Optomechanical mass spectrometry

The study addresses the limitations of conventional and nanomechanical mass spectrometry (MS) for analyzing large supramolecular assemblies. Traditional MS relies on ionization and measures ensemble mass-to-charge ratios, but commercially available instruments struggle above a few megadaltons (MDa). Nanomechanical resonators enable single-particle mass sensing by correlating frequency shifts to accreted mass, achieving extremely low limits of detection. However, early nanomechanical systems inherited constraints from ion-based methods (limited mass range, long analysis time, low capture cross-section). Later, neutral nanomechanical MS extended the measurable range to MDa–GDa without ionization, aided by arrays of resonators to improve capture efficiency. Yet, prior devices were one-dimensional beams/cantilevers where the mass-induced frequency shift depends on landing position, and can also depend on particle stiffness, size, and shape. This necessitates complex multi-mode readouts and degrades average mass resolution, while hampering the analysis of elongated biological particles (e.g., tailed viruses, fibrils). The research question is whether optomechanical nanoresonators, enabled by advanced VLSI optomechanics, can perform single-particle MS in a single mode with position- and shape-insensitive response, improving capture area and simplifying readout without sacrificing resolution.

• Conventional MS has progressed toward larger assemblies but remains limited above a few MDa.
• Nanomechanical resonators have demonstrated yoctogram sensitivity and enable neutral particle MS in the MDa–GDa range; arrays increased capture cross-section and enabled analysis of ~100 MDa virus capsids with high resolution.
• One-dimensional flexural resonators (beams, cantilevers) suffer from position-dependent frequency shifts and sometimes require accounting for particle stiffness, size, and shape, necessitating multi-mode detection and degrading average resolution and capture efficiency.
• Cavity optomechanics offers extreme displacement sensitivity and high bandwidth for sensing applications, including mass sensing, suggesting potential for lighter, higher-frequency resonators and new device topologies that could overcome the limitations of electrical transduction and 1D mechanics.

Device and mode topology: The "nano-ram" optomechanical resonator comprises a thin sensing platform supported by four beams, operating in an in-plane rigid-body translation mode such that the platform does not deform. Consequently, the mass-induced frequency shift is independent of particle landing position, stiffness, size, or aspect ratio.

Design and fabrication: Fabricated on 200 mm SOI wafers with a VLSI process. The platform and supports were thinned to 60 nm to reduce resonator mass while maintaining a large capture area (platform ~1.5 µm × 3 µm; supports ~80 nm × 500 nm). The optical subsystem is a single-mode silicon ring resonator (radius 10 µm, diameter 20 µm) coupled to a waveguide with a 200 nm gap; ring-to-platform gap ~100 nm. Electrostatic actuation uses a side-gate 250 nm from the nanoresonator. The SOI top layer (220 nm) is partially etched for grating couplers. Silicon is locally highly doped for low-resistance electrical contacts. After device patterning and metallization, a planarized oxide sacrificial layer and ~200 nm amorphous Si protective layer are deposited; the protection is opened only above the sensing platform with sub-100 nm alignment tolerance. Release is by vapor HF etch.

Optomechanical readout and actuation: The mechanical motion modulates the optical cavity transmission in the bad-cavity regime. Output optical power modulation per displacement scales as ΔP ∝ Qopt·Cr·gom. The ring operates under-coupled with contrast Cr≈0.23 and loaded optical Q≈5×10^4; losses dominated by scattering at ring spokes (tapered to ~100 nm). Laser wavelength is set on the steepest slope of the cavity resonance; transmitted power is demodulated near the mechanical resonance. Optomechanical coupling rate gom≈0.4 GHz/nm. Motion is electrically actuated (electrostatic) and optically detected, providing high signal-to-background decoupling.

Operating conditions and performance characterization: Experiments at ~77 K and ~10^-5 Torr. The first in-plane mode at ~44.9 MHz shows mechanical Q≈1700. Thermomechanical noise is resolved; the detector noise floor corresponds to displacement resolution ~2.8×10^-14 m Hz^-1. Frequency is tracked with a PLL; Allan deviation shows a ~1 ppm stability limit over integration times due to intrinsic mechanical frequency fluctuations; stability improves with drive until this limit, below the onset of mechanical nonlinearity.

Mass sensitivity and limit of detection: Minimum detectable mass δm_min derived as δm_min = 2·M_res·(δf_min/f0). With M_res≈608 fg and ~1 ppm frequency stability, δm_min≈0.7 MDa (~1.2 ag). Single-mode operation eliminates the need to resolve multiple modes and maintains uniform sensitivity over the entire capture area.

Optical packaging and protection: To ensure stable transduction during particle deposition, optical and electrical components are covered by an ~200 nm amorphous Si layer; only the sensing platform is exposed to particles, preventing optical detuning or scattering loss from accretion. Light is coupled via on-chip grating couplers (pitch 0.6 µm, width 0.3 µm, optimized for ~1550 nm at ~10° incidence). Fiber-transposer chips (Teem Photonics) are UV-glued to grating couplers for robust, compact coupling compatible with vacuum; standard feedthroughs route optical and electrical signals.

Mass spectrometry experiments: A multichamber vacuum system integrates a sputtering gas-aggregation source (for metallic nanoclusters) and an in-line TOF-MS. Tantalum clusters (dense, ρ≈16.6 g/cm^3) with tunable size are generated, pass through differential pumping, and impinge on the device in a deposition chamber (~10^-5 Torr) on a retractable, cryogenic stage. Retracting the stage allows sequential TOF-MS measurement of the same cluster population. TOF detection of masses up to ~7 MDa is enabled by increasing ion acceleration from 3 to 3.6 kV.

Data acquisition and processing: The resonator frequency is monitored in closed loop (PLL response ~10 ms). Frequency jumps from single-particle landings are converted to mass via Δm = 2·M_res·(Δf/f0). Time traces are averaged (10–20 ms depending on event rate), scanned for abrupt steps; jump heights are quantified by pre/post-step frequency levels. Masses are binned into histograms. To mitigate false positives near the detection limit, frequency noise is characterized for 30 s before and after each run; noise-derived "mass" histograms (scaled to run duration) are subtracted bin-wise from deposition histograms.

• Demonstrated single-particle mass spectrometry using an optomechanical nanoresonator operating in a single, in-plane rigid-body mode that is insensitive to particle landing position, stiffness, size, and shape.
• Achieved threefold increase in capture area compared to top-down beam resonators with electrical readout, without degrading mass resolution, despite relying on a single resonance mode.
• Optomechanical transduction resolved thermomechanical noise and reached the mechanical stability limit: mechanical resonance at ~44.9 MHz with Q≈1700; frequency stability ≈1 ppm over integration times; optomechanical coupling gom≈0.4 GHz/nm; optical Q≈5×10^4 with contrast ≈0.23.
• Minimum detectable mass ≈0.7 MDa (~1.2 ag) for a resonator mass of ~608 fg.
• Performed MS of tantalum clusters with optomechanical spectra showing mean masses from ~2.7 to 7.7 MDa (diameters ~8–11.3 nm). For one population (~5.7 MDa), recorded 1140 events in 5 minutes, depositing ~1% of the resonator mass without performance degradation.
• TOF-MS detected charged clusters up to ~7 MDa (improved via higher ion acceleration), but underrepresented high-mass tails; optomechanical MS, with mass-independent detection limit, captured high-mass events with better relative precision.
• Device and optical transduction remained stable during prolonged deposition (up to ~10% of initial resonator mass in <1 hour) with no measurable degradation in optical Q, transmission, or resonance detuning.

The work addresses longstanding limitations of nanomechanical MS based on one-dimensional flexural resonators, where mass readout requires multi-mode monitoring and the resolution depends strongly on landing position and, in some cases, particle stiffness and shape. By engineering an in-plane rigid-body mode, the optomechanical "nano-ram" provides uniform mass sensitivity across the entire platform and eliminates position and shape dependencies. The single-mode readout simplifies instrumentation and data processing while enabling the full capture area to be used, improving throughput. Optomechanical transduction affords the displacement sensitivity and bandwidth necessary for thin, light, high-frequency resonators. The combination of optomechanical readout and electrostatic actuation allows operation near the intrinsic frequency stability limit, translating to sub-megadalton mass resolution. Experimental validation with neutral tantalum clusters demonstrates accurate single-particle mass measurements across 2.7–7.7 MDa, capturing high-mass distributions more faithfully than TOF-MS, whose detection efficiency drops for very massive ions. The system shows robustness to substantial mass loading without degradation, supporting practical use. These results show that optomechanical resonators are a superior alternative to electrical nanoresonators for high-mass, single-particle MS, unlocking the analysis of non-spherical, elongated biological and synthetic nanoparticles that are problematic for 1D devices.

This study demonstrates high-resolution, single-particle mass spectrometry using a VLSI-fabricated optomechanical nanoresonator operating in a single, in-plane rigid-body mode. The platform provides position- and shape-insensitive mass readout, a threefold larger capture area than state-of-the-art electrically read beam resonators, and comparable or better mass resolution, with a minimum detectable mass of ~0.7 MDa. The device successfully measured tantalum cluster populations up to ~7.7 MDa with high throughput and stability. Future directions include: (i) multiplexing large arrays using wavelength-division multiplexing for high-throughput analyses; (ii) further thinning the resonators to ~10 nm and stiffening anchors to push resonance frequencies >500 MHz and mass resolution below 100 kDa; and (iii) extending applications to structurally complex, non-spherical biological nanoparticles (e.g., amyloid fibrils, tailed viruses) that are challenging for current nanomechanical MS.

• Current minimum detectable mass is ~0.7 MDa; measurements below ~3 MDa approach the resolution limit and require careful noise characterization and statistical subtraction to avoid false positives.
• Experiments conducted at cryogenic temperature (~77 K) and high vacuum (~10^-5 Torr), which may complicate deployment in some settings.
• Optical cavity operated under-coupled with scattering losses from anchoring spokes limiting optical Q; although sufficient, this contributes to modest contrast (≈0.23).
• Frequency stability is limited by intrinsic mechanical resonance frequency fluctuations to ~1 ppm, setting the present mass resolution ceiling.
• Validation experiments used metallic (tantalum) clusters; biological particle demonstrations remain to be shown within this platform (though design is intended to be geometry-insensitive).