The seawater is collected by Niskin bottles mounted on a CTD rosette. Upon recovery to the ship's deck, the water is transferred to pre-cleaned 20 L carboys through silicone tubing, directly from the Niskin bottle sample valve to a bulkhead fitted carboy cap. Transfer takes only a few minutes through one of the larger sampling valves. These steps have proven contamination-free.

The sealed carboys are returned to the shore laboratory immediately after collection, acidified with blank-checked, ultra-pure hydrochloric acid (2 ml HCl per liter of water). The water is tested at each transfer step to ensure absence of contamination. Pre-combusted vials are filled from a silicone tube attached to the carboy valve. The other end of the silicone tube is attached to a clean, long-nosed pipette. The pipette tip is placed cleanly and without flow at the bottom of the vial, and then the flow is started.

Low carbon water is generated in a Milli-Q A10 system. The system reduces DOC in deionized water to very low levels by utilizing high intensity, multi-wavelength (185 and 284 nm) UV light. Specifications call for the unit to generate water with a DOC content of <0.4 µM C (<5 ppb). The water is homogenized, acidified and distributed from a clean carboy as described above.

During vial filling, we randomly but regularly remove vials for immediate analysis to evaluate for contamination. Evaluation of these subsets ensures that there is little variability between batches of material prepared on different days. The volumes (20 and 40 ml) are suitable for DOC analyses by direct injection (normally 100 µl) into high temperature combustions systems and by chemical or photo-oxidation techniques.

Reference Material Consensus – Consensus on the concentration of organic carbon in the deep sea reference water is performed by a group of independent expert laboratories. We maintain a pool of 4-6 labs for each assessment. The CRM program is intended to run for several years, but the longevity of any one laboratory may not be that long. A group of 4-6 certification labs allows for loss of one or two labs, and replacement by other labs, while maintaining continuity over the years. Currently, the laboratories of Craig Carlson, Dennis Hansell, Hiroshi Ogawa, Elizabeth Minor, Chiara Santinelli and Rainer Amon provide the service.

Shortly after batch preparation, and upon successful reanalysis of the CRM’s by our laboratory, each expert/consensus laboratory receives 10-20 samples of the low carbon water (LCW) and of reference seawaters. Analyses of the concentration of DOC in the seawater is relative to the LCW, with each laboratory using its own standards and calibration procedures. The value presently assigned to the LCW is 1 µM.

CRM Storage and Distribution - The newly prepared samples are stored in the dark at room temperature. Samples are not frozen because it is unnecessary, because we wish to avoid alternating freezing/thawing/freezing cycles likely for the more remote recipients of the material, and because freezing weakens (through expansion of the water) the integrity of the glass vial walls.

Shipment to the users is by ‘overnight’ delivery in shock-protected boxes. Shipping costs are borne by the participating laboratory unless special circumstances make such payments (and, therefore, participation in the program) difficult. For example, some laboratories in less developed countries may not be able to afford the shipping costs. We find ways to keep such laboratories in the program.

CRM Stability - Our experience is that changes in DOC have not been seen for either the seawater or the LCW on the time frame of at least 2 years. It could be, however, that the samples undergo change on longer time scales, perhaps due to absorption of DOM onto the glass walls. At the time of distribution, we assign shelf lives of 1 year to the reference material.

CRM Use - We distribute a document describing the appropriate handling of reference materials received by the participating laboratories. In it, the purpose of the reference materials, a description of the material distributed, the recommended protocols for handling the materials, cautions, and appropriate statistical tools for evaluating results are listed.