A GPU ACCELERATED CDMS TRAP TRAJECTORY SIMULATOR AND OPTIMISER

Significance of the topic

CDMS integrates charge and m/z measurement to analyze large ions where conventional mass spectrometers struggle. Optimizing the ELIT geometry and voltages maximizes m/z resolution, underpinning accurate mass determination in heterogeneous samples.

Objectives and study overview

This work presents a GPU-accelerated trajectory simulator and voltage‐geometry optimiser for CDMS traps. The goal is to rapidly explore electrode configurations and operating parameters to achieve substantially higher m/z resolution than existing three‐electrode designs.

Methodology and instrumentation

Trajectories are computed using a fourth‐order Runge‐Kutta solver implemented in PyCUDA, with electric field maps generated by SIMION 2020. Single‐ion runs model axial oscillations without FFT transients, extracting mean crossing periods for resolution estimates. Optimization employs a modified Nelder‐Mead simplex across multiple GPU threads, varying electrode voltages and tube length. Initial ion phase space (axial KE spread, radial position/angle) is simulated based on upstream quadrupole optics. The platform runs on NVIDIA GV100 GPUs, achieving ~100× speedup versus single‐core CPU simulations.

Main results and discussion

Optimized four‐electrode traps with extended tube lengths (~155–175 mm) deliver Gaussian resolutions exceeding 600 k and FWHM resolutions >900 k, a five‐ to ten‐fold improvement over three‐electrode counterparts. A 14 mm electrode geometry outperformed a 12 mm layout in peak shape and resolution (~1.7× gain). Resolution strongly depends on initial phase‐space constraints, with tighter energy and angular spreads boosting performance to >2 million resolution. Moreover, four‐electrode designs exhibit enhanced tolerance to axial KE shifts, enabling multi‐ion trapping with minimal resolution loss.

Benefits and practical applications

• Rapid optimisation of trap configurations accelerates design cycles for high‐resolution CDMS.
• Superior m/z resolution enhances analysis of megadalton complexes and heterogeneous assemblies.
• Increased tolerance to energy spread supports parallel ion trapping, improving throughput.
• GPU acceleration democratizes complex trajectory studies by reducing computation time.

Future trends and opportunities

Automating physical geometry adaptation, integrating phase‐space feedback from upstream optics, and expanding optimisation to 3D asymmetric traps promise further resolution gains. Coupling with multi‐ion space‐charge models will facilitate real‐time trap tuning for high‐throughput CDMS in structural biology, polymer science, and nanomaterials.

Conclusion

The GPU‐based trajectory simulator and optimiser achieve m/z resolutions over three orders greater than current ELIT designs, with ~100× computational acceleration. Optimized four‐electrode traps offer high stability, resolution, and tolerance to kinetic energy variations, paving the way for advanced CDMS applications and parallel ion analysis.

References

1. Hogan JA, Jarrold MF. Optimized Electrostatic Linear Ion Trap for Charge Detection Mass Spectrometry. J Am Soc Mass Spectrom. 2018;29(10):2086–2095.
2. Zhao Q, Urosević D, Mladenović N, Hansen P. A restarted and modified simplex search for unconstrained optimization. Comput Oper Res. 2009;36(12):3263–3271.
3. Yavor MI, Pomozov TV, Kirillov SN, Khasin YI, Verenchikov AN. High performance gridless ion mirrors for multi-reflection time-of-flight and electrostatic trap mass analyzers. Int J Mass Spectrom. 2018;426:1–11.