Company: Thermo. Link to website("Gas Isotope Systems" only, no link to their H-device).

The H-device is a chromium reduction furnace that operates in viscous flow mode for the H-D analysis of water (and small organic molecules). The system is pumped out, and a CTC Pal autosampler is used to inject the sample through the septa and into the evacuated hot Cr. The water reacts with the Cr forming CrOx and releasing H as H2 gas. Once the reaction is complete (timed based), the valves open and the H2 is allowed into the evacuated bellows. After equilibration, the outer valves to the bellows close, and the pressure in the sample, standard gas, or both is adjusted to the set voltage and the analysis begins. This is run in classic dual inlet (viscous flow) mode.
The syringe must be rinsed thoroughly between runs, and allowed to dry. There is a memory effect, but that is mainly in the furnace itself. As the reactor gets used up, less H2 seems to be retained, so the apparent yields (initial pressure in bellows before auto adjusting) will go up--until the reactor fails. Typical runs use 0.5 to 1 uL sample injections plus more for needle rinses, though larger sample sizes, 1 mL +, are needed so the sample can be picked up by the syringe in the autosampler. A gastight syringe is needed since you are injecting into an evacuated furnace. You can get hundreds of runs in with a given furnace and septa, but in smaller batches (60+). The standard gas in the other bellows will run out, and the syringe will need cleaning.
The sample remaining in the vial after analysis can then be used for other analyses, e.g. 18O or redoing DH, with minimal contamination or drift--but you MUST test your set up. A 2 mL aliquot in the 2 mL sample vial is fairly stable, even after it's initial H-device run. A 500 uL sample or a 200 uL sample in an insert may drift a bit after the septum is pierced. This will be time, sample, septum and syringe needle dependent.

Read more about it:

DH on water:

CTC Pal autosampler

Tips and Tricks: