Dataset: Particulate bulk and amino acid d15N values from EXPORTS cruises on RRS Discovery (DY131) and RRS James Cook (JC214) in May 2021

Particle sampling was conducted onboard RRS Discovery (size-fractionated particles) and James Cook (sinking particles) in the Northeast Atlantic (around 49°N, 15°W) during a declining spring phytoplankton bloom in May 2021 as part of the EXPORTS campaign. Particles were collected by in situ pump filtration using large volume pumps (WTS-LV; McLane Research Laboratories, Inc) outfitted with mini-MULVFS (Multiple Unit Large Volume in situ Filtration System) filter holders (Bishop et al., 2012). At 30 meters (m) and 340 m, sequentially stacked filters were loaded on pumps of 5 filter pore sizes: 0.3-1 micrometers (μm) particles (combusted GF75 glass fiber filters), 1-6 μm particles (combusted quartz fiber filters), 6-51 μm particles, 51-335 μm particles, and >335 μm particles (acid-washed Nitex mesh). At 20, 50, 75 or 95, 125 or 145, 175 or 195, 330, and 500 m, filter stacks of two filter pore sizes were loaded onto pumps: 0.8-51 μm particles (Supor polyethersulfone filters) and >51 μm particles (acid-washed polyester mesh). Filters containing particles were stored in combusted foil and kept frozen at -80 degrees Celsius (°C) until further processing on land. Large particles were resuspended off Nitex mesh filters using low-nutrient, 0.2 µm- filtered seawater, and subsequently collected onto 47-mm diameter combusted glass fiber filters, which were then re-frozen, and lyophilized. Lyophilized, large particles (≥6 µm) were examined under a dissecting microscope, and materials that should be excluded from the passively sinking POM pool were removed. Sinking particles were collected using sediment traps, either surface tethered sediment traps (STT) or neutrally buoyant sediment traps (NBST). Sinking particles were collected in polycarbonate tubes with a collection area of 0.0113 square meters (m²). Sinking particles were kept frozen at -80°C until further processing on land, then particles were lyophilized. Lyophilized particles were quantitatively split by weight for bulk isotope analysis and compound-specific isotope analysis of amino acids (CSIA-AA).

Bulk nitrogen concentrations and δ¹⁵N values were determined using standard methods in the Close Laboratory (University of Miami). Subsamples from filters for bulk nitrogen analysis were quantitatively split by weight and packed directly into tin capsules for elemental and isotopic composition.

CSIA-AA preparation was guided by methods of Hannides et al. (2013), Hannides et al. (2020), and Popp et al. (2007) and were identical to those of Wojtal et al. (2023). Samples were hydrolyzed using 6N HCl at 110°C for 20 hours, filtered using disc filters, purified using cation exchange resin (50W-X8, 100-200 mesh, 1 milliliter (mL) bed volume), eluting in 2N ammonium hydroxide, followed by reprotonation using 0.2N HCl at 110°C for 5 minutes, and derivatized to trifluoroacetyl/isopropyl esters in a two-step process (isopropyl esterification and trifluoroacetylation; Popp et al., 2007). Excess salts were removed after derivatization using liquid-liquid extraction (phosphate buffer/chloroform), and samples were put through a second acetylation before storage at -20°C prior to analysis. Immediately before analysis, sample aliquots were dried under N₂ gas and reconstituted in ethyl acetate. All glassware was acid washed (10% HCl) and combusted at 500°C overnight before use.

Bulk nitrogen concentrations and δ¹⁵N values were determined using a Costech Elemental Analyzer (EA) interfaced with a MAT 253 Plus Isotope Ratio Mass Spectrometer (IRMS) via a Conflo IV open split interface. The EA oxidation reactor was held at 980°C, the reduction reactor held at 650°C, and the oven housing the gas chromatography column was held at 65°C. CSIA-AA samples were analyzed using a Thermo Fisher Scientific Trace 1310 Gas Chromatograph (GC) coupled to a MAT 253 Plus IRMS via a Thermo Fisher GC Isolink II system with a combined oxidation/reduction reactor held at 1000 °C, a liquid nitrogen trap, and a Conflo IV open split interface.

δ¹⁵N values (‰ vs. AIR) were calculated by the Thermo Isodat 3.0 software, using reference nitrogen gas peaks. Instrument accuracy was determined by analyzing a mixture of amino acids of known δ¹⁵N values that was prepared alongside samples. Individual amino acid δ¹⁵N values were calculated as the average across replicates, if applicable. The 1σ uncertainty was calculated as the standard deviation across samples replicates, if no replicates were possible (low sample concentration) a 1‰ error was applied. Amino acid concentrations were determined from the IRMS peak area response of individual amino acid standards and the amount of sample injected.