Integrated Hollow Waveguide Absorbers

• Integrated hollow waveguide absorbers are designs that embed absorptive structures directly within hollow guides to jointly control impedance, transmission, absorption, and thermal properties.
• They are implemented in diverse systems such as acoustic ducts using sidewall resonators, superconducting filters with leaky-wave absorbers, and radiometric sources with conical dielectric tapers.
• Design trade-offs include balancing precise mechanical tolerances, material properties, and integration with system-level requirements like ventilation, passband fidelity, and cryogenic isolation.

Searching arXiv for the cited papers to ground the article in current records. Integrated hollow waveguide absorbers are absorbing structures incorporated directly into hollow guiding geometries so that dissipation, impedance control, and wave transport are co-designed rather than separated. In the cited literature, the term encompasses sidewall-mounted hybrid membrane resonators in a clear acoustic waveguide for ventilation (Fu et al., 2016), hollow circular waveguides integrated into a rectangular coaxial stepped-impedance low-pass filter for superconducting circuits (Andersson et al., 4 Aug 2025), and a waveguide-mounted absorbing conical dielectric taper that forms a thermally isolated radiometric source inside a metallic guide (Rostem et al., 2013). Across these implementations, the absorber is not merely appended to a termination; it is structurally embedded in the waveguide system and used to shape transmission, reflection, absorption, emissivity, or thermal loading.

1. Conceptual scope and architectural variants

The central architectural feature is integration of the absorber into the hollow waveguide itself. In the acoustic implementation, a hybrid membrane resonator (HMR) is a decorated membrane resonator backed by a shallow sealed cavity, mounted flush on the sidewall of a clear waveguide so that the main flow channel remains fully open (Fu et al., 2016). In the superconducting-circuit implementation, circular hollow waveguides are embedded in the high-impedance sections of a rectangular coaxial stepped-impedance filter and oriented orthogonally to the main transmission axis (Andersson et al., 4 Aug 2025). In the radiometric implementation, the absorber is an absorbing conical dielectric taper centered inside an electroformed square copper waveguide, with the absorber thermally isolated while the waveguide walls are heat sunk to the cold bath (Rostem et al., 2013).

These variants realize different physical functions. The acoustic HMR creates a side-mounted monopolar response that can drive the waveguide boundary toward a soft condition, strongly suppress transmission, or, in a two-resonator arrangement, produce near-total absorption by destructive interference and impedance matching (Fu et al., 2016). The superconducting filter combines reflective near-band filtering with absorptive far out-of-band suppression: the stepped-impedance network defines the low-pass behavior, while the hollow waveguides become leaky-wave absorbers above cutoff and drain high-frequency energy into a carbon-loaded polyethylene foam termination (Andersson et al., 4 Aug 2025). The radiometric source is instead a near-ideal blackbody termination, where the absorber is integrated inside the guide to provide emissivity $\epsilon \approx 0.999$ over $33$–$50\ \mathrm{GHz}$ (Rostem et al., 2013).

A concise comparison is given below.

Platform	Integrated absorber form	Representative function
Clear acoustic duct	Sidewall-mounted HMRs	Asymmetric total absorption/reflection at $286.7\ \mathrm{Hz}$; multiple-band absorption below $1\ \mathrm{kHz}$ (Fu et al., 2016)
Rectangular coaxial microwave filter	Orthogonal circular hollow waveguides with absorber foam terminations	Low insertion loss in passband and more than $52.7\ \mathrm{dB}$ rejection above $17.3\ \mathrm{GHz}$ (Andersson et al., 4 Aug 2025)
Cryogenic radiometric source	Waveguide-mounted absorbing conical dielectric taper	Emissivity of $0.999$ over the full waveguide band (Rostem et al., 2013)

This range of embodiments suggests that “integrated hollow waveguide absorber” is best understood as a design class rather than a single mechanism. A plausible implication is that integration becomes valuable when the absorber must coexist with another system-level requirement, such as ventilation, passband fidelity, or cryogenic thermal isolation.

2. Acoustic sidewall absorbers based on hybrid membrane resonators

In the acoustic realization, the waveguide cross section is $90\ \mathrm{mm} \times 90\ \mathrm{mm}$, with characteristic acoustic impedance

$Z_0 = \rho_0 c_0 / A,$

where $33$0, $33$1, and $33$2 (Fu et al., 2016). The HMR front footprint is $33$3 with a circular opening of diameter $33$4 closed by a rubber membrane of thickness $33$5 and diameter $33$6. The cavity is rectangular, sealed, and has depth $33$7 for the single-frequency pair. The decoration masses are $33$8 for HMR-1 and $33$9 for HMR-2; a $50\ \mathrm{GHz}$0-mm-thick sponge was placed at the back of HMR-1’s cavity. The sidewall mounting leaves the air path unobstructed (Fu et al., 2016).

The physical mechanism is described in lumped form. The decorated membrane resonator surface response is

$50\ \mathrm{GHz}$1

the cavity impedance is

$50\ \mathrm{GHz}$2

the membrane impedance is

$50\ \mathrm{GHz}$3

and the HMR shunt impedance seen by the duct is

$50\ \mathrm{GHz}$4

For two HMRs separated by $50\ \mathrm{GHz}$5, the input impedance of the intermediate duct segment terminated by $50\ \mathrm{GHz}$6 is

$50\ \mathrm{GHz}$7

and the combined waveguide impedance at the HMR-1 plane is

$50\ \mathrm{GHz}$8

Near-total absorption requires impedance matching,

$50\ \mathrm{GHz}$9

with

$286.7\ \mathrm{Hz}$0

These relations formalize the sidewall resonator as a shunt load that can transition from strong reflection to strong absorption depending on detuning and spacing (Fu et al., 2016).

The single-frequency pair used center-to-center spacing $286.7\ \mathrm{Hz}$1. Simulated individual resonances were $286.7\ \mathrm{Hz}$2 for HMR-1 and $286.7\ \mathrm{Hz}$3 for HMR-2. For incidence from the HMR-1 side, at $286.7\ \mathrm{Hz}$4, $286.7\ \mathrm{Hz}$5 and $286.7\ \mathrm{Hz}$6, yielding measured absorption $286.7\ \mathrm{Hz}$7, low transmission, and a reflection minimum. For incidence from the HMR-2 side at the same frequency, $286.7\ \mathrm{Hz}$8, so the system behaves as a soft boundary with high reflection (Fu et al., 2016). Field evidence showed that at $286.7\ \mathrm{Hz}$9, incidence from HMR-1 produced velocities of approximately $1\ \mathrm{kHz}$0 incident in both cavities with opposite phases near each HMR, whereas incidence from HMR-2 strongly excited only HMR-2 at approximately $1\ \mathrm{kHz}$1, with HMR-1 weakly excited at approximately $1\ \mathrm{kHz}$2 (Fu et al., 2016).

The multiple-frequency pair, HMR-3 and HMR-4, used sidewall spacing $1\ \mathrm{kHz}$3, cavity front plate $1\ \mathrm{kHz}$4, cavity depth $1\ \mathrm{kHz}$5, and aperture $1\ \mathrm{kHz}$6 sealed by a rectangular membrane of thickness $1\ \mathrm{kHz}$7. Two semicircular platelets of radius $1\ \mathrm{kHz}$8 and thickness $1\ \mathrm{kHz}$9 were used on each membrane; HMR-3 had masses $52.7\ \mathrm{dB}$0 and $52.7\ \mathrm{dB}$1, while HMR-4 had masses $52.7\ \mathrm{dB}$2 and $52.7\ \mathrm{dB}$3 with positions detuned by $52.7\ \mathrm{dB}$4 (Fu et al., 2016). Measured absorption peaks for incidence from the HMR-3 side occurred at $52.7\ \mathrm{dB}$5, $52.7\ \mathrm{dB}$6, $52.7\ \mathrm{dB}$7, $52.7\ \mathrm{dB}$8, and $52.7\ \mathrm{dB}$9 with $17.3\ \mathrm{GHz}$0 ranging from approximately $17.3\ \mathrm{GHz}$1 to above $17.3\ \mathrm{GHz}$2. Theory reproduced four of the five peak frequencies and magnitudes well; peak-I showed approximately $17.3\ \mathrm{GHz}$3 offset, attributed to omitted structural subtleties such as membrane–platelet contact areas, exact mass distributions, and stress nonuniformities (Fu et al., 2016).

The acoustic study therefore establishes a clear version of integrated hollow waveguide absorption in which a sidewall device can provide strong, multi-band, and asymmetric absorption without obstructing airflow. This suggests that absorber effectiveness need not be tied to guide occlusion when a large-aperture side coupling mechanism is available.

3. Electromagnetic hollow-waveguide absorbers in superconducting filters

The superconducting-circuit realization is a $17.3\ \mathrm{GHz}$4st-order stepped-impedance low-pass filter implemented in a rectangular coaxial geometry, formed by two CNC-machined copper halves as the outer conductor and a flat copper strip as the center conductor (Andersson et al., 4 Aug 2025). Alternating high- and low-impedance sections are created by modulating the outer conductor width $17.3\ \mathrm{GHz}$5 and height $17.3\ \mathrm{GHz}$6, and the center-conductor height $17.3\ \mathrm{GHz}$7. High-impedance sections behave as series inductors and low-impedance sections as shunt capacitors when their electrical lengths are short compared to the wavelength. Hollow waveguides are embedded in the high-impedance sections and oriented orthogonally to the transmission axis, symmetrically above and below the center conductor at locations of large transverse electric field. Two radii are used, $17.3\ \mathrm{GHz}$8 and $17.3\ \mathrm{GHz}$9, with MACOR filling of relative permittivity $0.999$0 and depth $0.999$1 (Andersson et al., 4 Aug 2025).

The electromagnetic description combines stepped-impedance low-pass synthesis with hollow-waveguide cutoff physics. For short sections, the inductive and capacitive mappings are

$0.999$2

and the TEM impedance of the rectangular coax is

$0.999$3

In the prototype, $0.999$4 and $0.999$5 (Andersson et al., 4 Aug 2025). For the circular hollow waveguides, the cutoff frequency is

$0.999$6

with $0.999$7 for $0.999$8 and $0.999$9 for $90\ \mathrm{mm} \times 90\ \mathrm{mm}$0. Using the reported radii, the $90\ \mathrm{mm} \times 90\ \mathrm{mm}$1 cutoff estimates are approximately $90\ \mathrm{mm} \times 90\ \mathrm{mm}$2 for $90\ \mathrm{mm} \times 90\ \mathrm{mm}$3 and approximately $90\ \mathrm{mm} \times 90\ \mathrm{mm}$4 for $90\ \mathrm{mm} \times 90\ \mathrm{mm}$5; the $90\ \mathrm{mm} \times 90\ \mathrm{mm}$6 cutoffs are approximately $90\ \mathrm{mm} \times 90\ \mathrm{mm}$7 and approximately $90\ \mathrm{mm} \times 90\ \mathrm{mm}$8, respectively (Andersson et al., 4 Aug 2025).

Below cutoff, the axial propagation constant becomes evanescent and the fields decay as

$90\ \mathrm{mm} \times 90\ \mathrm{mm}$9

with $Z_0 = \rho_0 c_0 / A,$0. At $Z_0 = \rho_0 c_0 / A,$1, for $Z_0 = \rho_0 c_0 / A,$2 and $Z_0 = \rho_0 c_0 / A,$3, $Z_0 = \rho_0 c_0 / A,$4 so $Z_0 = \rho_0 c_0 / A,$5; for $Z_0 = \rho_0 c_0 / A,$6 and $Z_0 = \rho_0 c_0 / A,$7, $Z_0 = \rho_0 c_0 / A,$8 so $Z_0 = \rho_0 c_0 / A,$9 (Andersson et al., 4 Aug 2025). The stated interpretation is that this prevents parasitic in-band loss. Above cutoff, the same waveguides transition from evanescent to propagating behavior and act as leaky-wave absorbers that couple energy out of the transmission line and into a carbon-loaded polyethylene foam termination (Andersson et al., 4 Aug 2025).

The device was optimized using a hybrid parametric model. Inductive sections containing the hollow waveguides were simulated in full $33$00D COMSOL FEM as a function of physical length $33$01, interpolated to obtain continuous $33$02 and $33$03, and converted to ABCD matrices for cascading. Uniform capacitive sections used standard transmission-line ABCD matrices,

$33$04

A differential evolution global optimizer in SciPy minimized a cost function that penalized passband ripple, non-flat group delay in $33$05–$33$06, insufficient return loss up to approximately $33$07, and insufficient attenuation in approximately $33$08–$33$09 (Andersson et al., 4 Aug 2025).

Ten inductive sections incorporate hollow-waveguide apertures, symmetrically above and below the center conductor. Conductors are oxygen-free copper outer blocks and a $33$10 copper center strip; PTFE inserts define $33$11 alignment sections; absorber foam caps the waveguides away from the transmission line (Andersson et al., 4 Aug 2025). A calibrated vector network analyzer measured the prototype up to $33$12. Measurements confirmed a $33$13 cutoff frequency at $33$14, insertion loss below $33$15 for frequencies under $33$16, and more than $33$17 rejection above $33$18. For the final prototype, some hollow-waveguide apertures in sections $33$19, $33$20, and $33$21 were enlarged from $33$22 to $33$23 diameter to further suppress re-transmission peaks (Andersson et al., 4 Aug 2025).

The intended systems context is qubit protection. Superconducting quasiparticles are generated by photons with $33$24, corresponding to approximately $33$25 for thin-film Al, material-dependent. The integrated hollow waveguides absorb broadly above their $33$26 cutoffs, while the measured stopband attenuation exceeds $33$27 already above approximately $33$28 (Andersson et al., 4 Aug 2025). This does not imply direct absorption only at the pair-breaking threshold; rather, the architecture suppresses high-frequency radiation that can upconvert or propagate toward that regime.

4. Waveguide-mounted conical dielectric tapers as integrated blackbody terminations

A third implementation realizes integrated absorption as a waveguide-coupled thermal source (Rostem et al., 2013). The absorber is an absorbing conical dielectric taper of Eccosorb MF-117 mounted inside an electroformed square copper waveguide sized to be compatible with WR22.4 Q-band components. The characteristic dimension is $33$29, giving fundamental cutoff

$33$30

The square guide supports the two degenerate fundamental modes $33$31 and $33$32, enabling dual polarization, and the absorber is centered to preserve symmetry and avoid higher-order mode conversion (Rostem et al., 2013).

The cone is bonded to an annealed copper pin inserted $33$33 into the cone, then to a copper bobbin with conductive silver epoxy. A LakeShore Cernox CX-1050 thermometer and a $33$34 thin-film metal resistor heater are bonded to the bobbin, while $33$35 twisted-pair NbTi leads provide electrical connection and thermal isolation (Rostem et al., 2013). Static centering tolerance is $33$36 per side, achieved with $33$37 shims, and the absorber–wall separation is less than $33$38. This centering is identified as critical for equal coupling to both polarizations and for suppression of cross-polar and higher-order mode conversion (Rostem et al., 2013).

The absorber material properties are specified. MF-117 has relative permittivity $33$39, $33$40, and $33$41; the penetration depth in bulk at $33$42 is approximately $33$43. Bulk attenuation varies by approximately $33$44 from $33$45 to $33$46 due to disordered metallic loading. The measured in situ specific heat is

$33$47

between $33$48 and $33$49, with absorber mass $33$50 (Rostem et al., 2013).

The electromagnetic performance is summarized by the emissivity decomposition

$33$51

The tip-area upper bound on reflectance is

$33$52

which, for tip diameter $33$53 and square guide height $33$54, gives $33$55 or $33$56. Measurements show approximately $33$57 absorber reflection and flange reflection below $33$58 with careful clamping alignment; transmittance through the absorber mount is less than $33$59 (Rostem et al., 2013). The resulting emissivity is quoted as $33$60 over $33$61–$33$62, with the dominant deviation from unity attributed to waveguide ohmic loss rather than absorber reflectance (Rostem et al., 2013).

The conductor-loss model uses

$33$63

and

$33$64

with guide impedance

$33$65

Over a length $33$66 in the emitting region,

$33$67

Using electroformed copper with $33$68, the estimated emissivity bias is set by $33$69 across the band (Rostem et al., 2013).

Thermal isolation is achieved with a Kevlar-thread kinematic suspension under a steel-spring preload of $33$70, rigid in $33$71–$33$72, prohibiting rotation about $33$73, and leaving only $33$74 translation less than $33$75 unconstrained. The thermal link follows

$33$76

with $33$77, $33$78, and $33$79, implying

$33$80

The dominant thermal time constant is $33$81 at $33$82; finite-element modeling indicates $33$83 at $33$84 and approximately $33$85 at $33$86 (Rostem et al., 2013). In this device, integrated absorption is therefore inseparable from thermal engineering, because radiometric utility depends on controlling the absorber temperature while keeping wall emission subdominant.

5. Modeling frameworks and performance metrics

Although the three systems operate in different regimes, all are formulated through waveguide impedance, modal or cutoff behavior, and measurable scattering quantities. In the acoustic HMR system, the absorber is represented as a shunt impedance loading a duct; pairwise interaction is described through transmission-line relations for $33$87, $33$88, and $33$89, and absorption is defined by

$33$90

For a shunt-loaded section, an impedance retrieval from measured reflection is

$33$91

followed by

$33$92

Insertion loss is defined as

$33$93

and bandwidth may be characterized by the full-width at half-maximum of $33$94 or $33$95 peaks, with $33$96 (Fu et al., 2016).

In the superconducting filter, the principal observables are $33$97 and $33$98. Insertion loss is

$33$99

return loss is

$50\ \mathrm{GHz}$00

and the $50\ \mathrm{GHz}$01 cutoff frequency is the point where $50\ \mathrm{GHz}$02 drops by $50\ \mathrm{GHz}$03 relative to the low-frequency passband (Andersson et al., 4 Aug 2025). The hollow waveguides are interpreted as added frequency-dependent shunt admittances $50\ \mathrm{GHz}$04 above cutoff, while the stepped-impedance network sets the reflective low-pass response (Andersson et al., 4 Aug 2025).

In the radiometric source, characterization uses reflectance, transmittance, emissivity, and available thermal power. Reflectance extraction from a two-tier calibration with and without a quarter-wave shim is

$50\ \mathrm{GHz}$05

with

$50\ \mathrm{GHz}$06

The available single-mode thermal power is

$50\ \mathrm{GHz}$07

where

$50\ \mathrm{GHz}$08

In the Rayleigh–Jeans limit this reduces to

$50\ \mathrm{GHz}$09

These metrics differ in physical interpretation, but each quantifies how strongly the absorber modifies the modal energy carried by the guide (Rostem et al., 2013).

A plausible implication is that integrated hollow waveguide absorbers are best compared through system-level observables rather than through a single universal absorption metric. In one case the relevant figure is $50\ \mathrm{GHz}$10, in another $50\ \mathrm{GHz}$11 rejection, and in another emissivity.

6. Design trade-offs, misconceptions, and extension paths

The cited works present several recurring design trade-offs. In the acoustic HMR duct, increasing decoration mass raises $50\ \mathrm{GHz}$12 and lowers

$50\ \mathrm{GHz}$13

while larger cavity volume lowers the hybrid resonance through

$50\ \mathrm{GHz}$14

Controlled damping, such as sponge in the cavity back or grease on the membrane, widens bandwidth but can reduce peak $50\ \mathrm{GHz}$15 if overdone; the reported guidance is to start with $50\ \mathrm{GHz}$16 comparable to measured line widths, for example $50\ \mathrm{GHz}$17–$50\ \mathrm{GHz}$18 at approximately $50\ \mathrm{GHz}$19 (Fu et al., 2016). The asymmetric effect also relies on small detuning and precise spacing, with tolerance on $50\ \mathrm{GHz}$20 within a few millimeters at sub-kHz (Fu et al., 2016).

In the superconducting filter, the radius $50\ \mathrm{GHz}$21, dielectric loading $50\ \mathrm{GHz}$22, and depth $50\ \mathrm{GHz}$23 set the balance between passband invisibility and out-of-band absorptive coupling. The guidance is to require $50\ \mathrm{GHz}$24–$50\ \mathrm{GHz}$25 so that $50\ \mathrm{GHz}$26–$50\ \mathrm{GHz}$27 below cutoff, while larger $50\ \mathrm{GHz}$28 increases above-cutoff coupling but raises the risk of parasitic loading if cutoff approaches the passband (Andersson et al., 4 Aug 2025). The high ratio $50\ \mathrm{GHz}$29 sharpens roll-off but increases sensitivity to tolerances, and precise machining and transition design are identified as limitations (Andersson et al., 4 Aug 2025).

In the radiometric source, reducing the cone tip diameter lowers reflectance, but increased absorber size raises mass and therefore thermal time constant. Stronger thermal links shorten $50\ \mathrm{GHz}$30 but increase cryogenic loading, and higher-frequency scaling tightens centering and flange-alignment tolerances (Rostem et al., 2013). The design therefore trades electromagnetic match, thermal isolation, and mechanical repeatability against each other.

Several common misconceptions are not supported by the cited implementations. One is that strong absorption necessarily requires blocking the guide aperture. The acoustic HMRs are flush-mounted on the sidewall and keep the main flow channel fully open, yet achieve $50\ \mathrm{GHz}$31 at $50\ \mathrm{GHz}$32 in a $50\ \mathrm{GHz}$33 duct (Fu et al., 2016). Another is that adding absorbers to a waveguide must inevitably degrade useful transmission. The superconducting device instead combines insertion loss below $50\ \mathrm{GHz}$34 for frequencies under $50\ \mathrm{GHz}$35 with strong high-frequency rejection (Andersson et al., 4 Aug 2025). A third is that an absorber integrated into a hollow guide is merely a matched load. The radiometric source shows that the absorber can be an active metrological element whose thermal state is deliberately controlled while the surrounding waveguide is maintained near the bath temperature (Rostem et al., 2013).

The extension paths mentioned in the cited work are also diverse. For acoustic ducts, side-mounted HMR arrays can be distributed azimuthally or axially in multi-mode ducts, and adaptive designs may use tunable masses or variable-tension membranes (Fu et al., 2016). For superconducting circuits, proposed directions include elliptical or ridged hollow waveguides, graded $50\ \mathrm{GHz}$36 liners, metamaterial liners, direct co-optimization of hollow-waveguide diameter and termination in the differential-evolution loop, and co-design with device housings to eliminate box modes and minimize multipath (Andersson et al., 4 Aug 2025). For waveguide-coupled thermal sources, the absorber and suspension are stated to scale to higher GHz and sub-mm bands, with smaller cones reducing thermal mass and ring-center flanges helping maintain alignment (Rostem et al., 2013).

Taken together, these studies show that integrated hollow waveguide absorbers are not defined by a single absorber material or a single band of operation. They are defined by absorber-waveguide co-design: sidewall monopolar loading in clear ducts, above-cutoff leaky-wave extraction in superconducting filters, and waveguide-mounted lossy tapers in cryogenic radiometry. The underlying unifier is the deliberate use of the hollow guide as part of the absorbing mechanism rather than as a passive container.