Annual Metallic Flows in Roof Runoff from Different Materials: Test

Environ. Sci. Technol. 2009, 43, 5612–5618

Annual Metallic Flows in Roof Runoff from Different Materials: Test-Bed Scale in Paris Conurbation

Downloaded by UNIV OF SOUTH DAKOTA on September 10, 2015 | http://pubs.acs.org Publication Date (Web): July 1, 2009 | doi: 10.1021/es9002108

P . R O B E R T - S A I N T E , * ,†,‡ M . C . G R O M A I R E , † B . D E G O U V E L L O , †,‡ M . S A A D , † A N D G . C H E B B O †,§ Universite´ Paris-Est, Laboratoire Eau, Environnement, Syste`mes Urbains, UMR-MA102 - AgroParisTech: 6-8 avenue Blaise Pascal, Cite´ Descartes, Champs-sur-Marne, 77455 Marne La Valleé Cedex 2, France, Centre Scientifique et Technique du Baˆtiment (CSTB Building Scientific Technical Centre), and Faculte´ de Geńie, Lebanese University

Received January 23, 2009. Revised manuscript received June 10, 2009. Accepted June 14, 2009.

A substantial database of annual metal runoff loads, obtained from a 14-month field exposure campaign on 12 different metal roofing materials at two sites within Paris conurbation, is presented herein. Thirteen metallic species have been considered. A comparison among the various roofing materials yields a ranking of their runoff pollution potential, which highlights that aluminum, coated products, and stainless steel display the lower emission levels, before zinc and copper materials. Lead materials appear to release more metallic species, and tend to do so in quite large quantities. Whatever the material family considered, older materials apparently release more metallic species with higher levels of emission. In considering zinc emissions from zinc-based materials, it is clear that surface coatings significantly reduce zinc emissions (40% less for the Zn3(PO4)2 surface-treated Anthra zinc, compared to natural zinc; and 99% less for prepainted galvanized steel compared to standard galvanized steel). In the case of Anthra zinc however, surface treatment induces the release of Ni into the runoff, and Ni constitutes a priority pollutant in the European Water Framework Directive (2000/60 CE). A high level of consistency with literature data has been found for Zn runoff when considering runoff rates reported at the same inclination.

Introduction Metal materials are widely used in urban areas, especially for infrastructure such as buildings or street furniture. Exposed to atmospheric conditions, which include environmental pollutants and relative humidity, these materials gradually corrode. During a rainfall event, a portion of the corrosion products is retained on the surface, while the other is released into runoff and washed away. Several research programmes conducted since the 1990s have revealed high trace metal concentration levels in runoff from metal roofs (1-4). * Correspondingauthorphone:+33(0)164153630;fax:+33(0)164153764; e-mail: [email protected]. † Universite´ Paris-Est, LEESU, UMR-MA102 - AgroParisTech. ‡ Centre Scientifique et Technique du Baˆtiment (CSTB Building Scientific Technical Centre). § Faculte´ de Geńie, Lebanese University. 5612

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 15, 2009

Experiments performed on a 42-ha catchment in Paris have established that the atmospheric corrosion of roofing materials could provide a major source of zinc, cadmium, lead, and copper in wet weather flow (5). In central Paris, roughly 40% of all roofing surfaces are covered with rolled zinc. Architectural rules concerning the protection of historical monuments and their environment prevent any changes for many years (6). In addition, lead is introduced to ensure the watertight integrity of the roof; also, copper and zinc are conventionally used for gutters and downspouts on nonmetallic roofs (e.g., tiles). The potential impact on urban runoff-receiving environments is important: several studies on the speciation of metals from roofing materials have been carried out for zinc and cooper showing that emissions located directly at the bottom of the downspout are mainly in their labile form (7, 8). Moreover, should these forms exhibit high bioavailability and thus be harmful to aquatic organisms, speciation is likely to vary significantly during the journey to receiving environments. In the context of the European Water Framework Directive (2000/60 CE) (9), the aim of which is to attain a healthy ecological state of aquatic environments, it would appear necessary to reduce pollutant production. These sources must be identified, and concerning metallic species in particular, a better quantification of roof emissions, with respect to the type of material used, is needed. While many studies have focused on the atmospheric corrosion of building materials, only a few have been specifically concerned with metal runoff processes. Literature reports some values of annual runoff rates, especially for zinc and copper, as measured during field exposure campaigns in Sweden and Switzerland (10, 11), yet data are lacking for other materials, e.g. lead. All studies were performed on small samples of materials and considered only the main metallic species (e.g., Zn for Zn-based materials). No study has been conducted within a strictly Paris-based context, and a transposition of results obtained from other contexts, even if different types of atmospheric conditions had been tested (12, 13), seems difficult to undertake. The TOITEAU project was launched in 2005 with the aim of developing a methodology for estimating annual metallic flows from roofs at the catchment area scale (14). The initial objective of this program was to evaluate annual runoff rates for various roofing materials, to establish a prioritization of the runoff contamination risk relative to the type of roof. The second objective focuses on transposing results recorded to larger spatial scales, through the use of roof surface area data obtained from aerial photographs. This paper aims to provide annual runoff rates for various roofing materials and a wide array of metallic species, within the Paris context. These data will be used to classify materials tested by taking into account their metal emissions in runoff.

Materials and Methods Exposure Sites. This experimental work has been based on exposing test beds to actual atmospheric and pluviometric conditions (from December 2006 to January 2008) on two exposure sites, which differ from the standpoint of local activities and land uses. The first test bed was installed on a flat roof on top of a building at the University of Cre´teil (the city of Cre´teil offers an urban site with many highways and industries, located approximately 8 km southeast of Paris), while the second bed was placed on a building at the CSTB facility (Building Scientific Technical Centre), in the town of Champs-sur-Marne (a suburban residential site located some 20 km east of Paris). Each of these two sites was equipped 10.1021/es9002108 CCC: $40.75

 2009 American Chemical Society

Published on Web 07/01/2009

TABLE 1. Exposed Material Samples material family


FIGURE 1. Photograph of the zinc experimental device used in Cre´teil. for both pluviometric readings and atmospheric composition in NO2, SO2, and O3 monitoring (passive samplers distributed by the IVL Swedish Environmental Research Institute). The background amounts of trace metal from rainwater and dry deposition were evaluated using Plexiglas panels as blanks. Exposed Material Samples. While previous runoff studies (11, 13) had replicated exposure conditions imposed by the standardized corrosion test (45°, south-facing), the exposure conditions applied in this work have been guided by the final objective of the TOITEAU project. As noted by Odnevall Wallinder et al. (15), orientation proves to be an important parameter at the test-bed scale. At the catchment scale however, this effect becomes smoothed as all types of roof orientations are combined. Panel slope was shown to be determinant for the runoff rate per m2 of exposed material yet does not significantly influence the runoff rate per m2 of projected area (15). To avoid the effect of orientation on test-bed results and in consideration of the fact that the aerial photographs used thereafter will only provide a projected view of roof surface areas, the test beds were slightly inclined. The 0.5 m2 rectangular panels (Figure 1) were exposed on frames inclined by 5° off the horizon facing southeast, in alignment with the exposed buildings. All runoff water was collected in 30-L polyethylene containers. The device design allowed for testing materials according to different uses: metal roofing panels (1250 × 400 mm), metal gutters (400 mm long) collecting water from a Plexiglas panel (1250 × 400 mm), and metallic tightness elements stacked onto a Plexiglas panel (the distribution between the two sites is given in Table 1). Five families of materials (zinc, copper, lead, aluminum, and steel) were chosen from a market research analysis focused on roofing materials used in the Paris context. Runoff Water Collection and Runoff Rate Evaluation. All runoff water was collected in containers, which were changed about once a month. Field exposure was performed for a period extending beyond 14 months, divided into 13 exposure intervals. Due to the speciation of metals stemming from roofs, mainly in labile form (7), complete sample mineralization proved unnecessary and the acid-soluble fraction was evaluated. Container contents were acidified to pH 1 with HNO3 to dissolve metal complexes possibly formed on the container walls. These contents were then stirred and a waiting time of 30 min was observed prior to sampling. Two 100-mL samples were collected for each test-bed. Samples were filtered on 0.45 µm cellulose acetate membranes to remove particles, and metal concentrations were determined using inductively coupled plasma-atomic emission spectroscopy (ICP-AES, Varian Vista MPX); in all, 13 metallic species were considered.

exposure site type

Cre´teil

Champs

panels

gutters x x x

zinc

new olda Anthrab

x x x

x x x

x x x

steel

stainlessc galvanizedd prepainted, galvanizede

x x x

x x

x x x

lead

new oldf

x x

x x

copper

new oldg

x x

x x

aluminum

standard prepaintedh

x

x x

x x x

x x

a Approximately 40 years old, previously exposed in Paris (15th arrondissement for panels and 1st arrondissement for gutters). b Surface treatment with Zn3(PO4)2, thickness ) 4.7 ( 0.3 µm. c Stainless steel 304. d Metal coating: 70-100 µm Zn. e Metal coating: 20 µm Zn + surface treatment (5-µm primer and 20-µm topcoat, both polyester-based). f Approximately 50 years old, previously used as tightness element on a historical monument in the Paris area. g Approximately 25 years old, previously exposed in Paris. h Approximately 80-µm topcoat.

Atmospheric inputs, evaluated on Plexiglas panels, were directly deduced from concentration levels before the actual calculation of metal flows. The runoff rates for each sampling period were calculated from concentration values by considering both precipitation amount and exposure time. For a period denoted i, the runoff rates of each element Fbed,i were evaluated as indicated in the following equation, in which Vi (in L) stands for the collected volume of runoff water during exposure period Ti (in years), Smat,bed (in m2) is the test-bed surface area, and [element]bed,i and [element]ref,i are the measured concentrations (in g · L-1) of runoff water for the considered test bed and Plexiglas reference panel, respectively. Fbed,i(element) is expressed in g · m-2yr-1. Fbed,i(element) ) Vi ([element]bed,i - [element]ref,i) Smat,bed × Ti

(1)

All results are given at the annual time scale, obtained by compiling data from the 13 time intervals. Fbed,annual(element) is also expressed in g · m-2yr-1. j)13

Fbed,annual(element) )

∑ (F i)1

j)13 bed,i(element)

× Ti)/

∑T

i

i)1

(2) Uncertainties on Both Concentrations and Annual Runoff Rates. The uncertainties on concentrations are due to both sampling and analysis steps. To evaluate an overall uncertainty on concentrations for each metallic species and each concentration range, repeatability tests were conducted (n ) 10) on 4 different samples. The uncertainties on annual runoff rates were evaluated using the Monte Carlo method, which consists of both drawing random errors (from 500 runs) on concentrations using standard deviations obtained and calculating annual flows by incorporating these errors. Uncertainties on flows are then determined by analyzing the variability of calculated values. VOL. 43, NO. 15, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5613

TABLE 2. Annual Runoff Rates Measured on Panels (in mg · m-2 · yr-1; First Row Cre´ teil; Second Row Champs), with Atmospheric Inputs (Top Two Rows) Being Subtracteda panel code atm. new zinc old zinc Anthra-zinc galv. steel


prepainted galv. steel stainless steel (Cre´teil only) aluminum (Cre´teil only) new lead old lead a

Al

Ba

Cd

Co

Cr

Cu

Fe

Li

Mn

Ni

26.0 19.4 4.6 4.6 35.5 12.0 15.0 4.3 16.9 4.7 9.0

2.6 1.7 0.4 0.1 3.7 1.3 0.7 0.1 0.8 0.1 0.8

0.1 0.05 2.0 1.5 -

0.3 0.4 0.1 0.1 -

0.2 0.2 0.06 0.04 0.03 0.07 -

4.4 2.9 1.6 0.9 2.0

34.7 21.7 13.4 8.9 35.4 14.0 19.0 7.3 21.3 3.8 11.9

0.1 0.1 0.1 0.1 -

2.9 2.6 0.2 1.3 0.6 0.4 0.5 0.1 0.5

0.4 0.3 6.0 5.5 -

3.6 5.6

0.2 0.5

-

-

0.07

0.4 0.6

4.3 7.4

-

0.1 -

0.1

-

67.7

0.3

-

-

0.02

0.4

11.9

-

0.2

-

35.4 3.7 99.7 15.6

2.3 7.4 7.8

6.0 6.2

1.0 0.5

-

6.9 4.0 11.9 14.5

63.4 12.3 152 26.6

0.1 0.1

3.4 15.9 12.0

1.1 -

Sr

Ti

Zn

2.1 1.3 0.5 0.3 19.1 15.8 0.7 0.6 0.3 0.4 0.4

3.1 4.1 3.9 1.8 0.2

0.8 0.6 0.2 0.1 1.6 0.5 0.6 0.2 0.6 0.1 0.3

34.2 57.0 3868 3299 4517 4145 2295 2105 1966 1867 24.6

4.7

-

0.1 0.2

7.4 -

2.1

-

0.2

33.7

2.5 42.4 13.4

1.4 3.7 0.3

52.4 76.9 145.2 87.4

7233 9924 24517 14593

Metallic species appearing in alloys composition have been listed in bold characters.

The concentration levels of metallic species in our samples have varied according to the specific species under consideration. For each element, one or two concentration level ranges have been investigated, depending on the various levels found in our samples (Table S1 in the Supporting Information (SI)). Relative uncertainties on concentrations lie above 20% for low ranges of Cd and Cu, and these become lower for other elements: between 10% and 20% for Al, Co, Cr, Ti and low ranges of Fe, Ni, Pb, and Zn; less than 10% for Ba, Li, Mn, Sr and high ranges of Fe and Ni; and, last, less than 1% for high ranges of Cu, Pb, and Zn. The main source of this error stems from the sampling and filtration stages. When calculating the annual flows emitted by metal panels, the errors committed in concentrations tend to offset one another, with coefficients of variation being generally less than those found during the concentration uncertainty evaluation. For metallic species present in very low quantities (e.g., Cr, less than 1.5 µg · L-1), uncertainty specific to the annual runoff rate is relatively high, i.e. with a coefficient of variation exceeding 20%. The results presented for those metallic species present in trace amounts within our samples (Cr, Co, Ti, and Cu for low-level concentrations) must be treated with care.

Results and Discussion Environmental Parameters. Annual rainfall values (as calculated from values measured between November 2006 and January 2008) equal 638 and 656 mm at the Cre´teil and Champs sites, respectively. The SO2 levels remain low, respectively 3 and 2 µg · m-3 in Cre´teil and Champs. This pollutant has been constantly decreasing in the Paris metropolitan area for the last 15 years. In 1994, the SO2 level was recorded at 14 µg · m-3, but by 2007 it had dropped to just 4 µg · m-3 (16). Moreover, the other atmospheric pollutant levels are posted at: 34.8 and 25.1 µg · m-3 for NO2 and 36.1 and 39.7 µg · m-3 for O3, as reported for Cre´teil and Champs, respectively. These levels are fairly consistent with values measured in central Paris: 38 µg · m-3 for NO2 and 33 µg · m-3 for O3. As expected, the amounts of atmospheric metal deposition (Table 2) are on the whole higher in Cre´teil than in Champs, by a factor of between 1.1 and 1.7. Three metallic species (Al, 5614

Pb

9


Fe, Zn) present atmospheric deposits higher than 25 mg · m-2 · yr-1, in relation with their high presence in environment. Atmospheric zinc deposits, however, are 1.7 times higher in Champs than in Cre´teil, a result possibly explained by the fact that the experimental device support in Champs is a duckboard made of galvanized steel, which leads to a local atmospheric zinc contamination. Concerning Al and Fe, levels measured are higher in Cre´teil than in Champs, in relation with the flat gravel roof in Cre´teil which might have trapped some dusts rich in Al and Fe. These dusts can then have been placed back into suspension by air turbulence, leading to an overestimation of these kinds of deposits. For other elements, levels of atmospheric deposits are lower than 5 mg · m-2 · yr-1, which is in many cases considerably lower than metal emissions from materials, even though levels involved can sometimes be categorized in the same order of magnitude. Metal Emission Profiles of the Tested Materials. The exposure of various roofing materials over more than 1 year has enabled both constituting a substantial database of runoff rates for many metallic species used under different conditions (panels and gutters) and comparing materials, whether belonging to the same metal family or not. This step serves to highlight differences in emission profiles from one material to the next and from one exposure site to the other. Comparison Among Materials. Tables 2 and 3 provide the complete quantitative results (according to material use) of annual runoff rate values measured at both exposure sites, with atmospheric inputs being subtracted. Metallic species appearing in alloys composition have been listed in bold characters. The metal emission profiles differ depending on the material considered; this variation concerns both number of metallic species and quantities emitted. Aluminum and Stainless Steel Roofing Materials. Metal flows from aluminum and stainless steel panels (Table 2) and gutters (Table 3) are very low, reflecting the good corrosion resistance characteristics of these materials. Nonetheless, Al runoff rates from aluminum panels are still higher here than some results published in the literature: Faller et al. (4) reported values of less than 10 mg · m-2yr-1. On the other hand, Odnevall Wallinder et al. (17) recorded emission

TABLE 3. Annual Runoff Rates Measured on Gutters (In mg · lm-1 · yr-1; First Row Cre´ teil; Second Row Champs), with Atmospheric Inputs (First Row) Being Subtracted (lm = Linear Meter)a gutter code new zinc old zinc Anthra-zinc new copper old copper aluminum (Champs only) prepainted aluminum (Champs only)


a

Al

Ba

Cd

Co

Cr

Cu

Fe

Li

Mn

Ni

Pb

Sr

Ti

Zn

43.9 2.6 28.4 8.7 18.9 1.1 16.9 3.1 26.5 8.6 6.9

2.2 0.1 2.1 1.1 1.1 0.1 1.2 0.2 1.8 0.4 0.2

0.1 0.1 -

0.1

0.03 0.08 -

2.3 0.6 3.3 2.3 1.4 416 323 631 429 0.4

44.9 3.6 42.3 17.8 23.9 2.0 28.6 6.2 43.7 33.1 1.7

0.1 0.1 -

2.5 0.3 2.4 1.4 1.1 0.3 1.0 0.4 1.8 0.8 -

0.9 0.6 -

0.9 0.2 2.4 1.9 0.3 0.1 1.4 0.2 0.9 0.4 0.2

1.7 0.5 1.3 2.4 0.6 0.4 0.4 0.4 0.8 0.6 0.4

1.5 0.1 1.1 0.5 0.6 0.1 0.6 0.1 0.8 0.2 -

871 685 805 798 499 330 20.2 21.4 20.6 35.7

5.3

0.2

-

-

-

1.6

8.8

-

0.9

-

-

1.6

-

11.2

Metallic species appearing in alloys composition have been listed in bold characters.

TABLE 4. Comparison of Zinc Runoff Rates with Data from the Literature (Values Are Expressed Considering the Projected Area) corrected runoff rates (g · m-2 · yr-1) new zinc old zinc Anthra zinc galvanized steel prepainted galvanized steel

Cre´teil Champs Stockholm this study this study (22) 3.9 4.5 2.3 1.9 0.02

3.3 4.2 2.1 1.9 0.01

3.3 4.9 1.8 3.8 0.1

levels of Cr and Ni from stainless steel to be roughly three times higher than those measured in this study. Lead Roofing Materials. Two lead materials were tested as tightness elements. Their emission spectra (Table 2 and Figure S1 in the Supporting Information) are wide, especially for the older material: 9 species were found in runoff water from the new lead (Pb . Zn > Fe, Al > Cu > Mn, Sr, Ba, Ti), and 13 species in the runoff from the old lead (Pb . Zn > Fe, Al > Sr, Cu, Mn > Ba, Cd > Ti, Ni, Co > Li). This finding may be due to the fact that lead used in roofing mainly originates from recycled material and that purification steps are not sufficient to eliminate all species present in initial alloys. Old lead runoff rates are higher than the new lead rates for all metallic species, an outcome most likely correlated with the higher quantity of corrosion products available on the old lead sheet, along with the accumulation of atmospheric deposits in patina during previous exposure and the initial alloy quality. Differences in runoff rates are significant with Pb (14-24 g · m-2yr-1 for the old material vs 7-10 g · m-2yr-1 for the new), as well as with Cd (6 mg · m-2yr-1 for the old material, no Cd emissions for the new) and Cu (12-14 mg · m-2yr-1 for the old material vs 4-7 mg · m-2yr-1 for the new). In literature, data on lead runoff rates are scarce and fragmented: Schultze-Rettmer (18) reported annual Pb runoff rates lying between 1 and 4 g · m-2yr-1, as calculated from a theoretical corrosion rate and Matthes et al. (19) reported values between 2.9 and 4.1 g · m-2yr-1. Runoff rates evaluated in this study are higher in both cases. Copper Materials. Both new and old copper materials have been tested on gutters; the runoff rates derived are expressed in grams per linear meter of gutter per year (g · lm-1yr-1). The emission spectra of the two materials considered are very similar (Table 3), yielding the emission profile: Cu . Fe, Zn, Al > Pb, Ba, Mn > Sr, Ti > Cr.

Copper emissions are higher for the older material (34% in Cre´teil and 25% in Champs). This observation is probably due to the presence of more soluble corrosion products on the surface of material, which is correlated with the corrosion layer formed during previous exposure. The emission levels observed for copper lie between 0.32 and 0.63 g · lm-1yr-1. Data reported in the literature focus on materials used as panels, with the runoff rates evaluated for an urban or industrial environment equaling between 0.3 and 3.9 g · m-2yr-1 (10, 11, 15, 20, 21). In introducing the exposed gutter surface area, the runoff rates are calculated to be 2.2 and 1.7 g · m-2yr-1 for the newer material and 3.3 and 2.3 g · m-2yr-1 for the older at the Cre´teil and Champs sites, respectively. The results obtained in this study are indeed consistent with literature. Zinc Roofing Materials. Five zinc-based materials have been tested: new zinc, old zinc, new Anthra zinc, galvanized steel, and prepainted galvanized steel. The differences in emission profiles among these materials are quite significant (Tables 2 and 3, and Figure S2). Metal emissions from the prepainted galvanized steel are very low, as a result of surface coating. Zn emissions from this material lie between 0.7 × 10-2 and 2.4 × 10-2 g · m-2yr-1, as compared to 1.8-4.5 g · m-2yr-1 for other zinc-based materials. In comparing the standard emission profiles for both new zinc (Zn . Fe > Al > Pb, Ba, Ti, Mn) and Anthra zinc (Zn . Fe, Al g Ni > Pb, Ba, Ti, Mn > Cr), it appears that prepainted materials tend to release more species. The Zn3(PO4)2 surface treatment on Anthra zinc reduces zinc runoff rates by 40% compared to standard zinc panels. The levels of aluminum and iron however are higher for the Anthra zinc, which likely relates to the difference in surface affinity between the material and these elements (atmospheric depositions might be more readily attached to the Anthra zinc panel, thus leading to an overestimation of emissions). The main difference noted is the significant amount of nickel emissions (6.0 mg · m-2yr-1) from Anthra zinc. This element, included on the list of priority pollutants in the European Water Framework Directive, gets introduced during the coating process. The emission profile of galvanized steel (Table 2) is similar to that of new zinc; the main difference concerns the level of Zn emissions: these are 45% lower on galvanized steel panels than on zinc panels. Aluminum and iron levels tend however to be higher for the galvanized material (iron is VOL. 43, NO. 15, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5615

TABLE 5. Threshold Values Considered for the Classification of Materials and Annual Average Concentrations (in µg · L-1) of Cadmium, Lead, Nickel, Copper, and Zinc Stemming from the Various Materials Exposed at the Cre´ teil Site

thresholds values


panels

gutters

ref/material

Cd

Pb

Ni

Cu

Zn

drinking water SEQ-Eau EQS values Dutch MPC U.S.EPA new zinc old zinc Anthra-zinc galvanized steel prepainted galv. steel stainless steel aluminum new lead old lead new zinc old zinc Anthra-zinc new copper old copper aluminum prepainted aluminum

5 0.04 0.09 0.42 0.25

50 5.2 7.2 11 2.5 0.5 30.2 1.1 0.3 0.5 0.4 3.5 11353 38439 0.6 3.3 0.4 0.4 0.9 0.4

50 6.2 20 5.1 52

1000 1 8.2 1.5 9

9.6

2.2 0.1

5000 4.3 7.8 9.4 120 6064 7080 3597 3081 31 39 37 278 375 1044 1217 503

3.2

9.9 0.2

rank MCIa

8: aluminum

1

7: stainless steel 6: prepainted galvanized steel 5: galvanized steel 4: Anthra-zinc 2: new zinc 3: old zinc 9: new lead 10: old lead

2 3

a

4 5 6 7 8 9

gutters

9 17: prepainted aluminum 10 16: aluminum 10 13: Anthra-zinc 717 838 1410 1735 2285 7717

11: 12: 14: 15:

new zinc old zinc new copper old copper

rank MCIa 1

3

2 3

8 119

4 5 6 7

243 292 493 658

MCI calculated with reference to SEQ-Eau thresholds.

present in steel alloys; the higher level of aluminum can be explained in the same way as for Anthra zinc runoff). Old zinc releases more species (Zn . Fe, Al g Pb > Sr, Ba, Cd > Cu, Ti, Mn > Li, Co > Cr) than new zinc material. Zn runoff rates amount to between 14% and 20% higher for the older material. Given the significant age difference between the two materials (40 years) in addition to the cumulative previous exposure of the older material, the difference observed in terms of zinc quantity released remains quite moderate, albeit noticeable for Pb (15.8-19.1 mg · m-2yr-1 for the old zinc vs 0.3-0.5 mg · m-2yr-1 for the new), Cu (0.9-1.6 mg · m-2yr-1 for the old zinc, no Cu emissions for the new) and Cd (1.5-2 mg · m-2yr-1 for the old zinc, no Cd emissions for the new). This result can be explained by both initial quality of materials and presence of elements fixed in the patina during previous exposure. Cadmium released by older zinc is clearly due to the poorer initial alloy quality. Lead detected in older zinc runoff may result from the fixation of atmospheric lead, which has been present in large quantities over the past decades due to leaded gasoline use. Research conducted in Stockholm (22, 23) has produced zinc runoff rate data for similar materials. The atmospheric conditions of exposure were quite comparable with those of this study (Table 3): the SO2, NO2, and O3 levels measured were, respectively, 3, 50, and 60 µg · m-3, with 540 mm of rainfall quantity reported. In Stockholm, samples were inclined 45° from the horizon, compared with just 5° in this 5616

9

4.4

1.3

2.9 0.6 0.2 37 39 0.5 3.4 1.4 493 653 2.4

TABLE 6. Hierarchy of Materials Tested with Respect to Their Runoff Emissions panels

0.3


19 32.8 2.5

work. To avoid the bias with respect to this inclination difference, all runoff rates initially evaluated by considering material surface area (runoff rate Smat) were expressed per projected surface area (runoff rate Sproj), with the calculation being performed using the inclination θ. Data from literature were transformed by means of the following relation and then reported in Table 4: Runoff rate Sproj )

Runoff rate Smat cos(θ)

(3)

The runoff rates per projected area appear to be comparable for all sites and materials, except results specific to galvanized steels: levels described in literature are higher than those recorded during this study. Comparison between the Two Test Sites. Emissions from materials measured at Cre´teil are higher than those at Champs, especially for higher emission levels (e.g., Zn for zinc materials). This difference equals roughly 8-15% for zinc materials, 22-32% for copper, and 40% for older lead material. Even though levels of atmospheric contaminants (NO2, SO2, and O3) are quite similar at the two sites (Table 3), this variability in results may still be explained by the difference in atmospheric corrosiveness, especially in terms of acidity and atmospheric deposits. Atmospheric corrosion would seem to be faster at Cre´teil, as highlighted by the higher runoff rates measured on this site. Both aluminum and iron emissions are also higher at Cre´teil, and this is so for all materials (Table 2). The phenomenon involved can be explained by different ways: the overestimation of atmospheric deposits previously mentioned, coupled with some differences of interactions between material surfaces and atmospheric dusts from one material to another (some materials retain much more dusts in patina layer than that measured on the reference Plexiglas panel), leading to an uncertainty on atmospheric inputs evaluation and, consequently, on material emissions as well. This explanation is also correct for some species present in trace amounts (e.g., Sr, Ti). Material Hierarchy Based on Emission Levels. Materials are placed in hierarchical order according to the emissions of five metallic species, three of which (Cd, Ni, and Pb) fall within the European Water Framework Directive (9), while the other two (Zn and Cu) have been considered due to their high concentration levels in runoff waters of certain materials.


Five kinds of reference values have been chosen: (1) threshold values in drinking water (French Decree No. 2001-1220, adopted on December 20, 2001); (2) values (green level) from the SEQ-Eau (March 21, 2003), established in France by the River Basin Agencies (24) to evaluate river water quality; (3) values of environmental quality standards or EQS (adopted on June 17, 2008) (25) for priority metal pollutants cited in the Framework Directive (values for Cu and Zn presented by Crane et al. (2007) (26) have been added); (4) and (5) the maximum permissible concentrations allowed for freshwater by both Dutch regulation and U.S.EPA (27, 28). The threshold values corresponding to each of these standards are listed in Table 5, which also reports annual average concentrations of cadmium, lead, nickel, copper, and zinc stemming from the various materials tested. These concentrations have been calculated from annual flows of each metallic species, by taking into account annual rainfall amounts as well. These values correspond to the immediate release situation. Materials classification step has been performed separately for both panels and gutters. The classification method consists of calculating a “Metal Contamination Index” (MCI), which provides an indication of the emission level of each material. This MCI is calculated, with respect to the chosen threshold values (i.e., drinking water, SEQ-Eau, EQS, Dutch MPC, U.S.EPA), as the linear sum for all five considered metals of the ratios of annual average measured concentration to the reference value. MCI )

[Pb] [Ni] [Cu] [Zn] [Cd] + + + + [Cd]ref [Pb]ref [Ni]ref [Cu]ref [Zn]ref (4)

The lowest MCI values correspond to those materials displaying the lowest metal emission levels. This method allows incorporating the type of metallic species emitted, their relative toxicity, and the corresponding level of emissions, via threshold values determined according to various environmental regulations. Values obtained are influenced by high concentration levels of certain elements (e.g., Zn for zinc materials). MCI values calculated using the different reference patterns lead, on the whole, to the same material classification when considering materials used for both panels and gutters. These classifications are given in Table 6, along with corresponding MCI values calculated with the reference patterns SEQ-Eau (24). For panels, materials can be divided into four groups, distinguished by increasing emission levels. The first group contains aluminum, coated galvanized steel, and stainless steel, which constitute the most neutral materials with regard to metal emissions. The second group is composed of Anthra zinc and galvanized steel, and the third contains new zinc, old zinc, and new lead. Their MCI values increase across the groups, in correlation with both level of Zn emissions for zinc-based materials and level of Pb emissions for new lead. The fourth and last group contains only old lead material. For gutters, just three groups have been formed: the first encompasses aluminum products, which are all highly ranked, while the second group contains zinc materials, and the last group contains copper. To draw conclusions on the environmental impact of metal materials, other parameters would need to be taken into consideration, including not only the dilution factor between roof runoff and receiving waters, but also the speciation of metallic species once in the receiving body, which depends on metal emissions as well as receiving body characteristics. An analysis of the impact on both the environment and aquatic species has not been undertaken as part of this work program.

All data generated during the field exposure described in this paper will be used as input data into a predictive model, which is currently under development and will run image classification software for assessing flows of roofing metals from an aerial photograph. The calibration step will be performed using data measured at the actual roof scale, which is an intermediate scale between the test bed and the catchment area.

Acknowledgments We gratefully acknowledge the Seine-Normandy Water Agency and the Ecole Nationale des Ponts et Chausseés Engineering School for their financial support; thanks are also addressed to the CSTB Building Scientific Technical Centre for financial as well as technical assistance.

Supporting Information Available Details on the uncertainty values for concentrations and annual runoff rates, plus the annual emission profiles for lead and zinc materials. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Fo¨rster, J. Heavy metal and ion pollution patterns in roof runoff; 7th ICUSD; Hannover, Germany, 1996. (2) Gromaire-Mertz, M.-C.; Garnaud, S.; Gonzalez, A.; Chebbo, G. Characterisation of urban runoff pollution in Paris. Water Sci. Technol. 1999, 41 (2), 1–8. (3) Chang, M.; McBroom, M. W.; Beasley, R. S. Roofing as a source of nonpoint water pollution. J. Environ. Manage. 2004, 73, 307– 315. (4) Faller, M.; Reiss, D. Runoff behaviour of metallic materials used for roofs and facades - a 5-year field exposure study in Switzerland. Mater. Corros. 2005, 56 (4), 244–249. (5) Gromaire-Mertz, M.-C.; Garnaud, S.; Saad, M.; Chebbo, G. Contribution of different sources to the pollution of wet weather flows in combined sewers. Water Res. 2001, 35 (2), 521–533. (6) Gromaire, M.-C.; Chebbo, G.; Constant, A. Incidence of zinc roofing on urban runoff pollutant loads. The case of Paris. Water Sci. Technol. 2002, 45 (7), 113–122. (7) Heijerick, D. G.; Janssen, J. R.; Karlen, C.; Odnevall Wallinder, I.; Leygraf, C. Bioavailability of zinc in runoff water from roofing materials. Chemosphere 2002, 47, 1073–1080. (8) Karlen, C.; Odnevall Wallinder, I.; Heijerick, D.; Leygraf, C. Runoff rates, chemical speciation and bioavailability of copper released from naturally patinated copper. Environ. Pollut. 2002, 120, 691–700. (9) Parlement Europeén. Directive Cadre2000/60/CE sur l’Eau; 2000; Journal Officiel L327 du 22/12/2000: 0001-0073. (10) He, W.; Odnevall Wallinder, I.; Leygraf, C. A comparison between corrosion rates and runoff rates from new and aged copper and zinc as roofing material. Water, Air, Soil Pollut. Focus 1\ 2001, 67–82. (11) Leuenberger-Minger, A. U.; Faller, M.; Richner, P. Runoff of copper and zinc caused by atmospheric corrosion. Mater. Corros. 2002, 53, 157–164. (12) Oesch, S.; Faller, M. Environmental effects on materials: the effect of air pollutants SO2, NO2, NO and O3 on the corrosion of copper, zinc and aluminium. A short literature survey and results of laboratory exposures. Corros. Sci. 1997, 39 (9), 1505– 1530. (13) Odnevall Wallinder, I.; Verbiest, P.; He, W.; Leygraf, C. The influence of patina age and pollutant levels on the runoff rate of zinc from roofing materials. Corros. Sci. 1998, 40 (11), 1977– 1982. (14) Robert-Sainte, P. Contribution des mate´riaux de couvertures a` la contamination me´tallique des eaux de ruissellement. PhD Thesis, Univ. Paris-Est., 2009. (15) Odnevall Wallinder, I.; Verbiest, P.; He, W.; Leygraf, C. Effects of exposure direction and inclination on the runoff rates of zinc and copper roofs. Corros. Sci. 2000, 42, 1471–1487. (16) Airparif. Bilan de la qualite´ de l’air en Ile-de-France en 2007.2008; http://www.airparif.asso.fr/airparif/pdf/bilan_2007.pdf. (17) Odnevall Wallinder, I.; Leygraf, C. Environmental effects of metal induced by atmospheric corrosion. In Outdoor and Indoor Atmospheric Corrosion, ASTM STP 1421; Townsend, H. E., ed.; 2002. VOL. 43, NO. 15, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5617


(18) Schultze-Rettmer, R. Lead roofing and rainwater. Du ¨ sseldorf. A scientific study commissioned by Bleiberatung; 1995. (19) Matthes, S. A.; Cramer, S. D.; Covino, B. S.; Bullard, S. J.; Holcomb, G. R. Precipitation runoff from lead. In Outdoor and Indoor Atmospheric Corrosion, ASTM STP 1421; Townsend, H. E., ed.; 2002. (20) Jouen, S.; Hannoyer, B.; Barbier, A.; Kasperek, J.; Jean, M. A comparison of runoff rates between Cu, Ni, Sn and Zn in the first steps of exposition in a french industrial atmosphere. Mater. Chem. Phys. 2004, 85, 73–80. (21) Zhang, X.; He, W.; Odnevall Wallinder, I.; Pan, J.; Leygraf, C. Determination of instantaneous corrosion rates and runoff rates of copper from naturally patinated copper during continuous rain events. Corros. Sci. 2002, 44, 2131–2151. (22) Odnevall Wallinder, I.; Leygraf, C.; Karlen, C.; Heijerick, D.; Janssen, C. R. Atmospheric corrosion of zinc-based materials: runoff rates, chemical speciation and ecotoxicity effects. Corros. Sci. 2001, 43, 809–816. (23) Bertling, S.; Odnevall Wallinder, I.; Leygraf, C.; Berggren Kleja, D. Occurrence and fate of corrosion-induced zinc in runoff water from external structures. STOTEN 2006, 367, 908–923.

5618

9


(24) SEQ-Eau. Syste`me d’Evaluation de la Qualite´ des cours d’Eau, Version 2; Agences de l’Eau, Ministe`re de l’Environnement et du De´veloppement Durable, 2003; 106 p. (25) Parlement Europeén; 2008; www.europarl.europa.eu/RegData/ seance_pleniere/textes_consolides/2006/0129/EP-PE_TC2COD(2006)0129_FR.doc. (26) Crane, M.; Kwok, K. W.; Wells, C.; Whitehouse, P.; Lui, C. S. Use of field data to support European Water Directive Quality Standards for dissolved metals. Environ. Sci. Technol. 2007, 41 (14), 5014–5021. (27) National Institute of Public Health and the Environment. Maximum permissible concentrations and negligible concentrations for metals, taking background concentrations into account; The Netherlands, 1997; 260 pp. (28) U.S. Environmental Protection Agency. National Recommended Water Quality Criteria; Office of Water: Washington, DC, 2006; 25 pp.

ES9002108

Annual Metallic Flows in Roof Runoff from Different Materials: Test

Recommend Documents