Optimizing toroidal ion traps for miniature, field portable GC/MS systems

Transcription

1 Optimizing toroidal ion traps for miniature, field portable GC/MS systems S.A. Lammert, PerkinElmer, Inc. American Fork, UT R.H. Jackson, III Industrial Design Physics, LLC, Littleton, MA Presented at the 11th Harsh-Environment Mass Spectrometry Workshop Oxnard, California September 2017

2 BUT,more appropriately.. Down the Rabbit Hole Trapping Adventures in the Toroidal Coordinate System 2

3 The problem of Coordinates 3

4 The problem of Coordinates 4

5 How we got to where we are Conventional Ion Trap Toroidal Ion Trap 5

6 r-field E z-field E Asymmetric Toroidal Ion Trap Corrected electrode shapes restore mass resolution and performance Changed asymptote angles - Finnigan Stretched endcaps - Bruker/Franzen Sublinear ->Linear -> Superlinear field in ejection dimension only SYMMETRIC TOROID ASYMMETRIC TOROID 1/ 2 1/ 2 6 S. Lammert, et.al. Design, optimization and initial performance of a toroidal rf ion trap mass spectrometer International Journal of Mass Spectrometry 212 (2001) 6

7 GC/MS Performance: Figures-of-Merit * * GC Peak WIdth ~1 sec wide * * 90 sec Scan speed: ~10 scans/s ~1 s Perfluorotributylamine (FC-43) Mass Range: 43 - > 500 da. 7

8 Mass Resolution Performance d 8 - toluene [M-H] +, [M] + Tetrabromoethane [M-Br] Bromoform [M-Br] Dibromotetrafluorobenzene [M]

9 RF Trapping Devices and their Approximations 3D/Paul Trap Linear Ion Trap Toroidal Ion Trap 9

10 RF Trapping Geometries - Scalability 3D/Paul Trap Linear Ion Trap Toroidal Ion Trap r 0 R R/r 0 =3 10

11 But our Toroid Mass Analyzer is Not Fully Optimized (yet) Miniaturization induces compromises (for all devices) Space Charge/Ion loading (sensitivity/performance) Fabrication tolerances in hyperbolic devices Field compromises in approximation devices => Increased complexity in ion motion Toroidal corrections were only a first-order attempt. NO idea how to proceed further toroidal mathematics are much more difficult Problem is not in trapping, but in ejecting Any 3-D field with a saddle can trap ions (= Stability ) Ejecting ions with high mass resolution and high efficiency is harder - especially as the fields become less ideal and the ion motion becomes more complex. The Ion Dynamics problem. It was only later we ran into the cold hard truth that our device was not fully optimized. MS/MS resonance techniques were problematic. 11

12 PKI-Torion Production Trap - Ejection Simulation Simulation of the production trap 50 ions each m/z 50, 200, 350 Center birth Thermal ( room T ) Simple collision model (He) Low ejection efficiency (<25%) Remaining ions provide nothing except space charge. m/z 50,200, 350 Cold 12

13 SIMION Analysis of Ion Splat positions and time 13

14 Approach We needed a set of tools (analogous to those available to researchers using traditional 2D/3D quadrupole devices. Descriptive (a priori field equations, expansion equations) Predictive (allow accurate description of stability, ion motion, etc. Proscriptive (point the way to optimization) 14

15 Phase I: Wonderland Torodial coordinate system considerably more mathematically complex than Cartesian or cylindrical/polar systems Separable Laplace solutions for a toroid where τ=ln(d1/d2) Define a new term (ξ) ξ = e -τ = d2/d1 15

16 Toroidal Multipoles Toroidal functions analyzed using MatLab Solutions to Laplace equation for the first 4 toroidal harmonics ( poles ) are shown Up to T10 have been calculated cos sin 16

17 MatLab Results: TORION-9 Analyzer Fit Using only the first 10 terms, a least squares fit approximation of the production toroid trap was constructed 17

18 Toroidal Expansion (first 10 terms) for PKI Trap 18

19 Trapping Time (s) Some T2 Surprises in Wonderland z 2D r Toroidal Quadrupole QIT RF Amplitude (V 0-p ) 19

20 Curious 20

21 Curiouser and curiouser Studied trap stability diagram for cylindrical toroidal ion trap Large non-linear content Found prominent areas of instability ASMS 2017 and recent Int. J. Mass Spectrom. by Ailin Li (Daniel Austin Group Brigham Young Univ. 21

22 cyl-toroid Ion Trap Stability Diagram with Applied AC (Resonance ejection frequency = 380 khz) 22

23 Look closer Continued magnification of regions inside the chasm show fractal nature. (b) (c) Axial instability Radial instability (i) Radial instability(o) Stable Heavily dependent on (d) E, direction, ion position, RF phase (f) (g) (e) (h) (j) (i) 23

24 Sensitivity to KE, Velocity and Starting Position 0.2 V 0.2 V 0.2 V (a) (b) (c) Figure 4 Kinetic energy Velocity vector Starting position RF phase at ion birth (a) 0.05 ev (2,0,0) 0.1 mm offset in r 0 (b) 0.1 ev (0,1,0) 0.2 mm offset in r 0 (c) 0.1 ev (0,1,0) 0.3 mm offset in r 0 24

25 Phase II: Through the Looking Glass Understanding and narrowing down the entire field of toroidal functions to those that apply to our particular device and how to compute them Eliminating the complication that one of the toroidal variables goes to infinity at the trapping center. Parameterization of SIMION GEM files (Programmatic investigation of geometry variables) Tools to determine trapping centers, zero lateral field lines Tool to automatically make the trapping center equal to zero potential SIMION Extensions Ability to calculate the toroidal expansion coefficients for any arbitrary geometry file. Ability to calculate ion trajectory harmonics in real time. Ability to fly ions in electrodeless systems (mathematically derived) 25

26 Current PKI (Torion) Trap 26

27 Effect of slits Ability to characerize difference, coupled with ability to custom fabricate a candidate device may allow us to inoculate a slitless candidate for the eventual incorporation of slits. T harmonics for trap with slits Difference in T harmonics For Slit v. No Slit trap 27

28 Slit Effect on Lateral E-field in Production Toroid Field minima where lateral E-field (E x ) = 0 as f (y-distance) 28

29 Characterization of Current Production Trap Showing contours of constant E-field magnitude (blue) and contours of zero potential (red) 29

30 Zoomed examination of slit effects Both slit, no-slit potential arrays loaded simultaneously 30

31 Ion Frequency Components at Varied RF V peak. (Left) SIMION trajectory output for m/z 99 (time sample points shown in inset for V RF =110V. (Right). Harmonic frequency values at each V RF step.(colorized where the harmonic amplitude (A) is indicated by the color map. 31

32 V RF v. Harminv Analysis of Sampled Ion Trajectories RF only scan; Ions driven toward instability boundary. β=2/3 (667 khz) data set and initial parameters β=1 (f=1 MHz at Stability Boundary) Vrf-peak Ω=2 MHz Signed Frequency(MHz) "harminv_out2.csv" using 4:1:

33 Conclusions Opportunities still exist to improve the performance of the current toroidal ion trap Improved dynamic trapping range - Easier tuning, longer operational life Simpler ion motion - MS/MS, other resonance experiments May enable even smaller versions (or larger!) Simulations using only the T2 toroidal multipole shows both expected as well as very unusual behavior. Perhaps a T2 is not the best majority field for a toroidal device Strong indications that an odd order harmonic (e.g. T3) needed Exploring the effects of individual toroidal harmonics and reducing the complexity of the ion motion may lead to a path to optimization. We ve just begun to scratch the surface of this problem 33

34 Acknowledgements Bob Jackson (consultant, Instrumental Design Physiscs David Manura (Scientific Instrument Services SIMION) Expanded SIMION extensions to include Gnu Scientific Libraries (GSL) - Hypergeometric functions, FFT, Recompiled HarmINV. Prof. Daniel Austin (Brigham Young University, Dept of Chemistry) Jessica Higgs (graduate student) Dr. Ailin Li (recent graduate) Prof. Karl Warnick (Brigham Young University, Department of Electrical Engineering) Prof. Dennis Tolley (Brigham Young University, Department of Statistics) Ed Lee, Randy Waite (PerkinElmer) Joe Oliphant (Torion, retired) 34

35 Contact Info. Stephen A. Lammert, Ph.D. Senior Principal Research Scientist (office)