Article pubs.acs.org/est
Zinc Leaching from Tire Crumb Rubber Emily P. Rhodes,† Zhiyong Ren, and David C. Mays* Department of Civil Engineering, University of Colorado Denver, Campus Box 113, PO Box 173364, Denver, Colorado, United States S Supporting Information *
ABSTRACT: Because tires contain approximately 1−2% zinc by weight, zinc leaching is an environmental concern associated with civil engineering applications of tire crumb rubber. An assessment of zinc leaching data from 14 studies in the published literature indicates that increasing zinc leaching is associated with lower pH and longer leaching times, but the data display a wide range of zinc concentrations, and do not address the effect of crumb rubber size or the dynamics of zinc leaching during flow through porous crumb rubber. The present study was undertaken to investigate the effect of crumb rubber size using the synthetic precipitation leaching procedure (SPLP), the effect of exposure time using quiescent batch leaching tests, and the dynamics of zinc leaching using column tests. Results indicate that zinc leaching from tire crumb rubber increases with smaller crumb rubber and longer exposure time. Results from SPLP and quiescent batch leaching tests are interpreted with a single-parameter leaching model that predicts a constant rate of zinc leaching up to 96 h. Breakthrough curves from column tests displayed an initial pulse of elevated zinc concentration (∼3 mg/L) before settling down to a steady-state value (∼0.2 mg/L), and were modeled with the software package HYDRUS-1D. Washing crumb rubber reduces this initial pulse but does not change the steady-state value. No leaching experiment significantly reduced the reservoir of zinc in the crumb rubber. significant of which is zinc.9−15 Zinc is added to tires during the vulcanization process and represents approximately 1−2% of tires by weight.16−19 At elevated concentrations, zinc has been shown to cause a range of reproductive, developmental, behavioral, and toxic responses in a variety of aquatic organisms.20 Under the Clean Water Act, the EPA’s freshwater quality criteria specify a maximum zinc concentration (both acute and chronic) of 0.12 mg/L for protection of aquatic life assuming hardness of 100 mg/L.21 Similarly, the EPA’s stormwater discharge benchmark value for zinc is set to 0.117 mg/L for multisector general permits for industrial activities.22 Accordingly, recent research has focused on the transport of zinc through soils23−25 and the importance of zinc in urban water quality.26,27 Wik and Dave28 provide a review of the ecotoxicological effects of tire wear particles in the environment. Their review focuses on the toxicity of tire leachate using various leaching procedures and test organisms. Other studies also research the effects of tire leachate on differing aquatic organisms.10,12−14,29−34 These studies indicate a link between tire-derived zinc and toxicity, so the primary focus in the present study will be zinc leaching. Evidence in the literature suggests that zinc leaching from tire crumb rubber depends on zinc oxide levels in tires, particle size
1. INTRODUCTION According to the U.S. Environmental Protection Agency (EPA), 290 million tires are disposed annually in the United States,1 nearly one tire per person per year. Although the waste tire stockpile has been reduced by over 87% since 1990, the Rubber Manufacturers Association reports that the states of Alabama, Arizona, Colorado, Massachusetts, Michigan, New York, and Texas have over 85% of the remaining waste tire stockpile.2 Landfilling tires is problematic because tires tend to rise to the surface, which can harm landfill covers. Stockpiled tires can create breeding grounds for mosquitoes and other pests, and although tires are not categorized as hazardous material, waste tire facilities that have caught fire have been categorized as Superfund sites. These tire fires generate toxic emissions that damage both human and environmental health.1 To avoid the problems associated with waste tire disposal, tire crumb rubber is used in various applications including infill for turf fields, mulches, crumb rubber modified asphalt and other civil engineering applications, as well as molded and extruded products.3 Tire crumb rubber is produced when tires are ground and the fiber and steel belts are removed. Tire crumb rubber has a granular texture and ranges in size from very fine powder to 1 cm pieces.4 Tire crumb rubber has a density between 1.13 and 1.16 kg/L,5 and 4−5 kg of tire crumb rubber can be derived from one passenger tire.6 Applications utilizing tire crumb rubber are steadily increasing; however, crumb rubber can have negative impacts on air quality7 and tire leachate contains several chemicals of concern,8 the most © 2012 American Chemical Society
Received: Revised: Accepted: Published: 12856
February 9, 2012 September 14, 2012 November 12, 2012 November 12, 2012 dx.doi.org/10.1021/es3024379 | Environ. Sci. Technol. 2012, 46, 12856−12863
Environmental Science & Technology
Article
distribution, and surface area of rubber granulates.35 As pH decreases, zinc leaching increases.12,17 Zinc leaching also increases from rubber with greater surface area per mass,28 although for small particles, increasing the solid-to-liquid ratio can decrease zinc leaching due to aggregation of tire wear particles that decreases the effective surface area.12,36,37 As salinity increases, zinc leaching decreases17 and tire leachate becomes less toxic.32 Decreasing zinc leaching rates with time have been observed (in order of decreasing size) from tires,10,31 tire shreds,18 tire crumb rubber,19 and tire wear particles,17 and the mortality of aquatic organisms has been observed to decrease after sequential leaching periods.32,36 In contrast, a more aggressive test caused approximately 50% of the zinc in the tire crumb rubber to be leached during a HNO3−-H2O2 digestion test using microwave heating followed by centrifuge separation.37 Increased leaching during more aggressive leaching tests were also noted when comparing the EPA’s synthetic precipitation leaching procedure (SPLP), which specifies 18 h of end-over-end agitation at lowered pH, to a column leaching test in which 30 cm of simulated rainfall passed through a 5.1 cm column of the same crumb rubber; in this case the SPLP caused over 16 times more leaching than the column test.38 A new comparison of tire zinc leaching data from the literature (Table 1) highlights a large degree of variability. Accordingly, the present study reports (1) a statistical meta-analysis of the zinc leaching data in Table 1, (2) new leaching experiments, and (3) interpretation of results using a kinetic leaching model. Specifically, the SPLP was performed to measure how zinc leaching depends on crumb rubber size, quiescent batch leaching tests were performed to measure how zinc leaching depends on time, and column leaching tests (with unwashed and washed crumb rubber) were performed to measure the dynamics of zinc leaching. To our knowledge, this is the first study to synthesize such a broad collection of zinc leaching data from the literature, to report zinc leaching as a function of crumb rubber size, time, and flow dynamics, and to interpret results with a kinetic leaching model.
0.033 mg/L of background zinc. Zinc concentrations were measured by the private contract laboratory Colorado Analytical (Brighton, CO) or by the analytical laboratory at Denver’s Metro Wastewater Reclamation District (Denver, CO). Both laboratories measured zinc by inductively coupled plasma mass spectrometry (ICP-MS) following EPA Method 200.8 Rev. 5.4.41 To investigate effect of crumb rubber size, the SPLP was performed on each sample in accordance with EPA Standard Method 1312.41 In brief, unwashed tire crumb rubber was leached for 18 h in an end-overend agitator with an extraction fluid selected in accordance with the waste type and location. The extraction fluid used in the present study had pH 5, representative of locations west of the Mississippi River. The SPLP was selected because its low pH and fixed duration provided results in a short time frame, and using unwashed crumb rubber provided an upper bound on the expected degree of leaching. To prepare for the SPLP, the tire crumb rubber was sieved into four different sizes with diameters of 0.42, 0.84, 1.67, and 3.35 mm, where each size is the geometric mean of the opening space on the passing and retaining sieves. Tests were performed in duplicate. To investigate the effect of time, the required long-duration tests were conducted in quiescent water, which better simulates environmental leaching compared to the end-over-end SPLP. These quiescent batch leaching tests used washed crumb rubber in order to isolate leaching effects from the crumb rubber itself, rather than from any associated tire wear particles. 600 mL beakers were filled at ambient temperature with 500 mL tap water (1st replicate at pH 6.3) or deionized water (2nd replicate at pH 6.7), to which 25 g of crumb rubber was added. Deionized water used in this study was derived from tap water, and therefore had less than 0.033 mg/L of background zinc. The six beakers were not stirred, and samples were taken after 6, 18, 24, 48, and 96 h. To investigate the effect of flow dynamics, column leaching tests were performed using an acrylic column 152 cm tall with an inner diameter of 10.2 cm. In each experiment, the column was dry-packed with crumb rubber to a depth of 51.4 cm. Drypacking allowed a stainless steel wire mesh to be placed atop the crumb rubber, which was required to prevent floating. To ensure saturation, the column was then filled with tap water from the bottom at a rate of 50 mL/min, which required 46.9 min. Once the water level reached 2.5 cm above the crumb rubber, the column was filled with tap water from the top at a rate of 540 mL/ min. Once the water reached an overflow valve placed to impose 91.4 cm of ponding above the crumb rubber, release valves at the base were opened to allow the water to flow through the column at a steady rate of 540 mL/min confirmed by a flow meter. Column leaching tests 1 and 2 were conducted with unwashed crumb rubber, and column leaching test 3 was conducted with washed crumb rubber. The pH was 6.4 in column test 1, 6.0 in column test 2, and 6.5 in column test 3.
2. MATERIALS AND METHODS 2.1. Statistical Methods. For each study in Table 1, a linear model regressing the available variables (e.g., pH) on zinc concentration was performed. For each model, a one-sided test was used to determine overall significance of the association of a selected variable on zinc concentration (i.e., a nonzero slope for the variable in the linear model). Then, p-values from multiple studies were combined in a meta-analysis to obtain an overall pvalue for the selected variable using the Stouffer test (ref 39, page 45, footnote 15). All calculations were implemented in the statistical software package R 2.7.1.40 2.2. Experimental Methods. While it is recognized that the chemical composition of waste tires composed primarily of auto tires will vary from that composed primarily of truck tires, for this study a representative sample was used to study a typical auto− truck waste tire mix from a temperate climate in the continental U.S. The crumb rubber used in this study was provided by the now-defunct tire recycling company AcuGreen (Commerce City, CO), who donated 54 kg of tire crumb rubber, from which subsamples were randomly removed. The crumb rubber had a d60 of 1.8 mm (viz., 60% of the material by weight was smaller than 1.8 mm) and a coefficient of uniformity Cu = d60/d10 = 3.3 (Figure S1 in the Supporting Information). When washed, tire crumb rubber was rinsed with tap water between no. 16 and no. 200 sieves, and then air-dried. Tap water used in this study had
3. RESULTS 3.1. Statistical Results. Statistical meta-analysis of the data in Table 1 identifies three variables with significant association to zinc concentration, based on linear associations for individual studies: pH has a negative association with zinc concentration (p = 2.8 × 10−5), based on significant negative associations (p ≤ 0.05) in two studies37,42 and nonsignificant negative association in one study.12 The solid-to-liquid ratio (i.e., mass of crumb rubber per volume of water) has a positive association with zinc 12857
dx.doi.org/10.1021/es3024379 | Environ. Sci. Technol. 2012, 46, 12856−12863
Environmental Science & Technology
Article
Table 1. Zinc Leaching from Tire Crumb Rubber in 14 Studies in the Literature
study Bocca
11
California 46 Davis 47 Gualtieri 12
T [oC]
pH
d50 [μm]
solid liquid ratio [g/L]
25 ↓ 37 20b 20b ↓
5 n/a 2.3 4.3 3 ↓
3500a ↓ n/a n/a 45a ↓
62.5 ↓ 200 1 50 ↓
100 ↓
23 ± 2 ↓
31
Hartwell
37
Kanematsu
10 ↓ 25 ↓ 40 ↓
Mattina
48
Nelson 10 New York
Park
38
20b ↓ 20b ↓
49
Ronchak
42
San Miguel
23 ± 2 ↓ 20b 23 ± 2 ↓
50
Stephensen 51 Wik 36
20 20b 20b 20 ± 2 ↓
4 5 6 7 4.9 ↓ 5 7 9 5 7 9 5 7 9 4.2 7.0 8.36 4.2 ↓
6.9 ↓ 3 ↓ 5 ↓ 8 ↓ 6.2 7 7.69 7.5 ↓
50 ↓
10000 ↓
50 ↓
590 ↓
50 ↓
3000 ↓ n/a n/a ↓
50000 ↓ 25400 ↓
420c ↓ n/a n/a ↓
50 ↓ 181 50 ↓
n/a ↓ 428.9 437.4 469.0 469.0 250.0 250.0 50 0.1 100 0.1 ↓
t [hr] 24 ↓ 21 24 24 48 72 96 240 24 48 72 96 240 24 ↓
168 336 504 72 ↓
18 ↓ 744 column column 18 18 0.017 0.10 24 ↓
48 24 n/a 120 ↓ 168 ↓ 216 ↓ 264
C [mg/L]
study
T [oC]
pH
d50 [μm]
solid liquid ratio [g/L]
t [hr] ↓
2.55 0.966 17 3.4 44.73 14.47 13.3 8.49 28.57 35.28 13.02 12.37 9.26 20.15 10.5 6.5 5.2 1.2 0.026 0.021 0.034 12.483 4.177 2.544 18.93 5.597 2.544 27.839 3.263 2.082 2.8 1.05 0.755 0.291 0.2141 1.947 1.15 1.3 0.14 18.6 23.5 8.53 3.80 0.005 0.005 0.578 0.55 2.29 0.07 0.08 0.07 0.47 0.4 0.09 0.25 0.16 0.17 0.12
480 ↓ 1 ↓
120 ↓
168 ↓ 216 ↓ 264 ↓ 480 ↓ 10 ↓
120 ↓
168 ↓ 216 ↓ 264 ↓ 480 ↓
C [mg/L] 0.1 0.08 0.41 0.38 0.4 1.14 0.13 0.44 0.09 0.5 0.1 0.54 0.32 0.16 1.49 0.39 0.42 0.18 0.12 0.16 1.44 0.47 0.66 1.46 1.75 2.24 0.52 2.34 0.71 2.57 1.08 1.6 2.56 2.28 3.66 2.16 0.82 1.44 4.46 2.92 5.01
a c
12858
Arithmetic average of given range. bAssumed ambient temperature. Maximum size given. dx.doi.org/10.1021/es3024379 | Environ. Sci. Technol. 2012, 46, 12856−12863
Environmental Science & Technology
Article
concentration (p = 2.2 × 10−3), based on significant positive association in one study 36 and nonsignificant negative association in two studies.12,42 Leaching time is positively associated with zinc concentration (p = 2.0 × 10−3), based on significant positive association in one study,36 nonsignificant positive association in two studies,31,38 and nonsignificant negative association in one study.12 No significant association was identified between zinc concentration and temperature over the temperature ranges studied. No study in Table 1 varied crumb rubber diameter. 3.2. Experimental Results. The SPLP test indicates that the amount of zinc leaching from the tire crumb rubber decreases with increasing tire particle radius (Figure 1). This is consistent
Figure 2. Quiescent batch leaching results showing zinc concentration increasing linearly with time. Error bars are one standard deviation calculated from duplicate measurements, and are smaller than the plotting symbol at t = 48 h. No replicate was taken at t = 24 h.
Figure 3. Column leaching results for unwashed crumb rubber (test 1 in white diamonds, and test 2 in gray squares) and washed crumb rubber (test 3 in black triangles), showing an initial spike in effluent concentration before steady-state leaching. The EPA freshwater criterion of 0.12 mg/L is shown as a dashed horizontal line, and the inset shows detail of first 60 min. The first sample in column test 3, at t = 0, was discarded due to contamination.
mg/L after 5.0 min. After 20 h, the zinc concentration fell to 0.115 mg/L. Column test 3 was conducted to explore the difference in leaching from washed and unwashed rubber. Compared to column leaching tests 1 and 2, the peak in column test 3 was much lower, 0.60 mg/L, suggesting that washing the tire crumb rubber can reduce the initial leaching concentration. However, washing the tire crumb rubber did not reduce leaching at late time: Zinc concentration in column test 3 was 0.149 mg/L after 24 h, similar to column tests 1 and 2.
Figure 1. Synthetic precipitation leaching procedure (SPLP) results: (a) Zinc concentration versus crumb rubber diameter; (b) mass transfer rate coefficient α in eq 6 versus crumb rubber surface area. Data labels indicate sieve size, for example, R-50 is the sample retained on a no. 50 sieve. Error bars are one standard deviation calculated from duplicate measurements, and are smaller than the plotting symbol for R-16.
with the fact that smaller tire crumb rubber has greater surface area per mass than larger rubber particles. Therefore, the smaller the tire particle, the greater the concentration of zinc leached. The highest concentration of zinc leached in the SPLP tests was 1.3 mg/L. The quiescent batch leaching tests show that zinc concentration increased linearly with time, reaching a maximum of 2.7 mg/L after 96 h (Figure 2). This linear increase with time indicates similar leaching kinetics throughout the quiescent batch leaching test, suggesting that zinc release was not limited by source depletion or by accumulation of zinc ions near the crumb rubber surface. In column test 1 (Figure 3), the maximum concentration of zinc leached was 2.63 mg/L after 5.0 min. After 2 h and 30 min, the zinc concentration approached the EPA’s freshwater criteria of 0.12 mg/L. In column test 2, the leaching again spiked at 2.55
4. MODELING Zinc leaching and transport are modeled with the advectiondispersion equation (ADE) with an additional term for kinetic mass transfer and a corresponding expression for the solid concentration: αρb ∂C ∂ 2C ∂C (KdC − S) =D 2 −v − ∂t ∂x θ ∂x
(1)
and ∂S = α (K d C − S ) ∂t
(2)
where t is time [min], x is position [cm], C is aqueous zinc concentration [μg/mL] = [mg/L], S is solid phase zinc concentration [μg/g] = [mg/kg], D is the dispersion coefficient 12859
dx.doi.org/10.1021/es3024379 | Environ. Sci. Technol. 2012, 46, 12856−12863
Environmental Science & Technology
Article
[cm2/min], v is pore water velocity [cm/min], θ is dimensionless water content, α is a mass transfer rate coefficient (min−1), ρb is the bulk density of the crumb rubber [g/mL] = [kg/L], and Kd is a linear equilibrium partitioning coefficient [mL/g] = [L/kg]. The pore water velocity is v = q/θ, where q is volumetric flux [cm/min]. These equations are simplified by assuming plug flow (D = 0), and that leaching is far from equilibrium (KdC ≪ S). With these assumptions, (1) and (2) simplify to
αρb S ∂C ∂C = −v ∂t θ ∂x
In the column leaching tests, zinc leaching is conceptualized as the superposition of two processes. This rationale for this conceptualization is discussed in Section 5. The first process describes steady-state zinc leaching at late time, implemented with eq 6, based on a large solid concentration S1 and a small mass transfer rate coefficient α1. The second process describes dynamic zinc leaching at early time, in which a small mass of quickly mobilized zinc m2 gives a small solid concentration S2 associated with a large mass transfer coefficient α2. The second process is implemented numerically with eqs 3 and 4. This conceptual model was applied separately for column leaching tests with unwashed crumb rubber (column tests 1 and 2) and washed crumb rubber (column test 3). The following are assumed to be constant in column tests 1−3: Bulk density ρb = 0.495 kg/L and water content 0.562 were determined by measuring the mass of crumb rubber required to fill the column of depth 51.4 cm. The volumetric flux q = 6.66 cm/min was determined from the known discharge of 540 mL/min, so one pore volume is flushed every 4.34 min. To model dynamic zinc leaching, eqs 3 and 4 were implemented using the software package HYDRUS-1D, which was chosen because it allows the initial spatial distributions of aqueous concentration Co(x) and of solid concentration So(x) to be specified independently. The spatial resolution Δx = 2.572 cm and variable time step Δt ≥ 0.2 min were chosen to comply with the Courant number limitation of Cre ≤ 1 given in the HYDRUS1D (3.0) user’s manual (ref 44, Section 8.4.5). The assumption of plug flow (D = 0) prevented compliance with the corresponding grid Peclet number limitation of Peg < 5, but sensitivity analysis showed that results were insensitive to D. Steady flow was modeled by arbitrarily setting zero pressure at the upper and lower boundaries (i.e., unit gradient flow), and then setting the hydraulic conductivity in HYDRUS-1D to 6.66 cm/s in order to match the actual volumetric flux. All other specifications were identical to Example 5 in the HYDRUS-1D (3.0) user’s manual. The HYDRUS-1D model for dynamic zinc leaching was implemented by choosing a mass of initially mobilized zinc m2, calculating the corresponding solid concentration So(x), which was assumed to be uniformly distributed along the column depth, and then using HYDRUS-1D’s parameter estimation utility to fit the mass transfer rate coefficient α2 to the difference between the effluent concentration and the steady effluent concentration, C(t) − C(t ≥ 12 h), where the steady effluent concentration was modeled separately using eq 6. For unwashed crumb rubber, concentrations used for model fitting were averaged at t = 0, 5, and 10 min, where column tests 1 and 2 recorded effluent concentrations at equivalent times. Fitting was optimized by performing a grid search on m2 (every 0.3 mg) to identify the value that minimized the root mean squared error (RMSE) of the final superposition of the steady and dynamic zinc leaching models. For unwashed crumb rubber with 1.6% zinc oxide, the steady effluent concentration, corresponding to water exposed to crumb rubber for 4.34 min, is C(t ≥ 12 h) = 0.1078 mg/L. According to eq 6 this corresponds to a steady mass transfer rate coefficient α1 = 2.0 × 10−6 min−1. Fitting the mass of initially mobilized zinc m2 = 11.1 mg and the corresponding α2 = 0.19 min−1 gave a minimum RMSE of 0.13 mg/L (Figure 4). For washed crumb rubber, the steady effluent concentration is C(t ≥ 12 h) = 0.1595 mg/L, corresponding to a steady mass transfer rate coefficient α1 = 3.3 × 10−6 min−1. Choosing m2 = 7.1 mg and fitting α2 = 0.031 min−1 gave a minimum RMSE of 0.056 mg/L (Figure 5).
(3)
and
∂S = − αS ∂t
(4)
Equation 3 is further simplified for quiescent batch leaching tests (v = 0). Furthermore, by assuming S ≈ constant, the same simplification (that v = 0) can be applied to the steady phase (t ≥ 12 h) of column leaching tests by adopting a moving coordinate frame in which v = 0. With these assumptions, eq 3 simplifies to
αρb S dC = dt θ
(5)
For initially zinc-free water, this gives a simple linear model for zinc leaching: C=
αρb S θ
t
(6)
This model will be used to interpret the zinc leaching results for the SPLP (Figure 1), the quiescent batch leaching tests (Figure 2), and the column leaching tests (Figure 3). In the SPLP, 100 g of unwashed crumb rubber was added to make a total volume of 2000 mL. Assuming crumb rubber solid density 1.13 kg/L,43 this gives bulk density ρb = 0.050 kg/L and water content θ = 0.956. Assuming crumb rubber is 1.6% zinc oxide by weight (Chris Madden, Michelin Americas Research Corporation, personal communication, 2010) gives an initial solid concentration of S = 12 775 mg/kg. Using the standard SPLP agitation time of 18 h, the mass transfer rate parameter α was then calculated for each of the concentration measurements in Figure 1a, and ranged from a minimum of 5.1 × 10−7 min−1 for the largest crumb rubber to a maximum of 1.8 × 10−6 min−1 for the smallest crumb rubber. These results are plotted versus crumb rubber surface area in Figure 1b, where surface area is calculated by assuming monodisperse spherical crumb rubber: A=
6m ρs d
(7) 2
where A is surface area [m ], m is the mass of crumb rubber [kg], ρs is solid density [kg/m3], and d is the geometric mean of the opening space of the passing and retaining sieves [m]. Figure 1b shows that the mass transfer rate parameter depends linearly on crumb rubber surface area. In the quiescent batch leaching tests, 25 g of washed crumb rubber was added to 500 mL of water. Assuming solid density of 1130 kg/m3 as above gives total volume of 522 mL, bulk density ρb = 0.048 kg/L, and water content θ = 0.958. Linear regression of the (t, C) data in Figure 2, where the regression line was forced to pass through the origin, gives a slope of dC/dt = 0.0281 (mg/ L)/h, from which the average mass transfer rate coefficient α = 7.3 × 10−7 min−1 was calculated from (5). Figure 2 shows that the mass transfer rate coefficient is constant over time. 12860
dx.doi.org/10.1021/es3024379 | Environ. Sci. Technol. 2012, 46, 12856−12863
Environmental Science & Technology
Article
concentration in batch samples increased as t1/2, consistent with diffusion-limited leaching. We speculate the difference in kinetics is a consequence of our experiments being far from equilibrium, (i.e., KdC ≪ S), such that the driving force for zinc leaching remains nearly constant. Mass transfer rate coefficients fitted in this study are summarized in Table 2. Excluding α2, these coefficients range Table 2. Summary of Fitted Mass Transfer Rate Coefficients (α) SPLP quiescent batch leaching tests column leaching tests, steady, α1 column leaching tests, initial, α2
Figure 4. Fitted model for column leaching with unwashed crumb rubber (test 1 in white diamonds, and test 2 in gray squares), with an initial pulse of 11.1 mg and a mass transfer rate coefficient α2 = 0.19 min−1, giving a root mean squared error (RMSE) of 0.13 mg/L. The inset shows detail of first 60 min.
unwashed
washed
1.2 × 10−6 min−1 not available 2.2 × 10−6 min−1 0.19 min−1
not available 7.3 × 10−7 min−1 3.3 × 10−6 min−1 0.031 min−1
from a minimum of 7.3 × 10−7 min−1 for quiescent batch leaching tests to a maximum of 3.3 × 10−6 min−1 for steady column effluent from washed crumb rubber, with a geometric mean of 1.5 × 10−6 min−1. These differ by a factor of 4.7, which is much less than the ratio of the maximum to the minimum zinc concentrations of 8940 reported in Table 1. One speculative interpretation of the results in Table 2 is that leaching in the quiescent batch samples was slower than in steady phase of the column studies because the diffusion length scale is larger in stagnant water than in flowing water. Similarly, one may speculate that leaching from washed crumb rubber in the steady phase of the column studies was faster because washing removed detritus from the crumb rubber surface. However, these speculative interpretations do not explain why the fitted leaching rate in the SPLP is within the range of other rates, rather than being the fastest, which might have been expected considering the low pH and agitation used in the SPLP. In this light, perhaps a better way to interpret Table 2 is that the range of zinc leaching rates is within a factor of 2 or 3 of the geometric mean of α = 1.5 × 10−6 min−1. The width of this range is consistent with the wellknown chemical heterogeneity of tire crumb rubber.45 In contrast to the SPLP, quiescent batch leaching, and steadystate column leaching results that are all consistent with the linear model for zinc leaching given in (6), the initial column leaching results required an empirically fitted additional mass of 11.1 mg for unwashed crumb rubber and 7.1 mg for washed crumb rubber. These results are analogous to the literature cited above that indicate decreased leaching with time.10,17−19 The initial masses required to fit the initial column leaching results are equivalent to 0.042% of the total mass of zinc in the unwashed crumb rubber and 0.027% of the total mass of zinc in the washed crumb rubber. For the case of washed crumb rubber, this additional mass can be explained in part as follows: To prepare for each column leaching test, and to ensure saturation with water, the crumb rubber was filled from below at 50 mL/min, which required 46.9 min. An additional 13 min were required to fill the top of the column before the leaching test began. Therefore, at the beginning of the column leaching test, the water at the top of the column had been exposed to crumb rubber for 59.9 min. Using the steady mass transfer rate coefficient of α = 3.3 × 10−6 min−1, this corresponds to an initial concentration of 2.2 mg/L, with correspondingly lower initial concentrations lower in the column. Using this initial concentration profile Co(x) as an initial condition in HYDRUS-1D rather than a uniformly distributed additional mass of 11.1 mg produces a breakthrough curve (Figure S2 in the Supporting Information)
Figure 5. Fitted model for column leaching with washed crumb rubber (test 3 in black triangles), with an initial pulse of 7.1 mg and a mass transfer rate coefficient α2 = 0.031 min−1, giving a root mean squared error (RMSE) of 0.055 mg/L. The inset shows detail of first 60 min.
5. DISCUSSION This study and the literature summarized in Table 1 indicate that pH, crumb rubber size, and leaching time all play important roles in zinc leaching from tire crumb rubber. pH was not investigated in this study, but linear regression of the data in Table 1 indicates that dC/d(pH) ranges from −4.337 to −6.3,42 confirming the well-known result that zinc leaching increases with decreasing pH. With regard to crumb rubber size, the results in Figure 1 indicate increased leaching from smaller crumb rubber, with a linear correlation between the mass transfer rate coefficient and the crumb rubber surface area. The quiescent batch results in Figure 2 and the steady-state results in Figure 3 are all consistent with the simple model of eq 6 that indicates linear zinc leaching with time. The importance of leaching time also reflected in the significant correlation between zinc leaching and time evident in the data of Wik et al.36 shown in Table 1. The leaching model of eq 6 depends on the assumption that the solid phase zinc concentration S is constant with time, which appears to be reasonable based on the linear leaching versus time up to 96 h in Figure 2, and the steady effluent concentrations up to 24 h in Figure 3. The assumption of constant S was evaluated with a hypothetical HYDRUS-1D model having So(x) = 12 775 mg/kg and α = 3.3 × 10−6 min−1 (for washed crumb rubber) which resulted in a final value of S = 12 710 mg/kg, a drop of only 0.5% after 24 h of continuous leaching. The linear kinetics reported here are in contrast to the previous study of Degaffe and Turner,17 who reported that zinc 12861
dx.doi.org/10.1021/es3024379 | Environ. Sci. Technol. 2012, 46, 12856−12863
Environmental Science & Technology
Article
that is consistent with the first measured effluent concentration of 0.60 mg/L at 5 min, but results in a larger RMSE of 0.10 mg/L, and requires a fitted value of DL = 276 cm, which is probably unrealistic considering that the column is only 51.4 cm tall. Similarly, for the experiments with unwashed crumb rubber, using α = 2.2 × 10−6 min−1 gives an initial concentration of 1.5 mg/L for the water at the top of the column, which is too low to explain the peak of 2.6 mg/L measured at 5 min. Based on these lines of evidence, it appears unlikely that the initial pulse of elevated zinc concentration is simply a consequence of additional leaching time during column filling. The additional mass reflected in the peak may have resulted from mobilization of colloidal tire wear particles28 during the first few pore volumes of the column leaching test. Such particles would have been absent in the experiment using washed crumb rubber. This explanation remains speculative, however, lacking confirmation by turbidity measurements or other means to confirm the presence of colloidal tire wear particles. It is apparent, however, that washing the tire crumb rubber reduces the zinc leaching, at least in the initial pulse. Comparison with a similar column leaching study by Lee18 highlights several unique aspects of the current study and emphasizes the importance of EPA’s freshwater quality criteria. Lee measured zinc concentration versus time at the effluent of a 20 cm column leached with deionized water at pH 4, reporting a rapid drop in zinc concentration up to 10 pore volumes. This rapid drop in zinc concentration was modeled with an analytical solution of the ADE, which assumed instantaneous equilibrium between the solid and liquid phases, and which predicts zero effluent concentration at late time. In contrast, the present study reports effluent concentrations up to 24 h, equivalent to approximately 330 pore volumes. After an initial pulse of elevated zinc concentrations, these long-term leaching experiments reveal steady-state leaching that require the kinetic leaching model described above. The fact that zinc concentration does not go to zero at long time emphasizes that compliance with the EPA’s freshwater quality criteria for zinc is more difficult than compliance with the EPA’s maximum contaminant level (MCL) of 5 mg/L for zinc. Although steady zinc leaching concentrations approached the EPA’s 0.12 mg/L freshwater criteria, it should be stated that this correspondence is likely coincidental. Had the time required to transmit a pore volume been greater than 4.34 min, presumably the steady zinc leaching concentration would have been higher. The time required to transmit a pore volume, in turn, depends on the hydraulic head gradient and on the hydraulic conductivity, the latter of which depends on crumb rubber size. The hydraulic conductivity also depends on the degree of physical, chemical, or biological clogging, any of which could reduce flow rates, increase detention times, and consequently increase the concentration of zinc leached into the environment.
■
ASSOCIATED CONTENT
■
AUTHOR INFORMATION
Present Address †
URS Corporation, 8181 East Tufts Avenue, Denver, CO 802372579. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We thank Jeffrey Gee and Phil Robinson for assistance with the laboratory experiments, Randy Ray for assistance with machining, Katerina Kechris for the statistical meta-analysis, AcuGreen for donating the tire crumb rubber used in this study, and three anonymous reviewers for constructive feedback. Funding for this work was provided by the Colorado Department of Public Health and Environment through their Advanced Technology Grant Program.
■
REFERENCES
(1) U.S. Environmental Protection Agency. Wastes, Resource Conservation, Common Wastes and Materials, Scrap Tires Website; 2010. http://www.epa.gov/wastes/conserve/materials/tires/faq.htm (accessed March 9, 2011). (2) Rubber Manufacturers Association. Scrap Tire Markets in the United States: 9th Biennial Report; Rubber Manufacturers Association: Washington DC, 2009. (3) Gugliemotti, A.; Lucignano, C.; Quadrini, F. Production of rubber parts by tyre recycling without using virgin materials. Plast., Rubber Compos. 2012, 41 (1), 40−46. (4) Texas Natural Resource Conservation Commission. The Many Uses of Crumb Rubber; Waste Tire Recycling Program, Office of Permitting: Austin, TX, 1999. (5) Tang, Z. J.; Butkus, M. A.; Xie, Y. F. F. Crumb rubber filtration: A potential technology for ballast water treatment. Mar. Environ. Res. 2006, 61 (4), 410−423. (6) Pierce, C. E.; Blackwell, M. C. Potential of scrap tire rubber as lightweight aggregate in flowable fill. Waste Manage. 2003, 23 (3), 197− 208. (7) Simcox, N. J.; Bracker, A.; Ginsberg, G.; Toal, B.; Golembiewski, B.; Kurland, T.; Hedman, C. Synthetic Turf Field Investigation in Connecticut. J. Toxicol. Environ. Health, Part A 2011, 74 (17), 1133− 1149. (8) Li, X. L.; Berger, W.; Musante, C.; Mattina, M. I. Characterization of substances released from crumb rubber material used on artificial turf fields. Chemosphere 2010, 80 (3), 279−285. (9) Gualtieri, M.; Rigamonti, L.; Galeotti, V.; Camatini, M. Toxicity of tire debris extracts on human lung cell line A549. Toxicol. In Vitro 2005, 19, 1001−1008. (10) Nelson, S. M.; Mueller, G.; Hemphill, D. C. Identification of tire leachate toxicants and a risk assessment of water-quality effects using tire reefs in canals. Bull. Environ. Contam. Toxicol. 1994, 52 (4), 574−581. (11) Bocca, B.; Forte, G.; Petrucci, F.; Costantini, S.; Izzo, P. Metals contained and leached from rubber granulates used in synthetic turf areas. Sci. Total Environ. 2009, 407 (7), 2183−2190. (12) Gualtieri, M.; Andrioletti, M.; Vismara, C.; Milani, M.; Camatini, M. Toxicity of tire debris leachates. Environ. Int. 2005, 31 (5), 723−730. (13) Wik, A.; Dave, G. Acute toxicity of leachates of tire wear material to Daphnia magnaVariability and toxic components. Chemosphere 2006, 64 (10), 1777−1784. (14) Wik, A. Toxic components leaching from tire rubber. B. Environ. Contam. Tox. 2007, 79 (1), 114−119. (15) Mantecca, P.; Gualtieri, M.; Andrioletti, M.; Bacchetta, R.; Vismara, C.; Vailati, G.; Camatini, M. Tire debris organic extract affects Xenopus development. Environ. Int. 2007, 33 (5), 642−648. (16) Gunasekara, A. S.; Donovan, J. A.; Xing, B. S. Ground discarded tires remove naphthalene, toluene, and mercury from water. Chemosphere 2000, 41 (8), 1155−1160.
* Supporting Information S
The Supporting Information provides two figures showing the grain size distribution for the tire crumb rubber used, and results from an alternative HYDRUS-1D model. This material is available free of charge via the Internet at http://pubs.acs.org. Corresponding Author
*Phone: +1-303-352-3933; fax: +1-303-556-2368; e-mail: david.
[email protected]. 12862
dx.doi.org/10.1021/es3024379 | Environ. Sci. Technol. 2012, 46, 12856−12863
Environmental Science & Technology
Article
(17) Degaffe, F. S.; Turner, A. Leaching of zinc from tire wear particles under simulated estuarine conditions. Chemosphere 2011, 85 (5), 738− 743. (18) Lee, T. Leaching characteristics of bottom ash from coal fired electric generating plants, and waste tire; individually and mixtures when used as construction site fill materials. Waste Manage. 2011, 31 (2), 246−252. (19) Taheri, S.; Khoshgoftarmanesh, A. H.; Shariatmadari, H.; Chaney, R. L. Kinetics of zinc release from ground tire rubber and rubber ash in a calcareous soil as alternatives to Zn fertilizers. Plant Soil 2011, 341 (1− 2), 89−97. (20) Councell, T. B.; Duckenfield, K. U.; Landa, E. R.; Callender, E. Tire-wear particles as a source of zinc to the environment. Environ. Sci. Technol. 2004, 38 (15), 4206−4214. (21) U.S. Environmental Protection Agency. National Recommended Water Quality Criteria 2011. http://www.epa.gov/waterscience/ criteria/wqctable (accessed September 10, 2012). (22) U.S. Environmental Protection Agency. Technical Correction: Final National Pollutant Discharge Elimination System Storm Water Multi-Sector General Permit for Industrial Activities. Fed. Regist. 1996 61 Federal Register 5248, Document Number 96-2722. (23) Khaokaew, S.; Landrot, G.; Chaney, R. L.; Pandya, K.; Sparks, D. L. Speciation and release kinetics of zinc in contaminated paddy soils. Environ. Sci. Technol. 2012, 46 (7), 3957−3963. (24) Lin, Y. T.; Weng, C. H.; Lee, S. Y. Spatial distribution of heavy metals in contaminated agricultural soils exemplified by Cr, Cu, and Zn. J. Environ. Eng. 2012, 138 (3), 299−306. (25) Wiatrowska, K.; Komisarek, J.; Kozlowski, M. Efficiency assessment of various amendments in the immobilization process of Cu, Pb and Zn in artificially contaminated soil. Fresenius Environ. Bull. 2011, 20 (9), 2193−2202. (26) Ingvertsen, S. T.; Jensen, M. B.; Magid, J. A minimum data set of water quality parameters to assess and compare treatment efficiency of stormwater facilities. J. Environ. Qual. 2011, 40 (5), 1488−1502. (27) Roy, J. W.; Bickerton, G. Toxic groundwater contaminants: An overlooked contributor to urban stream syndrome? Environ. Sci. Technol. 2012, 46 (2), 729−736. (28) Wik, A.; Dave, G. Occurrence and effects of tire wear particles in the environmentA critical review and an initial risk assessment. Environ. Pollut. 2009, 157 (1), 1−11. (29) Wik, A.; Dave, G. Environmental labeling of car tiresToxicity to Daphnia magna can be used as a screening method. Chemosphere 2005, 58 (5), 645−651. (30) Day, K. E.; Holtze, K. E.; Metcalfesmith, J. L.; Bishop, C. T.; Dutka, B. J. Toxicity of leachate from automobile tires to aquatic biota. Chemosphere 1993, 27 (4), 665−675. (31) Hartwell, S. I.; Jordahl, D. M.; Dawson, C. E. O.; Ives, A. S. Toxicity of scrap tire leachates in estuarine salinities: Are tires acceptable for artificial reefs? Trans. Am. Fish. Soc. 1998, 127 (5), 796−806. (32) Hartwell, S. I.; Jordahl, D. M.; Dawson, C. E. O. The effect of salinity on tire leachate toxicity. Water, Air, Soil Pollut. 2000, 121 (1−4), 119−131. (33) Birkholz, D. A.; Belton, K. L.; Guidotti, T. L. Toxicological evaluation for the hazard assessment of tire crumb for use in public playgrounds. J. Air Waste Manage. 2003, 53 (7), 903−907. (34) Sheehan, P. J.; Warmerdam, J. M.; Ogle, S.; Humphrey, D. N.; Patenaude, S. M. Evaluating the risk to aquatic ecosystems posed by leachate from tire shred fill in roads using toxicity tests, toxicity identification evaluations, and groundwater modeling. Environ. Toxicol. Chem. 2006, 25 (2), 400−411. (35) Pysklo, L.; Pawlowski, P.; Parasiewicz, W.; Piaskiewicz, M. Influence of the zinc oxide level in rubber compounds on the amount of zinc leaching. Kautsch. Gummi Kunstst. 2006, 59 (6), 328−331. (36) Wik, A.; Nilsson, E.; Kallqvist, T.; Tobiesen, A.; Dave, G. Toxicity assessment of sequential leachates of tire powder using a battery of toxicity tests and toxicity identification evaluations. Chemosphere 2009, 77 (7), 922−927.
(37) Kanematsu, M.; Hayashi, A.; Denison, M. S.; Young, T. M. Characterization and potential environmental risks of leachate from shredded rubber mulches. Chemosphere 2009, 76 (7), 952−958. (38) New York State Department of Environmental Conservation. An Assessment of Chemical Leaching, Releases to Air and Temperature at Crumb Rubber Infilled Synthetic Turf Fields; New York State Department of Health: Albany, NY, 2009. (39) Stouffer, S.; Suchman, E.; DeVinney, L.; Star, S.; Williams Jr., R. The American Soldier, Vol. 1, Adjustment During Army Life; Princeton University Press: Princeton, NJ, 1949. (40) R Development Core Team. R: A Language and Environment for Statistical Computing: Reference Index Version 2.7.1; Foundation for Statistical Computing: Vienna, Austria, 2011. (41) Keith, L., Compilation of EPA’s Sampling and Analysis Methods, 2nd ed.; CRC Press: Boca Raton, FL, 1996. (42) Ronchak, A., Waste Tires in Sub-Grade Road Beds; Minnesota Pollution Control Agency: St. Paul, MN, 1990. (43) Tang, H. L.; Regan, J. M.; Clark, S. E.; Xie, Y. F. Prediction of clean-bed head loss in crumb rubber filters. J. Environ. Eng. 2011, 137 (1), 55−62. (44) Šimůnek, J.; van Genuchten, M.; Šejna, M. The HYDRUS-1D Software Package for Simulating the One-Dimensional Movement of Water, Heat, and Multiple Solutes in Variably-Saturated Media (Version 3.0); University of California Riverside: Riverside, CA, 2005. (45) Epps, J. Uses of Recycled Rubber Tires in Highways, Synthesis of Highway Practice 198, National Cooperative Highway Research Program, Transportation Research Board, National Research Council; National Academy Press: Washington, DC, 1994. (46) California Integrated Waste Management Board. Evaluation of Health Effects of Recycled Waste Tires in Playground and Track Products; Office of Environmental Health Hazard Assessment: Sacramento, CA, 2007. (47) Davis, A. P.; Shokouhian, M.; Ni, S. B. Loading estimates of lead, copper, cadmium, and zinc in urban runoff from specific sources. Chemosphere 2001, 44 (5), 997−1009. (48) Mattina, M.; Islyen, M.; Berger, W.; Ozdemir, S. Examination of Crumb Rubber Produced from Recycled Tires; Connecticut Agricultural Experiment Station: New Haven, CT, 2007. (49) Park, J. K.; Edil, T. B.; Kim, J. Y.; Huh, M.; Lee, S. H.; Lee, J. J. Suitability of shredded tyres as a substitute for a landfill leachate collection medium. Waste Manage. Res. 2003, 21 (3), 278−289. (50) San Miguel, G.; Fowler, G. D.; Sollars, C. The leaching of inorganic species from activated carbons produced from waste tyre rubber. Water Res. 2002, 36 (8), 1939−1946. (51) Stephensen, E.; Adolfsson-Erici, M.; Celander, M.; Hulander, M.; Parkkonen, J.; Hegelund, T.; Sturve, J.; Hasselberg, L.; Bengtsson, M.; Forlin, L. Biomarker responses and chemical analyses in fish indicate leakage of polycyclic aromatic hydrocarbons and other compounds from car tire rubber. Environ. Toxicol. Chem. 2003, 22 (12), 2926−2931.
12863
dx.doi.org/10.1021/es3024379 | Environ. Sci. Technol. 2012, 46, 12856−12863
