Monday, September 24, 2012

Enter George (we've spent many nights together)

by Stacy Carolin
Time to finally deal with the beastie: mass spectrometer, aka "George." Dr. Jess Adkins, my extra adopted advisor/professor/mentor/friend, is the head of the Caltech geochemistry lab into which I was invited, which includes two (!!) ThermoScientific NEPTUNE Multicollector ICP-MS (Inductively Coupled Plasma Mass Spectrometers). The older NEPTUNE that I work with is named George, which was given by postdoc extraordinaire Guillaume Paris referencing George Milton in Of Mice and Men.

Before giving George samples to eat, I think it's important to talk a little about how a mass spectrometer actually works, which will hopefully augment the limited information we all learned from CSI. Most simply, a mass spectrometer measures the mass-to-charge ratio of charged particles. The steps are:

1. Liquid sample is vaporized

2. The components of the sample (uranium and thorium) are ionized (this means given electric charge) by a plasma

3. The charged particles ("ions") are separated by their mass-to-charge ratio by electromagnetic fields (magnet)

4. The different mass ions are detected

5. The signal is processed

Ok. Step one. The liquid sample is first vaporized using a nebulizer, a little plastic guy that uses compressed gas to turn the liquid into an aerosol. The aerosol then flows through a curved heated spray chamber and "membrane desolvator module," which catches larger solvent droplets (oxides and hydrides, aka bad guys we don't want), and allows the smallest aerosol vapors to continue to the ICP-MS.

Next, the sample is ionized using an inductively coupled plasma (ICP) (ooo, cool! plasma!!) which is produced using an induction coil, argon gas flow, and an electric spark. The plasma is sustained in a torch and consists mostly of argon atoms with a small fraction of free electrons and argon ions (argon atoms that lost an electron, so positively charged). And its temperature is on the order of >5,000 K! As the nebulized sample enters through the ICP, it evaporates and any solids that were dissolved in the liquid aerosol vaporize and then break down into individual atoms ("atomization"). Then the plasma ionizes these atoms (steals off an electron to make the the atom a "positive" ion).

Ok, now we have a bunch of charged atoms with different masses ready to continue on through the mass spectrometer. I am running our uranium samples first, so we have U-238 (most common in sample), U-236 (from the spike), U-235 (not nearly as common as U-238, but small percentage occur naturally), and U-234 (radiogenic daughter of U-238, tiny amount). So the "mass-to-charge" ratio of these, assuming each has a +1 charge after passing through the plasma, equals 238, 236, 235, and 234 (note that they are all different values, but very close!). Following the plasma ionization, they are sent through a curved magnet, and based on electromagnetic laws including kinetic energy, centripetal force, and magnetic field strength, we can determine exactly where to place the collector cups for each ion mass using the equation:

This is what makes the multi-collector ICP-MS so amazing. It is able to measure several different masses all at the same time using multiple cups. This fact is imperative for our age calculations, which require ratios of several different atoms to be known.

The final step is detection. U-238, U-236, and U-235 are more abundant, and can be detected using Faraday cups, which are metal conductive cups that catch charged particles in a vacuum and then produce a current. The measured current can be used to determine the number of ions that hit the cup (in our case 1 volt = 62.5 million charged atom counts per second). U-234 is not great enough in our sample to be able to produce enough current, so it is instead measured on a "secondary electron multiplier" (SEM) which is basically an amplifier that produces an avalanche of electrons from a single ion, which can then be measured in counts per second.

Alright, now that we understand a little about how this all works, I can get started bringing the samples to the mass spectrometer. I add about 0.4-1.0 mL of 5% nitric acid to our tiny solid uranium samples, then transfer the liquid samples into "autosampler" vials to be loaded into the ICP-MS. 

Next I spend about half a day preparing the ICP-MS for a "run" by completing several tests that would take way too long to discuss. Just trust that I did them all right ;) I then prepare a sequence in the editor program in the NEPTUNE software, and once everything is in place, hit GO! Yay! Now we just let George do his thing, checking on him frequently, and hopefully no errors come up. Each sample takes about 1 hour for measurement, and then the same will have to be done to set up a run of thorium tomorrow. But once the run is complete we will finally have DATA!!, which can actually be analyzed for stalagmite AGES! Finally! Get pumped :) 

Friday, September 21, 2012

A brush with machine learning

by Kim Cobb
For the last two days I have been at the 2012 Climate Informatics workshop at the National Center for Atmospheric Research (affectionately referred to as NCAR). It was a wonderful collection of fellow paleoclimate geeks, including Cobb lab alum Julien Emile-Geay, as well as a mixture of 20+ computer and statistics nerds.

The goal of getting this diverse group of experts together was to probe the intersection of these fields, where novel or not-yet-developed techniques may lead to breakthroughs in climate science. A major motivation for this kind of venture are new federal initiatives aimed at tackling science and technological problems that leverage "Big Data"(see NSF example here). Such initiatives recognize that the explosion of increasingly large and complex datasets presents unique challenges and opportunities. They call for fundamental advances in computational and informational sciences – advances that will enable the use of Big Data to solve some seriously stubborn problems, like those involved in climate change.
View of NCAR, nestled into the foothills of the Rockies.

Al Kellie, the Director of the Computational and Information Systems Laboratory at NCAR, put the Big Data problem this way:  if you were to store the latest collection of climate model simulations run in support of the new IPCC assessment on 32GB iPads, you’d need a 6-mile-high wall of them stretching from Atlanta to Alaska! Our existing bag of data analysis tricks simply won’t work anymore.

So what’s a climate scientist to do? Adapt some whiz-bang tools from the computational and statistical folks to our questions. This basically involves writing algorithms to “teach” a computer how to extract the information you want from a huge archive of data. “Machine learning” is the term, and in this game, speed and accuracy are both of paramount importance. Whole papers have been written concerning a faster or more accurate algorithm for the detection of hand-written numbers – something my 3-yr-old son Isaac can do in less than 1 second (if he so desires). Here I mean no offense to Tony Jebara here, who gave a shockingly lucid presentation on various techniques in machine learning (and bought me lunch!). His talk, along with the others, can be viewed here.

One example from climate/comp sci hybrid Amy McGovern concerns the detection of tornadoes in an insanely high-resolution model (500m x 500m) of the atmosphere over Oklahoma. By amassing some rules and relationships about hurricanes from her dozens of simulations of historical data (her “training set”), she can build a hierarchical decision “tree” (yes, this is a technical term) that the computer will move through in order to assess the risk of a future tornado given real-time atmospheric data inputs. Her computer has “learned” how to predict tornados with Big Data. Thankfully for Amy, Oklahoma has a wonderfully rich network of meteorological observing stations that inform her super-high-resolution tornado model. Lucky her. She probably has more data in that Oklahoma network that the entire NOAA NCDC paleoclimate repository that covers the whole globe through all time!

In my talk, I tried to convince the computational and statistics crowd that helping us out with our paleoclimate problems i) would provide much-needed constraints on key climate change uncertainties, and ii) would benefit enormously from the types of Big Data techniques that they are developing. You might not think of paleoclimate as a Big Data problem, and you’d be right. We still have relatively few reconstructions of past climate. But, the latest IPCC model data archive contains upwards of 40 simulations of past climate variability, for the first time ever! The challenge for us is to develop efficient, objective, and well-suited analysis tools to perform paleodata-model data intercomparisons on the road to better estimates of climate change uncertainties.  

Perhaps the most thought-provoking talk was by Steve Easterbrook, who looked at the evolution of several IPCC climate models over the last decades as growing lines of computer code, as well as the apportionment of code to each component of the model (i.e. atmosphere, ocean, land, etc). He stressed that every model cannot be everything to everybody, advocating a model he called “Fitness for Purpose”. Some models were developed for weather prediction, while others were developed for long paleoclimate simulations. Some are easier to update and improve, while some are modified only rarely. Model center people management structures provide another unique model descriptor, which likely projects onto the model characteristics themselves. A neat meta-analyses of models, a rare glimpse at the models themselves, rather than their collective output plotted to 2100.

The NCAR supercomputer, "Bluefire". Wow.
I left feeling hopeful and inspired that there may be a path “from data to wisdom” (Ackoff et al., 1989, as quoted by Steve in his talk), as we risk drowning in bytes. We will clearly need students who are trained in climate science as well as computer science - a tall order. My graduate students struggle to learn Matlab their first year, but I hope they recognize that it's for their own good!

Science + Dance = The Perfect Mix

by Stacy Carolin
As you can see, this whole process takes a lot, a LOT of time and effort, and it's easy to get overwhelmed and claustrophobic (well, at least to me). As one who blatantly chooses to not comply with any expendable schedules and rules, I've doctored this up by choosing to morph my own schedule so that each workday still comprises everything that I love to do. As I've said before, science is not 100% of my life (an essential point, as I believe diversity in many fields is what ultimately leads to inspired ideas and creations), so I don't let it consume 100% of my day.

I thought this time I'd share how I (somewhat oddly) have chosen to set up my life in LA. So here we go, last day of chemistry before we transition into the mass spectrometer instrument room. (Note: this is usually completed over two days, but we are in crunch mode because I fly back to Atlanta in less than a week... so don't get scared about the intensity, it's not too normal)

7:30am: Ok, so what we left in the lab yesterday was our "dried-down" solid samples. Today we are going to use iron as our helper element to grab up all the uranium and thorium atoms for us so that we can dump everything remaining away. First I add 1.5 mL of hydrochloric acid to the Teflon beakers to turn the sample back into liquid acidic form, then transfer it into a centrifuge tube (a centrifuge is that machine that spins around really fast and makes any solids in the tubes collect together at the bottom). Next I add a drop of iron chloride, which makes the acidic solution turn yellow, then add drops of ammonia solution slowly. Ammonia solution is basic (opposite of acidic) so when it is added to hydrochloric acid it reacts to neutralize the acid, then once the solution has turned from acidic to slightly basic, it reacts with iron chloride to produce iron hydroxide (a reddish solid).

Iron hydroxide is the compound that happens to grab onto uranium and thorium!

So now, to get rid of everything else, I stick it into the centrifuge so that all the iron (and its grab-on buddies uranium and thorium) gets solidified at the bottom. Then I dump out all the remaining liquid ("supernate") from the tube. Next, to transfer our solid iron sample out of the centrifuge tube, I dissolve it in a little nitric acid and pour it back into its Teflon beaker.

10:20am: Time for dance! Jump on my bike and ride 15 miles from Pasadena to EDGE Performing Arts Center in Hollywood (fun!!)

11:30am: Class with Malaya at EDGE, absolutely amazing, I am so fortunate!!!

1:00pm: Ride back to Caltech from Hollywood (Fun note: After a recent serious bike accident related to too much exercise strain on my body, I now only ride back if I feel up to it and the weather isn't over 100 degrees; there is also a subway system that takes the same amount of time that I can use.)

2:30pm: Suit up and back into the lab! I put all the samples back on the hot plate to dry down into solid again, then add 1 mL of nitric acid to re-dissolve. While the samples are drying, I start preparing my "columns" to perform the final "column chromatography." Hmm, let's see if I'm able to explain how this works... Right now we have nice individual liquid samples each filled with iron, uranium, and thorium molecules all in little bonds together. Column chromatography is the method I use to get rid of that iron, then separate the uranium and thorium into two separate Teflon beakers so that they can be measured on the mass spectrometer separately. A "column" looks like a 5-inch tiny plastic straw with a wider diameter 1/2-inch long top section, which is what we load our liquid samples into. Inside the length of the thin part of the column is column resin, little solid beads that look like sand, which the sample's iron, uranium, and thorium grab onto.

After our sample is poured into the column and being held onto by the resin, I add different "eluents" (liquid solutions) one at a time to detach each molecule type separately. This is pretty cool!-- as a particular eluent travels down the column straw, it can either pass by the molecules holding onto the solid resin, or it can replace the molecules, aka kicking them off and grabbing onto the solid resin in their place. So, first, I add nitric acid which replaces all the iron bonds and makes the iron molecules drip out of the column straw and into a waste beaker underneath. Next, I add hydrochloric acid, which detaches the thorium, which then falls into a clean Teflon beaker below. Finally, I add water, which releases the uranium atoms into a second Teflon beaker below. And we are now done with column chromatography (about 8 hours later, haha, dripping is slow!) Congrats! ;)

To finish, I add a drop of perchloric acid one last time to each of the uranium and thorium Teflon beakers and dry down on the hot plate again (which should remove any remaining organics), then repeat with a drop of nitric acid. We now have a batch of Teflon beakers filled with a tiny thorium solid and a batch of Teflon beakers filled with a tiny uranium solid. YES! We're ready! Time to move into the instrument room with the incredibly complex and expensive inductively-coupled plasma mass spectrometer. Boom!

12:00am: And time to ride home :)
Fun night-bike-ride through Pasadena.
Great day!

Thursday, September 20, 2012

Gloves and Goggles, etc.

by Stacy Carolin
 Lab suit on. Lab hat on. Lab shoes on. Goggles on. Gloves on. Music on. :)

Alright, we made it into the super clean room! Finally! But wait, why do we need to be all dressed up in here? Because we are attempting to measure TRACE elements in those stalagmite powder samples we created from drilling. TRACE means that there are very very VERY few of these guys in the sample (for example, U-238, the by-far most abundant element out of the three we're looking for, usually has a concentration in our stalagmites of only about 0.1 to 1 parts per million, aka 0.0001%!) So, if while I am processing these samples to prepare them for the mass spectrometer they are contaminated by an unknown amount of air particles in the room (i.e. tiny pieces of dust which contain some thorium happen to fall into them) the entire batch will have to be thrown out. Yikes! So to prevent this I am working in a clean room, where air is constantly being recirculated and filtered, there are absolutely NO metals so no opportunity for corrosion, floors and workstations are constantly cleaned, and any scientists inside the room must be completely covered with special garments so they don't shed fibers or dead skin cells.

Inevitably, a tiny amount of uranium and thorium atoms that are floating around as part of "escapee" dust or somehow present in not-perfectly-cleaned reusable lab vials are going to slightly contaminate our samples. To correct for this, I am going to add six "blanks" to the batch. I do the EXACT same chemistry to these "blanks" as I do to our individual samples, except that they have no stalagmite powder in them to begin with. That way, we know the average number of uranium and thorium atoms found in these "blanks" must have come from any overall room contamination, so we can subtract that value from our sample values before calculating the age. The larger the variability of contamination in our "blanks," the greater the age error will be in our calculations, so it is extremely important that I work super carefully in a very clean environment while I process all of these samples (real and blank).

Ok yay! Time to start prepping! FIRST, I need to determine the exact weight of powder I have for each. To do this I weigh each one 3 times on a digital lab balance. SECOND, I pour the sample powder into a labeled 22 mL teflon beaker, then add about 15 drops of concentrated nitric acid until all of the calcite powder is dissolved. THIRD, I do what's called "spiking" our samples, which is adding a known amount of U-236 and Th-229 "spike" into each of our now liquid samples. U-236 and Th-229 are atoms that do not naturally occur, so any that the mass spectrometer measures we will know came from our "spike" and not from the stalagmite. And since I know exactly how many I put in (based on the weight and known concentration of the "spike"), the "spike" is what will allow us to convert whatever the mass spectrometer spits out for those atoms we're interested in (U-238, U-234, and Th-230) into actual values.

FOURTH (and last for today) I add 20 drops of perchloric acid to each of our samples, then put them on a hot plate to "dry down" (turn back to solid). Perchloric acid is a very potent oxidizing agent of organics, and is therefore added and then heated in order to remove organic material such as bat guano from our samples.

Hooray! First day of chemistry done! Just to recap, today we (1) "spiked" our samples, which is a step needed in order to complete our age calculations later on, and (2) dissolved and heated our samples in concentrated oxidizing acids in order to remove any organic material waste. Our goal is to eventually reduce these samples down to JUST uranium and thorium atoms floating around in acid, so we will continue the chemical "purifying" process tomorrow. Phew! :)

Wednesday, September 19, 2012

A Lesson on Dating...

by Stacy Carolin
Hmm, it's time to start explaining why we care so much about these crazy sounding elements, uranium (U) and thorium (Th). Remember the periodic table in high school? Uranium and thorium are both large metallic elements near the bottom of the chart. For example, while oxygen weighs only about 16 units, uranium weight averages about 238 units and thorium 232 units! Most importantly- both these elements are also radioactive, which means that they decay at a certain KNOWN RATE over TIME.

Lucky for us, these radioactive elements are found in tiny amounts in rocks and soil (and therefore stalagmites! yes!) And to make things a little more complicated, they exist in a few different sizes. We will be dealing with U-238, U-234, Th-232, and Th-230 (those numbers after the element symbol are their masses, aka total number of protons and neutrons in the atom's nucleus).

Ok, so just a quick lesson on how radioactive elements can be used to date a material: Each radioactive element decays over TIME at a given RATE (can be labeled as it's particular "half-life") that has already been scientifically determined. Let's say you have a pile of 1000 radioactive parent-blobs, whose "half-life" is known to be 7 days. That means that one week from now, about 500 parent-blobs will have decayed into daughter-blobs, while the other 500 parent-blobs are left in your pile. Two weeks from now, only 250 parent-blobs will be left. Three weeks from now, 125 left. If you were a scientist that walked into the room and saw 125 parent-blobs and 875 daughter-blobs, you'd know that to begin with there must have been 125+875=1000 parent-blobs, and you'd be able to calculate that 3 weeks had passed.

"U-series" dating is a bit more complicated because we are dealing with a chain of events (U-238 decays into U-234, which then decays into Th-230) but the idea is the same. What we need to know is how many U-238, U-234, and Th-230 atoms are present in our rock samples, and then using math equations we can calculate the exact age.

A mass spectrometer is perfect for this task- it can count elements based specifically on their different masses. So, for the next three days we will be doing a LOT of chemistry on our samples in order to convert the stalagmite powder into a liquid sample that can be loaded into the mass spectrometer (the mass spectrometer is VERY picky about what it likes to eat).

Ok, so now we can start! First step, get dressed to enter the wet chemistry super clean room. Ah, beautiful! :)

Monday, September 17, 2012

Slice It!

by Stacy Carolin
Here we go! DAY 1! (well, actually this was done a few months ago in Atlanta in prep for this Caltech trip, but let's use our imaginations.... :)

Today we are going to remove from our candidate stalagmites the small rock samples to be analyzed. Remember, we need one from the bottom of the stalagmite, which will give us the oldest date (aka when it started growing), and one from the top, which will give us the youngest date (aka when it stopped growing).

The best way to keep age ERRORS in our final calculations AS SMALL AS POSSIBLE is to (1) not let any of the stalagmite's dirt or mud contaminate our clean pure-rock samples, and (2) to only extract our sample from a SINGLE stalagmite growth layer (so we're not averaging many different ages together). In order to do this, we need to cut open the stalagmite to be able to see the many layers and access the cleanest area of growth: it's center.

I am using a big tile saw with a thin diamond-edge blade for slicing (oh boy!) First, I slice any stalagmites that are too long into smaller sections. Next, I slice each section down the center (longways) to see it's interior, and then finally, I slice the back half of the stalagmite off in order to turn the end product into a "slab." Thin section slabs are very helpful because when you hold them against a backlight the stalagmite's layers, and especially any dirt squished in between, become extremely visible.

Ok, now we have something very beautiful to work with! We are going to need about 250 mg of rock for a sample, which if drilled into powder is about 1/4th a teaspoon in size. In order to drill precisely out of only one of the stalagmite's top layers (and one of the bottom layers), I'm going to attach the stalagmite slab to a micromill that reads out the x-y-z placement of the drill bit. I can then first map out exactly where I'm going to drill into and write down all these coordinates, THEN go back and drill into every point that I picked out. And as I've done this several times before, I know approximately how many points to drill AND how deep to drill in order to have about 250 mg of powder when I'm finished. Just a side note, the drill bit I use is 1.6 mm in diameter, so the layer I'm drilling out, if done perfectly, will be 1.6 mm thick (from top to bottom- remember the stalagmite slab is laying on its side on the drill stage). Stalagmite growth rates vary over time, but just for example, if it grew very slowly at 300 years/mm, that layer that I just removed covers 480 years, which we should keep in mind.

For this exciting (and hopeful!) venture we're going to test 4 different stalagmites, so I repeat this 8 times (top and bottom for each) to get 8 total samples (which took about 8 x 2 hours each = 16 hours, sigh! welcome to science! ;).

Remember, our goal is to determine the AGE of each of these now 250 mg powder samples, and to do that we need to somehow determine the URANIUM and THORIUM atomic ratios that, right now, are invisibly hidden in each. It's time to switch from geology to chemistry... on to Caltech! Yes!