During the cruise SO296-1 on board of the German RV SONNE, we collected samples from the ship pumping system, connected to a filtration unit for microplastic analyses along the shelf of the eastern Pacific between Port Hueneme (USA) and Talcahuano (Chile). On the 0.7 µm GF filters between 300 and 560 L of surface water was filtered. The microplastic detection was carried out in the IOW laboratory using binocular microscope and the FTIR (LUMOS II, Bruker Optik GmbH, Germany). This data is reported together with surface temperature and salinity based on the continuous registration by thermosalinograph (https://doi.org/10.1594/PANGAEA.968975). Additional details about the cruise can be found in the cruise report (https://doi.org/10.48433/cr_so296_1), additional surface environmental data can be found at https://doi.org/10.1594/PANGAEA.988025.

Microplastic sampling, Detection, Polymer Identification: A total of 50 surface water samples were collected within the Pacific North-South transect of which 20 were analyzed for microplastics. The samples were collected via the ship's seawater intake pump, which was connected to an internal filtration system (0.7 μm glass fiber filter by Satorius, ∅ = 14.2 cm). A flow-meter monitored the volume of water filtered, in general 300 – 560 L were filtered. Following the filtration, each filter was immediately placed into an aluminum container with an aluminum-coated cardboard lid to prevent airborne contamination and dried in an oven at 50 ◦C for 24 h.For further analyses, two smaller sections (5 x 2.5 cm) were cut out of the obtained glass fiber filters loaded with the sample using a scalpel and a pre-rinsed cover glass. One of the sample sections were direly transferred to a glass slide and fixed with paper masking tape for microscopic identification of microfibers which were counted and photographed under a Binokular microscope equipped with a camera, while measuring the size with the open source program ImageJ 1.53o.The second sample section was transferred into a pre-rinsed glass beaker for further separation steps and microplastic analyses. To remove natural organic materials hydrogen peroxide was added into the beaker, covered with aluminum foil and incubated in the oven at 50°C for 24 h. During the last two hours, the samples were stirred using a magnetic stirrer, to homogenize the sample and the filter fibers. Subsequently, the same amount of pre-filtered acetic acid was added and left to react for 2 h at room temperature to remove calcareous components. After the reaction time, the samples were filtered using a stainless steel filter and rinsed with pre-filtered ultrapure water thoroughly. The obtained filter residues were then rinsed back into the glass beaker using the dense solution sodium polytungstate and transferred into a separation funnel to allow settling for 24 h after shaking the sample suspension thoroughly. At the following day, the lower sample part was discharged and the glass fiber free supernatant was filtered onto an Anodisc 25 filter. To ensure a higher recovery, the lower sample part containing heavier components was filled into the separation funnel to separate the supernatant for a second time. The obtained Anodisc filter was transferred into a glass Petri dish to store until further analyses.Microplastic identification was implemented by using a fully automated Fourier transform infrared spectroscope (LUMOS II, Bruker Optik GmbH, Germany), equipped with an 32x32 pixel Focal Plane Array (FPA) imaging detector cooled with liquid nitrogen and operated by using the Opus software (version 8.8). To obtain the most reliable results, five section covering 62.5% of each sample filter were selected and scanned in transmission mode at a spectral resolution of 4 cm−1 with 4 scan repetitions in a range of 3600 - 1250 cm−1. The detection limit of 1 pixel corresponds to a particle size of about 5 μm, but the lower microplastic size limit was set to 10 μm (minimum of 2 pixels) for reliability reasons. Because this method is strongly dependent on a plane filter area, 3 dimensional twisted microfibers are difficult to detect with the FPA-FTIR (Primpke et al., 2019, doi:10.1039/C9AY00126C). In addition to the spectroscopic measurement, microfibers were counted separately. The measured areas were further processed using the Purency software (version 4.17), which returns the microplastic numbers, polymer types, sizes, the specific correlations with internal reference spectra (similarity), as well as the reliability of the integrated machine learning algorithm (relevance). Only particles larger than 1 pixel (> 10 μm) and with similarity and relevance values ≥ 0.6 were accepted. Particles resulting in at least one of both values ≥ 0.6 were manually evaluated for their spectral fit.Temperature SBE 21 SEACAT from Thermosalinograph, range -5 to +35°C for 1st and 2nd sensor, accuracy 0.01 1st, and 0.001 2nd sensor, resolution 0.001 and 0.0003, respectivelyConductivity for salinity calculation SBE 21 SEACAT from Thermosalinograph, range 0 to 7, accuracy0.001, resolution 0.0001