Physical and Chemical Properties

This chapter describes the physical and chemical properties of asphalt products. Section 3.1 discusses physical and chemical properties and how manufacturing processes influence the chemical composition of asphalt products and how chemical composition, in turn, influences physical properties. Section 3.2 describes the different types and uses of asphalt products. Section 3.3 notes the use of asphalt modifiers and additives. Section 3.4 examines vapors and fumes and the differences in their chemical composition, as well as the difficulties involved in producing asphalt fumes in the laboratory that are representative of fumes produced in the field. Section 3.5 discusses the usefulness of various analytical sampling and analysis methods used to characterize asphalt exposures, and section 3.6 is a brief summary.

3.1 Properties

Table 3–1 is a summary of the physical properties, chemical names and synonyms, and numbers from the Chemical Abstract Service (CAS) and the Registry of Toxic Effects of Chemical Substances (RTECS) for asphalt, asphalt fumes, and asphalt-based paints.

Asphalt is the residuum produced by the distillation of crude petroleum at "atmospheric and under reduced pressures in the presence or absence of steam" [Puzinauskas and Corbett 1978]. Performance specifications (physical properties) and not chemical composition direct asphalt production. To meet performance specifications, asphalt may be air blown or further processed by solvent precipitation or propane deasphalting. In addition, the products of other refining processes may be blended with asphalt to achieve the desired performance specifications. Therefore, the exact chemical composition of asphalt depends on the chemical complexity of the original crude petroleum plus the manufacturing processes involved in creating the product.

Crude petroleum consists mainly of aliphatic compounds; cyclic alkanes; aromatic hydrocarbons; heterocyclic compounds containing nitrogen, oxygen, and sulfur atoms; and metals, e.g., iron, nickel, and vanadium. The proportions of these chemicals can vary greatly because of significant differences in crude petroleum from oil field to oil field or even from different locations in the same oil field [AI 1990a]. Consequently, because of their complexity, no two asphalts are chemically identical, and chemical analysis cannot be used to define the exact chemical structure or chemical composition of asphalt.

Elemental analyses indicate that most asphalts contain 79 to 88 weight percent (wt %) carbon, 7 to 13 wt % hydrogen, traces to 8 wt % sulfur, 2 to 8 wt % oxygen, and traces to 3 wt % nitrogen (see the examples in Table 3–2) [Speight 1992a]. Although heteroatoms (i.e., nitrogen, oxygen, and sulfur) make up only a minor component of most asphalts, they profoundly influence the differences in the physical properties of asphalts from different crude petroleum sources. The heteroatoms cause differences in physical properties by forming functional groups and imparting polarity to the asphalts; in turn, these functional groups and differences in polarity cause a variety of chemical interactions among asphalt molecules [Roberts et al. 1996; Speight 1992a].

CAS=Chemical Abstract Service.
RTECS=Registry of Toxic Effects of Chemical Substances.
*CAS number 64742-93-4.

3.2 Types and Uses of Asphalt

3.2.1 Paving Asphalts

Paving asphalts are manufactured to meet performance specifications that are based on the physical properties of the asphalt product and not on chemical properties.

3.2.1.1 Manufacturing Processes

To produce a paving asphalt, crude petroleum is heated from 340 to 400 °C (644 to 752 °F) and introduced at atmospheric pressures into a distillation tower in which the most volatile components will vaporize. The volatile components rise in the distillation tower and slowly cool. More volatile components will rise higher in the tower than less volatile components. When temperatures drop below the boiling point of a specific component, that component will condense and be collected in a tray. The remaining residuum is called "straight-reduced asphalt" [Federal Aviation Administration (FAA) 1991; Roberts et al. 1996; Speight 1992a].

However, because distillation is an inefficient separation process, considerable amounts of volatile components may remain in the residuum. Components with higher boiling points may need to be removed to meet the desired physical specifications. The residuum may be transferred to a vacuum distillation tower in which the distillation process is repeated at a reduced pressure. As pressure is reduced, the less-volatile components can vaporize at lower temperatures, and cracking (thermally breaking apart the asphalt molecules) is less likely to occur. The resulting residuum may be used to produce a "vacuum-processed asphalt." If steam is used during distillation, the resulting residuum is called a "vacuum-processed, steam-refined asphalt" [FAA 1991; Roberts et al. 1996; Speight 1992a].

While physical properties may change dramatically during the manufacturing process, the chemical nature of an asphalt does not change unless thermal cracking occurs. Raising the temperature to 400 to 565 °C (752 to 1049°F) will increase the likelihood of cracking and cause the more-volatile components (and even the components with higher boiling points) to be released from the residuum [Roberts et al. 1996; Speight 1992a].

Other common manufacturing processes include solvent precipitation, air blowing, and blending of asphalts or crude petroleum from different sources. Solvent precipitation (usually using propane or butane) removes highboiling-point components from vacuum-processed asphalt; these components are then used to make other products. Solvent precipitation results in a harder asphalt that is less resistant to temperature changes. This asphalt is often blended with straight-reduced or vacuum-processed asphalts. Paving asphalt is not usually air blown, but air can be introduced to a vacuum-processed asphalt to form a more viscous product that is more resistant to weather and temperature changes. The air-blowing process can be a continuous or a batch operation. Because a continuous operation is faster and results in a softer asphalt, it is preferred for processing paving asphalts. Crude petroleum from different sources can be blended before refining so that the resulting asphalt meets required specifications; similarly, a higher viscosity asphalt can be blended with a lower viscosity asphalt to produce an asphalt of intermediate viscosity [Roberts et al. 1996; Speight 1992a].

3.2.1.2 Types, Uses, and Grades

Three types of asphalt are used in paving: asphalt cements, cutback asphalts, and emulsified asphalts. Cutback and emulsified asphalts also are called liquid asphalts because they are liquid at ambient temperatures [Roberts et al. 1996; Speight 1992a].

Asphalt cement refers to a straight-reduced or vacuum-processed asphalt manufactured according to paving specifications. Asphalt cements are used mainly as binders (4% to 10% of the mixture) in hot-mix asphalts and serve to hold the aggregate together [Roberts et al. 1996; Speight 1992a]. The asphalt cement is heated to about 149 to 177°C (300 to 350 °F) and mixed with mineral aggregate heated from 143 to 163 °C (290 to 325 °F). Once transported to the worksite, the hot-mix asphalt is applied to the roadway. The temperature at application is generally between 112 and 162 °C (235 and 325 °F) [AI 1990a; FAA 1991; Roberts et al. 1996; Speight 1992a].

The grade of asphalt cement is measured by either penetration or viscosity and depends on the amount of the higher boiling-point components that have been removed from the residuum. Penetration grade is determined by the depth a standard needle will sink in a 5 HUMAN HEALTH EFFECTS 3 PHYSICAL AND CHEMICAL PROPERTIES sample of asphalt cement at a given temper-ature, for a given time, and under a given load. There are five penetration grades: 40–50, 60–70, 85–100, 120–150, and 200–300 dmm (0.1 mm). The hardest asphalt cement (40–50 dmm) will allow the least penetration, while the softest (200-300 dmm) will allow the most penetration.

Viscosity grade can be based on the original asphalt cement (AC-2.5, AC-5, AC-10, AC-20, AC-30, and AC-40) or on the asphalt residue (AR-4000, AR-8000, and AR-16000). Asphalt residue is asphalt cement aged in a rolling-thin-film oven. Both the AC number and the AR number indicate viscosity in hun-dreds of poises (gram per centimeter second) at 60 °C (140 °F) [Roberts et al. 1996; Speight 1992a]. Performance grades as defined by the Strategic Highway Research Program [Roberts et al. 1996] are not included here because this information adds little to understanding the health effects of asphalt exposures.

To achieve the same density of the final pave-ment, a harder asphalt cement requires more compaction by a roller than does a softer (i.e., less-viscous) asphalt cement or it must be laid at a higher temperature [FAA 1991; Roberts et al. 1996; Speight 1992a]. Even if two asphalt cements have the same penetration or viscosity grades at one temperature, they may not have the same grade at a different temperature when the underlying chemistries of the two are different [Roberts et al. 1996; Speight 1992a].

A cutback asphalt is made by adding a diluent (typically a petroleum distillate) to an asphalt cement. Because cutback asphalts are liquids at or near ambient temperatures, they are often applied by spraying them on a surface. Cut-back asphalts are graded depending on their viscosity at 60 °C (140 °F). Cutback asphalts are further classified according to the type of solvent used to liquefy the asphalt cement. Rapid-curing cutback asphalts are made by adding gasoline or naphtha and are mainly used as surface treatments, seal coats, and tack coats. Kerosene is added to produce medium-curing cutback asphalts, and diesel or other gas oils are added to produce slow-curingcutback asphalts. Medium- and slow-curing cutback asphalts are used mainly as surface treatments, primer coats, tack coats, mix-in-place road mixtures, and patching mixtures [Roberts et al. 1996; Speight 1992a].

Emulsified asphalt is a mixture of two normally immiscible components (asphalt cement and water) and an emulsifying agent (soap is an example). Since emulsified asphalts are liquids, they are often applied at ambient temperatures up to 150 °C (300 °F) simply by spraying them on a surface. Emulsified asphalts are graded as either cationic (electro-positively charged mi-celles containing asphalt molecules) or anionic (electro-negatively charged micelles containing asphalt molecules), depending on the emul-sifying agent. Emulsified asphalts are further graded on the basis of their setting rate, i.e., rapid, medium, or slow. Rapid-setting grades are used for surface treatments, seal coats, and penetration macadams; medium-setting grades are used for patching mixtures; and slow-setting grades are used in mix-in-place road mixtures, patching mixtures, tack coats, fog coats, slurry sealants, and soil stabilizers[Roberts et al. 1996; Speight 1992a].

3.2.2 Roofing Asphalts

Roofing asphalts are manufactured to meet roofing performance specifications on thebasis of the physical properties and not the chemical properties of the asphalt product.

3.2.2.1 Manufacturing Processes

Although straight-reduced or vacuum-proc-essed asphalts are used to make roofingproducts, much of the asphalt used in roofing operations is made by air blowing these asphalts. Air-blown asphalts are called oxidized asphalts, air-refined asphalts, or roofing asphalts. In the air-blowing process, asphalt hardens as it comes into contact with air at 204 to 288 °C (400 to 550 °F [Corbett 1979]). Once the asphalt has the desired specifications, it is either held in storage tanks at elevated temperatures or is cooled before it is pumped into storage containers, where it solidifies. The air-blowing process can be a continuous or a batch operation. A batch operation is slower and produces a harder asphalt, and is preferred for processing roofing asphalts [Puzinauskas and Corbett 1978; Corbett 1979; Roberts et al. 1996; Speight 1992a].

Air blowing combines oxygen with the hydrogen in the asphalt to produce water vapor. This decreases saturation and increases cross-linkages within and among different asphalt molecules. The process is exothermic (produces heat) and may include a series of chemical reactions, such as oxidation, condensation, dehydration, dehydrogenation, and polymerization. These reactions cause the amount of asphaltenes (hexane-insoluble materials) in the asphalt to increase, the amounts of polar aromatics (hard resins), cycloalkanes, and nonpolar aromatics to decrease (soft resins), while the amount of aliphatic compounds (oils and waxes) remains about the same (Table 3–3); the oxygen content of the asphalt increases (Moschopedis and Speight 1973; Corbett 1975; Puzinauskas and Corbett 1978; Boduszynski 1981; Roberts et al. 1996; Speight 1992a].

The effect of air blowing also can be facilitated with chemical compounds. Ferric chloride, aluminum chloride, zinc chloride, phosphorus pentoxide, copper sulfate, or boric acid have been used to produce catalytic asphalts. Moreover, sulfur or chlorine can be added to the asphalt to react with hydrogen, yielding hydrogen sulfide or hydrogen chloride, respectively [Puzinauskas and Corbett 1978; Corbett 1979; Roberts et al. 1996; Speight 1992a].

Roofing asphalt specifications also can be influenced by blending crude petroleum from various sources or asphalts. Crude petroleum can be blended before refining and air blowing to meet needed specifications; similarly, a high-viscosity roofing asphalt can be blended with a low-viscosity roofing asphalt to produce an intermediate-viscosity roofing asphalt [Speight 1992a].

3.2.2.2 Types, Uses, and Grades

Although the focus in this section is on BUR products, a variety of asphalts are used in roofing products. The asphalts used to pro-duce shingles, roll goods, felts, and under-layments may or may not be air blown and are shipped hot and kept hot until they are used in a manufacturing process [AREC 1999].

Most of the air-blown asphalt used for roofing is shipped as solid kegs in cardboard cartons and heated in a kettle at the worksite until it becomes a liquid. Asphalt also may be de-livered as a hot liquid in a tanker truck, but this practice is becoming less common be-cause of higher costs, regulatory constraints, and product supply considerations. Asphalt delivered by tanker may be also heated to the desired temperature in the tanker or trans-ferred to a kettle for heating, after which it is pumped to the roof [AREC 1999].

Some cutback and emulsified asphalts are also used in roofing operations [Speight 1992a]. Although these asphalts represent only a small amount of the asphalt used in roofing, a recent study suggests their use may be increasing [Herbert et al. 1995].

Mopping-grade roofing asphalts in the United States are classified into four types—I, II, III, and IV—based on increasing hardness. The type of asphalt to be used is determined by the grade or slope of the roof. Type I roofing asphalt, often called “dead level,” has the low-est softening point and is used on surfaces with a grade of 0.5 in/ft or less. Types II and III roofing asphalts are typically used on roofs with grades of 0.5 to 1.5 and 1 to 3 in/ft, respectively. Type IV roofing asphalt has the highest softening point and is used on roofs with a grade of 2 to 6 in/ft [ASTM 1997].

A mopping-grade roofing asphalt is best applied when the asphalt is at its equiviscous temperature, or the temperature at which the viscosity of the asphalt is either 125±25 centistokes for hand mopping or 75±25 centistokes for mechanical spreaders [Intec 1998]. Table 3–4 lists recommended appli-cation temperatures for mopping-grade roof-ing asphalts [AI 1990a, Appendix C]. How-ever, to achieve these application temper-atures, the asphalt must be heated to even higher temperatures in the kettle. During recent surveys of roofing operations in which Type III roofing asphalts were used, the kettles were often maintained at 288 °C(550 °F). Several kettlemen stated that if the demand on the roof for asphalt is high, they will heat the asphalt to 316 °C (600 °F) [Hayden 1998; Mead 1998].

Care must be exercised when operating kettles at high temperatures. Temperatures in excess 5 HUMAN HEALTH EFFECTS 3 PHYSICAL AND CHEMICAL PROPERTIES 10 Health Effects of Occupational Exposure to Asphalt of the flash point of the asphalt can result in fires that cause serious burns. If not quickly extinguished, kettle fires can spread rapidly to the exterior of the kettle and engulf equipment, including propane tanks (and gasoline tanks on some models), with catastrophic results. An explosion hazard may be created when the kettle lid is closed and the atmosphere in the headspace of the kettle is within explosive limits [AREC 1999].

3.2.3 Asphalt-Based Paints

Asphalt-based paints are specialized cutback asphalt products that may contain a small amount of lampblack, aluminum flakes, or mineral pigments. Asphalt-based paints are used as protective coatings in waterproofing operations and in other similar applications [AI 1990b]. The asphalt used to make an asphalt-based paint may or may not be air blown [Speight 1992a,b]; the only requirement is that the final product has the flow and drying characteristics of a paint. Basically, asphalt-based paints may be applied at or near ambient temperatures by spraying or brushing. Once the asphalt-based paint is applied to a surface, it should not flow, and it should harden quickly. This is achieved either by manipulating the manufacturing process or by the addition of diluents.

3.3 Asphalt Modifiers and Additives

Although the subject of asphalt modifiers and additives is beyond the scope of this document, it would be remiss not to mention their use because a worker may be exposed to a modifier or an additive or even to their decomposition products, and their presence in asphalt may affect the composition of asphalt fumes. Asphalt modifiers and additives are used to enable asphalt products to meet desired performance specifications and serve a variety of functions, as described in Table 3–5 [Roberts et al. 1996; Speight 1992a].

3.4 Asphalt Vapors and Fumes

When asphalt products are heated, vapors are released; as these vapors cool, they condense. By definition, the condensate is asphalt fume; however, because the components in the vapor do not condense all at once, workers are exposed not only to asphalt fumes, but also to vapors. The physical nature of fumes and vapors has not been well characterized, but fumes should be fairly viscous. When liquid asphalt products are used at ambient temperatures, workers are exposed to the liquid product and to vapors, but not to fumes. Fume particles may collide and stick together, making it difficult to characterize fume particle size. Some of the vapors may condense only to the liquid phase, thus forming a viscous liquid containing some solids.

3.4.1 Cutback Asphalts, Emulsified Asphalts, and Asphalt-Based Paints

Cutback asphalts, emulsified asphalts, and asphalt-based paints are liquids and are applied at or near ambient temperatures [Roberts et al. 1996; Speight 1992a]. Because these products are liquids, workers may be exposed through both respiratory and dermal contact. These products are applied in a variety of ways, including by spraying, and if the spray nozzle becomes clogged, a worker may face increased dermal exposure and clothing contamination when cleaning the nozzle. Petroleum distillates added to asphalt products can dry the skin, weakening the protective barrier skin provides and facilitating the entry of various compounds into the body. Furthermore, petroleum distillates introduce the hazard of exposure to volatile organic compounds (VOCs).

The composition of the vapors released from these asphalt products during drying can be explained with Raoult’s Law: The composition of the vapor phase above a solution is directly proportional to the mole fraction and vapor pressure of each component in the solution. Other factors influencing the composition of the vapor phase are that (1) as temperature increases, vapor pressure increases, which may allow appreciable quantities of certain compounds to exist in the vapor phase, and (2) the chemical composition of a chemical class will generally increase in complexity in the vapor phase. Generally, the smaller compounds in a given chemical class will have higher vapor pressures.

As liquid asphalt products harden from the outside surface in, the added diluents slowly evaporate from the outside surface, thus trapping part of the diluent inside the asphalt layer. However, if these asphalt products are heated even slightly, not only will the same compounds vaporize faster, but higher concentrations of the same compounds will vaporize along with other compounds that do not vaporize appreciably at ambient temperatures. In the absence of significant increases in temperature, these asphalt products are expected to release primarily vapors from the evaporating solvent.

No reports discussing the chemical analysis of cutback asphalts or emulsified asphalts were found. However, Robinson et al. [1984] reported on the analysis of select polycyclic aromatic compounds (PACs) in several a s p h a l t -b a s e d paints u s i n g ga s chromatography-mass spectroscopy (GC/MS). Benz[a]anthracene, benzo[a]pyrene (B(a)P), benzo[e]pyrene, chrysene, and phenanthrene were measured, but only trace amounts of phenanthrene (<0.01%) were detected.

3.4.2 Comparison of Vapors and Fumes from Paving and Roof-ing Asphalts

Information presented in the previous section indicates that (1) higher temperatures increase the chemical complexity of paving and roofing asphalts, (2) paving and roofing vapors and fumes are chemically more complex than liquid asphalt vapors and fumes, (3) vapors and fumes from paving asphalts are different from those of roofing asphalts because of differences in application temperatures, i.e., roofing asphalts are applied at higher temperatures (166 to 229 °C [340 to 455 °F]) than paving asphalts (112 to 162 °C [235 to 325 °F]), and (4) differences in manufacturing processes affect the composition of asphalt and consequently the composition of fumes. Compared to air-blown roofing asphalts, nonoxidized asphalts generally contain more aliphatic compounds, about the same amount of cycloalkanes and nonpolar aromatics, and smaller amounts of polar aromatic compounds and asphaltenes (Table 3–3). This does not mean, however, that air-blown roofing asphalts contain more aromatics than nonoxidized asphalts.

Differences in the way paving and roofing asphalts are handled also probably influence the composition of vapors and fumes. For example, a hot-mix asphalt begins to cool from the moment it leaves the plant and may not be used immediately when it arrives at a work-site, whereas roofing asphalts are heated continuously and stirred occasionally at a worksite until they are needed.

Using GC/MS, several investigators reported on chemical analyses of paving and roofing asphalt fumes [Reinke and Swanson 1993; Niemeier et al. 1988; Lunsford and Cooper 5 HUMAN HEALTH EFFECTS 3 PHYSICAL AND CHEMICAL PROPERTIES Health Effects of Occupational Exposure to Asphalt 13 1989; Hatjian et al. 1995a, 1997]. Others have used liquid chromatography (LC) methods to analyze for polycyclic aromatic hydrocarbons (PAHs) in asphalt vapors and fumes. Because of methodology limitations, LC methods should not have been used to analyze PAHs in asphalt fumes; therefore, results from these studies are not discussed here. However, LC methods and their limitations are discussed in section 3.5.

Reinke and Swanson [1993] collected paving asphalt fumes from a straight-reduced, vacuum-processed 85-100 grade asphalt cement. Using the NIOSH protocol [Thayer et al. 1981; Sivak et al. 1989], the authors collected fumes from a storage tank at a hotmix plant at a temperature of 149 °C (300 °F) and from laboratory generation at temperatures of 149 and 316 °C (300 and 601 °F). The fume samples were then analyzed for selected PACs (Table 3–6). The results indicate that only two- and three-ring PACs were present in the fumes from the storage tank, but that the chemical classes identified in the laboratory-generated fumes were mostly two- and three-ring PAHs and a few three-ring sulfur-PACs (S-PACs) and four-ring PAHs. Several of the four-ring PAHs are carcinogenic, i.e., benz[a]anthracene and chrysene. Methylated chrysenes, pyrenes, and fluoranthenes were also detected and may be of concern because of their structural relationship to known carcinogens. The concentration of four-ring PAHs was highest in fumes generated in the laboratory at the highest temperature. However, the concentration of tworing PAHs was lowest in laboratory fumes generated at the highest temperature, most abundant in fumes from the storage tank, and lower in the fumes generated in the laboratory at the ambient temperature of the storage tank.

In the laboratory tests, once the asphalt melted, the mixture was stirred constantly until it reached the desired temperature. The higher the generation temperature, the longer the mixture had to be stirred before the desired temperature was reached, which caused more of the two-ring PAHs to reach the surface of the liquid and escape before collection of the fumes began. This stirring procedure might explain why some three-ring PACs were found in the laboratory-generated fumes, but not in the fumes from the storage tank. In the storage tank, not enough of the three-ring PACs were brought to the surface to escape in sufficient concentrations to be detected. Of interest is that a summation of the measured PAHs (Table 3–6) only accounts for 0.8% to 1.3% of the total asphalt fumes, assuming the density of asphalt to be 1 gram per milliliter (g/mL) [Speight 1992a]. This is not surprising since aliphatic compounds compose the majority of the compounds present in asphalt fumes.

Thayer et al. [1981] and Niemeier et al. [1988] collected asphalt fumes generated in the laboratory from Type I and Type III roofing asphalts and Type I and Type III roofing coal-tar pitches. The fumes were generated attemperatures of 232 and 316 °C (450 and 601 °F) and analyzed for PAHs by GC/MS. The results of the analysis and information regarding which PAHs are considered carcinogenic are given in Table 3–7. The results indicate that asphalt fumes had much lower concentrations of PAHs than the coal-tar-pitch fumes and con-sisted mainly of two- to four-ring PAHs with small amounts of five-ring PAHs. Concen-trations of two-, three-, and some four-ring PAHs were generally lowest in the laboratory fumes generated at the highest temperatures. Concentrations of the two-, three- and some four-ring PAHs decreased before collection began, because the laboratory-generated fumes were allowed to escape until the entire mixture reached the desired temperature. Moreover, once the asphalt melted, the mixture was stirred con-stantly, causing more of the two-, three-, and four-ring PAHs to reach the surface of the liquid and escape before collection began.

Also, the higher the generation temperature, the longer the mixture would be stirred before the desired temperature was reached.

Because concentrations of the remaining four-and five-ring PAHs were low and similar in amount at the two generation temperatures, similar trends in concentration were not observable. Air blowing appears to have had little effect on concentrations of the higher molecular weight PAHs, but decreased concentrations of the lower molecular weight PAHs (Table 3–7).

Niemeier et al. [1988] reported that analysis by nuclear magnetic resonance spectroscopy (NMR) indicated that asphalt fume condensates were less than 1% aromatic and more than 99% aliphatic, whereas the coal-tar-pitch condensates were more than 90% aromatic. Assuming 13C NMR was used, these percentages are indicative of the carbon atom character. These results also indicate that in asphalt fumes, most of the carbon atoms are contained in aliphatic groups, while in the coal-tar-pitch condensates, most of the carbons are contained in aromatic groups.

Sivak et al. [1989, 1997] heated Type III roofing asphalt from the same lot used by Niemeier et al. [1988] at 316 °C (601 °F), generated fume condensates, and separated them into fractions by high-performance liquid chromatography (HPLC) [Belinky et al. 1988]. Using GC/MS, Lunsford and Cooper [1989] characterized the chemical classes present in these asphalt fume fractions. Results (Table 3–8) indicate the relative abundance of each compound class in the frac-tions, but not in the classes. The results also indicate that many of the compounds are ali-phatic and many compound classes containalkylated isomers. Fraction A constituted64.1% of the asphalt fume condensate, whilefractions B, C, D, and E constituted 8.3%, 10.5%, 11.5%, and 5.6%, respectively. Mostcompound classes were found in more thanone fraction, indicating the complexity createdby the addition of alkyl groups and thepresence of many different isomers for eachcompound class. Despite this complexity, onlyfractions B and C caused tumors in a mouse-skin painting study (see section 6.2.1). FractionB contained mainly alkylated two- and three-ring PAHs, oxygen-PACs (O-PACs), and S-PACs; a few alkylated four-ring PAHs (py-renes and fluoranthenes); and a variety ofketones. Little is known about the toxicity ofmost of these compounds except that some O-PACs may cause cancer and some S-PACsmay cause mutations [Tennant and Ashby1991; Pelroy et al. 1983; McFall et al. 1984].The two- and three-ring O-PACs were notdetected in fraction C, but the other PACs and ketones found in fraction B were also found in fraction C. Fraction C also contained alkylatedand unalkylated isomers of chrysene andbenz[a]anthracene, a few four-ring S-PACswith and without alkyl groups attached, a widevariety of ketones, and alkanoic acids.

Chrysene and benz[a]anthracene are known carcinogens, but little is known about the toxicity of the other observed PACs except that some S-PACs may cause mutations [Pelroy et al. 1983; McFall et al. 1984]. Many alkylated isomers of PACs have been identified in asphalt fumes, but although little is known about their carcinogenic and genotoxic activity, these PACs are a cause for concern because of their structural similarity to known carcinogens and genotoxins.

Probably because an air-blown roofing asphalt was used, numerous oxidized compounds were found; however, if paving fumes had been studied, it is possible that not as many oxidized compounds would have been detected. The presence of alkanoic acids probably indicates that the starting crude petroleum source contained aldehydes because aldehydes are easily oxidized to carboxylic acids. Because of the process (addition of ferric chloride and air blowing) used to prepare the roofing asphalt, it is unlikely that very many aldehydes survived the manufacturing process. However, if any of the aldehydes did survive, it is unlikely that workers would have been exposed to them in the field, because aldehydes are easily oxidized at ambient temperatures, and the elevated temperatures needed at a field site would only hasten the oxidation process. While it is more likely that aldehydes would not oxidize as easily during manufacturing of paving asphalts, it is also likely most of the aldehydes would oxidize before asphalt pavers could be exposed. Because only a few nitrogen-containing compounds were found, the asphalt may have been manufactured from a crude petroleum containing low amounts of nitrogen; therefore, other asphalts with a higher nitrogen content may yield more nitrogen-containing compounds.

Hatjian et al. [1995a, 1997] reported on the GC/MS analysis of nine PAHs in personalbreathing- zone air samples. The samples were collected from two sets of asphalt pavers (P1 and P2), two sets of asphalt roofers (R1 and R2), manual laborers (M) who had no apparent occupational exposure to PAHs, and office workers (C) who were used as controls. The median percentage for each PAH determined, the number of rings in each PAH, and the number of samples for each group are given in Table 3–9. No results for the control group are included in Table 3–9 because none of the samples contained detectable amounts of any of the measured PAHs.

Naphthalene accounted for 80% to 90% of the measured PAH exposure for each group, except in group R2 (60%). Naphthalene, acenaphthene, and phenanthrene accounted for 98% to 99% of the measured PAH exposure for each group, with the exception of groups R1 and R2 (94% and 84%, respectively). The four- and five-ring PAHs each accounted for less than 1% of the measured PAH exposure for all groups except R1 and R2. For group R1, the four- and five-ring PAHs (except pyrene) each accounted for less than 1% of the measured PAH exposure; pyrene accounted for 1%. For group R2, benz[a]anthracene and pyrene accounted for 2% and 7% of the measured PAH exposure, respectively, while the five-ring PAHs accounted for less than 1% each. Because roofing asphalts are heated to hotter temperatures and applied at higher temperatures than paving asphalts, which increases the amount of the larger PAHs in the fumes, the median percentage values for two-and three-ring PAHs were lower (Table 3–10).

Table 3–10 contains a summary of the PAH data and smoking habits for each group. For each PAH, a mean concentration was calculated on the basis of a 3-day geometric mean of the air samples expressed as an 8-hour time-weighted average (TWA). These mean concentrations were summed for different groups of PAHs: all PAHs (3 nine PAHs), all PAHs except naphthalene (3 eight PAHs), and four- and five-ring PAHs (3 four- and five-ring PAHs). Summations of four- and five-ring PAHs were not calculated for groups other than roofers, nor were summations of mean concentrations calculated for the control group because in both instances there were not enough data recorded at detectable levels to allow these calculations.

B(a)P was detected in personal-breathing-zone samples as follows: manual laborers (5.9% of the samples), group P1 (5.6%), group P2 (3.3%), group R1 (28%), group R2 (25%), and the control group (<1%). The highest B(a)P concentrations were 0.17 [Hatjian 1995], 0.02 and 0.20 :g/m3 for manual laborers, pavers, and roofers, respectively. Hatjian et al. [1995a, 1997] stated that the manual laborers had no apparent occupational exposure to B(a)P. Their reported exposures probably resulted from the environment, an unidentified source, or a sampling and analytical error. The authors did not provide background concentrations of PAHs downstream of the paving or roofing operations. However, in his review of the NIOSH draft hazard review, Hatjian [1999] stated that "background concentrations of PAHs downstream of the roofing operations in Hatjian’s studyshowed non-detectable levels of B(a)P on the day of monitoring the roofing operation."

When reviewing these data, consideration must be given to B(a)P concentration and sampling variability. Since the highest B(a)P concentration for a paver was only twice the reported detection limit for B(a)P, this determination is not reliable. Among all the groups, only one roofer had more than one of three personalbreathing-zone samples with detectable concentrations of B(a)P [Hatjian 1995].

Environmental and personal factors as well as work practices could contribute to pavers’ and roofers’ exposures to B(a)P. For example, 22% to 75% of the workers in the paving and roofing groups were smokers, and at least one paving group was exposed to diesel exhaust. In addition, the highest B(a)P concentration for a roofer may be atypically high because of work practices at both roofing sites (workers knelt while spreading roofing asphalt with a trowel or brush) and the high kettle temperature (300 °C [572 °F] at the R1 site). This temperature is about 70 °C (158 °F) higher than the highest recommended temperature for roofing application (Table 3–4).

3.4.3 Field-Generated Versus Laboratory-Generated Asphalt Fumes

When a large quantity of an asphalt fume is needed, collecting the fumes in a laboratory setting is more practical than collecting fumes at worksites. Niemeier et al. [1988], Lunsford and Cooper [1989], and Reinke and Swanson [1993] studied asphalt fumes generated in the laboratory using the procedure described in section 3.4.2. This procedure involves placing asphalt in a vessel, heating it, and stirring it at least 200 revolutions per minute once it had melted sufficiently. The fumes were allowed to escape until the desired temperature was reached, at which point collection began by pulling air at 10 liters per minute (L/min) through a series of cold traps. (This procedure could account for the differences in chemical composition often noted between asphalt fumes collected in the field and those generated in the laboratory.) The fumes were collected for at least 8 hours.

The following studies evaluated under what conditions laboratory-generated asphalt fumes could mimic asphalt fumes generated in the field.

Kriech and Kurek [1993] showed how genera-tion conditions can affect the composition of fumes. Using a variety of analytical techniques (gas chromatography with flame ionization detection [GC/FID], gas chromatography with flame photometric detection [GC/FPD], gas chromatography with atomic emission detection [GC/AED], and GC/MS), they compared asphalt fumes generated in the laboratory with fume samples collected from the headspace in a storage tank at a hot-mix plant (paving asphalt), from the headspace in kettles (roofing asphalt), and from personal-breathing-zone samples. Both the field and laboratory fumes were collected with a series of cold traps, while the personal-breathing-zone samples were collected at the field sites on a membrane filter backed up with a sorbent tube.

Kriech and Kurek concluded that temperature, rate of stirring, and pulling versus pushing the collection air all affected the chemical compo-sition of the fumes. Based on simulated distill-ation data and analyses of high-molecular-weight S-PACs, they also concluded that the storage-tank samples resembled the personal-breathing-zone samples more closely than did the laboratory-generated samples. However, the S-PAC data also indicated that the storage-tank samples contained more S-PACs than the personal-breathing-zone samples.

Similarly, Reinke and Swanson [1993] con-sidered asphalt fumes collected from a storage tank at a hot-mix plant to be representative of asphalt fumes from a field paving site. How-ever, Reinke and Swanson [1993] did not give a detailed analysis of the storage-tank samples or the personal-breathing-zone samples. Therefore, these questions remain: Are storage-tank fumes truly representative of the asphalt fumes to which workers are exposed in the field? Are storage-tank fumes more representative of field fumes than of fumes generated in the laboratory using the NIOSH protocol [Thayer et al. 1981; Sivak et al. 1989]?

In another study, Brandt et al. [1985] collected field and laboratory asphalt fume samples and analyzed them for total particulates, the benzene-soluble fractions of total particulates, and PAHs. In the field, point-emission and personalbreathing-zone (using personal-type samplers) samples were collected at both roofing and paving operations, while the laboratory samples were collected under a variety of conditions. The intent was to identify the conditions under which asphalt fumes generated in the laboratory would be similar to those collected at actual worksites.

Results indicated that temperature and heating time affected chemical composition of the fumes. Higher temperatures and longer heating times resulted in higher exposures to total and benzene-soluble particulates and changed the chromatographic elution profiles of the PAHs. Comparing analyses of the field samples and the laboratory-generated samples led Brandt et al. to conclude that their "laboratory rig" couldproduce laboratory fumes representative of field fumes. In this study, fumes were collected when the center of the sampler was placed 13 cm above the level of the asphalt. Brandt et al. felt this distance prevented the sampler frominfluencing the chemical equilibrium above the asphalt surface when the sampler flow rate was 2 L/min. Collection times had to be short (15 to 60 minutes), and generation temperature had to be close to that used in the field.

3.5 Analytical Sampling and Analysis Methods

This section is not intended to be an all-inclusive list of the analytical sampling and analysis methods available for characterizing asphalt vapor and fume exposures. Most of the methods are nonspecific, and none can be used to characterize total asphalt fume exposure.

3.5.1 Total and Respirable Particulates

NIOSH Method 0500 can be used to determine total particulates, and NIOSH Method 0600 can be used to determine respirable particulates [NIOSH 1994]. The only difference between these two methods is that NIOSH Method 0600 uses a size-selective inlet. Both methods are nonspecific; consequently, whatever is deposited on the sampling medium and remains until the sample is analyzed is included in thedetermination. Moreover, when matrices such as an asphalt fume are sampled, air stripping can cause volatile fume components to be lost from the sampling medium. Because both methods use a membrane filter as the sampling medium, these methods are not useful for collecting vapors.

3.5.2 Benzene-Soluble Fraction ofTotal Particulates

NIOSH Method 5042 can be used to determine both total particulates and the benzene-soluble fraction of total particulates employing a single sampler [NIOSH 1998]. Previously, benzene solubles and total particulates were determined using different samplers, thus making a comparison of results questionable. Also, the methods used to determine benzene solubles were originally developed for coal-tar-pitch volatiles, and the results were correlated to adverse health effects [Occupational Safety and Health Administration (OSHA) Method 58, 1986].

These methods of determining benzene solubles have commonly been used with other matrices where the results were only suspected of relating to an adverse health effect. These methods are nonspecific because most organic compounds are soluble in benzene and because asphalt fumes contain many organic compounds and compound classes not found in coal-tar-pitch volatiles. Anything in addition to asphalt fumes that is deposited on the sampling medium and is benzene soluble will interfere with the determination. Air stripping can cause volatile fume components to be lost from the sampling medium. Because NIOSH Method 5042 uses a membrane filter for the sampling medium, it is not useful for collecting vapors.

3.5.3 Polycyclic Aromatic Hydrocarbons

NIOSH Method 5506 uses liquid chroma-tography with ultraviolet and fluorescence detection (LC/UV/Fl) to determine selected PAHs, and NIOSH Method 5515 uses gas chromatography with a flame ionization de-tector (GC/FID) to determine selected PAHs [NIOSH 1998; NIOSH 1994]. In NIOSH Meth-od 5506, some of the PAHs (acenaphthene, acenaphthylene, anthracene, chry-sene, fluorene, naphthalene, and phenanthrene) are determined by UV detection, and the other PA H s ( b e n z[a ] a n th racen e ,benzo[b]fluoranthene, benzo[k]fluoranthene, benzo[ghi]perylene, B(a)P, benzo[e]pyrene, dibenz[a,h]anthracene, fluoranthene, indeno[1,2,3-cd]pyrene, and pyrene) are de-termined by fluorescence. NIOSH Methods 5506 and 5515 have been used to determine selected PAHs in matrices that contain only a few alkylated PAHs in relatively low con-centrations compared with the unalkylated PAHs of interest, and possibly a few alkylated and unalkylated N-, O-, and S-PACs. These matrices would include most coal derived products and combustion by-products. However, because asphalt fumes are composed of many alkylated isomers (e.g., mono, di, tri, and tetra-methyl) of PAHs, along with O-PACs and S-PACs, with the exception of naphthalene and some 3-ring PAHs, they are so chemically complex that they cannot be separated into discrete compounds (see section 3.4.2). The greater the lack of resolution between compounds, the less reliable are the quantification results. Because of the poor resolution obtained with asphalt fume samples, quantification is unreliable when these or other HPLC or GC/FID methods are used.

Also, the limitations of NIOSH Methods 5506 and 5515 require that an alternative method (such as GC/MS) be used to confirm the identity of any suspected PAHs, including naphthalene and other possible baseline resolved PAHs. Any compounds reported using NIOSH Methods 5506 and 5515, or similar methods, are tentative identifications at best, and the more complex the matrix, the more unreliable these identifications become. Furthermore, since chromatographic software programs assign peak identification based on the largest peak in a given time window and not on retention time, the wrong peak may be assigned and analyzed. Since these methods use a gradient elution (e.g., the mobilephase composition varies during the chromatographic run), retention times may vary, thus, further complicating the selection of the correct peak for identification and analysis. For reasons stated above, these problems can be overcome for matrices consisting of coal-derived products or combustion by-products. However, for asphalt fumes, these are formidable problems, because the alkylated PAHs are more abundant and are in higher concentrations than the PAHs of interest.

Figure 3–1 shows a typical asphalt fume chromatogram obtained usingliquid chromatography and a fluorescence detector. Figure 3–1 indicates base line resolution is not achieved; hence, this analytical technique should not be used for determining the concentration of PAHs in asphalt fume samples. Therefore, excluding results for naphthalene and some 3-ring PAHs, the analytical results for PAHs previouslyreported using NI OSH Method 5506 or similar HPLC methods are unreliable. Moreover, since many previously reported studies do not include chromatograms or sufficient experimental details, the methodology and data cannot be critically reviewed; therefore, results fornaphthale n e and 3-ring PAHsshould also be considered suspect. Furthermore, tangent skimming along the oscillations in the chromatogram for resetting the baseline would not be meaningful, because the remaining peak is too small and the peak widths are too wide to represent a single compound.

While NIOSH Method 5506 does not allow for varying the fluorescence excitation andemission wavelengths, these wavelengths can be varied to improve sensitivity; however, varying these wavelengths will introduce addedconcerns. If the fluorescence response at the new wavelength is not roughly zero, theaccompanying autozero that occurs will distort the chromatogram, and the data. For these reasons, PAH analyses in asphalt fumes by HPLC/fluorescence techniques are considered unreliable.

The UV chromatogram obtained by usingNIOSH Method 5506 is even more complex because all the PACs and other chemical classes in the asphalt fume sample absorb UV light at the wavelength being monitored (254nanometers [nm]). NIOSH Method 5515 would produce an even more complex chromatogram since the FID responds to everything passing through it.

Because of these limitations and a growing concern that all PACs in asphalt fume may play a role in adverse health effects, a method is needed to monitor all PAC material. A NIOSH investigation used a modification of NIOSH Method 5506 (i.e., Method 5800) to monitor all PAC material in asphalt fumes [Hanley and Miller 1996a,b; Almaguer et al. 1996; Miller and Burr 1996a,b, 1998; Kinnes et al. 1996; NIOSH 1998]. Basically, the same analytical equipment is used, except the LC column has been removed and the UV detector has been replaced with a second fluorescence detector. Because no LC column is used, the entire sample reaches the flow cell at once, resulting in a rapid and sensitive analysis of the sample. The two fluorescence detectors monitor different excitation and emission wavelengths. One set of wavelengths is more sensitive to two-and three-ring PACs, and the second set of wavelengths is more sensitive to four-ring and higher ring PACs.

3.5.4 Selected MethodsSelected Solvent Methods

NIOSH Method 1550 can be used to determine exposure to naphthas [NIOSH 1994]. The term naphthas includes petroleum ether, rubbersolvent, petroleum naphtha, VM&P naphtha, mineral spirits, Stoddard solvent, kerosene, and coal tar naphtha. This method may be useful because some liquid asphalt products contain petroleum distillates for which exposure limits have been established. The samples are collected on a sorbent tube and analyzed using GC/FID. Because these solvent mixtures are chemically complex and the components elute over a wide temperature range, interferences from other substances are possible.

Other NIOSH methods can be used to determine selected solvents that may be present in asphalt vapors and fumes [NIOSH 1994]. NIOSH Methods 1300 and 1301 have been used to determine ketones, and NIOSH Method 1501 has been used to determine total PAHs [Hanley and Miller 1996a,b; Almaguer et al. 1996].

3.6 Conclusions

An analysis of the chemical data indicates that paving and roofing asphalts arequalitatively and quantitatively different; therefore, the vapors and fumes from these asphalt products may also be presumed to be different. Chemical composition of asphalt vapors and fumes varies and depends on crude petroleum sources, type of asphalt, temperature and mixing during themanufacturing process, and temperature and extent of mixing during either laboratory generation or field operations. Although asphalt vapors and fumes have not been well characterized, the analysis of selected PAHs in asphalt vapors and fumes from asphalt products has been of interest. Many studies have been directed to the identification of PAHs in asphalt fume samples. The most meaningful of these studies used GC/MS for the analysis. PAH data obtained by HPLC/ fluorescence techniques are not included, be-cause the PAH identifications are uncertain and the results unreliable, see section 3.5.3.

Robinson et al. [1984] used GC/MS to analyze several asphalt-based paints for chrysene,benz[a]anthracene, B(a)P, benzo[e]pyrene, and phenanthrene; they detected only phenanthrene (0.01%). Several other investigators havereported on the chemical analysis of paving and roofing asphalt fumes [Niemeier et al. 1988; Lunsford and Cooper 1989; Reinke andSwanson 1993; Hatjian et al. 1995a, 1997]. Low levels of carcinogenic PAHs have been detected in laboratory-generated asphalt fumes. Reinke and Swanson [1993] detected 0.02 µg/m3 chrysene in fumes generated in the laboratory at 149 °C (300 °F). Niemeier et al. [1988] meas-ured low concentrations of several carcinogenic PAHs in roofing asphalt fumes generated in the laboratory at both 232 and 316 °C (450 and 651 °F). Most of the PAHs in the Niemeier et al. study were two-, three-, and four-ring PAHs. Lunsford and Cooper [1989] reported finding two- to four-ring PAHs along with many alkylated PAHs, O-PACs with and without alkyl groups, and S-PACs with and without alkyl groups in laboratory-generated asphalt fume fractions that caused tumors in a mouse-skin-painting study. The presence of O-PACs and S-PACs is a cause for concern, since some O-PACs may cause cancer, and some S-PACs may cause mutations [Tennant and Ashby 1991; Pelroy et al. 1983; McFall et al. 1984]. Also, because little is known about the carcinogenic and genotoxic activity of most of the alkylated PACs, these PACs are a cause for concern because of their structural similarity to known carcinogens and genotoxins.

Few studies have been directed at the iden-tification and measurement of PAHs in asphalt fumes generated at U.S. worksites. Reinke and Swanson [1993] collected paving asphalt fumes from a storage tank at 149 °C (300 °F) at a hot-mix plant (Table 3-6). Although they had detected chrysene in laboratory-generated asphalt fumes, they did not detect chrysene in asphalt fumes collected from the storage tank, and although two- and three-ring PAHs were found in the storage tank fumes, four-ring PAHs were not.

Hatjian et al. [1995a, 1997] reported on a GC/MS analysis for selected PAHs in asphalt paving and roofing fumes collected at several worksites. Two- and three-ring PAHs accounted for 99% of PAH exposures in the two paving asphalt groups. In the two roofing asphalt groups, two- and three-ring PAHs accounted for 84% and 94% of PAH exposures, respectively. Naphthalene accounted for 60% to 90% of PAH exposures for all work groups. B(a)P was detected in less than 6% of the personalbreathing-zone air samples from asphalt road pavers and manual laborers who had nooccupational exposure to PAHs and in 28% or 25% of the personal-breathing-zone air samples obtained from asphalt roofers, R1 and R2, respectively.

In a NIOSH study, environmental samples from paving operations were analyzed for PACs as a class, but no individual PAHs were determined [Hanley and Miller 1996a,b; Almaguer et al. 1996; Miller and Burr 1996a,b, 1998; Kinnes et al. 1996].

While data regarding the presence of carcinogens in asphalt fumes generated at U.S. worksites are limited, the occasional detection of B(a)P at these sites [Hatjian et al. 1995a, 1997] and more frequent detection of B(a)P and other carcinogenic PAHs in laboratory-generated asphalt fumes indicate that under some conditions, known carcinogens are likely to be present [Niemeier et al. 1988; Lunsford and Cooper 1989; Reinke and Swanson 1993]. Moreover, asphalt fumes generated at high temperatures are probably more hazardous than fumes generated at lower temperatures. Because asphalt fume samples collected in the field have not been well characterized, additional research is needed to better characterize them. Also, laboratory generation methods need to be evaluated to identify those that produce asphalt fume samples representative of fumes to which workers are exposed.

The presence of numerous alkylated PAHs, O-PACs, and S-PACs in asphalt fumes is cause for concern. Although little is known about their toxicologic activity, their structural similarity to known carcinogens and genotoxins is troublesome. If these compounds are a health concern, new sampling and analytical methods specifically for these compounds would need to be developed. Given the chemical complexity of asphalt fume samples, the most likely methods would utilize GC/MS techniques.

Various NIOSH methods have been used for characterizing asphalt vapor and fume exposures. However, most of the methods are nonspecific, and none are useful for characterizing total asphalt fume exposure. New or improved analytical sampling methods need to be developed.