Results from a round robin test for the ecotoxicological evaluation of construction products using two leaching tests and an aquatic test battery

Accepted Manuscript Results from a round robin test for the ecotoxicological evaluation of construction products using two leaching tests and an aquatic test battery Stefan Gartiser, Ines Heisterkamp, Ute Schoknecht, Michael Burkhardt, Monika Ratte, Outi Ilvonen, Frank Brauer, Jan Brückmann, André Dabrunz, Philipp Egeler, AndreaMaria Eisl, Ute Feiler, Ines Fritz, Sabina König, Herbert Lebertz, Pascal Pandard, Gabriele Pötschke, Dirk Scheerbaum, Frank Schreiber, Přemysl Soldán, Roland Weiß, Reinhilde Weltens PII:

S0045-6535(17)30165-0

DOI:

10.1016/j.chemosphere.2017.01.146

Reference:

CHEM 18759

To appear in:

ECSN

Received Date: 18 September 2016 Revised Date:

21 December 2016

Accepted Date: 31 January 2017

Please cite this article as: Gartiser, S., Heisterkamp, I., Schoknecht, U., Burkhardt, M., Ratte, M., Ilvonen, O., Brauer, F., Brückmann, J., Dabrunz, A., Egeler, P., Eisl, A.-M., Feiler, U., Fritz, I., König, S., Lebertz, H., Pandard, P., Pötschke, G., Scheerbaum, D., Schreiber, F., Soldán, P., Weiß, R., Weltens, R., Results from a round robin test for the ecotoxicological evaluation of construction products using two leaching tests and an aquatic test battery, Chemosphere (2017), doi: 10.1016/ j.chemosphere.2017.01.146. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT CHEMOSPHERE

Results from a round robin test for the ecotoxicological evaluation of construction products using two leaching tests and an aquatic test battery

RI PT

Authors Stefan Gartiser1 *, Ines Heisterkamp1, Ute Schoknecht 2, Michael Burkhardt3, Monika Ratte4, Outi Ilvonen5

SC

Frank Brauer6, Jan Brückmann7, André Dabrunz8, Philipp Egeler9, Andrea-Maria Eisl10, Ute Feiler11, Ines Fritz12, Sabina König13, Herbert Lebertz14, Pascal Pandard15, Gabriele Pötschke16, Dirk Scheerbaum17, Frank Schreiber18, Přemysl Soldán19, Roland Weiß20, Reinhilde Weltens21 1

Hydrotox GmbH, Freiburg, Germany BAM Federal Institute for Materials Research and Testing, Berlin, Germany 3 HSR University of Applied Sciences Rapperswil, Institute for Environmental and Process Engineering (UMTEC), Rapperswil, Switzerland 4 ToxRat Solutions GmbH, Alsdorf, Germany 5 Federal Environment Agency, Dessau, Germany 6 Federal Environment Agency FG III 2.5, Berlin, Germany 7 Institut Dr. Nowak, Ottersberg, Germany 8 Eurofins Agroscience Services Ecotox GmbH, Niefern-Öschelbronn, Germany 9 ECT Oekotoxikologie GmbH, Flörsheim/Main, Germany 10 Lenzing AG - Safety, Health and Environment Department, Lenzing, Austria 11 Federal Institute of Hydrology, Koblenz, Germany 12 University of Natural Resources and Life Sciences Vienna - IFA-Tulln, Austria 13 WESSLING GmbH, Altenberge, Germany 14 SGS Institut Fresenius GmbH, Taunusstein, Germany 15 INERIS, Verneuil en Halatte, France 16 IDUS - Biologisch Analytisches Umweltlabor, Ottendorf-Okrilla, Germany 17 Dr. U.Noack-Laboratorien, Sarstedt, Germany 18 Niedersächsischer Landesbetrieb für Wasserwirtschaft, Küsten- und Naturschutz; Hildesheim, Germany 19 Masaryk Water Research Institute, Ostrava, Czech Republic 20 Hygiene-Institut des Ruhrgebiets - Institute for Environmental Hygiene and Toxicology, Gelsenkirchen, Germany 21 VITO Environmental Toxicology, Mol, Belgium

EP

TE D

M AN U

2

AC C

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42

ABSTRACT

43

A European round robin test according to ISO 5725-2 was conceptually prepared,

44

realised, and evaluated. The aim was to determine the inter-laboratory variability of

45

the overall process for the ecotoxicological characterization of construction products

46

in eluates and bioassays. To this end, two construction products BAM-G1 (granulate)

47

and HSR-2 (roof sealing sheet), both made of EPDM polymers (rubber), were

48

selected. The granular construction product was eluted in a one stage batch test, the

ACCEPTED MANUSCRIPT planar product in the Dynamic Surface Leaching test (DSLT). A total of 17

50

laboratories from 5 countries participated in the round robin test: Germany (12),

51

Austria (2), Belgium (1), Czech Republic (1) and France (1). A test battery of four

52

standardised ecotoxicity tests with algae, daphnia, luminescent bacteria and

53

zebrafish eggs was used. As toxicity measures, EC50 and LID values were

54

calculated. All tests, except the fish egg test, were basically able to demonstrate toxic

55

effects and the level of toxicity. The reproducibility of test results depended on the

56

test specimens and the test organisms. Generally, the variability of the EC50 or LID

57

values increased with the overall level of toxicity. For the very toxic BAM-G1 eluate a

58

relative high variability of CV = 73% to 110% was observed for EC50 in all biotests,

59

while for the less toxic HSR-2 eluate the reproducibility of EC50 varied with

60

sensitivity: it was very good (CV = 9.3%) for the daphnia test with the lowest

61

sensitivity, followed by the algae test (CV = 36.4%). The luminescent bacteria test,

62

being the most sensitive bioassay for HSR-2 Eluate, showed the highest variability

63

(CV = 74.8%). When considering the complex overall process the reproducibility of

64

bioassays with eluates from construction products was acceptable.

M AN U

SC

RI PT

49

65 Keywords

67

Round robin test, construction products, leaching tests, eluates, ecotoxicity tests

68

TE D

66

INTRODUCTION

70

Construction products that come into contact with rain, seepage water or

71

groundwater in their intended use may release hazardous substances through

72

leaching. The use of waste materials in construction has been recognized as a

73

relevant source of heavy metal pollution (Cenni et al., 2001; Flyhammar and Bendz,

74

2006; Galvin et al., 2012). Leaching methods have been standardised and

75

regulations drawn up to address the issue (Eikelboom et al., 2001; Hage and Mulder,

76

2004; Susset and Grathwohl, 2011; Nebel and Spanka, 2013). So far construction

77

products releasing mainly hazardous organic substances have received much less

78

attention (Burkhardt et al., 2011; Wangler et al., 2012; Baderna et al, 2015). The

79

identification of hazardous substances by chemical analysis may not cover all

80

contaminants present in leachates from construction products. With bioassays the

81

joint effects of ingredients are detected by their effects to living organisms. Bioassays

82

are especially suited to assess effects of organic substances for which reference or

AC C

EP

69

ACCEPTED MANUSCRIPT limit values for water quality often do not exist like for heavy metals. This has been

84

acknowledged in research work addressing waste and road runoff (Pandard et al.,

85

2006; Waara and Färm, 2008).

86

The variety of different construction products on the market is huge. For both the

87

manufacturers and the users of construction products it is important that reliable and

88

easy to use methods for the assessment of the products’ environmental assessment

89

and comparison of product performance are available. Thus, a test battery has been

90

elaborated for the ecotoxicological characterisation of eluates from construction

91

products (Gartiser et al. submitted). Such a harmonised test battery is intended to

92

facilitate the development of a common understanding for the assessment of

93

leaching from the building sector for both regulatory purposes and voluntary

94

initiatives by manufacturers or ecolabels.

95

By the combination of leaching tests with ecotoxicity tests the overall variability of the

96

ecotoxicity test results is expected to increase. However, the extent of this variability

97

was not known so far. Therefore, an interlaboratory round robin test with 17

98

laboratories and two construction products relevant for leaching of organic

99

contaminants in their intended use (one granulate and one sheet like) was carried out

100

in 2015, while using biotests with algae, daphnia, fish eggs, and luminescent

101

bacteria. The testing strategy is also in-line with the technical CEN guidance CEN/TR

102

17105 (draft) on the use of ecotoxicity tests applied to construction products currently

103

being developed by the European Committee for Standardization (CEN). The

104

International Organization for Standardization has published several biological

105

methods under the technical committee ISO/TC 147 for testing ecotoxicological

106

effects.

107

An adequate reproducibility of ecotoxicity data is a prerequisite for the validation of

108

the methodology before it can be used for regulatory purposes. Standardised

109

performance tests for construction products have been mandated by the European

110

Commission to facilitate the removal of technical barriers for trade. The objective of

111

the round robin test was to obtain quantitative figures about the reproducibility and

112

robustness of data obtained from ecotoxicity testing of eluates.

AC C

EP

TE D

M AN U

SC

RI PT

83

113 114

MATERIALS AND METHODS

115

Participating laboratories and organization of the ring test

ACCEPTED MANUSCRIPT A total of 17 laboratories from 5 countries participated in the round robin test:

117

Germany (12), Austria (2), Belgium (1), Czech Republic (1) and France (1). The

118

laboratories belong to governmental institutes, contract laboratories, research

119

institutes and one university (see attribution of authors to their institutes, laboratory

120

codes L01 – L18 are anonymised). Most laboratories maintain a quality assurance

121

system, although the studies themselves have not been subjected under these

122

systems. Because only data fulfilling all validity criteria of the different tests were

123

considered in the evaluation, a high quality of the data set is assured.

124

The round robin test was organised by the Hydrotox GmbH laboratory

125

(www.hydrotox.de) within a research project on the ecotoxicological characterisation

126

of construction products funded by the German Environment Agency. The boundary

127

conditions for the preparation of the eluates, the sample pre-treatment and the

128

implementation of bioassays were described in a detailed study plan. Participants

129

were provided with templates for data input and demonstration videos for performing

130

the leaching tests.

131

Construction products. The first product, “BAM-G1”, was an ethylene propylene

132

diene rubber (EPDM) granulate of 0.5 mm to 2 mm grain size used as subbase and

133

draining layer for artificial turf of sports grounds. The material was brand-new and not

134

made up from recycled tire rubber waste, which has been previously addressed in

135

leaching studies and whose eluates showed high ecotoxicity in the algae and

136

daphnia tests (Krüger et al. 2013). The second product, “HSR-2”, was also made up

137

by an elastomeric EPDM rubber sheet for roof sealing with the following dimensions:

138

25 cm x 18 cm x 1.45 mm. The product is root-resistant without the use of growth

139

preventing herbicides such as mecoprop. It is manufactured ready for use and does

140

not require lamination. Details of the characteristics and ingredients of both products

141

are given in the preceding study (Gartiser et al. submitted).

142

were sent between the end of August and early September 2015 by BAM and HSR

143

to the participating laboratories. The laboratories were requested to consider

144

additional control blanks without the products in both leaching tests which include all

145

experimental steps (i.e. dilution water, leaching and sampling vessels, pretreatment,

146

storage). These additional blanks from the DSLT and the one stage batch test were

147

examined in parallel in all ecotoxicity tests at a 1:2 dilution (D=2). Three laboratories

148

did not provide blank controls for all bioassays. Some laboratories found some

149

toxicity in the blank controls at D=2, but the LID could not be determined, because no

AC C

EP

TE D

M AN U

SC

RI PT

116

The test specimens

ACCEPTED MANUSCRIPT follow-up tests at higher dilutions were performed. The leaching tests and bioassays

151

were carried out between August 25th and November 24th, 2015.

152

Leaching tests. Each participating laboratory performed the elution process for the

153

specimens BAM-G1 and HSR-2 on its own (only exception: L10 tested parallel

154

eluates provided by the organising laboratory). The granulate BAM-G1 was eluted in

155

the one stage batch test according to DIN EN 12457-1 (2003). This method was

156

developed for the characterization of granular waste with particle sizes below 4 mm

157

and uses a liquid to solid ratio of 2 L kg-1 using an overhead shaker. The technical

158

specification CEN/TS 16637-1 provides guidance on the assessment of release of

159

dangerous substances from construction products. Here, the on stage batch test DIN

160

EN 12457-1 is referred to as an example of an indirect method, which may be easier

161

and/or cheaper to apply for a specific application. The study plan recommended

162

adding 250 g material to 500 mL deionized water with a conductivity of < 5 µS cm-1.

163

The elution took place over 24 h (+/- 0.5 h) in a darkened room using an overhead

164

shaker at a revolution speed of 7 +/- 1 rounds per minute at 20 °C +/- 5 °C. According

165

to the study plan, the phase separation after shaking should preferably be performed

166

by sieving through a 0.5 +/- 0.2 mm stainless steel or porcelain sieve and subsequent

167

centrifugation at 2000 g for 30 minutes. Alternatively, a pressure or vacuum filtration

168

could be performed with pre-rinsed glass-fiber filters (0.45 µm).

169

A pre-treatment of eluates in order to remove particulate matter before ecotoxicity

170

testing was only required for BAM-G1 eluate while the DSLT-eluate with HSR-2 was

171

clear. Not all laboratories followed the recommended pre-treatment described in the

172

study plan (sieve followed by centrifugation). Filtration of eluates and all possible

173

combinations of the different pre-treatment procedures have also been carried out

174

(see table 1).

175

The sheet like product “HSR-2” was eluted with the horizontal dynamic surface

176

leaching test (DSLT) according to CEN/TS 16637-2. It consists in a simple tank test

177

for assessing the release of dangerous substances from monolithic or plate-like

178

construction products of > 40 mm edge length and for plates with a surface area

179

exposed to the eluate of > 100 cm2. In the standard procedure this test is carried out

180

for 64 days, while the eluate water is replaced at distinct time intervals (after 6 h, 24

181

h, 2.25 d, 4 d, 9 d, 16 d, 36 d and 64 d). Two adjustments have been conducted for

182

the preparation of eluates for ecotoxicity testing. First, only the first two elution steps

183

after 6 h and additional 18 h have been carried out and both eluates were unified for

AC C

EP

TE D

M AN U

SC

RI PT

150

ACCEPTED MANUSCRIPT ecotoxicity testing. Second, the liquid / surface area relation (L/A) was set to 20 L m-2

185

(2 mL cm-2). This corresponds to the lower limit of the DSLT. For this, deionized

186

water with a conductivity of < 5 µS cm-1 and preferably all-glass aquaria (e.g. 30 cm x

187

20 cm x 12 cm) were used as elution medium and leaching vessel. The distance

188

between all exposed surfaces of the construction product to the glass wall should be

189

> 2 cm and the test specimen should be submerged by a 2 cm high water column.

190

Because the last requirement could not always be fulfilled at this L/A-ratio a water

191

column of < 2 cm was also considered as being acceptable. The aquaria were

192

covered with glass plates and kept in the dark at 19 - 25 °C. According to the study

193

plan after 6 h +/- 15 minutes the 1st water exchange was realised and the 1st eluate

194

stored overnight in a refrigerator (ca. 4 °C). The 2nd water sampling was done after

195

18 h +/- 15 minutes. Both eluates, representing a total elution time of 24 h, were joint

196

and mixed thoroughly. No filtration or other pre-treatment of the eluates was

197

envisaged.

198

For both the one stage batch test and the DLST additional blank controls were run

199

with deionized water in additional leaching vessels with the same treatments as for

200

the eluate. Immediately after sampling the total volumes of eluates were divided into

201

several containers for the different ecotoxicological tests and frozen at < -18 °C for up

202

to 2 months until testing in accordance to ISO 5667-16. In this way, the different tests

203

could be carried out independently from each other. The laboratories were asked to

204

use sample containers (PP, PE or glass), which were thoroughly cleaned and rinsed

205

with deionized water.

EP

TE D

M AN U

SC

RI PT

184

206

Ecotoxicity tests. A test battery of four standardised ecotoxicity tests with algae,

208

daphnia, luminescent bacteria and zebrafish eggs was used. The testing strategy is

209

in-line with the technical CEN guidance on the use of ecotoxicity tests applied to

210

construction products currently being developed (see Gartiser et al., submitted).

211

The Algae growth inhibition test was carried out according to ISO 8692 (2012) with

212

the

213

subcapitata). The initial inoculum density should not exceed 104 algae mL-1 (nominal

214

or measured start concentration). The study design consisted in 3 replicates per

215

concentration and 6 controls, the light intensity should be in the range of 60-120 µmol

216

m-2·s-1. Usually, the test is carried out without pH adjustment, but it may be adjusted

217

with extremely alkaline or acid samples. The inhibition of the growth rates was

AC C

207

algae

species

Raphidocelis

subcapitata

(formerly

Pseudokirchneriella

ACCEPTED MANUSCRIPT determined after 72 hours via measurements of algae cell counts, chlorophyll-

219

fluorescence or absorbance.

220

The acute Daphnia toxicity test was used according to ISO 6341 (2012) with Daphnia

221

magna using the synthetic medium as dilution water. The pH was only adjusted to

222

the pH of the synthetic dilution water when it was outside pH 6 - 9. The test design

223

consisted in 2 to 8 beakers per concentration with 2 - 8 daphnids each (Annex F of

224

ISO 6341). The evaluation was realised after 24 h and 48 h. The results of a

225

sensitivity check with potassium dichromate were asked to be reported (valid range

226

between 0.6 – 2.1 mg L-1).

227

The Luminescent bacteria test was realised according to DIN EN ISO 11348 part 1-3

228

(2009). Preferably, freshly grown luminous bacteria (part 1) or liquid-dried

229

luminescent bacteria (part 2) should be used, but freeze-dried luminescent bacteria

230

(part 3) were also accepted. The evaluation of the decrease of luminescence of the

231

marine bacterium Aliivibrio fischeri (formerly Vibrio fischeri) was determined after an

232

exposure time of 30 minutes. As far as the pH was between pH 6.0 und 8.5 no pH-

233

adjustment was required. If the pH was below 6.0 or above 8.5 the pH should be

234

adjusted to 7.0 +/- 0.2. The test design considered two replicates per concentration.

235

The sensitivity of the bacteria batch used should be checked with reference

236

substances (by the supplier or by own measurements).

237

The fish egg test was carried out according to ISO 15088 (2007). The pH of the

238

eluate was adjusted to 7.0 +/- 0.2, if necessary. The fertilized eggs were exposed in

239

24-well cell culture plates at 26 °C +/- 1 °C. The study design consisted in 10

240

replicates (wells) per concentration, four internal negative controls per plate, an

241

additional external negative control and a positive control with 10 replicates each.

242

The evaluation after 24 h and 48 h (+/- 30 minutes) considered the endpoints

243

“coagulated eggs”, "no tail detachment" and "no heartbeat".

244

Most laboratories followed the study plan which requested that the eluates should be

245

stored below -18 °C. However, 6 from 17 laboratorie s used freshly prepared or

246

cooled eluates in all or some biotests. The storage duration at -18 °C was between

247

one and 55 days (median 9 – 20 days) while a maximum storage duration of 2

248

months was recommended. Some laboratories tested the recommended dilution

249

levels, but did not perform follow up testing if no definite test result was obtained.

250

AC C

EP

TE D

M AN U

SC

RI PT

218

ACCEPTED MANUSCRIPT Evaluation. The participants received Excel and Word templates for raw data

252

documentation. The statistical evaluation was centrally realised according to ISO

253

5725-2 (1994), using Microsoft Excel®, and the statistical software ToxRat®

254

Professional 3.2.1 (ToxRat Solutions GmbH, Alsdorf, Germany).

255

As a first step of the evaluation of the bioassays, the validity of the test results was

256

assessed according to the distinct criteria specified in the test guidelines. The validity

257

criteria refer to the maximum mortality in the controls (daphnia and fish egg test), the

258

effects observed with a reference substance (all tests), the oxygen content (daphnia

259

test, fish egg test), the variability of the growth rates of the controls (algae test), the

260

variability of replicates (luminescent bacteria), and the pH shift (algae test).

261

For the valid data, two effect parameters were calculated: The EC50 values (volume

262

percentage causing 50% effects) were calculated with ToxRat Professional via non-

263

linear regression (algae test) or linear regression (Probit analysis) (daphnia, bacteria,

264

fish eggs). If no significant dose response relationship was found the data were

265

further processed via linear regression with the Weibull function or with the

266

Spearman-Kärber or binomial interpolation methods (Environment Canada 2005).

267

The latter applied mainly to the fish egg data. Additionally, the lowest ineffective

268

dilution levels (LID) causing no inhibition or mortality, or only effects below the effect

269

specific threshold were determined. The LID corresponds to the lowest dilution factor

270

D, at which less than 10% mortality (daphnia, fish eggs) is observed, respective at

271

which the inhibition is below 5% (growth rate algae) or 20% (luminescence bacteria).

272

LID-values indicated as "higher than x", were not included in the statistical

273

evaluations. In these cases the geometric mean (and hence toxicity) tend to be

274

somewhat underestimated.

275

In a second step, an outlier analysis was performed. Mandels-h-statistic and Dixon-

276

test were performed using log transformed data. Additionally, the data were checked

277

for values beyond the 95% and 99% tolerance limits (warning charts concept,

278

Guidance document of Environment Canada (2005)), (see figures 1 and 2). Test

279

results identified as outlier on the 99% significance level by at least two methods or

280

as almost-outlier (stragglers) (95% significance level) by two methods and as outlier

281

by one method, were excluded from further analyses.

282

Because of the logarithmic characteristics of EC50 and LID values, all data were log

283

transformed (Y = ln (X)) prior to the following calculations:

AC C

EP

TE D

M AN U

SC

RI PT

251

ACCEPTED MANUSCRIPT mean value µ = y_ =

285

standard deviation- σ = sy =

286

It should be stated, that for calculation of standard deviation according to ISO 5725-

287

2, the data measured by each laboratory should be weighted depending on the

288

numbers of repeated measurements. Since in the present round robin test no

289

repetitions were performed, the simplified formula was applied.

290

By re-transformation (anti-log) of µ und σ using the formulas for log normal

291

distribution the following statistical characteristics are obtained for the original scale

292

X:

293

M AN U

SC

RI PT

284

294 295

geometric mean =

296 297 298 299 300 301

95% confidence limit =

302 303 304

expected value EW (X) =

The expected value is the probability-weighted average of all possible values. For

305

normal distribution expected value and geometric mean are identical, whereas for log

306

normal distributions the expected value exceeds the geometric mean. For log normal

307

distributed data, both standard deviation and variation coefficient (CV%) are related

308

to the expected value rather than to the geometric mean:

310

EXP (µ +

standard deviation Std (X) =

311 312

variation coefficient CV [%] =

313

All calculations were performed using MS Ecxel.

314

*1,96)

EXP (µ + σ * z) with z= 1,96 for 95%; z=2,57 for 99% EXP ( µ +

)

EP

TE D

95% resp. 99% tolerance limits =

AC C

309

EXP (µ)

EW (X) *

ACCEPTED MANUSCRIPT ISO 5725-2 (1994) distinguishes between intra-laboratory variability (repeatability)

316

and inter-laboratory variability. The corresponding variances sum up to the total

317

variability (reproducibility). Since the ring test participants performed each test only

318

once, the intra-laboratory standard deviation of repeatability could not be calculated.

319

Thus, the reproducibility of test results is approximated by the inter-laboratory

320

variability as the lower limit of the expected total variability of the examined process.

RI PT

315

321 322

RESULTS

323

The elution tests showed an acceptable variability of the pH, electric conductivity and

325

TOC, with coefficients of variation of 9% - 46% (electric conductivity) and 12% - 23%

326

(TOC). Obviously, the variability of the conductivity was higher than that of the TOC.

327

With a few exceptions no toxicity was observed in additional blank controls tested in

328

parallel, which included all experimental steps (dilution water, leaching test, pre-

329

treatment of eluates, storage of eluates before testing). Only in 9 from 106 tests

330

performed with the blank controls some toxicity was found at D=2, mostly in the effect

331

range of 10 - 20%. However, the exact LID could not be determined, because no

332

follow-up of higher dilutions was carried out. Thus, the occurrence of false positive

333

tested samples due to artificial effects can be excluded. The overall results are

334

shown in tables 2 and 3. The overall data basis considered for the statistical

335

evaluation is shown in table 4. All daphnia and fish egg tests submitted were valid,

336

while for the luminescent bacteria test L13 had to be excluded, because the validity

337

criteria were not fulfilled. For the algae test an exceptional high number of 9 tests

338

was found to be invalid. In 7 tests the minimum growth rate in the controls of 1.4 day-

339

1

340

controls was above 5% and in 4 tests the pH shift was above 1.5. The outlier tests

341

resulted in the exclusion of three additional biotests from further evaluation: one

342

luminescent bacteria and fish egg test (BAM-G1) and one daphnia test (HSR-2). The

343

reasons for these outliers could not be identified.

344

For the daphnia, fish egg, and luminescent bacteria tests, the remaining data base of

345

11 - 16 valid bioassays was meaningful. In contrast, for the algae test, there were

346

only 7 (HSR-2) or 8 (BAM-G1) valid bioassays at all. Furthermore two out of seven

347

valid algae tests with HSR-2 showed no dose-response relationship (L04, L08), thus

AC C

EP

TE D

M AN U

SC

324

was not achieved, in 4 tests the coefficient of variability of the growth rate in the

ACCEPTED MANUSCRIPT no EC50 values could be determined and the data base was reduced to 5. Due to the

349

poor data base, the results for the algae should be regarded as preliminary.

350

The fish egg test with HSR-2 did not show any effects in seven laboratories. In four

351

laboratories the maximum effects were 10% to 40% at the highest concentration and

352

no dose-response relationship could be derived. Thus, no EC50 values were

353

obtained at all.

354

The results of all bioassays are shown in figures 1 and 2. The EC50 and LID values

355

from invalid bioassays are indicated and were found to be generally in the same

356

range as the valid test results. Nevertheless, non-valid test results and outliers were

357

not considered for calculation of means, confidence and tolerance limits.

358

The overall conclusions resulting from these data are the following:

359

All bioassays except the fish egg test showed considerable toxic effects. For BAM-

360

G1 the sensitivity of test organisms was daphnia > luminescent bacteria > algae >>

361

fish-eggs. In contrast, for HSR-2 the sensitivity of test organisms was luminescent

362

bacteria >> algae > daphnia > fish-eggs. The fish egg test with HSR-2 resulted in

363

such low effects that no meaningful EC50 can be calculated and the median LID is 1

364

which means non-toxic in the undiluted eluate.

365

Generally, the HRS-2 eluate showed much lower toxicity (i.e. higher EC50 and lower

366

LID values) than the BAM-G1 eluate.

TE D

M AN U

SC

RI PT

348

367

The LID results are mostly within the range of ± 2 dilution steps the EC50 values

369

showed reproducibilities between 75% and 110% (BAM-G1) and 9% - 75% (HSR-2).

EP

368

AC C

370 371

DISCUSSION

372

With the exception of the algae test most results from the daphnia, fish egg and

373

luminescent bacteria tests complied with the validity criteria of the respective

374

guidelines. With the algae test, the validity criterion for the growth rate of the ISO

375

8692 test guideline was often not met. One reason for the unexpected high number

376

of invalid algae tests may be that the algae species Raphidocelis subcapitata is not

377

routinely used by all participating laboratories. In Germany Desmodesmus

378

subspicatus is more often used for environmental samples. Another reason could be

379

the fact that laboratories experience with Raphidocelis subcapitata is often based on

380

the test performance according to OECD 201, which is used for chemical testing.

ACCEPTED MANUSCRIPT OECD 201 only requires a biomass increase in the controls of a factor of 16 (growth

382

rate 0.92 day-1) while ISO 8692 demands for an increase of a factor of 67 (growth

383

rate 1.4 day-1).

384

There was no uniform ranking of bioassays with respect to their reproducibility. While

385

the daphnia test was the most sensitive species to BAM-G1, for HSR-2 this was the

386

Luminescent bacteria test. For the BAM-G1 eluate a relative high variability of CV =

387

73% to 110% irrespective of the bioassays was observed, while the reproducibility for

388

the HSR-2 eluate was very good for the daphnia test with the lowest sensitivity (CV =

389

9.3%), followed by the algae test (36.4%). The luminescent bacteria test, being the

390

most sensitive bioassay, showed the highest variability (74.8%). Thus, the

391

reproducibility of the EC50 or LID value depends on the level of toxicity. The more

392

toxic the eluates, the higher was the variability. One reason might be that for

393

construction products containing highly toxic ingredients even small differences in the

394

elution process can lead to large differences of effects to the respective organisms.

395

With respect to the luminescent bacteria test part of the variability might also be

396

explained with the different bacteria sources. Most laboratories (10) used liquid-dried

397

bacteria, but freeze-dried (3) and freshly grown (1) bacteria were also used.

398

Although mostly no toxicity was observed in the blank controls it is recommended to

399

consider such parallel blank controls carried out over the whole process in order to

400

detect possible artefacts.

401

The reproducibility of ecotoxicity tests from different eluates is acceptable and within

402

the expected range and a distinction of highly ecotoxic construction products and

403

non-ecotoxic construction products is possible. The variability of results corresponds

404

to the lower limit of the actual variability of the whole process since only the inter-

405

laboratory variance of the elution process and biotest performance was covered. The

406

intra-laboratory variance of repeated tests in one laboratory could not be determined.

407

The observed variance of the round robin tests also includes further factors such as

408

the eluate pre-treatment and sample storage as well as possible blank effects, which

409

were overall very heterogeneous (figure 3).

410

FINAL CONCLUSION

411

The results from the round robin test demonstrate, that combining leaching and

412

ecotoxicity tests contributes to the characterizing of environmental hazards from

413

complex construction products with a sufficient accuracy. Given the extremely

414

complex overall process the bioassays with eluates from construction products

AC C

EP

TE D

M AN U

SC

RI PT

381

ACCEPTED MANUSCRIPT showed an acceptable reproducibility. Although the reasons for differing results could

416

not be identified, the varying procedures followed by different laboratories in terms of

417

eluate preparation, pretreatment and storage certainly contributed to the overall

418

variability. By further restricting the range of options allowed in the different standards

419

validity and reproducibility is expected to be improved. A further validation of the

420

leaching tests for the generation of eluates planned to take place in 2017 will

421

contribute to the reliability of both chemical and ecotoxicological tests carried out on

422

the eluates.

RI PT

415

423 AUTHOR INFORMATION

425

* Corresponding Author

426

Phone: ++49-761-45512-24, fax: ++49-761-45512-34, e-mail: [email protected].

427

ACKNOWLEDGEMENT

428

This study was carried out on behalf of the Federal Environment Agency (UBA)

429

within the framework of the “Umweltforschungsplan“ – FKZ 3712 95 309 and

430

supported by federal funds.

TE D

431

M AN U

SC

424

REFERENCES

433

Baderna, D., Lomazzi, E., Passoni, A., Pogliaghi, A., Petoumenou, M. I., Bagnati, R.,

434

Fanelli, R., 2015. Chemical characterization and ecotoxicity of three soil foaming

435

agents used in mechanized tunneling. Journal of hazardous materials, 296, 210-220.

436

EP

432

Burkhardt, M., Zuleeg, S., Vonbank, R., Schmid, P., Hean, S., Lamani, X., Boller, M.,

438

2011. Leaching of additives from construction materials to urban storm water runoff.

439

Water science and technology, 63(9), 1974-1982.

440

AC C

437

441

Cenni, R., Janisch, B., Spliethoff, H., Hein, K. R. G., 2001. Legislative and

442

environmental issues on the use of ash from coal and municipal sewage sludge co-

443

firing as construction material. Waste Management, 21(1), 17-31.

444 445

CEN/TR 16110, 2010. Characterization of waste - Guidance on the use of ecotoxicity

446

tests applied to waste

447

ACCEPTED MANUSCRIPT 448

CEN/TR 17105, draft, 2016. Construction products - Assessment of release of

449

dangerous substances - Guidance on the use of ecotoxicity tests applied to

450

construction products

451 CEN/TS 16637-1, 2014. Construction products - Assessment of release of

453

dangerous substances - Part 1: Guidance for the determination of leaching tests and

454

additional testing steps

RI PT

452

455

CEN/TS 16637-2, 2014. Construction products - Assessment of release of

457

dangerous substances - Part 2: Horizontal dynamic surface leaching test

SC

456

458

DIN EN 12457-1, 2003. Characterization of granular waste: „Compliance test for

460

leaching of granular waste materials and sludges. One stage batch test at a liquid to

461

solid ratio of 2 L/kg for materials with high solid content and with particle size below 4

462

mm (without or with size reduction)

463

M AN U

459

Eikelboom, R. T., Ruwiel, E., Goumans, J. J. J. M. 2001. The building materials

465

decree: an example of a Dutch regulation based on the potential impact of materials

466

on the environment. Waste Management, 21(3), 295-302.

TE D

464

467

EN ISO 11348, 2007-12. Water quality - Determination of the inhibitory effect of

469

water samples on the light emission of Vibrio fischeri (Luminescent bacteria test) -

470

Part 1: Method using freshly prepared bacteria, Part 2: Method using liquid-dried

471

bacteria, Part 3: Method using freeze-dried bacteria

AC C

472

EP

468

473

Environment Canada, 2005. Guidance Document on Statistical Methods for

474

Environmental Toxicity Tests. Report EPS 1/RM/46, Ottawa, Ontario

475

http://publications.gc.ca/collections/collection_2012/ec/En49-7-1-46-eng.pdf

476 477

Flyhammar, P., Bendz, D., 2006. Leaching of different elements from subbase layers

478

of alternative aggregates in pavement constructions. Journal of hazardous materials,

479

137(1), 603-611.

480

ACCEPTED MANUSCRIPT 481

Galvín, A. P., Ayuso, J., Jiménez, J. R., Agrela, F., 2012. Comparison of batch

482

leaching tests and influence of pH on the release of metals from construction and

483

demolition wastes. Waste management, 32(1), 88-95.

484 Gartiser, S., Heisterkamp, I., Schoknecht, U., Burkhardt, M., Ratte, M., Ilvonen, O.,

486

Recommendation for a test battery for the ecotoxicological evaluation of the

487

environmental soundness of construction products. Chemosphere, submitted

RI PT

485

488

Grubbs, F. E., 1969. Procedures for detecting outlying observations in samples.

490

Technometrics 11, 1-21

491

SC

489

Hage, J. L. T., Mulder, E., 2004. Preliminary assessment of three new European

493

leaching tests. Waste management, 24(2), 165-172.

494

M AN U

492

495

ISO 5725-2, 1994. Accuracy (trueness and precision) of measurement methods and

496

results - Part 2: Basic method for the determination of repeatability and reproducibility

497

of a standard measurement method

498

EN ISO 8692, 2012-06. Water quality - Fresh water algal growth inhibition test with

500

unicellular green algae.

501

TE D

499

EN ISO 6341, 2013-1. Water quality - Determination of the inhibition of the mobility of

503

Daphnia magna Straus (Cladocera, Crustacea) - Acute toxicity test

EP

502

504

EN ISO 15088, 2007-10. Water quality - Determination of the acute toxicity of waste

506

water to zebrafish eggs (Danio rerio)

507

AC C

505

508

ISO 5667-16, 1998. Water Quality-Sampling-part 16: Guidance on biotesting of

509

samples

510 511

Krüger, O., Kalbe, U., Richter, E., Egeler, P., Römbke, J., Berger, W., 2013. New

512

approach to the ecotoxicological risk assessment of artificial outdoor sporting

513

grounds. Environmental Pollution 175, 69-74

514

ACCEPTED MANUSCRIPT 515

Nebel, H., Spanka, G., 2013. Harmonisation of test methods for the execution of the

516

European construction products Directive (CPD). Validation of a European leaching

517

test for construction products. Waste and Biomass Valorization, 4(4), 759-767.

518 Pandard, P., Devillers, J., Charissou, A. M., Poulsen, V., Jourdain, M. J., Férard, J.

520

F., Bispo, A., 2006. Selecting a battery of bioassays for ecotoxicological

521

characterization of wastes. Science of the Total Environment, 363(1), 114-125.

RI PT

519

522

Susset, B., & Grathwohl, P., 2011. Leaching standards for mineral recycling

524

materials–A harmonized regulatory concept for the upcoming German Recycling

525

Decree. Waste management, 31(2), 201-214.

526

SC

523

Waara, S., & Färm, C., 2008. An assessment of the potential toxicity of runoff from

528

an urban roadscape during rain events. Environmental Science and Pollution

529

Research, 15(3), 205-210.

530

M AN U

527

Wangler, T. P., Zuleeg, S., Vonbank, R., Bester, K., Boller, M., Carmeliet, J., &

532

Burkhardt, M., 2012. Laboratory scale studies of biocide leaching from façade

533

coatings. Building and Environment, 54, 168-173.

536 537 538 539 540

EP

535

AC C

534

TE D

531

ACCEPTED MANUSCRIPT Table 1. Chemical parameters of eluates Lab code

BAM-G1 pH

HSR-2

Conductivity

TOC

Pre-

[µS/cm]

[mg/L]

treatment

TOC

[µS/cm]

[mg/L]

163

48,4

s/c

6,9

8,6

1,5

7,4

180

43,0

s/c

6,6

7,6

2,9

L04

7,7

180

N/A

s/c/f

6,4

9,0

N/A

L05

8,3

190

43,6

s/c

6,6

9,2

2,1

L06

7,9

189

45,8

s/c

7,1

8,5

1,8

L07

8,2

204

62,6

f

6,9

13,6

1,8

L08

8,1

183

N/A

s/c

6,1

10,3

1,8

L09

8,1

194

48,0

s/c

7,1

7,5

1,8

L10

8,1

179

47,1

s/c

L11

8,3

183

56,7

f

L12

7,9

202

54,0

s/c

L13

8,5

195

44,0

s/f

L14

8,2

195

N/A

L15

7,5

246

N/A

L16

7,5

200

51,0

L17

8,1

182

N/A

L18

7,8

191

N/A

N

17

17

11

Mean

8,0

191,6

49,5

Std

0,3

17,3

TE D

EP AC C

SC

RI PT

8,2

L02

6,1

Pretreatment: s=sieve; c = centrifugation; f=filtration

544

Conductivity

L01

N/A: not analysed

543

pH

6,7

8,5

1,5

7,1

20,0

2,2

7,2

10,2

1,4

7,4

8,8

1,5

c

7,1

22,0

N/A

s/c

7,9

24,2

N/A

s/f

6,8

9,0

1,9

c

7,9

10,0

N/A

s/c

6,9

8,8

N/A

17

17

12

7,0

11,5

1,8

0,5

5,3

0,4

M AN U

541 542

ACCEPTED MANUSCRIPT 545 546

Table 2. Interlaboratory test results with BAM-G1 BAM-G1 Algae 12

Invalide tests

0

4

Outlier No ECx/LID obtained

0

0

1

0

n (initial)

Daphnia

Algae

Bacteria

Fish

17

12

14

12

1

0

0

4

1

0

1

1

0

0

1

1

1

0

2

2

2

0

6

10

11

8

11

11

15

Expected value

2,39

1,07

24,56

552,2

196,9

995,0

10,86

Geometric mean

0,52

1,61

0,76

19,64

396

152,4

804,0

8,66

95% CI

0.34 - 0.81

0.87 - 2.98

0.46 - 1.23

13.2 - 29.2

262 -598

85.9 - 270

536 - 1204

5.82 -12.9

95% PI

0.09 - 3.02

0.28 - 9.23

0.15 - 3.85

5.3 - 72.8

80 - 1961

37.5 - 620

223 - 2891

2,32 - 32.4

99% PI

0,05 - 5,21

0.16 - 15.9

0.09 - 6.39

3.5 - 109.5

49 - 3227

24.2 - 959

150 - 4305

1,54 - 48.8

sR

0.86

2.64

1.1

18.4

538

161

725

8.22

CVR %

111

110

99.7

75.1

97.4

81.8

72.9

75.7

M AN U

SC

16 0,78

n (final)

547 548 549 550 551 552 553

LID Fish 12

RI PT

Daphnia 17

EC50 Bacteria 14

n=number of tests; CI = confidence interval; PI= prediction interval; sR = standard deviation of reproducibility; CVR% = coefficient of variation of reproducibility in %. n (final) =number of individual results available for statistics = results with dose response (ECx) and clear LID(i.e. except “>x”), respectively, after validity check and outlier rejection

Table 3. Interlaboratory test results with HSR-2

HSR-2

n (initial)

Bacteria 14

Fish 12

Daphnia

Algae

Bacteria

Fish

17

12

14

12

5

1

0

0

5

1

0

0

0

0

1

0

0

0

2 5

1 12

0 12

3

0

1

0

13

7

12

12

65.1

45.6

15,86

n.d.

2,39

8,31

59,46

1,25

64.8

42.8

12,70

n.d.

2,34

5,56

47,73

1,2

61.7 - 68.0

31.4 - 58.3

8.7 - 18.5

n.d.

2.09 - 2.61

2.6 -12.2

32.8 - 69.5

1.0 – 1.4

54.0 - 77.8

21.5 - 85.4

3.4 - 46.9

n.d.

1.56 - 3.50

1.0 -32.0

13.0 - 175

0.6 – 2.2

51.0 - 82.3

17.3 - 106

2.3 - 70.5

n.d.

1.38 - 3.96

0.6 - 55.2

8.7 - 262

0.5 – 2.7

6.1

16.6

11.9

n.d.

0.0

9.14

44.2

0.4

9.3

36.4

74.8

n.d.

20.7

110

74.3

32.1

0

Outlier No ECx/LID obtained

1

95% CI 95% PI 99% PI sR CVR %

AC C

Geometric mean

2 14

EP

Expected value

LID

Algae 12

Invalide tests

n (final)

554 555 556 557 558 559 560 561 562 563 564

Daphnia 17

TE D

EC50

n=number of tests; CI = confidence interval; PI= prediction interval; sR = standard deviation of reproducibility; CV% = coefficient of variation in %. n (final) =number of individual results available for statistics = results with dose response (ECx) and clear LID(i.e. except “>x”), respectively, after validity check and outlier rejection

ACCEPTED MANUSCRIPT 565

Table 4. Data basis for the round robin test BAM-G1 Daphnia

Bacteria

HSR-2

Algae

L01

L01

L01

L02

L02 b

Fish

Algae

Fish

L01

L01

L01 c

L02

L02 b

L02

L02

L02 b

L02

L04

L04

L04 a

L04

L04

L05 a

L05 a,b*, c

L05

L05 a

L05 a b c

L05

L06

L06

L06

L06 b

L06

L06 b

L06

L07

L07

L07

L07

L07

L07

L07

L07

L08 a

L08

L08 a

L08 a

L08 a

L08 a,b

L09 a

L09 a

L09 a

L10

L10

L10

L10

L11

L11 a

L11 b

L11

L11 a

L12

L12

L12 c

L12

L12

L13

L13 b* c

L13 b c

L13

L13

L14a

L14a

L14a

L14a

L15

L15

L16

L16

L15 d L16

L16 L17

L18

L18 a,b

17

14

12

16

12

8

L18 12

17

11

15

L09 a

L10

L11 a

L10

L11 b

L11 a

L12

L12 c

L12

L13 b c

L13 b* c

L14a

M AN U

L18

L09

L15 b*

SC

L15 c

L09 a

RI PT

L06

L13 L14a

L15 c

L16

L16

L18

L18 b

14

12

12

13

7

12

EP

TE D

Grey shaded tests: data not considered in evaluation. Blue shaded tests: outliers, not considered for evaluation. Index a) Storage of eluates differing from study plan, b) blank control >10% , b* = no blank control tested, c) not valid, d) nearly identical results reported than for BAM-G1; probably data were mixed up

AC C

566 567 568 569 570

Bacteria

L04 b

L17 No of tests performed No of tests for statistical evaluation

Daphnia

ACCEPTED MANUSCRIPT

EP

TE D

M AN U

SC

RI PT

Figure 1. EC50 and LID from all tests with BAM-G1

AC C

571

EP

TE D

M AN U

SC

Vertical error bars (whiskers): 95% confidence interval of the EC50, Open symbols: invalid tests, red symbols: outliers, green symbols: LID only determined as “greater than”. Only data indicated as blue diamonds were considered for calculation of means and tolerance limits. Blue line: geometric mean, the red dotted / dashed line: 95% / 99% tolerance range. Missing values: test not performed or missing dose response.

AC C

572 573 574 575 576 577

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT

EP

TE D

M AN U

SC

RI PT

Figure 2. EC50 and LID from all tests with HSR-2

AC C

578

ACCEPTED MANUSCRIPT EC50, HSR-2, Fish-Test

SC

Vertical error bars (whiskers):95% confidence interval of the EC50, Open symbols: invalid tests, red symbols: outliers, green symbols: LID only determined as “greater than” Daphnia test EC50, pink diamond: data were nearly identical than that reported for BAM-G1, data judged as probably erroneous. Only data indicated as blue diamonds were considered for calculation of means and tolerance limits. Blue line: geometric mean, the red dotted / dashed line: 95% / 99% tolerance range. Missing values: test not performed or missing dose response. Algae test EC50: The blue triangle refers to L08, which provided a corrected data set after the final evaluation (with the first data set, no EC50 could be derived). The new data could not be considered for the overall statistical evaluations anymore, but is shown here.

M AN U

579 580 581 582 583 584 585 586 587 588 589 590 591

RI PT

No meaningful EC50 could be calculated due to the low sensitivity of the fish egg test to this eluate.

Figure 3. Factors influencing the total variability of test results

EP

Variance of Sub-samples from construction products

AC C

=

TE D

Total variance of test results

592 593 594

+

Variance of Leaching tests

+

Storage Stability of eluates

+

Variance of Sample preparation

+

Pre-treatment of eluates (filtration, centrifugation)

+

Variance of Ecotoxicitcy tests

ACCEPTED MANUSCRIPT

RI PT SC M AN U TE D

• •

EP

• •

A European round robin test has been carried out with 17 laboratories from 5 countries. Two construction products were eluted in an on-stage batch test or a tank test. Ecotoxicity against algae, daphnia, luminescent bacteria and zebrafish eggs was determined according to ISO standards. The more toxic the eluates, the higher was the variability. The inter-laboratory variability of the overall process for the ecotoxicological characterization of construction products in eluates and bioassays was acceptable.

AC C

•