Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
3 C H A P T E R 1 A. Introduction Hot mix asphalt (HMA) is a well-established paving material with over 100 years of proven performance. It has been used to pave 94% of the more than 2.5 million miles (4.0 million km) of hard-surfaced roads in the United States [FHWA 2008; National Asphalt Pavement Association (NAPA) 2010]. It is combined at an elevated temperature in either batch or drum mixing plants and then compacted at temperatures ranging from 250°F (121°C) to 325°F (163°C) (Kuennen 2004; Newcomb 2005a). High temperatures are necessary to ensure complete drying of the aggregate and subsequent bonding with the binder, coat- ing of the aggregate by the binder, and workability for adequate handling and compaction. All of these processes contribute to good pavement performance in terms of durability and resis- tance to permanent deformation and cracking. In the past few decades, advances in asphalt technology such as polymer-modified binders and the use of angular aggregate that improve resistance to permanent deformation (stone matrix asphalt, for example), and an emphasis on compac- tion for quality assurance, have resulted in further increases in HMA mixing temperatures up to a limit of 350°F (177°C), where polymer breakdown in the binder may occur. These high temperatures are linked to increased emissions and fumes from HMA plants (Stroup-Gardiner et al. 2005). In addition, the HMA process consumes considerable energy in drying the aggregate and heating all materials prior to mixing and com- pacting. The use of warm mix asphalt (WMA) technology results in reduced production and paving temperatures with- out sacrificing the quality of the final product. This has led to a wider range of available production temperatures that may be employed by the contractor. Economic, environmental, and possible performance ben- efits motivate the reduction of asphalt mixing and compac- tion temperatures. Past efforts that date back to the late 1950s include binder foaming processes (using either steam or water), binder emulsification, and incomplete aggregate dry- ing (Kristjansdottir 2006; Zettler 2006). The latest technol- ogy is WMA, where temperatures range from 175°F (79°C) to 295°F (146°C), and for which the following benefits have been cited [Koenders et al. 2002; Jones 2004; National Center for Asphalt Technology (NCAT) 2005; Newcomb 2005b; McKenzie 2006; Button et al. 2007]: ⢠Decreased energy consumption of up to 30% to 40% ( Jenkins et al. 2002; Kuennen 2004; Prowell et al. 2011a). ⢠Reduced emissions and odors at the plant [30% reduction in carbon dioxide (CO2)] (Kuennen 2004; Prowell et al. 2011a). ⢠Reduced fumes and improved working conditions at the construction site (fumes below detection limits) (Newcomb 2005b; Prowell et al. 2011a). ⢠Decreased plant wear and costs. ⢠Extended haul distances, a longer pavement construction season, and a longer construction day than if produced at typical HMA temperatures (NCAT 2005; Kristjansdottir 2006; Ursich 2008; Prowell et al. 2011b). ⢠Reduced construction time for pavements with multiple lifts (Kuennen 2004). ⢠Improved workability and compactability (Bistor 2008; Prowell et al. 2011a, 2011b). ⢠Reduced initial costs in some cases. ⢠Reduced aging and subsequent susceptibility to cracking and raveling. ⢠Decreased life-cycle costs in some cases. There have been a number of products and processes introduced to the marketplace to produce WMA over the last several years (Prowell et al. 2011a). These include waxes, surfactants, mineral additives, and mechanical foaming pro- cesses. Waxes, such as those from the Fischer-Tropsch pro- cess, montan, and polyolefin, are highâmelt-point materials that become fluid at mixing and placement temperatures and then harden at service temperatures. Surfactants reduce the surface tension of the liquid binder, allowing it to better coat Background
4the aggregate at low temperatures and to remain workable at reduced placement temperatures. WMA production in the United States has increased expo- nentially in recent years, from 19.2 million tons [17.4 million metric tons (mt)] in 2009 to 86.7 million tons (78.7 million mt) in 2012, an increase of more than 400% in only 3 years (Hansen and Copeland 2013). The work in this research study focused on central plantâproduced foamed WMA because foaming asphalt is the largest segment of the WMA market. Accord- ing to a survey done by NAPA (Hansen and Copeland 2013), mechanical foaming units were responsible for about 88% of all WMA produced in 2012. Binder foaming has become the process of choice for most contractors using WMA. There are different techniques used in the production of plant-produced foamed asphalt mix- tures, including the use of zeolite, wet sand, and mechani- cal foaming units. Zeolite, a mineral additive, has a small amount of water contained in its interstices that is released in the form of steam when the material is exposed to the hot asphalt mixture. The steam then foams the liquid binder, increasing its volume and allowing it to coat the aggregate at a lower temperature. The warm asphalt mixture (WAM) foam process used in Norway consists of pre-coating the aggre- gate with low-viscosity binder and then adding a foamed hard binder to provide cohesion to the mixture. In the low- emission asphalt (LEA) process used by McConnaughy in New York, coarse aggregate is dried and coated with binder, after which wet sand and an additive are introduced to create foam with the already heated binder. In current mechanical plant foaming processes in the United States, cold water is injected into a hot binder stream that may be anywhere from 285°F to 340°F (140°C to 171°C). The cold water turns to steam when it comes in contact with the hot binder, and the water expands creating an increased volume of binder. The amount of water used in producing the foam has normally varied between 1.0% and 3.0% by weight of binder. Foam- ing works in two ways to promote mixing at lower tempera- tures: (1) it increases the volume of the binder, which makes it easier to coat particles, and (2) it reduces the overall viscos- ity of the binder through shear thinning, which makes the mix more workable (Fort et al. 2011). All of these processes and products work to allow asphalt mixing and placement temperatures to be reduced, but they introduce a new set of conditions that are not readily accounted for in the selection of materials and mixture design. The changes brought about by WMA in mixture compo- nents, mix processing, and plant design have left many paving technologists questioning the validity of current mix design methods in adequately assessing the volumetric needs of asphalt mixtures and the physical characteristics required to meet per- formance expectations. While a variety of other NCHRP proj- ects (9-43, 9-47A, 9-48, 9-49, 9-49A, and 9-52) attempted to address many of these issues, this study considered the impact of WMA foaming technology on the volumetric and perfor- mance characteristics of binders and mixtures. While foaming is popular, there have been a number of questions surrounding the use of water in WMA. For instance: ⢠Will the presence of water in the mix have detrimental effects on performance in terms of its strength and durability? ⢠How long will the effect of the foam last? ⢠How can a distinction be made between the foaming abili- ties of different binders? ⢠How much mix production and paving temperature reduction may be realistically achieved by foaming? ⢠Will all foaming techniques produce the same quality and quantity of foam? ⢠How will the presence of additives in the liquid binder affect the ability of the binder to foam? ⢠How will polymer-modified binders behave in the pres- ence of foaming? ⢠Will mix design and evaluation procedures need to be modified to accommodate foaming? These questions were addressed during the course of this study. B. Foaming Theory and Applications This section presents a review of foaming technology in both the asphalt mixture and other industries. This is an essential first step prior to researching and understanding the foaming characteristics of binders. B.1. Definitions It is important to identify a few common terms used in the industry and literature to characterize foams. The term âfoamabilityâ is often used to describe the extent to which a liquid can be foamed, both in terms of volume and stability. Foamability is a property of the liquid, although it is governed by external factors such as gas concentration and temperature. There are types of foams prepared by processes not involving the direct dispersion of a gas in a liquid phase (Klempner et al. 2004). These may be prepared by the leaching of a fugitive phase such as a water-soluble salt, sintering small particles dis- persed in a heat-stable matrix, fusing initially discrete polymer particles that initially entrap air or other gases, and forming a polymer matrix around hollow spheres. These processes do not follow the same steps of gas dispersion, bubble growth, and stabilization as are found in the binder foaming process and, thus, are not considered in this review. The terms âfoamâ and âfrothâ are often used interchangeably. However, the term âfoamâ is used to describe a two-component
5 system composed of a gas and a liquid (typically), which when broken down leaves a homogenous liquid phase (Pugh 2005). In some cases, foam is also regarded as an emulsion of a liquid and a gas. The term âfrothâ is used to described a three- component system, typically a gas, a fluid, and a solid, which when broken down results in a two-component composite. Not all foams are the same. Foams are typically classified and studied in two different categories (Adamson and Gast 1997; Pugh 2005): ⢠Polyederschaum (polyhedral foam): In this type of foam, the volume of gas is much larger as compared to the vol- ume of the fluid. The fluid exists in the form of very thin films separating the gas, which are also referred to as lamella. The name is derived from the fact that the gas cells are polyhedral in shape (Adamson and Gast 1997). ⢠Kugelschaum (spheroidal foam): In this type of foam, the gas volume is relatively low compared to the polyeder- schaum. A relatively thicker film of the fluid separates the gas bubbles. Figure 1-1 shows a schematic of the two types of foam. These two types of foam could coexist in the same material, given the highly dynamic process of foaming. Bubbles could start as isolated and then begin impinging on each other and form polyhedral foam as they grow. In the context of foamed asphalt mixtures, the latter (kugelschaum) type of foam appears to be more relevant as the bubbles would emerge to the surface and collapse before a thin wall between bubbles could be formed. This was substantiated by Hailisilassie et al. (2014) by means of X-rays of binder foam in a study con- ducted at the Swiss Federal Laboratories for Materials Science and Technology. B.2. Foaming in Other Industries There are a number of industries that require the design and use of foamed materials. The applications range from food, pharmaceutical and health care products, polymers, and in some cases even metals (Koehler et al. 1998). Sev- eral different methods are used to produce foam in the food industry, including whipping, injection, sparging (bubbling a gas through a liquid), and shaking (German et al. 1985). The characteristics of the foam produced using each of these methods are different. There are several different variables associated with each of these four methods that can result in the production of foams with very different characteristics. Arzhavitina and Steckel (2010) present a detailed review of the different types of foams typically used in the pharmaceu- tical sector. One of the types of foam that is commonly used is the pressurized aerosol foam. Aerosol foam is produced by using a propellant vapor to drive the liquid through a nozzle from a storage can. In the packaging industry, foaming is typically achieved using a blowing agent, which upon controlled energy input (e.g., thermal) causes a desired level of foaming of a polymer. In a typical foaming process, the blowing agents gasify through chemical reaction or thermal decomposition under the foam- ing process conditions. The blowing agent is mixed with pel- letized polymer and loaded into an extruder. It decomposes in the extruder barrel at an elevated temperature, resulting in gas formation in the polymer melt. The gas formed in the poly- mer melt leads to bubble nucleation and growth. On exiting the extruder, the polymer profile expands in volume because of the bubble growth. In addition to chemical reaction and thermal decomposition, gas can be delivered to the extruder through either pre-saturation of the polymer pellets or high- pressure injection right into the polymer melt inside the extruder. Foams produced in this way have a smaller bubble size and a more uniform foam structure. B.3. Metrics to Characterize Foam in Other Industries The two most common attributes that are used to charac- terize foams across different industries are the volume and stability of the foam. In fact, as will be evident in the sum- mary that follows, these two attributes and concomitant met- rics are similar to what is currently being used to characterize binder foaming: expansion ratio (ER) and half-life (HL). Foamability of wines is an important characteristic that is measured to select base wines for the production of sparkling wines. Over the past few years, researchers in this area have developed two different sets of parameters to characterize the foamability of different wines. The first set has three different parameters: (1) maximum height (HM) reached by the foam Figure 1-1. Schematic to illustrate two types of foam (adapted from Pugh 2005).
6after carbon dioxide injection for a specific interval of time; (2) foam stability height (HS) during carbon dioxide injec- tion, which represents persistence of the foam; and (3) foam stability time (TS), identified as the time it takes for all the bubbles to collapse after carbon dioxide injection has stopped. The second set has two different parameters: (1) E, or foam expansion, which is very similar to the ER used to character- ize expansion of binders; and (2) Lf, or foam stability, which is broadly defined as the area under the foam height versus time curve following the peak expansion divided by the maximum height of the foam. These two parameters are sometimes accom- panied with a characteristic bubble size. Andrés-Lacueva et al. (1996) demonstrated that the first set of parameters, HM and HS, were strongly correlated, and therefore only HM and TS were adequate to describe the foamability of different wines. Gallart et al. (1997) compared the relative benefits between the two sets of parameters and reported that the precision of each set of parameters was dependent on the extent to which the wine foamedâthe second set being more appropriate for low-foaming wines. Food products are routinely characterized to determine the influence of factors such as production method and ingredi- ents (proteins in particular) on their foaming characteristics. The two parameters that are often used in the food industry are (1) foam overrun and (2) foam stability. Foam overrun is roughly a measure of the volume of foam produced. For example, Phillips et al. (1987) defined it as the difference between the weights of 100 mL of protein and 100 mL of foam divided by the weight of 100 mL of the foam. The measurements were made at a specific time after foaming the subject material. Foam stability was defined as the ability of the foam to resist rupture or collapse under the influence of gravity. In more recent studies, Raymundo et al. (1998) tried to modify these two metrics to come up with a single param- eter, referred to as the foaming index. In this case, overrun was measured not at a prespecified point in time but con- tinually as a function of time using automated equipment. The integral of the overrun with respect to time was used as a single parameter (the foaming index). Raymundo et al. tried to demonstrate the sensitivity and advantages of using this single parameter to characterize the foam. In pharmaceutical applications, two metrics that are used to describe the quality of the foam are foam density and break- ability (Arzhavitina and Steckel 2010). Foam density is simply the relative density of the foam with respect to water. Stability of the foam is characterized in terms of three different types of breaking behavior: (1) quick-breaking foams, or foams that are thermally unstable and collapse on contact with skin; (2) lathers, or stable foams that demonstrate a tendency to increase in volume when subjected to shearing action; and (3) breakable foams, or foams that are stable at skin tempera- tures but collapse and spread easily upon application of shear forces. Other metrics and methods, similar to those used in the food sector, are also used to describe binder foam stability. In addition to foam density and stability, the rheology of the foam is considered of importance in pharmaceutical applica- tions. However, there are several challenges associated with measuring the rheological properties of foam, even when the foam is relatively stable compared to foamed binder. Kealy et al. (2008) conducted a detailed study on the methods that can be used to measure the rheological properties of several different types of pharmaceutical foams. They cited several difficulties associated with the use of standardized methods, such as use of the dynamic shear rheometer (DSR), to measure the rheo- logical properties of the foam. For example, they report that the small gaps in typical test geometries [0.03 to 0.08 in. (1 to 2 mm) with parallel plates or a few micrometers with the cone and plate geometries] are inappropriate because of the presence of large bubbles in the foam. Even if the bubble sizes were small, there were problems associated with slippage between the foam and the end plates. In their study they tried to overcome these limitations by using larger gaps and serrated plates. They also used vane geometry to measure the rheological properties. They were able to measure the yield stress for the different foams by applying a monotonically increasing shear stress until the speci- men failed or started to flow (Figure 1-2). They also measured the complex shear modulus of different foams in the frequency domain and demonstrated that at low frequencies the rheologi- cal properties measured using the vane geometry were similar to the properties measured using the serrated parallel plate. In the context of foamed binder, this is an important finding because it may be possible to use geometry similar to the vane (e.g., a portable paddle viscometer) to obtain real-time viscosity and rheological measurements of the foam as it collapses. One other characterization procedure that is worth men- tioning is foam drainage. Drainage tests are frequently con- ducted to evaluate the structure and stability of different kinds of foam (Pugh 2005). There are several different methods that are used to drain foams to determine their characteris- Figure 1-2. Typical results for the yield strength of shaving foam (adapted from Kealy et al. 2008).
7 tics. Koehler et al. (1998) identified four different methods of draining foams: (1) free drainage under gravity; (2) wetting of the foam by placing the foam in contact with a liquid bath; (3) forced drainage, where a constant liquid supply is pumped onto the top of the foam; and (4) response of the foam to an external pulse. Parameters that are measured in a foam drain- age experiment may include the rate at which the liquid flows out of the foam and changes in pressure across the boundaries of the foam. Measurements are usually carried out using opti- cal or electrical-based instrumentation attached to a column designed to create and drain foams (discussed in the following section). Koehler et al. (1998) present the analytical solutions that describe the rate at which foams drain as a function of the properties of the foam. However, based on the literature, it appears that the drainage method is more appropriate to characterize the polyederschaum category of foams (foams with very thin films of the fluid separating the gas). B.4. Mechanisms of Bubble Formation in Fluids The mechanics of bubble formation in fluids have much to do with how gas is delivered in the fluid. Two different situa- tions have been studied in the literature: (1) gas delivered in the fluid with the use of an orifice, nozzles, capillaries, and porous plates; and (2) gas dissolved in the fluid through dif- fusion. In the first case, the nozzle is submerged in the fluid, and the gas is blown into it. As the rate of gas flow is increased, three regions of bubble formation can be obtained. At very low rates of gas flow, bubble formation is almost a static prob- lem. Bubble size is primarily determined by orifice diame- ter, surface tension, and the fluid density. Bubble formation under this condition is also the basis of the drop-volume method for measuring surface tension. At intermediate gas- flow rates, bubbles are formed at a constant frequency, and the bubble size is uniform, dependent on gas flow rate, ori- fice diameter, and the volume of the gas chamber upstream from the orifice. At high gas-flow rates, a normal distribu- tion with respect to the logarithm of the bubble diameter is established for the turbulent flow of gas through the orifice. As turbulence becomes fully developed, a marked decrease in the bubble diameter occurs. For bottom entrance flow, the bubble diameter appears to be independent of the orifice diameter (Leibson et al. 1956). While bubble formation and detachment at low flow rates is well understood, the forma- tion of a large quantity of fine bubbles at turbulent condi- tions appears to be due to the shattering of large bubbles by turbulent swirling discontinuous jets. It is also observed that, in general, lower surface tension helps create smaller bubbles. The second case, where gas is delivered in fluid through diffusion, can be illustrated with the microcellular plastics injection molding process (Xu 2010). In the microcellular injection molding process, a gas-polymer solution is first cre- ated in an extruder barrel. The gas at the supercritical fluid state is metered and injected into the barrel and then dissolved into molten polymer. As the gas flows into molten polymer, it forms large gas droplets (bubbles) since it takes time for the gas to be dissolved. The large droplets of gas are sheared, elongated, and broken into smaller bubbles with the rotation of the extruder screw. Then the gas quickly diffuses into the molten polymer due to the increased polymer-gas interfacial area and the reduction of diffusion length. The gas dissolu- tion process obviously depends on many factors, including gas pressure, temperature of the polymer melt, and diffusivity and solubility of the gas. It is possible that the gas is not completely dissolved in the polymer and small bubbles still exist during the gas-polymer mixing stage. These small bubbles will grow bigger in the next stage when the mixture is released through the exit of the extruder. Nucleation can also happen elsewhere in the mixture as long as the foaming condition is met. In the microcellular injection process, sudden release of pressure is the cause of bubble nucleation and subsequent bubble growth. In addition to direct gas injection, foaming agents can be mixed with polymers as additives, which will then decompose and give off the needed gas inside the extruder barrel. This is how gas is delivered in regular foam extrusion processes that will yield larger bubble sizes than the microcellular injection molding process. Bubble growth in polymer foaming with CO2 has been explicitly modeled, and the significance of material proper- ties has been demonstrated (Wang and Li 2008; Kim and Li 2011). In CO2 polymer foaming, a small weight percent of CO2 is injected into the polymer matrix. The bubble growth process is modeled as a quasi-static diffusion-driven pressure and momentum-balanced process in a viscoelastic fluid. Fig- ure 1-3 shows a single bubble model for a polymer-gas system. The foaming of liquid binder is a dynamic process where water is used as the foaming agent in the hot binder. As water Figure 1-3. Schematic of the single bubble model for polymer-gas systems.
8turns to steam in atmospheric pressure, it expands by a fac- tor of about 1,600, but in binder the steam is confined by the pressure of the liquid and the size of the pipe or vessel in which the foaming is initiated. Once the foamed liquid is released in the asphalt mixing chamber (drum or pugmill), the expansion of the binder reaches its full potential as it is mixed with the aggregate. The degree of expansion of the binder is dependent on a number variables that will be dis- cussed later, but the major factors are the temperature and viscosity of the binder as well as system characteristics such as the external pressure and rate of flow. The dynamic nature of binder foaming is illustrated in Fig- ure 1-4, where it can be seen that expansion of the binder on the order of about 18 times happens almost instantaneously, and its subsequent collapse to its HL (where the volume equals half its maximum expansion) occurs within the first few seconds. At the maximum expansion, the density of the binder is reduced from 0.96 g/cm3 to about 0.05 g/cm3, which places it generally between a liquid and a gas for most substances. This unstable state quickly moves toward liquid initially as the gas in the bubbles is released. The rate of dissipation slows considerably after the initial collapse of the foam and continues for minutes afterward. It is believed that this region of stable dissipation is critical in terms of providing coating and workability in foamed asphalt mixtures, as will be discussed later. At this point, a dis- cussion of the theory of foaming provides the background needed to understand the mechanisms of binder foaming. The momentum equation of the polymer shell surround- ing the gas bubble can be expressed as: 2 2 0 (1-1)bubble system shell P P R r drrr R Râ«â â γ + Ï â Ï =θθ Where: Pbubble = the gas pressure inside the bubble. Psystem = the system pressure around the bubble. r = the radial distance from the center of the bubble. R = the bubble radius. Rshell = the bubble shell radius representing the poly- mer around the bubble. g = the surface tension at the bubble surface. trr and tqq = the stress components in the shell along the r and q directions, respectively. The viscoelastic rheology of the polymer-CO2 system can be modeled as: Ï = â λ + +  ï£ï£¬   Ï â η λ +  ï£ï£¬   Ï = â λ â +  ï£ï£¬   Ï + η λ +  ï£ï£¬   θθ θθ 1 4 4 (1-2) 1 2 2 2 3 2 3 2 3 2 3 d dt R R y R R R y R d dt R R y R R R y R rr rr o o Where: l = the relaxation time of the polymer/gas solution. ho = the zero-shear-rate viscosity of the polymer-gas system. y = the transformed Lagrangian coordinate, defined as: = â (1-3)3 3y r R From these equations, the bubble pressure Pbubble can be calculated by assuming a known bubble size at a given time step. On the other hand, it is clear that the bubble growth process is driven by gas diffusion into the bubble. The mass transfer equation can be modeled as: ( )ââ = ââ ââ â ââ (1-4)2 2 22Ct Dr r r Cr RRr CrG Where: C = the local gas concentration in the polymer matrix. DG = the gas diffusivity. The law of conservation of mass requires that the rate of change of the mass in the gas bubble be equal to the mass of gas diffusing into the bubble through the bubble surface. Thus, the bubble pressure can be related to the concentration gradient at the bubble surface by: ( )pi = pi ââ =43 4 (1-5)bubble 3 2ddt P RKT R D Cr r R Where: K = the universal gas constant. T = the temperature that is estimated from the heat trans- fer model for the current time step. Using these two equations, the pressure inside the bubble Pbubble can be related to the bubble size r as well. By comparing the bubble pressures Pbubble calculated from both the momentum and diffusion equations, the bubble size r at each time step can be determined using a recursive proce- Figure 1-4. Expansion ratio versus time curve of foamed binder.
9 dure. Such a procedure relates the bubble growth process with the material properties such as surface tension, relaxation time, viscosity, and gas diffusivity, as well as process parameters such as temperature, gas concentration, and the system pressure. The knowledge obtained from CO2 gas foaming of poly- mers can be readily extended to better understand the factors that affect the characteristics of water foaming of binders. The amount of water carried by the binder depends on the pressure of water and binder within the nozzle as well as the diffusivity of water or steam in the binder. While the pressure within the nozzle varies with the type of nozzle, the ability of the water to diffuse within the binder depends on the type of binder. The type of nozzle may also determine the water droplet size, which in turn determines how fast water converts to steam and the size of the steam bubbles in the binder. The importance of these two variables (type of nozzle and binder) has been recognized in previous studies ( Castedo and Wood 1983). Ozturk and Kutay (2014a; 2014b) concluded that water content and the driving pressure behind the foam have profound effects on bubble size, ER, and HL of the binder. They inferred the bubble size distribution through the appli- cation of Stokesâ law. It is also expected that the size and life of the bubbles (related to ER and HL) within the binder coming out of the nozzle will depend on the type of the binderâ more specifically, the surface tension of the binder surround- ing the bubble and its viscosity. This preliminary understanding of the relationship between material properties and foaming helps explain the findings from other previous studies. For example, Fu et al. (2011) dem- onstrated that binder grade is not related to its ability to foam. This is expected because the binder grade does not reflect the material properties such as viscosity and surface tension that are related to foaming. Another example is that the addition of liquid anti-strip agents improves the ability of the binder to foam (Abel and Hines 1979; Engelbrecht 1999; Fu et al. 2011). This is also expected because Bhasin et al. (2007) have dem- onstrated that the addition of liquid anti-strip agents reduces the surface tension of the binder. A more detailed explana- tion of the influence of binder properties on foam expansion and applied foaming theory to binder foaming is included in Appendix A. B.5. Summary Based on the survey of literature on foaming in binder and other industries, it is expected that a number of factors influ- ence binder foamability, including but not limited to: ⢠Surface tension of the binder (which is governed by the chemical composition of the crude oil and processing of the crude oil to produce the binder, including chemical and polymer modifications). ⢠Temperature of the binder. ⢠Viscosity of the binder (which is governed by the same fac- tors as surface tension). ⢠Water content used to produce foam. ⢠Size and dispersion of the water droplets introduced in the binder (which is governed by the characteristics of the nozzle delivering the water). ⢠Quality of the water and the presence of any additives in the water that influence its surface tension. ⢠Presence of anti-foaming agents in the binder. ⢠Use of foam promoters or other additives in the binder and the concentration of such additives. ⢠Atmospheric conditions (humidity, air pressure). C. Binder Foaming Technology and Applications Most of the available literature dealing with foamed binder concerns its application in the stabilization of soils or base materials, which began with Csanyi (1957). Although Csanyi used steam injected into the hot liquid binder, Mobil Oil Australia acquired the patent and modified it so that cold water was introduced into a stream of hot binder and then the foamed binder was mixed with cold, wet aggregate or soil (Muthen 1998). The desired outcome for stabilization was the coating of the fine particles by binder and the spot welding of the coarse aggregate to achieve some measure of cohesion. This type of stabilization is usually done in place but can also be accomplished through the use of a mixing plant (Muthen 1998). In recent years, this has been increasingly applied to in-situ recycling of pulverized asphalt pavements (Fu 2011). Fortunately, the process for WMA foaming, regardless of whether it is done mechanically through a nozzle, by intro- ducing wet sand, or by using zeolite, all amounts to adding a small quantity of water to the hot binder (1.0% to 3.0% by weight) (Fort et al. 2011) and allowing the generation of steam to expand the binder through the formation of voids. Thus, lessons learned in base and soil stabilization apply to WMA production using foam. C.1. Mechanical Foaming There are a variety of methods to disperse a foaming agent such as water into a medium such as liquid binder. Water in a liquid state is introduced to the hot binder stream, wherein it turns to steam. Mechanical systems, which could be used for foaming and applied in the field, have been identified by the authors as mechanical mixing, Venturi mixing, expansion chamber, shear/colloid mill, air-atomized water, and high- pressure atomized water. In some instances, more than one mechanism may be employed, or the system actually uses a hybrid of methods. The commercially available laboratory
10 units use either air-atomized water or pressurized water in an expansion chamber. C.1.1. Mechanical Mixing Systems using mechanical mixing (Figure 1-5) have an inlet port for introducing water into the binder stream. Downstream of the water introduction, the system may have paddles, baffles, or other means of mixing the water/ binder mixture. In this approach, some dispersion of the water into the binder occurs as the cold water turns to steam and creates bubbles, followed by additional agitation that may serve to more finely divide the bubbles and enhance the foaming action. C.1.2. Venturi Mixing Venturi mixing (Figure 1-6) is accomplished by introduc- ing the water into the binder line ahead of a constriction in the pipe. This reduced cross-section increases the pressure in the line, which is released as the cross-section opens. This creates turbulence in the fluid, which acts to mix the steam and liquid binder. C.1.3. Expansion Chamber In an expansion chamber (Figure 1-7), binder and water are introduced simultaneously. The cold water comes into con- tact with the hot binder and converts to steam, and expansion of the binder occurs. The foam is then forced out of a nozzle and into the mix. Expansion chambers can be configured in a manifold system where several are placed in parallel. C.1.4. Shear/Colloid Mill Shear/colloid mills (Figure 1-8) are used to mix substances such as binder and water. However, they are normally used to suspend binder in water for asphalt emulsions. In this case, cold water is introduced into a chamber with hot binder. The water turns to steam as it is forced with the binder through a very small opening between a rotor and a stator, which shears the water into very small particles. As the mixture exits the colloid mill, it expands and is introduced into the mix. C.1.5. Air-Atomized Water The use of air-atomized water (Figure 1-9) is a variation on the expansion chamber discussed previously. In this case, a stream of air is forced into the stream of water to break it into finer droplets. This should disperse the moisture throughout the binder stream, and as the moisture comes into contact with the hot binder, it expands. The chamber allows for the expansion of the foamed binder, which is then forced out through a nozzle into the mixture. C.1.6. High-Pressure Atomized Water In this type of system (Figure 1-10), water is forced through a very small orifice under very high pressure into the binder Figure 1-5. Mechanical mixing. Water Asphalt Paddles Foam Figure 1-6. Venturi mixing. Asphalt Water Foam Figure 1-7. Expansion chamber. Expansion Chamber Asphalt Foam Nozzle Water Figure 1-8. Shear/colloid mill.
11 line. At this point the binder line may be enlarged to accom- modate the increased volume, or its cross-sectional area may remain the same, in which case the pressure is considerably increased. A survey was sent to manufacturers of asphalt plant foam- ing equipment to gain an understanding of what types of systems exist and their operating characteristics. Of the seven manufacturers for which information was either found or who returned the survey, two used atomized water, one used an expansion chamber, one used multiple expansion cham- bers, one used mechanical mixing, one used a colloid mill, and one used a combination of mechanical mixing, Venturi mix- ing, and atomized water. The maximum water pressure varied from 60 to 2,000 psi (410 to 14,000 kPa), and only one manu- facturer identified a maximum binder flow rate of 120 gal/min (450 L/min). Water meters varied, with Coriolis, orifice, Ven- turi, positive displacement pump, sparling magna flow- meter, and turbine systems being identified. None of the plant foaming manufacturers had laboratory-scale foaming units, and none had a procedure for identifying the opti- mum water content to use in their foaming process. It is clear that plant foaming systems cover a span of rela- tively simple systems to complex systems, and thus the costs vary widely as well. The three laboratory foaming units inves- tigated in this study use either an expansion chamber or air- atomized water to produce foaming. As will be shown later in this report, the optimum water content for foaming is rela- tively independent of the type of foaming unit used. C.2. Zeolite Zeolite is a mineral additive that has a small amount of water contained in its interstices that is released in the form of steam when the material comes in contact with the hot asphalt mix. It is introduced into the pugmill of a batch plant or the mix- ing zone of a drum plant as the binder and aggregate are being combined. The steam then foams the liquid binder, increas- ing its volume and allowing it to coat the aggregate at a lower temperature. When an additive such as zeolite is used to provide mois- ture release, one must consider the potential residual impact of the additive once the moisture release has taken place. The release rates, moisture release quantities, and temperatures over which the moisture release takes place are important. There appears to be a gradual release of moisture that extends beyond the period required for improved workability during construction. D. Review of Past Work There are numerous views on the importance of foamed binder properties on the final mixture properties and char- acteristics. Bowering and Martin (1976) maintained that the cohesion and compressive strength of foam-produced mixes were greater when the ER was on the order of 15:1. Fu et al. (2011) reported that small changes in binder temperature and foaming moisture could significantly change the charac- teristics of foamed binder while only having minimal effects on the final mixture properties. However, there was a correla- tion between the foaming characteristics and the dispersion of binder in mixtures. The improved workability of foamed binder in WMA applications has been attributed to its shear-thinning char- acteristics by Fort et al. (2011). They claim this helps to maintain a greater film thickness without drain-down dur- ing storage and transport, yet helps the workability during paving and compaction. However, Clark and Rorrer (2011) noted that challenges with foamed binder included difficulty with handwork and temperature segregation, but that lower temperatures allowed for compaction of mixes sooner and for improved density. Clark and Rorrer (2011) found that in Virginia, typical mixture temperature reductions achieved with foaming sys- tems was between 25°F and 50°F (14°C and 28°C), but this depended on a number factors, including weather conditions, type of paving job, haul distance, and moisture content of the aggregate. It was common for the temperature of the mix to be increased if the paving required a significant amount of handwork, and for temperature segregation to be addressed by the use of a material transfer device. In the construction of I-55 and I-57 in Missouri, Fort et al. (2011) found that Figure 1-9. Air-atomized water. Foam Water Asphalt Air Figure 1-10. High-pressure atomized water. WaterWater Cross-Secon Asphalt Plan View Foam
12 the use of a polymer-modified binder (PG76-22) required that the average temperature of HMA production needed to be at 350°F (175°C) while the WMA average temperature was 293°F (145°C). The compaction temperature range was 320°F to 230°F (160°F to 110°C) and 275°F to 212°F (135°C to 100°C) for HMA and WMA, respectively, and the density of the final mat was 94% for both (Fort et al. 2011). The formation of steam bubbles within the liquid binder increases the volume of binder, and foam reduces the mass viscosity from that of the liquid at the same temperature. External work in the form of shearing may destroy some or all of the bubbles. External work includes mixing; dropping into the slat conveyor, the silo, and the truck; dumping from the truck into the paver; and movement through the paver to the spreader auger, the spreader auger to screed, and finally to the compactor. At the same time, the mix is cooling, and as this progresses, the volume of steam contracts to the point of liquid, below the boiling point. Ideally, the mix would remain at a relatively constant level of workability up through its compaction, at which point the binder mat would assume its final density and stiffness. Hajj et al. (2011) stated that Marshall mixture and volu- metric properties for an asphalt mixture that contained an unmodified foamed binder were met if the lab compaction occurred within 4 hours of plant production. They also observed that the effect of foaming in a WMA project is lost somewhere between 4 and 15 hours after short-term oven aging at 250°F (121°C). They reported problems in meeting Marshall mix design requirements when polymer- modified foamed WMA was compacted in the laboratory at the project placement temperature of 255°F (124°C), and that these could only be obtained when the compaction temperature was raised to 305°F (152°C). The temperature observations agree with Fort et al. (2011), who suggested that polymer-modified binders may require an elevated temperature. However, the time that the foaming charac- teristics are reported to last in Hajj et al. (2011) and that was also noted by Prowell et al. (2011b) is far greater than the matter of minutes reported by Muthen (1998). The observa- tions by Muthen (1998) are similar to the results obtained in the current effort. For WMA applications, Allosta et al. (2011) and Abbas et al. (2013) found that the characteristics of asphalt mix- tures produced by foaming were not very distinguishable from HMA. Nazzal et al. (2011) reported similar results for a foamed binder project in Ohio, except that the foamed mix seemed to have a greater tendency for rutting in the asphalt pavement analyzer. They speculated that the reason for the increased rutting may have been the presence of natural sand in the mix. This agrees with observations by Sakr and Manke (1985) and Abbas and Ali (2011). Fu et al. (2011) stated that binder temperature and foaming moisture could signifi- cantly change the foaming characteristics of the binder while only having minimal effects on the final mixture properties. However, they did find a correlation between the foaming characteristics and the dispersion of binder in mixtures. E. Study Objectives The objectives of this study were to (1) determine the properties of foamed binders that relate to asphalt mixture performance and (2) develop laboratory foaming and mixing protocols that may be used to design asphalt mixtures. This report documents the selected test methods for determining properties of binder foaming and foamed asphalt mixtures as well as the laboratory and field results and analysis used to characterize the foamed binders and their relation to asphalt mixture characteristics. When developing the testing protocols and equipment, the goals were to (1) minimize the complexity of any test- ing protocols and methods, (2) make any test applicable to both laboratory and field conditions, (3) simplify any testing equipment and design it to be as rugged as possible while being able to detect sensible differences in measurements, and (4) minimize, to the extent possible, the cost of the equipment and testing. F. Study Scope In concert with the project panel, a decision was made to focus on mechanical foaming techniques and the use of zeo- lite for foaming. These approaches to foaming make up the vast majority of processes used by agencies and contractors in the United States. While the WAM-foam and LEA pro- cesses for foaming are both valid and have been successfully employed, they have not been used to a significant degree in this country. Readers who have an interest in WAM-foam are referred to Koenders et al. (2002) and DâAngelo et al. (2008). For those desiring more information on the LEA approach to WMA, Romier et al. (2006), Harder et al. (2008), and DâAngelo et al. (2008) are recommended. G. Report Organization This report is composed of five chapters. Chapter 2 dis- cusses the approach used in the laboratory binder foam- ing and foamed mixture studies. The test methods used to characterize binder foaming are discussed in addition to the laboratory mixture approach, including the development of measures for workability and coatability and the proposed foamed mix design procedure. Chapter 3 describes the results of the laboratory binder and mixture studies, including the
13 effect of selected variables on binder foaming characteristics (expansion/collapse, bubble size, and shear behavior), a com- parison of laboratory foamers, and a description of the effects of foaming on workability and coatability of asphalt mix- tures. The validation of the proposed mix design approach using various foaming units is also included in this chapter. Chapter 4 presents the results of foaming at field sites in Texas and Ohio, including an initial trial to explore the binder foaming and foamed mixture sampling and testing. A field comparison of plant foaming units is also discussed, along with a field evaluation of the proposed mix design approach. Performance test results using the Hamburg Wheel Track Test (HWTT), resilient modulus (MR), and indirect tensile (IDT) strength are included for all mixtures. Conclusions stem- ming from the data generated in this study are reported in Chapter 5.
