Evaluating Galvanometric Mirror Dynamics for High Speed Multimodal Laser Scanning Microscopy

In a comparative evaluation of galvanometric mirror scanners integrated into operational multimodal laser scanning microscopy platforms, researchers found that a Cambridge Technology scanner system demonstrated roughly two-fold faster transient step response, significantly smaller phase delays, and superior amplitude retention across driving frequencies from 200 to 800 hertz compared to an alternative Thorlabs scanner setup. Conducted by Marta Bukumira, Stanko Nikolić, Nikola Stojanovic, Lidija Mančić, Nataša V. Todorović, Ana Senkić, Mihailo D. Rabasović, and Aleksandar J. Krmpot, the experimental investigation directly assessed how electromechanical scanner performance limits imaging speed, geometric precision, and field fidelity within fully assembled instruments. Galvanometric mirrors serve as central beam-steering components in laser scanning systems such as confocal, two-photon, and harmonic generation microscopes, where their physical motion governs beam deflection across biological and material specimens. While scanner specifications are often evaluated through isolated component datasheets, this study focused on real-world microscope assemblies, revealing how the coupled influences of mirror inertia, closed-loop driver electronics, and feedback damping produce measurable spatial distortions and scanning nonlinearities during rapid optical data acquisition.

Galvanometric scanners operate by deflecting an incoming laser beam using a small mirror with carefully engineered rotational inertia mounted onto an electromechanical actuator shaft. When incorporated into laser scanning microscopes, these devices provide an effective balance between operational speed, deflection precision, and equipment cost, while their metallic reflective surfaces accommodate broad laser spectral bandwidths required for multiphoton and nonlinear optical imaging. The dynamic behavior of a galvanometric scanning unit can be modeled as a second-order linear time-invariant system governed by classic mass-spring-damper mechanical characteristics. In this electromechanical framework, the balance established between the mirror assembly rotational inertia, electromagnetic torque generated by the actuator, and damping in the closed-loop feedback network establishes the natural resonance frequency and stability boundaries of the scanner. From a signal processing perspective, the finite mechanical bandwidth causes the galvanometric mirror to function fundamentally as a low-pass filter, which progressively attenuates the higher frequency components of commanded driving waveforms.

To establish how these physical principles govern real microscopy performance, the researchers examined two operational platforms designated as System A and System B. System A was built as an upright non-descanning microscope equipped with a pair of Cambridge Technology 6215 H closed-loop galvanometer optical scanners driven by dedicated Cambridge driver electronics. System B was configured as an inverted descanning microscope utilizing a pair of Thorlabs GVS002 closed-loop galvanometric scanners coupled to a Thorlabs driver module, incorporating a transimpedance amplifier and photomultiplier tube detection. Both instrument platforms arranged the paired galvanometric mirrors in a conventional orthogonal two-axis configuration, where a horizontal fast axis executes repetitive line scanning and a vertical slow axis controls frame progression. In both optical designs, the mirrors were relayed to the objective back focal plane through a four-f telecentric lens configuration to guarantee uniform telecentric beam deflection across the specimen plane. Both scanner pairs were driven and managed through an identical unified software interface adapted to match manufacturer operational limits.

The experimental methodology focused on three central dynamic properties that directly dictate imaging acquisition speed, trajectory precision, and system stability: transient step response, frequency-dependent amplitude and waveform reproduction, and dynamic phase delay. Mirror positional tracking was acquired directly from internal position detectors within the driver electronics, yielding a true galvo output signal that is proportional to actual mirror deflection angle and records real-time deviations from commanded trajectory profiles. Testing began with step response measurements driven by a 1.5-volt input step signal to evaluate mirror agility during sudden position changes. A rapid, well-damped step response is vital for repositioning laser scanning mirrors between successive frames while minimizing dead time and avoiding unnecessary beam exposure in sensitive samples, such as fungal cell walls during laser microsurgery procedures. Evaluating agility using the standard ten-to-ninety percent rise time criterion demonstrated that System A responded roughly twice as fast as System B, exhibiting a steeper transition and faster rise to steady state.

To replicate realistic image acquisition regimes, both scanner pairs were tested using triangular drive signals across an operational frequency range of 200 to 800 hertz. Triangular signals emulate practical line scanning, wherein mirror assemblies must accelerate and decelerate at turning points while preserving nearly constant angular velocity across the central field of view. The researchers evaluated two primary operational states shared by both microscopes: an 800-hertz live mode optimized for real-time sample positioning and signal optimization at 256-by-256 pixel resolution with a 2.5-microsecond pixel dwell time and 0.64-millisecond half-period oscillations, and a 200-hertz high-quality acquisition mode collecting 1024-by-1024 pixel frames at the identical 2.5-microsecond dwell time, corresponding to 2.5 milliseconds per scanned line. The recorded galvo output waveforms displayed noticeable distortions relative to commanded inputs, retaining input frequency but exhibiting altered amplitudes and rounding at turning points that worsened as driving frequency increased due to the physical inertia of the mirror assemblies.

Because ideal constant-velocity deflection occurs only along the linear portion of the triangular waveform, waveform rounding directly truncates the usable linear imaging region. The authors adopted an empirical threshold defining acceptable image quality as a maximum four percent deviation from the ideal triangular command waveform. Under this four percent criterion, approximately two-thirds of the field of view remains undistorted during 800-hertz live mode operation, whereas ninety-five percent of the image area remains undistorted during 200-hertz high-quality scanning. Beyond waveform deformation, the experimental data revealed significant frequency-dependent amplitude attenuation when driving the systems with a constant 1-volt input amplitude. Tracking peak-to-peak output voltage averaged across five consecutive cycles demonstrated that both systems experienced gradual amplitude decreases as frequency rose. This attenuation was noticeably more severe in System B, which suffered substantial amplitude loss at higher oscillation frequencies, directly shrinking the achievable physical scanning angle and reducing the effective field of view along the fast scanning axis.

Measurements of frequency-dependent phase delay further highlighted the operational contrast between the two integrated platforms across the 200 to 800 hertz span. Phase shift was calculated from mean time delay measurements across five consecutive cycles using the relationship multiplying time delay by frequency and 360 degrees. While both scanners exhibited low phase delay and faithful signal tracking at lower frequencies, dynamic delay grew progressively larger as driving frequencies climbed. At an intermediate frequency of 400 hertz, System A registered a phase delay of approximately 23 degrees, whereas System B reached over 30 degrees. At the maximum operational frequency of 800 hertz, the divergence became pronounced, with System A recording a phase delay of 46.7 degrees compared to 62.8 degrees for System B. The authors noted that commercial galvanometer specifications are traditionally defined under sinusoidal driving conditions, meaning that when driven by triangular signals, the higher harmonic components fall outside the frequency envelope where manufacturer performance guarantees apply.

The physical consequences of these electromechanical limitations were directly confirmed through representative optical imaging experiments. In two-photon fluorescence imaging of SCC-25 oral cancer cells acquired at high resolution, finite mirror inertia produced distinct spatial distortions near the perimeter of the scanning field, where scanner trajectories exhibited localized stretching and bending. Because high-resolution acquisition allowed roughly ninety-five percent undistorted field coverage, edge stretching remained confined to peripheral regions where the actuator required higher torque to turn the mirror assembly. Scanner-induced distortions were further demonstrated using a physical calibration target featuring five-micrometer groove spacings scanned with System B at oscillation frequencies of 130 hertz and 700 hertz. Comparing raw, unprocessed images confirmed that operating at 700 hertz substantially compressed the observed field of view compared to 130 hertz. This reduction demonstrated how dynamic amplitude attenuation and transient loss of synchronization near maximum deflection angles physically restrict the accessible imaging region during high-speed imaging.

In both microscope platforms, image acquisition executes bidirectional line scanning, wherein each horizontal line is scanned in both directions and averaged to construct the final composite image. Without calibration, the phase delay and rotational inertia of the galvanometric mirrors introduce a spatial misregistration between the forward and backward passes, manifesting as severe horizontal discontinuities and structural blurring. To correct this offset, the researchers implemented a half-image offset parameter, which represents the precise pixel shift needed to register forward and backward scans into structural alignment. The optimal half-image offset was determined experimentally by computing the two-dimensional spatial autocorrelation of the raw misregistered image, where the offset corresponds to the distance from the origin to the first autocorrelation intensity peak. The researchers demonstrated this calibration using second harmonic generation microscopy of tungsten disulfide monolayers grown by chemical vapor deposition. At 1024-by-1024 pixel resolution, proper offset adjustment restored structural fidelity, while tests across frequencies at 256-by-256 pixels demonstrated that the required offset scales directly with driving frequency.

By benchmarking galvanometric scanners as integrated assemblies within functioning optical microscopes rather than standalone benchtop components, the study establishes a practical framework for configuring and troubleshooting high-speed laser scanning systems. The findings demonstrate that dynamic bandwidth limitations, feedback loop responsiveness, and mirror inertia collectively dictate achievable line rates, usable field dimensions, and spatial linearity. While System A demonstrated superior agility, smaller phase lag, and higher dynamic bandwidth across the full testing envelope, the characterization methodology demonstrates how software-based compensation techniques, such as autocorrelation-derived half-image offset correction, can mitigate phase-induced misregistration in slower scanning systems. The authors noted that integrating dynamic characterization protocols into automated scanner correction routines represents a critical next step for high-speed imaging of dynamic biological phenomena. Such automated compensation would maintain spatial fidelity at maximum frame acquisition rates, allowing researchers to observe rapid physiological and cellular events without sacrificing structural accuracy or incurring excessive laser exposure.

Subject of Research: Galvanometric mirror dynamic performance in multimodal laser scanning microscopy

Article Title: Performance evaluation of galvanometric mirrors for fast scanning in multimodal laser microscopy

Article References: Bukumira, M., Nikolić, S., Stojanovic, N., Mančić, L., Todorović, N. V., Senkić, A., Rabasović, M. D., & Krmpot, A. J. (2026). Performance evaluation of galvanometric mirrors for fast scanning in multimodal laser microscopy. Applied Physics B, 132(6), Article 83. https://doi.org/10.1007/s00340-026-08693-2

Image Credits: AI Generated

DOI: 10.1007/s00340-026-08693-2

Keywords: laser scanning microscopy, galvanometric mirrors, multimodal imaging, step response, phase delay, waveform distortion, multiphoton microscopy, image fidelity, Performance, evaluation, galvanometric, mirrors