-fluxe bio-fluxe(after (after2828 days)

Figure 3. Cont.

500500

500500

0 0

Materials 2017, 10, 1058

6 of 12

Materials 2017, 10, 1058

6 of 12

0.0

b) Materials 2017, 10, 1058

6 of 12

0.1 0.0

0.00

0.6 4500

0.2

4000

3440

0.6 4500

4000

3500

neat 35/50

3500

2940

3000

0.06

2000

869 812 746

559

869 812 746

1160 1105 1052 1026

1315 1500

0.07

2500 2000 2000 Wavenumber, cm-1

35/50 + 5.0% bio-fluxe 3000

1000

1458

0.5

0.05 0.07

2924

2940

2924

3440

14581377

0.06 0.04

0.1

0.2

1160 1105 1052 1026

1744

0.05 0.03

2953

0.4

1315 1377

1978

2855

0.04

0.02 0.30.1 0.0

2953

0.5

0.03

0.01

3314

0.4

3314

0.2

0.00

1601

1744

0.02 0.0

0.3

Absorbance

Absorbance

0.1

559

b)

1601

0.01

1978

2855

0.2

1500

1000

1500

1500

500

1000

35/50 + 5.0% bio-fluxe (after 28 days)

2500 2000 Wavenumber, cm-1

500

1000

0

500

500

0

spectra ofneat neat bitumen bitumen by bio-fluxe rapeseed (a) 70/100 Figure 3. 3. FTIR FTIR spectra of bitumen and fluxed by oilester: methyl ester: 35/50neat 35/50and + 5.0% fluxed bio-fluxe 35/50bitumen + 5.0% (after rapeseed 28oil days)methyl bitumen; (b) 35/50 (a) 70/100and bitumen; andbitumen. (b) 35/50 bitumen. Figure 3. FTIR spectra of neat bitumen and fluxed bitumen by rapeseed oil methyl ester: (a) 70/100 bitumen; and (b) 35/50 bitumen.

The formation formation of of oxygen-bearing oxygen-bearing compounds compounds is is evident evident in in the the FTIR FTIR spectrum spectrum of of fluxed fluxed bitumen bitumen The −1 (νO–H) and after 28 days of hardening, which exhibits a broad absorbance region at 3500–3000 cm − 1 after 28 days of hardening, which exhibits a broad absorbance region 3500–3000 cm bitumen (νO–H) and The formation of oxygen-bearing compounds is evident in the FTIRatspectrum of fluxed is not not observed inthe the original sample. Additionally, the spectrum hardening bitumen after 28 daysinof hardening, which exhibits a broad absorbance region at of 3500–3000 cm−1 fluxed (νO–H) and is observed original sample. Additionally, the spectrum of hardening fluxed bitumen shows −1 shows a broadening of the bands centred at 1160 cm , corresponding to νC–O and δO–H vibrations − 1 is not observed in the original sample. Additionally, the spectrum of hardening fluxed bitumen a broadening of the bands centred at 1160 cm , corresponding to νC–O and δO–H vibrations of ethers of ethers alcohols. cm−1 is−1,almost alwaystoto assigned to δO–H theof formation of C– showsand a broadening ofaround thesignal bands centred atalmost 1160 cm corresponding νC–O and vibrations and alcohols. The signalThe 560around cm−1 is560 always assigned the formation C–C=O bonds −1C– −1 is of ethers and alcohols. The signal around 560 cm almost always assigned to the formation of C=O bonds in ketones and aldehydes. The visible broadening of the band around 1740 cm in the in ketones and aldehydes. The visible broadening of the band around 1740 cm−1 in the FTIR spectrum −1 C=O bonds inofketones aldehydes. The visible broadening of the bandstretching around 1740 cm in the FTIR spectrum fluxedand bitumen (Figure 3a,b), corresponds to the vibrations of the of fluxed bitumen (Figure 3a,b), corresponds to the stretching vibrations of the carbonyl group (νC=O) FTIR group spectrum of fluxed bitumen (Figureesters 3a,b), added corresponds to the stretching vibrations of the carbonyl (νC=O) in fatty acid methyl to bitumen. in fatty acid methyl esters added to bitumen. carbonyl group (νC=O) esters added to bitumen.rings of bitumen compounds. The The absorption absorption band in atfatty 1600acid cm−−1methyl the aromatic 1 derives from The band at 1600 cm fromthe thearomatic aromaticrings rings bitumen compounds. The absorption band at 1600 cm−1 derives derives from of of bitumen compounds. The The −1 region 2960–2850 2960–2850 cm cm−1isistypical for CH 2 and CH3 stretching of saturated compounds. region typical for CH and CH stretching of saturated compounds. −1 region 2960–2850 cm is typical for CH22 and CH3 3stretching of saturated compounds. The effect effectofofhardening hardening the chemical composition of fluxed binders are in present in the The on on the composition of fluxed binders are present theinTLC-FID The effect of hardening on chemical the chemical composition of fluxed binders are present the TLC-FID chromatograms chromatograms (Figure 4).(Figure 4). TLC-FID chromatograms (Figure 4). 15 15

a) a)

13 13

Resins Resins

FID signal / mVolt

FID signal / mVolt

11 11 9 7 5

9

Asphaltens

Asphaltens

7 Aromatics

5

Aromatics

3

3

Saturates 1

Saturates

1 -1 0.0

-1

0.1

0.0

0.1 neat 70/100

neat 70/100

0.2

0.2

0.3

Time, min

Time, min 70/100 + 5.0% bio-fluxe 70/100 + 5.0% bio-fluxe

Figure 4. Cont.

0.3

0.4

0.5

0.4

70/100 + 5.0% bio-fluxe (after 28 days)

70/100 + 5.0% bio-fluxe (after 28 days)

0.5

Materials 2017, 10, 1058

7 of 12

Materials 2017, 10, 1058

7 of 12

15

Resins

b)

13

FID signal / mVolt

11 9

Asphaltens

7 5

Aromatics

3 Saturates 1 -1 0.0

0.1

0.2

0.3

0.4

0.5

Time, min neat 35/50

35/50 + 5.0% bio-fluxe

35/50 + 5.0% bio-fluxe (after 28 days)

Figure 4. TLC-FID chromatograms andfluxed fluxedbitumen bitumen rapeseed oil methyl Figure 4. TLC-FID chromatogramsofofneat neatbitumen bitumen and byby rapeseed oil methyl ester:ester: (a) 70/100 bitumen; and (b)(b) 35/50 (a) 70/100 bitumen; and 35/50bitumen. bitumen.

The hardening of fluxed bitumen generally results in an increase of the resins content and a The hardening of fluxed bitumen generally results in an increase of the resins content and decrease of the aromatics and asphaltenes. However, the content of saturates change slightly due to a decrease of the aromatics and asphaltenes. However, the content of saturates change slightly due to their inert nature to oxygen.

their inert nature to oxygen.

3.3. Physical and Viscoelastic Properties of Fluxed Bitumen

3.3. Physical and Viscoelastic Properties of Fluxed Bitumen

70/100 road bitumen was modified with two bio-fluxes (Version A and B) to compare the binder 70/100 road bitumen modified two and B) to compare theasbinder hardening process overwas time. Figure 5with shows thebio-fluxes results of(Version dynamicAviscosity tests at 60 °C a hardening process over time. Figure 5 shows the results of dynamic viscositytheir testsviscosity at 60 ◦ Cover as atime. function function of air influence time. It can be observed that both samples increase initial viscosity values the twothat binders not different other because the The of airThe influence time. It can be of observed bothare samples increasefrom theireach viscosity over time. bio-fluxes in the initial phase (once added to the bitumen) causes the same fluxing effect. During the in initial viscosity values of the two binders are not different from each other because the bio-fluxes exposure of the binder to air there is a gradual reaction of saturating unsaturated bonds in the the initial phase (once added to the bitumen) causes the same fluxing effect. During the exposure bio-flux viscosity reaction of the bitumen. Over 50% viscosity increase 28 days was and of the binderand to increasing air there isthe a gradual of saturating unsaturated bondsafter in the bio-flux found for both the cobalt modifier, as well as the cobalt modifier with cumene hydroperoxide. increasing the viscosity of the bitumen. Over 50% viscosity increase after 28 days was found for both Hardening of bitumen is caused by the presence of bio-flux. As shown in another publication [28], the cobalt modifier, as well as the cobalt modifier with cumene hydroperoxide. Hardening of bitumen hardening of neat bitumen (in previous research evaluated based on the complex modulus G*), after is caused by was the presence of bio-flux. another publication hardening of shown neat bitumen 28 days equal to only 8%. DueAs toshown the plotinclarity, data for 70/100[28], bitumen were not in (in previous research evaluated based on the complex modulus G*), after 28 days was equal to (191 only 8%. Figure 5, due to a high difference in the dynamic viscosity at 60 °C between neat 70/100 bitumen Due Pa∙s) to theand plotfluxed clarity, data for bitumen werebetween not shown in Figure 5, due a high difference bitumen (970/100 Pa∙s). The difference the cure efficiency of to both fluxes is 7%. in ◦ C between neat 70/100 bitumen (191 Pa·s) and fluxed bitumen (9 Pa·s). the dynamic viscosity Due to the danger atof60 explosions at high temperature of the cumene hydroperoxide containing bio-flux, a bio-flux 90 °C and only cobalt siccative usedoffor further at The difference betweenproduced the cure at efficiency of containing both fluxes is 7%. Due to the was danger explosions studies (version B). high temperature of the cumene hydroperoxide containing bio-flux, a bio-flux produced at 90 ◦ C and

containing only cobalt siccative was used for further studies (version B). The efficiency of fluxing and curing of bio-fluxed binders was investigated using a dynamic shear rheometer (DSR). The tests were performed over a wide temperature range of 0 ◦ C to 100 ◦ C. Figure 6 shows the results of a complex modulus (G*) and phase shift (δ) determination for two binders with different consistencies, i.e., for soft 70/100 bitumen in Figure 6a, and for harder 35/50 bitumen in Figure 6b.

Dynamic viscosity at 60 °C, Pa·s Dynamic viscosity at 60 °C, Pa·s

16

70/100 + 5% (RME + 0.1% Co + 1.0% Cumene Hydroperoxide) - version A 70/100 + 5% (RME + 0.1% Co) - version B

14

Materials 2017, 10, 1058 Materials 2017, 10, 1058

8 of 12 8 of 12

12 16

70/100 + 5% (RME + 0.1% Co + 1.0% Cumene Hydroperoxide) - version A

10

70/100 + 5% (RME + 0.1% Co) - version B 14 8 0

1

3

7

12

14

28

time, day

Figure 5. Preliminary tests of reactivity of bio-fluxes in time based on dynamic viscosity test at 60 °C. 10

The efficiency of fluxing and curing of bio-fluxed binders was investigated using a dynamic 8 shear rheometer (DSR). The tests were performed over a wide temperature range of 0 °C to 100 °C. 0 1 3 7 14 28 Figure 6 shows the results of a complex modulus (G*) and phase shift (δ) determination for two time, day binders with different consistencies, i.e., for soft 70/100 bitumen in Figure 6a, and for harder 35/50 ◦ C. Figure on dynamic dynamic viscosity viscosity test testat at60 60°C. bitumen in 5. Figure 6b. tests Figure 5.Preliminary Preliminary testsof ofreactivity reactivityof of bio-fluxes bio-fluxes in in time time based based on

0.01 100

a)

40 90

fluxing process

Phase angle, °

Complex modulus, kPa

100 a) The efficiency of fluxing and curing of bio-fluxed binders was investigated90using a dynamic shear rheometer (DSR). The tests were performed over a wide temperature range80of 0 °C to 100 °C. 10 Figure 6 shows the results of a complex modulus (G*) and phase shift (δ) determination for two 70 binders with different1 consistencies, i.e., for soft reversible 70/100 bitumen in Figure 6a, and60 for harder 35/50 process bitumen in Figure 6b.0.1 50

30 80

0.001 10

10

20

30

40

50

60

70

Temperature, C 0.01

90

100

0 50

40 [G*] 70/100 + 5.0% bio-flux 30 [δ] neat 70/100 [δ] 70/100 + 5.0% bio-flux (after 28 days) 20

[G*] neat 70/100 fluxing [G*] 70/100 + 5.0%process bio-flux (after 28 days) [δ] 70/100 + 5.0% bio-flux

0.001

80

0.0001 100

10

0.00001 10

0 0

10

20

30

40

50

60

70

80

90

100

Complex modulus, kPa

Temperature, C

Complex modulus, kPa

b)

1

reversible [G*] 70/100 + 5.0% bio-flux process[δ] neat 70/100

[G*] neat 70/100 [G*] 70/100 + 5.0% bio-flux (after 28 days) [δ] 70/100 + 5.0% bio-flux

0.1

Phase angle, °

0

90 80 70 60

[δ] 70/100 + 5.0% bio-flux (after 28 days)

50 40 90

100 0.01

30 80

fluxing process

10 0.001

20 70

0.00011

reversible process

10 60

0.1 0.00001

050 0

10

20

30

40

50

60

70

80

90

100 40

Temperature, °C

0.01 [G*] neat 35/50 fluxing [G*] 35/50 + 5.0% bio-flux (after 28 days) [δ] 35/50 + 5.0% bio-flux process

0.001

Phase angle, °

b)

0.00001 0.1

10 60

reversible process

[G*] 35/50 + 5.0% bio-flux [δ] neat 35/50 [δ] 35/50 + 5.0% bio-flux (after 28 days)

0.0001

Phase angle, °

Complex modulus, kPa

20 70 0.00011

30 20 10

Figure 6. bitumen based based on on the the dynamic dynamic shear shear test: test: Figure 6. Fluxing and reversibility processes of bio-fluxed bitumen 0.00001 0 (a) bitumen; and(b) (b)35/50 35/50bitumen. bitumen. (a) 70/100 70/100 bitumen; and 0

10

20

30

40

50

60

70

80

90

100

Temperature, °C neatbe 35/50 [G*] 35/50 + 5.0% bio-flux Based on Figure 6 it[G*] can stated that the bio-flux additive reduces the stiffness of the bituminous [G*] 35/50 + 5.0% bio-flux (after 28 days) [δ] neat 35/50 binder in a wide temperature range for both 70/100 and 35/50 binders. Binder [δ] 35/50 + 5.0% bio-flux [δ] 35/50 + 5.0% bio-flux (after 28 days)modification with 5% addition of the bio-flux results in a decrease in the value of the complex modulus by an order of

Figure 6. Fluxing and reversibility processes of bio-fluxed bitumen based on the dynamic shear test: (a) 70/100 bitumen; and (b) 35/50 bitumen.

Materials 2017, 10, 1058

9 of 12

magnitude (G*) of 1 × 10−1 kPa over the entire range of tested temperatures. It should be emphasized that the G* change by 1 × 10−1 kPa corresponds approximately to one change of one bitumen grade (according to the EU bitumen classification system). Taking into account the logarithmic nature of changes in the properties of asphalt binders, this level of change should be regarded as typical. The results of the phase shift (δ) determinations shown in Figure 6 also indicate that the addition of 5% bio-flux causes a change in its value throughout the wide temperature range. This change is different as compared to the aforementioned, however, and depends on the temperature value. In the temperature range of 0 ◦ C to 40 ◦ C, bio-flux produces a much higher increase in the phase angle than in temperatures above 40 ◦ C in the bitumen binder. The increase in the proportion of the viscous component in the low and medium binder temperatures is favorable due to the potential increase of the fatigue resistance of the asphalt mixture with such binders. 4. Discussion Catalytic oxidation of the rapeseed oil methyl ester results in a consistency increase, which should be attributed to the disappearance of the most reactive double bonds after two hours of oxidation. This finding is confirmed by the decrease in the iodine value. The observed fall in the content of double bonds in rapeseed oil methyl ester upon oxidation gives evidence that some part of these bonds was engaged in oligomerization and/or was oxidatively decomposed. The viscosity weight is about 1.3 and 1.8 times that of the raw material, which substantiates only a partial oligomerization of the molecules of unsaturated fatty acid methyl esters. Oligomers are formed mainly at the stage of peroxide generation. It can be assumed that peroxides are intermediate structures in the oxyoligomerization of unsaturated fatty acid methyl esters. A higher viscosity of the liquid oxidation products is attained when the reaction is carried out in the presence of cumene hyperoxide. When the flux containing this hydroperoxide is heated during flash point determination, cumene hydroperoxide decomposes (its half-life temperature is decreased by the cobalt catalyst), and the hydroxyl radicals formed initiate scission of the double bonds. This results in the formation of low molecular compounds with a flash point lower than those of the esters. Oxidized rapeseed oil methyl ester is a reactive flux, and hardening is possible once mixed with bitumen. After 28 days of conditioning, structural changes are confirmed by effective formation of oxygen organic compounds in fluxed binders. The absorption bands of hydroxyl, aldehydes, ketones, ether, and carboxylic group are evidenced in the FTIR spectra. The hardening of the binder after application is obtained by further polymerization of the fluxes’ vegetable-based components during consumption of molecular oxygen from the air. The reaction schemes for the formation of higher molecular and decomposition products are presented in Figure 7. The kinetic reaction of oxidative polymerization in the bituminous medium depends on the diffusion of oxygen and it is slower in binders with higher consistency. The observed intensity bands of the oxygen functional group are more noticeable in the spectrum of fluxed 70/100 bitumen than for harder 35/50 binder. Similar to the FTIR observation are the results of the fractional analysis of the fluxed binders. After 28 days of conditioning the increased concentration of resins suggests higher molecular polarity in fluxed binder. The changes in the fractional composition with reduction of the asphaltenes content improves the colloidal stability of the fluxed bitumens after stabilization. Based on the rheological curve shown in Figure 6 (dashed lines), changes in the complex modulus curve (G*) and phase angle (δ) can be observed for the binders exposed to air for 28 days. Exposure to the air of the thin layer of bitumen causes the oxypolymerization reaction. It can be seen that both asphalt 70/100 and asphalt 35/50 are partially reconstructed for the original properties of the bitumen. This can be determined by analyzing the position of the binder curves after 28 days (dashed line), which runs between clean bitumen and bitumen with added bio-flux tested immediately after mixing. The shift of G* and δ curves for bio-fluxed bituminous binders towards the clean bitumen curve

Materials 2017, 10, 1058

10 of 12

demonstrates overlapping polymerization in the modified bituminous binder under the influence of oxygen the air. Materials 2017,from 10, 1058 10 of 12 O H 3C O -H* O CH* H 3C O O2 O O*

O

CH H3C O RH O OH

O

CH H 3C O

O

O

OH

CH H 3C O

O*

O

CH

+

H 3C O

O

RH

*OH

O*

O

OH

O*

O

CH

CH CH

H 3C O

H3C O

H3C O

+

O

CH* H3C O

CH

*CH 2

H 3C O CH

H 3C O

O

O O

O2

O

O

O

CH

C

H 3C O

H 3C O

Polymerization

RH O

Formation oxygen compounds

OH CH 3

Decomposition

Figure 7. Reaction pathways of catalytic oxidation of oleic acid methyl ester.

Figure 7. Reaction pathways of catalytic oxidation of oleic acid methyl ester.

5. Conclusions

Based on the rheological curve shown in Figure 6 (dashed lines), changes in the complex on(G*) the conducted and(δ) analysis, following be drawn: modulusBased curve and phasetests angle can be the observed forconclusions the binderscan exposed to air for 28 days. Exposure to the air of the thin layer of bitumen causes the oxypolymerization reaction. It can be seen The most characteristic changes in the rapeseed oil methyl ester structure upon catalytic oxidation 1. that bothare asphalt 70/100 of and asphalt 35/50 are reconstructed for the originaland properties the formation organic peroxides andpartially hydroperoxides (intermediate products) the loss of the bitumen. This can be determined by analyzing the position of the binder curves after 28 days of the unsaturated bonds in the fatty acid methyl ester. (dashed line), which runs between bitumen andcatalyst bitumen added bio-flux 2. The cumene hydroperoxide in theclean presence of a cobalt is a with more effective promoter tested of immediately after mixing. The shift of G* and δ curves for bio-fluxed bituminous binders towards oxidative polymerization. the 3.cleanThe bitumen curve demonstrates in the with modified bituminous pocess is time-dependent. Theoverlapping hardening ofpolymerization the bio-fluxed binders time results in binder under the influence of oxygen from the air. oxidative polymerization process of unsaturated fatty acid methyl ester with the formation of oxygen compounds. 5. Conclusions 4. The application of oxidized rapeseed oil methyl esters to bitumen reduces the asphaltene content and stabilizes the colloidal structure of the binder. Based on the conducted tests and analysis, the following conclusions can be drawn: 5. Twenty-eight day oxygen exposure to bio-fluxed bitumen results in approximately 25% recovery 1. Theofmost characteristic changes in the rapeseed oil methyl ester structure upon catalytic the original binder stiffness. are the formation of organic peroxides and hydroperoxides 6.oxidation The fuxing effect is particularly advantageous in the temperature range of(intermediate 0 ◦ C to 40 ◦ C, products) which the fatigue lifeinofthe thefatty bitumen andcan thepotentially loss of theincrease unsaturated bonds acidbinder. methyl ester.

2. 3.

4. 5. 6.

The cumene hydroperoxide in the presence of a cobalt catalyst is a more effective promoter of oxidative polymerization. Acknowledgments: The research leading to these results received funding from the European Union’s Seventh Program for research, technological development, and demonstration under grant agreement No.time 603862. This in The pocess is time-dependent. The hardening of the bio-fluxed binders with results work was also supported by funds for science for years 2014–2017 granted by the Polish Ministry of Science and oxidative polymerization processprojects. of unsaturated acid ester with the Higher Education to support international This articlefatty reflects onlymethyl the author’s views, and theformation European of Union is not liable for any use that may be made of the information contained. The authors wish to express their oxygen compounds. gratitude to Piotr Radziszewski and Irena Gaweł for their inspiration. The application of oxidized rapeseed oil methyl esters to bitumen reduces the asphaltene content and stabilizes the colloidal structure of the binder. Twenty-eight day oxygen exposure to bio-fluxed bitumen results in approximately 25% recovery of the original binder stiffness. The fuxing effect is particularly advantageous in the temperature range of 0 °C to 40 °C, which

Materials 2017, 10, 1058

11 of 12

Author Contributions: Jan B. Król, and Karol J. Kowalski conceived and designed the experiments; Łukasz Niczke analyzed the physiochemical changes, FTIR results, and described the oxypolimerysation reaction; Jan B. Król and Karol J. Kowalski analyzed the viscoelastic properties of bitumen; and Jan B. Król, Łukasz Niczke, and Karol J. Kowalski wrote and edited the paper. Conflicts of Interest: The authors declare no conflict of interest. The founding sponsors had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

References 1.

2. 3. 4. 5.

6. 7. 8. 9.

10. 11. 12.

13. 14. 15.

16.

17. 18. 19. 20.

Simonen, M.; Pellinen, T.; Valtonen, J.; Blomberg, T. Biofluxed bitumens—Experiences from field tests and further developments. In Proceedings of the 5th Eurasphalt & Eurobitume Congress, Istanbul, Turkey, 13–15 June 2012; Paper No. A5EE-318. Peralta, J.; Williams, R.C.; Rover, M.; Silva, H.M. Development of a rubber-modified fractionated bio-oil for use as noncrude petroleum binder in flexible pavements. Transp. Res. Circ. 2012, E-C165, 23–36. Peralta, J.; Silva, H.M.; Williams, R.C.; Rover, M.; Machado, A.V. Development of an innovative bio-binder using asphalt-rubber technology. Int. J. Pavement Res. Technol. 2013, 6, 447–456. Guarin, A.; Khan, A.; Butt, A.A.; Birgisson, B.; Kringos, N. An extensive laboratory investigation of the use of bio-oil modified bitumen in road construction. Constr. Build. Mater. 2016, 106, 133–139. [CrossRef] Tanghe, T.; Lemoine, G.; Nösler, I.; Kloet, B. Influence of rejuvenating additives on recycled asphalt (RAP) properties. In Proceedings of the 5th Eurasphalt & Eurobitume Congress, Istanbul, Turkey, 13–15 June 2012; Paper No. P5EE-416. Kowalski, K.J.; Król, J.B.; Bankowski, ´ W.; Radziszewski, P.; Sarnowski, M. Thermal and fatigue evaluation of asphalt mixtures containing RAP treated with a bio-agent. Appl. Sci. 2017, 7, 2016. [CrossRef] Majid, Z.; Esmaeil, A.; Hallizza, A.; Mohamed, R.K. Investigation of the possibility of using waste cooking oil as a rejuvenating agent for aged bitumen. J. Hazard. Mater. 2012, 233, 254–258. Hugener, M.; Partl, M.N.; Morant, M. Cold asphalt recycling with 100% reclaimed asphalt pavement and vegetable oil-based rejuvenators. Road Mater. Pavement Des. 2014, 15, 239–258. [CrossRef] Zaumanis, M.; Mallick, R.B.; Poulikakos, L.; Frank, R. Influence of six rejuvenators on the performance properties of Reclaimed Asphalt Pavement (RAP) binder and 100% recycled asphalt mixtures. Constr. Build. Mater. 2014, 71, 538–550. [CrossRef] Hassan, N.A.; Khan, R.; Raaberg, J.; Presti, D.L. Effect of mixing time on reclaimed asphalt mixtures: An investigation by means of imaging techniques. Constr. Build. Mater. 2015, 99, 54–61. [CrossRef] Porot, L.; Broere, D.; Wistuba, M.; Grönniger, J. Asphalt and binder evaluation of asphalt mix with 70% reclaimed asphalt. Road Mater. Pavement Des. 2017, 18, 66–75. [CrossRef] Bocci, E.; Grilli, A.; Bocci, M.; Gomes, V. Recycling of high percentages of reclaimed asphalt using a bio-rejuvenator—A case study. In Proceedings of the 6th Eurasphalt & Eurobitume Congress, Prague, Czech, 1–3 June 2016; Paper No. EE.2016.334. Chen, M.; Leng, B.; Wu, S.; Sang, Y. Physical, chemical and rheological properties of waste edible vegetable oil rejuvenated asphalt binders. Constr. Build. Mater. 2014, 66, 286–298. [CrossRef] Lei, Z.; Bahia, H.; Yi-Qiu, T. Effect of bio-based and refined waste oil modifiers on low temperature performance of asphalt binders. Constr. Build. Mater. 2015, 86, 95–100. [CrossRef] Behnood, A.; Olek, J. Rheological properties of asphalt binders modified with styrene-butadiene-styrene (SBS), ground tire rubber (GTR), or polyphosphoric acid (PPA). Constr. Build. Mater. 2017, 151, 464–478. [CrossRef] Somé, C.; Pavoine, A.; Chailleux, E.; Andrieux, L.; DeMarco, L.; Da Silva Philippe, B.S. Rheological behavior of vegetable oil-modified asphaltite binders and mixes. In Proceedings of the 6th Eurasphalt & Eurobitume Congress, Prague, Czech, 1–3 June 2016; Paper No. EE.2016.222. [CrossRef] Fini, E.H.; Hosseinnezhad, S.; Oldham, D.J.; Chailleux, E.; Gaudefroy, V. Source dependency of rheological and surface characteristics of bio-modified asphalts. Road Mater. Pavement Des. 2017, 18, 408–424. [CrossRef] Pasquier, M. Bituminous Binder, Composition and Use. U.S. Patent 6,156,113, 5 December 2000. Wexler, H. Polymerization of drying oils. Chem. Rev. 1964, 64, 591–611. [CrossRef] Mallégol, J.; Gonon, L.; Lemaire, J.; Gardette, J.L. Long-term behaviour of oil-based varnishes and paints 4. Influence of film thickness on the photooxidation. Polym. Degrad. Stab. 2001, 72, 191–197. [CrossRef]

Materials 2017, 10, 1058

21. 22. 23.

24. 25. 26. 27. 28. 29.

30.

12 of 12

Stiepanowicz-Orchimienko, I.; Wasiljewicz-Wiercholancew, W. Chemia i Technologia Substancji Błonotwórczych; WNT: Warszawa, Poland, 1982; pp. 298–334. (In polish) Mallégol, J.; Lemaire, J.; Gardette, J.L. Drier influence on the curing of linseed oil. Prog. Org. Coat. 2000, 39, 107–113. [CrossRef] Mallégol, J.; Gonon, L.; Commereuc, S.; Verney, V. Thermal (DSC) and chemical (iodometric titration) methods for peroxides measurements in order to monitor drying extent of alkyd resins. Prog. Org. Coat. 2001, 41, 171–176. [CrossRef] Muizebelt, W.J.; Hubert, J.C.; Nielen, M.W.F.; Klaasen, R.P.; Zabel, K.H. Crosslink mechanisms of high-solids alkyd resins in the presence of reactive diluents. Prog. Org. Coat. 2000, 40, 121–130. [CrossRef] Gorski, W.; Ostaszewski, W.; Wislicki, B. Chemical composition of vegetable oils as potential biomass for production of fuel and lubricants oils. Paliwa Oleje Smary Eksploat. 2002, 96, 14–18. Bhat, S.V.; Nagasampagi, B.A.; Sivakumar, M. Chemistry of Natural Products; Narosa: New Delhi, India, 2005; pp. 206–213. Gaweł, I.; Piłat, J.; Radziszewski, P.; Niczke, Ł.; Krol, J.; Sarnowski, M. Bitumen fluxes of vegetable origin. Polimery 2010, 55, 55–60. Król, J.B.; Kowalski, K.J.; Niczke, Ł.; Radziszewski, P. Effect of bitumen fluxing using a bio-origin additive. Constr. Build. Mater. 2016, 114, 194–203. [CrossRef] Kowalski, K.J.; Król, B.; Radziszewski, P.; Piłat, J.; Sarnowski, M. New concept of sustainable road structure with RAP binder course using bio-agent. In Proceedings of the 6th Eurasphalt & Eurobitume Congress, Prague, Czech, 1–3 June 2016; Paper No. EE.2016.260. [CrossRef] Kowalski, K.J.; Król, J.; Radziszewski, P.; Casado, R.; Blanco, V.; Pérez, D.; Viñas, V.M.; Brijsse, Y.; Frosch, M.; Le, D.M.; et al. Eco-friendly materials for a new concept of asphalt pavement. Transp. Res. Procedia 2016, 14, 3582–3591. [CrossRef] © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).