Marine Pollution Bulletin 198 (2024) 115795

Contents lists available at ScienceDirect

Marine Pollution Bulletin

journal homepage: www.elsevier.com/locate/marpolbul

Detecting small microplastics down to 1.3 μm using large area ATR-FTIR

Yuanli Liu a, b, c, d, *, Stephan Lüttjohann e, Alvise Vianello a, Claudia Lorenz a, Fan Liu a,
Jes Vollertsen a
a
Department of the Built Environment, Aalborg University, Thomas Manns Vej 23, 9220 Aalborg, Denmark
b
College of Environmental and Biological Engineering, Putian University, Putian 351100, China
c
Fujian Provincial Key Laboratory of Ecology-Toxicological Effects and Control for Emerging Contaminants, Putian University, Putian 351100, China
d
Key Laboratory of Ecological Environment and Information Atlas, Fujian Provincial University, Putian 351100, Fujian, China
e
Bruker Optics GmbH & Co. KG, Rudolf-Plank-Straße 27, 76275 Ettlingen, Germany

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

Keywords: Large area attenuated total reflectance-Fourier transform infrared spectroscopy (LAATR-FTIR) is introduced as a
LAATR novel technique for detecting small microplastics (MPs) down to 1.3 μm. Two different LAATR units, one with a
ZnSe zinc selenide (ZnSe) and one with a germanium (Ge) crystal, were used to detect reference MPs < 20 μm, and
Ge
MPs in marine water samples, and compared with μ-FTIR in transmission mode. The LAATR units performed well
Microplastics
Spectroscopy
in identifying small MPs down to 1.3 μm. However, they were poorly suited for large MPs as uneven particle
thickness resulted in uneven contact between crystal and particle, misinterpreting large MPs as many small MPs.
However, for more homogeneous matrices, the technique was promising. Further assessment indicated that there
was little difference in spectra quality between transmission mode and LAATR mode. All in all, while LAATR
units struggle to substitute transmission mode, it provides additional information and valuable information on
small MPs.

1. Introduction 2012; Käppler et al., 2018). Fourier Transform Infrared Spectroscopy

(FTIR) is by far the most widely used chemical detection technology,
Microplastics (MPs) have gained much attention over the last decade providing information about the chemical bonds of the analyzed mate­
or so (Thompson et al., 2004), and thousands of studies have been done rial (Cowger et al., 2020; Valls-Conesa et al., 2023). The FTIR spectra
to investigate the distribution, source, and fate of MPs to better under­ can be compared to reference libraries for identification (Gaffney et al.,
stand their potential threat (Bank and Hansson, 2019; Li et al., 2018; 2012), which are typically commercially available or custom-built for
Simon-Sánchez et al., 2022; Wang et al., 2021). Studies have revealed specific purposes. There are three signal acquisition modes commonly
that the presence of microplastics (MPs) in aquatic environments can used for the detection of MPs, mostly depending on the size range of the
have adverse impacts on ecosystems (Ding et al., 2022b), including particles targeted: attenuated total reflectance (ATR) (Cowger et al.,
potential implications for human health (Blackburn and Green, 2021), 2020; Veerasingam et al., 2021), transmission (Käppler et al., 2015;
highlighting the importance of understanding the occurrence of MPs in Löder et al., 2015), and reflectance (Harrison et al., 2012; Vianello et al.,
natural and anthropogenic environments. However, a lack of harmoni­ 2013). ATR-FTIR is routinely used to analyze larger particles (>300 μm)
zation of methods and quality control procedures makes the comparison mainly because of particle’s thickness and surface irregularity (Chand
among studies extremely challenging, hampering the scientific com­ et al., 2022; Liu et al., 2019). It requires limited sample preparation and
munity from obtaining a more comprehensive understanding of such allows the analysis of thick and irregularly shaped materials producing
multifaceted pollutants (Cui et al., 2022; Ding et al., 2022a; Liu et al., high-quality spectra not influenced by the thickness of the particle.
2021). However, this approach requires a manual pre-sorting of putative MPs,
The methodologies for detecting MPs have undergone significant after which each particle is analyzed by ATR-FTIR, resulting in a long
technological advancements, moving from simple visual inspection to analysis time. Smaller particles (<300 μm) are typically analyzed by
complex chemical detection (Fahrenfeld et al., 2019; Hidalgo-Ruz et al., μFTIR, a combination of microscopy and FTIR that characterizes MPs in

* Corresponding author.
E-mail address: [email protected] (Y. Liu).

https://doi.org/10.1016/j.marpolbul.2023.115795
Received 6 August 2023; Received in revised form 18 October 2023; Accepted 12 November 2023
Available online 25 November 2023
0025-326X/© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Y. Liu et al. Marine Pollution Bulletin 198 (2024) 115795

(a) (d)

Screws to control stage

Screws to fix crystal Screws to control stage

(e)
(b)
Screws to control stage
Screws to fix crystal

stage
stage

Screws to fix crystal

Screws to control stage

(c) (f)

Ø = 1800 m Ø = 1000 m

Fig. 1. Schematic of the LAATR set-up. LAATR with ZnSe crystal (a) set-up ready for mapping; (b) top view of the structure of the unit; (c) bottom view of the crystal;
LAATR with Ge crystal (d) set-up ready for mapping; (e) top view of the structure of the unit; (f) bottom view of crystal.

a contactless way (Corami et al., 2020; Possenti et al., 2021). Such solid immersion, which increases the magnification of the optics by a
analysis is typically carried out in transmission or reflection and allows factor of n, where n is the crystal’s refractive index. Prior studies have
the analysis of samples down to around 20 μm. The spatial resolution in demonstrated the efficacy of LAATR-FTIR in cultural heritage (Guarnieri
μFTIR is ultimately limited by the diffraction limit, which is determined et al., 2023), forensic science (Patterson and Havrilla, 2016), biology
by the wavelength of light and the numerical aperture of the microscope and medical science (Song and Kazarian, 2020) and material science
objective. Most modern instruments also allow chemical mapping or (Patterson et al., 2016). However, there is a dearth of research exploring
imaging, providing both spatial and chemical information of the the potential of LAATR-FTIR in analyzing small MPs or investigating the
analyzed sample (Löder et al., 2015). μFTIR-imaging allows to analyze performance of other crystal materials.
and characterize particles as small as, and sometimes even a bit smaller Small MPs (<100 μm) appear to dominate in the marine environ­
than, 10 μm when applying high-magnification Cassegrains (Corami ment (e.g., Liu et al., 2023), which highlights the importance of un­
et al., 2020; Kirstein et al., 2021; Possenti et al., 2021; Ye et al., 2022). derstanding small MPs in the environment as they are available to
To further enhance the performance of μFTIR-imaging, multi-detectors marine organisms and human beings (Matluba et al., 2023). When
such as linear array (LA) and focal plane array (FPA) detectors can be analyzing these samples from Kattegat, Denmark, with μ-FTIR in
coupled with the μ-FTIR system. This combination enables an automatic transmission mode, an increase in false positives was noticed when MPs
analysis of a sample over a large area and can collect up to millions of were below 20 μm, probably due to the thinness of such MPs. This made
spectra within a few hours (Dorling and Baker, 2013). μFTIR-imaging these small MPs suitable subjects to test the performance of the LAATR
operates with the same three collection modes as single-point μFTIR method with both zinc selenide (ZnSe) and germanium (Ge) crystals. To
(transmission, reflection, ATR). The transmission mode is the most address this knowledge gap, LAATR-FTIR with both crystals was
widely used collection mode when detecting environmental MPs (Löder employed to analyze reference MPs smaller than 20 μm and the above-
et al., 2015; Mintenig et al., 2017; Molazadeh et al., 2023). However, mentioned marine samples. The primary objective was to assess the
this process requires the sample as well as the substrate to be IR trans­ effectiveness of various analytical processes for identifying small MPs
parent and the sample must be sufficiently thin (roughly <100 μm) applying this technique.
(Käppler et al., 2016). Furthermore, a weak signal is acquired for thin
MPs, as the amount of absorbed IR light decreases with decreasing 2. Materials and methods
thickness, leading to a decreasing signal to noise ratio. This then in­
creases the risk of false detections for small MPs. 2.1. Sample and preparation
Large area ATR-FTIR (LAATR-FTIR) is a combination of ATR-FTIR
and μ-FTIR. It is a contact measurement that requires optical contact The study used reference MPs (<20 μm in size) produced by cryo-
between the sample and the unit. An ATR crystal is placed on top of a milling larger plastic fragments in the lab, as well as environmental
sample, then scanned in reflection mode through the crystal. The com­ samples of filtered marine water from Kattegat, Denmark (Liu et al.,
bination of ATR-FTIR and μ-FTIR enables the LAATR-FTIR to provide 2023). The reference MPs were polyethylene (PE), polypropylene (PP),
chemical mapping of samples and high-quality spectra upon optical and polyamide (PA) which were the most common polymer types found
contact, a typical advantage of ATR-FTIR. Another benefit that LAATR as fragments in the Kattegat (Liu et al., 2023). They were mixed and
inherits from ATR-FTIR is the enhanced spatial resolution, which is stored in 10 mL vials containing 50 % ethanol (HPLC grade, Sigma). The
significantly higher than μ-FTIR. This is due to the crystal’s refractive environmental samples previously mentioned were collected on 10 μm
index located on top of the sample. It functions as an additional lens or a stainless steel filters, and the MPs isolated by a series of biochemical and

2
Y. Liu et al. Marine Pollution Bulletin 198 (2024) 115795

physical treatment steps (Liu et al., 2023), after which the extracts were sample.
analyzed by μ-FTIR. These samples including polyester, PE, PP, PA and
other polymers, and similarly preserved in 10 mL vials containing 50 % 2.2. FTIR system and measurement parameters
ethanol (Liu et al., 2023).
Both the reference MPs and marine samples were homogenized on a 2.2.1. Transmission
vortex mixer and subsampled in increments of 50 and 100 μL using a The μ-FTIR used was a Bruker Lumos II microscope with 32 × 32
glass pipette (micro-classic, GmbH, Germany). The subsamples were pixels FPA Mercury Cadmium Telluride detector). The microscope was
then deposited on an optical substrate, a barium fluoride (BaF2) window equipped with an 8× Cassegrain objective, providing a 5.0 μm pixel
(Korth Kristalle GmbH, Ø 25 × 4 mm), using a metal ring (Ø = 2 mm) resolution in transmission and reflection. All scans were conducted with
and dried at 50 ◦ C. a spectral range of 4000–750 cm− 1 at a 4 cm− 1 resolution and were
The marine samples analyzed previously by Liu et al. (2023) and generated by 10 co-added scans. A background was created by co-adding
used in this study to look for smaller particles, were rather low on MPs. 10 scans.
For some analyzed windows no MPs were detected in neither trans­
mission mode nor LAATR mode. In total 14 windows with marine 2.2.2. LAATR units
samples were analyzed, and the sample with the highest abundance used Two commercially available LAATR units were applied. One which
to illustrate the LAATR results. This led to 4 windows being included in was equipped with a ZnSe crystal (PIKE technologies) and one which
the results, namely two with reference MPs, and two with environmental was equipped with a Ge crystal (Bruker). The LAATR units are

(a) (b) Transmission (c) LAATR (ZnSe)

50 m
50 m 50 m

(d) Transmission (e) LAATR (ZnSe)

PA
PE

50 m 50 m

(f) (g) (h)

50 m
Transmission 50 m LAATR (Ge) 50 m

(i) (j)

PA
PE
PP

Transmission 50 m 50 m
LAATR (Ge)

Fig. 2. (a) Visual image of reference MPs. Heat map (absorbance at 2900 cm− 1) of reference MPs detected by (b) transmission mode and (c) LAATR with ZnSe crystal.
Hyperspectral map (fit between spectra and reference library leading to detected MPs) of reference MPs detected by (d) transmission mode and (e) LAATR with ZnSe
crystal. (f) Visual image of reference MPs. Heat map of reference MPs detected by (g) transmission mode and (h) LAATR with Ge crystal. Hyperspectral map of
reference MPs detected by (i) transmission mode and (j) LAATR with Ge crystal.

3
Y. Liu et al. Marine Pollution Bulletin 198 (2024) 115795

(a) (b) (c)

50 m Transmission 50 m LAATR (ZnSe) 50 m

(d) (e)

PE

Transmission 50 m LAATR (ZnSe) 50 m

(f) (g) (h)

10 m 10 m 10 m

(i) (j)

PE
polyester

10 m 10 m

Fig. 3. (a) Visual image of marine sample. Heat map (absorbance at 2900 cm− 1) of marine sample detected by (b) transmission mode and (c) LAATR with ZnSe
crystal. Hyperspectral map (fit between spectra and reference library leading to detected MPs) of marine sample detected by (d) transmission mode and (e) LAATR
with ZnSe crystal. (f) Visual image of marine sample. Heat map of marine sample detected by (g) transmission mode and (h) LAATR with Ge crystal. Hyperspectral
map of marine sample detected by (i) transmission mode and (j) LAATR with Ge crystal.

accessories that are mounted on the stage of the FTIR instrument. 2.2.2.1. LAATR with ZnSe crystal (ZnSe unit). The setup of the LAATR
Spectra are acquired by scanning through the so-mounted unit. The two unit with a ZnSe crystal is shown in Fig. 1a, and the top view of the unit
units differ in detection areas and refractive indices, the latter affecting is shown in Fig. 1b. The top part with the ZnSe crystal is screwed to the
the achievable pixel size. stage and can be dismounted from the stage as shown in Fig. 1b. The
The detection limit of an LAATR unit can be calculated as follows: stage comprises various metals parts, with the former designated for
holding the sample, while the latter is rotatable for adjusting the sample
DFTIR
DLAATR = (1) (argent part) up and down (Fig. 1b). Additionally, two side screws can
RI LAATR
be used to move the sample stage around. The bottom part of the ZnSe
where DLAATR is the detection limit of the LAATR, DFTIR is the detection crystal has a diameter of approximately 1800 μm (Fig. 1c), which also
limit of the μ-FTIR, and RILAATR is the refractive index of the LAATR serves as the detection area. To make sure the data is reliable, only the
crystal. centermost 60 % was taken as detection area. With the settings applied,
Since the pixel size of the μ-FTIR used in this study was 5 μm, the it took around half hour to scan an area of 600 × 600 μm. The LAATR
detection limit of the ZnSe and Ge unit were 2.1 and 1.3 μm, given that had to be manually cleaned between each application.
the refractive index of ZnSe and Ge are 2.4 and 4, respectively (Eq. 1).
2.2.2.2. LAATR with Ge crystal (Ge unit). The setup of the LAATR unit

4
Y. Liu et al. Marine Pollution Bulletin 198 (2024) 115795

with a Ge crystal is shown in Fig. 1d. Like the ZnSe crystal setup, the top of the analysis, including the polymer type, Feret major and minor di­
part with Ge crystal can be dismounted from the stage, and two screws mensions, correlation score (maximum score), and mean signal-to-noise
can be used to fix the top crystal part (Fig. 1e). The stage is in the center ratio (S/N), were compiled in a table, along with hyperspectral maps
of the unit to hold the sample, and the surrounding components are and heat maps. The S/N was calculated by the ratio between the max
designed to secure the sample in place. The up and down movement of signal peak and the noise. The max signal peak was for all polymer types
the stage is controlled by the left bottom screw (Fig. 1e). The other two found in the wavenumber range 4000–950 cm− 1, while the noise
side screws allow the stage to move. The crystal has a diameter of around determination range was 3800–3600 cm− 1. The imaged MPs were
1000 μm (Fig. 1f). Similar as the ZnSe unit, only the centermost 60 % composed of pixels, and each pixel was assigned a value for the score
was taken as detection area. With the settings applied, it took about 1 h and the S/N. The pixel with the max score was defined as the max score
to scan an area of 600 × 600 μm. The LAATR had to be manually cleaned pixel, and the spectrum of this pixel displayed as the max score spec­
between each application. trum. The average spectrum was calculated by taking the average of all
spectra behind the pixels that constitute an MP. The mean S/N ratio for a
2.2.3. Detection process particle was calculated as the average value of all S/N values of the
As mentioned in Section 2.1, the window enriched with reference spectra that constitute an MP. After the analysis, all data were visualized
MPs or marine samples was first analyzed by transmission mode, fol­ using R (v4.0.3).
lowed with LAATR mode (ZnSe unit or Ge unit) to compare the results
from different modes. The transmission detection process covered the 3. Results and discussion
whole sample area (Ø 2000 μm), while the LAATR only covered Ø 1080
μm or Ø 600 μm. Since the detection area differed, the study area full of 3.1. Performance of LAATR units in detecting MPs
particles (the study areas differed because the distribution of particles
was uncontrollable) was selected to better explain the difference be­ To explore the effectiveness of LAATR units for MP identification, a
tween different detection modes. The comparison was done within the series of experiments were performed. The results of reference MPs
same window and on the same particles but with different detection detected by transmission mode and LAATR unit are shown in Fig. 2. For
modes. Two results are hence presented for each sample, labelled with the ZnSe unit, the study area is shown in Fig. 2a, while heat maps
transmission result and LAATR result, respectively. Then specific par­ (absorbance at 2900 cm− 1) and hyperspectral maps (fit between spectra
ticles were selected in the visual image, corresponding to the trans­ and reference library leading to detected MPs) are presented in
mission and LAATR result. Finally, the performance was compared Fig. 2b–e. The number of particles discernable in the heat maps was
based on these two results. similar between the two analytical methods (Fig. 2b–c). Applying
identical settings of the analytical software (e.g., thresholds), 5 MPs
were identified from spectra acquired in transmission mode, and 7 MPs
2.3. Data analysis (siMPle) for spectra acquired by the ZnSe unit (Fig. 2d–e). This result indicated
that the LAATR unit was more sensitive in detecting small MPs than the
All spectral data collected in this study were analyzed using the transmission mode. On the other hand, the MP sizes detected by the
software siMPle, which was developed by Aalborg University, Denmark, ZnSe unit were 161 % larger (on average) than those of the transmission
and Alfred Wegener Institute, Germany (Liu et al., 2019; Primpke et al., mode, except for one MP of PA (Table S2). This size difference can be
2020). To ensure comparability between all scan types, the effective explained by the pressure applied to the sample’s particles to obtain
wavenumber range was set to 4000–900 cm− 1. The lower range was optical contact, a necessary requirement when performing LAATR
chosen because the BaF2 windows increasingly absorb IR light below independently of the crystal used (Fig. 1h) causing the particles to be
approx. 900 cm− 1. Threshold parameters for comparing acquired squeezed and, therefore, reshaped. The results obtained from the com­
spectra to those of the references database, as well as all other software parison between Transmission-μFTIR and LAATR with the Ge unit were
settings used for all three modes, were identical (Table S1). The results

Fig. 4. Distribution of mean S/N, max score, and major dimension collected in the detection process including both reference MPs and marine samples by three
different modes. Each point represents an MP, and same MPs will have different max score and mean S/N in different modes. Each window was detected by both
transmission mode and LAATR mode (cf. Section 2.1), hence, the same particle was detected by both transmission and LAATR mode.

5
Y. Liu et al. Marine Pollution Bulletin 198 (2024) 115795

Fig. 5. Representative spectra of a) PA, b) PE from MPs detected by transmission mode and LAATR with ZnSe crystal; representative spectra c) PP, d) PA from MPs
detected by transmission mode and LAATR with Ge crystal. Colored spectra refer to max score spectra, and black spectra refer to average spectra, pink colored refer to
spectra from transmission, green colored refer to spectra from the ZnSe unit, blue colored refer to the Ge unit. Spectra in the same figure come from the same MP in
different modes. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)1

opposite (Fig. 2f–j). More MPs were identified in transmission mode, fragmentation. The Ge unit could not identify some large MPs of PE
which is also observed in the respective heatmap (Fig. 2g), while all identified by transmission mode, and the size of an MP of polyester was
these MPs were missing in the heat map obtained with Ge unit (Fig. 2h). smaller than found in transmission mode (Fig. 3f–j, Table S4). This result
This might be due to the difference in thickness of the particles used to again addressed the importance of homogenous pressure during the
perform the experiment, which led to a poor optical contact for some detection process. In addition, the MP size detected by the LAATRs was
MPs, resulting in a poor spectral quality and non-identification. The Ge on average 117 % larger than that in transmission mode (Table S4).
unit detected a small PP (major dimension: 21.6 μm, minor dimension: In summary, the LAATR units showed good performance in detecting
6.2 μm) that was overlooked in transmission mode (Table S3), high­ small MPs (Fig. 2d, e and i, j) and significant limitations in detecting
lighting the positive performance of the Ge unit in detecting small MPs larger ones (Fig. 3d, e and i, j). The main reason was the optical contact
when achieving sufficient optical contact. when performing LAATR, which lead to a seeming fragmentation of
The experiment conducted using marine environmental samples large particles, overlooking of small MPs when together with large MPs,
revealed an issue related to the LAATRs in detecting large MPs (Fig. 3). and reshapeing of MPs. These findings show figure and weaknesses of
The ZnSe unit failed to accurately define the borders of the large MPs, LAATR, which can provide higher spatial resolution (2.1 μm for ZnSe
leading to the incorrect identification of these large MPs as if they were and 1.3 μm for Ge) than transmission-μFTIR, but they only performed
broken down into multiple smaller ones (Fig. 3a–e). This can be well when samples consisted of small particles of homogeneous size.
explained by the pressure difference during the detection process, i.e.,
varying degree of optical contact, which then led to this seeming
3.2. Spectra quality exploration with LAATR mode and transmission
mode

1
Fig. 5(a) and (b) are not align with each other, neither with (c) and (d). Another advantage of LAATR over transmission-μFTIR is linked to

6
Y. Liu et al. Marine Pollution Bulletin 198 (2024) 115795

(a) PA (b)
33.647 0.576

5.176 0.089

(c) PE 18.814 (d) 0.567

2.894 0.087

(e) polyester (f)

24.605 0.698

3.785 0.107
(g) PP (h)
16.252 0.511

2.500 0.079
Fig. 6. Pixel analysis of S/N value (blue) and score value (red) within MPs. The lighter the blue color, the lower the signal to noise ratio of the spectra. The lighter the
red-orange color, the poorer the score between reference spectra and pixel spectrum. (a, b) PA from reference MPs with ZnSe unit; (c, d) PE from marine sample with
ZnSe unit; (e, f) polyester from marine sample with Ge unit; (g, h) PP from marine sample with Ge unit. (For interpretation of the references to color in this figure
legend, the reader is referred to the web version of this article.)

the fact that ATR is a surface technique and hence not affected by par­ no specific pattern between S/N ratio, max score, and particle size for
ticle thickness, allowing to analyze also very thin particles. While μ-FTIR MPs detected in transmission mode and LAATR modes (ZnSe and Ge
in transmission mode provides increasingly uncertain results when crystal). Low maximum scores were observed in cases where the mean
particles get very small (and thin). This was supported by experimental S/N ratio was high in transmission. Conversely, it was not possible to
observations: when dealing with increasingly small MPs, transmission- attain high max scores when the mean S/N was very low. To further
μ-FTIR increasingly misidentified natural material as MPs (Fig. S1), explore this, two representative spectra were selected from each ZnSe
while no false positives were found for the LAATR units, even when and Ge unit dataset (Fig. 5). The spectra of transmission and LAATR unit
basing the identification on a single pixel (Fig. S2). This is a consequence come from the same MP. For the mean S/N, three of the four trans­
of the better spectral quality obtained via LAATR, which ensures a more mission spectra had a higher S/N value than the corresponding LAATR
reliable chemical identification even for very fine particles. spectra. With respect to the max score, higher scores were observed in
To further assess if and to which extent LAATR offers better chemical the LAATR mode for two out of the four spectra. In some cases, a higher
mapping and higher quality spectra, two indicators, max score and mean mean S/N resulted in a lower max score (Fig. 5b and d), suggesting that
S/N ratio, were explored. All data of max score and mean S/N ratio there was no relationship between max score and mean S/N ratio.
collected in the detection process, including both reference MPs and Further exploration based on pixel analysis of S/N and score values of
marine samples, were summarized in Fig. 4. It is assumed that the max the pixels constituting maps of MPs was carried out, as presented in
score will increase with the increase of mean S/N. However, there was Fig. 6. The blue colors of the maps to the left indicate the S/N of each

7
Y. Liu et al. Marine Pollution Bulletin 198 (2024) 115795

pixel, while the red-orange color scale of the maps to the right indicates Credit authorship contribution statement
the score between the reference spectra used in the analysis and the
spectra acquired for each pixel. The distributions of these two parame­ Yuanli Liu: Writing – original draft preparation, Experimental sec­
ters, i.e., the color distribution, are clearly different, which validated the tion, and Visualization; Stephan Luettjohann: Supervision, Writing –
conclusion that there was no relationship between the S/N ratio and the reviewing and editing; Alvise Vianello: Supervision, Writing – reviewing
score value. The numerical data for this comparison is presented in and editing; Claudia Lorenz: Supervision, Writing – reviewing and
Fig. S3, which indicated that high scores could not be acquired with very editing; Fan Liu: Supervision, Writing – reviewing and editing; Jes
poor spectrum quality. Vollertsen: Supervision, Funding acquisition; Writing – reviewing and
editing.
3.3. Improvement of ZnSe and Ge units

Although the LAATR units showed good performance in detecting Declaration of competing interest
small MPs, some offset in the position of MPs in the acquired images was
noticed compared to what was detected in transmission mode. The ZnSe The authors declare that they have no known competing financial
crystal produced a larger offset than the Ge crystal, which could be interests or personal relationships that could have appeared to influence
related to the structure of the LAATR units (Fig. 1). For the ZnSe unit the work reported in this paper.
(Fig. 1a–c), both screws used to fix the crystal were rotatable, which
means the tightening process could have changed the position of the Data availability
unit. However, the sample remained stationary, resulting in an offset.
Additionally, the connection part on the side of the unit was too large for Data will be made available on request.
the screws, allowing the unit to move too easily. The sample stage also
had a rotatory control, which could raise and lower the sample. When Acknowledgements
raised, the ZnSe crystal received an upward force, and the rotatory nut
caused the crystal to move slightly, contributing to the offset. This project was supported by MONPLAS (European Union’s Horizon
Conversely, the Ge crystal had only one rotatable screw, limiting the 2020 research and innovation programme under the Marie Skłodowska-
offset to that screw and potentially explaining why the offset for the Ge Curie grant agreement No. 860775 (H2020-MSCA-ITN-2019).
unit was smaller than that of the ZnSe unit. Like the ZnSe unit, the
sample stage was controlled by a rotatory nut to regulate up-and-down Appendix A. Supplementary data
motion.
Considering the aforementioned comparisons, some suggestions for Supplementary data to this article can be found online at https://doi.
improvement are proposed. For the ZnSe unit, one screw could be fixed, org/10.1016/j.marpolbul.2023.115795.
as was the case for the Ge unit, to control the offset of the sample.
Additionally, the other screw could be made thicker to reduce the space References
between the screw and the rotary nut. The sample stage could also be
improved by incorporating a design that fixes the optical window with Bank, M.S., Hansson, S.V., 2019. The plastic cycle: a novel and holistic paradigm for the
anthropocene. Environ. Sci. Tech. 53 (13), 7177–7179. https://doi.org/10.1021/acs.
the sample, similar to the Ge unit. In this study, double-sided tape and est.9b02942.
modeling clay was used to secure the window and maintain the sample Blackburn, K., Green, D., 2021. The potential effects of microplastics on human health:
horizontality, serving the purpose of the experiments but limiting the what is known and what is unknown. Ambio 51 (3), 518–530. https://doi.org/
10.1007/S13280-021-01589-9.
application of the setup for more reliable and reproducible experiments.
Chand, R., Kohansal, K., Toor, S., Pedersen, T.H., Vollertsen, J., 2022. Microplastics
Moreover, the materials used to secure the samples, have the potential to degradation through hydrothermal liquefaction of wastewater treatment sludge.
cause cross-contamination of the sample. Thus, a better sample holder is J. Clean. Prod. 335, 130383 https://doi.org/10.1016/J.JCLEPRO.2022.130383.
Corami, F., Rosso, B., Bravo, B., Gambaro, A., Barbante, C., 2020. A novel method for
needed to improve the usability of the setup.
purification, quantitative analysis and characterization of microplastic fibers using
Micro-FTIR. Chemosphere 238, 124564. https://doi.org/10.1016/J.
4. Conclusion CHEMOSPHERE.2019.124564.
Cowger, W., Gray, A., Christiansen, S.H., DeFrond, H., Deshpande, A.D.,
Hemabessiere, L., Lee, E., Mill, L., Munno, K., Ossmann, B.E., Pittroff, M.,
This study is the first to investigate LAATR units to analyze MPs. It Rochman, C., Sarau, G., Tarby, S., Primpke, S., 2020. Critical review of processing
compared the performance of two different LAATR units (ZnSe and Ge and classification techniques for images and spectra in microplastic research. Appl.
crystal) to transmission mode in detecting reference MPs (<20 μm) and Spectrosc. 74 (9), 989–1010. https://doi.org/10.1177/0003702820929064.
Cui, Y., Liu, M., Selvam, S., Ding, Y., Wu, Q., Pitchaimani, V.S., Huang, P., Ke, H.,
MPs from marine water samples. The results showed that the LAATR Zheng, H., Liu, F., Luo, B., Wang, C., Cai, M., 2022. Microplastics in the surface
units were suitable for detecting small MPs due to their lower size limit waters of the South China Sea and the western Pacific Ocean: different size classes
(2.1 μm for ZnSe and 1.3 μm for Ge), while they performed poorly in reflecting various sources and transport. Chemosphere 299, 134456. https://doi.
org/10.1016/J.CHEMOSPHERE.2022.134456.
detecting large MPs. This result can be explained by the optical contact, Ding, J., Sun, C., He, C., Zheng, L., Dai, D., Li, F., 2022b. Atmospheric microplastics in
or lack hereof, during the detection process, which might lead to the Northwestern Pacific Ocean: distribution, source, and deposition. Sci. Total
reshapeing, seeming fragmentation, and overlooking of particles if the Environ. 829, 154337 https://doi.org/10.1016/J.SCITOTENV.2022.154337.
Ding, L., Huang, D., Ouyang, Z., Guo, X., 2022a. The effects of microplastics on soil
particles in the sample are not homogeneous in size. All this indicated
ecosystem: a review. Curr. Opin. Environ. Sci. Health 26, 100344. https://doi.org/
that LAATR units are limited in analyzing MPs in complex matrices, 10.1016/J.COESH.2022.100344.
usable for analyzing MPs in some clean matrices, and applicable as a Dorling, K.M., Baker, M.J., 2013. Rapid FTIR chemical imaging: highlighting FPA
detectors. Trends Biotechnol. 31 (8), 437–438. https://doi.org/10.1016/J.
supplementary technology together with μ-FTIR in transmission mode.
TIBTECH.2013.05.008.
The spectrum quality was quite similar between transmission and Fahrenfeld, N.L., Arbuckle-Keil, G., Naderi Beni, N., Bartelt-Hunt, S.L., 2019. Source
LAATR-acquired spectra, but the LAATR units had a better accuracy in tracking microplastics in the freshwater environment. TrAC Trends Anal. Chem. 112,
terms of avoiding false positives. In conclusion, this study demonstrates 248–254. https://doi.org/10.1016/J.TRAC.2018.11.030.
Gaffney, J.S., Marley, N.A., Jones, D.E., 2012. Fourier transform infrared (FTIR)
the potential applicability of LAATR units to, in our case, detect MPs spectroscopy. Mater. Sci. 1–33 https://doi.org/10.1002/0471266965.COM107.
down to 1.3 μm, offering a promising analytical approach to study small PUB2.
MPs in environmental matrices. Guarnieri, N., Ghirardello, M., Goidanich, S., Comelli, D., Dellasega, D., Cotte, M.,
Fontana, E., Toniolo, L., 2023. Imaging and micro-invasive analyses of black stains
on the passepartout of Codex Atlanticus Folio 843 by Leonardo da Vinci. Sci. Rep. 13
(1), 1–12. https://doi.org/10.1038/s41598-023-31129-2.

8
Y. Liu et al. Marine Pollution Bulletin 198 (2024) 115795

Harrison, J.P., Ojeda, J.J., Romero-González, M.E., 2012. The applicability of reflectance Molazadeh, M., Liu, F., Simon-Sánchez, L., Vollersten, J., 2023. Buoyant microplastics in
micro-Fourier-transform infrared spectroscopy for the detection of synthetic freshwater sediments — how do they get there? Sci. Total Environ. 860, 160489
microplastics in marine sediments. Sci. Total Environ. 416, 455–463. https://doi. https://doi.org/10.1016/J.SCITOTENV.2022.160489.
org/10.1016/J.SCITOTENV.2011.11.078. Patterson, B.M., Havrilla, G.J., 2016. Attenuated total internal reflection infrared
Hidalgo-Ruz, V., Gutow, L., Thompson, R.C., Thiel, M., 2012. Microplastics in the marine microspectroscopic imaging using a large-radius germanium internal reflection
environment: a review of the methods used for identification and quantification. element and a linear array detector. Appl. Spectrosc. 60 (11), 1256–1266. https://
Environ. Sci. Tech. 46 (6), 3060–3075. https://doi.org/10.1021/es2031505. doi.org/10.1366/000370206778998941.
Käppler, A., Windrich, F., Löder, M.G.J., Malanin, M., Fischer, D., Labrenz, M., Patterson, B.M., Havrilla, G.J., Marcott, C., Story, G.M., 2016. Infrared
Eichhorn, K.J., Voit, B., 2015. Identification of microplastics by FTIR and Raman microspectroscopic imaging using a large radius germanium internal reflection
microscopy: a novel silicon filter substrate opens the important spectral range below element and a focal plane array detector. Appl. Spectrosc. 61 (11), 1147–1152.
1300 cm− 1 for FTIR transmission measurements. Anal. Bioanal. Chem. 407 (22), https://doi.org/10.1366/000370207782596969.
6791–6801. https://doi.org/10.1007/S00216-015-8850-8/FIGURES/11. Possenti, E., Colombo, C., Realini, M., Song, C.L., Kazarian, S.G., 2021. Time-resolved
Käppler, A., Fischer, D., Oberbeckmann, S., Schernewski, G., Labrenz, M., Eichhorn, K.J., ATR-FTIR spectroscopy and macro ATR-FTIR spectroscopic imaging of inorganic
Voit, B., 2016. Analysis of environmental microplastics by vibrational treatments for stone conservation. Anal. Chem. 93 (44), 14635–14642. https://doi.
microspectroscopy: FTIR, Raman or both? Anal. Bioanal. Chem. 408 (29), org/10.1021/acs.analchem.1c02392.
8377–8391. https://doi.org/10.1007/S00216-016-9956-3/TABLES/3. Primpke, S., Cross, R.K., Mintenig, S.M., Simon, M., Vianello, A., Gerdts, G.,
Käppler, A., Fischer, M., Scholz-Böttcher, B.M., Oberbeckmann, S., Labrenz, M., Vollertsen, J., 2020. Toward the systematic identification of microplastics in the
Fischer, D., Eichhorn, K.J., Voit, B., 2018. Comparison of μ-ATR-FTIR spectroscopy environment: evaluation of a new independent software tool (siMPle) for
and py-GCMS as identification tools for microplastic particles and fibers isolated spectroscopic analysis. Appl. Spectrosc. 74 (9), 1127–1138. https://doi.org/
from river sediments. Anal. Bioanal. Chem. 410 (21), 5313–5327. https://doi.org/ 10.1177/0003702820917760.
10.1007/S00216-018-1185-5/TABLES/2. Simon-Sánchez, L., Grelaud, M., Lorenz, C., Garcia-Orellana, J., Vianello, A., Liu, F.,
Kirstein, I.V., Hensel, F., Gomiero, A., Iordachescu, L., Vianello, A., Wittgren, H.B., Vollertsen, J., Ziveri, P., 2022. Can a sediment core reveal the plastic age?
Vollertsen, J., 2021. Drinking plastics? Quantification and qualification of Microplastic preservation in a coastal sedimentary record. Environ. Sci. Technol.
microplastics in drinking water distribution systems by μFTIR and Py-GCMS. Water https://doi.org/10.1021/ACS.EST.2C04264.
Res. 188, 116519 https://doi.org/10.1016/J.WATRES.2020.116519. Song, C.L., Kazarian, S.G., 2020. Micro ATR-FTIR spectroscopic imaging of colon biopsies
Li, J., Liu, H., Paul Chen, J., 2018. Microplastics in freshwater systems: a review on with a large area Ge crystal. Spectrochim. Acta A Mol. Biomol. Spectrosc. 228,
occurrence, environmental effects, and methods for microplastics detection. Water 117695 https://doi.org/10.1016/J.SAA.2019.117695.
Res. 137, 362–374. https://doi.org/10.1016/J.WATRES.2017.12.056. Thompson, R.C., Olson, Y., Mitchell, R.P., Davis, A., Rowland, S.J., John, A.W.G.,
Liu, F., Vianello, A., Vollertsen, J., 2019. Retention of microplastics in sediments of urban McGonigle, D., Russell, A.E., 2004. Lost at sea: where is all the plastic? Science 304
and highway stormwater retention ponds. Environ. Pollut. 255, 113335 https://doi. (5672), 838. https://doi.org/10.1126/SCIENCE.1094559/SUPPL_FILE/
org/10.1016/J.ENVPOL.2019.113335. THOMPSON.SOM.PDF.
Liu, M., Ding, Y., Huang, P., Zheng, H., Wang, W., Ke, H., Chen, F., Liu, L., Cai, M., 2021. Valls-Conesa, J., Winterauer, D.J., Kröger-Lui, N., Roth, S., Liu, F., Lüttjohann, S.,
Microplastics in the western Pacific and South China Sea: spatial variations reveal Harig, R., Vollertsen, J., 2023. Random forest microplastic classification using
the impact of Kuroshio intrusion. Environ. Pollut. 288, 117745 https://doi.org/ spectral subsamples of FT-IR hyperspectral images. Anal. Methods 15 (18),
10.1016/J.ENVPOL.2021.117745. 2226–2233. https://doi.org/10.1039/D3AY00514C.
Liu, Y., Lorenz, C., Vianello, A., Syberg, K., Nielsen, A.H., Nielsen, T.G., Vollertsen, J., Veerasingam, S., Ranjani, M., Venkatachalapathy, R., Bagaev, A., Mukhanov, V.,
2023. Exploration of occurrence and sources of microplastics (>10 μm) in Danish Litvinyuk, D., Vethamony, P., 2021. Contributions of Fourier transform infrared
marine waters. Sci. Total Environ. 865, 161255 https://doi.org/10.1016/J. spectroscopy in microplastic pollution research: a review. Crit. Rev. Environ. Sci.
SCITOTENV.2022.161255. Technol. 51 (22), 2681–2743. https://doi.org/10.1080/10643389.2020.1807450.
Löder, M.G.J., Kuczera, M., Mintenig, S., Lorenz, C., Gerdts, G., 2015. Focal plane array Vianello, A., Boldrin, A., Guerriero, P., Moschino, V., Rella, R., Sturaro, A., Da Ros, L.,
detector-based micro-Fourier-transform infrared imaging for the analysis of 2013. Microplastic particles in sediments of Lagoon of Venice, Italy: first
microplastics in environmental samples. Environ. Chem. 12 (5), 563–581. https:// observations on occurrence, spatial patterns and identification. Estuar. Coast. Shelf
doi.org/10.1071/EN14205. Sci. 130, 54–61. https://doi.org/10.1016/J.ECSS.2013.03.022.
Matluba, M., Ahmed, M.K., Chowdhury, K.M.A., Khan, N., Ashiq, M.A.R., Islam, M.S., Wang, X., Bolan, N., Tsang, D.C.W., Sarkar, B., Bradney, L., Li, Y., 2021. A review of
2023. The pervasiveness of microplastic contamination in the gastrointestinal tract microplastics aggregation in aquatic environment: influence factors, analytical
of fish from the western coast of Bangladesh. Mar. Pollut. Bull. 193, 115145 https:// methods, and environmental implications. J. Hazard. Mater. 402, 123496 https://
doi.org/10.1016/J.MARPOLBUL.2023.115145. doi.org/10.1016/J.JHAZMAT.2020.123496.
Mintenig, S.M., Int-Veen, I., Löder, M.G.J., Primpke, S., Gerdts, G., 2017. Identification of Ye, Y., Yu, K., Zhao, Y., 2022. The development and application of advanced analytical
microplastic in effluents of wastewater treatment plants using focal plane array- methods in microplastics contamination detection: a critical review. Sci. Total
based micro-Fourier-transform infrared imaging. Water Res. 108, 365–372. https:// Environ. 818, 151851 https://doi.org/10.1016/J.SCITOTENV.2021.151851.
doi.org/10.1016/J.WATRES.2016.11.015.