Results Dissemination/Achievements
Project Information
Founding Source: Unitatea Executiva pentru Finantarea Invatamantului Superior, a Cercetarii, Dezvoltarii si Inovarii (UEFISCDI)
Project Code: PN-II-RU-TE-2014-4-1010
Project No.: 282/01.10.2015
Project Title: Influence of the chromatographic analysis of flame retardants on the estimation of human exposure to organohalogenated compounds
RESULTS
Theoretical computations: density functional study of bond dissociation energies in highly brominated diphenyl ethers
Bond dissociation enthalpies (DEs) relevant to thermal dissociation of brominated diphenyl ethers were investigated in a computational approach at the density functional theory (DFT) level. In a preliminary assessment of eight of the most popular exchange-correlation functionals (including B3P86, PBE1PBE, mPW1PW91, wB97xD and two of each of the M05 and M06 families) the M06 meta-hybrid is shown to perform the best in reproducing two experimental C-O and C-Br BDEs, with errors bellow 1 kcal/mol and less dependent on basis set. The M06/cc-pVDZ is chosen as a good compromise between cost and accuracy for computing DEs of seven brominated diphenyl ethers. In the case of decabromodiphenyl ether we report a DE of 68.7 kcal/mol for the homolysis of the ether group and 74 to 77 kcal/mol for bromine cleavage. Compared to the corresponding values of the fully brominated compound, in lower brominated congeners we predict a substantial increase of both DEs with the decrease of bromine content [Maftei et al., Studia UBB Chemia, 2016].
DEs for relevant bonds in bromobenzene (C6H5–Br) and diphenyl ether (C6H5O–C6H5), computed with each of the considered functionals in conjunction with eight basis sets are collected in Tables 1 and 2, respectively. First, it could be noticed the strong variation of the computed values with the exchange-correlation functional, with up to 8-12 kcal/mol (10-15%) discrepancy between M05 values, the lowest overall, and the double exchange M05-2X counterparts. In addition, values computed with the same two functionals are in qualitative disagreement, the later (M05-2X) foreseeing a DE for the C6H5O–C6H5 bond at 3 kcal/mol below the corresponding value for the C6H5–Br, whereas the M05 functional reproduces the same quantities in the reverse order.
Table 1. C6H5–Br bond dissociation enthalpies (in kcal/mol, at 298.15K) computed at different DFT levels
| Exchange-correlation functional | ||||||||
| Basis set | M05 | M05-2X | M06 | M06-2X | B3P86 | PBE0 | mPW1PW91 | wB97xD |
| 6-311G(d) | 74.1 | 85.6 | 79.6 | 86.1 | 79.0 | 78.9 | 77.0 | 81.6 |
| 6-311+G(d) | 72.5 | 84.6 | 78.4 | 85.0 | 77.9 | 77.7 | 75.9 | 80.4 |
| 6-311+G(d,p) | 72.7 | 84.7 | 78.6 | 85.0 | 78.1 | 77.9 | 76.0 | 80.5 |
| 6-311++G(d,p) | 72.6 | 84.6 | 78.5 | 84.9 | 78.0 | 77.8 | 76.0 | 80.4 |
| cc-pVDZ | 75.6 | 86.7 | 80.6 | 87.7 | 80.2 | 80.0 | 78.1 | 82.7 |
| cc-pVTZ | 74.2 | 86.1 | 80.1 | 86.3 | 79.5 | 79.3 | 77.3 | 81.8 |
| aug-cc-pVDZ | 73.3 | 86.0 | 80.2 | 86.3 | 78.9 | 78.5 | 76.7 | 81.7 |
| aug-cc-pVTZ | 73.3 | 86.1 | 80.1 | 86.3 | 78.9 | 78.5 | 76.7 | 81.5 |
While the eight functionals considered generally disagree, even on the ordering of the two bond DEs, values computed with each of the functionals are almost independent on the basis set. In the case of the first four (Pople) basis sets, further addition of diffuse and/or polarization functions to the 6-311G(d) yields C–Br DEs up to 1 kcal/mol lower and has essential no effect (-0.2 kcal/mol) on C–O DEs. Using correlation-consistent (cc, Dunning) basis sets result in even lower differences, DEs computed at DFT/aug-cc-pVTZ being less than 0.5 kcal/mol apart from the corresponding values computed using the cc-pVDZ basis set, the latter at a substantially reduced cost. Provided the methodology adopted in computing DEs (see Experimental Section, Computational protocols), prone to basis set incompleteness or superposition errors, ensuring that values computed are essentially converged values with respect to the basis set size is of particular importance. Therefore, from a methodological perspective, we emphasized that using either the cc-pVDZ or the aug-cc-pVDZ variant, the latter at a substantial cost when used on highly brominated systems, should provide converged DEs in conjunction with any of the above functionals.
Table 2. C6H5O–C6H5 bond dissociation enthalpies (in kcal/mol, at 298.15K) computed at different DFT levels
| Exchange-correlation functional | ||||||||
| Basis set | M05 | M05-2X | M06 | M06-2X | B3P86 | PBE0 | mPW1PW91 | wB97xD |
| 6-311G(d) | 73.8 | 81.3 | 77.1 | 78.4 | 80.5 | 79.8 | 78.3 | 79.4 |
| 6-311+G(d) | 73.6 | 81.2 | 76.9 | 78.2 | 80.4 | 79.7 | 78.2 | 79.3 |
| 6-311+G(d,p) | 73.6 | 81.2 | 76.9 | 78.2 | 80.4 | 79.7 | 78.2 | 79.3 |
| 6-311++G(d,p) | 73.6 | 81.2 | 76.9 | 78.2 | 80.4 | 79.7 | 78.2 | 79.3 |
| cc-pVDZ | 74.9 | 82.2 | 77.8 | 79.8 | 81.0 | 80.3 | 78.8 | 80.1 |
| cc-pVTZ | 75.1 | 82.6 | 77.9 | 79.1 | 81.3 | 80.6 | 79.0 | 80.2 |
| aug-cc-pVDZ | 74.9 | 82.6 | 78.0 | 80.1 | 81.2 | 80.5 | 79.0 | 80.3 |
| aug-cc-pVTZ | 74.9 | 82.7 | 77.8 | 79.1 | 81.3 | 80.5 | 79.0 | 80.2 |
Finally, to assess the accuracy of each density functional one needs to compare the computed DEs with the experimental values of either 80 kcal/mol [23] or 80.3 kcal/mol [24] reported for the corresponding C–Br bond and 77.8 kcal/mol [24] for C–O, respectively. Differences between computed and experimental DEs are graphically presented in Figure 1. For the C–Br bond we choose the recent value of 80.3 kcal/mol. One may note the sharp discrepancies of M05 and M05-2X values that underestimate and overestimate, respectively, the experimental quantities with up to 7-8 kcal/mol (10%). In addition, the M05, PBE0 and mPW1PW91 functionals reverse the order of the two bond DEs, B3P86 and M06-2X recover the two bond DEs values with different accuracy, whereas including dispersion corrections (as in wB97xD) improves marginally the computed DEs.
Among the selected functionals, the accuracy required in modeling the two bond DEs (2.5 kcal/mol apart) is achieved by M06 when used with any of the correlation-consistent cc-pVnZ basis sets. M06/cc-pVTZ values are within 0.1-0.2 kcal/mol of their experimental counterparts, but chemical accuracy (< 1 kcal/mol deviation) is achieved already at M06/cc-pVDZ, with a substantially reduced computational effort. Therefore we have further assumed comparable accuracy of the latter in modeling the C–O and C–Br bond energetics in PBDEs.
Table 3. Selected bond lengths (Å), valence and torsion angles (°) in BDE 209, computed at M06/cc-pVDZ level vs. experimental data
| Parameter | Computed | Experimental | ||
| Bond lengths | C1-O | 1.363 | 1.397 | 1.386 |
| C2-Br | 1.888 | 1.894 | 1.879 | |
| C3-Br | 1.885 | 1.887 | 1.877 | |
| C4-Br | 1.887 | 1.901 | 1.885 | |
| Valence angles | C1-O-C1’ | 125.0 | 120.9 | 120.7 |
| C2-C1-O | 123.8 | 125.8 | 124.7 | |
| Torsion angle | C2-C2-O-C1’ | 49.7 | 47.5 | 46.2 |
In Table 3 we compared a subset of internal coordinates from the geometry of BDE 209, optimized at M06/cc-pVDZ level in gas phase, with average experimental (XRD) values from selected references. Excepting the notable underestimation of the C–O bond (-0.2 to -0.3 Å), which may result in overestimated DE, and the 4° overestimation of the C–O–C angle, most of the remaining internal coordinates were reproduced within experimental uncertainties, in spite of the fact that M06 functional is not generally regarded as providing accurate geometries.
In a thermal decomposition scenario, degradation of highly brominated PBDEs may be initiated by either a debromination step, that explain the formation of nona-brominated congeners, the homolysis of C–O bond yielding brominated phenols or both. Energetics of the two pathways, as standard bond DEs computed at M06/cc-pVDZ level, is depicted in Figure 2. Compared to the DE computed for the corresponding bond in bromobenzene, our computations foresee a 12 kcal/mol weaker C–O bond in BDE 209. This may be rationalized on the basis of mutual repulsion between the bromine atoms in position adjacent to the ether group, but also considering the additional stabilization of the resulting brominated radicals compared to the unbrominated analogues [27]. In contrast, C–Br bonds are predicted at only 4-6 kcal/mol lower DEs in BDE 209 than in diphenylether. The computed 5-8 kcal/mol difference between the energetics of the C–Br and C–O bonds may be however underestimated, on the basis of an anticipated overestimation of the latter (see above). Hence, while most of the previous studies emphasize on debromination, photo-oxidation and subsequent intramolecular processes [27–29] in lower brominated PBDEs, results reported herein suggest that thermal decomposition toward polybrominated phenoxyl radicals and polybrominated phenyls is slightly favored in decabrominated diphenyl ether. However, the overall thermal decomposition mechanism may be influenced by several other factors, including radical stabilization or bimolecular steps not considered in our current approach.
Figure 3 depicts C–O and C–Br bonds DEs computed at the same M06/cc-pVDZ level of theory for six PBDE congeners: 2,2’,3,3’,4,4’,6,6’-octabromodiphenyl ether (BDE 197), 2,2’,3,4,4’,5’,6-heptabromodiphenyl ether (BDE 183), 2,2’,4,4’,5,5’-hexabromodiphenyl ether (BDE 153), 2,2’,4,4’,5,6’-hexabromodiphenyl ether (BDE 154), 2,2’,4,4’,5-pentabromodiphenyl ether (BDE 99), 2,2’,4,4’,6-pentabromo-diphenyl ether (BDE 100), 2,2’,4,4’-tetrabromodiphenyl ether (BDE 47) and 2,4,4’-tribromodiphenyl ether (BDE 28). The above mentioned PBDE congeners were selected based on previous literature on their levels reported for indoor dust samples, one of the most important matrices when addressing human exposure to such compounds, while together with BDE 209 they generally consist of more than 95% of the total PBDEs measured in such samples [3,4].
Values predicted for the C–O and C–Br bond DEs may be rationalized not only on the basis of bromine content, but also of the position of bromine atoms on the two rings. The overall trend shows that C–O bonds strengthens with the reduction of bromine content, from less than 70 kcal/mol in BDE 209 (see Figure 2) or about 70 kcal/mol in octa-brominated BDE 197 to more than 77 kcal/mol in the tri-brominated congener BDE 28. However, in asymmetric congeners the two C–O bonds do not follow the same trend. For instance, the homolytic cleavage of the C–O bond leading to 2,4,5-tribromo-phenyl radical is slightly favored to that leading to 2,4,5-tribromo-phenoxyl, the result being most likely related to a different stabilization energy of the two radicals. In the case of C–Br bonds, while the overall variation show a similar trend, the lower the number of bromine atoms, the higher the DE, mutual repulsion between bromine atoms in adjacent positions appears to weaken the corresponding C–Br bond with up to 6 kcal/mol when compared to the anticipated values. Given the smaller differences between C–Br and C–O bond DEs in congeners with lower bromine content it results in comparable strengths for the two bonds. However, regardless of the ordering of the two bond DEs, in deriving a kinetic model for the thermal decomposition of PBDEs one should account for the multiplicity of the C–Br bonds.
Thermal degradation studies on brominated flame retardants
We have investigated the thermal degradation behavior of decabromodiphenyl ether (BDE 209) by simultaneous TG/DTA under various conditions. Due to low volatility, high elution temperatures are needed when applying chromatographic techniques for this class of flame retardants (FR), resulting in thermal degradation affecting analysis results. Kinetic analysis was performed by the non-parametric (NPK) method, and in correlation with evolved gas analysis (EGA) information and molecular modeling results, provided valuable insight into the kinetics and mechanism of BDE 209 thermal degradation [Dumitras et al., Revista de Chimie, 2017]. The conclusions were aimed at designing proper specific GC analytical methods for the selected FRs [Dumitras et al., Acta Chemica Iasi, 2016] and serving for a better estimation of the human exposure to such compounds.
Thermal degradation experiments
The TG/DTG curves for BDE 209 thermal degradation under different atmospheres (nitrogen, helium and air) are presented in Figure 4, while Figure 5 depicts the corresponding DTA recordings. Analyzing the TG/DTG curves it can be seen that thermal degradation of BDE 209 occurs in a single step, with 100% weight loss, regardless of heating rate and atmosphere, with an initial degradation temperature of 310 0C, slightly shifted towards higher values with increasing heating rate (316 0C for 20 K/min).
DTA recordings suggest a partial overlapping of melting (the endothermic peak at 304 ºC, regardless of heating rate) and thermal degradation in the initial stages, followed by a second endothermic degradation step at higher temperatures that shifts with increasing heating rate (438, 456 and 471 ºC for 2, 10 and 20 K/min, respectively). This hypothesis is supported by the fact that the first peak is completely reversible if consecutive heating and cooling cycles are run, as shown in Figure 6. For the specific melting enthalpy a value of 34.6 J/g was found by integrating the DTA melting peak for multiple runs and averaging the results.
Marked asymmetry and an inflexion point are noted in the DTA degradation peak, suggesting a change in the reaction mechanism after maximum degradation rate is attained, without any correspondence in the TG curves. As DTA and TG measurements are based on different fundaments, this apparent disagreement does not necessarily have a physical significance. DTA is much more sensitive than TG and the inflexion could indicate a variation in the degradation mechanism, as it is also observed in DTG (Figure 4) and, as it will be shown later, in the variation of the apparent activation energy with temperature.
NPK analysis
For the kinetic analysis of the TG data, the T(t) and m(T) dependencies provided by the thermo balance were normalized as α(T), where T is sample temperature, t is the time passed since the beginning of the degradation, m is sample mass and α is the conversion degree. The reaction rate is a two-variable function, as it depends on temperature and on conversion degree. The reaction rate dependence on temperature and on the conversion degree along the experimental TG curves is depicted in Figure 7.
After computing the isothermal and isoconversional vectors, the nature of the f(α) and k(T) functions can be investigated directly. They can also be used for fitting in mechanistic kinetic modelling.
Following this procedure, the experimental data in Figure 7 were used to construct the three-dimensional surface representing the reaction rate by thin plate spline multivariate interpolation, as depicted in Figure 8. However, experimental data are available on a restricted range of temperature and conversion degree. To avoid using extrapolated data outside the experimental range, adaptive NPK was used by extracting a set of partially overlapped submatrices that were then submitted to SVD, each submatrix providing a pair of individual vectors (u and v). In all cases the first singular value exceeded 90% of the sum of all sv’s, proving the separability of T and α. Because SVD is unique up to a multiplication scalar, the individual vectors obtained were shifted vertically with respect to each other and they were scaled based on a continuity criterion to construct the isothermal and isoconversional vectors.
The isoconversional vector is presented in the Arrhenius-type plot in Figure 9, as ln(k) vs (RT)-1. Two thermal degradation regimes are made evident, with a turning point at 405 ºC and different apparent activation energies: an initial stage between 310 and 400 ºC, with Ea = 60.8 kJ/mol, followed by a second stage between 410 and 490 ºC, with Ea = 185.9 kJ/mol, each exhibiting remarkable linearity in these coordinates. The transition from one regime to the other is continuous, with a continuously varying apparent activation energy. These observations indicate that competitive reaction pathways coexist, with different contributions to the global kinetics of the degradation, as previously reported for other systems.
Figure 9. Arrhenius-type plot of the isoconversional vector for BDE 209 thermal degradation
The balance between the contributions of parallel reaction paths to global kinetics is expected to shift with temperature since the effective rate constant is a linear combination of the rate constants of the individual pathways, and the temperature dependence of the degradation process could be modeled as a sum of contributions, each of them being of Arrhenius type.
An analysis of the conversion function (Figure 10) reveals similar behavior, with two different reaction regimes separated by a transition range between 0.4 and 0.65 conversion degree values. A detailed analysis of the nature of the isothermal vector would require a mechanistic approach and it was beyond the scope of this study, but some information can be extracted by fitting with the Sestak-Berggren equation. As noted in Figure 10, the initial degradation regime is very close to zero order kinetics, indicating a predominantly heterogeneous process. Non-integer coefficients obtained in the second stage are indicative of complex degradation mechanism.
Figure 10. Isothermal vector for BDE 209 thermal degradation
These findings suggest that in the initial stage evaporation and thermal scission of molecules from the surface with vaporisation of the fragments occurs simultaneously. With increasing temperature an alternate reaction route with higher energy barrier, probably involving intermolecular chain transfer and occurring in the entire mass of the sample, becomes predominant and eventually takes control over the global kinetics.
Identification and quantification of selected flame retardants in indoor dust samples collected from various environments from Eastern-Romania.
In order to evaluate the type of samples and analytes to be considered for analysis there were evaluated previously published results on BFRs, in general, and PBDEs in particular from dust samples collected from various locations worldwide, but also from Europe and Romania. Therefore, all previous studies suggested that for dust samples collected from European countries when PBDEs were targeted for analysis, BDE 209 was the most prominent congener with percent contribution to the total PBDE levels from indoor dust samples ranging from 80.8% and up to 97.9%. The only exceptions from such observations referred to samples collected from other continents, namely America or Australia (detailed information on such review data is presented in Table 4).
Table 4. Percentage contribution of BDE 209 to the total PBDE levels reported for indoor dust samples collected from EU and non-EU countries.
| Locations, Year | BDE 209 | ΣPBDEs | %BDE 209 |
| Romania, 2010 | 275 | 340 | 80.8 |
| Romania, 2006 | 470 | 480 | 97.9 |
| Belgium, 2009 | 313 | 355 | 88.1 |
| UK, 2008 | 2800 | 2919 | 95.9 |
| Canada, 2008 | 560 | 1219 | 45.9 |
| USA, 2008 | 1300 | 2990 | 43.4 |
| USA, 2008 | 1350 | 2980 | 45.3 |
| Australia, 2007 | 291 | 460 | 63.2 |
The following considerations where the most important when selecting the target analytes to be further investigated from samples collected from Iasi, Romania:
i) availability of the standard solutions to be sold by the companies in this field across Romania;
ii) major BDE 209 contribution when compared to other lower brominated congeners from indoor dust samples, especially reported for Romanian samples;
iii) previously reported issues on the thermal stability of BDE 209 when compared to other PBDE congeners targeted for analysis.
As a consequence, given the details presented above, BDE 209 was selected for analysis from dust samples targeted for collection, namely computer dust samples collected mainly from lab-facility across Department of Chemistry, „Alexandru Ioan Cuza” University from Iasi, Romania. Therefore, there were selected for sampling a number of computers, all being located in the same lab-facility, while for comparison purposes there were also selected for sampling a number of computers located as single workstations located mainly in offices from the same Department.
Quantification of selected BFRs from dust samples while modifying analytical protocol parameters.
The analytical methodology to be implemented in order to measure BFRs was based on few sample-preparation and instrumental analysis steps, as follows:
a) Dust sampling and sample pre-treatment
For dust sampling, the methodology implies the use of a vacuum cleaner equipped with a filtration membrane to be attached in front of the vacuum tubing. Therefore, dust from the computer mainboards was vacuumed and collected in filtration membranes, individually for each sampled computer and later transferred to an amber glass recipient and further kept in lab conditions prior to analysis. Later, each sample was sieved on a 500 μm sieve and later weighted and coded. In total, following this procedure there were sampled a number of 17 computers, out of which 11 were located in the students lab-facility Communication Techniques and Programming Software of the Department of Chemistry, „Alexandru Ioan Cuza” University of Iasi, Romania, 5 samples were collected for comparison purposes from office-computers and 1 sample has been previously collected during 2008 as dust from lab-furniture from the same Department. After sieving, samples were weighted being generally within 50 to 100 mg of dust.
Dust samples were weighted in glass recipients and immediately after that they were treated with internal standards, as follows: 5 ng (100 μL of mix standards solution) of 50 pg/μL from ε-HCH (ε-hexachlorocyclohexane) and respectively 10 ng PCB 143 (2,2′,3,4,5,6′-Hexachlorobiphenyl) (100 μL of standards solution of 100 pg/μL). Although, BDE 209 quantification was performed based on PCB 143, the other internal standards were added for future analysis of other contaminants (α-, β-, γ-, δ-HCH, hexachlorobenzene) potentially present in samples (depending on the standards availability) through instrumental re-injecting of the final extracts.
b) Sample extraction
For extraction optimization there were considered two of the most commonly applied methodologies in this research field, namely: classical Soxhlet extraction and ultrasonic/vortex extraction. In order to choose the suitable technique for this study, the extraction solvents were also considered for the optimization process, later being decided to use a mixture of n-hexane:dichloromethane (1:1, v/v). Therefore, all dust samples were processed through a combination of vortex and ultrasonication: three cycles of 2 minutes vortex (Velp, Italy, 1200 rpm) combined with 30 minutes room temperature ultrasonication (ISOLAB, 6 L, 40 KHz, 180W) followed by centrifugation of the extracts (5000 rpm) after each cycle and transferring of the extracts after each extraction in a separate tube with combining of the organic phase together for each sample. The total extraction volume after following this extraction procedure was initially reduced by the use of a rota-evaporator and later using a concentrator/evaporator under a gentle nitrogen flow (Thermo Scientific, Reacti-Vap™ Evaporator) operated at 25 degrees Celsius. Thus, the extract was reduced to below 3 mL, the target analytes being concentrated at this stage in n-hexane.
c) Clean-up of the extracts
Due to the non-polar nature of the extraction solvents, other compounds might also be co-extracted from collected samples contributing to the increasing of the S/N ratio and interfering with the target analytes for this study. Therefore, there was applied a clean-up step on the obtained extracts from previous stage of the analysis protocol. The clean-up of the extracts was based on using column chromatography (3 mL, propylene, Supelco) filled with a layer of acidified silicagel (44% sulphuric acid, Merck), inferior layer, and anhydrous sodium sulphate on top. The extracts were applied on the clean-up cartridges and later analytes were eluted with 10 mL mixture of n-hexane:dichloromethane (1:1, v/v). Later the obtained extracts were again concentrated to near dryness and redissolved in 100 μL iso-octane. Prior to instrumental analysis, the extracts were filtered on Eppendorf filtration cartridges (20 μm pore-size) followed by centrifugation, 7000 rpm. The final extracts were transferred to injection vials in limited volume inserts of 250 μL.
Image 1. Columns prepared for clean-up of the obtained extracts.
Image 2. Vacuum manifold system used for extracts clean-up and filtration.
Both standard solutions and final extracts obtained after sample preparation procedures were injected under optimized conditions in two gas-chromatographic systems with mass spectrometer detection. The instrumental characteristics and operation parameters used for analysis are given in Tables 5 and 6.
Table 5. Characteristics of the instruments used for identification/quantification of BDE 209 and PCB 143
| Parameter |
Characteristics |
|
| Chromatographic system | Gas-chromatograph Agilent 6890 | Gas-chromatograph QT-2010 Shimadzu |
| Detector | Mass spectrometer, HP 5973Mass analyzer: quadrupole | Mass spectrometerMass analyzer: quadrupole |
| Injector | Standard split/split-less | Standard split/split-less |
| Chromatographic column | AT-5, Length – 30 m, Diameter – 0.25 μm, df – 0.25 μm | AT-5, Length – 30 m, Diameter – 0.25 μm, df – 0.25 μm |
Table 6. Gas chromatograph operating conditions for identification/quantification of BDE 209 and PCB 143
| Parameter | Characteristics | |
| Injector | Temperature (oC) | 280 |
| Operating mode | Spit-less | |
| Chromatographic columnColumn oven | Initial temperature (oC) | 90 |
| Ramp (oC/min) | 10 | |
| Final temperature | 310 | |
| Hold time final temperature | 10 min | |
| Detector | Operating mode | Full-scan m/z: 70-900 |
| Operating mode | SIM: m/z=207, 800, 360, 290 | |
| Mass analyzer temperature | 300 | |
Standard solutions of each tested analyte, namely BDE 209 (concentration of 1.25 ng/μL), ε-HCH (50 pg/µL), PCB 143 (100 pg/µL) where injected in GC systems with detector operated in full-scan in order to select the proper retention time, identification/quantification ions for each compound. Standard solution of BDE 209 were prepared starting with solid analyte of 99.9% purity (Dr. Ehrenstorfer Laboratories, Augsburg, Germany) and later consequently diluted up to the required concentration level. In order to quantify the target analyte from extracted dust samples, a calibration set of solutions were prepared based on data presented in Table 7.
Table 7. Standard solutions used for preparing the calibration curve (internal calibration based on PCB 143)
| Nr. crt. | Cstd. stoc (ng/µL) | Vstd (µL) | mstd (ng) | Vis (µL) | mis (ng) | Mass Ratio(manalyte/mIS) |
| 1 | 1 | 50 | 50 | 100 | 10 | 5 |
| 2 | 1 | 100 | 100 | 100 | 10 | 10 |
| 3 | 5 | 100 | 500 | 100 | 10 | 50 |
| 4 | 50 | 50 | 2500 | 100 | 10 | 250 |
| 5 | 50 | 100 | 5000 | 100 | 10 | 500 |
Calibration solutions and final extracts of the dust samples were injected into GC-MS system operated accordingly with conditions included in Table 6 and later the peaks were manually integrated in order to extract the areas of each signal associated with target analytes and later by the use of the internal calibration equation the analyte mass was calculated. Using the processed sample mass there could be obtained the analyte concentrations expressed in ng/g of sample.
Method applicability for determination of BDE 209 from dust samples
The above described protocol was applied for all extracts obtained after preparation of the collected dust samples, but in order to quantify the BDE 209 there were applied two different approaches: with and without considering the presence of PCB 143 with role of internal standard and thus with and without corrections for the analyte loss during sample preparations procedures.
The obtained results on the BDE 209 levels (ng/g sample) are given in Table 8.
Table 8. BDE 209 concentrations (ng/g sample) from dust samples collected from working stations (computers) sampled from the Department of Chemistry, UAIC, Romania
| Sample type | Sample code | BDE 209 concentration (internal calibration based on PCB 143) | BDE 209 concentration (external calibration) |
| Working station from computer Lab-facility, UAIC | D-01 | 124.0 | 151.3 |
| D-02 | 4562.0 | 5422.1 | |
| D-03 | 1902.2 | 2193.2 | |
| D-04 | 3071.5 | 3704.4 | |
| D-05 | 1277.2 | 1501.6 | |
| D-06 | 385.2 | 472.2 | |
| D-07 | 541.6 | 654.9 | |
| D-08 | 713.6 | 810.4 | |
| D-09 | 302.0 | 334.4 | |
| D-10 | 992.0 | 1203.7 | |
| D-11 | 1283.4 | 1450.1 | |
| Office Computer | B-01 | 150.4 | 184.5 |
| B-02 | 936.2 | 1120.4 | |
| B-03 | 1040.6 | 1272.8 | |
| B-04 | 315.0 | 367.9 | |
| B-05 | 1501.9 | 1706.3 | |
| Lab of Analytical Chemistry, 2008 | A-01 | 453.2 | 522.8 |
| Median concentration (ng/g) | 936.2 | 1120.4 | |
| Range (min-max) | 124 – 4562 | 151.3 – 5422.1 |
Estimation of the human exposure to BDE 209 through dust ingestion
In order to estimate the human exposure to various contaminants through dust ingestion, the existing exposure models are based on some assumptions which were considered also during the following of this study. One of these assumptions is based on the fact that 100% of the analytes contained by dust are adsorbed at the gastro-intestinal tract; therefore they are assumed to be 100% bioavailable.
Table 9. Estimation of the BDE 209 human (adults) exposure (ng/day and ng/kg bw/day respectively) through dust ingestion
| Adults(ng/day) | Adults(ng/kg bw/day) | |||
| Average dust exposure(20 mg/day) | High dust exposure(50 mg/day) | Average dust exposure(20 mg/day) | High dust exposure(50 mg/day) | |
| D-01 | 2.5 | 6.2 | 0.04 | 0.1 |
| D-02 | 91.2 | 228.1 | 1.30 | 3.3 |
| D-03 | 38.0 | 95.1 | 0.54 | 1.4 |
| D-04 | 61.4 | 153.6 | 0.88 | 2.2 |
| D-05 | 25.5 | 63.9 | 0.36 | 0.9 |
| D-06 | 7.7 | 19.3 | 0.11 | 0.3 |
| D-07 | 10.8 | 27.1 | 0.15 | 0.4 |
| D-08 | 14.3 | 35.7 | 0.20 | 0.5 |
| D-09 | 6.0 | 15.1 | 0.09 | 0.2 |
| D-10 | 19.8 | 49.6 | 0.28 | 0.7 |
| D-11 | 25.7 | 64.2 | 0.37 | 0.9 |
| B-01 | 3.0 | 7.5 | 0.04 | 0.1 |
| B-02 | 18.7 | 46.8 | 0.27 | 0.7 |
| B-03 | 20.8 | 52.0 | 0.30 | 0.7 |
| B-04 | 6.3 | 15.7 | 0.09 | 0.2 |
| B-05 | 30.0 | 75.1 | 0.43 | 1.1 |
| A-01 | 9.1 | 22.7 | 0.13 | 0.3 |
| 5% | 2.9 | 7.3 | 0.04 | 0.1 |
| median | 18.7 | 46.8 | 0.27 | 0.7 |
| 95% | 67.4 | 168.5 | 0.96 | 2.4 |
Table 10. Estimation of the BDE 209 human (toddlers) exposure (ng/day and ng/kg bw/day respectively) through dust ingestion
| Toddlers(ng/day) | Toddlers(ng/kg bw/day) | |||
| Average dust exposure (50 mg/day) | High dust exposure (200 mg/day) | Average dust exposure (50 mg/day) | High dust exposure (200 mg/day) | |
| D-01 | 6.2 | 24.8 | 0.3 | 1.0 |
| D-02 | 228.1 | 912.4 | 9.5 | 38.0 |
| D-03 | 95.1 | 380.4 | 4.0 | 15.9 |
| D-04 | 153.6 | 614.3 | 6.4 | 25.6 |
| D-05 | 63.9 | 255.4 | 2.7 | 10.6 |
| D-06 | 19.3 | 77.0 | 0.8 | 3.2 |
| D-07 | 27.1 | 108.3 | 1.1 | 4.5 |
| D-08 | 35.7 | 142.7 | 1.5 | 5.9 |
| D-09 | 15.1 | 60.4 | 0.6 | 2.5 |
| D-10 | 49.6 | 198.4 | 2.1 | 8.3 |
| D-11 | 64.2 | 256.7 | 2.7 | 10.7 |
| B-01 | 7.5 | 30.1 | 0.3 | 1.3 |
| B-02 | 46.8 | 187.2 | 2.0 | 7.8 |
| B-03 | 52.0 | 208.1 | 2.2 | 8.7 |
| B-04 | 15.7 | 63.0 | 0.7 | 2.6 |
| B-05 | 75.1 | 300.4 | 3.1 | 12.5 |
| A-01 | 22.7 | 90.6 | 0.9 | 3.8 |
| 5% | 7.3 | 29.0 | 0.3 | 1.2 |
| median | 46.8 | 187.2 | 2.0 | 7.8 |
| 95% | 168.5 | 673.9 | 7.0 | 28.1 |
Additionally, the human exposure scenarios to dust ingestion are based on the existence of two accepted models of exposure for both adults and toddlers, namely:
– one scenario of average human exposure to dust which considers that an adult (70 kg) ingests 20 mg of dust, while a toddler (6-24 months old, 24 kg) ingests 50 mg of dust daily;
– a second scenario of high exposure to dust which is based on the following exposure figures: an adult (70 kg) ingests 50 mg of dust, while a toddler (6-24 months old, 24 kg) ingests 200 mg of dust daily.
Table 11. Estimation of the differences recorded for the human (adults) exposure scenarios to BDE 209 through dust ingestion due to instrumental protocol modification
| Adults(ng/day) | Adults(ng/kg bw/day) | |||
| Average dust exposure(20 mg/day) | High dust exposure(50 mg/day) | Average dust exposure(20 mg/day) | High dust exposure(50 mg/day) | |
| D-01 | 0.5 | 1.4 | 2.1 | 2.1 |
| D-02 | 17.2 | 43.0 | 76.2 | 74.2 |
| D-03 | 5.8 | 14.6 | 30.8 | 30.0 |
| D-04 | 12.7 | 31.6 | 52.0 | 50.7 |
| D-05 | 4.5 | 11.2 | 21.1 | 20.5 |
| D-06 | 1.7 | 4.4 | 6.6 | 6.5 |
| D-07 | 2.3 | 5.7 | 9.2 | 9.0 |
| D-08 | 1.9 | 4.8 | 11.4 | 11.1 |
| D-09 | 0.6 | 1.6 | 4.7 | 4.6 |
| D-10 | 4.2 | 10.6 | 16.9 | 16.5 |
| D-11 | 3.3 | 8.3 | 20.3 | 19.8 |
| B-01 | 0.7 | 1.7 | 2.6 | 2.5 |
| B-02 | 3.7 | 9.2 | 15.7 | 15.3 |
| B-03 | 4.6 | 11.6 | 17.9 | 17.4 |
| B-04 | 1.1 | 2.6 | 5.2 | 5.0 |
| B-05 | 4.1 | 10.2 | 23.9 | 23.3 |
| A-01 | 1.4 | 3.5 | 7.3 | 7.1 |
| 5% | 0.7 | 1.6 | 2.5 | 2.4 |
| median | 3.7 | 9.2 | 15.7 | 15.3 |
| 95% | 13.6 | 33.9 | 56.9 | 55.4 |
Table 12. Estimation of the differences recorded for the human (toddlers) exposure scenarios to BDE 209 through dust ingestion due to instrumental protocol modification
| Toddlers(ng/day) | Adults(ng/kg bw/day) | |||
| Average dust exposure (50 mg/day) | High dust exposure (200 mg/day) | Average dust exposure(20 mg/day) | High dust exposure(50 mg/day) | |
| D-01 | 1.4 | 5.5 | 2.1 | 2.1 |
| D-02 | 43.0 | 172.0 | 76.2 | 74.2 |
| D-03 | 14.6 | 58.2 | 30.8 | 30.0 |
| D-04 | 31.6 | 126.6 | 52.0 | 50.7 |
| D-05 | 11.2 | 44.9 | 21.1 | 20.5 |
| D-06 | 4.4 | 17.4 | 6.6 | 6.5 |
| D-07 | 5.7 | 22.7 | 9.2 | 9.0 |
| D-08 | 4.8 | 19.4 | 11.4 | 11.1 |
| D-09 | 1.6 | 6.5 | 4.7 | 4.6 |
| D-10 | 10.6 | 42.3 | 16.9 | 16.5 |
| D-11 | 8.3 | 33.3 | 20.3 | 19.8 |
| B-01 | 1.7 | 6.8 | 2.6 | 2.5 |
| B-02 | 9.2 | 36.8 | 15.7 | 15.3 |
| B-03 | 11.6 | 46.4 | 17.9 | 17.4 |
| B-04 | 2.6 | 10.6 | 5.2 | 5.0 |
| B-05 | 10.2 | 40.9 | 23.9 | 23.3 |
| A-01 | 3.5 | 13.9 | 7.3 | 7.1 |
| 5% | 1.6 | 6.5 | 2.5 | 2.4 |
| median | 9.2 | 36.8 | 15.7 | 15.3 |
| 95% | 33.9 | 135.7 | 56.9 | 55.4 |
Based on the above mentioned exposure scenarios and on measured levels of BDE 209 from collected dust samples the estimation of the human exposure to BDE 209 through dust ingestion data is presented in Tables 9 and 10 (adults and toddlers) and differences recorded for such exposure scenarios based on analytical protocol modification are presented in Tables 11 and 12 (adults and toddlers).
Published/Submitted research articles
1. Dan Maftei, Mihai Dumitras, Dragos L. Isac, Alin C. Dirtu. Density functional study of bond dissociation energies in highly brominated diphenyl ethers. Studia UBB Chemia, LXI, 4, 137-146, 2016. (http://chem.ubbcluj.ro/~studiachemia/chemia2016_4.html)
2. Mihai Dumitras, Dan Maftei, Dragos L. Isac, Anton Airinei, Alin C. Dirtu. Thermal degradation study of decabromodiphenyl ether. Translating thermo-analytical results into optimal chromatographic conditions. Acta Chemica Iasi, 24 (2), 76-87, 2016. (https://www.degruyter.com/view/j/achi.2016.24.issue-2/issue-files/achi.2016.24.issue-2.xml)
3. Mihai Dumitras, Dan Maftei, Anton Airinei, Nita Tudorachi, Alin C. Dirtu. NPK analysis of the thermal degradation of decabromodiphenyl ether. Revista de Chimie, 68 (11), 2017 – article in press. (http://www.revistadechimie.ro/)
4. Mihai Dumitras, Liviu Leontie, Dan Maftei, Alin C. Dirtu. Kinetic analysis of the thermal degradation of two brominated flame retardants. Journal of Thermal Analysis and Calorimetry, 2017 – article submitted. (https://link.springer.com/journal/10973)
5. Dan Maftei, Dragos L. Isac, Mihai Dumitras, Stefan Bucur, Alin C. Dirtu. Trends in Bond Dissociation Energies of Brominated Flame Retardants from Density Functional Theory. Structural Chemistry, 2017 – article submitted. (https://link.springer.com/journal/11224)
Conference participation
1. Alin C. Dirtu, Dan Maftei, Mihai Dumitras, Mirela Suchea, Adrian Covaci. Influence of the chromatographic analysis of flame retardants on the estimation of human exposure to organohalogenated compounds. Nanoscience in Chemistry, Physics, Biology and Mathematics – NanoMathChem, 12-14 November 2015, Cluj, Romania. (http://www.esmc.ro/#!nanomathchem2015/c6ve)
2. Alin C. Dirtu, Mihai Dumitras, Dan Maftei, Anton Airinei, Nita Tudorachi, Dragos L. Isac. Thermal degradation study of brominated flame retardants with emphasis on the estimation of human exposure through indoor dust ingestion. 12th International Conference on Colloid and Surface Chemistry – ICCSC, 16-18 May 2016, Iasi, Romania. (https://iccsc2016.wordpress.com/)
3. Alin C. Dirtu, Dan Maftei, Mihai Dumitras, Adrian Covaci. Challenges in gas-chromatographic analysis of thermal unstable compounds. Case study for brominated flame retardants. 3rd International Conference on Analytical Chemistry – RO-ICAC, 28-31 August 2016, Iasi, Romania. (http://roicac.acadiasi.ro/)
4. Dan Maftei, Mihai Dumitras, Dragos L. Isac, Alin C. Dirtu. Thermal degradation pathways of selected brominated flame retardants from first principles. International Conference of Physical Chemistry – ROMPHYSCHEM, 21-23 September 2016, Galati, Romania. (http://gw-chimie.math.unibuc.ro/romphyschem/)
5. Mihai Dumitras, Dan Maftei, Anton Airinei, Alin C. Dirtu. NPK analysis of the thermal degradation of selected brominated flame retardants. International Conference of Physical Chemistry – ROMPHYSCHEM, 21-23 September 2016, Galati, Romania. (http://gw-chimie.math.unibuc.ro/romphyschem/)
6. Mihai Dumitras, Dan Maftei, Dragos L. Isac, Alin C. Dirtu. Study on the kinetics and mechanism of decabromo-diphenyl ether thermal degradation. XXXIV-th Romanian Chemistry Conference, 04-07 October 2016, Calimanesti-Caciulata, Valcea, Romania. (http://conferinta.oltchim.ro/)
7. Dan Maftei, Mihai Dumitras, Dragos L. Isac, Alin C. Dirtu. Computational thermochemistry of brominated flame retardants. Case study of decabromo-diphenyl ether. XXXIV-th Romanian Chemistry Conference, 04-07 October 2016, Calimanesti-Caciulata, Valcea, Romania. (http://conferinta.oltchim.ro/)
8. Dragos L. Isac, Mihai Dumitras, Dan Maftei, Anton Airinei, Alin C. Dirtu. Study on the thermal degradation of brominated flame retardants: computational thermochemistry, kinetics and degradation mechanism. “Alexandru Ioan Cuza” University Days, Faculty of Chemistry Conference, 27-28 October 2016, Iași, România. (http://www.chem.uaic.ro/ro/manifestari/program-zu-2016.html)
9. Alin C. Dirtu, Mihai Dumitras, Dan Maftei, Adrian Covaci. Challenges in brominated flame retardants analysis in dust samples from personal computers. Case study in a university campus from Eastern Romania. 19th International Conference – Materials, Methods & Technologies, 26-30 June 2017, Elenite, Bulgaria. (https://www.sciencebg.net/en/conferences/materials-methods-and-technologies/)
10. Dan Maftei, Dragos L. Isac, Stefan Bucur, Alin C. Dirtu. Bond dissociation enthalpies in selected brominated flame retardants from density functional theory. 19th International Conference – Materials, Methods & Technologies, 26-30 June 2017, Elenite, Bulgaria. (https://www.sciencebg.net/en/conferences/materials-methods-and-technologies/)
11. Mihai Dumitras, Dan Maftei, Dragos L. Isac, Alin C. Dirtu. Comparative study of the thermal degradation of two brominated flame retardants. 19th International Conference – Materials, Methods & Technologies, 26-30 June 2017, Elenite, Bulgaria. (https://www.sciencebg.net/en/conferences/materials-methods-and-technologies/)
12. Stefan Bucur, Dragos L. Isac, Dan Maftei, Mihai Dumitras, Alin C. Dirtu. New insights into brominated flame retardants structure: computational assessment of σ-hole effect. Scientific Communications Session for Bachelor, Master of Science and PhD Students, 30 June 2017, Faculty of Chemistry, “Alexandru Ioan Cuza” University of Iasi, Romania (http://www.chem.uaic.ro/ro/manifestari/program-scsmd-2017.html).
13. Dragos-Lucian Isac, Stefan Bucur, Dan Maftei, Mihai Dumitras, Alin C. Dirtu. Aspecte teoretice privind structurile chimice ale compusilor ignifugi bromurati. Metode noi de evidentiere a efectelor regiunii σ si π. Zilele Academice Ieșene (ZAI), 5 – 6 Octombrie 2017, Iași, România (http://www.icmpp.ro/zai/program.html).
14. Mihai Dumitras, Dan Maftei, Dragos L. Isac, Alin C. Dirtu. NPK analysis of the thermal degradation of two brominated flame retardants. Colloque Franco-Roumain de Chimie Médicinale (CoFr-RoCM) – 4th Edition, 5-7 October 2017, Iasi, Romania (http://www.chem.uaic.ro/cofrrocm-2017/).
Recent Comments