Reply to Materić: Appropriate blanks should avoid major contamination sources in the lab

We agree with Materić that nanoplastics measurements must have appropriate blanks (1). However, we disagree with his conclusion that “Milli-Q water was a procedural blank that should have been used against the other bottled water samples” (1, 2).

Milli-Q water was not and should not have been used as a procedure blank in our original study (2). Milli-Q water was found to be contaminated with plastic particles itself (2) and contributed to up to half of the total plastic contaminants (3), which was also cited by Materić (1). This is consistent with our results: Milli-Q water has comparable plastic particles than all three brands of bottled water. When comparing Milli-Q water with Brand B, the polymer composition profile of micro- and nano-plastics (MNPs) differs significantly (Fig. 1), where Milli-Q water has more MNP counts than bottled water for certain plastic types. Therefore, Materić’s suggestion of using Milli-Q water to prepare the procedure blank and estimate the detection limit is invalid.

Fig. 1.

Parallel comparison of results from Brand B and Milli-Q based on the MNP quantification analysis using fluorescent polystyrene (PS) beads as the internal standard. From Left to Right, the relative number abundance of polyethylene terephthalate (PET), polyvinyl chloride (PVC), and polyamide (PA) plastic particles compared to the added PS beads. Error bars, mean ± SEM. Data are originated from SI Appendix, Fig. S20 (2). P values are calculated from generalized linear mixed model analysis.

The blank utilized in our study provided better control over the potential contamination sources than Milli-Q water (2). In our preparation procedure (2), only two steps were involved: sample filtration and filter preparation for stimulated Raman scattering (SRS) imaging. Bottled water was directly filtered, contacting only the air and the thoroughly cleaned glassware from chromic acid and excessive sample water, making airborne particles the possible contamination source to control. Airborne particles contributed to minimal contamination (3) and were controlled by the blank Anodisc filters exposed to the same lab air. PET has been found to be the dominant atmospheric polymer indoors (4), and the fact that no PET particles were detected in the Milli-Q water sample also supports that airborne plastic contamination is minimal in our procedure (Fig. 1). In the filter preparation step, the Anodisc filters were prepared “in the same way as the bottled water samples” for SRS imaging (3), providing control for all contamination sources.

Our measurements provided thoroughly characterized results. Three bottled water brands possessed a distinct profile of plastic composition (2), excluding the possibility of a common lab-based contamination. The prevalent detection of polyamides is consistent with many other bottled water studies (5, 6), possibly from polyamide-based reverse osmosis (2, 7). Moreover, our results on the microplastic abundances are comparable to previous literature using different techniques (2, 5, 6). Finally, the PET particles detected in the bottled water samples, rather than Milli-Q, support PET packaging release as one contamination source.

Although, ideally, certified clean water would be used if available, we adopted the blank method following the common US Environmental Protection Agency practice—exposing the filters to the same environment without drawing air or water through (8). This serves as an appropriate blank for simple procedures such as filtration. However, when analyzing complicated samples needing more processing and additional reagents, developing plastic-clean water is necessary. As MNPs are ubiquitous, it is vital to examine each experimental component. Blindly adopting any “clean” water, equipment, or reagents in the analytical procedure will also mislead the field from better understanding MNPs in our environments.