**5. Results**

All the laboratories processed the data as follows:


#### **5.1 Data analysis and visualization**

After the collection of the data from each participant, the results were statistically elaborated. Actually, graphical descriptive analysis was used to compare the data obtained from the laboratories. They were not always homogenous as sometimes diverse working conditions were applied, and in few cases, some data were missing or incomplete.

#### **5.2 Available data**

**Table 3** shows all the missing data in red. Eight laboratories could not perform the study due to several drawbacks, such as instrumental breakdown or unavailability of filtering systems.

#### **5.3 Rate calculation**

It was required to report all the microplastic materials found in each replicate. Since many microplastics not belonging to standard microfilaments were found in the suspensions, the recovery rate was evaluated by dividing the microplastics found in each sample by the number of standard microfilaments. This number was interpreted as the fraction of the total standard microfilaments over their theoretical quantity.

#### **5.4 Recovery rates**

**Figure 9** shows the fraction of standard microfilaments found in each sample (Sample 1, Sample 2, and Sample 3) and for each replicate [1–3] obtained from all the different laboratories.


#### **Table 3.**

*Data collected from each laboratory. Red (missing data).*

Only the relative fraction of the target material identified was calculated, and the absolute values were not reported.

If no evident problems of contamination occurred, the fraction of standard microfilaments was just above 0.5. However, some significant differences were found between the laboratory data. Lab 1 did not obtain accurate data for Sample 3 because during the filtration of the 1st replicate of the 3rd sample, the filter broke. The data of Lab 2 were not processable for each reference sample due to the incorrect acquisition mode of the spectra of the microplastics collected on the chosen filter. Moreover, the use of an automatic analysis system and the lack of a good reference spectral library could have amplified the mistakes. Lab 4 did not find any microplastic fibers in Sample 1 for all the replicates due to problems of working conditions during micro-FTIR automatic analysis identification, as indicated by the lab itself, for example,. selection of brightness and contrast. In fact, in this case, Lab 4 obtained data only for PP particle contaminants and not for microplastic fibers. This drawback determined a lower fraction target for all other samples Labs 6, 11 (replicate 1 of sample 2), and 16 had a recovery rate above 1, which means that they found a number of standard microfilaments higher than the present quantity. Laboratory 10 performed only one replicate for each sample and did not report anything different from the target material. Laboratory 17 did not carry out the analysis on two of the three samples and reported only the data of replicate 3 of Sample 2. The lack of results did not depend on the quality of the standard samples.

Labs 3, 7, 9, 11, and 13 reached good recovery rates for all the three samples analyzed. Labs 7 and 9 reached a recovery rate of 80–90%. The results highlighted that the labs that followed all the steps in the standard protocol reached the target fraction, confirming the quality and reproducibility of the method.

Moreover, the analysis of the results highlighted that the μ-FTIR and μ-Raman techniques also allow the identification, counting, and monitoring of the pollutants and their sources of contamination.

*Round Robin Test on Microplastic Counting and Identification Method DOI: http://dx.doi.org/10.5772/intechopen.109757*

**Figure 9.** *Fraction of target material identified in the samples.*

The main contaminants identified in analyzed samples were polypropylene (particles), natural fibers deriving from the storage system, sample preparation, and laboratory environment.

In particular, polypropylene particles were found by all the labs as a result of the containers and caps used in sample preparation. The presence of cotton fibers was identified, but it was not possible to understand whether it depended on demineralized water or dissolved wool tops.

However, when the nature of contamination was different, the source was the result of incorrect procedure. For example, one lab carried out additional sample manipulation before IR analysis, thus increasing the risk of microplastic loss and contamination (**Figure 10**).

Finally, a few participants did not perform the dimensional analysis of the microplastics identified in all the samples as suggested in the standard protocol. However, some dimensional data showed that the identified microplastic fibers were classified in the length range between 100 μm and 500 μm and diameters between 40 μm and 66 μm, confirming the control data.

**Figure 10.** *Fraction of contaminant material identified in the laboratories.*
