This article has been updated to clarify statements Schubert made regarding the company's manufacturing relationship with Agilent.
[IMGCAP(1)] At A Glance
Name: Michael Schubert
Position: Executive vice president, Bruker Daltonics, since 2005.
Background: Vice president of R&D, Bruker Daltonics, 2004-2005; Head of R&D for Bruker Daltonik GmbH, Germany, 2000 — 2004; Scientist and project manager for ion trap product lines, Bruker Daltonics, 1991-2000.
PhD in physics, Hamburg University, Germany, 1992.
Michael Schubert has been involved in the development of ion trap mass spectrometers at Bruker for more than 15 years. ProteoMonitor spoke with Schubert to find out about the history of nonlinear ion traps and the advantages and disadvantages of nonlinear traps compared with linear traps.
When did you get involved in developing ion trap mass spectrometers?
I graduated from Hamburg University in 1992 with a PhD in physics. I was working on ion traps at that time already. That was a reason to go to Bruker, and I joined Bruker Daltonik GmbH in Germany in 1991. At Bruker, I was a project manager for the ion traps for a long time — approximately until 2000. And then I went into R&D management at various levels. Finally, I became the vice president of R&D in 2004, and then on Sept. 1, 2005, I moved over to the US as executive vice president.
When you first started working with ion traps, were you using them for analysis, or were you developing the instrument?
At university, I was using them as well as developing them for quantum optics applications. So that was not mass spectrometry — that was laser spectroscopy of quantum optics of a single ion or very few stored ions in a [Wolfgang] Paul trap. Its key features were the ability to store and localize individual quantum particles and have them interact with laser beams. We also built and operated large traps for large clouds of ions, and tried quadrupole traps (Paul traps) as well as octopole traps. We also studied their motion by probing them with laser light, and we studied the photon statistics of the fluorescence from these ions.
At Bruker, I worked on using ion traps for mass spectrometry; not for quantum optics anymore. That was a very exciting time, because a little bit before I joined, Dr. Jochen Franzen, together with a few others, made the key invention of the nonlinear ion trap at Bruker. Another key, additional invention was the ion ejection via phase-coupling the main RF and the ejection frequency. These are very fundamental, basic inventions around ion trap mass spectrometers, and they are the basis for today's Bruker Daltonics ion trap product line. Nonlinear resonance ejection with phase coupling is the key method, and still today the best-known method to operate an ion trap mass spectrometer. It is the reason why the ion traps that Bruker has on the market today are superior to the ones of Thermo Finnigan in terms of scan speed … and in terms of mass range and several other aspects.
Bruker has a very large patent portfolio on ion traps.
Could you go over the difference between linear and nonlinear ion traps?
The nonlinear ion trap inventions happened around the end of the 80s and early 90s. So the nonlinear ion trap was around way before the linear ion trap. The linear ion trap was invented maybe four to six years ago.
After I joined Bruker, the instruments that we brought out were electrospray ion trap instruments — so we worked on exploiting these inventions and making these ion traps work very well, especially in terms of scan speed. Bruker ion traps even today have the highest scan speed at a given mass resolution. I have to say 'highest speed' at a given resolution, because the scan speed can be set as high as you want it, but then the resolution goes down. So at a given resolution, we have the highest scan speed.
Then we launched our Esquire series of ion traps that we co-developed together with Agilent. Then, one of the next key innovations in ion traps was the high-capacity trap, which we brought out. Now there are several high-capacity traps. A linear ion trap is also a high-capacity trap — it can trap a lot of ions. But a Paul trap, or what some people today call a 3D ion trap, can hold very many ions too. It can hold about the same number of ions as a linear ion trap, which is very surprising at first glance, because it seems counterintuitive. However, the explanation is in physics of the nonlinear ion ejection. Once that fact is optimized and exploited at the next level, the capacity can be increased a lot.
The HCT was introduced about a year before Thermo introduced the LTQ.
So there is talk in the market about 2D ion traps and 3D ion traps. These are terms that have been coined, but they don't make much sense from a fundamental point of view. I mean, any ion trap is meant to be a device to keep an ion in all three dimensions. If it keeps ions in only two dimensions, it's not a trap, because the ion goes away in the third dimension.
The term 2D ion trap is used for an ion trap that is elongated in one special direction. So it's an ion trap where the trapping volume is like a long cigar, rather than a sphere.
What are the advantages and disadvantages of having a spherical versus an elongated ion trap?
Fundamentally, not much. The advantage that is claimed for the linear ion traps is that they have higher capacity than previous generation Paul ion traps like the LTQ, and that is true. However, the capacity problem can be solved in various different ways. It is just one way to go to a linear ion trap design.
Another way is to go to a high-capacity trap design, which increases the effective trapping volume in all three dimensions. Both result in the same capacity, and both solve the same problem. So the linear trap solves a problem, but it's not the only solution to that problem. The high-capacity trap is an equally good solution to this problem as well.
On another note, linear ion traps have a disadvantage versus high-capacity traps in that they have difficulty trapping ions of positive and negative charge at the same time. Paul ion traps, as well as the high-capacity traps, can do that by their very design, because they use only RF voltages to effect the trapping, and the RF does not distinguish between positive and negative ions. As a consequence, because they can easily trap positive and negative charges at the same time, high-capacity traps are very well suited to do electron transfer dissociation. This is pretty important because ETD is a process where positively and negatively charged ions are brought together to effect the dissociation reaction.
In linear ion traps, the trapping normally is effected by having RF on the rods, and DC on the end plates to effect the trapping. As soon as there is a DC component to the trapping voltages, the trapping becomes ion polarity dependent, and the device can trap ions of only one polarity. There are ways around that, but all linear ion traps that came on the market use this DC trapping method. This makes it pretty difficult to implement ETD on these devices.
Thermo Finnigan had to modify [its] ion trap at its very heart, in the trapping method and voltages, to make ETD work in it. In short, they need to switch back and forth between RF and DC on the end plates during the scan cycle.
As you ask for advantages and disadvantages, I note that another key difference is the scan speed that is achieved. In high-capacity traps, you get a scan speed of 26,000 Daltons per second at pretty good unit resolution, often allowing the identification of doubly charged ions as doubly charged; while with linear ion traps the scan speed that just gives unit resolution is between 4000 Daltons per second and 16,700 Daltons per second, depending on design. Hence, speed is a key advantage of high-capacity traps using non-linear ion ejection with phase coupling, and it is a disadvantage of linear ion traps.
Also, in high-capacity ion traps we have a direct mass range of 3,000 Daltons, which can be extended to 6,000 Daltons, while the linear ion traps have a mass range of 2,000 Daltons, which on some models can be extended to 4,000 Daltons. And that clearly is a limitation, especially, again, for ETD. ETD can very well do sequencing of large peptides or small proteins, and in that, mass range obviously is a big issue.
What is it that makes a high-capacity trap high capacity?
That is basically making sure that the nonlinear resonances are exploited in a very different way. We had in all the years since the early 90s exploited nonlinear resonances to get very high scan speeds, and it took us some time to realize that by combining the nonlinear resonances in a different way, we were able to not only increase the speed even further, which we did from the Esquire to the HCT series — we increased the speed from 13 kiloDaltons to 26 kiloDaltons per second — but we were also able to drive up the capacity by large factors — 10, 20, 30. That was a key step, and it was very surprising when we found that out.
When I say that we combined the nonlinear resonances in a different way, that means we changed the shapes of the electrodes in a very special mode for the high-capacity ion traps, and that gives us the factor of 10, 20, 30 in additional capacity versus the Esquire series.
Why is that? Basically, what we do with the nonlinear resonances is we shield the ions that get ejected from the space charge of the other ions. The space charge is the limitation normally for the capacity of the ion traps. We can use the nonlinear resonance to shield the ions that get ejected at a particular moment in time from the space charge of the other ions which are still in the trap and will be ejected at a later point in time. Because of that, we get that large increase in capacity.
Other people have said that high-capacity traps are still not as high capacity as linear ion traps. Would you say that's not true?
That is not true. That is an incorrect statement. I mean, I understand why the statement is being made — it's for marketing purposes — but the statement is not correct.
Can you compare the sensitivities of a high-capacity ion trap versus a linear ion trap?
If you look at the sensitivity specs for an LTQ as well as an HCT Ultra, then they have in the MS/MS a sensitivity spec of 250 femtograms, 50:1. Both instruments have that spec, so there is no difference in sensitivity.
There's another point I'd like to bring up. There was an Association of Biomolecular Resource Facilities study on protein identification in a complicated mixture, and the data from the HCT Ultra came out extremely favorably. There were linear traps, quite a few of them, participating in the study, but they came out much lower in score. I think that also tells something.
Are there any fundamental caps or limits to what either linear or nonlinear traps can do?
These are two different ways to approach a topic. In terms of capacity, I don't see a difference. In terms of mass range and scan speed, we see the difference very clearly.
My personal opinion is that a linear or nonlinear setup is not really the question. The key question is 'What is the analytical performance of the instrument?'
If you ask, for example, for a linear trap — there are several linear traps on the market from various vendors. The instruments have very different performance characteristics. So whatever performance characteristics you look at [are] not determined by whether it's a linear trap or not. If so, the instruments should be very similar in performance, which, in fact, they are not.
So I think it's the wrong label to look for. The right label to look for is overall performance, or scan speed, or mass range, or sensitivity, or what happens in an ABRF study, or how complete is the bioinformatics suite. Linear versus not linear is not the key question.
How is the competition within the nonlinear ion trap market?
There are two vendors for nonlinear ion traps — that's Agilent and us. And [Bruker does the final assembly and the final tests for] the Agilent traps. They are not identical in performance. We have instituted some performance differential, of course, in order to have a different product positioning. But the nonlinear ion trap technology is the key Bruker technology. And since Bruker assembles Agilent's traps, Agilent is not a true competitor in the market.
What kind of directions are you looking at to improve your instruments?
It is clear we are looking at many ways to improve the instruments. We have an ongoing policy to improve all the product lines that we have — the orthogonal time of flight product line, the TOF and TOF/TOF product lines, the FTMS product line, as well as the Esquire and HCT product lines. We have done that over the years.
I would rather not speculate about what's around the corner, because I prefer to talk about facts, rather than potentialities.