use of crumb rubber in road paving applications: …...the asphalt rubber supplier provided both hot...

Sardinia 2013, Fourteenth International Waste Management and Landfill Symposium

USE OF CRUMB RUBBER IN ROAD

PAVING APPLICATIONS: WORKERS’

HEALTH RISK ASSESSMENT

M.C. ZANETTI, S. FIORE, B. RUFFINO, E. SANTAGATA

Department of Environmental, Land and Infrastructure Engineering,

Politecnico di Torino – Corso Duca degli Abruzzi, 24, 10129 Torino, Italy

SUMMARY: Crumb rubber (CR) derived from end-of life tyres (ELTs) is widely employed for

the production of high-performance bituminous mixtures for paving applications, but no

information is available on the effects that CR may have on health-related risks which workers

may be exposed to on site. In such a context, this paper shows the results obtained in an

investigation that considered gaseous emissions sampled during the laying of bituminous

mixtures containing CR in two pavement test sections. By referring to the contents of Volatile

Organic Compounds (VOCs) and Polycyclic Aromatic Hydrocarbons (PAHs), it was found that

the overall diffusion scenario of gaseous emissions was coherent with CR composition but was

also strongly dependent upon several other material- and site-specific factors. Results were

compared with data from other pavement construction sites and were used within a sanitary-

environmental risk analysis procedure. The influence of CR was highlighted in differential

terms, comparing calculated risks associated to the laying of mixtures with and without CR.

1. INTRODUCTION

European Law 1999/31/CE establishes that wastes characterized by a lower heating value (LHV)

higher than 13,000 kJ/kg cannot be landfilled. Moreover, European Law 2000/53/CE requires

that starting from 2015 at least 95% b.w. of each end-of-life vehicle (ELV) must be recycled or

recovered. As a result of these legal constraints, currently end-of-life tyres (ELTs) are mainly

employed for energy production or matter recovery and massive research efforts are devoted to

the identification of the most efficient destinations.

Based on several years of experience worldwide, it has been proven that crumb rubber (CR)

coming from the grinding of ELTs can be employed in the production of high-performance

bituminous mixtures for road paving applications (Caltrans, 2005). Available technologies are

known as the “wet” and “dry” processes. In the “wet” process, CR is preliminarily mixed with

bitumen, thus obtaining a very ductile and elastic modified binder, also known as “asphalt

rubber” (ASTM D6114, 2009), that is then combined with aggregates in the hot mix plant. In the

“dry” method, CR is introduced in the production flow of bituminous mixtures as a

supplementary component, substituting part of the aggregates and providing enhanced elastic

response under loading (Santagata and Zanetti, 2012a; Santagata et al., 2013b).

Research has focused on a number of performance-related issues of bituminous mixtures

containing CR, including the assessment of resistance to permanent deformation, fatigue

cracking, oxidative ageing and water damage (Hicks, 2002). Moreover, several aspects to be

taken into account during mix design and quality control operations have been addressed and


have led to the definition of reliable technical specifications (Way et al., 2012; Caltrans, 2003).

Nevertheless, the effects of the use of CR on gaseous emissions produced during laying

operations have not been subjected to extensive investigations, with a limited number of

experimental studies which have not yet yielded quantitative information on the potential health

risks which workers are exposed to on site (Watts et al., 1988; Burr et al., 2001; Stout and

Carlson, 2003).

The abovementioned issue was considered of crucial importance while planning two research

projects which were recently launched in Italy on the use of CR in bituminous mixtures (“wet”

and “dry” technologies): POLIPNEUS (2012-16), funded by Ecopneus, a non-profit company

which has the mission of managing the entire flow of ELTs in Italy, and TYREC4LIFE

(2011-14), supported by the European Commission as part of the LIFE+ funding program

(Santagata et al., 2012b; Province of Turin, 2011). The preliminary work presented in this paper

is based on monitoring activities which were carried out within these projects on two

construction sites (“A” and “B”) in which bituminous mixtures containing CR were laid for the

formation of surface wearing courses. The main goal of the investigation was to assess the

potential health impact on workers of gaseous emissions produced during paving operations. For

such a purpose, fumes were sampled at the paver and were then subjected to laboratory analyses

for the determination of the concentration of Volatile Organic Compounds (VOCs) and

Polycyclic Aromatic Hydrocarbons (PAHs). Results were compared with data from other

pavement construction sites and used within a procedure for the assessment of health risks which

workers are exposed to. In such a context, the influence of CR was highlighted in differential

terms, comparing calculated risks associated to the laying of mixtures with and without CR.

2. MATERIALS AND METHODS

The experimental investigation was carried out by considering bituminous mixtures which were

laid for the formation of surface wearing courses of two urban roads. In the case of site “A”,

located in Zambra (Pisa), the surface course was 4 cm thick and the bituminous mixture was

prepared in a nearby plant by employing freshly-produced asphalt rubber binder. Site “B” was

located in Caserta, with a surface course 4 cm thick composed by a mixture which was obtained

by means of a hybrid production process. In fact, the same asphalt rubber employed at site “A”

was diluted with neat bitumen and the resulting binder was thereafter used for the production of

the bituminous mixture. Such a process is similar to what occurs during the so-called “field

blending” operations, which lead to bituminous binders in which CR dosage is generally limited

to below 10% b.w. of total binder.

Mixtures were laid in test sections of similar widths (of the order of 10 m) and different

lengths (equal to 500 m and 250 m, respectively for site A and B). Component materials and

mixtures were sampled during production and laying and were then subjected to

characterization. Moreover, gaseous emissions were sampled at the paver and were then

analyzed in the laboratory.

2.1 Crumb rubber (CR)

CR employed for the production of asphalt rubber was supplied by a ELT treatment plant located

in Southern Italy which operates according to the ambient size-reduction process. Samples of the

material were subjected to laboratory tests for the determination of particle size distribution,

evaluation of contents of heavy metals (Al, As, Ba, Cd, Co, total Cr, Cu, Fe, Mn, Ni, Pb, Sb, Ti,

Zn), VOCs (Volatile Organic Compounds), PAHs (Polycyclic Aromatic Hydrocarbons) and

elemental analysis (C, H, N, S) (Zanetti et al., 2011).


Particle size distribution analysis was performed by making use of sieves of the Tyler series

in dry conditions. Metals were determined by using a Perkin-Elmer Optima 2000 ICP-OES after

digestion of a 100 mg sample in a Milestone 1200 Mega microwave oven with nitric acid (65%,

Sigma Aldrich, 3 ml) and perchloric acid (70%, Merck, 1 ml). VOC and PAH analyses were

carried out by means of solvent extraction and subsequent gas-chromatographic analysis. A 2 g

sample was extracted with 20 ml of CH2Cl2 for 20 minutes in a Milestone 1200 Mega microwave

oven set at 600 W. Gas-chromatographic analyses were performed by using an Agilent

7890/5975 GC-MS equipped with a HP5-MS capillary column (30m0.25mm0.25m)

(Santagata and Zanetti, 2012a,b). Carbon, hydrogen, nitrogen and sulphur contents were

determined by employing a Flash 2000 ThermoFisher Scientific CHNS analyzer.

2.2 Bituminous mixtures and binders

Bituminous mixtures laid at the two construction sites were designed by referring to different

technical specifications. The mixture employed at site A was conceived as a classical “gap-

graded” material (GG, with a non-continuous particle size distribution and high binder content),

the specifications of which were made available by the asphalt rubber supplier (ARI, 2012). In

the case of site B the mixture was proportioned in order to obtain an “open-graded friction

course” (OGFC), characterized by a moderately discontinuous aggregate size distribution, high

void content after compaction (above 10%) and enhanced surface friction in service (CIRS,

2001). Aggregates available in the two hot mix plants were quite different both in terms of size

distribution and mineralogical composition. The asphalt rubber supplier provided both hot mix

plants with the same product (made by the same components combined with a target CR dosage

of 20% b.w. of total binder). However, in the case of site B the binder was further diluted with a

neat bitumen (50/70 penetration grade), different from that employed for asphalt rubber

production. According to the information retrieved from plant operators, field blending led to a

final CR percentage of the order of 5% b.w. of total binder.

At the production plants, samples of the bituminous binders were taken from storage tanks.

Materials taken into consideration included asphalt rubber binder (which was common to the two

sites), its base bitumen and the bitumen employed for dilution in the case of site B. While asphalt

rubber was characterized as part of other activities included in the project, neat bitumens were

subjected in the laboratory to fractionation analysis for the determination, according to the

simplified schematization of colloidal system (Nellensteyn, 1928), of relative amounts of

saturates, aromatics, resins and asphaltenes (SARA analysis). This was done by employing a

technique which combines Thin-Layer Chromatography (TLC) with Flame Ionization Detection

(FID) in a procedure which has often been employed for the characterization of paving binders

(Leroy, 1989; Ecker, 2001).

During construction of test sections, temperature of the mixtures was continuously monitored

behind the paver’s screed by means of hand-held immersion thermometers. Furthermore, fume

samples were taken at the driver’s seat of the paver and at the screed for the subsequent

evaluation of VOCs and PAHs. This was done by employing a pump (0.5 l/min flow rate, 5

minutes total sampling time) by means of which fumes were adsorbed on active granular carbon

cartridges which were then stored at freezing temperature until analysis. These matrixes were

subjected to solvent extraction (with methylene chloride, Fluka, HPLC grade) in an ultrasound

bath for 60 minutes (EN 13649, 2002; Lindberg et al., 2008). Subsequent analysis was carried

out in an Agilent 7890/5975, equipped with a HP5-MS capillary column

(30m0.25mm0.25m) (Santagata and Zanetti, 2012a,b).

Site-specific conditions were also recorded by measuring air temperature, pressure and wind

speed by means of a multi-purpose hand-held probe which was employed throughout

construction operations.


Samples of the two bituminous mixtures taken at the construction sites were subjected to tests

for the determination of binder content, aggregate size distribution and volumetric properties of

laboratory-compacted specimens.

Binder content was determined by means of ignition tests carried out according to EN

12697-39 (2012). Size distribution of the aggregates recovered from ignition tests was evaluated

in wet conditions by employing the standard set of sieves indicated in technical specifications

(CIRS, 2001). Laboratory-compacted specimens were prepared by employing both the classical

Marshall equipment (EN 12697-30, 2004) and the more simulative gyratory shear compactor

(GSC) (EN 12697-31, 2007) at a reference temperature of 170°C. Marshall specimens were

prepared by applying 75 standard blows on each face. Gyratory specimens were fabricated in 150

mm moulds by imposing 100 gyrations (600 kPa vertical pressure, 1.25° gyration angle). All

specimens were then subjected to the determination of bulk density by means of the vacuum

sealing procedure (EN 12697-6, 2012). Volumetric properties (void content, voids in the mineral

aggregate, voids filled with bitumen, EN 12697-8, 2003) were thereafter calculated by making

use of the theoretical maximum density values of the two mixtures derived from measurements

performed with the pycnometer method (EN 12697-5, 2010).

3. RESULTS AND DISCUSSION

3.1 Tests on crumb rubber

Average results of the analyses carried out for the physical and chemical characterization of CR

samples are synthesized in Figure 1 (particle size distribution) and Tables 1 and 2 (chemical

analyses). In both cases, obtained results are compared with the database which the Authors have

referred to in the past for the certification of similar ELT-derived granular products (Santagata et

al., 2013c).

Figure 1. Particle size distribution of CR.

0

25

50

75

100

0.01 0.1 1 10

Per

cen

t p

ass

ing

(%

)

Diameter (mm)

CR

Reference database


Table 1. Metals and C, H, N, S content of CR.

CR Database CR Database

Na (mg/kg) 178 198 - 252 C (%) 80.1 76.2 - 81.9

K (mg/kg) 422 300 - 586 H (%) 7.36 7.00 - 7.38

Ca (%) 0,191 0.130 - 1.120 N (%) 0.513 0.390 - 0.530

Mg (mg/kg) 441 246 - 1240 S (%) 2.01 1.69 - 2.42

Fe (%) 0,172 0.042 - 0.915

Mn (mg/kg) 17,7 5.1 - 59.6

Ba (mg/kg) 9,58 6.3 - 211.0

Al (mg/kg) 649 372 - 800

Cd (mg/kg) 3,13 2.17 - 6.30

Cr (mg/kg) 6,09 2.29 - 12.3

Ni (mg/kg) 5,35 3.84 - 13.2

Pb (mg/kg) 22,8 26.6 - 194

Cu (mg/kg) 450 64 - 1218

Zn (%) 1,89 1.16 - 2.26

Co (mg/kg) 305 151 - 418

Ti (mg/kg) 30,4 33.6 - 67.4

Sb (mg/kg) 385 151 - 608

Table 2. VOC and PAH contents (mg/kg) of CR.

CR Database CR Database

benzene 9.62 8.94 - 33.68 naphtalene 0.50 0.21 - 0.53

toluene 0.46 0.01 - 4.04 acenaphtylene 0.42 0.20 - 0.74

ethylbenzene 1.39 0.59 - 3.70 phenantrene 1.19 1.45 - 3.36

styrene 0.12 0.05 - 0.52 anthracene 1.94 0.32 - 1.80

1,2,4-trimethylbenzene 0.94 1.60 - 4.77 fluorantene 0.16 2.36 - 4.08

1,3,5-trimethylbenzene 0.12 0.61 - 1.85 pyrene 7.10 10.02 - 17.20

p-xylene 0.04 0.92 - 1.88 benzo[a]anthracene 0.20 0.80 - 3.02

1,3,5-trichlorobenzene 0.94 1.19 - 5.84 chrysene 0.54 0.12 - 0.88

1,2,4-trichlorobenzene 0.35 0.31 - 0.62 benzo[a]pyrene 0.48 0.45 - 1.85

Total VOCs 13.98 20.8 - 45.6 benzo[b]fluorantene 0.66 0.31 - 1.56

benzo[g,h,i]perylene 0.32 0.84 - 4.56

Total PAHs 13.51 20.40 - 30.74

From the data plotted in Figure 1 it can be noticed that the CR sampled at the asphalt rubber

production plant is entirely passing at the 0.8 mm sieve and has a continuous size distribution

which is approximately positioned at the center of the database range.

Table 1 shows that contents of heavy metals are particularly low when compared to database

limits, especially in the case of iron, probably as a result of an efficient separation of the rubber

matrix from other tyre components. The high degree of cleanliness is confirmed by the contents

of lead and titanium which are below the lower limits of database ranges. As expected, in the

case of elemental analysis (C, H, N, S), results were coherent with those obtained for all other

reference materials.

Table 2 shows that VOC and PAH values are very low when compared to those of all other

CRs previously subjected to analysis. This observation applies both to total values and to the

values of almost each compound, with the only exception of anthracene (above range). Such a

result may be related either to the kind of tyres which are fed to the CR production plant (since

truck tyres have a high percentage of natural rubber with a low content of organic aromatic

substances) or to the specific characteristics of the grinding equipment which may have caused

by itself the early release of organic substances in the environment during processing.


3.2 Tests on bituminous mixtures and binders

Average SARA composition of the neat bitumens employed for the production of the plant-

produced asphalt rubber binder (site A) or for the formation of the “field blended” binder (site B)

are given in Table 3. The blend was obtained by combining the base and dilution bitumen in the

proportions which simulate what was obtained during the previously described hybrid process. It

can be observed that all binders are quite similar to each other, with variations of fraction

percentages which may be considered within the inherent uncertainty of test results.

Table 3. Composition of base bitumens (fractionation analysis).

Base Dilution Blend

Saturates (%) 2.6 4.3 2.3

Aromatics (%) 67.3 66.5 67.0

Resins (%) 16.3 14.5 14.1

Asphaltenes (%) 13.8 14.7 16.6

Results obtained from ignition tests and subsequent sieve analyses revealed that composition of

the two mixtures met the requirements of technical specifications (ARI, 2012; CIRS, 2001). This

is clearly shown in Figure 2, where the two aggregate size distributions are plotted and compared

to acceptance ranges. In the case of binder content, values equal to 8.5% and 5.6% b.w. of dry

aggregates were respectively recorded for mixtures laid in sites A and B, in both cases within

acceptance intervals indicated in technical specifications (7.5÷8.5% for the GG mixture,

4.8÷5.8% for the OGFC mixture).

Figure 2. Aggregate size distribution of the bituminous mixtures containing CR.

0

25

50

75

100

0.01 1 100

Per

cen

t p

ass

ing

(%

)

Diameter (mm)

ARI limits

GG - Site A

0

25

50

75

100

0.01 1 100

Per

cen

t p

ass

ing

(%

)

Diameter (mm)

CIRS limits

OGFC - Site B


Volumetric characteristics of GSC and Marshall specimens (percent voids %v, voids in the

mineral aggregate VMA and voids filled with bitumen VFB) are listed in Table 4. It was

observed that for both mixtures, volumetric conditions reached after compaction were totally

coherent with design expectations. Moreover, by analyzing GSC densification curves it was

found that the volumetrics of Marshall specimens was equivalent to that of GSC specimens

compacted with different numbers of gyrations depending upon mixture type (50 and 75

gyrations for the GG and OGFC mixture respectively).

Table 4. Volumetric properties of laboratory-compacted specimens.

GG mixture (A) OGFC mixture (B)

%v

(%)

VMA

(%)

VFB

(%)

%v

(%)

VMA

(%)

VFB

(%)

GSC specimens, 100 gyrations 4.4 22.5 80.4 9.4 20.9 55.0

Marshall specimens, 75 blows 6.9 24.5 71.9 10.9 22.2 51.0

Results of the analyses performed on gaseous emissions sampled at the test sections are given in

Tables 5 and 6. Listed compounds are those which are considered toxic or carcinogenic among

all substances potentially detectable by means of gas-chromatographic techniques. These are also

the compounds which were considered within the sanitary-environmental risk analysis procedure

described in paragraph 4.

Experimental results were compared to the data obtained in other pavement construction sites

monitored in the context of previous or ongoing research projects (Santagata and Zanetti,

2012a,b; Santagata et al, 2012; Santagata et al, 2013a). They are divided in two groups, which

respectively refer to the laying of bituminous mixtures containing asphalt rubber (only two sites,

“asphalt rubber database”) and to the laying of traditional mixtures with neat or polymer-

modified bitumen (eight sites, “traditional database”).

Information regarding the sites under consideration (A and B) and the two reference ones

belonging to the asphalt rubber database are synthesized in Table 7. Symbols hc, TL, TA, pA and

vw respectively indicate layer thickness (in cm), mixture temperature during laying operations (in

°C), air temperature (in °C), air pressure (in bar) and wind speed (in m/s).

Table 5. VOCs of gaseous emissions (g/m3) at the paver (D: driver’s seat; S: screed).

Site A Site B Asphalt rubber database Traditional database

D S D S D S D S

benzene 11.89 3.66 1.21 2.66 0.94 - 2.60 <0.10 - 2.21 0.57 - 10.80 0.59 - 8.40

toluene 24.11 21.09 23.47 31.56 5.40 - 12.27 <0.10 - 13.40 0.74 - 38.86 0.55 - 20.51

ethylbenzene 7.16 6.07 13.26 7.73 <0.10 - 5.24 <0.10 - 6.41 2.22 - 10.69 1.44 - 16.08

bromobenzene 1.91 2.86 1.36 2.49 <0.10 - 1.38 <0.10 - 1.68 0.93 - 4.36 0.69 - 6.83

styrene 0.56 0.25 0.55 0.56 <0.10 - 1.71 <0.10 0.23 - 0.93 0.22 - 0.68

1,2,4-trimethylbenzene 42.20 16.88 24.54 38.82 7.60 - 18.68 6.10 - 40.15 6.49 - 54.97 4.67 - 32.99

1,3,5-trimethylbenzene 39.98 66.08 36.48 55.01 5.90 - 45.88 17.10 - 81.27 9.16 - 118.75 31.27 - 66.19

p-xylene 15.99 1.88 15.10 5.10 11.80 - 32.66 41.08 - 157.30 0.90 - 28.01 1.47 - 12.95

1,3,5-trichlorobenzene <0.10 <0.10 <0.10 <0.10 <0.10 <0.10 <0.10 - 0.76 <0.10

1,2,4-trichlorobenzene <0.10 <0.10 <0.10 <0.10 <0.10 <0.10 <0.10 <0.10

p-isopropyltoluene 15.50 3.95 10.35 14.04 1.40 - 4.16 12.25 - 57.30 1.55 - 24.50 0.99 - 14.65

butylbenzene 8.51 1.90 2.86 17.96 <0.10 - 2.66 3.10 - 9.64 2.13 - 41.07 1.33 - 15.68

Total VOCs 167.8 124.6 129.2 175.9 34.7 - 125.6 207.7 - 254.3 96.4 - 215.3 70.5 - 163.1


Table 6. PAHs of gaseous emissions (g/m3) at the paver (D: driver’s seat; S: screed).

Site A Site B Asphalt rubber database Traditional database

D S D S D S D S

naphtalene 1.99 3.34 1.36 2.38 1.28 - 7.90 1.41 - 5.20 0.86 - 4.51 0.70 - 4.14

acenaphtylene 0.14 0.17 <0.10 0.09 1.80 - 1.84 1.20 - 2.45 0.07 - 1.17 0.08 - 0.99

1-bromonaphtalene 2.42 3.87 2.31 3.05 10.30 - 17.81 9.20 - 11.94 0.76 - 8.89 0.54 - 8.68

acenaphtene <0.10 0.08 <0.10 <0.10 2.30 - 23.73 1.40 - 3.78 0.05 - 1.50 0.07 - 1.81

fluorene 1.82 2.23 2.28 1.30 1.30 - 15.70 1.00 - 8.12 0.43 - 3.48 0.37 - 2.65

phenanthrene 0.37 0.15 <0.10 0.10 2.00 - 4.44 0.53 - 1.10 0.48 - 2.21 0.26 - 1.26

anthracene 0.17 <0.10 <0.10 <0.10 1.80 - 2.72 0.38 - 0.60 0.05 - 1.74 0.25 - 0.68

fluoranthene 1.67 0.08 <0.10 0.09 2.32 - 3.10 0.22 - 0.90 0.08 - 2.04 0.09 - 1.59

pyrene 0.65 0.38 0.23 0.32 2.38 - 3.00 0.90 - 0.96 0.06 - 2.07 0.12 - 2.08

benzo[a]anthracene <0.10 <0.10 <0.10 <0.10 <0.10 - 3.60 <0.10 - 0.80 <0.10 <0.10

benzo[a]pyrene 2.39 8.08 4.14 6.74 3.60 - 15.96 0.80 - 27.07 1.75 - 19.66 1.68 - 27.73

benzo[b]fluoranthene 3.15 1.60 2.13 1.58 4.14 – 5.30 3.09 - 3.50 1.64 - 52.76 1.16 - 32.30

dibenzo[a,h]anthracene <0.10 <0.10 <0.10 <0.10 <0.10 - 8.90 <0.10 - 2.15 <0.10 <0.10

indeno[1,2,3-cd]pyrene <0.10 <0.10 <0.10 <0.10 <0.10 <0.10 <0.10 <0.10

benzo[ghi]perylene 0.26 <0.10 <0.10 <0.10 3.60 - 68.87 5.78 - 19.88 0.45 - 3.56 0.34 - 3.60

Total PAHs 15.0 20.0 12.5 15.6 58.5 - 161.2 34.5 - 79.8 15.5 - 75.7 12.5 - 64.2

Table 7. Synthesis of information regarding construction sites and gaseous emissions.

Site Bituminous mixture hc

(cm)

TL

(°C)

TA

(°C)

vw

(m/s)

pA

(bar)

VOCtotal

(g/m3)

PAHtotal

(g/m3)

D S D S

A Gap-graded wearing course

8.5% asphalt rubber 4 145 28 0.4 1014 167.8 124.6 15.0 20.0

B Open-graded friction course

5.6% binder (field blending) 4 170 26 0 1004 129.2 175.9 12.5 15.6

BV Gap-graded wearing course

8.5% asphalt rubber 3 170 9 n.a. n.a. 34.7 254.3 58.5 34.5

SF Gap-graded wearing course

8.0% asphalt rubber 5 170 24 n.a. n.a. 128.7 207.7 161.2 79.8

Interpretation of the data listed in Tables 5 and 6 is not a trivial task since the composition of

fumes is affected by several material-specific (i.e. mixture composition, CR type and base

bitumen type) and site-specific (i.e. layer thickness, temperature, laying temperature, wind, air

pressure) factors. Moreover, significant differences may be found depending upon considered

compounds (being VOCs much more volatile than PAHs and therefore more sensitive to

temperature variations) or on sampling position (D or S).

It can be observed that the low PAH values of the employed CR (Table 2) are reflected, for

both mixtures, in PAH values (either total or referred to a single compound) in gaseous emissions

which are lower than minimum database limits in both sampling positions. The same observation

only partially applies to VOCs, which are below the reference interval at the screed (S), but are

higher than the recorded upper limit in the case of fumes sampled at the driver’s seat (D).

The above described scenario can be explained by taking into account the contribution of

bitumen to gaseous emissions. In fact, since a peak in emissions is observed for most volatile

compounds (VOCs) in the sampling position which is closer to mixture discharge (D), it can be

hypothesized that in this early phase of construction most of the VOCs may have originated from

bitumen due to its composition and to the very high delivery temperature. In the case of PAHs

this did not occur possibly as a result of their lower volatility and to temperature reduction which

takes place while the mixture is laid down and is then fully exposed behind the moving paver.


Data interpretation provided above is coherent with the observations which stem from the

comparison between the two construction sites (A and B) and mixtures (GG and OGFC). Even

though most of the material- and site-specific factors are quite different (Table 7), measured

VOC and PAH contents are quite similar (Tables 5 and 6), with total and compound-specific

values only slightly higher for the GG mixture (site A). This is due to the fact that this mixture is

characterized by a higher binder content (8.5% vs. 5.8%) the effects of which are however

partially compensated by lower temperature (145°C vs. 170°C). It should also be observed that

the unit contribution of bitumen to gaseous emissions in the two cases is expected to be very

similar, given that the composition of the three analyzed binders (base bitumen of asphalt rubber,

dilution bitumen, blend of the two according to proportions adopted in site B) were found to be

very similar between each other (Table 3).

With respect to the two sampling positions (i.e. at the driver’s seat, D, and at the screed, S), no

specific trends were observed. This clearly indicates that during construction works the diffusion

of gaseous emissions around the paver is strongly influenced by site-specific conditions, the

effects of which are quite difficult to take into account in analyzing experimental data.

4. SANITARY-ENVIRONMENTAL RISK ASSESSMENT

Assessment of toxicological and carcinogenic risks which workers are exposed to on site as a

result of the presence of gaseous emissions coming from hot bituminous mixtures containing CR

was carried out by means of a sanitary-environmental risk analysis procedure (ASTM E1739,

2010; ASTM E2081, 2010). This was developed based on previous work carried out for the

evaluation and remediation of contaminated sites (Marescalco and Zanetti, 2010; Zanetti et al.,

2013) and for the approval of CR use in artificial turf sports fields (Ruffino et al., in press).

Given that both the contaminant source (i.e. bituminous mixture) and the potential receptors

(i.e. paving workers) were clearly identified and that direct measurements were performed (i.e.

fume sampling and analysis), risk evaluation was developed by comparing experimental data

with threshold values (the so-called “level 1” of analysis) and by analytically modelling them in

each local scenario (“level 2”). Analysis at “level 3”, based on numerical simulation models

capable of fully reproducing contamination dynamics, was not included in the investigation.

With respect to “level 1” analysis, reference can be made to limits which have been expressed

in other terms, such as those of the American Conference of Governmental Hygienists (ACGIH,

2001) and of the German regulation system (German Bitumen Forum, 2006). In the first case, the

maximum exposure limit is fixed by considering the benzene-soluble fraction of inhaled aerosol,

which cannot exceed 0.5 mg/m3 (by assuming 8 hours/day, 5 days/week). German regulations

refer to a maximum concentration of total hydrocarbons in emissions equal to 10 mg/m3, without

distinguishing between compounds characterized by different degrees of toxicity.

Data collected at the test sections cannot be directly compared to the limits illustrated above

since measurements were carried out by focusing on potentially harmful VOCs and PAHs.

Nevertheless, by considering the sum of total VOC and PAH contents, and by conservatively

assuming a worst-case scenario of continuous exposure, composition of analyzed emissions

seems to be compatible with the two abovementioned thresholds.

Risk analysis developed at “level 2” required the choice and use of an adequate model for

exposure evaluation and of dose-response curves relative to toxic and carcinogenic substances.

These two key elements of risk assessment are given in Equation (1) and Figure 3, respectively.

ATBW

EDEFCREM

(1)


where EM is calculated exposure (or effective exposure flow, in m3·kg-1·day-1), CR is the so-

called contact factor (dependent on the type of exposure, in this case inhalation and therefore

assumed equal to 3.6 m3/day), EF is exposure frequency (in days/year, hypothesized equal to 250

for construction workers), ED is exposure duration (in years, assumed equal to 25 years), BW is

average body weight (in kg, fixed at 70 kg), AT is the average mediation period of exposure

(corresponding to ED in the case of non-carcinogenic substances, equal to 70 years in the case of

carcinogenic substances).

Figure 3. Dose-response curves of carcinogenic and toxic (non-carcinogenic) substances.

As shown in Figure 3, the shape of the dose-response curves is different when comparing toxic

(non-carcinogenic) and carcinogenic substances. In the first case, a dose threshold can be

identified below which it has been experimentally verified that there are no harmful effects of

that substance. However, risk calculations require the use of a “reference dose” (RfD) which is

obtained by reducing the threshold in order to take into account uncertainty in the extrapolation

of dose levels from animals to humans and to consider specific characteristics of human

response. In the case of carcinogenic substances, the concept of a threshold is no longer valid

since health of human beings is damaged at any considered dose. For the description of such

effects it can therefore be assumed that in a wide dose range the response curve is linear, with the

consequent identification of a single gradient value, also known as “slope factor” (SF).

In the investigation described in this paper, RfD and SF values of compounds listed in Tables

5 and 6 were retrieved from the database of the Istituto Superiore di Sanità (ISS, 2012), the

leading technical and scientific public body of the Italian National Health Service. The dose

value (D, in mg·kg-1·d-1) to use as an input in risk analysis evaluations was calculated, for each

compound, by means of Equation 2, where C is compound concentration (in mg/m3) in the

gaseous emissions sampled on site.

EMCD (2)

Toxicological and carcinogenic risks were quantified by respectively considering the so-called

Hazard Quotient (HQ) and the Individual Excess Life Cancer Risk (IELCR), calculated as

indicated in Equations 3 and 4.

Res

pon

se

Dose (mg·kg-1·d-1)

Carcinogenic substance

Toxic (non-carcinogenic) substance

SF

RfD Threshold


RfDDHQ (3)

SFDIELCR (4)

HQ or IELCR contributions due to each compound (or group of compounds) were summed

together, thus obtaining final HQ or IELCR values which provide a synthetic description of the

potential impact of a specific paving site on workers’ health, both from a toxicological and

carcinogenic viewpoint. According to ASTM recommendations, in order for the working

conditions to be acceptable, it is required that HQ be lower than 1 and IELCR be lower than 10-4

(ASTM E1739, 2010; ASTM E2081, 2010).

For the A and B test sections considered in this study, since the focus of research was to

highlight the effects of CR on workers’ health, risk parameters were expressed in differential

terms by considering their ratio (HQratio and IELCRratio) with respect to reference values. These

were calculated for the eight paving sites which form the previously mentioned reference

database of traditional mixtures (containing neat or polymer-modified bitumen) monitored for

the assessment of gaseous emissions. In particular, as shown in Table 8, ratios were calculated

with respect to the minimum, maximum and average value of HQ and IELCR, both at the

driver’s seat (D) and at the screed (S).

Table 8. Relative risk parameters of the two pavement test sections.

Site A Site B

D S D S

HQratio with respect to minimum HQ 1.19 1.43 1.13 1.81

with respect to maximum HQ 0.67 0.64 0.64 0.81

with respect to average HQ 0.89 0.97 0.85 1.22

IELCRratio with respect to minimum IELCR 1.40 4.21 2.23 3.52

with respect to maximum IELCR 0.11 0.27 0.18 0.23

with respect to average IELCR 0.25 0.55 0.40 0.46

At first glance, data listed in Table 8 clearly indicate that HQratio and IELCRratio change

significantly depending upon which reference parameters are adopted for their calculation.

However, it is interesting to note that they are all centered around the unit value and are in any

case comprised within a single order of magnitude (ranging from 0.11 to 4.21). This means that

the sanitary-environmental risk associated to the use of bituminous mixtures containing CR is

comparable to that of standard paving materials (produced by employing neat or polymer-

modified bitumen).

In order to carry out the relative risk evaluation in the most conservative conditions, analysis

needs to focus on HQratio and IELCRratio values obtained by comparing risk parameters of sites A

and B to the those of the paving sites in which HQ and IELCR were found to be the smallest in

all the database. In such a case, values of HQratio and IELCRratio are, as expected, all greater than

1. When comparing different sampling points (i.e. different potential positions of workers), it can

be observed that in both sites toxicological and carcinogenic risks tend to be higher at the screed

than at the driver’s seat. Moreover, in each position, HQratio and IELCRratio values of the two sites

are very similar, thus leading to the conclusion that regardless of variations in mixture

composition and of site-specific factors (Table 7), most of the sanitary-environmental risk stems

from the characteristics of employed bitumen. In fact, as mentioned previously, in the two

construction sites asphalt rubber had the same base bitumen and the dilution binder added in site

B had a very similar composition (Table 3).


If the above described conservative approach to analysis is progressively loosened,

considering HQratio and IELCRratio values calculated with respect to the average or maximum

values of the reference database, it is found that the coherency between the two sites is

maintained and that in relative terms sanitary-environmental risks are significantly reduced. This

observation simply proves that the use of CR in bituminous mixtures does not necessarily

increase toxic and carcinogenic impact on workers. In fact, as demonstrated by the results of the

laboratory analyses described in previous paragraphs, gaseous emissions are strongly dependent

upon site-specific factors and among those which are material-specific, bitumen type and

composition certainly plays a dominant role.

5. CONCLUSIONS

Results obtained in the investigation described in this paper represent a preliminary contribution

to the assessment of health risks which workers may be exposed to on site during paving

operations involving the use of bituminous mixtures containing crumb rubber (CR). Even though

more research is certainly needed on this topic in order to develop a sound and fully validated

know-how, several conclusions can already be drawn.

Chemical characterization of materials employed in the production of bituminous mixtures is

an essential step of analysis which should be mandatory in technical specifications. In the case of

bitumen, preliminary information on the potential release of harmful substances during laying

operations may stem from the results of fractionation analysis. Characterization of CR, as shown

in this paper, requires a wider array of laboratory tests. These are based on the use of specific

procedures for which it is hoped that in the near future official standards will be defined and

shared within the scientific community.

Sampling of gaseous emissions at paving sites and subsequent laboratory analyses are the key

elements of the workers’ health risk assessment. Even in this case, it is crucial that codified

procedures are followed in order to have reliable data to feed into evaluation models. Results

obtained on the two sites described in this paper, and those already available from previous and

ongoing research projects, show that composition of fumes is affected by several material-

specific (i.e. mixture composition, CR type and bitumen base type) and site-specific (i.e. layer

thickness, laying and air temperature, wind, air pressure) factors. Nevertheless, relative

contribution of bitumen type and composition seems to be the most relevant.

Modelling of the experimental data within a sanitary-environmental risk analysis procedure

requires conservative assumptions to be made, mainly with respect to the work load of paving

operators. As shown in this paper, in order to highlight health-related effects caused by the use of

CR, it is convenient to express toxic and carcinogenic risk parameters (Hazard Quotient, HQ, and

Individual Excess Life Cancer Risk, IELCR) in relative terms. This can be done by comparing

absolute values with analogous data obtained in other paving sites where no CR was employed.

Results of the sanitary-environmental assessment indicate that the toxic and carcinogenic risk

to which workers are exposed on site in the case of bituminous mixtures containing CR is

comparable to that of standard paving materials (produced by employing neat or polymer-

modified bitumen). Moreover, calculated risk seems to be extremely dependent upon site-specific

factors and, among material-specific ones, mainly upon bitumen type and composition.

In order to validate the preliminary conclusions synthesized above, the Authors are currently

addressing a number of issues related to the use of CR in bituminous mixtures and on its

consequent possible impact on workers’ health. In particular, current databases are being

expanded by characterizing additional CRs and by continuing field monitoring activities.

Furthermore, a laboratory test procedure is being developed in order to measure the maximum

emission potential of any bituminous mixture in standard controlled conditions.


ACKNOWLEDGEMENTS

The investigation described in this paper was carried out as part of the POLIPNEUS project,

funded by Ecopneus S.c.p.A. In such a context, the support of G. Corbetta and D. Fornai is

gratefully acknowledged. Part of the work was also performed within the TYREC4LIFE

research project, supported by the European Commission through the LIFE+ funding program.

REFERENCES

ACGIH (2001). TLV and BEI 2001. American Conference of Governmental Industrial

Hygienists, Cincinnati, OH, USA.

ARI (2012). Technical specifications for gap-graded asphalt concrete with asphalt rubber.

Asphalt Rubber Italia, Agliana, Italy, June 2012.

ASTM D6114-09 (2009). Standard specification for asphalt-rubber binder. ASTM International,

West Conshohocken, PA, USA.

ASTM E1739-09(2010)e1 (2010). Standard guide for risk-based corrective action applied at

petroleum release sites. ASTM International, West Conshohocken, PA, USA.

ASTM E2081-00(2010)e1 (2010). Standard guide for risk-based corrective action. ASTM

International, West Conshohocken, PA, USA.

Burr G., Tepper A., Feng A., Olsen L. and Miller A. (2001). Crumb-rubber modified asphalt

paving: occupational exposures and acute health effects. NIOSH Health Hazard Evaluation

Report 2001-0536-2864. National Institute for Occupational Safety and Health, Cincinnati,

OH, USA.

Caltrans (2003). Asphalt rubber usage guide. State of California Department of Transportation,

Sacramento, CA, USA.

Caltrans (2005). Use of scrap tire rubber. State of the technology and best practices. State of

California Department of Transportation, Sacramento, CA, USA.

CIRS (2001). Performance-based technical specifications. InterUniversity Road Research

Center, Ancona, Italy (in Italian).

Cooley L.A., Brumfield J.W., Mallick R.B., Mogawer W.S., Partl M., Poulikakos L. and Hicks

G. (2009). Construction and maintenance practices for permeable friction courses, NCHRP

Report 640. Transportation Research Board, Washington, D.C., USA.

Ecker A. (2001). The application of Iatroscan-technique for analysis of bitumen. Petroleum and

Coal, vol. 1, n. 1, 51-53.

EN 12697-5 (2010). Test methods for hot mix asphalt - Determination of the maximum density.

European Committee for Standardization (CEN), Brussels, Belgium.

EN 12697-6 (2012). Test methods for hot mix asphalt - Determination of bulk density of

bituminous specimens. European Committee for Standardization (CEN), Brussels, Belgium.

EN 12697-8 (2003). Test methods for hot mix asphalt - Determination of void characteristics of

bituminous specimens. European Committee for Standardization (CEN), Brussels, Belgium.

EN 12697-30 (2012). Test methods for hot mix asphalt - Specimen preparation by impact

compactor. European Committee for Standardization (CEN), Brussels, Belgium.

EN 12697-31 (2007). Test methods for hot mix asphalt - Specimen preparation by gyratory

compactor. European Committee for Standardization (CEN), Brussels, Belgium.


EN 12697-34 (2012). Test methods for hot mix asphalt - Marshall test. European Committee for

Standardization (CEN), Brussels, Belgium.

EN 12697-39 (2012). Test methods for hot mix asphalt - Binder content by ignition. European

Committee for Standardization (CEN), Brussels, Belgium.

EN 13649 (2002). Stationary source emissions - Determination of the mass concentration of

individual gaseous organic compounds - Active carbon and solvent desorption method.

European Committee for Standardization (CEN), Brussels, Belgium.

German Bitumen Forum (2006). Progress report 2006.

Hicks R.G. (2002). Asphalt rubber design and construction guidelines. Northern California

Rubberized Asphalt Concrete Technology Center - California Integrated Waste Management

Board, Sacramento, CA, USA.

ISS (2012). ISS-INAIL database for sanitary environmental risk analysis. Istituto Superiore di

Sanità, Rome, Italy (in Italian).

Leroy G. (1989). Bitumen analysis by thin layer chromatography (Iatroscan), Proceedings, 4th

Eurobitume Symposium, Madrid, Spain, 4-6 October, 1989.

Lindberg H. K., Väänänen V., Järventaus H., Suhonen S., Nygren J., Hämeilä M., Valtonen J.,

Heikkilä P. and Norppa H. (2008). Genotoxic effects of fumes from asphalt modified with

waste plastic and tall oil pitch. Mutation Research/Genetic Toxicology and Environmental

Mutagenesis, 653, (1-2), 82-90.

Marescalco P.A.C. and Zanetti M.C. (2010). Impact of the contamination due to chlorinated

solvents and chromium VI on the human receptors resident near an industrial area.

Proceedings, Hazardous and Industrial Waste Management, Chania, Crete, Greece, 5-8

October, 2010.

Nellensteyn F.J. (1928). Relation of the micelle to the medium in asphalt. Journal of the Institute

of Petroleum Technologists, vol. 14.

Province of Turin (2011). TYREC4LIFE: Development and implementation of innovative and

sustainable technologies for the use of scrap tyre rubber in road pavements, Research

proposal. Province of Turin, Turin Italy.

Ruffino B., Fiore S. and Zanetti M.C. (in press). Environmental–sanitary risk analysis procedure

applied to artificial turf sports fields. Accepted, November 2012.

Santagata E. and Zanetti M.C. (2012a). The use of products from end-of-life tyres in road

pavements. ECOPNEUS, Milan, Italy (in Italian).

Santagata E. and Zanetti M.C. (2012b). Evaluation of the possible use in road pavements of

crumb rubber from end-of-life tyres, Final Report. Province of Turin, Turin, Italy (in Italian).

Santagata E., Riviera P.P., Dalmazzo D., Lanotte M., Zanetti M.C., Fiore S. and Ruffino B.

(2012). Design and construction of a full-scale test section with asphalt rubber gap-graded

wearing course mixture, Procedia: Social & Behavioral Sciences, vol. 53, 524-534.

Santagata E., Zanetti M.C., Corbetta G. and Fornai D. (2013a). “POLIPNEUS Project. Science

and technology of bituminous binders and mixtures containing crumb rubber from end-of-life

tyres”, Strade & Autostrade, vol. 97/XVII, n. 1, January/February 2013, 62-68 (in Italian).

Santagata E., Lanotte M., Dalmazzo D. and Zanetti M.C. (2013b). Potential performance-related

properties of rubberized bituminous mixtures produced with dry technology. Proceedings,

12th

International Conference on Sustainable Construction Materials, Pavement Engineering

and Infrastructure, Liverpool, UK, 27-28 February, 2013.


Santagata E., Zanetti M.C., Dalmazzo D., Lanotte M., Ruffino B. and Fiore S. (2013c).

Development of a performance-based certification system and database for paving

applications of crumb rubber derived from end-of-life tyres. Proceedings, First Asia Pacific

Rubber Conference (APRC) 2013, Surat Thani, Thailand, 5-6 September, 2013.

Stout D. and Carlson D.D. (2003). Stack emissions with asphalt rubber. A synthesis of studies.

Proceedings, Asphalt Rubber Conference 2003, Brasilia, Brazil, 2-4 December, 2003.

Watts R.R., Wallingford K.M., Williams R.W., House D.E. and Lewtas J. (1988). Airborne

exposures to PAH and PM2.5 particles for road paving workers applying conventional asphalt

and crumb rubber modified asphalt. Journal of Exposure Analysis and Environmental

Epidemiology, April- June 1998, 8(2), 213-229.

Way G.B., Kaloush K.E. and Biligiri K.P. (2012). Asphalt-rubber standard practice guide,

Second edition. Rubber Pavements Association, Tempe, AZ, USA.

Zanetti M.C., Fiore S., Ruffino B. and Santagata E. (2011). End of life tyres: crumb rubber

physical and chemical characterization for reuse in paving applications. Proceedings, Sardinia

2011, Thirteenth International Waste Management and Landfill Symposium, S. Margherita di

Pula, Italy, October 2011.

Zanetti M.C., Marescalco P.A.C., Fiore S. and Ruffino B. (2013). The site of national interest

Ecolibarna: characterization, risk analysis with RACHEL software and remediation.

Proceedings, SiCon 2013, Rome, Italy, 21-23 February, 2013, 233-247 (in Italian).

use of crumb rubber in road paving applications: …...the asphalt rubber supplier provided both hot...

Documents