Triple-switching mirror cell pushes compact methane laser sensor to new precision

Researchers have built a methane sensor that squeezes a 53.2-meter optical path into a cell with an effective optical volume of just 268.27 milliliters, and used it to reach a detection precision of 12.6 parts per billion at an integration time of 9.4 seconds. The device, described in Applied Physics B, is based on a tunable diode laser absorption spectroscopy (TDLAS) architecture whose heart is a newly designed multipass cell the authors call the triple-switching matrix multipass cell, or TSMC.

Methane detection at trace levels matters across several fields. The authors point to natural gas pipeline leakage detection, greenhouse gas monitoring and noninvasive disease diagnosis as applications that demand ppm-level precision in chemically complex environments, while also requiring the portability needed for onsite inspection or regional scanning. TDLAS, which rests on the Beer-Lambert law, offers high selectivity and high sensitivity, and its non-contact measurement character and miniaturization potential make it attractive for trace methane work. The multipass gas cell is the core component that extends the laser’s path length and thereby improves the detection limit, which is why multipass cell design remains a major research focus in the field.

The problem the team set out to solve is a persistent tension in existing designs. Multipass cells used in current TDLAS methane sensors generally suffer from limited optical path length and poor beam quality when constrained to a small volume, which restricts overall sensor performance. As the authors put it in their introduction, existing multipass cells generally struggle to simultaneously meet the multiple requirements of a long optical path, small volume, high signal-to-noise ratio and long-term stability, limiting their use in portable detection devices.

The most familiar miniaturized design, the Herriott cell, has a simple structure, a regular circular spot pattern and low cost, which has made it popular in compact TDLAS sensors. Its weakness is low mirror utilization efficiency: to reach an effective optical path length on the order of tens of meters, the required gas cell volume is usually greater than one liter. Researchers have proposed rectangular-like spot and double-helix spot arrangements based on independent circles to improve mirror utilization and lengthen the path in smaller volumes, but those structures involve complex spherical mirror layouts and make spot pattern recognition difficult, significantly increasing the difficulty and complexity of system alignment and assembly. A second technical route uses dual-spherical-mirror structures with dense spot patterns, such as independent circles and concentric circles. Compared with traditional Herriott cells, these dense-spot multipass cells improve mirror utilization and can reach ratios of optical path length to volume (RLV) generally around 20 cm−2, but each spatial position of the spot is reflected only once, capping further improvement. Achieving a long path also forces the spacing between spots to shrink, and under nonparaxial conditions the spots are prone to distortion, which can cause the beam to exit the cell prematurely and induce interference effects that raise system noise and introduce detection errors.

The new cell builds on the Pickett-Bradley White cell (PBWC), a confocal cavity of three spherical mirrors with equal curvature radii that offers excellent beam quality and large spot spacing. The PBWC’s limitation is structural: it can only form two columns of spots, and the number of columns cannot be extended, so mirror utilization stays low. In 1991, Chernin and colleagues addressed this by proposing the three-objective and four-objective Chernin matrix multipass cells, which translate the PBWC reflection mode in the row direction and expand the original two columns into a matrix with flexibly adjustable rows and columns. The four-objective version lets the laser beam complete two reflections at almost every spatial position on the field mirror side, significantly improving mirror utilization, and its RLV exceeds that of traditional improved Herriott cells and the PBWC. Even so, the authors argue, its optical path length still needs further enhancement under small-volume constraints, and the small distance between the incident and exit positions tends to cause spatial interference between the incident collimation system and the detector. Later investigations achieved decoupling of the incident and exit positions, but the authors conclude that matrix multipass cell structures still require optimization, with increasing the reuse rate of reflection spot positions as the main objective.

The TSMC is the team’s answer. It uses six spherical mirrors with the same radius of curvature: one main field mirror and two auxiliary field mirrors on the field mirror side, and three objective mirrors on the opposite side. Taking the reflection structure of the conventional PBWC as the basic unit, the cell automatically switches the reflection mode along three directions. The three objective mirrors pair with the main field mirror to form three PBWCs: mirrors Mo1 and Mo2 serve as the objective mirrors for a vertical PBWC, Mo2 and Mo3 for an oblique PBWC, and Mo1 and Mo3 for a horizontal PBWC. The key insight is that if a single spatial position is to be reused three times, a traditional scheme would need six objective mirrors, but by sharing the centers of curvature of the three objective mirrors pairwise, the count is simplified from six to three, with each objective mirror participating in two sets of PBWC reflections. The centers of curvature are arranged in an isosceles right triangle, and two auxiliary field mirrors handle the switching, with their spherical centers located at the midpoints between the spots in the objective mirrors of the corresponding PBWC configurations.

The consequence is that almost every spatial position of the spot on the main field mirror undergoes three reflections, raising mirror utilization and enabling both a small volume and a long optical path. The spot pattern is a matrix type that the authors say is easily identified, and the design needs only six standard spherical mirrors, a number comparable to other matrix-type multipass cells. Compared with the four-objective Chernin cell, which also uses six spherical mirrors and yields a 6-row by 5-column spot arrangement corresponding to 90 reflections, the TSMC achieves 130 reflections under the same number of rows and columns, providing a longer effective optical path while keeping the exit position far from the incident position and thereby avoiding spatial interference between the incident collimation device and the detector. The confocal cavity structure contributes excellent beam quality: the exit position of a confocal multipass cell is insensitive to small perturbations in the incident beam position, and the spot pattern suffers little distortion. The relatively large spot spacings also avoid the premature beam exit problem seen in dense-spot cells, which otherwise introduces interference noise and degrades beam quality.

The geometry is also expandable and analytically tractable. Because the basic reflection mode expands in units of two rows and two columns, the field-mirror-side spot pattern can only form an even number of rows and an odd number of columns, while the pattern on the objective mirror side stays unchanged. For the general case of m rows and n columns, the total number of reflections follows the formula N = 6mn − 6m − 4n + 6. The authors also analyze optical stability, which they treat as an important indicator because it is closely related to long-term reliability. Under the paraxial approximation, the position of each spot follows from the symmetry of the spot two steps earlier about the center of curvature of the mirror holding the intervening spot, so as long as the spots remain on the mirrors rather than falling on the edges or outside them, small displacements of the spherical mirrors do not affect the spot distribution or the operation of the cell. Tilt-angle errors introduced during assembly shift the centers of curvature and are the primary factor affecting stability; on the objective-mirror side the three-row, two-column distribution is preserved under small shifts, and on the field-mirror side the cell still ensures normal beam output as long as the spots do not fall onto the gaps between adjacent mirrors.

To validate the design, the team fabricated a prototype with an 8-row by 7-column spot pattern on the field mirror side, giving 266 reflections. All mirrors are spherical with a radius of curvature of 20 cm and a high-reflectance coating of 99.9% reflectivity. The beam enters through a fiber collimator at the input aperture, a rectangular clear opening of 3.3 mm by 3.3 mm, and exits at a separate output aperture. The main field mirror measures 25 mm by 20 mm, each auxiliary field mirror 21 mm by 5 mm, and each objective mirror 14 by 14 mm, with mirror spacings held within 1 mm on the field mirror side and 3 mm on the objective mirror side. The resulting main cell body has external dimensions of 207 mm by 36 mm by 36 mm, corresponding to the 268.27 mL effective optical volume, defined as the minimum volume of the gas-interaction space enclosed by all spherical mirrors, excluding detachable optical alignment and assembly fixtures. In the 266-reflection mode this yields the 53.2 m effective optical path and an RLV of 19.83 cm−2. The authors note transparently that the laboratory prototype carries three-axis adjustment stages on both sides for assembly and calibration, and that the total prototype volume including these structures is 485.33 mL, giving an RLV of approximately 10.96 cm−2; they argue these adjustable mounting structures are temporary tooling for prototype debugging rather than inherent components of the optical design, and can be replaced by monolithic packaging in engineering applications.

The sensor itself uses a distributed feedback laser with a central wavelength of 1653 nm, locked near the methane absorption peak at 6046.95 cm−2 by a temperature of 25.64 °C and a current of 60.73 mA, with the exiting light detected by a Ge photodetector. Wavelength modulation spectroscopy was employed, with the laser scanned by a 10 Hz sawtooth wave and modulated by a 20 kHz sinusoidal signal, and the second-harmonic signal extracted in software. In calibration measurements over 0 to 30 ppm, prepared by mixing 30 ppm methane standard gas with high-purity nitrogen using mass flow controllers, a quadratic polynomial corrected a slight nonlinear effect, and linearity tests at 0, 6, 12, 18, 24 and 30 ppm gave a fitting residual R² above 0.9999. Allan-Werle deviation analysis on a continuous 300 s measurement of 30 ppm methane showed white-noise-dominated behavior up to the optimal integration time of 9.4 s, where the detection precision reaches 12.6 ppb, with the concentration distribution following a Gaussian profile of half width at half maximum 150.78 ppb. At a gas flow rate of 1 L/min, the average response time over three rising edges, from 0 to 90% of the final concentration, was 20.5 s. The authors present the TSMC as a new approach for improving high-precision detection of trace gases, with the design strategy of switching among three PBWC reflections using two auxiliary field mirrors offered as a reference for developing higher-performance multipass cells.

Subject of Research: A compact TDLAS methane sensor based on a novel triple-switching matrix multipass cell with high optical path length to volume ratio

Article Title: Compact TDLAS methane sensor based on a triple-switching matrix multipass cell

Article References: Li, S., Cao, L., Ou, X., Dong, K., Liu, P., & Zhou, X. (2026). Compact TDLAS methane sensor based on a triple-switching matrix multipass cell. Applied Physics B, 132(10), Article 105. https://doi.org/10.1007/s00340-026-08719-9

Image Credits: AI Generated

DOI: 10.1007/s00340-026-08719-9

Keywords: methane sensing, TDLAS, multipass cell, optical path length, wavelength modulation spectroscopy, trace gas detection, Pickett-Bradley White cell, Chernin cell, Allan-Werle deviation, miniaturization, infrared laser, gas sensor