Systematic Study of Microplastic Fiber Release from 12 Different Polyester Textiles during Washing

Abstract

Microplastic fibers (MPFs) have been found to be a major form of microplastics in freshwaters, and washing of synthetic textiles has been identified as one of their main sources. The aim of this work was to use a panel of 12 different textiles of representative fibers and textile types to investigate the source(s) of the MPF during washing. Using standardized washing tests, textile swatches tailored using five different cutting/sewing methods were washed up to 10 times. The MPF quantity and fiber length were determined using image analysis. The 12 textiles demonstrated great variability in MPF release, ranging from 210 to 72,000 MPF/g textile per wash. The median MPF length ranged from 165 to 841 μm. The number of released MPF was influenced by the cutting method, where scissor-cut samples released 3-21 times higher numbers of MPF than the laser-cut samples. The textiles with mechanically processed surfaces (i.e., fleece) released significantly more (p-value < 0.001) than the textiles with unprocessed surfaces. For all textiles, the MPF release decreased with repeated wash cycles, and a small continuous fiber release was observed after 5-6 washings, accompanied by a slight increase in the fiber length. The decrease in the number of MPF released is likely caused by depletion of the production-inherited MPFs trapped within the threads or the textile structure. The comparison of MPF release from laser-cut samples, which had sealed edges, and the other cutting methods allowed us to separate the contributions of the edge- and surface-sourced fibers from the textiles to the total release. On an average, 84% (range 49-95%) of the MPF release originated from the edges, highlighting the importance of the edge-to-surface ratio when comparing different release studies. The large contribution of the edges to the total release offers options for technical solutions which have the possibility to control MPF formation throughout the textile manufacturing chain by using cutting methods which minimize MPF formation.

Full text

Systematic Study of Microplastic Fiber Release from 12 Different
Polyester Textiles during Washing
Yaping Cai, Tong Yang, Denise M. Mitrano, Manfred Heuberger, Rudolf Hufenus, and Bernd Nowack*
Cite This: https://dx.doi.org/10.1021/acs.est.9b07395
Read Online
ACCESS Metrics & More Article Recommendations *
sıSupporting Information
ABSTRACT: Microplastic fibers (MPFs) have been found to be a major
form of microplastics in freshwaters, and washing of synthetic textiles has
been identified as one of their main sources. The aim of this work was to use
a panel of 12 different textiles of representative fibers and textile types to
investigate the source(s) of the MPF during washing. Using standardized
washing tests, textile swatches tailored using five different cutting/sewing
methods were washed up to 10 times. The MPF quantity and fiber length
were determined using image analysis. The 12 textiles demonstrated great
variability in MPF release, ranging from 210 to 72,000 MPF/g textile per
wash. The median MPF length ranged from 165 to 841 μm. The number of
released MPF was influenced by the cutting method, where scissor-cut samples released 3−21 times higher numbers of MPF than
the laser-cut samples. The textiles with mechanically processed surfaces (i.e., fleece) released significantly more (p-value < 0.001)
than the textiles with unprocessed surfaces. For all textiles, the MPF release decreased with repeated wash cycles, and a small
continuous fiber release was observed after 5−6 washings, accompanied by a slight increase in the fiber length. The decrease in the
number of MPF released is likely caused by depletion of the production-inherited MPFs trapped within the threads or the textile
structure. The comparison of MPF release from laser-cut samples, which had sealed edges, and the other cutting methods allowed us
to separate the contributions of the edge- and surface-sourced fibers from the textiles to the total release. On an average, 84% (range
49−95%) of the MPF release originated from the edges, highlighting the importance of the edge-to-surface ratio when comparing
different release studies. The large contribution of the edges to the total release offers options for technical solutions which have the
possibility to control MPF formation throughout the textile manufacturing chain by using cutting methods which minimize MPF
formation.
1. INTRODUCTION
The ubiquitous presence of microplastics in the environ-
ment
1−6
as well as in biota
7,8
has been reported by hundreds of
studies. A mixture of fibers, beads, and fragments were usually
found in environmental samples, where microplastic fibers
(MPFs) were reported to account for a large proportion of all
microplastics.
9−11
A modeling study has shown that fibers
released from textiles are one of the most important sources of
microplastics in freshwaters.
12
One pathway is likely from
synthetic clothing where domestic washing may release a high
number of fibers.
13,14
Although the removal efficiency of
microplastics in wastewater treatment plants can reach 98% or
higher,
15,16
a few percent of MPFs which are not captured can
make a contribution of about 20% to the microplastic release
to freshwaters.
12
Alternatively, MPF captured during water
treatment can enter the terrestrial environment when sewage
sludge is applied on soil.
17−20
Several studies have investigated the general release of MPF
from textiles during washing, addressing a few factors in
isolation that were hypothesized to control the release.
13,21−32
The MPF release was mostly found to decrease with increasing
number of wash cycles,
21,27−32
sometimes reaching a plateau
between the 4th and 5th wash.
30,31
However, a relative stable
release over time was also observed in one study, which
proposed that MPF may have been formed during the textile
manufacturing process prior to washing.
24
The addition of
liquid or powder detergents was reported to either
increase
22,24,27
or have no impact
28,32
on the MPF release as
compared to pure water, while textile softeners were found to
significantly reduce the MPF release.
22
A recent study revealed
that a higher water volume can elevate the release of MPF
during washing because of the increasing hydrodynamic
pressure on fibers.
32
In addition, other factors, including
temperature, duration of the wash cycle, and mechanical stress
(number of steel balls added to the container) were
investigated, but there was no significant influence of these
factors on the MPF release.
22,24,32
Received: December 6, 2019
Revised: March 18, 2020
Accepted: March 23, 2020
Articlepubs.acs.org/est
© XXXX American Chemical Society A
https://dx.doi.org/10.1021/acs.est.9b07395
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Downloaded via LIB4RI on April 6, 2020 at 14:20:54 (UTC).
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

The textile samples tested in previous studies were either
garments purchased from the market
14,23,26,27,29
or consisted of
only a limited types of textile samples.
21,22,24,25
Although
efforts have been made to compare the MPF release between
different washing systems,
33
considering the variability of
experimental protocols adopted by the different studies, it is
still difficult to comprehensively compare the results on MPF
release from textiles with different properties (e.g., the surface
treatment, textile structure, and yarn type) and to gain an
integrated insight into how these properties influence the MPF
release through the washing process. Furthermore, the cutting
method has been identified as being one of the most critical
factors affecting the number of MPF released from textiles
during washing.
25
The most commonly used cutting approach
in the industry is mechanical cutting where knives are vertically
guided to cut multiple layers of textiles.
34
In addition, thermal
cutting is also applied during textile tailoring.
34
Then, cut
edges are often sewn with stiches in the later step. Similar to
the industry, many studies adopted either mechanically cut
swatches with sewn edges
31
or thermally cut swatches
21,32
to
reduce MPF shedding from edges. However, it is still unclear
whether those procedures can really prevent MPF release from
the edge and how the cutting/sewing method affects MPF
release.
Therefore, the aim of our study was to systematically
investigate the influence of textile properties and tailoring
methods on the MPF release during washing. In particular, we
place special emphasis on studying the influence of industrially
used cutting/edge sewing methods. We suspect that: (1) MPFs
are already present in the textile before washing and are only
released but not generated through the washing procedure; (2)
MPF release from textiles during washing will be affected by
textile properties such as the surface treatment, textile
structure, and/or yarn type; (3) higher mechanical stress will
lead to increased MPF release; (4) textiles with tailored edges
will release fewer MPFs than the textiles with scissor-cut edges
without any treatment. The results from this work will not only
form a basis for the understanding of the mechanisms which
influence MPF release from textiles during washing but also
provide evidence to help the industry to develop better
mitigation strategies.
2. MATERIALS AND METHODS
2.1. Textiles. A set of 12 commercially available polyester
textiles were obtained directly from different manufacturers
(Table 1). Seven of the textiles were made of filament yarns
and five of spun staple fiber yarns. The suffix“S”or “F”was
used to distinguish textiles made from spun yarns (staple-
length fibers) or filament yarns (endless fibers), respectively.
Six woven and six knitted textiles were used. The production of
some types of textiles can include a mechanical surface
treatment, where the surface fibers are cut to create a fuzzy or
soft texture of the final product.
34,35
For the fleece, there is a
shearing process where the surface fibers are cut by a blade.
For textiles with brushed surfaces, the break of surface fibers is
usually carried out by a metal brush in the brushing process. In
our study, a fleece textile and a plain textile with brushed
surfaces were investigated, and they were referred to as textiles
with the processed surface. The suffix“B”was given to the
plain textile with the brushed surface. A microfiber cloth
containing fibers with a much smaller diameter than all the
other textiles was also included. All samples were dark colored
and ranged in density from 75 to 294 g/m2(determined by
weighing three pieces of 36 cm2samples). Fourier-transform
infrared spectroscopy analysis (Varian 640-IR) confirmed the
chemical composition (polyester) of all textile samples.
Scanning electron microscopy (SEM) (Hitachi S6200, 2.0
kV, ×40) was used to characterize the textile structure and
fiber diameter (Tables 1 and S1). Before observation under the
SEM, a layer of 7 nm Au/Pd was sputtered on the samples’
surface by a high vacuum sputter coater (LEICA EM ACE600)
to enhance the contrast. To determine the fiber diameter, 10
fibers from the textile were randomly measured on the SEM
image.
2.2. Sample Preparation. Textiles were cut into small
pieces with sizes depending on the requirement of the
experiments (Figure 1). For the first three sets of experiments,
the swatches were cut either with textile scissors or a laser
cutter (tt-1300, Times technology) into pieces of 4 ×10 cm
according to the ISO standard 105-C06.
36
The weight of these
samples ranged between 0.29 and 1.30 g. To discriminate the
influence of the cutting/sewing methods, an additional set of
textile swatches were cut to a larger size of 10 ×10 cm to have
enough area available to sew the edges. Three additional types
of cutting/sewing methods commonly used in the industry
were investigated: model cutting (MC) (SAMCO type SB 25
A), overlock sewing (OS), and double-folded sewing (DS). An
OS machine (BERNINA 2000D) was used to sew the edges
with overlock stiches (width: 5 mm; length: 2.5 mm; and
density: 10 stiches per inch), which is one of the most
Table 1. Characterization of Textile Physical Properties
surface structure type yarn color density [g/m2]fiber diameter [μm]
unprocessed knit interlock spun black 209 ±1 12.2 ±0.8
jersey spun black 226 ±1 12.8 ±0.8
rib spun black 294 ±2 12.7 ±1.1
rib filament black 199 ±1 15.9 ±2.2
terry spun black 208 ±2 13.0 ±1.3
woven plain spun black 100 ±0 12.7 ±0.5/13.4 ±0.9
a
plain filament black 149 ±1 7.5 ±0.6/7.9 ±0.5
a
twill filament black 154 ±1 12.4 ±1.8/19.9 ±1.7
a
satin filament black 75 ±0 13.0 ±0.7/16.4 ±1.7
a
processed knit fleece filament black 185 ±1 11.7 ±1.3
woven plain brushed filament black 131 ±0 9.0 ±1.2/10.1 ±1.5
a
woven filament grey 191 ±3 19.9 ×8.9/7.7 ×2.2
b
(microfiber)
a
The diameters of the weft and the warp yarns of the woven textiles.
b
The width and length of the weft yarn (19.9 ±1.1 ×8.9 ±1.2 μm) and the
warp yarn (7.7 ±0.9 ×2.2 ±0.5 μm) with a rectangular cross section for the microfiber sample.
Environmental Science & Technology pubs.acs.org/est Article
https://dx.doi.org/10.1021/acs.est.9b07395
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
B

commonly used stiches to tailor the edges in garment
manufacturing. For the double-sewn samples, the edges were
folded twice and then sewn with simple stiches (length: 1.5
mm and density: 17 stiches per inch), as described in our
previous study.
24
Both overlock- and double-sewn samples
were originally scissor-cut, and white cotton threads were used
to sew the samples so that it would not interfere with our MPF
analysis later in the study. The contribution of the sewing
thread to the total weight of the finished sample was between 3
and 6%. The size of a finished double-sewn swatch was
approximately 5 ×5 cm with an average weight of 1.47 and
2.29 g per piece for plain F and interlock S, respectively.
Examples of the finished samples can be found in Figure S1.
2.3. Washing Experiments. Domestic washing was
simulated by a Gyrowash lab washing machine (James Heal,
GyroWash model 1615). Before each wash cycle, the steel
vessels were rinsed with deionized water three times, and a 5
min rinse cycle with linear alkylbenzene sulfonic acid (LAS)
solution (0.75 g/L, pH 9.2 ±1) was performed. The LAS
solution was used to simply mimic the liquid laundry detergent
commonly containing a more complex composition used in
households, as was demonstrated in our previous study.
24
Two
of the eight steel vessels were used as blanks containing only
the LAS solution in each experimental round, and the blank
vessels were always randomly chosen for the next round of
experiments. No prewash step was performed on the textile
samples. All experiments were performed in triplicates. A
schematic of the work flow is shown in Figure 1. The washing
experiments were conducted following a standardized wash
procedure from ISO (105-C06, 1994) with a slight
modification of the washing solution (Table S3), and no
adjacent textile was used. One piece of textile was added to a
steel vessel (500 mL) with 150 mL of the LAS solution
together with 10 steel balls (diameter of 6 mm). Each wash
cycle lasted for 45 min at a temperature of 40 °C. This
standardized washing condition was applied once to all 12
textiles in the experiment which investigated the influence of
textile properties on MPF release. To determine the influence
of repeated wash cycles on MPF release, four representative
textile variants (fleece, interlock S, plain F, and plain S) with
scissor-cut and laser-cut edges were washed for 10 rounds
following the washing condition, as described above. The wash
liquid from 1st, 2nd, 3rd, 5th, 8th, and 10th round was filtered
and analyzed. The influence of mechanical stress on the MPF
release was investigated by adding different number of steel
balls (0, 10, and 20) for the four selected samples (fleece,
interlock S, plain F, and plain S). To study how the edge
treatment influenced the MPF release, samples of two selected
textiles (plain F and interlock S) were prepared with five
different cutting/sewing methods, as described in the previous
section. The experiments followed the standardized conditions,
and three wash cycles were performed for each treatment.
2.4. Water Collection and Filtration. After each wash
cycle, the textile sample was taken out of the vessel with
tweezers and allowed to drip for 15 s to drain excess liquid.
The washing liquid remaining in the vessel was continuously
stirred to keep the suspension homogeneous. Then, the liquid
was transferred by a 10 mL pipet to a filtration system
consisting of a filtration unit and a vacuum pump. The wash
water was filtered through a cellulose nitrate membrane (GE
Whatman diameter 4.7 cm, pore size 0.45 μm). To avoid too
many overlapping fibers on the filters, the volume of the
filtered solution was adjusted to be between 1 and 50 mL
depending on the MPF concentration, aiming for 100 to 200
MPFs per filter. All filters were dried in separate Petri dishes
(VWR, diameter 90 mm, height 16 mm) covered with the
corresponding caps at room temperature overnight.
2.5. Filter Imaging and Analysis. The filters were imaged
by a single-lens reflex camera (Nikon D850) with a macro lens
(Nikon 105 mm/2.8) together with a calibration slide for a
microscope with a smallest scale of 100 μm (VWR Catalog
number: 630-1123). The images (8256 ×5504 pixels) were
edited in the software Adobe Lightroom CC (version:
2015.14) to enhance the contrast. All images were analyzed
with the software FiberApp (version: 1.51)
37
for MPF number
and length. The semiautomatic software allowed manual
selection of the starting and the end points of each fiber and
automatically recorded the fiber number and calculated the
fiber length. In total, the individual lengths of 52,800 MPF
were recorded from 530 filters with a detection limit of 7−8
pixels, corresponding to a length of about 90 μm. Examples of
filter images can be found in Figures S2 and S3. Additionally,
to determine the mechanism of fiber breakage, fibers released
from interlock S during the 1st and 10th washing cycle on the
filter were sampled by adhesive tapes, and nine fibers were
randomly selected to image the fiber ends with SEM (Hitachi
S6200, 2.0 kV, ×1.80). The experimental procedure was
assessed in terms of level of contamination and reliability of a
standardized workflow, which can be found in the Supporting
Information.
The mass of the released MPF from knit textiles was
calculated by multiplying the density of polyester (1.38 g/
cm3)
38
by the measured length and the known diameter of the
fibers (Table 1). For woven textiles, the mass cannot be
calculated because there are two types of yarns with different
diameters which we could not separate with the chosen image
analysis method.
2.6. Statistics. For the experiment investigating textile
properties, the influence of three factors, including the surface
treatment (unprocessed and processed), the textile structure
(knit and woven), and the yarn type (spun and filament) on
the number and length of MPF released from scissor- and
laser-cut textiles was determined by a linear mixed model
(package “lmerTest”) in R (version 3.4.3.), respectively. The
three factors and the textile types (interlock, plain, fleece, etc.)
were considered to exhibit fixed and random effects,
respectively. For the other experiments, a one-way ANOVA
Figure 1. Flow chart of the study; 12 types: 12 types of textiles, as
shown in Table 1; 4 types: fleece, interlock S, plain F, plain S; 2 types:
plain F and interlock S. SC: scissor cutting; LC: laser cutting; DS:
double-folded sewing; MC: model cutting; and OS: overlock sewing.
Environmental Science & Technology pubs.acs.org/est Article
https://dx.doi.org/10.1021/acs.est.9b07395
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
C

test in IBM SPSS software (version 25) was used to determine
the impact of different treatments on the number of MPF
released in the following groups: (1) mechanical stress:
experiments with different steel balls and (2) cutting/sewing
methods: five different ways of cutting/sewing technique. A
nonparametric Kruskal−Wallis (K−W) one-way ANOVA test
in IBM SPSS software (version 25) was applied to compare the
length distribution of the mentioned groups. The p-value of
any statistical test below 0.05 was considered to indicate
statistical significance.
3. RESULTS
3.1. MPF Release from 12 Polyester Textiles. The
number of released MPF per wash demonstrated a large variety
ranging from 210 MPF/g for laser-cut Twill F to 72,000 MPF/
g for scissor-cut microfiber (Figure 2A, Table S4). The
majority of the MPF exhibited a length between 100 and 1000
μm(Figure 2B, Table S5). The shortest MPFs were released
from laser-cut microfiber (median: 165 μm), while the longest
MPF were found from scissor-cut terry S (median: 841 μm).
Furthermore, the cutting method and the surface treatment
of the textile were found to have significant impact on the MPF
release. The scissor-cut samples released 3−21 times higher
numbers of MPF than the laser-cut textiles (Figure 2A). The
length distribution was displayed in a notched box plots, where
there is a strong evidence (95% confidence) of different
medians if two boxes’notches do not overlap.
39
For most of
the samples, the MPF shed from the scissor-cut samples
exhibited longer length (Figure 2B). The textiles with
mechanically processed surfaces released significantly more
MPF (p-value < 0.001) than textiles with unprocessed surfaces
for both scissor-cut and laser-cut samples. On the other hand,
the linear mix model shows that the textile structure and the
yarn type were found to have no significant influence on the
number of MPF released. The MPF length seemed to be
influenced by the textile structure and surface treatment
depending on the cutting methods. The scissor-cut woven
textile and the laser-cut textile with the processed surface shed
significantly shorter MPFs than the knit textile (p-value =
0.0297) and the textile with the unprocessed surface (p-value <
0.001), respectively.
3.2. MPF Release during Repeated Wash Cycles. To
investigate the influence of repeated wash cycles on MPF
release, four textiles were washed for 10 wash cycles, and the
wash liquid from selected rounds were filtered and analyzed. A
decrease in the number of MPF released in the first three wash
cycles was observed for all the samples (Figure 3,Tables S6
and S7). Depending on the type of textiles, 6 to 120 times
higher numbers of MPF were released in the 1st wash cycle
than the 10th wash cycle. After three to five wash cycles, the
number of released MPF either stabilized at a lower level or
slightly increased. The number of MPF released after the 10th
wash cycle was still above the blank (4 MPF per filter) with an
average between 10 and 1200 MPF/g. In terms of the length,
most of the samples released significantly shorter MPFs in the
1st wash cycle compared to the 10th cycle (Figure 4,Table
S8), especially for fleece. To determine the mechanism of fiber
breakage, MPFs were randomly selected to characterize the
morphology of the fiber ends with SEM, and the images
suggest that the most MPFs released from the 1st (Figure 5A)
and 10th (Figure 5B) wash cycles had ends similar to those
caused by the scissor cutting (Figure 5D). In addition, a
comparison between the MPFs shed from scissor- and laser-cut
samples (Figure 5A,C) shows that both samples shed some
MPFs with distorted ends, but MPFs possessing molten ends
were exclusively found for laser-cut samples.
To calculate the cumulative release of MPF during the 10
wash cycles, the number of MPF released from 1st to 10th
wash cycles was summed up. An interpolation was made for
the cycles where no analysis was performed (4th, 6th, 7th, and
9th wash cycles). We assumed that the number of released
MPF in the wash cycles which were not analyzed was the mean
number of its neighboring cycles. For instance, to estimate the
release during 4th wash cycle, we averaged the number of MPF
released from 3rd and 5th wash cycles. The resulting
cumulative release is shown in Table S9. The release during
the 1st wash cycle accounted for 30% (plain S) to 90% (plain
F) of the cumulative release.
3.3. MPF Release with Different Mechanical Stress
and Cutting/Sewing Methods. The number of steel balls
added (0, 10, and 20) did not show a statistically significant
impact on the number of MPF released for the majority of
tested samples, including plain F, interlock S, and fleece
(Figure S7 and Table S10). On the other hand, a similar MPF
length distribution was found for most of the textiles regardless
of the level of mechanical stress during the experiments.
Five different cutting and sewing methods were tested to
examine their influence on MPF release. Only laser cutting
Figure 2. (A) Number of MPF released from 12 polyester textiles
during the 1st wash cycle in number of MPF per gram of textile. The
mean values are represented by the black dots (scissor-cut) and red
dots (laser-cut). Error bars show the standard deviation for three
experimental replicates. The mean of the blanks for this set of
experiments is displayed as a horizontal line. S/L stands for the scissor
to laser ratio, which was calculated as number of released MPF from
the SC sample divided by the LC sample. (B) Length distribution of
MPF released during the 1st wash cycle. The values presented here
are a summation of triplicate experiments. The number of MPF
plotted per box was between 182 and 1498, with an average of 560.
The boxes represent 25th and 75th percentiles with the median
indicated by a line. The notches represent 95% confidence interval for
the median. Whiskers and outliers are not shown on the graph. The
complete datasets can be found in Figure S5. Asterisks are given to
those where the median length was significantly different between
laser-cut and scissors-cut samples.
Environmental Science & Technology pubs.acs.org/est Article
https://dx.doi.org/10.1021/acs.est.9b07395
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
D

significantly reduced the amount of the MPF release from
textiles during washing (Figure 6,Table S11). Laser-cut
interlock S released a cumulative number of 1300 MPF/g in
the first three washes which was about 7 to 12 times lower than
the other cutting/sewing methods. Similar results were found
for plain F, where the laser-cut sample shed fewer MPFs (370
MPF/g) than the other treatments (3760−7,400 MPF/g). No
significant difference was found for the number of MPF
released between the other four types of cutting/sewing
methods (scissor-cut, model-cut, double-sewn, and overlock-
sewn) for interlock S. However, for plain F, the overlock-sewn
sample released 1.5 to 20 times more MPF (7,400 MPF/g)
than the other cutting/sewing methods. In addition, all
treatments decreased the number of MPF released through
multiple wash cycles (Figure S9).
The length of MPF released from the textile with different
cutting/sewing methods exhibited a larger variability for plain
F than for interlock S (Figure 6B and S10). Scissor-cut plain F
released approximately two times longer MPF (median: 434
μm) than the double-sewn samples (median: 204 μm) during
the 1st wash cycle, while the interlock S released MPFs with
lengths ranging from 363 (double-sewn) to 528 μm (model-
cut). For both textiles, the shortest MPF were found for
double-sewn samples.
4. DISCUSSION
Our work explored the hypothesis that a significant fraction of
MPF released during washing of textiles were already present
in the textile during the manufacturing process and not
produced during the number of wash cycles investigated here.
We found that the number of MPF released was greatly
influenced by the number of wash cycles, with fewer MPF
observed in the wash water after the first cycle, as has been
observed in other studies.
21,27−32
This strongly indicates that
MPF were present in textile products before washing and that
the act of washing released the MPF which were on/in the
textile. The morphology of the fiber ends can inform us about
the processes responsible for fiber breakage.
40,41
For example, a
mushroom head indicates that the fracture was caused by a
high tensile force (Figure S11). For washing, the most relevant
Figure 3. Number of MPF released per cycle, as a function of the number of wash cycles for plain F, plain S, interlock S, and fleece. (A,B) For
scissor-cut and laser-cut samples. All values are presented in MPF per gram of textile. Error bars show the standard deviation for three experimental
replicates. The mean of blanks (n= 4 MPFs) for this set of experiments is displayed as a horizontal line.
Figure 4. Length distribution of MPF released during the 1st, 2nd, 3rd, 5th, 8th, and 10th wash cycle. NA: less than five MPFs. The values here
were a summation of triplicate experiments. The number of MPF plotted per box was between 10 and 1277, with an average of 318. The boxes
represent 25th and 75th percentiles with the median indicated by a line. Whiskers and outliers are not shown on the graph. The complete datasets
can be found in Figure S6. Asterisks are given to those samples where the median length was significantly different from the 1st release.
Environmental Science & Technology pubs.acs.org/est Article
https://dx.doi.org/10.1021/acs.est.9b07395
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
E

scenario is breakage by fatigue failure, resulting in the fiber
ends that have been split into finer fibers or fibrils.
41
Another
form of fiber failure is the surface wear and peeling, which
often leads to tapered ends or multiple splits.
41
In our
experiments, the SEM images showed no sign of fibrillation or
splits at the fiber end. Most of the MPF released from scissor-
cut samples during the 1st and 10th wash cycles showed a solid
end similar to the scissor-cut ones. Some MPFs exhibited
distorted ends, which seem to originate from the manufactur-
ing process where mechanical stresses, such as transverse
pressure, can happen either deliberately or accidentally.
The relatively steady release of MPF after a few washings
suggests that a part of the fibers have a slow release dynamics.
The MPF released during the 10th wash cycle were
significantly longer as compared to the MPF from the 1st
wash cycle for most of the samples. Because longer fibers are
more entangled within the textile structure, they may require
more time to work their way out from textiles. It may be
possible that additional breakage and release of fibers could be
expected during laundering with older textiles as they
experience more wear during use, but this was not evaluated
in this present study.
The textiles with mechanically processed surfaces exhibited
higher MPF releases than those with untreated surfaces, which
was also observed in some other studies.
21,30
Processed
surfaces investigated in this study are produced by cutting
surface fibers by a blade or brush. On one hand, possibly
several loosely entangled MPF are generated during this
process. On the other hand, such processes can loosen both
yarn and surface structures, therefore making the release of
MPF from within the textile structure easier. Statistical analysis
showed that the MPF released from the laser-cut processed
textiles had shorter lengths suggesting that shorter MPF may
be generated during the surface treatment process. Another
point worth noting is that although textiles with the processed
surface shed higher amounts of MPFs at the beginning, after 10
rounds of washing, the MPF release decreased to the same
level as the unprocessed textiles.
We hypothesized that higher mechanical stress will lead to
higher MPF release during washing because the steel balls may
contribute to the abrasive friction on the fibers or to increase
textile bending and consequently loosening the structure to
release more pre-existing MPF. However, the mechanical stress
enacted by adding different number of steel balls did not
influence the number and length of MPF released for most
samples in our experiment. This may be because the timeframe
of the experiment was too short to observe any significant
impact induced by adding steel balls.
Figure 5. Characterization of MPFs released from interlock S (A) scissor-cut, 1st wash cycle; (B) scissor-cut, 10th wash cycle; (C) laser-cut, 1st
wash cycle; and (D) examples of scissor-cut and molten fiber ends. SEM images were taken under 2 kV, and the magnification was 1.80k.
Figure 6. (A) Influence of the cutting/sewing method on the number of MPFs released from two textiles during the 1st wash cycle. The mean
values are represented by black (laser-cut), red (double-sewn), blue (model-cut), green (scissor-cut), and purple (overlock-sewn) symbols. Error
bars indicate the standard deviation for three experimental replicates. (B) Length distribution of MPF released during the 1st wash cycle. The boxes
represent 25th and 75th percentiles with the median indicated by a line. Whiskers and outliers are not shown on the graph. The complete datasets
can be found in Figure S8.
Environmental Science & Technology pubs.acs.org/est Article
https://dx.doi.org/10.1021/acs.est.9b07395
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
F

Two aspects are suspected to contribute to the fact that the
scissor-cut samples released more MPFs than the laser-cut
textiles. On one hand, the cutting process may open up the
yarn at the end which results in the release of MPF present
inside. On the other hand, it is also possible that some new
MPFs are generated during the cutting processes. This can be
derived from the fact that some scissor-cut samples released
significantly longer MPFs than the laser-cut ones, which is an
indication that a new population of MPFs may be formed. For
the laser-cut samples, the yarns are melted together at the end
which largely prevent the “opening up”at the yarn end.
Moreover, even though the laser cutting can generate some
new MPFs similar to scissor cutting, the majority of those
newly generated MPFs will be probably melted/bonded into
“solid”ends. Therefore, most of them cannot be released in the
future washing process.
Most of the cutting/sewing methods (MC, DS, and OS) did
not significantly reduce the MPF release from textile during
washing, which is against our hypothesis that textiles with sewn
edge release less MPF. The overlock-sewn sample even
released higher numbers of MPF than the samples with raw
edges. The elevated level of release may be because of an
additional damage by cutting or needling during the OS
process. Because OS is one of the most commonly used
hemming/seaming methods in the textile industry, this finding
is highly relevant for those who want to identify the stages
responsible for the generation of MPF during the textile
manufacturing.
The evaluation of the results using the different cutting/
sewing methods can be used to gain additional insights into the
origin of MPF released from within the textile. The SEM
images show that the edges of laser-cut samples were molten
and sealed, while the scissor-cut edges were loose and open
(Table S2). Therefore, we can assume that all the MPFs
released from laser-cut samples originate from the textile
surface rather than the edges. For the scissor-cut textiles, it can
then be expected that release from the textile surface should be
similar to the laser-cut samples and that the additional fibers
measured are released from the edges. Therefore, a “release
equation”can be proposed for each textile (eq 1)
=
×
×
+
−
×
×
Release MPF
area mass area
MPF MPF
perimeter mass perimeter
textile
LC
sample sample
textile
SC LC
sample sample
textil
e
(1)
where “MPFSC
”and “MPFLC
”are the MPF releases from the
scissor- and laser-cut samples, respectively. The first term
describes the release from the surface, the second term the
release from the edge, and “releasetextile
”is the mass-specific
MPF release from a textile (in MPF/g). With this equation, the
number of MPF released from a larger piece of textile can be
predicted, as shown in Table 2. For 10 out of 12 textiles, the
MPFs shedding from edges contributed to more than 80% to
the total release. It is worth noting that the majority of MPFs
released from textiles with the processed surface still originated
from the cut edge, especially for plain B where the edge was
responsible for more than 90% of the MPF release. For the
edge-to-surface ratio, no statistically significant influence was
found by the surface treatment, the textile structure, and the
yarn type.
The 12 textiles investigated in our work demonstrated great
variability in MPF release. The upper limit observed in our
study is much higher than that found in other studies reporting
a range from 4 to 13,000 MPF (Figure S12). The length range
of the fibers released in our study (100−1000 μm) is in
accordance with the previous measurements.
22,24,27,29
The
MPF release from the microfiber (72,000 MPF/g) is the
highest among all the published studies in terms of the fiber
number. The generally higher number of MPF released in our
study can be explained by the higher edge-to-surface ratio
(0.70 cm) of our samples when compared to other Gyrowash
studies (with a ratio of 0.34−0.43 cm).
21,22,25
Because we have
found that the edge released much more MPFs than the
surface of the textile, the higher edge-to-surface ratio will result
in higher MPF release when the result is normalized to the
textile weight. Some previous studies reported lower release in
terms of number but higher mass release per wash as compared
to our study (Table S12, Figures S12, and S13). This may be
because a few studies only measured the weight of the released
MPFs and estimated the fiber number based on the average
length of a few MPFs. The mean value is influenced by
extremes which may not be accurate enough. Therefore, we
believe that it is conceptually better to work with a number-
based MPF release method.
Because our study has shown that the majority of the
released MPFs during washing likely originate from edges, as
opposed to the surface of the textile, adopting cleaner cutting
techniques is important for the industry to help reduce MPF
release. Furthermore, a prewash of tailored garments at the
factory could effectively collect a large portion of the
production-inherited MPFs. One thing to be kept in mind is
that the MPF release quantified here were only those released
during the laundering process. This may not accurately reflect
a real-world scenario because other processes such as wearing
processes are also involved and may account for additional
MPF releases in subsequent washes as the textile ages. In
addition, when standardizing the quantification method of
MPF release from textiles, it is extremely important to be
cautious about the cutting methods and the edge-to-surface
ratio. In addition, although the image analysis method is time
consuming, it is conceptually better than the mass-based
Table 2. Estimated Mass-Specific MPFs Released per cm2
Surface Area and per cm Perimeter From the Textile
Samples (4 ×10 cm2) during Washing When Fabric
Swatches Have a Scissor-Cut Edge
a
textile releaseedge
[MPF/(cm·g)] releasesurface
[MPF/(cm2·g)] releaseedge/releasesurface
[cm]
satin F 26 19 1
twill F 100 5 19
rib F 193 7 30
plain F 186 19 10
terry S 221 9 23
plain S 235 28 9
rib S 249 31 8
jersey S 291 23 12
interlock S 391 15 26
fleece 475 253 2
plain B 1251 58 22
microfiber 2001 408 5
a
Release from the fabric surface is measured from fabric swatches with
laser-cut edges. The ratio of the MPF release from the edges and
surface was calculated for the sample that has an area of 40 cm2and a
perimeter of 28 cm. releasesurface and releaseedge were calculated using
eq 1
Environmental Science & Technology pubs.acs.org/est Article
https://dx.doi.org/10.1021/acs.est.9b07395
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
G

method if we want to gain a complete profile of the MPF
release (number and length).
■ASSOCIATED CONTENT
*
sıSupporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.est.9b07395.
SEM images for 12 textiles; examples of cutting/sewing
methods; SEM images for edges, discussion, and
assessment of washing procedure; number of MPF
released during 1st wash cycle; length of MPF released
during 1st wash cycle; number of MPF released during
repeated wash cycles; median length of released MPF
during 1st and 10th wash cycles; estimated cumulative
release during 10 wash cycles; influence of mechanical
stress on number and length of MPF released from
textile; number and length of MPF released with
different cutting/sewing methods; comparison of MPF
release between washing and ultrasonication; and
comparison of results from different washing studies
(PDF)
■AUTHOR INFORMATION
Corresponding Author
Bernd Nowack −Technology and Society Laboratory, Empa
Swiss Federal Laboratories for Materials Science and
Technology, 901 St. Gallen, Switzerland; orcid.org/0000-
0002-5676-112X; Phone: +41 58 765 7692;
Email: nowack@empa.ch
Authors
Yaping Cai −Technology and Society Laboratory, EmpaSwiss
Federal Laboratories for Materials Science and Technology, 901
St. Gallen, Switzerland
Tong Yang −Technology and Society Laboratory, EmpaSwiss
Federal Laboratories for Materials Science and Technology, 901
St. Gallen, Switzerland
Denise M. Mitrano −Process Engineering Department,
EawagSwiss Federal Institute of Aquatic Science and
Technology, 8600 Dübendorf, Switzerland; orcid.org/0000-
0001-8030-6066
Manfred Heuberger −Laboratory for Advanced Fibers,
EmpaSwiss Federal Laboratories for Materials Science and
Technology, 9014 St. Gallen, Switzerland; orcid.org/0000-
0001-5799-3785
Rudolf Hufenus −Laboratory for Advanced Fibers, Empa
Swiss Federal Laboratories for Materials Science and
Technology, 9014 St. Gallen, Switzerland
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.est.9b07395
Notes
The authors declare no competing financial interest.
■ACKNOWLEDGMENTS
We would like to thank Melanie Halter from Empa for
assistance in the preparation of textile samples. This research
was partially supported by funds from the Zürcher Stiftung für
Textilforschung. D.M.M. was supported by the Swiss National
Science Foundation, Ambizione Grant number
PZP002_168105.
■REFERENCES
(1) Desforges, J.-P. W.; Galbraith, M.; Dangerfield, N.; Ross, P. S.
Widespread distribution of microplastics in subsurface seawater in the
NE Pacific Ocean. Mar. Pollut. Bull. 2014,79,94−99.
(2) Woodall, L. C.; Sanchez-Vidal, A.; Canals, M.; Paterson, G. L. J.;
Coppock, R.; Sleight, V.; Calafat, A.; Rogers, A. D.; Narayanaswamy,
B. E.; Thompson, R. C. The deep sea is a major sink for microplastic
debris. R. Soc. Open Sci. 2014,1, 140317.
(3) Kaya, A. T.; Yurtsever, M.; Bayraktar, S. C. Ubiquitous exposure
to microfiber pollution in the air. Eur. Phys. J. Plus 2018,133, 488.
(4) Ng, K. L.; Obbard, J. P. Prevalence of microplastics in
Singapore’s coastal marine environment. Mar. Pollut. Bull. 2006,52,
761−767.
(5) Zarfl, C.; Matthies, M. Are marine plastic particles transport
vectors for organic pollutants to the Arctic? Mar. Pollut. Bull. 2010,
60, 1810−1814.
(6) Rillig, M. C. Microplastic in terrestrial ecosystems and the soil?
Environ. Sci. Technol. 2012,46, 6453−6454.
(7) Van Cauwenberghe, L.; Claessens, M.; Vandegehuchte, M. B.;
Janssen, C. R. Microplastics are taken up by mussels (Mytilus edulis)
and lugworms (Arenicola marina) living in natural habitats. Environ.
Pollut. 2015,199,10−17.
(8) Nelms, S. E.; Galloway, T. S.; Godley, B. J.; Jarvis, D. S.;
Lindeque, P. K. Investigating microplastic trophic transfer in marine
top predators. Environ. Pollut. 2018,238, 999−1007.
(9) Anderson, P. J.; Warrack, S.; Langen, V.; Challis, J. K.; Hanson,
M. L.; Rennie, M. D. Microplastic contamination in Lake Winnipeg,
Canada. Environ. Pollut. 2017,225, 223−231.
(10) Su, L.; Xue, Y.; Li, L.; Yang, D.; Kolandhasamy, P.; Li, D.; Shi,
H. Microplastics in Taihu Lake, China. Environ. Pollut. 2016,216,
711−719.
(11) Baldwin, A. K.; Corsi, S. R.; Mason, S. A. Plastic Debris in 29
Great Lakes Tributaries: Relations to Watershed Attributes and
Hydrology. Environ. Sci. Technol. 2016,50, 10377−10385.
(12) Kawecki, D.; Nowack, B. Polymer-Specific Modeling of the
Environmental Emissions of Seven Commodity Plastics As Macro-
and Microplastics. Environ. Sci. Technol. 2019,53, 9664−9676.
(13) Henry, B.; Laitala, K.; Klepp, I. G. Microfibres from apparel and
home textiles: Prospects for including microplastics in environmental
sustainability assessment. Sci. Total Environ. 2019,652, 483−494.
(14) Browne, M. A.; Crump, P.; Niven, S. J.; Teuten, E.; Tonkin, A.;
Galloway, T.; Thompson, R. Accumulation of microplastic on
shorelines woldwide: sources and sinks. Environ. Sci. Technol. 2011,
45, 9175−9179.
(15) Sun, J.; Dai, X.; Wang, Q.; van Loosdrecht, M. C. M.; Ni, B.-J.
Microplastics in wastewater treatment plants: Detection, occurrence
and removal. Water Res. 2019,152,21−37.
(16) Schmiedgruber, M.; Hufenus, R.; Mitrano, D. M. Mechanistic
understanding of microplastic fiber fate and sampling strategies:
Synthesis and utility of metal doped polyester fibers. Water Res. 2019,
155, 423−430.
(17) Keller, A. S.; Jimenez-Martinez, J.; Mitrano, D. M. Transport of
nano-and microplastic through unsaturated porous media from
sewage sludge application. Environ. Sci. Technol. 2020,54, 911−920.
(18) Nizzetto, L.; Futter, M.; Langaas, S. Are agricultural soils
dumps for microplastics of urban origin? Environ. Sci. Technol. 2016,
50, 10777.
(19) Corradini, F.; Meza, P.; Eguiluz, R.; Casado, F.; Huerta-
Lwanga, E.; Geissen, V. Evidence of microplastic accumulation in
agricultural soils from sewage sludge disposal. Sci. Total Environ. 2019,
671, 411−420.
(20) Zubris, K. A. V.; Richards, B. K. Synthetic fibers as an indicator
of land application of sludge. Environ. Pollut. 2005,138, 201−211.
(21) Almroth, B. M. C.; Åström, L.; Roslund, S.; Petersson, H.;
Johansson, M.; Persson, N.-K. Quantifying shedding of synthetic
fibers from textiles; a source of microplastics released into the
environment. Environ. Sci. Pollut. Res. 2018,25, 1191−1199.
(22) De Falco, F.; Gullo, M. P.; Gentile, G.; Di Pace, E.; Cocca, M.;
Gelabert, L.; Brouta-Agné
sa, M.; Rovira, A.; Escudero, R.; Villalba, R.;
Environmental Science & Technology pubs.acs.org/est Article
https://dx.doi.org/10.1021/acs.est.9b07395
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
H

Mossotti, R.; Montarsolo, A.; Gavignano, S.; Tonin, C.; Avella, M.
Evaluation of microplastic release caused by textile washing processes
of synthetic fabrics. Environ. Pollut. 2018,236, 916−925.
(23) Hartline, N. L.; Bruce, N. J.; Karba, S. N.; Ruff, E. O.; Sonar, S.
U.; Holden, P. A. Microfiber Masses Recovered from Conventional
Machine Washing of New or Aged Garments. Environ. Sci. Technol.
2016,50, 11532−11538.
(24) Hernandez, E.; Nowack, B.; Mitrano, D. M. Polyester Textiles
as a Source of Microplastics from Households: A Mechanistic Study
to Understand Microfiber Release During Washing. Environ. Sci.
Technol. 2017,51, 7036−7046.
(25) Jonsson, C.; Arturin, O. L.; Hanning, A. C.; Landin, R.;
Holmstrom, E.; Roos, S. Microplastics Shedding from Textiles-
Developing Analytical Method for Measurement of Shed Material
Representing Release during Domestic Washing. Sustainability 2018,
10, 2457.
(26) McIlwraith, H. K.; Lin, J.; Erdle, L. M.; Mallos, N.; Diamond,
M. L.; Rochman, C. M. Capturing microfibers−marketed technolo-
gies reduce microfiber emissions from washing machines. Mar. Pollut.
Bull. 2019,139,40−45.
(27)Napper,I.E.;Thompson,R.C.Releaseofsynthetic
microplastic plastic fibres from domestic washing machines: Effects
of fabric type and washing conditions. Mar. Pollut. Bull. 2016,112,
39−45.
(28) Pirc, U.; Vidmar, M.; Mozer, A.; Kržan, A. Emissions of
microplastic fibers from microfiber fleece during domestic washing.
Environ. Sci. Pollut. Res. 2016,23, 22206−22211.
(29) Sillanpää, M.; Sainio, P. Release of polyester and cotton fibers
from textiles in machine washings. Environ. Sci. Pollut. Res. 2017,24,
19313−19321.
(30) Belzagui, F.; Crespi, M.; Álvarez, A.; Gutié
rrez-Bouzá
n, C.;
Vilaseca, M. Microplastics’emissions: Microfibers’detachment from
textile garments. Environ. Pollut. 2019,248, 1028−1035.
(31) De Falco, F.; Di Pace, E.; Cocca, M.; Avella, M. The
contribution of washing processes of synthetic clothes to microplastic
pollution. Sci. Rep. 2019,9, 6633.
(32) Kelly, M. R.; Lant, N. J.; Kurr, M.; Burgess, J. G. Importance of
Water-Volume on the Release of Microplastic Fibers from Laundry.
Environ. Sci. Technol. 2019,53, 11735−11744.
(33) De Falco, F.; Gentile, G.; Di Pace, E.; Avella, M.; Cocca, M.
Quantification of microfibres released during washing of synthetic
clothes in real conditions and at lab scale. Eur. Phys. J. Plus 2018,133,
257.
(34) Gries, T.; Veit, D.; Wulfhorst, B. Textile Technology: An
Introduction; Carl Hanser Verlag GmbH Co KG, 2015.
(35) Choudhury, A. K. R. Principles of Textile Finishing; Woodhead
Publishing, 2017.
(36) ISO. 105-C06:2010: TextilesTests for Colour Fastness Part
C06: Colour Fastness to Domestic and Commercial Laundering, 2010.
(37) Usov, I.; Mezzenga, R. FiberApp: an open-source software for
tracking and analyzing polymers, filaments, biomacromolecules, and
fibrous objects. Macromolecules 2015,48, 1269−1280.
(38) Kallay, N.; Grancari, A. M.; Tomi, M. Kinetics of polyester fiber
dissolution. Text. Res. J. 1990,60, 663−668.
(39) Chambers, J. M.; Cleveland, W. S.; Kleiner, B.; Tukey, P. A.
Graphical Methods for Data Analysis; Wadsworth and Brooks/Cole,
1983.
(40) Morton, W. E.; Hearle, J. W. S. Physical Properties of Textile
Fibres, 4th ed.; Woodhead Publishing Series in Textiles, 2008; pp
509−558.
(41) Hearle, J. W.; Lomas, B.; Cooke, W. D. Atlas of Fibre Fracture
and Damage to Textiles; Elsevier, 1998.
Environmental Science & Technology pubs.acs.org/est Article
https://dx.doi.org/10.1021/acs.est.9b07395
Environ. Sci. Technol. XXXX, XXX, XXX−XXX
I