Nested Herriott Cells

• Nested Herriott cells are multipass optical architectures characterized by multiple interleaved spot families that increase effective interaction length while preserving compactness.
• They combine conventional Herriott designs with controlled perturbations to improve mirror-area utilization and reduce spot overlap for more uniform interaction.
• Applications in bulk amplification, gas post-compression, and atomic sensing demonstrate the potential of nested designs for optimizing high-pass optical performance.

Searching arXiv for recent and directly relevant work on Herriott and nested-like multipass cell architectures. Nested Herriott cells are multipass optical architectures in which the beam path is organized into more than one coupled or interleaved Herriott-like spot family within a compact footprint. In the strict architectural sense, they are distinct from a single re-entrant Herriott orbit, a folded two-mirror Herriott cell, or a single intracavity focal-plane imaging scheme. The recent literature surveyed here shows that genuinely “nested” implementations remain uncommon in arXiv reports, but several closely related geometries develop the same underlying objectives: increasing pass count, extending effective interaction length, improving mirror-aperture utilization, distributing interaction sites, and preserving compactness. In that sense, nested Herriott cells are best understood as a broader design direction within advanced Herriott-derived multipass optics rather than as a single standardized topology. The most relevant recent studies are a quasi-waveguide bulk amplifier based on a single symmetric Herriott-type cell (Meyer et al., 2024), an integrated alkali-vapor recirculating multipass cell that explicitly behaves like “multiple Herriott cells within a single, compact cavity” (Kee et al., 17 Jun 2025), a folded compact multipass cell that adds planar mirrors to compress a long Herriott-type path (Schönberg et al., 2024), a compact vapor-cell Herriott cavity with same-hole re-entrant routing (Hao et al., 2021), and an analytical optimization of nonlinear phase accumulation in finite-aperture gas-filled symmetric Herriott cells (Meyer et al., 29 Apr 2026).

1. Terminology and scope

The term nested Herriott cell is often used loosely to describe any Herriott-derived geometry that achieves high path length in a compact package by means beyond the canonical two-mirror circular spot pattern. The papers considered here make a sharper distinction.

In “Quasi-waveguide amplifiers based on bulk laser gain media in Herriott-type multipass cells” (Meyer et al., 2024), the reported system is a single symmetric Herriott-type cell with a gain crystal placed at the refocus plane. The beam forms a sequence of laterally shifted foci inside the crystal and a circular spot pattern on the mirrors, but the paper explicitly does not present a true nested or compound Herriott architecture. The separated focal spots are generated by one periodic orbit, not by multiple coupled sub-cells (Meyer et al., 2024).

“Compact, folded multi-pass cells for energy scaling of post-compression” (Schönberg et al., 2024) likewise does not present multiple nested Herriott sub-cells. Its compactness is obtained by folding the long propagation segment of a symmetric Herriott-type cell with two planar mirrors. The underlying mode and re-entrant condition remain defined by the two concave mirrors, so the architecture is more precisely a folded Herriott-type cell than a nested one (Schönberg et al., 2024).

By contrast, “Spin Correlations in Recirculating Multipass Alkali Cells for Advancing Quantum Magnetometry” (Kee et al., 17 Jun 2025) is the closest recent arXiv work to a nested-Herriott concept. It presents a three-mirror recirculating alkali-vapor multipass cell and explicitly characterizes it as behaving like multiple Herriott cells within a single, compact cavity. The optical pattern does not remain confined to a single Lissajous or elliptical family; instead, small controlled mirror tilts cause the beam to migrate through displaced spot families, increasing coverage and reducing overlap (Kee et al., 17 Jun 2025). This suggests that, in current usage, nested Herriott cells are best distinguished from simpler Herriott variants by the presence of multiple coordinated spot families or substructures rather than merely by high pass count.

A further boundary case is the “Herriott-Cavity-Assisted Closed-Loop Xe Isotope Comagnetometer” (Hao et al., 2021), where a single Herriott cavity is integrated into a compact vapor cell. It uses a re-entrant same-hole topology and 21 reflections, but it does not introduce multiple spot families, multiple coupled cells, or nested reuse of mirror area beyond a single geometric trajectory (Hao et al., 2021).

2. Geometric principles

At the core of all Herriott-derived systems is periodic beam evolution between curved reflectors. In a symmetric two-mirror Herriott cell, the discrete orbit is parameterized by the number of reflections on one mirror, $N$, and a progression parameter, $M$, which determines how the reflection sequence advances through the spot pattern (Meyer et al., 2024, Meyer et al., 29 Apr 2026).

For the symmetric bulk-amplifier geometry, the eigenmode waist radius is given as

$$w_{0,\mathrm{eigen}}=\sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)},$$

and the mirror-plane mode radius is given as

$$w_{m,\mathrm{eigen}} = \frac{w_{0,\mathrm{eigen}}}{\cos\!\left(\frac{\pi M}{2N}\right)} = \sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)\sec\!\left(\frac{\pi M}{2N}\right)}.$$

The relation between the mirror spot-pattern radius $r_m$ and the interaction-plane radius $r_0$ is

$r_0 = r_m \cos\!\left(\frac{\pi M}{2N}\right).$

These expressions are central to any discussion of nested or quasi-nested cells because they map discrete orbit geometry onto beam size, mirror occupancy, and the distribution of internal interaction sites (Meyer et al., 2024).

A related single-cell formulation appears in the 2-$\mu$m gas-filled MPC optimization study, which uses the symmetric-cell relation

$L = 2R \sin^2\!\left(\frac{\pi M}{2N}\right),$

with the spot-to-spot progression angle

$\vartheta = 2\pi \frac{M}{N}.$

That work emphasizes finite mirror size as the dominant constraint on achievable pass count and shows that maximizing per-pass intensity near the concentric limit does not generally maximize total nonlinear phase shift for a given mirror set (Meyer et al., 29 Apr 2026). This is directly relevant to nested architectures because nested designs are typically motivated by the same finite-aperture bottleneck.

The recirculating three-mirror cell reformulates the problem with an ABCD description. One round trip is represented by

$M$0

and in the simplified plane-plus-concave limit the angular step between successive spots is

$M$1

Without tilt, the spot coordinates follow harmonic recurrences,

$M$2

which are standard Herriott-type periodic motions (Kee et al., 17 Jun 2025). The nested-like behavior emerges only after controlled perturbation of this base motion.

3. From single-orbit cells to nested-like spot families

The most important conceptual distinction in the recent literature is between repeated imaging along a single orbit and genuine generation of multiple coupled spot families.

The quasi-waveguide amplifier repeatedly refocuses the beam into a Ti:sapphire crystal placed at the cell focus. The foci appear at distinct lateral positions inside the crystal while the mirror spots lie on the familiar circular Herriott pattern. The paper explicitly states that this is not a nested or compound Herriott architecture; the mirror ring and crystal ring are merely two conjugate planes of the same periodic orbit (Meyer et al., 2024). Its significance lies in demonstrating that a single Herriott cell can generate a distributed set of interaction sites within an internal plane.

The recirculating alkali cell goes further. Small opposite tilts of the two mirrors $M$3 and $M$4 introduce a drift term into the ordinary Herriott-like motion. For one step,

$M$5

and after $M$6 reflections,

$M$7

The first term is ordinary periodic motion; the second is a cumulative offset that causes recirculation into new displaced spot families (Kee et al., 17 Jun 2025). The corresponding offset is written as

$M$8

This is the strongest recent analytical description of a nested-like Herriott strategy: controlled perturbation is used to convert one orbit family into a sequence of offset families, thereby increasing mirror-area utilization and reducing self-overlap.

The paper’s exit criterion,

$M$9

with total reflections

$$w_{0,\mathrm{eigen}}=\sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)},$$0

shows explicitly how pass count depends not only on the base Herriott angular increment $$w_{0,\mathrm{eigen}}=\sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)},$$1 but also on the family-shift parameter $$w_{0,\mathrm{eigen}}=\sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)},$$2 (Kee et al., 17 Jun 2025). A plausible implication is that nested Herriott design can be viewed as controlled management of family drift in addition to ordinary re-entrant closure.

4. Compactification strategies

One of the principal motivations behind nested and related Herriott architectures is compactification: preserving long optical path length while reducing physical size.

The folded compact multipass cell provides the clearest recent demonstration of this objective. It begins from a symmetric Herriott-type cell with re-entrant condition

$$w_{0,\mathrm{eigen}}=\sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)},$$3

and introduces a folding ratio $$w_{0,\mathrm{eigen}}=\sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)},$$4 by forcing the beam to undergo multiple reflections between two planar mirrors during each nominal transit between the concave focusing mirrors. The physical end-to-end size becomes approximately

$$w_{0,\mathrm{eigen}}=\sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)},$$5

In the demonstrated system, $$w_{0,\mathrm{eigen}}=\sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)},$$6, $$w_{0,\mathrm{eigen}}=\sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)},$$7, $$w_{0,\mathrm{eigen}}=\sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)},$$8, and $$w_{0,\mathrm{eigen}}=\sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)},$$9, giving an unfolded inter-focusing-mirror length of $$w_{m,\mathrm{eigen}} = \frac{w_{0,\mathrm{eigen}}}{\cos\!\left(\frac{\pi M}{2N}\right)} = \sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)\sec\!\left(\frac{\pi M}{2N}\right)}.$$0, an effective folded length of $$w_{m,\mathrm{eigen}} = \frac{w_{0,\mathrm{eigen}}}{\cos\!\left(\frac{\pi M}{2N}\right)} = \sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)\sec\!\left(\frac{\pi M}{2N}\right)}.$$1, packaged to about $$w_{m,\mathrm{eigen}} = \frac{w_{0,\mathrm{eigen}}}{\cos\!\left(\frac{\pi M}{2N}\right)} = \sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)\sec\!\left(\frac{\pi M}{2N}\right)}.$$2, while maintaining about $$w_{m,\mathrm{eigen}} = \frac{w_{0,\mathrm{eigen}}}{\cos\!\left(\frac{\pi M}{2N}\right)} = \sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)\sec\!\left(\frac{\pi M}{2N}\right)}.$$3 of total propagation (Schönberg et al., 2024). This is not nesting in the strict sense, but it fulfills a major nested-Herriott design goal: dense long-path routing in a short device.

The same-hole re-entrant comagnetometer cell achieves compactness differently. It uses two crossed cylindrical mirrors, each with curvature 100 mm, diameter 12.7 mm, relative angle 52°, and mirror separation 19.3 mm. The linearly polarized probe enters through a 2.5 mm hole in the center of the front mirror and exits through the same hole after 21 reflections (Hao et al., 2021). This same-port re-entrance is highly relevant to nested thinking because it demonstrates aperture reuse within a sealed compact vapor cell, even though the architecture remains single-family rather than nested.

The bulk quasi-waveguide amplifier also offers a compactification mechanism of a different sort. By repeatedly refocusing seed and pump into the crystal, it creates a waveguide-like long effective interaction length within a bulk medium. In the experimental configuration, $$w_{m,\mathrm{eigen}} = \frac{w_{0,\mathrm{eigen}}}{\cos\!\left(\frac{\pi M}{2N}\right)} = \sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)\sec\!\left(\frac{\pi M}{2N}\right)}.$$4 reflections per mirror were geometrically possible, but 17 crystal passes were used, giving an effective crystal interaction length of $$w_{m,\mathrm{eigen}} = \frac{w_{0,\mathrm{eigen}}}{\cos\!\left(\frac{\pi M}{2N}\right)} = \sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)\sec\!\left(\frac{\pi M}{2N}\right)}.$$5, compared with a single-pass reference of $$w_{m,\mathrm{eigen}} = \frac{w_{0,\mathrm{eigen}}}{\cos\!\left(\frac{\pi M}{2N}\right)} = \sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)\sec\!\left(\frac{\pi M}{2N}\right)}.$$6 and a hypothetical single-pass bulk design near the Rayleigh-range scale of $$w_{m,\mathrm{eigen}} = \frac{w_{0,\mathrm{eigen}}}{\cos\!\left(\frac{\pi M}{2N}\right)} = \sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)\sec\!\left(\frac{\pi M}{2N}\right)}.$$7 (Meyer et al., 2024). This suggests that internal repeated imaging can compactify interaction length even without true nested mirror topology.

5. Optimization under finite-aperture and loss constraints

A persistent misconception in Herriott-cell design is that operation as close as possible to the concentric resonator always maximizes performance. Recent work shows that this is not generally true once finite mirror size, clipping margin, and cumulative loss are included.

The 2-$$w_{m,\mathrm{eigen}} = \frac{w_{0,\mathrm{eigen}}}{\cos\!\left(\frac{\pi M}{2N}\right)} = \sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)\sec\!\left(\frac{\pi M}{2N}\right)}.$$8m gas-filled Herriott study derives the total nonlinear phase shift in an ideal symmetric cell as

$$w_{m,\mathrm{eigen}} = \frac{w_{0,\mathrm{eigen}}}{\cos\!\left(\frac{\pi M}{2N}\right)} = \sqrt{\frac{\lambda_0}{2\pi}R\sin\!\left(\pi\frac{M}{N}\right)\sec\!\left(\frac{\pi M}{2N}\right)}.$$9

with approximate critical power

$r_m$0

Because finite aperture limits feasible $r_m$1 pairs, maximizing total nonlinear phase becomes a problem of maximizing $r_m$2 under clipping constraints, not of maximizing per-pass focusing intensity (Meyer et al., 29 Apr 2026).

The no-clipping condition is expressed as

$r_m$3

with safety factor $r_m$4 typically between 2 and 3. From this, the paper identifies an optimum

$r_m$5

rather than the near-concentric limit $r_m$6 (Meyer et al., 29 Apr 2026). For $r_m$7 mm, the optimized max-$r_m$8 configuration can yield about 2× higher nonlinear phase shift than the longest $r_m$9 geometry, and if the shorter geometry also permits about 1.5× higher gas pressure under the same pressure-volume limit, the theoretical improvement can reach about 3× (Meyer et al., 29 Apr 2026).

Losses further modify the optimum. With mirror reflectivity $r_0$0,

$r_0$1

Mirror fluence scales as

$r_0$2

so high pass count can rapidly become impractical (Meyer et al., 29 Apr 2026). This tradeoff applies directly to nested cells: extra structural complexity does not remove the mirror-area, fluence, and reflectivity constraints; it only redistributes them.

The folded CMPC reaches a similar conclusion in a different application. Because planar folding mirrors can intercept the beam anywhere along the inter-mirror path, the limiting criterion becomes focus fluence rather than only fluence on the concave mirrors. This yields

$r_0$3

and, using the folding ratio,

$r_0$4

The new degree of freedom is thus $r_0$5, which trades physical length against folding density (Schönberg et al., 2024). A plausible implication is that a true nested Herriott design could be analyzed analogously by introducing stage-coupling or family-multiplicity parameters alongside the usual $r_0$6 or $r_0$7 relations.

6. Applications and performance regimes

Recent Herriott-derived architectures have been developed across bulk amplification, gas-based post-compression, and atomic sensing, and these application domains highlight different reasons for pursuing nested-like designs.

In bulk amplification, the quasi-waveguide geometry uses collinearly overlapped pump and seed beams, both mode-matched to the cell eigenmodes, so that each pass yields a similar waist at a different lateral location in the gain crystal. In the reported Ti:sapphire demonstration, the mirrors had ROC $r_0$8 and reflectivity $r_0$9, the seed waist at 800 nm was $r_0 = r_m \cos\!\left(\frac{\pi M}{2N}\right).$0, the pump waist at 532 nm was $r_0 = r_m \cos\!\left(\frac{\pi M}{2N}\right).$1, and 93% pump absorption was reached after 17 passes through a 2 mm crystal, corresponding to 34 mm effective crystal propagation. Maximum measured gain was 13%, with optimized values at full power typically 9%–13%, compared with 1%–4% for a single-pass reference (Meyer et al., 2024). The paper attributes the higher gain to the longer effective interaction length and distributed thermal load, though thermal degradation remained severe at high pump power (Meyer et al., 2024).

In gas-filled post-compression, the folded CMPC demonstrates that a 45 cm device can perform the role of a much longer cell. The reported system compressed 8 mJ, $r_0 = r_m \cos\!\left(\frac{\pi M}{2N}\right).$2 ps, 1030 nm pulses in atmospheric air to 51 fs, with measured throughput 89% and inferred average mirror reflectivity $r_0 = r_m \cos\!\left(\frac{\pi M}{2N}\right).$3 per reflection over 550 reflections (Schönberg et al., 2024). It also presents a scaling scenario for 200 mJ, 1 ps, 1030 nm pulses using $r_0 = r_m \cos\!\left(\frac{\pi M}{2N}\right).$4, $r_0 = r_m \cos\!\left(\frac{\pi M}{2N}\right).$5, and an effective setup length of 2.5 m, with simulated compression to 80 fs and transmission close to 80% (Schönberg et al., 2024). These results are directly relevant to any nested-Herriott discussion of high-energy compact propagation.

The 2-$r_0 = r_m \cos\!\left(\frac{\pi M}{2N}\right).$6m gas-filled single-cell study addresses a different regime: relatively low peak-power pulse compression where ordinary near-concentric design becomes suboptimal. Using $r_0 = r_m \cos\!\left(\frac{\pi M}{2N}\right).$7 mm, mirror diameter 50 mm, $r_0 = r_m \cos\!\left(\frac{\pi M}{2N}\right).$8, $r_0 = r_m \cos\!\left(\frac{\pi M}{2N}\right).$9, and $\mu$0 mm in a compact high-pressure vessel up to 30 bar, it demonstrated self-compression at 2120 nm from 197 fs to below 46 fs in a negative-dispersion regime, and compression to 55 fs at 2000 nm in a positive-dispersion regime after external dispersion compensation (Meyer et al., 29 Apr 2026). The work is not nested, but it provides a finite-aperture optimization principle likely to be foundational for nested high-pass-count designs.

In atomic sensing, the single Herriott-cavity Xe comagnetometer uses passive optical path-length enhancement rather than magnetic parametric modulation to improve readout sensitivity. The cavity-assisted system achieved an angle random walk of $\mu$1, bias instability of $\mu$2 $\mu$3, and closed-loop bandwidth of 1.5 Hz, while the Herriott cavity improved Rb magnetometer sensitivity by one order of magnitude compared with a conventional cell (Hao et al., 2021). The recirculating alkali cell extends this domain by connecting optical geometry to spin-correlation noise. It reports examples with 78 reflections and 120 reflections, an analytical-vs-Zemax average position error of 0.0122 mm over 120 reflection positions, and emphasizes improved beam coverage and reduced overlap relative to cylindrical multipass cells (Kee et al., 17 Jun 2025).

7. Design implications, misconceptions, and open directions

Several misconceptions are corrected by the recent literature.

The first is that any multi-spot or multi-plane Herriott-derived geometry should be called nested. The quasi-waveguide amplifier is explicitly not nested: its mirror ring and crystal focal ring are two conjugate planes of one periodic orbit (Meyer et al., 2024). The folded CMPC is likewise not nested; it is one Herriott-type path geometrically compressed by planar mirrors (Schönberg et al., 2024). The Xe comagnetometer cavity is re-entrant and compact but still a single-family geometry (Hao et al., 2021). The recirculating alkali cell is the closest current example to a nested-Herriott concept because it intentionally generates multiple displaced spot families within one compact cavity (Kee et al., 17 Jun 2025).

The second misconception is that maximum pass performance follows automatically from the most near-concentric geometry. The finite-aperture analysis at 2 $\mu$4m shows that this fails once spot size on the mirrors limits usable $\mu$5; the optimum under idealized single-cell assumptions is instead near $\mu$6 and $\mu$7 (Meyer et al., 29 Apr 2026). The recirculating cell makes a related point in a different language: pass count alone is not an adequate figure of merit if it is achieved at the cost of spot overlap, tightly focused hot regions, or poor active-volume coverage (Kee et al., 17 Jun 2025).

The third misconception is that compactification is “free.” Folded or nested routing trades reduced physical length against additional reflections, stricter coating requirements, greater sensitivity to displacement of folding optics or mirror tilts, and more complicated clipping margins (Schönberg et al., 2024, Kee et al., 17 Jun 2025). In atomic vapor cells, it must also contend with mirror contamination, thermal engineering, and field-gradient effects (Hao et al., 2021).

Three technical directions emerge from the recent literature. First, family engineering—using controlled perturbations such as mirror tilts to generate displaced spot families—appears to be a practical route toward nested-like behavior in compact devices (Kee et al., 17 Jun 2025). Second, global optimization rather than local focusing optimization is essential: the relevant objective may be total nonlinear phase, effective interaction length, active-volume coverage, or distributed thermal loading, but in each case it is a whole-system quantity rather than a single-pass intensity maximum (Meyer et al., 2024, Meyer et al., 29 Apr 2026). Third, compactness parameters beyond classical Herriott integers are becoming important. The folding ratio $\mu$8 in CMPCs is one example (Schönberg et al., 2024); a true nested-cell formalism would likely require analogous parameters for family multiplicity, inter-stage coupling, or shared-aperture routing.

Taken together, these studies indicate that nested Herriott cells are not yet represented on arXiv by a dominant canonical architecture. Instead, the field currently consists of a set of converging design strategies: repeated intracell imaging, recirculating multi-family spot patterns, same-hole re-entrant compact cavities, and folded long-path cells. The strongest current evidence suggests that the distinctive value of a nested Herriott approach would lie not merely in increasing pass count, but in coordinating multiple spot families or substructures so as to improve aperture utilization, interaction uniformity, compactness, and loss-limited performance beyond what a single periodic Herriott orbit can provide (Kee et al., 17 Jun 2025, Meyer et al., 29 Apr 2026).