Chemosphere xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Development of a material with reproducible emission of selected volatile organic compounds – l-Chamber study Michael Nohr a,⇑, Wolfgang Horn a, Katharina Wiegner a, Matthias Richter a, Wilhelm Lorenz b a b
BAM Federal Institute for Materials Research and Testing, Germany Martin-Luther-University Halle-Wittenberg, Institute of Chemistry, Food Chemistry and Environmental Chemistry, Germany
h i g h l i g h t s First approach of a reference material with emissions of VOCs less volatile than toluene. Micro-Chamber/Thermal Extractor (l-CTE™) used as test chamber. Lacquer as basic matrix were spiked with VOCs of interests. Mostly the VOC-emission was reproducible with a variation of less than 20% RSD.
a r t i c l e
i n f o
Article history: Received 13 August 2013 Received in revised form 13 December 2013 Accepted 18 December 2013 Available online xxxx Keywords: VOC Reference material Emission testing Indoor air Inter-laboratory study
a b s t r a c t Volatile organic compounds (VOCs) found indoors have the potential to affect human health. Typical sources include building materials, furnishings, cleaning agents, etc. To address this risk, chemical emission testing is used to assess the potential of different materials to pollute indoor air. One objective of the European Joint Research Project ‘‘MACPoll’’ (Metrology for Chemical Pollutants in Air) aims at developing and testing a reference material for the quality control of the emission testing procedure. Furthermore, it would enable comparison of measurement results between test laboratories. The heterogeneity of the majority of materials makes it difficult to find a suitable reference sample. In the present study, styrene, 2-ethyl-1-hexanol, N-methyl-a-pyrrolidone, lindane, n-hexadecane, 1,2dimethyl- and 1,2-di-n-butyl-phthalate were added to 12 commercially available lacquers (6 alkyd and 6 acrylic polymer based lacquers) serving as carrier substrate. After homogenization, the mixtures were loaded into a Markes Micro-Chamber/Thermal Extractor (l-CTE™) for curing and investigation of the emission behavior for each compound. For almost all of the investigated chemicals, the preferred glossy acrylic lacquer showed emissions that were reproducible with a variation of less than 20% RSD. Such lacquer systems have therefore been shown to be good candidates for use as reference materials in interlaboratory studies. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Several investigations have shown that building materials and furniture can emit substances that have the potential to cause negative health issues. The substances emitted to air are classified by the WHO (1989) as very volatile, volatile and semi-volatile organic compounds (VVOC, VOC, SVOC). The emission of these chemicals is usually determined either by field measurements of indoor air or by placing test specimens of indoor materials into emission test chambers under controlled conditions of temperature, air exchange and humidity, to simulate indoor conditions. The results of such investigations can be used for taking steps to improve ⇑ Corresponding author. Tel.: +49 30 8104 1421.
indoor air quality. For example, labelling schemes, such as THE BLUE ANGEL in Germany have been developed to evaluate building products and furniture in terms of their emission behaviour. Round robin tests are widely accepted as an essential tool for checking the performance of test laboratories. Reference materials, which provide representative and stable (constant) test samples for all participants, play a fundamental role (Namies´nik and Zygmunt, 1999) in round robin studies and analytical proficiency testing generally. They require consistent physical and chemical properties, so that any variations in results can be solely ascribed to the performance of the measurement methods and not to the material itself. In order to be able to draw reliable conclusions, it is also important that the true or accurate value of the parameter under test should be well characterised and validated for the reference material.
E-mail address: [email protected] (M. Nohr). 0045-6535/$ - see front matter Ó 2014 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.chemosphere.2013.12.047
Please cite this article in press as: Nohr, M., et al. Development of a material with reproducible emission of selected volatile organic compounds – l-Chamber study. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2013.12.047
2
M. Nohr et al. / Chemosphere xxx (2014) xxx–xxx
Reference materials with such properties are not yet commercially available for emission testing but the need for them is well understood. Previous round robin tests with different building materials have shown that coefficients of variation of up to 300% are to be expected between the participants (Howard-Reed et al., 2007). The heterogeneity of the materials used in these studies is often thought to contribute to these high variations. Cox et al. (2010) and Wei et al. (2012) published two approaches for reference materials for toluene. As part of the European Joint Research Project ‘‘MACPoll’’ (Metrology for Chemical Pollutants in Air), the present study aims at the development of a reference material for emission testing that is particularly targeted at compounds less volatile than toluene. As described below, liquid lacquers are spiked with selected substances, before being homogenized and loaded into micro-scale emission test chambers (l-CTEÒ) in which the lacquer is cured and the emission behaviour investigated. 2. Materials and methods
2.3. Emission test chamber The l-CTE (MARKES), which contains six 44 mL capacity microchambers in one heating block, was used for these investigations. Each microchamber was purged with the same constant flow of pure air (AIR LIQUIDE). The air passed over the samples in each microchamber sweeping emitted VOCs onto thermodesorption tubes connected to the outlets. The l-CTE temperature was set to 25 °C and a flow of approx. 30 mL min1 was applied. The size of the l-CTE and range of parameters it offered made it more convenient than normal emission test chambers for this investigation. Furthermore, the l-CTE allowed the emission experiments to be accelerated by raising the temperature. However, it must be borne in mind that the standard version of the l-CTE is normally used with dry gas and provides limited flexibility for humidification. VOCs were sampled onto the thermodesorption tubes for 8 min resulting in the collection of around 250 mL of vapour in each case. This was carried out on defined days – namely 1, 3, 7, 9, 11, 14, 16, 18 and 21 days after loading.
2.1. Chemicals
3. Experimental
For the investigation the five VOC styrene (100-42-5, ALFA AESAR, 99.5%), N-methyl-a-pyrrolidone (872-50-4, NMP, FLUKA, >99.9%), 2-ethyl-1-hexanol (104-76-7, E.H., ALDRICH, 99.6%), 1,2-dimethyl-phthalate (131-11-3, DMP, ALFA AESAR, 99%), n-hexadecane (544-76-3, C16, ALDRICH, 99%) and the two SVOC lindane (58-89-9, ALDRICH, 99.8%) and 1,2-di-n-butyl-phthalate (13111-3, DBP, ALDRICH >98%) were selected. All of these substances can be found in indoor air as they emit from different building materials. NMP and DBP (a plasticiser like DMP) are specified in the candidate list of ‘‘substances of very high concern’’ (ECHA, 2006) because of their toxicity for reproduction. E.H. and styrene are reactants for many synthetic materials.
3.1. Lacquer preparation
2.2. Chemical analysis Tenax-TAÒ-tubes (GERSTEL) were used to collect air samples. The tubes were desorbed in a GERSTEL TDS A (thermodesorption system) together with a GERSTEL CIS 4 (cold injection system) linked with an AGILENT 6890 gas chromatograph (GC) and an AGILENT 5973 N mass spectrometer (MS). The thermodesorption starts at 30 °C for 1 min to 260 °C with 30 °C min1 and holding for 5 min. The transfer line was held at 300 °C the whole time. The CIS was programmed with a starting temperature of 120 °C for 0.05 min followed by a heating period of 12 °C s1 to 300 °C and holding for 3 min. No split was selected and the injection liner was packed with glass wool. The GC was equipped with a Rxi-5ms column (RESTEK, 60 m, 25 mm, 0.25 lm) that was flushed with a constant flow of helium (ALPHAGAZ AIR LIQUIDE) at 1.4 mL min1. The GC oven was programmed to start at 40 °C for 4 min, then heating with 10 °C min1 to 150 °C for 1 min and again to 300 °C with 8 °C min1 for 5 min. The MS (transfer line at 300 °C, EIsource at 230 °C and quadrupole at 150 °C) operated in SIM-mode (single ion monitoring) with a solvent delay of 4 min. The quantifier (underlined) and the qualifiers were as follows: styrene: 104, 103, 78; E.H.: 57, 83, 70; NMP: 99, 98, 44; DMP: 163, 77, 164; C16: 57, 71, 43, 226; lindane: 181, 183, 219, 217; DBP: 149, 150, 223. The seven substances were externally calibrated with 12 solutions ranging in concentration from approximately 1 to 200 ng lL1 of the mixed analytes in methanol (MeOH, J.T. Baker). The tubes were each spiked with 1 lL of the stock solutions. The measured analyte amount divided by the sampling volume results in the air concentration, given in lg m3.
The samples were prepared by weighing portions of lacquer and decanting them into 100 mL screw cap bottles. Defined amounts of the pure analytes were then added using gas-tight syringes under simultaneous magnetic stirring. Because lindane is the only solid substance among the investigated compounds, it had to be dissolved in MeOH before adding it to the lacquer. Afterwards, the mixture was agitated for 1 h in the closed bottle, and defined amounts of the lacquer mixture were filled into Petri dishes (£ 33 mm, GREINER BIO ONE). After preparation, aliquots of each batch were immediately loaded into the l-CTE and left to cure for one day in the microchambers under the conditions described above. 3.2. Lacquer selection The 12 standard lacquers selected were all made by the same manufacturer and purchased from the same local building supplies store (to improve comparability). Six of them were acrylic based (incorporating water as solvent) and six were based on alkyd polymers (incorporating organic solvents). The acrylic lacquers contained water, a dispersion of acrylat and polyurethane, glycols, additives, preservatives (methyl-, benzyl- and chlor-isothiazolinone) and pigments (if it is a coloured lacquer). Also these lacquers carried the Blue Angel eco-label for low emission. The alkyd lacquers contained the alkyd resin, additives, white spirit and pigments (if coloured). Lacquers with matching colour and finish (matt or gloss) were selected from each of the two main groups, Fig. 1 shows the nomenclature applied to each selected lacquer: The following lacquers were purchased (WGI is similar to WGW, but WGW is not available): WGC, WGI, WGR, WMC, WMW, WMR, OGC, OGW, OGR, OMC, OMW and OMR. To test emission performance, 55 g of each lacquer was spiked with 50 lL of each tested substance except lindane. 2.5 mL of a methanolic lindane-solution of 0.02 mg lL1 was spiked into the lacquer. After a stirring time of one hour, 2 g of each lacquer mixture was transferred into two Petri dishes and loaded into two different microchambers for duplicate analysis. Vapour sampling was carried out 1, 3, 7, 9, 11, 14, 16, 18 and 21 days after loading the samples into the l-CTE. The nomenclature for the tests was extended by adding the total mass of each lacquer (e.g. 55 g) and the mass
Please cite this article in press as: Nohr, M., et al. Development of a material with reproducible emission of selected volatile organic compounds – l-Chamber study. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2013.12.047
3
M. Nohr et al. / Chemosphere xxx (2014) xxx–xxx
introduced into petri-dishes and loaded into the microchambers (e.g. 2 g), for example WGC55_2. The concentration or level of spiking of each analyte in the lacquer x (mg g1) is the quotient of the mass of each compound introduced (calculated from the spiked volume VA multiplied by the analyte density qA) and the mass of lacquer mL.
x¼
Table 1 Calculation of the added analyte amounts.
mA qA V A ¼ mL mL
For lindane, which is dissolved in MeOH, the concentration cl must be used instead qA. Table 1 shows the calculated x-values for the selected analytes. 3.3. Optimization of the emitted SVOC amounts To increase the emission of lindane and DBP two approaches were investigated. In the first experiment, 2 g (WGC55_2) and 4.5 g (WGC55_4.5) samples of spiked WGC, each containing the same concentration of DBP and lindane (Table 1) were loaded into microchambers in Petri dishes to check the influence of sample loading on emission behaviour. In the second experiment 2 g samples of lacquers loaded with higher concentrations of the two analytes were investigated. To prepare these samples, 25 g of WGC was loaded with higher levels of the 2 analytes leading to x-values of 4.08 mg g1 lindane and 6.72 mg g1 DBP (WGC25_2). Each sample was loaded into the l-CTE in duplicate and vapour sampling was carried out 1, 3, 7, 9, 11, 16, 18, 21, 23 and 25 days after loading. 3.4. Determination of repeatability and reproducibility of the emission To determine the reproducibility (relative standard deviation (RSD repro)) of analyte emissions from spiked lacquers, three 50 g samples of WGC were spiked with all seven substances. Table 4 shows the x-values for the analytes which were adjusted slightly, according to the outcome of previous experiments, to keep the expected vapour concentrations within the calibration range. Sixfold 2 g aliquots of each of these three spiked lacquer samples were then transferred to Petri dishes and loaded subsequently into the l-CTE such that all six chambers contained sample and any deviation in emissions between each of the six chambers could be determined. This deviation averaged over the three batches indicates the relative standard deviation of repeatability (RSD repeat). Samples were collected on days 2, 4, 7, 9 and 11 after loading. Fig. 2 summarizes the preparation steps and the investigated deviations. To compensate for slight flow-rate differences between each microchamber during the whole emission process, the results are calculated in area-specific emission rates: SERA (ISO 16000-9, 2006). These rates are calculated by multiplying the air concentration with the quotient of chamber flow (V) and emission surface (A):
SERA ¼ canalyte
V_ chamber Asurface
lg m2 h
Fig. 1. Nomenclature for the tested lacquers (w – acrylic, o – alkyd).
a
Analyte
qA (mg lL1)
VA (lL)
mA (mg)
mL (g)
x (mg g1)
Styrene E.H. NMP DMP C16 Lindane DBP
0.91 0.83 1.03 1.19 0.77 0.02a 1.04
50 50 50 50 50 2500 50
45.5 41.5 51.5 59.5 38.5 50 52.0
55 55 55 55 55 55 55
0.83 0.75 0.94 1.08 0.70 0.91 0.95
cl used instead qA.
The constant circular emission surface of the Petri dishes with a diameter of 33 mm was calculated to be 855.3 mm2. The repeatability and reproducibility are indicators of the homogeneity and stability of material emission rates. 4. Results and discussion 4.1. Lacquer selection All the alkyd lacquers tested in this study showed similar emission profiles. As expected, high concentrations of solvents were emitted completely masking the signals of NMP and E.H. and grossly overloading the GCMS analytical system. The solvent emission levels did not come down, even after three days. In fact the high levels resulted in significant carryover in the analytical system, which interfered with subsequent measurements and which could not be completely eliminated despite multiple heating cycles of the thermal desorption system. For these reasons this sort of lacquers were excluded from the study. In contrast to alkyd based lacquers, the acrylic ones show almost no inherent emissions, thus providing a much more suitable substrate for preparing a reference material for emission testing. The emission curves for styrene, E.H. and NMP in acrylic-based lacquers are typical of more volatile organic compounds showing a rapid initial increase of emission rate reaching a maximum between 1 and 3 days, followed by a period of decay – see the example of styrene in Fig. 3. The other compounds show emission curves that are more typical for SVOCs with a slower initial increase of emission rate followed by a very slow decay (Fig. 4). Table 2 summarizes the emission characteristics of each compound from each acrylic type lacquer by listing the test chamber concentrations on 1st and 21st day after loading. It was decided to select one acrylic lacquer for further, more detailed evaluation based on these results. Key factors in lacquer selection included emission levels and ease of handling. In general, the emission levels from WGC, WMC and WMW were the highest, and so the preferred lacquer had to be one of these three. For the SVOCs lindane and DBP in particular, the emitted amounts were very low, making it difficult to interpret results for these substances. Moreover WMW and the other coloured lacquers showed three major disadvantages compared to the clear lacquers WMC and WGC: 1. Even after having transferring them to screw-cap bottles they cured within a few months while the clear acrylic lacquers remained in the liquid form for at least a year. Fast curing would make the lacquers unsuitable for use as reference materials, e.g. for round robin tests. 2. The coloured acrylic lacquers tended to produce fissures while curing in the Petri dishes. This meant they produced a less reproducible and less well defined surface – factors which have a significant impact on compound release rate.
Please cite this article in press as: Nohr, M., et al. Development of a material with reproducible emission of selected volatile organic compounds – l-Chamber study. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2013.12.047
4
M. Nohr et al. / Chemosphere xxx (2014) xxx–xxx
Fig. 2. Preparation for repeatability and reproducibility investigations.
3. The coloured lacquers are also more viscous making decantation and stirring less effective and reproducible. For these reasons coloured lacquers were excluded from further investigation. A comparison of WMC and WGC shows that the lacquer with the glossy finish (WGC) is preferred, despite having slightly lower emission rates than WMC, because it is less viscous and because the surface remains smooth and homogeneous after curing. 4.2. Optimization of the emitted SVOC amounts Table 3 shows the highest (max) and lowest (min) micro-chamber vapour concentrations over the different sampling days for lindane and DBP to show that it is possible to produce emission rates above the limit of quantitation LOQ over the whole sampling period.
It is clear that increasing the amount of lacquer introduced to the microchamber does not boost emissions. On the other hand emission rates can be enhanced by increasing the concentrations of the analytes inside the lacquers. For example, increasing the concentration of lindane inside the lacquer (from 0.84 to 4.08 mg g1) increases emission levels by almost the same factor from 6 to 37 lg m3. DBP emissions respond similarly to increasing the concentration of analyte inside the lacquer. These phenomena are to be expected in accordance to Yang et al. (2001). The emission can be considered as a partitioning or distribution of the compounds between the solid lacquer-surface film and the gaseous air-phase The partition coefficient Kd of a compound A results in the concentration of A in the surrounding air (cair) divided by the concentration of A inside the lacquer surface film (clacquer-film):
K d ðAÞ ¼
cair ðAÞ clacquer-film ðAÞ
1400
air concentration in µg/m³
1200
1000
WMC WMR WMW WGC WGR WGI
800
600
400
200
0
1
3
7
9
11
14
16
18
21
sampling day Fig. 3. Using the l-CTE to evaluate emissions of styrene from different spiked acrylic lacquers over the course of several days. Two samples of each lacquer were loaded. Note that similar curves were obtained for E.H. and NMP.
Please cite this article in press as: Nohr, M., et al. Development of a material with reproducible emission of selected volatile organic compounds – l-Chamber study. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2013.12.047
5
M. Nohr et al. / Chemosphere xxx (2014) xxx–xxx
250
air concentration in µg/m³
200
150
WMC WMR WMW WGC WGR WGI
100
50
0
1
3
7
9
11
14
16
18
21
sampling day Fig. 4. Using the l-CTE to evaluate emissions of DMP from spiked acrylic lacquers over the course of several days. Two samples of each lacquer were loaded. Note that similar curves were obtained for C16, lindane and DBP.
Table 2 Test chamber concentrations in lg m3 on 1st and 21st day after loading.
Spiking (mg g1)
Styrene
E.H.
NMP
DMP
C16
Lindane
DBP
0.79
0.72
0.90
1.03
0.67
0.85
0.90
WMC
1 d (lg m3) 21 d (lg m3)
1300 110
2100 67
160 210
210 120
110 260
12 12
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Development of a material with reproducible emission of selected volatile organic compounds – l-Chamber study Michael Nohr a,⇑, Wolfgang Horn a, Katharina Wiegner a, Matthias Richter a, Wilhelm Lorenz b a b
BAM Federal Institute for Materials Research and Testing, Germany Martin-Luther-University Halle-Wittenberg, Institute of Chemistry, Food Chemistry and Environmental Chemistry, Germany
h i g h l i g h t s First approach of a reference material with emissions of VOCs less volatile than toluene. Micro-Chamber/Thermal Extractor (l-CTE™) used as test chamber. Lacquer as basic matrix were spiked with VOCs of interests. Mostly the VOC-emission was reproducible with a variation of less than 20% RSD.
a r t i c l e
i n f o
Article history: Received 13 August 2013 Received in revised form 13 December 2013 Accepted 18 December 2013 Available online xxxx Keywords: VOC Reference material Emission testing Indoor air Inter-laboratory study
a b s t r a c t Volatile organic compounds (VOCs) found indoors have the potential to affect human health. Typical sources include building materials, furnishings, cleaning agents, etc. To address this risk, chemical emission testing is used to assess the potential of different materials to pollute indoor air. One objective of the European Joint Research Project ‘‘MACPoll’’ (Metrology for Chemical Pollutants in Air) aims at developing and testing a reference material for the quality control of the emission testing procedure. Furthermore, it would enable comparison of measurement results between test laboratories. The heterogeneity of the majority of materials makes it difficult to find a suitable reference sample. In the present study, styrene, 2-ethyl-1-hexanol, N-methyl-a-pyrrolidone, lindane, n-hexadecane, 1,2dimethyl- and 1,2-di-n-butyl-phthalate were added to 12 commercially available lacquers (6 alkyd and 6 acrylic polymer based lacquers) serving as carrier substrate. After homogenization, the mixtures were loaded into a Markes Micro-Chamber/Thermal Extractor (l-CTE™) for curing and investigation of the emission behavior for each compound. For almost all of the investigated chemicals, the preferred glossy acrylic lacquer showed emissions that were reproducible with a variation of less than 20% RSD. Such lacquer systems have therefore been shown to be good candidates for use as reference materials in interlaboratory studies. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Several investigations have shown that building materials and furniture can emit substances that have the potential to cause negative health issues. The substances emitted to air are classified by the WHO (1989) as very volatile, volatile and semi-volatile organic compounds (VVOC, VOC, SVOC). The emission of these chemicals is usually determined either by field measurements of indoor air or by placing test specimens of indoor materials into emission test chambers under controlled conditions of temperature, air exchange and humidity, to simulate indoor conditions. The results of such investigations can be used for taking steps to improve ⇑ Corresponding author. Tel.: +49 30 8104 1421.
indoor air quality. For example, labelling schemes, such as THE BLUE ANGEL in Germany have been developed to evaluate building products and furniture in terms of their emission behaviour. Round robin tests are widely accepted as an essential tool for checking the performance of test laboratories. Reference materials, which provide representative and stable (constant) test samples for all participants, play a fundamental role (Namies´nik and Zygmunt, 1999) in round robin studies and analytical proficiency testing generally. They require consistent physical and chemical properties, so that any variations in results can be solely ascribed to the performance of the measurement methods and not to the material itself. In order to be able to draw reliable conclusions, it is also important that the true or accurate value of the parameter under test should be well characterised and validated for the reference material.
E-mail address: [email protected] (M. Nohr). 0045-6535/$ - see front matter Ó 2014 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.chemosphere.2013.12.047
Please cite this article in press as: Nohr, M., et al. Development of a material with reproducible emission of selected volatile organic compounds – l-Chamber study. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2013.12.047
2
M. Nohr et al. / Chemosphere xxx (2014) xxx–xxx
Reference materials with such properties are not yet commercially available for emission testing but the need for them is well understood. Previous round robin tests with different building materials have shown that coefficients of variation of up to 300% are to be expected between the participants (Howard-Reed et al., 2007). The heterogeneity of the materials used in these studies is often thought to contribute to these high variations. Cox et al. (2010) and Wei et al. (2012) published two approaches for reference materials for toluene. As part of the European Joint Research Project ‘‘MACPoll’’ (Metrology for Chemical Pollutants in Air), the present study aims at the development of a reference material for emission testing that is particularly targeted at compounds less volatile than toluene. As described below, liquid lacquers are spiked with selected substances, before being homogenized and loaded into micro-scale emission test chambers (l-CTEÒ) in which the lacquer is cured and the emission behaviour investigated. 2. Materials and methods
2.3. Emission test chamber The l-CTE (MARKES), which contains six 44 mL capacity microchambers in one heating block, was used for these investigations. Each microchamber was purged with the same constant flow of pure air (AIR LIQUIDE). The air passed over the samples in each microchamber sweeping emitted VOCs onto thermodesorption tubes connected to the outlets. The l-CTE temperature was set to 25 °C and a flow of approx. 30 mL min1 was applied. The size of the l-CTE and range of parameters it offered made it more convenient than normal emission test chambers for this investigation. Furthermore, the l-CTE allowed the emission experiments to be accelerated by raising the temperature. However, it must be borne in mind that the standard version of the l-CTE is normally used with dry gas and provides limited flexibility for humidification. VOCs were sampled onto the thermodesorption tubes for 8 min resulting in the collection of around 250 mL of vapour in each case. This was carried out on defined days – namely 1, 3, 7, 9, 11, 14, 16, 18 and 21 days after loading.
2.1. Chemicals
3. Experimental
For the investigation the five VOC styrene (100-42-5, ALFA AESAR, 99.5%), N-methyl-a-pyrrolidone (872-50-4, NMP, FLUKA, >99.9%), 2-ethyl-1-hexanol (104-76-7, E.H., ALDRICH, 99.6%), 1,2-dimethyl-phthalate (131-11-3, DMP, ALFA AESAR, 99%), n-hexadecane (544-76-3, C16, ALDRICH, 99%) and the two SVOC lindane (58-89-9, ALDRICH, 99.8%) and 1,2-di-n-butyl-phthalate (13111-3, DBP, ALDRICH >98%) were selected. All of these substances can be found in indoor air as they emit from different building materials. NMP and DBP (a plasticiser like DMP) are specified in the candidate list of ‘‘substances of very high concern’’ (ECHA, 2006) because of their toxicity for reproduction. E.H. and styrene are reactants for many synthetic materials.
3.1. Lacquer preparation
2.2. Chemical analysis Tenax-TAÒ-tubes (GERSTEL) were used to collect air samples. The tubes were desorbed in a GERSTEL TDS A (thermodesorption system) together with a GERSTEL CIS 4 (cold injection system) linked with an AGILENT 6890 gas chromatograph (GC) and an AGILENT 5973 N mass spectrometer (MS). The thermodesorption starts at 30 °C for 1 min to 260 °C with 30 °C min1 and holding for 5 min. The transfer line was held at 300 °C the whole time. The CIS was programmed with a starting temperature of 120 °C for 0.05 min followed by a heating period of 12 °C s1 to 300 °C and holding for 3 min. No split was selected and the injection liner was packed with glass wool. The GC was equipped with a Rxi-5ms column (RESTEK, 60 m, 25 mm, 0.25 lm) that was flushed with a constant flow of helium (ALPHAGAZ AIR LIQUIDE) at 1.4 mL min1. The GC oven was programmed to start at 40 °C for 4 min, then heating with 10 °C min1 to 150 °C for 1 min and again to 300 °C with 8 °C min1 for 5 min. The MS (transfer line at 300 °C, EIsource at 230 °C and quadrupole at 150 °C) operated in SIM-mode (single ion monitoring) with a solvent delay of 4 min. The quantifier (underlined) and the qualifiers were as follows: styrene: 104, 103, 78; E.H.: 57, 83, 70; NMP: 99, 98, 44; DMP: 163, 77, 164; C16: 57, 71, 43, 226; lindane: 181, 183, 219, 217; DBP: 149, 150, 223. The seven substances were externally calibrated with 12 solutions ranging in concentration from approximately 1 to 200 ng lL1 of the mixed analytes in methanol (MeOH, J.T. Baker). The tubes were each spiked with 1 lL of the stock solutions. The measured analyte amount divided by the sampling volume results in the air concentration, given in lg m3.
The samples were prepared by weighing portions of lacquer and decanting them into 100 mL screw cap bottles. Defined amounts of the pure analytes were then added using gas-tight syringes under simultaneous magnetic stirring. Because lindane is the only solid substance among the investigated compounds, it had to be dissolved in MeOH before adding it to the lacquer. Afterwards, the mixture was agitated for 1 h in the closed bottle, and defined amounts of the lacquer mixture were filled into Petri dishes (£ 33 mm, GREINER BIO ONE). After preparation, aliquots of each batch were immediately loaded into the l-CTE and left to cure for one day in the microchambers under the conditions described above. 3.2. Lacquer selection The 12 standard lacquers selected were all made by the same manufacturer and purchased from the same local building supplies store (to improve comparability). Six of them were acrylic based (incorporating water as solvent) and six were based on alkyd polymers (incorporating organic solvents). The acrylic lacquers contained water, a dispersion of acrylat and polyurethane, glycols, additives, preservatives (methyl-, benzyl- and chlor-isothiazolinone) and pigments (if it is a coloured lacquer). Also these lacquers carried the Blue Angel eco-label for low emission. The alkyd lacquers contained the alkyd resin, additives, white spirit and pigments (if coloured). Lacquers with matching colour and finish (matt or gloss) were selected from each of the two main groups, Fig. 1 shows the nomenclature applied to each selected lacquer: The following lacquers were purchased (WGI is similar to WGW, but WGW is not available): WGC, WGI, WGR, WMC, WMW, WMR, OGC, OGW, OGR, OMC, OMW and OMR. To test emission performance, 55 g of each lacquer was spiked with 50 lL of each tested substance except lindane. 2.5 mL of a methanolic lindane-solution of 0.02 mg lL1 was spiked into the lacquer. After a stirring time of one hour, 2 g of each lacquer mixture was transferred into two Petri dishes and loaded into two different microchambers for duplicate analysis. Vapour sampling was carried out 1, 3, 7, 9, 11, 14, 16, 18 and 21 days after loading the samples into the l-CTE. The nomenclature for the tests was extended by adding the total mass of each lacquer (e.g. 55 g) and the mass
Please cite this article in press as: Nohr, M., et al. Development of a material with reproducible emission of selected volatile organic compounds – l-Chamber study. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2013.12.047
3
M. Nohr et al. / Chemosphere xxx (2014) xxx–xxx
introduced into petri-dishes and loaded into the microchambers (e.g. 2 g), for example WGC55_2. The concentration or level of spiking of each analyte in the lacquer x (mg g1) is the quotient of the mass of each compound introduced (calculated from the spiked volume VA multiplied by the analyte density qA) and the mass of lacquer mL.
x¼
Table 1 Calculation of the added analyte amounts.
mA qA V A ¼ mL mL
For lindane, which is dissolved in MeOH, the concentration cl must be used instead qA. Table 1 shows the calculated x-values for the selected analytes. 3.3. Optimization of the emitted SVOC amounts To increase the emission of lindane and DBP two approaches were investigated. In the first experiment, 2 g (WGC55_2) and 4.5 g (WGC55_4.5) samples of spiked WGC, each containing the same concentration of DBP and lindane (Table 1) were loaded into microchambers in Petri dishes to check the influence of sample loading on emission behaviour. In the second experiment 2 g samples of lacquers loaded with higher concentrations of the two analytes were investigated. To prepare these samples, 25 g of WGC was loaded with higher levels of the 2 analytes leading to x-values of 4.08 mg g1 lindane and 6.72 mg g1 DBP (WGC25_2). Each sample was loaded into the l-CTE in duplicate and vapour sampling was carried out 1, 3, 7, 9, 11, 16, 18, 21, 23 and 25 days after loading. 3.4. Determination of repeatability and reproducibility of the emission To determine the reproducibility (relative standard deviation (RSD repro)) of analyte emissions from spiked lacquers, three 50 g samples of WGC were spiked with all seven substances. Table 4 shows the x-values for the analytes which were adjusted slightly, according to the outcome of previous experiments, to keep the expected vapour concentrations within the calibration range. Sixfold 2 g aliquots of each of these three spiked lacquer samples were then transferred to Petri dishes and loaded subsequently into the l-CTE such that all six chambers contained sample and any deviation in emissions between each of the six chambers could be determined. This deviation averaged over the three batches indicates the relative standard deviation of repeatability (RSD repeat). Samples were collected on days 2, 4, 7, 9 and 11 after loading. Fig. 2 summarizes the preparation steps and the investigated deviations. To compensate for slight flow-rate differences between each microchamber during the whole emission process, the results are calculated in area-specific emission rates: SERA (ISO 16000-9, 2006). These rates are calculated by multiplying the air concentration with the quotient of chamber flow (V) and emission surface (A):
SERA ¼ canalyte
V_ chamber Asurface
lg m2 h
Fig. 1. Nomenclature for the tested lacquers (w – acrylic, o – alkyd).
a
Analyte
qA (mg lL1)
VA (lL)
mA (mg)
mL (g)
x (mg g1)
Styrene E.H. NMP DMP C16 Lindane DBP
0.91 0.83 1.03 1.19 0.77 0.02a 1.04
50 50 50 50 50 2500 50
45.5 41.5 51.5 59.5 38.5 50 52.0
55 55 55 55 55 55 55
0.83 0.75 0.94 1.08 0.70 0.91 0.95
cl used instead qA.
The constant circular emission surface of the Petri dishes with a diameter of 33 mm was calculated to be 855.3 mm2. The repeatability and reproducibility are indicators of the homogeneity and stability of material emission rates. 4. Results and discussion 4.1. Lacquer selection All the alkyd lacquers tested in this study showed similar emission profiles. As expected, high concentrations of solvents were emitted completely masking the signals of NMP and E.H. and grossly overloading the GCMS analytical system. The solvent emission levels did not come down, even after three days. In fact the high levels resulted in significant carryover in the analytical system, which interfered with subsequent measurements and which could not be completely eliminated despite multiple heating cycles of the thermal desorption system. For these reasons this sort of lacquers were excluded from the study. In contrast to alkyd based lacquers, the acrylic ones show almost no inherent emissions, thus providing a much more suitable substrate for preparing a reference material for emission testing. The emission curves for styrene, E.H. and NMP in acrylic-based lacquers are typical of more volatile organic compounds showing a rapid initial increase of emission rate reaching a maximum between 1 and 3 days, followed by a period of decay – see the example of styrene in Fig. 3. The other compounds show emission curves that are more typical for SVOCs with a slower initial increase of emission rate followed by a very slow decay (Fig. 4). Table 2 summarizes the emission characteristics of each compound from each acrylic type lacquer by listing the test chamber concentrations on 1st and 21st day after loading. It was decided to select one acrylic lacquer for further, more detailed evaluation based on these results. Key factors in lacquer selection included emission levels and ease of handling. In general, the emission levels from WGC, WMC and WMW were the highest, and so the preferred lacquer had to be one of these three. For the SVOCs lindane and DBP in particular, the emitted amounts were very low, making it difficult to interpret results for these substances. Moreover WMW and the other coloured lacquers showed three major disadvantages compared to the clear lacquers WMC and WGC: 1. Even after having transferring them to screw-cap bottles they cured within a few months while the clear acrylic lacquers remained in the liquid form for at least a year. Fast curing would make the lacquers unsuitable for use as reference materials, e.g. for round robin tests. 2. The coloured acrylic lacquers tended to produce fissures while curing in the Petri dishes. This meant they produced a less reproducible and less well defined surface – factors which have a significant impact on compound release rate.
Please cite this article in press as: Nohr, M., et al. Development of a material with reproducible emission of selected volatile organic compounds – l-Chamber study. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2013.12.047
4
M. Nohr et al. / Chemosphere xxx (2014) xxx–xxx
Fig. 2. Preparation for repeatability and reproducibility investigations.
3. The coloured lacquers are also more viscous making decantation and stirring less effective and reproducible. For these reasons coloured lacquers were excluded from further investigation. A comparison of WMC and WGC shows that the lacquer with the glossy finish (WGC) is preferred, despite having slightly lower emission rates than WMC, because it is less viscous and because the surface remains smooth and homogeneous after curing. 4.2. Optimization of the emitted SVOC amounts Table 3 shows the highest (max) and lowest (min) micro-chamber vapour concentrations over the different sampling days for lindane and DBP to show that it is possible to produce emission rates above the limit of quantitation LOQ over the whole sampling period.
It is clear that increasing the amount of lacquer introduced to the microchamber does not boost emissions. On the other hand emission rates can be enhanced by increasing the concentrations of the analytes inside the lacquers. For example, increasing the concentration of lindane inside the lacquer (from 0.84 to 4.08 mg g1) increases emission levels by almost the same factor from 6 to 37 lg m3. DBP emissions respond similarly to increasing the concentration of analyte inside the lacquer. These phenomena are to be expected in accordance to Yang et al. (2001). The emission can be considered as a partitioning or distribution of the compounds between the solid lacquer-surface film and the gaseous air-phase The partition coefficient Kd of a compound A results in the concentration of A in the surrounding air (cair) divided by the concentration of A inside the lacquer surface film (clacquer-film):
K d ðAÞ ¼
cair ðAÞ clacquer-film ðAÞ
1400
air concentration in µg/m³
1200
1000
WMC WMR WMW WGC WGR WGI
800
600
400
200
0
1
3
7
9
11
14
16
18
21
sampling day Fig. 3. Using the l-CTE to evaluate emissions of styrene from different spiked acrylic lacquers over the course of several days. Two samples of each lacquer were loaded. Note that similar curves were obtained for E.H. and NMP.
Please cite this article in press as: Nohr, M., et al. Development of a material with reproducible emission of selected volatile organic compounds – l-Chamber study. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2013.12.047
5
M. Nohr et al. / Chemosphere xxx (2014) xxx–xxx
250
air concentration in µg/m³
200
150
WMC WMR WMW WGC WGR WGI
100
50
0
1
3
7
9
11
14
16
18
21
sampling day Fig. 4. Using the l-CTE to evaluate emissions of DMP from spiked acrylic lacquers over the course of several days. Two samples of each lacquer were loaded. Note that similar curves were obtained for C16, lindane and DBP.
Table 2 Test chamber concentrations in lg m3 on 1st and 21st day after loading.
Spiking (mg g1)
Styrene
E.H.
NMP
DMP
C16
Lindane
DBP
0.79
0.72
0.90
1.03
0.67
0.85
0.90
WMC
1 d (lg m3) 21 d (lg m3)
1300 110
2100 67
160 210
210 120
110 260
12 12
