Increasing the Service Life of Bridge Decks by Incorporating Phase Change Materials (PCMs) to Reduce Freeze/Thaw Cycles | NIST

During a freezing event, pore solution in cementitious bodies expands and creates stresses that can cause damage; therefore, reducing the number of freeze/thaw cycles experienced by a structure will extend the structure’s service life. The incorporation of phase change materials (PCMs) as a way to reduce the number of freeze/thaw cycles experienced by bridge decks has been investigated by modeling and mechanical testing, calorimetry, and x-ray microtomography. Models identified regions where freeze/thaw damage is not a significant concern, as well as regions where this technology may be practical, increasing the service life of a bridge deck by at least one year. The incorporation of PCM reduces strength by varying amounts, and for varying reasons, depending on which PCM is used and how it is introduced into the concrete. As a variety of methods exist to address this loss in strength, PCM incorporation is a promising technique for addressing one aspect of the impending infrastructure crisis in the United States. 1 – Introduction Although bridges account for roughly 40 % of national infrastructure maintenance costs (about $10 billion annually), in 2009 the American Society of Civil Engineers (ASCE) issued an overall grade of ‘C’, or ‘mediocre’, to the nation’s bridges [1, 2]. That same year, the American Association of State Highway and Transportation Officers (AASHTO) listed ‘age and deterioration’ first on a list of the ‘top five problems for bridges’ [3]. Over 150,000 bridges (roughly one in four) are either structurally deficient or functionally obsolete [4]. At an average price of $50/square foot, replacing the bridge decks of all structurally deficient or functionally obsolete bridges in the United States would cost nearly $49 billion, though the ASCE estimates that an annual $7 billion increase in funding every year would be required to “substantially improve bridge conditions” [4-6]. As infrastructure continues to age, the amount of money needed for maintenance is expected to rapidly increase: the average American bridge was constructed in 1966 and expected to last for 50 years [3]. Unless the nation is to face a similar crisis in the future, new technologies that extend the service life of bridge decks are needed. One of the most important factors affecting the service life of a bridge deck is the corrosion of reinforcing steel [7]. Exposure to freeze/thaw cycles can accelerate corrosion, as solutions in the pores of concrete expand during a freezing event and exert pressures that lead to the creation of cracks. These cracks provide aggressive media such as de-icing salts easy access to reinforcing steel [8]. The presence of water in cracks then further increases the amount of deterioration caused by successive freeze/thaw cycles. Although cracking can be caused by a variety of other mechanisms, and a variety of techniques have been developed to prevent or mitigate the effects of cracking, freeze/thaw damage still plays a significant role in limiting bridge deck service life [7]. The number of freeze/thaw cycles experienced by a bridge deck can be reduced by incorporating phase change materials (PCMs), materials with high enthalpies of phase change (H) [9-11]. At a given temperature (TMELT) PCMs solidify, releasing energy and maintaining surrounding temperatures at TMELT. If enough energy can be released in this way (dictated by H) a freezing event, and the associated damage, can be prevented [12, 13]. PCM can easily be introduced into a system by a variety of methods, most commonly as macroor micro-encapsulated ‘pellets’ [14, 15]. These pellets are, essentially, polymer spheres containing PCM and an appropriate amount of empty space for the accommodation of volumetric changes with temperature. Alternatively, lightweight aggregate (LWA) can be used; LWA is a porous aggregate that can absorb and hold liquids by capillary action [16]. These aggregates are commonly used for internal curing (supplying additional water to the cement paste so as to encourage more complete hydration), and are graded in such a manner that fine aggregate (i.e. sand) can simply be replaced by similarly sized LWA during mix design. This study investigates the use of PCM as a means by which to increase bridge deck durability by reducing freeze/thaw cycling. The effects of 12 different PCMs in 237 different locations were modeled using a combination of the CONCTEMP program developed at the National Institute of Standards and Technology (NIST) and a service life prediction model developed for the Indiana Department of Transportation (INDOT). Mortars containing PCM were produced to investigate the effects of PCM incorporation on mechanical properties, on the rate of hydration, and to investigate the 3D microstructure with x-ray microtomography. 2 – Methodology 2.1 – Experimental Four mortars were investigated: a control, a mix containing a paraffin wax (PCM6) in LWA, a mix containing a vegetable oil (PT4) in LWA, and a formula containing PT4 encapsulated in polymer. The composition of each formula can be found in Table 1. The cement used was a commercially available ASTM C150 Type II cement. The aggregate was a mixture of four different normal-weight sands which has previously been used in studies involving LWA [17]. The LWA used was a commercially available expanded clay corresponding to standard ASTM C330, “Specification for Lightweight Aggregate for Structural Concrete”. The mass ratio of water:cement was fixed at 0.35; the volume fraction of sand was 55 %. Tap water was used throughout. The quantity of LWA used was based on the LWA’s previously determined maximum absorption of 26.5 % by mass [17]. LWA was incorporated into the formulae simply by replacing normal-weight sand, ensuring that the gradation of aggregate particle sizes in all mortars was the same. Theoretically, the amount of LWA used could absorb up to 22 g more liquid than the amount of PCM used. Therefore, in the two mixes containing LWA-PCM, an additional 30 g of mix water was used to ensure that the LWA did not dehydrate the system. The LWA was saturated with PCM by immersion and agitation on a shaker-mixer for 1 hour. The amount of encapsulated PCM to be used was calculated with the knowledge that each pellet contains the mass ratio of PCM to encapsulation media of 4:1. Mortar preparation was performed in accordance with ASTM C109 with the exception that, when used, the PCM (in either LWA-PCM or encapsulated form) was added immediately after the aggregate. Specimens were placed in steel molds in plastic bags, placed in an environmental chamber, removed from the molds after 24 hours, and submerged in water in a temperature-controlled cabinet. Compression tests were performed at 3 d, 7 d, and 28 d on no fewer than 5 specimens. One cube specimen of each mortar was placed in a sealed plastic bag after failure in the compression test. The specimen was sawn into three equal sections using oil as the cutting lubricant, wiped dry, and placed in a plastic bag. The two outer sections of the specimen were used to determine thermal properties using Transient Plane Source (TPS) analysis, which has been described in [18, 19]. A 6.403 mm radius probe (Ni foil encased in Kapton) was sandwiched between two faces of the specimen (to minimize possible artifacts caused during the sawing, the outer surface faces were used in every experiment). Measurements were taken every 45 minutes after an equilibriation time of 45 minutes. Five measurements of thermal conductivity and volumetric heat capacity were recorded. Measurements were taken at either 22 °C or 23 °C with a power of 0.3 W applied over 10 s. The volumetric heat capacity, provided by the system, was then converted to specific heat capacity (mass basis) by dividing by average density (calculated by weighing and measuring cubes before compression tests). According to the manufacturer, thermal conductivity measurements on homogeneous materials are reproducible within ± 2 %, while heat capacity is reproducible within ± 7 %. After the dry measurements were made, sections were placed in a limewater bath for 7 days. The sections were removed, patted dry, and re-measured, with the entire apparatus placed inside a plastic bag to minimize drying over time. The middle section of the specimen was further sawn down to produce a small sample for analysis by x-ray microtomography. These tests were performed at a voltage of 200 kV and a tube current of 200 A, at an output resolution of 4000 pixels x 2096 pixels, a spatial resolution of about 5 m/pixel, scanned at 0.3° per step, 5 frames averaged, and a full 360 degree rotation. 2.2 – Modeling Bridge deck modeling was performed using the CONCTEMP program developed at NIST [20]. The model predicts time-of-wetness (due to both precipitation and condensation), surface temperature, and time-of-freezing for bridge decks based on a one-dimensional finite difference scheme that includes heat transfer by convection, conduction, and radiation. Radiation considers both solar radiation as a source and radiative cooling to the sky (greater at night) as a sink. A full description of CONCTEMP can be found in [20]. CONCTEMP requires the user to define several variables, including thermal and physical properties of both the concrete and PCM. The values for the concrete (heat capacity, thermal conductivity, and density) were set at 1000 J/kg·K, 1.5 W/m·K, and 2350 kg/m 3 , respectively. The latter two values are particularly conservative (see below). Of the 12 different PCMs investigated, two (PT4 and PCM6) are commercially available products 1 for which H and TMELT were supplied by the manufacturer. H and TMELT values used for the other 10 PCMs were based on a recent review by Cabeza et al. [9] (Table 2). Where a range was reported for a value, a simple average was used here. The 12 PCMs represent a relatively broad range of both H (153 J/g to 295 J/g), and TMELT (0 °C to 8 °C). The final user-defined