Obviously you know I really mean “what is abundance sensitivity”, as it is not a physical thing but a relative term. As a reader of this blog, it is more than likely that you already have knowledge of analytical instrumentation, and you may well already clearly understand the term abundance sensitivity. There are many wonderful resources available on the internet that describe it in detail. My blog is not one of those, this is really a discussion piece about how I’ve learned to understand its relative importance, and what impact it has on the isotope ratio world that I inhabit.
Abundance sensitivity is a measure of the contribution of the peak ‘tail’ of a major isotope on an adjacent minor isotope. The tailing is predominantly due to the collision of ions with residual gas molecules in the mass spectrometer causing a loss of ion energy. This is a direct quote from an application note published on our own website, so surely it must be true, right? It is, but this is a rather specific definition applied to isotope ratio analysis. In more general terms it is the (unwanted) tail contribution at one mass peak from another, much larger adjacent mass peak. In other words, it is not specific to isotopes but can be an important factor in all forms of mass spectrometry, including the measurement of complex compounds such as proteins.
To business then… why is abundance sensitivity important and how does it affect isotope ratio measurement? In most forms of mass spectrometry, the analyst is trying to determine the amount of a compound / element / isotope. Sometimes you want to measure that amount precisely (how much mercury is in my drinking water?), yet in some cases just an approximate measurement is fine (measuring the metabolites in my urine a few hours after I ingested a pharmaceutical product). Isotope ratio measurement often falls into the former group, and if you’ve invested heavily in an isotope ratio mass spectrometer then your application is definitely in the former group.
This means that any form of error, even at the fraction of a percent level, needs to be taken into consideration, and eliminated if possible. Abundance sensitivity can be a significant source of error when an analyte isotope peak is adjacent to either a much bigger isotope from the same element, or a large mass peak from something else (something in the matrix that your analyte resides in, or possibly something from the mass spectrometer itself).
What influences abundance sensitivity? The best way to think about this is to consider the ion beam in a mass spectrometer as a jumbled mess of ions, atoms and molecules rather than a rigid laser beam of ions with no deviation. The charged species (ions) don’t all leave the ion source with exactly the same energy – instead there is a small spread in energy across the ion population. The ions are then accelerated typically by using a large voltage to pass them into the mass analyser and on to the detector., On the way, the ions bump into each other and to the remaining neutral species that annoyingly manage to insinuate their way into the ion beam. These collisions cause the energy of the ion population to spread further. With magnetic sector mass spectrometers such as those made by Isotopx, this energy spread for the beam of a given ion (e.g. 238U+) causes those with higher energy to bend less in the magnetic field and those with lower energy to bend more.
Obviously, there is fancy Gaussian mathematics to predict the outcome of all of this, but of interest to me is the observation that when you use the mass spectrometer you don’t get a single sharp peak (like a spectral line in an optical instrument) but instead a distribution, with a central region of peak maximum and a tail on both the lower and higher mass sides of the peak. Simply put, if the peak we are talking about is big, then the tail can make a major contribution to adjacent (but smaller) peaks, in extreme cases swamping the small adjacent peak(s) entirely.
How do we improve it? As you might expect, the best way to improve the situation is to stop those ions colliding with all and sundry, meaning less smearing and less tailing. There are several ways to do this. The easiest way to improve things is simply to remove as much residual background “stuff” from the mass spec vacuum envelope as possible, which you achieve by improving the vacuum pressure. It’s very simple physics, if you remove all the residual gas that is aimlessly floating around your ion beam flight path then you get less collisions, a sharper mass peak and less tailing. I said easy but I didn’t say cheap – adding additional vacuum pumping to a mass spectrometer can be an expensive business and the gains get smaller per dollar you throw at it as the vacuum pressure gets lower.
At this point I was searching for a good analogy and my colleague Dave kindly provided one: Leaves. Yes really! Consider a tree, gently swaying in the breeze, it’s fall/autumn, so the leaves are falling from the tree. The falling leaves are the ion beam – a large tree (equivalent to a large ion beam) drops lots of leaves which scatter on their way down. If it is a still day (the analogue of high vacuum) then most of those leaves fall at the base of the tree, and certainly don’t collect at the base of nearby trees. But on a breezy day (the analogue of low vacuum) the scattering increases and some of the leaves end up at the base of a much smaller tree nearby. The leaves from the big tree that have blown to the base of the nearby tree are the “tail” of the mass peak. But anyway, enough of analogues, back to science…
Fortunately, you can also throw some electromagnetism at the problem. If you send your ion beam around a nicely curved beam path, with appropriate voltages applied, then as you deflect your ion beam, you magically throw away a lot of the ions with lower and higher energies, meaning mostly just the ions with exactly the right energy get through. It’s a bit like sending white light through a prism (think Pink Floyd’s “Dark Side of the Moon” album cover) and using a slit to throw away all the other colours except the one you want.
At a stroke this significantly reduces the tailing, bingo! Typically, this system is known as an electrostatic filter (ESF) and is mounted behind the array of individual detectors that you need for an isotope ratio mass spec that measures multiple isotopes simultaneously. Other analytical systems can achieve the same thing using quadrupole / multipole filters; the principle is always the same, find a physical way of throwing away the pesky low and high energy ions, just keeping the ones you want.
This was indeed my first encounter with the importance of abundance sensitivity. I was working with multicollector ICP-MS and I couldn’t understand why a customer might pay many thousands of dollars extra to have this little curved filter at the back of their mass spec. Then I was asked to measure the amount of 236U (an isotope of anthropogenic origin, typically at vanishingly low levels) and was astonished to see that the m/z 236 peak was totally obscured by the low mass tail from the 238U. And that’s two mass units away! Switching over to using the rear electrostatic filter to reduce the tailing clearly revealed the 236U peak. I later learned that these filters typically reduce the abundance sensitivity (degree of tailing) by about an order of magnitude. I had seen the light.
Abundance sensitivity is also an issue at the other end of the mass spectrum in ICP-MS. You, an avid reader of this blog, will know that ICP-MS instruments rely on some fairly heavy duty physics, using a super-hot plasma to rip apart most molecular species and ionize them. In theory this is great for a mass spectrometer, but unfortunately the plasma that achieves this task needs a source of ions, usually an inert gas such as argon. This means that the ion beam you created is also carrying a whole heap of argon and molecular species containing argon, such as oxides. (Side note: as physical chemists we think of argon as almost completely unreactive. Well, stick it in a 10,000-degree plasma and you’ll change your mind!).
The main isotope of argon is at m/z 40, and in an ICP-MS the peak there is biiiiiiiiiig. The problems that this creates are twofold. Firstly, and arguably most importantly, you have very little chance of measuring anything at m/z 40 (like calcium for example), because the 40Ca peak would be totally swamped by the 40Ar beam. A similar thing happens at the ArO peak (m/z 56). But additionally, any smaller peaks close by the argon and argon oxide peaks suffer from the tailing issues we’ve been discussing. No reduction in background vacuum pressure or addition of an ESF is going to be enough, so during the 1990’s a new technology emerged, the use of “reaction cells” or “collision cells” to use chemical and physical techniques to either break up molecular species, or to react them with other species to shove them to a different part of the mass spectrum than your target analyte.
Although a mature technology in quadrupole ICP-MS, these technologies did not make their way to MC-ICP-MS instruments until a bit later on. I spent a bit of time working with my trusty colleague Sergey on one of these in about 2001, and we were most intrigued to learn that not only could the collision cell be used to make/break compounds to shift those annoying species to another place, but also the abundance sensitivity improved. The phraseology of the time was that “collisional focusing” was taking place. In other words, the bumps and reactions inside the cell had the effect of reducing the spread in the energy of the ions. It was most unexpected at the time but with hindsight it is logical.
Decades later, multiple technologies involving forced molecular interactions are again back in fashion, and are making great headway with some of the more challenging applications in MC-ICP-MS. But of course, that’s a blog for another day.
That’s all for now. Hopefully this blog straddled the difficult tightrope walk between mildly informative and irritatingly patronizing. As always, send your e-mails to: (Stephen.guilfoyle@isotopx.com). More soon…
