We have developed techniques for accurately and precisely measuring samples containing less than a few hundred micrograms of carbon, using a compact AMS system (NEC 0.5 MV 1.5SDH-1). Detailed discussions of the sample preparation, measurement setup, data analysis and background corrections for a variety of standard samples ranging from 0.002 to 1 mgC are reported. Multiple aliquots of small amounts of CO₂ were reduced to graphite with H₂ over pre-baked iron powder catalyst. A reduction reaction temperature of 450 °C was adopted for graphite samples below 0.05 mgC, rather than the usual 550 °C used on samples of 0.1–1 mgC. In our regular reactors (∼3.1 cm³), this reduction in temperature improved the graphite yield from ∼60 to 90–100% for samples ranging from 0.006–0.02 mgC. The combination of lower reaction temperature with a reduced reactor volume (∼1.6 cm³) gave yields as high as 100% on graphite samples <0.006 mgC. High performance measurements on ultra-small samples are possible also due to a modified NEC MC-SNIC ion-source that generates C⁻ currents of 1 μA per μg of carbon for samples in the 0.002 to 0.010 mgC range, combined with on-line measurement of ¹²C and ¹³C (AMS δ ¹³C) to correct machine-induced isotopic fractionation. Source efficiencies are in excess of 10%, which enables 4–5% of the radiocarbon atoms in 0.005–0.010 mgC samples to be measured. Examination of the background samples revealed two components: (a) 0.2–1 μg of modern carbon and (b) 0.1–0.5 μg of dead carbon. The latter component can be ignored when measuring unknown samples paired to small standards of precisely identical size (matching size normalizing standard method). Otherwise, one must make corrections for both background components. Ultra-small samples from 0.002 to 0.01 mgC can be measured with accuracy and precision of a few percent, based on scatter in results for multiple aliquots of a primary standard and deviations of secondary standards from their known values.

We report profiles of concentrations and radiocarbon contents of suspended particulate organic carbon (POC_susp) and sedimentary organic carbon from an abyssal site (Stn M) in the northeast (NE) Pacific collected in September 1994 (a period of very high flux of particulate carbon in the deep sea) and June 1995, as well as from stations on the continental rise and slope off the coast of California in June 1995. We show that during a period of anomalously high sinking POC flux to the deep sea (September 1994), Δ¹⁴C of suspended POC did not decrease detectably between 85 and 1600 m depth. This is in contrast to depth profiles during low and moderate fluxes of sinking POC at this station where Δ¹⁴C-POC_susp decreased 50–60‰ in this depth range. One explanation for the constant Δ¹⁴C values of POC_susp between 85 and 1600 m is that large quantities of sinking POC could continuously release labile, ¹⁴C-enriched POC_susp during biological and chemical alteration of the sinking POC. The radiocarbon evidence further suggests that resuspension of organic carbon from the sediment surface, either locally or laterally transported from the slope to the deep sea, is likely, but is probably limited to depths within a few hundred meters of the bottom. Sorption of ‘old’ DOC by suspended particulate matter in the water column is also possible, especially at shallower depths (<3500 m), though proof of this mechanism cannot be demonstrated at this time.