Method Development: Yttrium

An ongoing topic of research in Oceanography involves the cycling and sequestration of carbon within the world’s oceans. While it should be no surprise that such a topic is the focus of many labs, including my own, it may be hard to believe that scientists are still grappling with how to best measure the exchange of carbon between the ocean’s surface and subsurface layers.

The classic and intuitive method involves placing a sort of “carbon rain gauge” below the surface layer. This device, called a sediment trap, captures the sinking carbon particles for layer analysis, which can provide information on the rate of carbon flux to depth, the composition and origin of the particles, as well as a wealth of chemical and genomic information. With such a tool at our disposal, it may seem that the question of how much carbon sinks out of the surface waters is solved; yet, as with most answers in science, there are many caveats. While we wont go into them here, the result is that scientists need additional methods and tools to help unify the bits and pieces that sediment traps provide us with.

A chemical technique that has been used for the past decade or two involves using a naturally occurring radioactive isotope to determine the amount of sinking particles. This method begins with Uranium 238, which occurs naturally in the crust and which exists at moderate (relative to other trace elements) concentrations in all seawater. Since the half-life of this isotope (U-238) is quite long lived, ~3.8 billion years, you can think of the concentration of U-238 as uniform in the oceans (ocean are mixed much quicker than it decays away). Additionally, Uranium in seawater forms a soluble and inert ion $U(OH)_4^-$ meaning that it doesn’t bond with particles or anything like that.

Thus U-238 decays down a chain of elements before finally becoming Lead: its stable end product. Along the decay chain, Thorium-234 is created which has a half-life of just 24 days. Since the oceans take much longer than 24 days to mix (they take 1,000s of years), there is an opportunity for the Thorium distribution to change from being the same everywhere (like Uranium) to existing in higher concentrations in some places and lower concentrations in others (heterogeneous).

Additionally, the chemical properties of Thorium are quite distinct from Uranium. While Uranium is virtually inert in seawater, Thorium is quite reactive and has a high affinity for organic matter. Due to this affinity, almost all the Thorium in seawater will bind to particles given the opportunity. So what scientists realized is that if you measure the amount of Th-234 in the surface waters, it will be lower than the concentration that you would expect based on how much U-238 is there. It will be lower since some of the particles, and therefore some of the Th-234, has sank out of the surface waters in the past month (i.e. in the past couple half lives). So this system calculates the number of sinking particles through the “disequilibrium” of Th-234 in the surface waters. While the Thorium method also has its drawbacks, it works well in collaboration with sediment traps by providing complementary and independent estimates of this export flux of carbon.

My most recent project is attempted to develop an extension of the Thorium method for use in most dynamic settings such as in coastal waters or at fronts. My new method uses Yttrium-90 instead of Thorium-234, yet the mechanics themselves are quite similar. A benefit of a Y90 method over the Thorium technique would be a much better temporal resolution in the measurement since the Thorium-derived signal comes from export over the past month (half-life of 24 days), the Y90 signal would come from the past few days (half-life of 3 days).

A shorter half-life is great and all, but it comes at a high, analytical cost: concentration. Whereas the concentration of Th-234 is such that 4L of seawater (~1 gal) is enough to get an accurate reading, the concentration of Y-90 is much less (it only lasts 3 days vs 24 days) so that anywhere from 30-60+ L (8-15+ gal) of seawater are required. The shorter half-life also implies that the time frame of any measurement must be much shorter than those 3 days (i.e. time is precious).

Nevertheless, progress for my method is being made and during this cruise I’ve had several opportunities to try it out. While I’m currently limited to ~20 L, it has been enough to start seeing a signal.

A preliminary (and I mean preliminary) analysis of my second sample[1] showed that I could back-out a Y-90 signal from the bulk beta-decay counts, which is the necessary analysis for this method. Details of this analysis and the method itself will be written up by itself in another post.

The next steps for my method will involve scaling up the system by ~2x and improving the physical engineering of the filtering rig to provide more control and a more robust setup. Ultimately I’m confident that I can make progress, and probably more importantly I know what my next steps need to be.

1. I used the first sample to work out the bugs and to improve some of my methods.