Note 59: Computer Modeling of a TOF Reflectron With Gridless Reflector Using SIMION 3D

By Steven M. Colby

Presented at ASMS, Palm Springs, CA., June 1997

INTRODUCTION

The ion simulation software SIMION 3D is used to model the scattering of ions by grids within a reflectron TOFMS. Our purpose is to examine the effects of ion scattering on the resolution and sensitivity of the instrument. This simulation demonstrates the use of SIMION software for the analysis of a common but difficult ion optics problem. This poster specifically addresses the effect of source grids in a reflectron instrument with a gridless reflector.

Background Software

SIMION 3D v6.0 (Scientific Instrument Services, Inc.) is the industry standard for the simulation of ion optics. The original software was developed in 1977 by D. C. McGilvery at Latrobe University, Australia. During the last 20 years, it has been greatly expanded by David Dahl at the Idaho National Engineering Laboratory and now shares little with the first versions [1]. The program supports full three dimensional modeling and potential arrays of up to 10,000,000 points. Specific features include, dynamic parameter variation, time dependent potentials, and space charge effects. The simulation involves the use of "Instances". These are three dimensional electrostatic and magnetic arrays used to model sections of an instrument. Each instance is independently defined and modeled. This allows the user to take advantage of symmetry elements that may vary between different sections of the instrument. It also permits the use of higher resolution arrays in more critical areas. Instances are positioned on an "Ion Optics Bench" for the simulation of an entire system. For example, in a reflectron TOFMS, the source, reflector, and detector can each be modeled separately and then positioned on an optics bench. The position and angle of each instance can be selected so as to easily examine a variety of instrument geometries. SIMION has a powerful user programming interface. This permits the simulation of time dependent fields (i.e. ion traps). It also enables the modeling of random effects such as collisions, ionization, and velocity distributions. We have previously shown how this feature can be used to examine the passing of an ion through grids (or mesh). [2] This requires a Monte Carlo simulation since the ion may pass through any part of a grid opening.

The Problem

 Fine mesh or grids are commonly used to establish and divide acceleration regions in TOFMS. When different electric fields are placed on each side of a grid, a small electrostatic lens is produced at each opening. The effects of these lenses on the ion throughput and time resolution of instruments has been a point of controversy. Two published papers have indicated that the grids have little or no impact on the performance of the instrument. [3,4] Two others suggest that grids can have a considerable effect and in some cases may even be the limiting factor in instrument resolution. [5,7] We have used SIMION 3D v.6.0 to simulate the flight of ions through a reflectron instrument containing grids in both the source and reflector [8]. Our results showed that there can be a considerable effect on the trajectory of the ion. Grid effects were most pronounced when an ion was decelerated after passing through a grid as occurs in the reflector. At this point, the perpendicular velocity introduced by the field non-homogeneity at the grid could become a significant fraction of the total velocity. The sensitivity and resolution of the simulated instrument depended greatly on the electric fields and grid density.

Figure 1. Fate of Ions With 70 and 333 lpi Grids (Reflector With Grids)

Figure 1 compares these results obtained for 70 and 333 line per inch grids (lpi). The sensitivity of the instrument (with grids in the reflector) is approximately 3 times greater when 333 lpi grids are used. The lower transmission of these grids (70%) compared to that of the 70 lpi grids (90%) is compensated for by a reduction in ion scattering as the ions pass through the grids. When ions are scattered by the 70 lpi grids over 62% are given a new trajectory that prevents them from reaching the detector, almost all of the ions simulated in an instrument with 333 lpi grids either strike a grid or reach the detector. This also has a significant impact on resolution. When ions are scattered at two different grid transitions there can be a significant difference in their path length and therefore time-of-flight.

2. Simulated Flight Times: A. 70 lpi; B. 333 lpi

The flight time distributions shown in Figure 2 are entirely due to grid effects. No other effects are included in the simulation.

Simulation:

In this work we examine the effects of grid scattering in a Reflectron TOFMS with a gridless reflector. We have examined this problem for two reasons. First, it was observed that a significant fraction of the grid scattering in our earlier work [2,8] occurred during the 4 transitions ions made through reflector grids. We also noted that many of the ions that reached the detector only did so because there were significantly scattered at two grid transitions. The second event happened to return the ion to a trajectory that terminated at the detector. These ions contributed significantly to the temporal width of the signal at the detector. It should be noted that the gridless reflectors have been commercially available for several years (Brucker).

Figure 3. Simulated Instrument With Gridless Reflector  Electrostatic contour lines and an ions trajectory without grid effects are shown.

A three dimensional view of the simulated instrument is shown in Figure 3. The reflector in this instrument was identical to that used in Reference 8, except that the reflector grids have been removed and a ground can has been place around the reflector and part of the drift region. A two stage source and a two stage reflector were used. In the source, the ion was assumed to start at the first electrode and be given 10% of its drift energy in the first stage. Most of the acceleration, therefore, took place in the second stage. Each source region was 12.5 mm long. The ion mirror (or reflector) was positioned 475 mm from the source at an angle of 0.73 degrees from the primary axis of the source (z-axis). The reflector had regions of 12 (front) and 127 mm (back). Within each of these regions the potential along the surface of the reflector varied linearly. The point between the two regions was held at a potential equivalent to 90 % of the drift energy. The back of the reflector was held at a potential 1 % greater than the starting point of the ions in order to turn them back toward the detector. A 20 mm diameter detector was placed 500 mm from the reflector. Each of the instrument elements, source, reflector, and detector, were modeled as a separate instance and placed on the Ion Optics Bench. They were positioned such that, when the effects of grids were not considered, an ion originating from the source would strike the center of the detector. This is illustrated by the ion trajectory shown in Figure 3. A series of equi-potential contours are also shown to illustrate the effect that grid removal had on the electrostatic fields within the reflector. The simulation of a small section of grid occurred in a fourth instance. This instance was modeled with a very high density array of 0.0017 mm between potential array points. The entire instance required over 5 million points. While the grid instance is included in Figure 3, it is too small to be seen on the scale shown. It included nine grid holes and the volume within a distance of 1.5 grid holes on either side of the grid. At that distance, the electric field was assumed to be uniform and a pair of solid electrodes were used to establish the fields in either direction. (See References 3 and 8 for further details.) The simulation involved flying a large number of ions through the instrument. Every time an ion experienced a large change in electric field, a user program was used to "jump" the ion to the grid instance. The ion was placed in a random position just above the plane of one of the solid electrodes. The electrical fields within the grid instance were then adjusted to reflect the fields that the ion had experienced just before it jumped. Thus, the single grid instance could be used to simulate all of the grids within the instrument. Since the ion was placed in a random position along a side of the grid instance, the simulation was able to account for ions passing through different parts of a grid opening. In the current calculation, only the field along the z-axis is considered so it is assumed that grids are normal to this axis.

Figure 4. Paths of ions with grid effects.

Results:

Figure 4 shows the paths of several ions with a grid jumping active. Each time the ion encounters a grid, it jumps to a point at the lower right of the figure where the very small grid instance is positioned. After passing through the grid instance, the ion is then placed back in its original position. This occurs once for each of the two grid transitions. The fate of the ions is shown in Figure 5. A very large fraction of the ions missed the detector. It was found that the exact trajectory of the ions entering the reflector was critical to its reaching the detector. This was a result of the strongly diverging electrostatic lens produced by our simple reflector design. It was found, for instance, that the reflector angle had to be within +/- 0.13 degrees for the ion to reach the detector when grid effects were not included. This range was +/- 0.22 degrees when grids were included in the reflector. (It is assumed that modifications to the shapes and potentials of the reflector electrodes would be used to reduce the strength of the diverging lens. However, it is not our current goal to optimize the reflector design. Our purpose is to demonstrate the need for proper simulation and demonstrate a method for doing so.) Use of 333 lpi grids increased the fraction of detected ions by 15 times. This is an even more dramatic result than was found in the simulation of the gridded reflector. It was hoped that the few ions detected would at least represent a narrow range in flight times. However, the detected signals were on the order of 33 ns wide.

Figure 5. Fate of Ions With Gridless Reflector

Conclusions

Grid scattering is having a dramatic effect on both the sensitivity and resolution of the instrument. This is true even in instrument with gridless reflectors. We have shown the need and a method of simulating these effects. We are currently planning to compare our simulated results with experimental data.

References

1. David A. Dahl 43ed ASMS Conference on Mass Spectrometry and Allied Topics, May 21-26 1995, Atlanta, Georgia, 717.

2. S. M. Colby; C. W. Baker; J. J. Manura Proc, 41st ASMS Conf. 1996. (Available at www.sisweb.com application note #47)

3. X. Tang, R. Beavis, W. Ens, F. Lafortune, B. Schueler and K. G. Standing, Int. J. Mass Spectrom. Ion Processes, 85 (1988) 43.

4. D. Ioanoviciu, Int. J. Mass Spectrom. Ion Processes, 131 (1994) 43.

5. T. Bergmann, T. P. Martin and H. Schaber, Rev. Sci. Instrum., 60 (1989) 347.

6. R. C. King, R. Goldschmidt and K. G. Owens 39th ASMS Conference on Mass Spectrometry and Allied Topics, May 19-24 1994, Nashville, TN, 717.

7. V. V. Laiko and A.F. Dodonov, Rapid Comm. Mass Spectrom. 8 (1994) 720-726. 8. S. M. Colby PittCon97 730, and 1996 Eastern Analytical Symposium.