Chemosphere 298 (2022) 134282

Contents lists available at ScienceDirect

Chemosphere
journal homepage: www.elsevier.com/locate/chemosphere

Monitoring microplastics in drinking water: An interlaboratory study to

inform effective methods for quantifying and characterizing microplastics
Hannah De Frond a, *, Leah Thornton Hampton b, Syd Kotar b, Kristine Gesulga b, Cindy Matuch b,
Wenjian Lao b, Stephen B. Weisberg b, Charles S. Wong b, Chelsea M. Rochman a, **
a
Department of Ecology & Evolutionary Biology, University of Toronto, 25 Willcocks Street, Room 3055, Toronto, Ontario, M5S 3B2, Canada
b
Southern California Coastal Water Research Project Authority, 3535 Harbor Blvd., Suite 110, Costa Mesa, CA, 92626, USA

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• A novel study measured the perfor­

mance of common methods for micro­
plastic analysis.
• Variation in recovery was likely due to
experience and particle contamination.
• Microscopy is an accurate method for
counting plastic particles down to 50
μm.
• Both FTIR and Raman spectroscopy
accurately identified spiked plastic
particles.
• Raman spectroscopy may be more suit­
able for particles smaller than 20 μm. .
• Methods to reduce time and identify
smaller particles need further develop­
ment. .

A R T I C L E I N F O A B S T R A C T

Handling Editor: Derek Muir California Senate Bill 1422 requires the development of State-approved standardized methods for quantifying
and characterizing microplastics in drinking water. Accordingly, we led an interlaboratory microplastic method
Keywords: evaluation study, with 22 participating laboratories from six countries, to evaluate the performance of widely
Microplastic used methods: sample extraction via filtering/sieving, optical microscopy, FTIR spectroscopy, and Raman
Method
spectroscopy. Three spiked samples of simulated clean water and a laboratory blank were sent to each laboratory
Monitoring
with a prescribed standard operating procedure for particle extraction, quantification, and characterization. The
Spectroscopy
Microscopy samples contained known amounts of microparticles within four size fractions (1–20 μm, 20–212 μm, 212–500
μm, >500 μm), four polymer types (PE, PS, PVC, and PET), and six colors (clear, white, green, blue, red, and
orange). They also included false positives (natural hair, fibers, and shells) that may be mistaken for micro­
plastics. Among laboratories, mean particle recovery using stereomicroscopy was 76% ± 10% (SE). For particles
in the three largest size fractions, mean recovery was 92% ± 12% SD. On average, laboratory contamination
from blank samples was 91 particles (± 141 SD). FTIR and Raman spectroscopy accurately identified

* Corresponding author.,
** Corresponding author.
E-mail addresses: hannah.defrond@utoronto.ca (H. De Frond), leahth@sccwrp.org (L. Thornton Hampton), sydk@sccwrp.org (S. Kotar), kristineg@sccwrp.org
(K. Gesulga), cindymatuch@gmail.com (C. Matuch), waynel@sccwrp.org (W. Lao), stevew@sccwrp.org (S.B. Weisberg), charlesw@sccwrp.org (C.S. Wong),
chelsea.rochman@utoronto.ca (C.M. Rochman).

https://doi.org/10.1016/j.chemosphere.2022.134282
Received 25 November 2021; Received in revised form 14 February 2022; Accepted 7 March 2022
Available online 10 March 2022
0045-6535/© 2022 Elsevier Ltd. All rights reserved.
H. De Frond et al. Chemosphere 298 (2022) 134282

microplastics by polymer type for 95% and 91% of particles analyzed, respectively. Per particle, FTIR spec­
troscopy required the longest time for analysis (12 min ± 9 SD). Participants demonstrated excellent recovery
and chemical identification for particles greater than 50 μm in size, with opportunity for increased accuracy and
precision through training and further method refinement. This work has informed methods and QA/QC for
microplastics monitoring in drinking water in the State of California.

1. Introduction 2. Materials and methods

Microplastics represent a potential threat to human health, with Participants were from 22 laboratories in the United States, Canada,
exposure through the air we breathe (Chen et al., 2020; Gasperi et al., Germany, China, Australia, and Norway. The study benefitted from di­
2015), the food we eat (Cox et al., 2019; EFSA Panel on Contaminants in versity in institutions, laboratory practice, and experience by including
the Food Chain (CONTAM), 2016), and the water we drink (Eerkes-­ laboratories from different countries. The study was conducted ‘blind’,
Medrano et al., 2019). While further research is needed to understand and participating laboratories did not know any details of the spiked
biological fate and toxicity of microplastics in humans (Barboza et al., particle content within each sample.
2018; Carbery et al., 2018; Rist et al., 2018), there is sufficient evidence Laboratories were asked to follow a strict protocol (Appendix A.1), in
to impose a precautionary approach for monitoring and mitigation of which particles were extracted, then counted and identified using ster­
microplastics in drinking water (Danopoulos et al., 2020; Leslie and eomicroscopy, followed by chemical identification using either FTIR
Depledge, 2020; Senathirajah et al., 2021). and/or Raman spectroscopy on a subsample of particles (Appendix B.3).
Drinking water exposure has become a focal point for the State of Laboratories self-identified as novice (level 0 = new to microplastics
California, which has a legislative mandate (Senate Bill 1422) to enact analysis, having participated in a three-day training course on these
routine monitoring, including adoption of standard methods for imple­ methods immediately prior to the study), intermediate (level 1 = < 1
menting that requirement (Wyer et al., 2020). Such methods do not year of experience) or experts (level 2 = > 1 year experience). Each
currently exist, though there is a growing body of research from which to laboratory was also asked to record aspects related to QA/QC proced­
draw. Microscopy is a useful pre-screening tool for enumerating ures, time, and cost.
microplastic particles and is a comparatively quick method with low
upfront and equipment costs (Primpke et al., 2020). Fourier-Transform
2.1. Creation of spiked samples
Infrared (FTIR) and Raman spectroscopy are often used to confirm
microplastic particle counts by providing chemical confirmation of
Spiked simulated clean water samples were prepared at Southern
material type. Several studies have previously demonstrated the capa­
California Coastal Water Research Project (SCCWRP). Each spiked
bilities of these methods (Cabernard et al., 2018; Käppler et al., 2016; Xu
sample contained a combination of 15 different types of plastic particles
et al., 2019).
(Table S1). Plastic fragments of polystyrene (PS), polyvinyl chloride
Establishment of standard methods for management application re­
(PVC), polyethylene terephthalate (PET) and polyethylene (PE) of
quires several additional steps beyond publishing methods in peer-
various sizes and PET microfibers came from the Norwegian Institute for
reviewed journals. Management application requires development of a
Water Research, each in individual gelatin capsules containing sodium
detailed standard operating procedure with sufficient specificity to
bicarbonate and malic acid to facilitate dissolution. Spheres of PE and PS
ensure repeatability across laboratories. It also requires a multi-
(Table S1 - particles 4–7) were sourced from Cospheric LLC© (Santa
laboratory method evaluation study to determine performance charac­
Barbara, California, USA). Spiked plastics included some of the most
teristics, including demonstration of whether the precision and accuracy
produced polymer types (PlasticsEurope, 2020) with a range of den­
of the chosen methods are suitable for the intended purpose. These data
sities. Spiked plastic particles were white, clear, blue, green, and orange,
serve as a foundation for the development of laboratory accreditation,
and morphologically were fragments, spheres, and fibers. Across all
which define expectations for well-performing laboratories.
samples, the mean plastic content was 609 (±132 SD) particles, with
There have been several interlaboratory studies assessing micro­
average particle content of 249 (±60 SD) particles when the 1–20 μm
plastic methods performance across laboratories with blind microplastic
size fraction is excluded. The standard deviation of the spiked particle
samples (Cadiou et al., 2020; Isobe et al., 2019; Michida et al., 2019;
count was calculated by summing all standard deviations across all
Müller et al., 2020; Van Mourik et al., 2021). However, none have been
spiked particle types.
conducted using specific prescribed methodologies, which results in
In addition, four types of natural or ‘false positive’ particles were
high of variability among labs and across methods. In addition, all these
added (mean 80 particles per sample) to represent particles commonly
studies except for one (Müller et al., 2020) have been limited to particles
found within real-world samples which might be mistaken for plastics.
larger than 50 μm, which is problematic because drinking water is
These were fibers (undyed and dyed cellulose, animal fur) and shell
typically filtered to at least this size and monitoring needs to focus on
fragments (Table S1 - Particles 16–19).
smaller particles that can potentially pass through filters (Na et al.,
Simulated clean water samples were created using 1 μm filtered
2021).
(PCTE, Sterlitech) deionized water, referred to as microplastics analysis
Here we present a multi-laboratory validation study using three
grade (MAG) water. Glass containers were previously washed with soap
microplastic measurement methods: stereomicroscopy, Raman spec­
and water, pre-ashed at 450 ◦ C to destroy any possible contaminants,
troscopy, and FTIR spectroscopy. For each, we establish performance
and triple-rinsed with 1 μm filtered deionized water. 250 mL of MAG
characteristics, including accuracy and precision. We also establish the
water was placed in a heated water bath. Pills containing microplastics
time requirements to implement such methods, necessary for use in
were added (12 total per sample), and jars were kept at 48 ◦ C for 2 h
required monitoring. The study included laboratories from around the
until all soluble inert components (gelatin, malic acid, sodium bicar­
world with a range of experience, and all received samples spiked with
bonate) had dissolved, and the plastic particles dispersed. Each sample
diverse particle shapes, densities, colors and sizes, a strict protocol to
then received an additional 155 mL of MAG water and 45 mL of 1 μm
follow for extraction, particle identification, counting and chemical
filtered (PCTE, Sterlitech) 10% Alcojet® solution to minimize particu­
identification. These results collectively inform recommendations for
late agglomeration. The remaining particles were manually added using
microplastic analysis in drinking water, for monitoring programs within
metal forceps. For blank samples, the same process was followed with 12
the State of California and beyond.
empty pills (i.e., without plastic particles) each. Prepared samples were

2
H. De Frond et al.
covered and stored at room temperature prior to shipping. Throughout
all stages of sample preparation, measures were in place to minimize
particle contamination (Appendix B.1). To confirm spiked particle
counts, three test samples were prepared, extracted, and examined via
stereomicroscope prior to sample distribution (Table S2). To quantify
contamination from sample preparation, a contamination check was
carried out for every sample batch (Appendix B.2). Each participating
laboratory received three spiked samples and one blank sample.
2.2. Sample processing instructions
Previous studies noted that recoveries became incomparable when
laboratories were given the freedom to choose their extraction tech­
nique (Isobe et al., 2019). Therefore, a standard operating procedure
(SOP) (Appendix A.1) for extraction was provided for participants to
follow precisely. This protocol included a recommended supplies list,
and instructions for sample extraction based upon commonly used
methods (Hidalgo-Ruz, 2012; Masura et al., 2015; Michida et al., 2019)
considering equipment requirements, cost, and time. Laboratories were
instructed to use sieving and vacuum filtration to extract particles from
the samples. Participating laboratories were encouraged to practice
QA/QC measures to minimize laboratory contamination, in line with
recommendations from previous studies (Prata et al., 2020a) and other
practices from personal experience and communication (Appendix A.2).
2.3. Sample extraction
Samples were fractioned via sieving and vacuum filtration into four
size fractions: 1–20 μm, 20–212 μm, 212–500 μm, and >500 μm. Par­
ticipants were given the option to rinse particles from each sieve into
glass jars, or to vacuum-filter each size fraction onto filter paper prior to
stereomicroscopy (hereafter referred to as microscopy).
2.4. Microscopy
Following sample extraction, laboratories used microscopy to count
all suspected microplastic particles, and to categorize them by color and
morphology based upon keys provided. Particles were also measured
(length and width). For chemical characterization, a subsampling
strategy was provided using methods outlined in De Frond et al. (2022,
in prep.) where up to 30 particles of each color and morphology com­
bination (e.g., white fragments, blue fragments) within each size frac­
tion were selected (Appendix B.3).
2.5. Spectroscopy
For laboratories using μ-FTIR (or FTIR spectroscopy, hereafter
collectively called FTIR) or μ-Raman spectroscopy (hereafter called
Raman) to analyze particle composition, all subsampled particles were
chemically identified. Specific methods for particle preparation and
instrument settings here were not specified, and the instrument used by
each lab differed by manufacturer and model. However, general in­
structions for conducting manual microplastic analyses using the in­
struments available at SCCWRP (Thermo Fisher Nicolet iN10 FTIR
spectrometer and HORIBA XploRA ™ Plus Raman spectrometer) to local
participants were provided.
2.6. Data submission
Participating laboratories submitted data via an online data entry
form (Appendix C). Data submission included a description of
morphology, color, and size fraction for each suspected microplastic
particle counted using microscopy. For subsampled particles, the length,
width, particle image, and the chemical identification result from
spectroscopic analyses were required. Additional data was also collected
on laboratory information (e.g., QA/QC measures, expertise), extraction
3
Chemosphere 298 (2022) 134282
methods (e.g., filter material), instrument specifics (e.g., manufacturer,
software), microscopy settings (e.g., magnification), and spectroscopic
method parameters. The full dataset is publicly available to download
via the SCCWRP webpage: microplastics.sccwrp.org.
2.7. Data synthesis and analysis
Laboratories were anonymized, and data was analyzed using R
(Version 4.0.3). Total particle counts encompass all suspected micro­
plastic particles counted via microscopy, including spiked particles, false
positive natural particles misidentified as plastic, and particles intro­
duced from background contamination, i.e., recovery data was not
corrected based upon blank sample contamination or spectroscopy.
Overall recovery was evaluated by comparing reported total particle
counts with spiked particle counts for each sample. Precision was
evaluated using standard error of recovery. Among samples, there was
some variation in the number of manually added particles. This varia­
tion was documented and accounted for when calculating recovery for
each sample.
Results were separated by particles counted above and below the 20
μm size fraction. For all statistical tests, the significance level was set at
p < 0.05. Normality of data was tested using the Shapiro-Wilk test and
homogeneity of variance was tested using Bartlett’s test. All recovery
data was non-normal, so Kruskal-Wallis tests were used. The standard
error of recovery within each lab was used to compare the precision
within individual laboratories. To test if there was a relationship be­
tween accuracy (difference between reported recovery and 100%) and
precision (standard error of recovery) among laboratories, Pearson’s
correlation coefficient test was used.
To determine the accuracy of microscopy for identifying plastic
particles, the mean proportion of particles identified as plastic using
spectroscopy was calculated for each laboratory. Some laboratories re­
ported that logistical limitations prevented them from following the
prescribed subsampling procedure. Therefore, these results are treated
as estimates, as the number of particles identified using spectroscopy
varied widely among laboratories and was not always representative.
To determine the accuracy of FTIR and Raman spectroscopy for
chemically identifying material types, the images and raw data sub­
mitted by each laboratory were checked. Each particle visually and
descriptively matching a spiked plastic particle by color and
morphology was identified. Results were then filtered to include only
confirmed spiked particles chemically identified by spectroscopy. As
mentioned, the number of particles chemically identified using spec­
troscopy varied among laboratories. Although a comparison of perfor­
mance among laboratories was intended, this was not possible without
introducing bias from the number of particles identified by each lab.
Therefore, results from all laboratories were pooled into one dataset,
and the overall proportion of correct chemical identification results was
determined. For example, to determine accuracy of FTIR for identifying
PET particles, all spiked PET particles chemically identified among
laboratories were pooled, and the proportion of correct results was re­
ported. Some particle types had the same visible appearance (e.g., Par­
ticles 14 and 15), therefore, it was not possible to determine the true
polymer type prior to spectroscopic analysis. In these cases, if the
chemical identification result was one of the possible polymer types (e.
g., PET or PS) the result was considered correct.
Logistical results for time (per sample and per particle) were
compared for all steps in the analyses. The mean result was calculated
among laboratories and compared between methods using Mann-
Whitney U tests. Scripts used to generate all statistical analyses, sum­
mary results and figures are available at microplastics.sccwrp.org.
3. Results
Twenty-two laboratories submitted results for at least one method.
All laboratories submitted results for microscopy and counted particles
H. De Frond et al.
in size fractions >20 μm. Of the 22 laboratories, 13 did not count par­
ticles in the 1–20 μm size fraction. Seventeen laboratories submitted
results for their blank sample. For chemical identification, 11 labora­
tories used FTIR spectroscopy and 9 used Raman spectroscopy.
3.1. Microscopy: particle recovery
We measured particle recovery by comparing the number of sus­
pected microplastic particles counted using microscopy to the number of
spiked plastic particles. Across all labs, five (23%) reported mean par­
ticle counts that were within the true range (total spike ± SD). Thirteen
laboratories (59%) reported total particle counts lower than the true
value (total spike – SD) and four (18%) reported particle counts above
the true value (total spike +SD). Mean recovery among all laboratories
and all size fractions was 76 ± 10% (SE), with a significant difference in
recovery among laboratories (Kruskal-Wallis test).
Given that some laboratories did not submit data for the smallest size
fraction, results are separated into >20 μm (n = 22) and <20 μm (n = 9)
size fractions. Within the >20 μm size fractions, actual particle sizes as
measured by visual microscopy ranged from 50 μm to 2000 μm. The
minimum and maximum particle sizes in the <20 μm size fraction were
3 μm and 20 μm. Mean recovery of particles in size fractions >20 μm was
92 ± 12% among laboratories (Fig. 1A) (mean = 223 particles, median
= 213 particles, Figure S1; Table S3). Recoveries within this group
ranged from 13 to 246%. For the <20 μm size fraction, mean recovery
was 32 ± 16% (Fig. 1B) (mean = 109 particles, median = 79 particles,
Figure S1; Table S3). Recoveries within this group ranged from 1 to
158%.
4
Chemosphere 298 (2022) 134282
For particles in size fractions >20 μm (n = 22 laboratories), the
standard error of recovery within each laboratory ranged from 1 to 35%
(mean SE = 11%, Fig. 1C and Figure S2). For the <20 μm size fraction (n
= 9 laboratories), standard error of recovery ranged from 1 to 120%
(mean SE = 19%. Fig. 1D and Figure S2). Among all laboratories, there
was not a statistically significant relationship between accuracy (the
difference between mean recovery per laboratory and 100% recovery,
dependant on the size fractions counted) and precision (SE) of recovery
per laboratory (Figure S3).
3.2. Procedural contamination
Among laboratories, mean blank sample particle count was 91 ± 141
(SD) particles (median = 45) and ranged from 7 to 511 particles (Fig. 2,
Table S3). The most common colors reported were black, clear, blue, and
brown (Figure S4). The least common colors were gold, pink, purple and
silver. The most common particle morphologies reported were fibers,
then fragments (Figure S5). All but one laboratory counted fibers within
their blank, and all but three laboratories reported fragments. By size
fraction (Figure S6), mean blank particle counts were greatest in the
20–212 μm size fraction at 34 particles per sample.
3.3. Accuracy of microscopy for identifying plastic particles
Spectroscopy results were used to assess the accuracy of using visual
microscopy to differentiate between plastic and non-plastic (i.e., natural
or other anthropogenic) particles, determined by the proportion of
particles identified as plastic or non-plastic from the subsample per
Fig. 1. A and B) Overall particle recovery.
Each individual point within the box plot is the
mean percent recovery from visual microscopy,
reported by each laboratory (n = 1-3). Results
are separated into A) recovery of particles in
size fractions > 20 μm (n = 22 laboratories),
and B) recovery of particles below the 20 μm
size fraction (n = 9 laboratories). C and D)
Summary of particle recovery per sample from
visual microscopy, reported by each laboratory.
Results are separated into C) recovery of par­
ticles in size fractions > 20 μm (n = 22), and
D) recovery of particles below the 20 μm size
fraction (n = 9). The dotted lines represent
100% recovery based upon mean spiked parti­
cle counts among labs, and short, solid lines
represent the mean within each laboratory.
Each individual point represents an individual
sample. The grey shaded area indicates the
standard deviation of the spike for the relevant
size fractions, converted to percent recovery.
Standard deviation among all particles was
calculated by summing the standard deviation
of each spiked particle.
H. De Frond et al.
Fig. 2. Summary of particle counts reported within blank samples. Each point
represents the particle count within the blank sample as reported by each
laboratory (n = 1 per lab), 17 laboratories submitted particle counts for
blank samples.
laboratory (Fig. 3). The number of particles chemically identified using
spectroscopy varied widely among laboratories and among samples
(Figure S7). In total, the number of particles analyzed ranged from 103
to 2068 particles per laboratory, and thus the number of particles sub­
sampled was not always representative. Of the particles that were
chemically identified via spectroscopy, among laboratories (n = 19), the
mean proportion of plastic particles was 73% (SD = 24) (Fig. 3). Within
laboratories, the proportion of plastic particles ranged from 13 to 100%
(Figure S7).
3.4. FTIR spectroscopy
Of the spiked particles (both plastic and natural) analyzed using FTIR
spectroscopy, 93% were correctly identified by material type (Table S4).
Of the spiked plastic particles analyzed in all size fractions, 95% were
correctly identified. Accuracy by polymer type was 100% for PVC (n =
52), 95% for PS (n = 332), 97% for PE (n = 661), and 90% for PET (n =
169). Accuracy for identifying plastic particles in the 1–20 μm size
fraction was 33% (n = 3). Of the incorrectly identified plastic particles
(n = 62), 44% were classified as a plastic, however the polymer type
identified was incorrect. Further, 30% of results were inconclusive, and
5
Chemosphere 298 (2022) 134282
26% were misidentified as natural or semi-synthetic materials.
Of the false positive particles (natural), 65% were correctly identi­
fied. Accuracy for undyed cellulose fibers (n = 15) and shell fragments
(n = 1) was 100%, accuracy was 68% for dyed cellulose fibers (n = 47),
and 20% for fur fibers (n = 14). Of the incorrectly identified false pos­
itive particles, 44% were inconclusive, 33% were misidentified as
another natural material, and 22% were misidentified as plastic.
3.5. Raman spectroscopy
For all particles, Raman spectroscopy identified 83% of materials
accurately (Table S5). Of spiked plastic particles in all size fractions,
91% were correctly identified. By polymer type, accuracy was 91% for
PVC (n = 94), 98% for PS (n = 545), 88% for PE (n = 648), and 91% for
PET (n = 266). Of the plastic particles incorrectly identified using
Raman spectroscopy, 92% of results were inconclusive or classified as
‘other’ i.e., not classified as plastic or natural, 7% were misidentified as a
different plastic type, and 1% were misidentified as a natural material.
Overall accuracy for false positive particles (natural) was 18% using
Raman spectroscopy. Accuracy was 100% for animal fur (n = 13), 97%
for shells (n = 31), 25% for undyed cellulose (n = 4), and 5% for dyed
cellulose (n = 168). Of the incorrectly identified false positives, 83% of
results were misidentified as plastic, 16% were classed as inconclusive
or ‘other’, and 1% were misidentified as a different natural material.
3.6. Time
The mean time among laboratories taken to process a single sample
prior to spectroscopic identification was 41 h (Table S6). The largest
proportion of this time was from visual microscopy and extraction (16
and 15 h per sample), varying widely among laboratories (SD = 22 and
26 respectively).
Time per sample varied widely among laboratories dependent on
how many particles were recovered, counted, and chemically identified.
Therefore, results were normalized by calculating time per particle.
Here, time spent on microscopy averaged 4 ± 5 min per particle (mean
± SD). Both spectroscopic methods required a similar amount of time,
with Raman slightly faster (10 ± 11 min per particle) compared to FTIR
(12 ± 9 min per particle), but not significantly so (Mann-Whitney U).
Fig. 3. Proportion of particle counts
confirmed as plastic, by lab (n = 19). Pro­
portions are calculated as the mean propor­
tion among test samples (n = 1–3 per lab).
Proportions were calculated from a sub­
sample of particles analyzed via spectros­
copy, with numbers ranging from 103 to
2068 particles analyzed per lab. Laboratories
that did not use spectroscopy to chemically
identify particles are excluded from these
results. Laboratories are separated by self-
reported experience level for using micro­
scopy to analyze microplastics; 0 = First
experience, 1 = Less than 1 year experience,
2 = Over 1 year experience.
H. De Frond et al.
4. Discussion
Overall, the methods used in this study are fit for purpose to quantify
and characterize microplastics in drinking water for most applications.
Below we discuss the variables that affected the recovery and precision
of our results and the applicability of this work to the monitoring of
microplastics in drinking water.
This study included a SOP for all labs to follow, which is unique to
other method evaluation studies (Cadiou et al., 2020; Isobe et al., 2019;
Michida et al., 2019; Müller et al., 2020; Van Mourik et al., 2021). The
inclusion of an SOP is likely the reason that variability across labora­
tories was less than in other multi-laboratory studies with blind samples,
which reported lower mean recovery among laboratories (Cadiou et al.,
2020; Michida et al., 2019). Michida et al. (2019) attributed underes­
timation to particle loss during sample extraction, which we expect was
reduced by following the SOP that was tested prior to use to ensure
minimal procedural particle loss (Table S2).
However, even using an SOP, we found that interlaboratory repeat­
ability was highly varied, and results were significantly different among
laboratories. As recovery results were not corrected based upon particle
counts within blank samples, it is likely that differences in particle
contamination may have contributed to the differences in particle re­
coveries among laboratories. Notably, this variation does not appear
strongly attributable to experience level, with laboratories of all expe­
rience levels having false positive particle counts where they incorrectly
identified non-plastic materials (Cadiou et al., 2020). It is important to
account for different experience levels in establishing monitoring pro­
grams as novice laboratories will be common in new microplastics
monitoring programs, at least initially. This can be done via standard­
ized particle identification protocols and training, which are discussed
further by Kotar et al. (this volume).
Although other factors are relevant, particle size was the most
important factor affecting method performance. For particles in size
fractions >20 μm, mean recovery was 92%, whereas performance fell to
32% for particles <20 μm. This is in line with findings of previous
interlaboratory method comparison studies which reported overlooking
of particles in smaller size fractions when using microscopy (Isobe et al.,
2019; Michida et al., 2019; Müller et al., 2020). Visual microscopy is
thus appropriate for identifying plastic particles above 50 μm, the
smallest measured particle size within the >20 μm size fractions. Rec­
ommendations for microscopy are discussed in Kotar et al., (2022, this
issue). Method recommendations for spectroscopy, including priorities
for future research are discussed in De Frond et al., (2022, this issue).
Particle size s imilarly affected ability to distinguish polymer types
via spectroscopy. FTIR accurately identified microplastics from non-
plastics and correctly identified polymer type for particles in size frac­
tions above 20 μm but was ineffective below 20 μm. This was expected,
as the reported limit of detection using FTIR ranges from 10 to 20 μm in
other studies due to wavelength limitations (Primpke et al., 2020; Xu
et al., 2019). Raman was 94% effective in identifying plastics, even
including particles in the smallest size fraction in which particles ranged
from 3 to 20 μm, consistent with other studies where Raman spectros­
copy was used to identify small particles (Anger et al., 2018; Araujo
et al., 2018; Oβmann et al., 2018). However, Raman performed poorly
for dyed cellulose fibers. Other works have reported fluorescence
interference from dyes and pigments for Raman spectroscopy, a draw­
back of this method (Araujo et al., 2018; Lenz et al., 2015; Munno et al.,
2020). In addition, our polymer identification rate probably represents a
maximum because we did not use environmentally-weathered particles,
which can change the visual appearance of microplastics (Song et al.,
2017; ter Halle et al., 2017; Veerasingam et al., 2020).
While particle size was the most important factor, color, and shape
also affected performance. White particles, that were most common in
smaller size fractions, were challenging to identify under the light of a
microscope (Shim et al., 2017) when on a clear glass Petri dish (Michida
et al., 2019) or when mounted on a white or an opaque filter surface. Of
6
Chemosphere 298 (2022) 134282
the spiked plastics and false positive particles analyzed, accuracy of
chemical identification using both Raman and FTIR spectroscopy was
lowest for fibers compared to other morphologies. The effects of particle
characteristics on method accuracy for both microscopy and spectros­
copy are discussed further in Kotal et al., (2022, this issue), and De Frond
et al., (2022, this issue).
Combined, these findings help establish whether the methods we
examined are fit for purpose. It is important to recognize there are
several purposes for monitoring microplastics in drinking water, and
methods should be evaluated based on their ability to meet each
objective. One of those objectives is assessment of source water
contamination, for which these methods work well since source water
generally involves description of larger sized particles (e.g., >45 μm in
size (Zhu et al., 2021)). Another objective may be to determine exposure
in a risk assessment. For humans, particles <20 μm are considered most
relevant for both exposure and health effects (Prata et al., 2020b).
Although difficulties in measuring these particles are evident, this study
did provide an estimate for the recovery rate of particles <20 μm (32%)
making it possible to correct for the bias.
While the primary goal of the study was to document performance,
the study also served to identify several areas in which methods could be
improved. One of those areas is in blank contamination. On average,
laboratory contamination from blank samples was 91 particles, although
this varied among labs. This variation was most likely due to differences
in protocols for contamination control and the range in time spent
processing samples. Besides the opportunity to reduce blank contami­
nation, our SOP did not require or prescribe a procedure for blank
correction, which added to variability across laboratories. This proced­
ure was intentionally left open-ended as correction can be complex, as
the amount of airborne laboratory contamination can differ among
shapes and colors of particles (Prata et al., 2020a). Further, limits of
quantitation are driven by variability in replicate response and blank
contamination. If adopted, the limit of quantitation will help determine
the volume of water that must be sampled to confidently distinguish a
measured value from zero. Recommendations for minimization and
reporting of blank particle counts and limits of quantitation are dis­
cussed further in Munno et al., (2022) (this issue).
Another consideration in determining fitness for purpose is the time
expenditures required for implementation, an area identified for
improvement. On average, laboratories spent more than 50 h per sample
to complete sample processing through to visual quantification followed
by Raman spectroscopy, which was the faster of the two spectroscopic
methods (per particle). The times recorded are likely an overestimate as
our samples were spiked with more plastics than would be found in a
typical drinking water sample. Wong et al., (2022, this issue) have made
corrections for sample density and arrived at an estimate that was still
more than 40 h per sample for a typical drinking water sample. This
large time investment leads to consideration of a tiered monitoring
system in which spectroscopic methods are used only when detailed
particle description is needed, and less expensive methods are used for
screening level questions (Coffin et al., 2022; this issue). In addition, it
prioritizes the need for developing method automation, which is in
progress already for both FTIR and Raman spectroscopy (Primpke et al.,
2017; Xu et al., 2019).
5. Conclusions
Method performance was highly dependent on particle size, with
good recovery for particles >50 μm. Both FTIR and Raman spectroscopy
were effective at identifying microplastic particles and differentiating
from non-plastics but there were performance differences based on
particle size. FTIR could accurately identify polymer types for particles
in size fractions above 20 μm whereas Raman did so for particles in size
fractions above 1 μm. This size issue is of concern for sampling drinking
water systems as they typically filter down to 20 μm, and the primary
target is smaller particles. However, the data provided here quantifies
H. De Frond et al.
that bias allowing for application of correction factors. Standardized
measures to minimize laboratory particle contamination will reduce
variation in particle counts among labs. Time and associated costs for all
stages of microplastic sample analysis are currently barriers to cost-
effective monitoring, requiring new approaches for automation and
incorporation of tiered monitoring that uses less expensive methodolo­
gies for screening level questions.
Author contributions
Hannah De Frond: Conceptualization, Methodology, Validation,
Formal analysis, Data curation, Writing – original draft, Writing – re­
view & editing, Visualization. Leah Thornton Hampton: Methodology,
Formal analysis, Data curation, Visualization, Writing – review & edit­
ing. Syd Kotar: Validation, Investigation, Resources, Data curation,
Writing – review & editing. Kristine Gesulga: Methodology, Validation,
Investigation, Writing – review & editing. Cindy Matuch: Data curation,
Investigation. Wenjian Lao: Methodology, Validation, Investigation,
Resources, Data curation, Writing – review & editing. Stephen B.
Weisberg: Conceptualization, Writing – review & editing, Supervision,
Funding acquisition. Charles S. Wong: Methodology, Resources, Visu­
alization, Writing – review & editing, Supervision, Project administra­
tion. Chelsea M. Rochman: Conceptualization, Methodology, Writing –
review & editing, Supervision
Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Acknowledgements
We thank all participants of the study including: Alfred-Wegener-
Institute; Algalita Marine Research and Education; Barnett Technical
Services; BASF; California Department of Public Health; California State
University (Bakersfield); California State University (Channel Islands);
Carollo Engineers Inc., East China Normal University; Eastman Chemical
Company; U.S. Environmental Protection Agency; Eurofins (Australia,
Norway, US); HORIBA Scientific; Innovations institut für Nano­
technologie und korrelative Mikroskopie; Institute of Hydrobiology
(Chinese Academy of Sciences); Jinan University; NatureWorks LLC;
Lever Photonics/University of California Davis, Metropolitan Water
District; Norwegian Institute for Water Research; Ontario Ministry of the
Environment, Conservation and Parks; Orange County Sanitation Dis­
trict; Oregon State University; Pennsylvania State University; Photo­
thermal Spectroscopy Corp.; RJ Lee Group; Southern California Coastal
Water Research Project Authority; SiMPore, Inc.; Thermo Fisher (US);
University of California (Riverside); University of California (Santa
Barbara); University of Minnesota (Duluth); University of Quebec at
Rimouski; University of Toronto. We also thank Keenan Munno for her
involvement in the study planning and methodology, and her time spent
developing the methods used in this work and providing training to
study participants on protocols for sample extraction and visual char­
acterization of particles. We thank Bridget O’Donnell, Eunah Lee and
Suja Sukumaran for providing method and instrument training for
participating laboratories. We thank Robert Butler, Paul Smith and Zaib
Quraishi for their work in creation and management of the data portal
and data submission process. Finally, we thank Cobie and Kiera, two pet
rabbits of C.S. Wong, for donating Particle type 18 during normal
grooming. Funding was provided by the California State Water Re­
sources Control Board and SCCWRP.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
7
Chemosphere 298 (2022) 134282
org/10.1016/j.chemosphere.2022.134282.
References
Anger, P.M., Esch, E. von der, Baumann, T., Elsner, M., Niessner, R., Ivleva, N.P., 2018.
Raman microspectroscopy as a tool for microplastic particle analysis. TrAC Trends
Analyt.l Chem. 109, 214–226. https://doi.org/10.1016/j.trac.2018.10.010.
Araujo, C.F., Nolasco, M.M., Ribeiro, A.M., Ribeiro-Claro, P.J., 2018. Identification of
microplastics using Raman spectroscopy: latest developments and future prospects.
Water Res. 142, 426–440.
Barboza, L.G.A., Vethaak, A.D., Lavorante, B.R., Lundebye, A.-K., Guilhermino, L., 2018.
Marine microplastic debris: an emerging issue for food security, food safety and
human health. Mar. Pollut. Bull. 133, 336–348.
Cabernard, L., Roscher, L., Lorenz, C., Gerdts, G., Primpke, S., 2018. Comparison of
Raman and fourier transform infrared spectroscopy for the quantification of
microplastics in the aquatic environment. Environ. Sci. Technol. 52 (22),
13279–13288. https://doi.org/10.1021/acs.est.8b03438.
Cadiou, J.-F., Gerigny, O., Koren, Š., Zeri, C., Kaberi, H., Alomar, C., Panti, C., Fossi, M.,
Adamopoulou, A., Digka, N., 2020. Lessons learned from an intercalibration exercise
on the quantification and characterisation of microplastic particles in sediment and
water samples. Mar. Pollut. Bull. 154, 111097.
Carbery, M., O’Connor, W., Palanisami, T., 2018. Trophic transfer of microplastics and
mixed contaminants in the marine food web and implications for human health.
Environ. Int. 115, 400–409.
Chen, G., Feng, Q., Wang, J., 2020. Mini-review of microplastics in the atmosphere and
their risks to humans. Sci. Total Environ. 703, 135504. https://doi.org/10.1016/j.
scitotenv.2019.135504.
Cox, K.D., Covernton, G.A., Davies, H.L., Dower, J.F., Juanes, F., Dudas, S.E., 2019.
Human consumption of microplastics. Environ. Sci. Technol. 53 (12), 7068–7074.
https://doi.org/10.1021/acs.est.9b01517.
Danopoulos, E., Twiddy, M., Rotchell, J.M., 2020. Microplastic contamination of
drinking water: a systematic review. PLoS One 15 (7), e0236838. https://doi.org/
10.1371/journal.pone.0236838.
Eerkes-Medrano, D., Leslie, H.A., Quinn, B., 2019. Microplastics in drinking water: a
review and assessment. Drinking Water Contaminants 7, 69–75. https://doi.org/
10.1016/j.coesh.2018.12.001.
EFSA Panel on Contaminants in the Food Chain (CONTAM), 2016. Presence of
microplastics and nanoplastics in food, with particular focus on seafood. EFSA J. 14
(6), e04501 https://doi.org/10.2903/j.efsa.2016.4501.
Gasperi, J., Dris, R., Mirande-Bret, C., Mandin, C., Langlois, V., Tassin, B., 2015. First
overview of microplastics in indoor and outdoor air. In: 15th EuCheMS International
Conference on Chemistry and the Environment. https://hal-enpc.archives-ouvertes.
fr/hal-01195546.
Hidalgo-Ruz, et al., 2012. Microplastics in the marine environment: a review of the
methods used for identification and quantification. Environ. Sci. Technol. 46 (6),
3060–3075. https://doi.org/10.1021/es2031505.
Isobe, A., Buenaventura, N.T., Chastain, S., Chavanich, S., Cózar, A., DeLorenzo, M.,
Hagmann, P., Hinata, H., Kozlovskii, N., Lusher, A.L., Martí, E., Michida, Y., Mu, J.,
Ohno, M., Potter, G., Ross, P.S., Sagawa, N., Shim, W.J., Song, Y.K., Zhang, W., 2019.
An interlaboratory comparison exercise for the determination of microplastics in
standard sample bottles. Mar. Pollut. Bull. 146, 831–837. https://doi.org/10.1016/j.
marpolbul.2019.07.033.
Käppler, A., Fischer, D., Oberbeckmann, S., Schernewski, G., Labrenz, M., Eichhorn, K.-
J., Voit, B., 2016. Analysis of environmental microplastics by vibrational
microspectroscopy: FTIR, Raman or both? Anal. Bioanal. Chem. 408 (29),
8377–8391. https://doi.org/10.1007/s00216-016-9956-3.
Lenz, R., Enders, K., Stedmon, C.A., Mackenzie, D.M.A., Nielsen, T.G., 2015. A critical
assessment of visual identification of marine microplastic using Raman spectroscopy
for analysis improvement. Mar. Pollut. Bull. 100 (1), 82–91. https://doi.org/
10.1016/j.marpolbul.2015.09.026.
Leslie, H.A., Depledge, M.H., 2020. Where is the evidence that human exposure to
microplastics is safe? Environ. Int. 142 https://doi.org/10.1016/j.
envint.2020.105807, 105807–105807.
Masura, J., Baker, J., Foster, G., Arthur, C., 2015. Laboratory Methods for the Analysis of
Microplastics in the Marine Environment: Recommendations for quantifying
synthetic particles in waters and sediments.
Michida, Y., Chavanich, S., Chiba, S., Cordova, M.R., Cozsar Cabanas, A., Glagani, F.,
Hagmann, P., Hinata, H., Isobe, A., Kershaw, P., 2019. Guidelines for Harmonizing
Ocean Surface Microplastic Monitoring Methods, Version 1.1.
Müller, Y.K., Wernicke, T., Pittroff, M., Witzig, C.S., Storck, F.R., Klinger, J.,
Zumbülte, N., 2020. Microplastic analysis—are we measuring the same? Results on
the first global comparative study for microplastic analysis in a water sample. Anal.
Bioanal. Chem. 412 (3), 555–560.
Munno, K., De Frond, H., O’Donnell, B., Rochman, C.M., 2020. Increasing the
accessibility for characterizing microplastics: introducing new application-based and
spectral libraries of plastic particles (SLoPP and SLoPP-E). Anal. Chem. 92 (3),
2443–2451.
Na, S.-H., Kim, M.-J., Kim, J.-T., Jeong, S., Lee, S., Chung, J., Kim, E.-J., 2021.
Microplastic removal in conventional drinking water treatment processes:
performance, mechanism, and potential risk. Water Res. 202, 117417. https://doi.
org/10.1016/j.watres.2021.117417.
Oßmann, B.E., Sarau, G., Holtmannspötter, H., Pischetsrieder, M., Christiansen, S.H.,
Dicke, W., 2018. Small-sized microplastics and pigmented particles in bottled
mineral water. Water Res. 141, 307–316. https://doi.org/10.1016/j.
watres.2018.05.027.
H. De Frond et al.
PlasticsEurope, 2020. Plastics The Facts. PlasticsEurope. https://plasticseurope.org
/knowledge-hub/plastics-the-facts-2020/.
Prata, J.C., Castro, J.L., da Costa, J.P., Duarte, A.C., Rocha-Santos, T., Cerqueira, M.,
2020a. The importance of contamination control in airborne fibers and microplastic
sampling: experiences from indoor and outdoor air sampling in Aveiro, Portugal.
Mar. Pollut. Bull. 159, 111522. https://doi.org/10.1016/j.marpolbul.2020.111522.
Prata, J.C., da Costa, J.P., Lopes, I., Duarte, A.C., Rocha-Santos, T., 2020b.
Environmental exposure to microplastics: an overview on possible human health
effects. Sci. Total 702, 134455. https://doi.org/10.1016/j.scitotenv.2019.134455.
Primpke, S., Christiansen, S.H., Cowger, W., Frond, H.D., Deshpande, A., Fischer, M.,
Holland, E.B., Meyns, M., O’Donnell, B.A., Ossmann, B.E., Pittroff, M., Sarau, G.,
Scholz-Böttcher, B.M., Wiggin, K.J., 2020. Critical assessment of analytical methods
for the harmonized and cost-efficient analysis of microplastics. Appl. Spectrosc. 74
(9), 1012–1047.
Primpke, S., Lorenz, C., Rascher-Friesenhausen, R., Gerdts, G., 2017. An automated
approach for microplastics analysis using focal plane array (FPA) FTIR microscopy
and image analysis. Anal. Methods 9 (9), 1499–1511. https://doi.org/10.1039/
C6AY02476A.
Rist, S., Almroth, B.C., Hartmann, N.B., Karlsson, T.M., 2018. A critical perspective on
early communications concerning human health aspects of microplastics. Sci. Total
626, 720–726.
Senathirajah, K., Attwood, S., Bhagwat, G., Carbery, M., Wilson, S., Palanisami, T., 2021.
Estimation of the mass of microplastics ingested – a pivotal first step towards human
health risk assessment. J. Hazard Mater. 404, 124004. https://doi.org/10.1016/j.
jhazmat.2020.124004.
Shim, W.J., Hong, S.H., Eo, S.E., 2017. Identification methods in microplastic analysis: a
review. Anal. Methods 9 (9), 1384–1391. https://doi.org/10.1039/C6AY02558G.
Chemosphere 298 (2022) 134282
Song, Y.K., Hong, S.H., Jang, M., Han, G.M., Jung, S.W., Shim, W.J., 2017. Combined
effects of UV exposure duration and mechanical abrasion on microplastic
fragmentation by polymer type. Environ. Sci. Technol. 51 (8), 4368–4376.
ter Halle, A., Ladirat, L., Martignac, M., Mingotaud, A.F., Boyron, O., Perez, E., 2017. To
what extent are microplastics from the open ocean weathered? Environ. Pollut. 227,
167–174. https://doi.org/10.1016/j.envpol.2017.04.051.
Van Mourik, L., Crum, S., Martinez-Frances, E., van Bavel, B., Leslie, H., de Boer, J.,
Cofino, W., 2021. Results of WEPAL-QUASIMEME/NORMANs first global
interlaboratory study on microplastics reveal urgent need for harmonization. Sci.
Total Environ. 772, 145071.
Veerasingam, S., Ranjani, M., Venkatachalapathy, R., Bagaev, A., Mukhanov, V.,
Litvinyuk, D., Mugilarasan, M., Gurumoorthi, K., Guganathan, L., Aboobacker, V.M.,
Vethamony, P., 2020. Contributions of Fourier transform infrared spectroscopy in
microplastic pollution research: a review. Crit. Rev. Environ. 1–63. https://doi.org/
10.1080/10643389.2020.1807450.
Wyer, H., Polhemus, D., Moore, S., Weisberg, S.B., Coffin, S., Rochman, C.M., 2020. Steps
scientists can take to inform aquatic microplastics management: a perspective
informed by the California experience. Appl. Spectrosc. 74 (9), 971–975.
Xu, J.-L., Thomas, K.V., Luo, Z., Gowen, A.A., 2019. FTIR and Raman imaging for
microplastics analysis: State of the art, challenges and prospects. Trac. Trends Anal.
Chem. 119, 115629.
Zhu, X., Munno, K., Grbic, J., Werbowski, L.M., Bikker, J., Ho, A., Guo, E., Sedlak, M.,
Sutton, R., Box, C., Lin, D., Gilbreath, A., Holleman, R.C., Fortin, M.-J., Rochman, C.,
2021. Holistic assessment of microplastics and other anthropogenic microdebris in
an urban bay sheds light on their sources and fate. ACS ES&T Water 1 (6),
1401–1410. https://doi.org/10.1021/acsestwater.0c00292.