Influence of Processing Temperature on the Modification Route and

Aug 15, 2011 - Phone: +34 959 21 99 89. ... This work deals with the effect that processing temperature exerts on the modification of bitumen by thiou...
1 downloads 6 Views 3MB Size
ARTICLE pubs.acs.org/EF

Influence of Processing Temperature on the Modification Route and Rheological Properties of Thiourea Dioxide-Modified Bitumen A. A. Cuadri, P. Partal,* F. J. Navarro, M. García-Morales, and C. Gallegos Dpto. Ingeniería Química, Facultad de Ciencias Experimentales, Campus “El Carmen”, Universidad de Huelva, Huelva 21071, Spain ABSTRACT: This work deals with the effect that processing temperature exerts on the modification of bitumen by thiourea dioxide (ThD). In that sense, shear and bending rheological tests, (thermal analysis thermogravimetric analysis/derivative thermogravimetric analysis/differential thermal analysis (TG/DTG/DTA) and modulated differential scanning calorimetry (MDSC)), and chemical characterization by thin layer chromatography coupled with a flame ionization detector (TLC-FID) were carried out on neat bitumen with a penetration grade (pen.) of 40/50 and corresponding 9 wt % ThD modified samples, prepared at 90, 130, and 180 °C. Tests carried out on those binders revealed the existence of a different modification mechanism at every temperature. In any case, viscosity was always seen to increase after ThD addition. Also, modification enhanced elastic properties and thermal susceptibility at high in-service temperatures and yielded improved binder resistance to thermal cracking by a decrease in its glass transition temperature. These results suggest that the use of thiourea dioxide may become a promising alternative to the use of other chemical modifiers for the paving industry.

1. INTRODUCTION Bitumen, as a residue from the distillation of heavy and light crude oils, is a complex material basically composed of hydrocarbons along with some other molecules, which contain small percentages of heteroatoms (sulfur, nitrogen, and oxygen). Bitumen compounds can be classified by chromatographic techniques into four different fractions (usually referred to as SARAs): saturates (S), aromatics (A), and resins (R), which make up the maltenes and asphaltenes (As). The complexity, aromaticity, heteroatom content, and molecular weight increase in the order S < A< R < As.1 A colloidal model, with asphaltenes being dispersed into an oily matrix of maltenes and surrounded by a shell of resins whose thickness is temperature-dependent, is traditionally used to describe bitumen behavior. Accordingly, the physicochemical and rheological properties of bitumen strongly depend on both temperature and the relative proportion of the SARAs fractions.2 On account of its properties, bitumen is the most suitable material to be used as a binder of mineral aggregates for paving applications,3 and consequently, roads are mainly constructed using a composite mixture of bitumen (∼5 wt %) and mineral aggregates. Despite its small proportion, bitumen forms the continuous matrix and is the only deformable component in the pavement. Consequently, it mainly controls the performance of a road.4 Unfortunately, even the best designed and constructed road pavements deteriorate over time under the combined effects of traffic loading and weathering, with the most common distresses being rutting5 (or permanent deformation of the pavement at high temperatures) and thermal cracking6 (or thermal fracture due to its lack of flexibility at low temperatures). Hence, bitumen performance has traditionally been improved through the addition of either virgin polymers (SBS, SBR, EVA, etc.) or waste polymers (plastics from agriculture, crumb tire rubber, etc.).7 An alternative to the above-commented modifiers is the use of nonpolymeric reactive agents, such as sulfur, polyphosphoric acid (PPA),8 a mineral acid,9 or organic molecules,10 which are r 2011 American Chemical Society

able to form chemical bonds with bitumen compounds. On these grounds, this work deals with the use of thiourea dioxide (ThD) as a potential chemical agent to be used in bitumen modification. Although well-known in many other industrial applications (as a reducing agent), it may become a promising approach in the asphalt industry. Following up PPA modification, which promotes the development of bridging/phosphorylation/condensation reactions,8 bitumen modification by ThD is expected to occur on a reducing/decomposition action that may lead to different pathways depending on the selected processing conditions. In this sense, ThD might represent an alternative to PPA, with the additional benefit of anticorrosion properties of ThD in opposition to storage tanks corrosion. A previous work demonstrated that modification by ThD, at 130 °C,11 which involves the decomposition of ThD into urea, sulfur, and sulfur dioxide, may enhance bitumen performance at both high and low temperatures. Looking further into this novel modifying agent, interest is now focused on the effect that temperature may exert on the modification mechanism observed, as well as the enhancement of properties thereof derived. With this aim, 40/50 pen. neat bitumen was modified, at three different temperatures (90, 130, and 180 °C), by adding 9 % ThD by weight of the modified binder. Viscous flow and oscillatory shear behavior, dynamic mechanical thermal analysis (DMTA), thermal analysis (DTG/DTA and modulated differential scanning calorimetry (MDSC)), and chemical composition (by thin layer chromatography coupled with a flame ionization detector (TLC/FID)) of the resulting binders were studied.

2. EXPERIMENTAL SECTION Bitumen with a penetration grade of 40/50, supplied by Construciones Morales, S.A. (Spain), has been used as the base material for the Received: May 31, 2011 Revised: August 12, 2011 Published: August 15, 2011 4055

dx.doi.org/10.1021/ef200801h | Energy Fuels 2011, 25, 4055–4062

Energy & Fuels

ARTICLE

Table 1. Penetration, Ring and Ball Softening Temperature, SARAs Fractions, and Colloidal Index for the Neat Bitumen Studied neat a

penetration (dmm)

49

ring and ball softening point (°C)b

53.5

saturates (wt %)

6.2

aromatics (wt %)

50.3

resins (wt %) asphaltenes (wt %)

24.5 19.0

colloidal index (CI) a

0.34

According to ASTMD5.12 b According to ASTMD36.13

modification. The results of technological tests (penetration grade and softening temperature, according to ASTM D512 and D36,13 respectively) and the chemical composition, in terms of SARAs fractions, are shown in Table 1. The bitumen SARAs fractions were determined by thin layer chromatography coupled with a flame ionization detector (TLC/FID), using an Iatroscan MK-6 analyzer (Iatron Corporation Inc., Japan). The elutions were performed in hexane, toluene, and dichloromethane/methanol (95/5), following the procedure outlined elsewhere.14 Thiourea dioxide (ThD), supplied by Sigma Aldrich, has been the selected chemical agent used for bitumen modification. It melts at approximately 124127 °C and has a molecular weight of 108.12 g/mol. Blends of bitumen with 9 wt % ThD were prepared for 1 h, in a cylindrical vessel (60 mm diameter, 140 mm height) and at three different temperatures (90, 130, and 180 °C). A low-shear mixer IKA RW20 (Germany), equipped with a four-blade impeller, was used at a rotating speed of 1200 rpm. Just after processing, a part of this binder was poured onto aluminum foil, forming a thin layer which was exposed to ambient conditions (curing period) for up to 60 days. Additionally, a reference bituminous binder containing 3 wt % commercial SBS Kraton D1101 was prepared, in the same mixing device, for 2 h, at 180 °C and 1200 rpm. The specific effects of temperature and agitation on neat bitumen were assessed by subjecting neat bitumen, without ThD addition, to the same processing conditions as described above (referred to as the “blank” sample hereinafter). Finally, because sulfur is formed during the ThD decomposition reaction, a 2 wt % sulfur modified bitumen was also processed under the same conditions (2 wt % sulfur is the equivalent quantity produced after reaction R1; see below). This study will allow us to establish its effect on the rheological response of the modified binder. Viscous flow measurements, at 60 °C, were carried out in a controlledstrain ARES rheometer (Rheometric Scientific, U.S.). Temperature sweep tests in oscillatory shear, at a heating rate of 1 °C/min, a frequency of 10 rad/s, and a 1% strain within the linear viscoelastic region, were conducted in a controlled-stress rheometer Physica MCR-301 (Anton Paar, Austria), between 30 and 100 °C. For both types of tests, a serrate plate-and-plate geometry (25 mm diameter and 1 mm gap) was used. All tests were carried out at least twice, in order to ensure the repeatability of the results. Dynamic mechanical thermal analysis (DMTA) tests were performed on neat and modified bitumens with a Seiko DMS 6100 (Seiko Instruments Inc., Japan) in double cantilever bending mode, using 50  10  3 mm3 specimens and liquid nitrogen as the cooling system. Temperature sweep tests were conducted at a constant frequency of 1 Hz and strains within the linear viscoelasticity region (LVE), by selecting a temperature ramp of 2 °C 3 min1 and a temperature interval from 35 to 35 °C. Modulated differential scanning calorimetry (MDSC) was performed with a TA Q-100 (TA Instruments, U.S.). Samples (510 mg) were

Figure 1. Viscous flow curves, at 60 °C, for 9 wt % ThD modified binders, processed at 90 °C (A) and 180 °C (B), as a function of curing time.

Table 2. Evolution with Curing Time of Zero-Shear Limiting Viscosity, η0, and Critical Shear Rate, γ_ C, for 9 wt % Thd Modified Binders, Processed at 90 and 180 °Ca

Tproc = 90 °C

neat blank noncured

Tproc = 180 °C

η0 (Pa 3 s)

γ_ c (s1)

598.81 622.28

3.6 3.1

720.55

2.4

20 days

1431.12

1.1

60 days

1744.06

1.0

696.00

3.2

noncured 20 days

1068.20 1663.07

2.0 1.2

60 days

2178.50

1.0

blank

a

Neat bitumen and the corresponding blanks at both temperatures included as reference.

subjected to the same testing procedure: a temperature interval between 80 and 100 °C; a heating rate of 5 °C/min; an amplitude of modulation (0.5 °C, with a period of 60 s; and nitrogen as purge gas, with a flow rate of 50 mL/min. To provide the same recent thermal history, all the samples were placed into hermetic aluminum pans for 24 h before measurement. Simultaneous DTG/DTA measurements, using a Seiko TG/DTA6200, both in T-ramp (10 °C 3 min1; between 40 and 600 °C), were carried out under inert atmosphere on 510 mg samples of ThD, neat bitumen, and modified bitumen.

3. RESULTS AND DISCUSSION 3.1. Influence on Viscous Flow Behavior. Viscous flow curves, at 60 °C, for 9 wt % ThD modified bitumen binders, processed at two different temperatures (Tproc = 90 and 180 °C), are presented in Figure 1, as a function of curing time. Neat bitumen and the corresponding blanks, at both temperatures, have been included as references. These samples always present a Newtonian region with a constant viscosity at low shear rates, η0, followed by shear-thinning region above a “critical” shear rate value, γ_ c. Table 2 reveals a decrease in the values of this parameter with curing time, and so, it reveals the development of a more complex microstructure, which is more susceptible to shear forces.15 4056

dx.doi.org/10.1021/ef200801h |Energy Fuels 2011, 25, 4055–4062

Energy & Fuels

ARTICLE

Figure 2. Evolution of the modification index with curing time for 9 wt % ThD modified binders at the three selected processing temperatures.

Figure 1 shows that, for any of the processing temperatures presented, ThD addition always leads to a viscosity increase, which is further enhanced when samples are stored at ambient conditions for a period of up to 12 months. Thus, after 60 days of curing, the sample processed at 90 °C presents viscosities very similar to those of the reference sample, containing 3 wt % SBS (commonly used in paving applications), and at 180 °C, its viscosity is even higher. After 12 months, any of the two processing temperatures studied gives rise to ThD-modified binders with higher viscosity than that measured for the 3% SBS modified bitumen. As can be observed for the blank sample processed at 180 °C, binder manufacture itself causes an increase in viscosity, as a result of the oxidation of the maltene fraction,16 showing very similar viscosities to those observed for the noncured 9 wt % ThD modified bitumen prepared at 90 °C, for which oxidation is negligible. In that sense, the evolution of η0, at 60 °C, with curing time, shown in Figure 1, describes the absolute degree of bitumen modification at high in-service temperature. Instead, viscosity changes due only to ThD addition (regardless of the expected bitumen oxidation during processing) can be quantified by means of a modification index (MI), which is defined as follows: MI ¼

η0, mod  η0, blank 100 η0, blank

ð1Þ

where η0,mod and η0,blank refer to modified and blank bitumen viscosities, respectively. As observed in Figure 2, which presents the evolution of MI with curing time for bitumen modification carried out at 90, 130, and 180 °C, a remarkable increase in MI, larger than that merely induced by bitumen oxidation, results from ThD addition. It can also be noticed that, for noncured samples, the MI increase undergone by the samples becomes more significant as temperature increases. However, Figure 2 demonstrates that long-term bitumen modification is quite a more complex issue, as deduced from the results obtained for the sample processed at 90 °C after 20 days of curing, which exceeds the MI value of the sample processed at 130 °C; after 60 days, it displays a higher modification degree than that found for the sample containing 3% SBS; and after 12 months, it shows an even higher modification index than that measured for the system processed at 180 °C. Consequently, this suggests, for every temperature studied, the formation of different reaction products that would evolve with time in a different way. 3.2. Influence on the Linear Viscoelastic Behavior. The linear viscoelastic behavior of the ThD modified binders was

Figure 3. Evolution with temperature of the (A) linear viscoelastic moduli in oscillatory shear (G0 and G00 ) for noncured 9 wt % ThD modified binders processed at different temperatures and (B) loss tangent for 9 wt % ThD modified binders, processed at 90 °C, as a function of curing time.

Figure 4. Evolution with temperature of the linear viscoelastic moduli, in dynamic bending (E0 and E00 ), for 9 wt % ThD modified binders cured for 60 days and processed at 130 °C (A) and 90 °C (B). Blank samples included as reference.

studied by means of temperature sweep tests in oscillatory shear (high temperatures) and bending (low temperatures). Figure 3A shows the evolution of G0 and G00 (oscillatory shear mode) with processing temperature (90, 130, and 180 °C) for noncured samples. It can be seen that G0 and G00 values monotonously decrease with increasing testing temperature from 35 up to 90 °C. Moreover, the viscous modulus clearly prevails over the elastic one in the entire temperature interval tested, which points out a predominantly viscous behavior. In addition, an increase in the viscoelastic moduli of the noncured samples is also observed as the processing temperature increases from 90 to 180 °C (results similar to those found for the viscosity, shown in Figure 1). The evolution of G0 and G00 was also studied as a function of curing time for ThD modified binders processed at 90 °C. Figure 3B makes clear the prevailing viscous behavior of the samples, with loss tangent values always higher than 1. However, an enhancement in the elastic features is observed with increasing curing time, as deduced from the lower values of tan δ after 20 and 60 days of curing. On the other hand, the ThD modified binders, after 60 days of curing, have proved to improve bitumen properties in the low temperature region, as can be deduced from Figure 4, which 4057

dx.doi.org/10.1021/ef200801h |Energy Fuels 2011, 25, 4055–4062

Energy & Fuels

ARTICLE

Table 3. DMTA Glass Transition Temperatures (Tg,1Hz), MDSC Glass Transition Temperatures (Tg1 and Tg2), Second Event Temperatures (T2nd) and Fourth Event Enthalpies (ΔH4th) for 9 wt % ThD Modified Binders, Processed at 90 and 130 °C, after 60 Days of Curing a

neat blank (T = 130 °C)

Tg,1Hz

Tg1

Tg2

T2nd

ΔH4th

(°C) b

(°C) c

(°C) c

(°C) d

(J/g) e

6.0 ( 0.5 2.0 ( 0.3

30.5 30.5

2.0 3.0

9.0 9.3

2.14 2.97

modified (T = 90 °C)

9.0 ( 0.7

32.5

0.5

12.6

2.38

modified (T = 130 °C)

10.0 ( 0.6

33.0

0.0

11.3

3.57

Neat bitumen and blank samples, processed at 130 °C, included as reference. b Obtained from the peak of E00 observed in DMTA curves (Figure 4). c Obtained from MDSC measurements (peaks of dCp/dT curves, Figure 5). d Obtained from the nonreversing heat flow curve peak marked in Figure 9 at low temperature. e Obtained from endothermic event located between 40 and 70 °C in Figure 9. a

shows the temperature dependence between 35 and 35 °C of the flexural moduli (E0 and E00 ), within the LVE strain limits, at 1 Hz. In that sense, below the crossover point between E0 and E00 , the evolution from the transition to the glassy region (this latter is characterized by high material brittleness and, therefore, high probability of thermal cracking under loading) with increasing temperature is apparent for every sample tested.17 A “mechanical” (and so, frequency-dependent) glass transition temperature, taken as the value at the maximum peak in the E00 curve, may result in a suitable parameter to establish a comparative analysis between the low temperature performance of the modified bituminous binders (after 60 days of curing) and their corresponding blank samples. Thus, values of the glass transition temperature measured from isochronal E00 curves, at 1 Hz, are shown in Table 3. A significant decrease of 8 °C in Tg values, from 2 °C for the blank sample at 130 °C to 10 °C for its corresponding modified sample, demonstrates that ThD addition could contribute to enhance binder performance at low inservice temperatures, and consequently, it is expected to have similar benefits on the asphalt mixtures thereof obtained. For samples processed at 90 °C, at which “primary ageing” during mixing is negligible, Tg corresponding to the neat sample can be taken as the reference value for comparison. In that sense, just a decrease in Tg of 3 °C is noticed. Looking further into the binder performance in the low temperatures range, MDSC corroborates the observations raised through dynamic flexural tests. Figure 5 displays curves of the heat capacity derivative (dCp/dT) for neat bitumen, its corresponding blank sample, and 9 wt % modified binders, processed at 90 and 130 °C, after 60 days of curing. As may be seen, two well-defined peaks are found between 40 and 20 °C, which result from the overlapping of at least two glass transition processes corresponding to saturates and aromatics in the maltenic fraction.18,19 Thus, Figure 5 reveals that the ThD addition brings about a decrease in the temperatures at which Tg1 and Tg2 peaks are located (see their corresponding values in Table 3). In that way, chemical modification by ThD seems to improve the low in-service temperatures of bituminous binders, as it is able to significantly counteract the increase in Tg caused by aging. Thus, interestingly, it induces effects similar to those of physical rejuvenating agents and, further, may be useful to prevent primary aging.20

Figure 5. Evolution with temperature of Cp derivative, obtained by MDSC, for 9 wt % ThD modified binders cured for 60 days and processed at 90 and 130 °C. Blank samples included as reference.

3.3. Modification Mechanisms. The above-mentioned enhancements in viscous and viscoelastic behaviors, at both high and low in-service temperatures, hint at significant changes in binder microstructure as a consequence of chemical reactions involving ThD, which seem to follow different pathways depending on the processing temperature. Hence, thermal analysis (DTG/DTA and MDSC tests) and thin layer chromatography (TLC-FID) may shed some light on this issue. DTG and DTA tests, which allow enthalpy and sample mass loss to be determined simultaneously in an open system,21 may provide valuable information about possible modification pathways and their resulting reaction products for ThD modified bitumens. Figure 6 reveals that an increase in the processing temperature leads to significant differences in the DTG/DTA curves obtained. Thus, neat bitumen presents a large mass loss process, extending over a wide temperature interval (from 300 to 500 °C, with a DTG peak at 455 °C), which involves the decomposition/ volatilization of chemical compounds with very different molecular weights. However, the noncured ThD modified bitumens shown in Figure 6 present some other peaks. First, it can be observed that the height of the peak at 455 °C decreases as processing temperature decreases, as a consequence of the existence of two new significant peaks: the first one at 220 and the second one at 300 or 310 °C for processing temperatures of 90 or 130 °C, respectively. Interestingly, those peaks almost vanish at the highest processing temperature (180 °C). Aiming to understand the origin of the news peaks, the thermal decomposition of thiourea dioxide was studied, and the result is included in Figure 6. According to Wang et al.,22 thiourea dioxide thermal decomposition occurs in two stages. The first one, located between 121 and 144 °C (with a marked DTG peak at 135 °C), is related to the release of SO2 and involves a significant mass loss of about 17 wt %: NH2 ðNHÞCSO2 HðsÞ f H2 NCONH2 ðsÞ

þ 0:5SðsÞ þ 0:5SO2 ðgÞ

ðR1Þ

The second stage in thiourea dioxide decomposition (shown by a shorter but broader DTG peak at around 220 °C) is governed by continuous melting, vaporization, and decomposition of urea produced by the first reaction. Although it is a 4058

dx.doi.org/10.1021/ef200801h |Energy Fuels 2011, 25, 4055–4062

Energy & Fuels

ARTICLE

Figure 6. TG, DTG, and DTA curves between 40 and 600 °C and under N2 atmosphere for thiourea dioxide, neat bitumen, and noncured 9 wt % ThD modified binders processed at the three selected processing temperatures.

Table 4. Evolution with Curing Time of the Colloidal Index (CI) for 9 wt % ThD Modified Binders Processed at 90 and 130 °C noncured

20 days

60 days

modified bitumen (Tproc = 130 °C)

0.21

0.16

0.16

modified bitumen (Tproc = 90 °C)

0.29

0.28

0.24

important is the decrease noticed for the binder processed at 90 °C. In any case, the contribution of each single SARAs fraction to the overall colloidal structure of the resulting modified bitumen can be quantified by means of a colloidal index (CI),20 which can be written as follows: Figure 7. Evolution of SARAs fractions with curing time for 9 wt % ThD modified binders processed at 90 °C (A) and 130 °C (B).

complex process, the overall chemical reaction can be written as H2 NCONH2 ðsÞ f NH3 ðgÞ þ HNCOðgÞ

ðR2Þ

Accordingly, the peak appearing at 220 °C in the modified binders should be attributed to urea decomposition. However, the second peak, at 300 or 310 °C, which cannot be related to any identified process, might correspond to new products formed through bitumen modification by ThD (or even upon heating over 120 °C in DTG/DTA tests). Figure 6 also reveals that they do not appear when bitumen modification has been carried out at 180 °C. Moreover, the evolution with curing time of the SARAs fractions for the ThD modified bitumens processed at 90 and 130 °C (see Figure 7) may be of some help to understand these results. Changes in the different bitumen fractions evidence chemical interactions. Also, a very significant decrease in the asphaltenic fraction concentration with increasing curing time is observed for the binder processed at 130 °C. Much less

CI ¼

saturates þ asphaltenes aromatics þ resins

ð2Þ

Table 4 presents the evolution of this parameter for ThD bitumen binders, processed at 90 and 130 °C. A decrease in CI with increasing curing time is observed that does not support the increase in viscosity and elastic functions at high in-service temperatures previously found (see Figures 13). However, eq 2 defines CI in terms of the SARAs fractions, determined through a chromatographic method that has a starting point consisting of dissolving a 5 mg sample in 10 mL of toluene. Hence, the CI only considers those compounds that dissolve in toluene. So, we may assume that an important part of those new products responsible for the mechanical property enhancements is basically insoluble in toluene. The filtration tests and thermogravimetric analysis in Figure 8 confirm this assumption. In this sense, photographs of residues collected after the filtration of modified bitumens (after 60 days of curing) dissolved in toluene are shown as an inset. Interestingly, the dark-colored residue of the sample prepared at 130 °C may indicate the presence of asphaltene-derived compounds, confirmed in Figure 7 by a significant reduction in the asphaltenic 4059

dx.doi.org/10.1021/ef200801h |Energy Fuels 2011, 25, 4055–4062

Energy & Fuels

ARTICLE

Table 5. DTA Peak Areas at 300310 °C for 9 wt % ThD Modified Binders Processed at 90, 130, and 180 °C (Noncured and after 60 Days of Curing)

Figure 8. DTG curve between 40 and 600 °C and under N2 atmosphere for thiourea dioxide, neat bitumen, and filtration residues of 9 wt % ThD modified binders cured for 60 days and processed at 90 and 130 °C. Inset: photographs of residues described in the text.

fraction of the sample processed at this temperature. On the contrary, a light-colored residue is observed when bitumen modification is carried out at 90 °C, and therefore, it demonstrates the existence of a different modification mechanism. Moreover, those residues were analyzed by means of DTG/ DTA tests. For the sample processed at 130 °C, the DTG curve points out that the dark-colored residue is mainly composed of urea (peak at 220 °C) and new reaction compounds (peak at 310 °C). In this case, only a small quantity of residual nonreacted ThD is noticed. Thus, the evolution with curing time of viscosity and elasticity, at high in-service temperatures, may be mainly attributed to the formation of products between polar bitumen compounds (above all, asphaltenes) and urea formed through reaction R1. Just to ensure that the sulfur formed after ThD decomposition is not contributing to the overall rheological response, it was added to neat bitumen in equivalent quantity to that produced in reaction R1 (about 2 wt % over the total binder weight). The decrease in viscosity shown in Figure 1B demonstrates that, far from being a viscosity enhancing agent, such a small percentage of sulfur can only produce a plasticizing effect.23 By contrast, for the sample processed at 90 °C, a much higher quantity of residual nonreacted ThD appears (peak at 135 °C), as reaction R1 does not happen during modification at that temperature. Also, a peak, at 200 °C, appears as a consequence of the urea formed upon heating over 120 °C in the DTG/DTA test, although no major peak due to formation of new products is observed at 300 °C, as the residue does not contain a significant quantity of asphaltenes. Thus, at 90 °C, neat ThD and urea are thought to be responsible for the modification observed. The presence of urea in the sample processed at 90 °C (below the ThD melting/decomposition temperature) would arise from a slow decomposition of ThD, typically found in acidic to weakly alkaline aqueous media, toward urea and sulfoxylic acid. In that sense, a slow diffusion into bitumen of water from the surrounding environment (i.e., as a result of air moisture) could explain the long term curing process observed for at least 12 months.24,25 Small molecules such as urea and thiourea derivatives have demonstrated strong hydrogen bonding activity and have emerged as potential organocatalysts. Similarly, polyphospheric acid (PPA) has been proposed as both a reaction catalyst and a reactant able to change the hydrogen bond network of bitumen.26

processing temp (°C)

noncured (μV 3 s/mg)

60 days (μV 3 s/mg)

90

22

28

130

35

66

180

0

0

In addition, DTA measurements may provide additional information about the new reaction products (peaks at 300 or 310 °C). Thus, decomposition enthalpies of new reaction products were calculated by measuring the area corresponding to the endothermic peaks in the DTA curves, and the results are presented in Table 5. In accordance with previous results, a significant increase in ΔH310°C with curing time is noticed for the samples prepared at 130 °C, which matches the decrease in asphaltenes content observed in Figure 7. Much less significant is the reduction found in samples prepared at 90 °C. Once again, no peak was detected at 180 °C, which suggests quite a different modification route. For the sample manufactured at 180 °C, no residue after filtration was observed, while none of the characteristic peaks appear in the DTG curve, other than that corresponding to the bitumen decomposition, which can be seen in Figure 6. In fact, the DTG results in this figure reveal that 150 °C is the onset for urea decomposition; so, both reactions R1 and R2 may be occurring during processing at 180 °C. Hence, urea finally decomposes into two gaseous components. However, although NH3 and HNCO represent the final products of reactions, urea thermal decomposition is known to be a complex process involving further decomposition of primary products through secondary series of reactions.27 Thus, at 152 °C, decomposition begins, accompanied by vigorous gas evolution from the melt. In this case, the overall reaction R2 is split into the two following partial reactions: H2 NCONH2 ðmÞ f NH4þ NCO ðmÞ

ðR3Þ

NH4þ NCO ðmÞ f NH3 ðgÞ þ HNCOðgÞ

ðR4Þ

On the basis of these facts, a modification mechanism via reaction between cyanic acid and polar bitumen compounds with the formation of urethane linkages (reaction R5) is herein proposed. In addition, a product of urea decomposition, HNCO, may also react with nonreacted urea to produce biuret-derived compounds (reaction R6), at approximately 160 °C: HNCO þ ROH f RNHCOOH

ðR5Þ

H2 NCONH2 ðmÞ þ HNCOðgÞ f H2 NCONHCONH2 ðmÞ ðR6Þ Finally, the above-commented chemically induced changes in the bitumen composition yield an important modification of the bitumen microstructure, which can be studied by means of the nonreversing component of the heat flow curve, obtained by modulated DSC. According to Masson and Polomark,18 saturates are semicrystalline, aromatics are amorphous, and resins and asphaltenes are mesophasic. They order in four stages upon cooling from the melt, yielding four specific thermal events in the nonreversing heat flow curve. Figure 9 shows the nonreversing 4060

dx.doi.org/10.1021/ef200801h |Energy Fuels 2011, 25, 4055–4062

Energy & Fuels

Figure 9. Nonreversing heat flow thermograms for 9 wt % ThD modified bitumens cured for 60 days and processed at 90 and 130 °C. Neat bitumen and blank samples included as reference.

thermograms for neat bitumen, as well as its corresponding blank sample and ThD-modified binders processed at 90 and 130 °C, after 60 days of curing. If attention is first focused on neat bitumen, a broad endothermic background extending, approximately, from 40 to 80 °C (first event) can be observed. Also, two exotherms, located at about 9 and 40 °C (second and third events, respectively), are noticed. Finally, an endothermic peak appears at around 50 °C (fourth event). The second and third thermal events are caused by a time-dependent cold-crystallization of low and high molecular weight saturated segments upon cooling, respectively. Hence, above the frozen state (T > Tg), chain mobility is increased and those segments can crystallize.16,18,19 Table 3, which gathers temperatures at which the second event (T2nd) appears, shows a similar value of about 9 °C for neat bitumen and its corresponding blank sample at 130 °C. By contrast, T2nd decreases to 11.3 and 12.6 °C after the addition of 9 wt % ThD and 60 days of curing, at processing temperatures of 130 and 90 °C, respectively. These results are supported by the lower Tg values after ThD addition observed in DMTA tests, which would increase the chain mobility, which would consequently reduce the temperature at which cold crystallization begins. On the other hand, the endotherm at 50 °C (fourth event) is related to the diffusion of relatively large and high molecular weight structures, such as those found in resins and asphaltenes, to form independent domains. In that sense, the area taken up by this event is clearly seen to increase with the ThD addition at 130 °C, suggesting the formation of new large structures. Thus, results in Table 3 indicate that ΔH4th (calculated from areas in Figure 9) increases from 2.14 to 3.57 J/g for this binder, while no significant changes are observed for the binder at 90 °C.

4. CONCLUSIONS The influence of processing temperature on the modification route and rheological properties of thiourea dioxide-modified bitumen has been evaluated. With this aim, bitumen was modified at three different temperatures (90, 130, and 180 °C) by adding 9 wt % ThD. A different modification mechanism for each of the three selected temperatures has been identified, by means of chromatography, filtration tests, and thermal analysis. Thus, for the samples processed at 90 °C, neat ThD and urea are

ARTICLE

thought to be responsible for the modification observed, as no thermal decomposition occurs at that processing temperature. On the other hand, a significant reduction in the asphaltenic fraction with curing time, for the samples processed at 130 °C, suggests that modification may be the result of new reaction products originated by asphaltenes and urea derived from ThD decomposition (reaction R1). Finally, for binders processed at 180 °C, no residue after filtration was observed, while none of the characteristic DTG peaks for the remaining ThD modified binders can be appreciated. On the basis of these facts, a modification mechanism via reaction between cyanic acid derived from urea decomposition (reaction R2) and polar bitumen compounds is herein proposed. In any case, viscosity was always seen to increase after ThD addition. Also, modification enhanced elastic properties and thermal susceptibility at high in-service temperatures and might yield improved binder resistance to thermal cracking by a decrease in its glass transition temperature. Consequently, these outcomes suggest thiourea dioxide may become a promising alternative to the use of other chemical modifiers for the paving industry.

’ AUTHOR INFORMATION Corresponding Author

*Phone: +34 959 21 99 89. Fax: +34 959 21 93 85. E-mail: [email protected].

’ ACKNOWLEDGMENT This work is part of a research project sponsored by an MECFEDER programme (Research Project MAT2007-61460) and by a Junta de Andalucía Programme (TEP6689). The authors gratefully acknowledge their financial support. A.A.C. also acknowledges the concession of an MEC FPU research fellowship (AP2008-01419). ’ REFERENCES (1) Claudy, P.; Letoffe, J. F.; King, G. N.; Brule, B.; Planche, J. P. Fuel Sci. Techonol. Int. 1991, 9, 71–92. (2) Lesueur, D.; Gerard, J. F.; Claudy, P.; Letoffe, J. F.; Planche, J. P.; Martin, D. J. Rheol. 1996, 40, 813–836. (3) Whiteaoak, D. The Shell Bitumen Handbook; Shell Bitumen U.K.: Surrey, U.K., 1990. (4) Adedeji, A.; Gr€unfelder, T.; Bates, F. S.; Mascoko, C. W.; StroupGardiner, M.; Newcomb, D. E. Polym. Eng. Sci. 1996, 36, 1707–1723. (5) Kandhal, P. S.; Cooley, L. A. Natl. Coop. Highw. Res. Program Rep. 2003, 508. (6) Lu, X.; Isacsson, J.; Ekblad, J. Mater. Struct. 2003, 36, 652–656. (7) Fawcett, A. H.; McNally, T.; McNally, G. M.; Andrews, F.; Clarke, J. Polymer 1999, 40, 6337–6349. (8) Masson, F. J. Energy Fuels 2008, 22, 2637–2640. (9) Giavarini, C.; Mastrofini, D.; Scarsella, M.; Barre, L.; Espinat, D. Energy Fuels 2000, 14, 495–502. (10) Martinez, A.; Paez, A.; Martin, N. Fuel 2008, 87, 1148–1154. (11) Cuadri, A. A.; Partal, P.; Navarro, F. J.; García-Morales, M.; Gallegos, C. Fuel 2011, 90, 2294–2300. (12) Standard test method for penetration of bituminous materials. ASTM D5-06e1; ASTM International: West Conshohocken, PA, 2011; DOI: 10.1520/D0005-06E01. (13) Standard test method for softening point of bitumen (ring and ball apparatus). ASTM D36/D36M-09; ASTM International: West Conshohocken, PA, 2009; DOI: 10.1520/D0036_D0036M-09. 4061

dx.doi.org/10.1021/ef200801h |Energy Fuels 2011, 25, 4055–4062

Energy & Fuels

ARTICLE

(14) Eckert, A. The application of Iatroscan-technique for analysis of bitumen. Pet. Coal 2001, 43, 51. (15) Martín-Alfonso, M. J.; Partal, P.; Navarro, F. J.; García-Morales, M.; Gallegos, C. Ind. Eng. Chem. Res. 2008, 47, 6933–6940. (16) Navarro, F. J.; Partal, P.; García-Morales, M.; Martín-Alfonso, M. J.; Martinez-Boza, F.; Gallegos, C.; Bordado, J. C. M.; Diogo, A. C. J. Ind. Eng. Chem. 2009, 15, 458–464. (17) Partal, P.; Martinez-Boza, F.; Conde, B.; Gallegos, C. Fuel 1999, 78, 1–10. (18) Masson, F. J.; Polomark, G. M. Thermochim. Acta 2001, 374, 105–114. (19) Masson, J. F.; Polomark, G. M.; Collins, P. Energy Fuels 2002, 16, 470–476. (20) Lesueur, D. Adv. Colloid Interface Sci. 2009, 145, 42–82. (21) Yang, J.; Roy, C. Thermochim. Acta 1999, 333, 133–140. (22) Wang, S.; Gao, Q.; Wang, J. J. Phys. Chem. B 2005, 109, 17281–17289. (23) Gawel, I.Sulphur-Modified Asphalts. In Asphaltenes and Asphalts; Yen, T. F., Chilingarian, G. V., Eds.; Developments in Petroleum Science 40B; Elsevier Science B. V.: Amsterdam, The Netherlands, 2000; Vol. 2, Chapter 19. (24) Martín-Alfonso, M. J.; Partal, P.; Navarro, F. J.; García-Morales, M.; Gallegos, C. Eur. Polym. J. 2008, 44, 1451–1461. (25) Fujiyasu, K.; Shigeki, Y.; Ito, H. Method of Activating Thiourea Dioxide. U.S. Patent 4,610,802, 1986. (26) Masson, F. J.; Gagne, M. Energy Fuels 2008, 22, 3402–3406. (27) Schaber, P. M.; Colson, J.; Higgins, S.; Thielen, D.; Anspach, B.; Brauer, J. Thermochim. Acta 2004, 424, 131–142.

4062

dx.doi.org/10.1021/ef200801h |Energy Fuels 2011, 25, 4055–4062