Fish were brought from a local market in Shandong Province, collected from the Yellow Sea (detailed information of the fish species is provided as Supplementary Material). The digestive tract characteristics of the fish we used are different, such as the fat content and the thickness of the digestive tract. Furthermore, because of different feeding habits the intestinal contents were quite different. For example, some contents were mainly small fish, shrimp, crab, and other arthropods; and some were mainly aquatic algae. Because of these differences, even if one uses the same experimental protocol, each kind of fish’s digestion effect (the experimental result of the designed digestion experiment) is quite different. To eliminate such errors as much as possible, the tissues used in our study were derived from a mixture of all the fish digestive tracts and subsequently stored at −20°C.

The 11 types of plastics tested in this study were supplied by Shanghai Yichen Industrial Co., Ltd. These include acrylonitrile–butadiene–styrene (ABS), acrylonitrile–styrene copolymer (AS), high-impact polystyrene (HIPS), high- and low-density polyethylene (HDPE and LDPE, respectively), polyamide (PA), polyethylene (PE), polyethylene phthalate (PET), polypropylene (PP), polystyrene (PS), and polyvinyl chloride (PVC). The following plastic samples differed in color: milky white (ABS, PE, LDPE, and HDPE), transparent (AS, PS, and PVC), semitransparent (PP, HIPS, and PA), and white (PET).

Our initial investigations used particle sizes from ∼3 to 4 mm in diameter. After confirming the digestion protocol, we used MPs (>60–500 μm in diameter) for the recovery experiments.

Subtle effects of the digestion system on the plastics were taken into consideration. A cutting knife was used to make an irregular cross-shaped incision on the surface of the particles. After marking, the particles were rinsed with Milli-Q water and dried to a constant mass at room temperature (RT).

For the recovery rate experiments, the aforementioned 11 plastic particles were polished with a pulverizer (HCP-100) for ∼1 min and a temperature below 40°C. The ground plastic particles were passed through a series of sieves with different mesh sizes (60, 100, 200, and 500 μm) and sorted into four grain size classes (<60, 60–100, 100–200, and 200–500 μm). The samples were measured with a microscope for further size determinations.

Hydrogen peroxide (H₂O₂), nitric acid (HNO₃), perchloric acid (HClO₄), sodium hydroxide (NaOH), and ferrous sulfate (FeSO₄) were purchased from Sinopharm Chemical Reagent Co., Ltd. SDS was supplied by Kalmar-reagent. Aqueous solutions of NaOH (0.5 M), HNO₃ (11.5% v/v), FeSO₄ (5 mM), and SDS (150 g/L) were prepared in Milli-Q water. All the digestion solutions, including the acid digests (HClO₄ and HNO₃) and oxidizing solutions (H₂O₂), were mixed with Milli-Q water to obtain the following proportions solutions: 1:0, 4:1, 2:1, and 1:1 (v/v), which is the high-concentration group [69, 55.2, 46, and 34.5% (v/v), respectively]; and 1:2, 1:3, 1:4, and 1:5 (v/v), which is the low-concentration group [23, 17.25, 13.8, and 11.5% (v/v), respectively]. The filter was obtained from Yancheng Prich Laboratory Instrument Co., Ltd. We used a filter of the following specifications: pore diameter, 5 μm; and filter diameter, 50 mm.

The effects of the digestion were determined by comparing photos taken before and after the reaction. Note that because the digestion solution bubbled as soon as the reagent was added, it was not possible to photograph the solution. Therefore, an equal volume of Milli-Q water was used to represent the digestive reagent. A corresponding photo was taken to represent the initial state. Then the Milli-Q water was poured out and replaced with a digestive reagent. At the end of the digestion, a photo was taken of the final state. Specifically, the mixtures were first added to 60 mL of Milli-Q water (start point) in a breaker to take photos to represent the state of the starting point of digestion. Next, the water was replaced by the same volume of various types of digestion solutions, and the digestion was maintained at RT for 12 h, if necessary, through ultrasonic (JP-040S, 180W) and/or SDS treatment (end point). At the ending point of digestion, the visual examination was performed and photographed, comparing with the starting point photo the initial digestion effect can be recorded. And the efficiency of the digestions was measured by the frequency of changing the filter membrane and the filtration duration.

The impact of digestion on the plastic particles was investigated in terms of their shape and chemical composition. The MPs were removed from the beaker, washed with Milli-Q water 3 × to remove the impurities attached to the surface, and dried at RT for further examination.

In the initial investigation (sections “Unitary Digestion System,” “Binary Digestion System,” and “Protocol Optimization for Dispersion of Fat Problem”), the particles before and after the tests for the various digestion protocols were analyzed by a visual examination to check for signs of subtle changes in the cross marker (section “Plastic Preparation”) which used by the digestion treatments. For unbiased comparisons, the observation point—set exactly at the intersection—was analyzed with a Raman spectrometer (Horiba, LabRAM, HR Evolution) with a 5 × Olympus objective.

The O-PTIR spectra were collected on the MP surface with an effective spectral resolution of 2 cm^–1 and co-averaged for five spectral scans.

In field application investigations, the O-PTIR microscope was used to identify plastic particles. It is worth noting that the O-PTIR images were obtained at a 500-nm step size and 100-nm step resolution by tuning the QCL device to the frequencies that correspond to the specific vibrations of extraneous materials at various wavenumbers (1,023, 1,075, 1,641, and 1,725 cm^–1). The system was equipped with IR laser settings where the wavenumber was 1,247 cm^–1, the IP power was 4%, the pulse rate was 104.63 kHz, the duty cycle was 5%, and the pulse width was 480 ns. Regarding the probe settings, the probe power was 5% and the detector gain was 2 ×.

In the recovery experiments, we counted the recovery rates for four particle sizes, and each particle size was spiked with 30 particles. The digests were then filtered with filter paper. The quantity of the MPs on the filter membrane was used to calculate recovery percentages using Equation 1:

where R_a = number of MPs recovered; and R_i = initial number of MPs.

Fish species were collected from the Bohai Sea north of the Shandong Peninsula, and processed with the digestion method previously described and analyzed for MPs. The O-PTIR microscope was used to identify plastic particles.

We explored each digestion reagent separately. Accordingly, we mixed HClO₄ (70% v/v) and HNO₃ (69% v/v) with water separately in various proportions, as detailed in section “Chemicals.” One can visually observe the digestion of fish tissue and plastic particles in an aqueous acid solution. During digestion as per HClO₄, most of the plastic particles corroded yet the tissues were completely dissolved at various acid concentrations (Supplementary Figure 1). The use of HClO₄ in high concentrations is detrimental to nylon fibers (Claessens et al., 2013). We suggested that HClO₄ strongly corrodes plastics, rendering it unsuitable as a digestive agent. Regarding HNO_3, the corrosion of plastic particles by high concentrations of HNO₃ was obvious; the damage to PA was the most drastic. Low concentrations of HNO₃ had no obvious effects on the plastic particles except for PVC; only PVC had bubbles on its surface. However, fish tissue was poorly digested (Supplementary Figure 2). Among the plastics, we observed the most destruction for PA, which was completely destroyed by aqueous HNO₃ at proportions of 1:0, 2:1 and 4:1 v/v. As the concentration of HNO₃ decreased, PA initially dissolved, softened at lower concentrations, and at sufficiently low concentrations was undamaged. The results for other plastics were similar; that is, they retained their original shape when the concentration of HNO₃ was sufficiently low.

Our results are consistent with the literature of completely dissolved PA by acid treatment (Enders et al., 2017). Other polymers were affected to lesser extents, mainly color leaching and softening. Karami et al. (2017) digested biological materials in concentrated HNO₃ and produced optimum digestion efficiency only at RT. Concentrated HNO₃ had the most destructive effect on melted LDPE and PP fragments, and altered the color of most of the tested plastic polymers. Dehaut et al. (2016) recommended against using acid for studies of MPs because of its degradation of polyamide and its yellowing tendency to plastics; acid digestion may result in miscataloging of MPs during the optical examination. Thus, the use of acid to digest fish tissues for MP surveys is not straightforward. If acid treatment is to be implemented in practice, the experimental parameters (such as concentration and temperature) should be optimized or refer to Enders et al. (2017) to combine the treatment with other digestion solutions.

Compared with the results of acid treatment, the results from H₂O₂ treatment were mild. The fish tissue gradually dissolved as per visual observations at the H₂O₂ proportions of 1:0 and 4:1 (v/v). The tissue was less well digested in accordance with decreasing H₂O₂ concentrations. Karami et al. (2017) reported that H₂O₂ did not completely dissolve biological material at either RT or 40°C. Li et al. (2015) dissolved bivalve tissues in 30% (v/v) aqueous H₂O₂, incubated the mixture at 65°C for 24 h, and then digested the mixture at RT for 24∼48 h. The tissues dissolved but the quantity of tissues is an important parameter. Cole et al. (2014) used H₂O₂ (35% v/v) at RT for 7 days, but did not observe much digestion. Avio et al. (2015) reported incubation with H₂O₂ (15% v/v) at 50°C overnight. Thus, temperature is an important parameter in digestion. Although increasing the temperature results in a greater extent of tissue digestion, there will be deterioration of the plastic particles.

In summary, the use of various concentrations of H₂O₂ was promising in that there was almost no effect on the plastics. Nevertheless, low concentrations of H₂O₂ poorly digested fish tissues, especially the digestive tract (Supplementary Figure 3). Therefore, the application of H₂O₂ as a digestive solution for extracting MPs from fish tissues is limited and requires improvement.

We first mixed 11.5% (v/v) HNO₃ with 30% (v/v) H₂O₂ in various proportions to form a mixture, and the tissue and plastic particles were digested by the added mixture. The digestion protocol did not substantially affect the plastic polymers; we observed no changes in shape. In accordance with increasing proportions of HNO₃, the digestion of the fish digestive tract epidermis gradually worsened. There were tissue residues in all the solutions except at the proportion of 1:0 v/v, especially in the low-concentration groups (Supplementary Figure 5).

It is odd that none of the binary digestive systems (HNO₃ and H₂O₂ added in combination) were as effective as any of the unitary digestion systems (Supplementary Figures 1–3). We hypothesized that these results for the binary digestion system may be because of a chemical reaction between HNO₃ and H₂O₂; after the reaction, the concentrations of H₂O₂ and HNO₃ decrease, resulting in an inferior digestion. To test this hypothesis, we calculated the reaction between H₂O₂ and HNO₃ (Equation 2) using Gaussian modeling. Figure 1 shows the change in energy and material over the course of the reaction. The hypothetical reaction process and intermediate products in the reaction are shown in Figures 1A,B. As for Figure 1A, the A represents reactants HNO₃ and H₂O₂, which are optimized before calculation; as the reaction progresses, HNO₃ bond breaks and formed intermediate product B; the C represents the TS1 (NO₂, H⋅, OH⋅, and -O₂H) which was the transition state of the reaction, the D was the product of phase 1 which forming the final product NO₂, H₂O and the reactant of phase 2. For Figure 1B, the A represents the reactants OH⋅ and -O₂H, which undergo chemical bond breaking and recombination (B, C) to form the final products O₂ and H₂O. The specific reaction path is shown in Figure 1C. The modeling results indicate that H₂O₂ and HNO₃ react under certain conditions, which will affect the concentration of the solution in the digestion system and ultimately affect the experimental results.

Accordingly, to avoid a chemical reaction between HNO₃ and H₂O₂ that hinders digestion, we added 11.5% (v/v) HNO₃ and 30% (v/v) H₂O₂ in succession, in various proportions for tissue digestion. We first added H₂O₂ and then HNO₃. These treatments did not affect the polymers, whereas compared with the results in section “Added in Combination” the efficiency of the tissue digestion improved; large masses of tissue were not evident (Supplementary Figure 6). The proportion of 2:1 (v/v) afforded the best experimental results.

Although 11.5% HNO₃ and 30% H₂O₂ separately added in two steps can partly avoid both direct mixing reactions. The experimental results also indicated that, compared with directly adding, the digestion effect of tissues was significantly improved (Supplementary Figure 6), the fish tissues were not completely dissolved. The digestion efficiency was not as ideal as possible. This may be because of residual H₂O₂ in the digestion system, similarly to why the combination treatments were ineffective. Thus, we sought to decrease the quantity of residual H₂O₂. Accordingly, on the basis of the Fenton reaction principle, we added Fe²⁺ and NaOH after digestion by H₂O₂, yet before digestion by HNO₃. Other studies have also shown evidence that both enzymes and Fenton’s reagent could achieve efficient, rapid, and inexpensive digestion of organic matter and other materials. Dissolving biological matrices to analyze the gut contents of freshwater fish has literature precedent (Rodrigues et al., 2018; Collard et al., 2019).

After adding Fe²⁺ and NaOH, digestion was slightly improved compared with simply adding H₂O₂, perhaps because NaOH and Fe²⁺ catalyze digestion. Subsequent digestion as per H₂O₂–HNO₃ was considerably improved compared with the simple sequential H₂O₂/HNO₃ digestions, in which HNO₃ is used to react with the residual H₂O₂ rather than digestion. The plastics were not damaged.

Thus far, we had obtained optimum results for the following binary digestion: add 30% (v/v) H₂O₂ and allow to stand for 1 h; then add 10 mL of 0.5 M NaOH and 1 mL of 5 mM Fe²⁺, and allow to stand for 4 h; then add 11.5% (v/v) HNO₃ and allow to stand for 6 h. Nevertheless, residual fat remained in the filter membrane, which decreased the filtration efficiency and recovery. We thus needed to optimize the protocol for dispersion of fat problem.

We developed an efficient fish digestion protocol by visual inspection; it is difficult to conceive of a clear-cut quantitative metric. The literature indicates that the gravimetric method is pertinent to evaluating the efficiency of digestion, but compared with our method the weight change is not the most important factor. We found that the most important factor was the type of residue rather than the weight that affected the final test result, such as digestive tract dyspepsia. However, compared with the residual fat particles after digestion, the undigested digestive tract did not considerably interfere with the test results; the residual digestive tract can be cleaned on its surface and removed without participating in filtration. Therefore, by observing the digestion of fish tissues with unaided eyes, we can preliminarily test whether our digestion plan of fish tissues is feasible.

We applied our protocol to the detection of 368 fish samples. We greatly improved the filtration speed. The last step of our experiment was to filter the digestion solution and observe the types of plastic particles on the filter membrane; thus, the filtration time is an important indicator to evaluate our method. After the application of our protocol, 40% roughly of the sample filtration time was controlled within 20 min, followed by 10 min (30%), 30–60 min (20%), and longer than 60 min (10%), thus proving that our binary digestion system improves efficiency and saves time.

To further verify the feasibility of our protocol, its impact on MPs (micromorphology and chemical structure) should be evaluated. In order to improve the efficiency of the experiment, we used different methods to evaluate plastics in different experimental exploration stages. In initial investigations we used microscopes to observe changes in the details of the plastic and rule out methods that clearly affected the plastic. After obtaining a clear experimental protocol, we confirmed that the digestion process had no effect on the chemical structure of plastics by O-PTIR. The results for the recovery rate and field application finally proved that the method is feasible.

In initial investigations, we performed particle analyses by static image analyses with a Raman spectrometer. Artificial marks provide more details than smooth plastic particles; we made cross marks with a knife in each plastic particle. We used the position of the mark to compare readily identifiable changes before and after digestion. The cross marks on the surface of 11 types of plastic particles did not change appreciably after digestion compared with before digestion (Figure 3). The digestion solution did not corrode or dissolve the plastic particles.

We observed the changes in the chemical composition of the polymers by comparing the O-PTIR spectra before and after the digestion (Figure 4). The peak positions of the 11 plastics before and after the reaction were not obviously changed, especially the HIPS, LDPE, PET, ABS, AS, and PVC particles. However, we observed a few differences in the spectra of individual plastic particles compared with the standard sample, such as for PA, PE, and PS. For PE, the peak intensity at the carbonyl stretching region around 1,720–1,770 cm^–1 or 1,730–1,750 cm^–1 was lower than that of the standard spectrum. Using O-PTIR Studio software, we determined that the corresponding stretching peak of these three positions represents two residues. One is attributable to the C = O stretching peak in ClCOOC and the other is attributable to the C = O stretching peak in COO. For PA, although an expected peak was not evident at 1,372 cm^–1, this peak represents the C = O stretching peak of RCOO; which indicates that the plastic has been reduced to produce carboxylic acids. Similarly, as for PS, the corresponding peak change was at 1,638 cm^–1, which indicates a C = O or C = N stretching peak. The functional group represented by the changed characteristic peak is either C = O or C = N. These two kinds of chemical bonds indicate that the plastic particles underwent a redox reaction over the course of digestion. This is an acceptable experimental error; the digestion reagents had no obvious effect on the molecular structure of the plastics and did not affect the particle quantitation results. Therefore, the structures of the plastics did not change appreciably. The digestion protocol had little impact on the plastic particles and did not affect the detection of MPs in fish tissue.

In our study (Figure 5), the recoveries of 11 types of plastic particles larger than 60 μm in diameter each reached more than 87%, as same as diameter < 60 μm. Larger-sized plastics (200–500 μm) each reached 90%; the recoveries of ABS, PE, PET, HIPS, and LDPE (200–500 μm) all reached 100%; PP reached 103% and PET reached 110%. The recoveries of PP, PVC, PE, and HDPE (100–200 μm) also each reached 100%; PE reached 103% and PET reached 107%. In the small size of 60–100-μm plastic particles, only ABS and HIPS reached 100%. Regarding sizes less than 60 μm, the maximal recovery rate was 97%, which refers to PE and PVC. A recovery rate greater than 100% may be because of the influence of preexisting microplastics in the digestive glands of the fish. Similar results also appeared in Maes et al. (2017); Fraissinet et al. (2021) which is a relatively common situation. The recovery rate ranged from 87 to 100%, which indicated that the digestion protocol used in our experiments generally had relatively little influence on the recovery rate of MPs as a function of the type of plastic.

Based on the verification of the protocol, we next analyzed field-collected fish from Shandong Tantai.

As shown in Figure 6, we obtained single-frequency O-PTIR images at a 500-nm step size and 100-nm step resolution by tuning the QCL device to the frequencies that correspond to the specific vibrations of extraneous materials at various wavenumbers (1,023, 1,075, 1,641, and 1,725 cm^–1); detailed analysis of eight materials in Figure 6. In addition to PA, PET, and other plastic particles, we detected some crustacean bones and plastic additives.

We thus comprehensively confirmed the feasibility of our protocol. Our method is not only applicable to the digestion of fish tissues in the laboratory, but also applicable to practical analysis of real fish samples from the field, which has a practical value and promotion significance.