Measuring tiny amounts of metal in seawater… not so easy on a metal ship

Metal-free bottles being deployed on a kevlar line to collect samples for dissolved metals in sea water
A towed ‘fish’ allows us to pump sea water onboard from a metal-free tube via a teflon pump.

Collecting seawater samples for trace metal analysis is different than sampling for other ocean parameters such as temperature, salinity, carbon and so on. Why? The concentration of metals in seawater is naturally very low (hence ‘trace’ metals) and it is important to avoid sample contamination from the ships environment which is full of possible contamination sources.

Niskin bottles are metal free sampling bottles designed for trace metal sampling. We hand-tie six bottles at spaced intervals to a Kevlar wire which allows us to collect water from different depths in the water column. In this vessel, the trace metal sampling cable can extend up to 2,000 meters and carry more than 2 000 kg’s. We cannot use the traditional rosette and winch system which is full of metal parts.

We lower the open bottles to the required depths and seal them with a messenger, a weight, which slides along the rope. Back onboard, the bottles are transported to a metal free ‘bubble’, a plastic room constructed within one of the ship’s labs. The bubble is kept at high pressure using filtered air, this way ship’s ‘dirty’ air does not enter the sampling environment. We also wear ‘clean suits’ to avoid introducing contamination from our clothing and hair.  

Sampling depths are chosen according to the depth of the seafloor, location of different water layers (water masses) and interesting biological features such as the depth of maximum phytoplankton abundance (assessed used fluorescence measurements). Trace metal sampling is always carried out after sensors have been deployed in the water column. This means that when the Niskin bottles are deployed we already have a good idea of the structure of the water column. Salinity, temperature, density, light, fluorescence, turbidity, dissolved oxygen and nitrate concentrations values are known from surface to bottom before trace metal sampling. Separate pressure and salinity sensors are attached to the Niskin bottles to confirm we sampled the intended depths.

The Niskin bottles are used to sample the water column but is not an efficient method to get high resolution surface data because the ship has to be stationery for each deployment. We therefore also use a ‘towfish’, a torpedo shaped sampler which is towed off the side of the ship while underway. Tubing connects the towfish to an onboard pump and finally to the bubble where a constant stream of surface seawater is flowing. The towfish is towed outside of the ships wake and all tubing is thoroughly cleaned with detergents and acid to avoid contamination. As a result, we can collect surface samples at rapid intervals without the ship stopping.