A restricted study organised by the Joint Research Centre
The Joint Research Centre (JRC - 12) is leading the development of an OECD Test Guideline (TG) for the determination of the volume specific surface area (VSSA) of manufactured nanomaterials. For this purpose, JRC is collaborating with a group of experts for the preparation, execution and evaluation of preparatory work, including a collaborative study, which will be taken into consideration in the preparation of the final draft test guideline. The final draft TG will be presented for discussion and eventual approval to the OECD Working Group of the National Coordinators of the Test Guidelines Programme.
The JRC organized and evaluated an interlaboratory comparison (ILC) on the determination of VSSA of nanomaterials by gas adsorption according to the Brunauer, Emmet and Teller (BET) method and gas pycnometry, with the participation of seven laboratories. The ILC was launched in May 2020 and concluded in November 2020 by presenting the evaluation of the results in a dedicated meeting of the group of experts. The ILC was designed as a semi-blind exercise with codes randomly attributed to samples and each of the seven participating laboratories.
The set of seven selected test materials included inorganic (metal oxides – zinc oxide, and two types of titanium) and carbon based materials (graphene and a multiwall carbon nanotube (MWCNT)) as well as one organic pigment. The selected materials also cover various shapes with one, two and three dimensions in the nanoscale (e.g. fibre, platelets and particles) and a range of size distributions and aggregation. Most of the selected test materials were subsampled in vials of 1 gram in the JRC Nanomaterials Repository facility.
Seven laboratories, from France, Germany, Italy, Spain, UK, Korea and the JRC-Ispra, took part in this study by analysing the test materials according to a defined ILC matrix that was constructed to decrease the resources needed per single laboratory. Hence, one of the materials was analysed by all laboratories and the rest of materials analysed by a minimum of three laboratories. The seven laboratories have a varying degree of experience with the techniques.
The seven laboratories reported results for the three measurands considered, i.e. (mass) specific surface area (SSA), density and VSSA. These data have been statistically evaluated by robust statistics according to the principles laid down in ISO 5725-5.
Overall results show good repeatability with relative standard deviation for repeatability (RSDr) values less than 2% for all materials except for the organic pigment with RSDr < 5%. The variability between laboratories was higher, with overall relative standard deviation for reproducibility (RSDR) values less than 20% for all materials and measurands, in particular RSDR values less than 10% were obtained for the analysis of four materials (graphene, MWCNT, zinc oxide and one type of titanium dioxide) while a higher variability, with RSDR values in the 10% - 20% range, was obtained in the analysis of the organic pigment, the fumed silica and one type of titanium dioxide.
The repeatability and reproducibility values worsen when all replicate measurements both for SSA and skeletal density are considered in the determination of the VSSA. Overall, the RSDr is less than 6.5 % and the RSDR is less than 20% for all tested materials. In this study, a higher variability of the within-laboratory measurement of the skeletal density was observed compared to the SSA determination, in particular for one of the seven laboratories that have reported relative standard deviations in the 5 - 20% range for skeletal density measurements while the values for SSA measurements fall in the 2% to 5% interval.
In the literature, there is limited information about ILCs on the determination of SSA and what exists is mainly related to titanium dioxide. Data on ILCs for density determination by gas pycnometry is even scarcer. This fact prevents a comprehensive appraisal, but, in general, when comparing the data obtained in this study with other ILCs, the repeatability for the determination of SSA is similar to the values reported in the few published ILCs (RSDr < 3.5%) and also to those obtained during the certification of some titanium dioxide reference materials (RSDr < 1.5%). With reference to the between laboratory variability associated to the VSSA determination in the present ILC, some laboratories deviated from the Standard Operating Procedure (SOP) mainly with regard to sample amount and degassing conditions. These might have had an influence in the observed variability between laboratories and prompt a revision of the SOP and consideration of a potential training phase for laboratories with less experience, which could lead to improved reproducibility. Concerning the bias of the method, for SSA it was observed that the deviations of the overall robust average from the reference value (where they exist), is in the range from 0.7% to 4.7% depending on the specific material. A similar observation was made for the skeletal density. In the case of the density, there is also a good match but an accurate evaluation is hampered by the fact that the measured value corresponds to the skeletal density while the theoretical value does not specify if it is referring to bulk density or skeletal density.
This study involved only a limited number of datasets per type of material. For this reason, in this study the evaluation of the results was carried out using robust statistics. Having a higher number of data sets per type of material would reinforce the statistical outcome. Regarding the aim of the ILC to prove that engineered nanomaterials are part of the general applicability domain, it is worth noting that organic materials were underrepresented, as were other types of material such as microporous materials or materials with very high specific surface area.
It can be concluded that state-of-the-art repeatability standard deviations have been obtained in this ILC for the analysis of all selected inorganic materials (RSDr < 2%), and also the value for the analysis of the organic (RSDr < 5%) may be considered acceptable. On the other hand, the reproducibility standard deviations are higher, with values between 4% and 20%. This indicates that some improvement is desirable and probably could be obtained by including additional details in the SOP on aspects related to the proper sample amount and degassing conditions for the analysis of the materials, especially for organic materials. Exchange of information on best practices among laboratories and some training activities for less experienced laboratories would probably also be beneficial and lead to a more reliable implementation of the method.
BARRERO Josefa;
SENALDI Chiara;
LA SPINA Rita;
RIEGO SINTES Juan;
2021-06-04
Publications Office of the European Union
JRC124644
978-92-76-37685-9 (online),
978-92-76-37684-2 (print),
1831-9424 (online),
1018-5593 (print),
EUR 30702 EN,
OP KJ-NA-30702-EN-N (online),
OP KJ-NA-30702-EN-C (print),
https://publications.jrc.ec.europa.eu/repository/handle/JRC124644,
10.2760/41115 (online),
10.2760/08026 (print),
