MFB: A Mid-Frequency-Band Space Gravitational Wave Observer for the 2020 Decade

Ilya Mandel; John Lipa; Peter F. Michelson; Robert L. Byer; Sasha Buchman; Shally Saraf

REVIEW 4 major objections 5 minor 2 references

MFB: A Mid-Frequency-Band Space Gravitational Wave Observer for the 2020 Decade

T0 review · 4 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A geosynchronous, 73,000 km-arm gravitational-wave observatory could reach LISA-like sensitivity in the 10 mHz–1 Hz band and launch in the 2020s for $500M–$1B.

desk verdict A credibly argued mission concept whose headline sensitivity claim rests on an unproven sensor goal, not a demonstrated requirement. read the letter →

arxiv 1908.02861 v1 pith:NPRF7KAC submitted 2019-08-07 astro-ph.IM physics.space-ph

Peter F. Michelson , Robert L. Byer , Sasha Buchman , Ilya Mandel , John Lipa , Shally Saraf This is my paper · ORCID

classification astro-ph.IMphysics.space-ph

keywords gravitationalwavesmid-frequencybandgeosynchronousorbitspaceinterferometrytime-delaydrag-freesensorbinaryblackholesLISAcomplement

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper argues that a space-based gravitational-wave observatory in geosynchronous orbit, with three spacecraft forming an equilateral triangle of about 73,000 km per side, would reach its best strain sensitivity close to that of LISA but at frequencies roughly a hundred times higher, from about 10 mHz to 1 Hz. That band sits between ground-based detectors like LIGO and Virgo and the millihertz LISA mission, so an MFB observatory would catch binary black holes and neutron-star binaries before they enter the LIGO frequency range, and would observe extreme-mass-ratio inspirals and massive black hole mergers in its main band. The authors claim such a mission could be developed in seven to ten years and flown in the 2020s at a cost between $500 million and $1 billion by using LISA-derived laser interferometry, off-the-shelf satellites, and a simpler single-sphere drag-free sensor. The case matters because the mid-frequency band is where medium-mass binary black holes and many electromagnetic counterparts would be seen long before coalescence, enabling sky localization to arcminutes and multi-messenger follow-up.

What carries the argument

The load-bearing object is the MFB constellation itself: three satellites in geosynchronous orbit forming an equilateral triangle with 73,000 km arms, exchanging laser beams and applying Time-Delay Interferometry (TDI), a data combination method that cancels laser frequency noise by comparing heterodyne measurements along the arms with appropriate time delays. The sensitivity curve is carried by two noise assumptions: residual acceleration noise of $3.0\times10^{-15}$ m s$^{-2}$ Hz$^{-1/2}$ (low-frequency noise) and residual position noise of $0.5$ pm Hz$^{-1/2}$ (high-frequency noise). The enabling hardware is the Modular Gravitational Reference Sensor (MGRS), a single spherical test mass about 7 cm in diameter, spun at about 10 Hz, read out optically to picometer level, with gaps an order of magnitude larger than LISA's cubes; its prototype has a requirement of $10^{-14}$ m s$^{-2}$ Hz$^{-1/2}$ and a goal of $3\times10^{-15}$ m s$^{-2}$ Hz$^{-1/2}$. TDI and the sensor together convert the short geosynchronous baseline into a detector whose frequency band is set roughly by arm length: shorter arms push the sensitive band upward by about a factor of 100 relative to LISA.

What would settle it

Test the full-scale MGRS prototype on the ground: if its measured acceleration noise floor cannot be brought down from the $10^{-14}$ requirement to the $3\times10^{-15}$ goal across 0.1 mHz–1 Hz, the MFB sensitivity curve would not reach LISA-close levels in its lower band, and the central scientific-rate claims would weaken accordingly.

Watch

Extended reading notes

Core claim

MFB's central claim is that a constellation of three spacecraft in geosynchronous orbit, separated by 73,000 km arms, can achieve a characteristic strain sensitivity close to LISA's over the frequency band 10 mHz–1 Hz, a factor of about 100 higher in frequency, by combining LISA-style time-delay interferometry with a simplified gravitational reference sensor. The resulting sensitivity curve, averaged over sky positions and polarizations, is computed from the TDI A, E, T combinations assuming residual acceleration noise of $3.0\times10^{-15}$ m s$^{-2}$ Hz$^{-1/2}$ and residual position noise of $0.5$ pm Hz$^{-1/2}$ per spacecraft. On this basis the authors project event rates of $10^3$–$10^6$ per year for stellar-mass binary black holes and neutron star binaries, parameter estimation that can track sources coherently from MFB into the LIGO band, and localization of binary neutron stars to a few arcminutes so their host galaxies can be identified to roughly 500 Mpc. The paper also argues the mission is implementable at $500M–$1B with a 2020s launch, making it a cheaper and earlier complement to LISA rather than a competitor.

Load-bearing premise

The sensitivity curve is only as good as an unproven hardware assumption: the spherical gravitational reference sensor must meet its goal of $3\times10^{-15}$ m s$^{-2}$ Hz$^{-1/2}$ acceleration noise, while the full-scale prototype is currently specified only to a $10^{-14}$ requirement, and the required $0.5$ pm Hz$^{-1/2}$ position noise must also hold on a 73,000 km arm.

Editorial extensions

If this is right

Sources are caught before merger: medium-mass binary black holes and binary neutron stars would be tracked coherently from the MFB band into the LIGO/Virgo band, improving parameter estimation and enabling electromagnetic follow-up of neutron-star mergers.
Binary neutron star signals stay quasi-constant for years in the MFB, so the observatory can use the 2 AU diameter of Earth's solar orbit as a baseline and localize sources to a few arcminutes, identifying host galaxies out to roughly 500 Mpc.
With roughly 100 neutron-star mergers, the gravitational-wave measurement of the Hubble constant would reach 5% or better over a much larger volume than current events allow.
Mergers of massive black holes in the $10^3$–$10^8$ solar mass range would be characterized with precision comparable to or better than LISA, because the shorter baseline gives higher harmonics that improve position determination.
The mission would also act as a technical pathfinder for LISA, retiring drag-free and interferometry risks on a shorter schedule and at under half the cost of the ESA LISA mission.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If MFB's technology-demonstration satellites fly before LISA, the same sensor and interferometry hardware would retire much of LISA's remaining technical risk; this follows from the paper's own pathfinder claim but is not developed there.
The single-sphere gravitational reference sensor, if it hits its goal, would be a generic precision drag-free reference; it could plausibly be reused by other precision space missions such as geodesy or fundamental-physics tests, which the paper does not discuss.
The frequency overlap with the predicted stochastic gravitational-wave background means MFB data could test inflationary and primordial-black-hole models even if no individual source is ever resolved; the paper lists these as possible detections but does not quantify the sensitivity to the background.
A null result on the predicted $10^3$–$10^6$ per year event rates would itself be informative, sharply constraining binary black hole and neutron star population models; this follows directly from the rate estimate the paper cites.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

This is a mission concept paper, not a new science result, and it should be read that way. The genuinely new contribution is the specific architecture: a geosynchronous 73,000 km laser interferometer using a single spherical gravitational reference sensor, with a concrete cost and schedule estimate. The paper is upfront about building on prior work, correctly citing Mandel et al., Graham et al., and the earlier gLISA proposal (ref 35). That citation honesty matters, and it's a point in the paper's favor.

What it does well: the orbit analysis is real and useful. The plots show inter-spacecraft velocities below 0.7 m/s, angle variations within ±3 arcmin over station-keeping intervals, and fractional arm-length changes below 5e-4. That's the kind of engineering feasibility evidence a concept study needs. The paper also states plainly that the critical technology is at roughly TRL 4 and lays out a small-satellite demonstration path. The cost table comes from a 2016 JPL A-Team study, and while it is not independently reproducible, it is at least an internal estimate rather than a hand-wave.

The soft spots are where the reader's stress test lands. The Figure 1 sensitivity curve is asserted, not derived. It assumes residual acceleration noise of 3e-15 m/s^2/sqrt(Hz) and position noise of 0.5 pm/sqrt(Hz). The position noise assumption is aggressive but plausible; the acceleration number is the problem. The MGRS section states that the full-scale prototype has a requirement of 1e-14 and only a goal of 3e-15. At 10 mHz, the low-frequency end of MFB's claimed band, meeting only the requirement worsens the strain sensitivity by roughly a factor of three, which directly widens the gap to LISA at exactly the frequency where the paper says MFB is best placed. No error budget, no scaling argument from the 1/3-scale prototype, and no analysis of Earth-tidal disturbances connect the demonstrated requirement to the assumed goal. That's load-bearing for the central claim.

The event rates of 10^3 to 10^6 per year are taken from published simulations without recomputation. For a concept paper, that is acceptable; the sources are standard. The cost estimates are similarly not independently auditable, but they are presented in enough detail to see where the numbers come from.

Who gets value from this paper: people working on the mid-band science case, mission designers comparing geocentric architectures, and anyone involved in decadal prioritization who needs a concrete counterpoint to LISA's schedule and cost. It deserves a serious referee because mission selection decisions are informed by exactly this kind of proposal, and the key technical risk, whether the MGRS can reach its goal, is specific and testable.

Recommendation: send it to peer review, but the reviewer should be asked to assess whether the MGRS goal should be reclassified as a requirement before the sensitivity curve is accepted. The paper's own arguments could survive that change, but the headline claim would need to be softened.

Referee Report

4 major / 5 minor

Summary. The paper proposes the Mid-Frequency-Band (MFB) gravitational-wave observatory, a three-spacecraft laser interferometer in geosynchronous orbit with an arm length near 73,000 km, targeting the frequency band 10 mHz to 1 Hz. The authors argue that MFB would bridge the gap between ground-based detectors and LISA, enabling pre-merger observations of stellar-mass binaries, EMRIs, IMBHs, and massive black hole mergers, with event rates of roughly 10^3 to 10^6 per year. The technical concept builds on LISA interferometry heritage and replaces the LISA cubic test masses with a single spherical gravitational reference sensor (MGRS). The paper claims a best strain sensitivity close to that of LISA but shifted to frequencies about a factor of 100 higher, at a cost of $500M-$1B and with a launch in the 2020s. The sensitivity curve in Figure 1 assumes a residual acceleration noise of 3e-15 m/s^2/rtHz and a position noise of 0.5 pm/rtHz, but these values are stated as assumptions rather than demonstrated performance.

Significance. If the assumed sensitivity is realized, MFB would occupy a scientifically valuable and currently unserved frequency band, enabling multi-band gravitational-wave observations and unique tests of black hole physics. The paper's strengths are its clear science motivation, the use of LISA technology heritage to argue for reduced cost and schedule, and the explicit, if briefly documented, cost estimate from a 2016 JPL Team-X study. The central quantitative claims, however, are not supported within the manuscript: the sensitivity curve lacks a derivation, the key acceleration-noise assumption is only a goal of the proposed MGRS rather than a demonstrated requirement, and the projected science yields are not computed from the MFB response. The concept is plausible and worth further development, but the paper as submitted is closer to a mission concept white paper than to a self-contained technical proposal.

major comments (4)

[Mission Performance / Figure 1] The sensitivity curve in Figure 1 is asserted rather than derived. The Mission Performance section specifies sqrt(S_a)=3.0e-15 m/s^2/rtHz and sqrt(S_L)=0.5 pm/rtHz as assumptions, but no formula relating these noise values to the characteristic strain h_c(f), no TDI response model, and no optical or phasemeter noise budget are provided. Using the stated parameters at 10 mHz with L=7.3e7 m, the position-noise strain is sqrt(S_L)/L about 6.9e-21 and the acceleration-noise strain is sqrt(S_a)/((2*pi*f)^2 L) about 1.0e-20, giving a total strain near 1.2e-20. Since the paper's headline claim begins at 10 mHz, the figure and its underlying noise model must be shown explicitly.
[MGRS / Mission Performance] The central sensitivity assumption sqrt(S_a)=3e-15 depends on meeting the MGRS performance goal, not its demonstrated requirement. The MGRS section states that the full-scale prototype has an acceleration-noise requirement of 1e-14 m/s^2/rtHz and only a goal of 3e-15 at frequencies between 0.1 mHz and 1 Hz. If only the requirement is met, the acceleration contribution at 10 mHz rises by about a factor of 3.3, worsening the total strain at the low-frequency edge of MFB's band by roughly a factor of three. The paper provides no error budget, no scaling analysis from the 1/3-scale prototype cited in ref. 78, and no analysis of geosynchronous-orbit disturbances to show that the goal is reachable. This is load-bearing because the 'close to LISA' claim at the band edge depends directly on meeting the goal.
[Key Science Goals and Objectives] The claim that signals requiring five years of integration for LISA are observable in a few months by MFB is not supported by any signal-to-noise calculation in the paper. Similarly, the event rates of 10^3-10^6 per year are quoted from external references without being mapped onto the MFB sensitivity curve. Because the science case rests on these numbers, the manuscript should include at least one representative SNR calculation for a canonical BBH or BNS source at the MFB design sensitivity, including the assumed source parameters and the TDI response.
[Cost Estimates / Schedule] The $500M-$1B cost claim rests on a 2016 JPL Team-X/A-TEAM study that is not documented in sufficient detail for the reader to verify. Table 4 gives a top-level breakdown, but the text does not state whether the quoted total includes the two small-satellite technology demonstration missions described in the Schedule section, nor the cost of developing and qualifying the MGRS to the 3e-15 goal. The cost estimate should be reconciled with the proposed schedule and technology-development plan, or the paper should clearly identify what is and is not included.

minor comments (5)

[Table 3] The table header contains a typo, 'RANN', which should be 'RAAN'; the table caption is also incomplete and should define the acronyms and the meaning of the margins.
[Key Science Goals and Objectives] The phrase '10^3-106 per year' should read '10^3-10^6 per year'.
[Figure 1] The figure caption does not define the axes or the units of h_c(f); the in-text placeholder only shows 'MFB f(Hz) h_c(f)', and the actual plot is not reproduced in the manuscript, making it impossible to assess the claimed sensitivity curve.
[References] Several references are incomplete or non-standard, including refs. 2, 3, 46, 47, 53, and 65; these should be given full bibliographic entries.
[Trajectory] The statement that a geostationary orbit analysis 'fairly well represents' the geosynchronous case is not quantitatively justified; because the proposed orbit relies on an inclination-drift strategy, the authors should explain why the geostationary results carry over to the geosynchronous trajectory.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular derivation: MFB sensitivity is an explicit assumption, science rates come from independent literature, and MGRS self-citations support feasibility without defining the predicted curve.

full rationale

The derivation chain is not circular. The central sensitivity curve is set by explicit inputs, not recovered from outputs: the Mission Performance section states 'The MFB sensitivity shown in Figure 1 assumes a residual acceleration noise in each spacecraft equal to sqrt(Sa)=3.0e-15 m s-2 Hz-1/2 ... and a residual position noise sqrt(SL)=0.5 pm Hz-1/2'; these are named assumptions for the TDI-based noise projection (ref 41), and the paper does not fit them to any target detection. Science yields are taken from external, independent literature (e.g., Belczynski et al. event rates, Mandel/Sesana/Vecchio localization), so no fitted input is relabeled as a prediction. The MGRS technology claims do rest partly on the authors' own analytical model (ref 77) and a 1/3-scale prototype (ref 78), but this is a self-citation supporting feasibility, not a definitional reduction: the paper explicitly discloses that the full-scale prototype is 'under development' with a requirement of 1e-14 and only a goal of 3e-15, so the gap between assumed noise and demonstrated hardware is visible rather than hidden. The remaining self-citations are therefore a programmatic/feasibility risk, not circularity. No uniqueness theorem is imported from the same authors, and no known result is renamed as a new law.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The central MFB concept rests on two design-target noise values that are not demonstrated, a set of standard physics assumptions, event rates taken from external literature, a self-referential MGRS validation, and an internal cost study. No fundamentally new physical entity is introduced.

free parameters (2)

residual acceleration noise = 3.0e-15 m/s^2/sqrt(Hz)
Chosen as the mission requirement for the MFB sensitivity curve; it is the MGRS performance goal, not demonstrated in this paper.
residual position noise = 0.5 pm/sqrt(Hz)
Chosen as the mission requirement for high-frequency noise; not derived or demonstrated in the paper, and cited as a design target.

assumptions (5)

standard math Standard gravitational wave interferometer response is given by TDI combinations as assumed in refs 35, 37-41.
The paper uses the standard TDI A, E, T combinations to define the sensitivity curves, but does not reproduce the derivation.
domain assumption LIGO-derived merger rates can be extrapolated to the MFB band to yield 10^3-10^6 events per year.
The event rate numbers are taken from Belczynski et al. (refs 11,12) and Sesana (ref 13) without recomputation or uncertainty bounds.
domain assumption The MGRS spherical test mass sensor can meet or exceed the LISA Pathfinder acceleration noise performance.
The full-scale MGRS prototype is under development with a requirement of 1e-14 and a goal of 3e-15 m/s^2/sqrt(Hz); the paper cites its own analytical model (ref 77) and a 1/3 prototype (ref 78).
domain assumption The geosynchronous orbit dynamics used in the sensitivity and pointing analysis are correct and representative.
The inter-spacecraft velocities, angle variations, and arm-length changes in Figures 4-6 are based on a prior geostationary orbit analysis (ref 37), which is assumed to represent the geosynchronous case.
ad hoc to paper The 2016 Team-X cost estimate is reliable and transferable to the MFB design.
The $500M-$1B cost claim rests on an internal JPL A-Team study (Table 4) that is not independently documented or audited in the paper.

how reviews work

0 comments

Cite this review

Pith. "Pith review of MFB: A Mid-Frequency-Band Space Gravitational Wave Observer for the 2020 Decade." pith.science (2026). https://pith.science/paper/NPRF7KAC

@misc{pith2026190802861,
  author       = {Pith},
  title        = {Pith review of: MFB: A Mid-Frequency-Band Space Gravitational Wave Observer for the 2020 Decade},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NPRF7KAC}},
  note         = {Machine review of arXiv:1908.02861}
}

read the original abstract

We make the case for the early development of a Mid-Frequency-Band (MFB) gravitational wave (GW) observatory in geosynchronous orbit (73,000 km arm), optimized for the frequency band 10 mHz to 1 Hz. MFB bridges the science acquisition frequencies between the ground observatories LIGO/VIRGO (4/3 km arm - as well as future planned ones 10/40 km arm), and the milli-hertz band of LISA (2.5 Gm arm)- with usable sensitivity extending to 10 Hz. We argue that this band will enable the timely development of this game-changing field of astrophysics, with observations of medium mass Binary Black Holes (BBH) and Binary Neutron Stars (BNS) sources prior to their mergers in the LIGO frequency range as well as Extreme Mass Ratio Inspirals (EMRI)s and mergers of supermassive BBH within the main detection band. MFB is better placed than LISA to access this exciting frequency region.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

2 extracted references · 1 canonical work pages

[1]

Optimal orbits for eLISA science

1 R. Weiss, Quarterly Progress Report of RLE, MIT 105, 54 (1972) 2 LIGO Caltech 3 VIRGO Pisa 4 O. Jennrich, et al. NGO assessment study report (Yellow Book) 〈hal-00730260〉(2012) 5 P.F. Michelson ASTRO 2020 Science White Paper (2019) 6 B. P. Abbott, et al. Astrophys. J. Lett. 848(2), L12 (2017) 7 C. M. Will, Living Rev. Relativity, 9 (2006) 8 I. Mandel, A....

arXiv 1972
[2]

A Geostationary Gravitational Wave Interferometer (GEOGRAWI)

Office, Washington, D.C., 689-690 (1984) 44 P. Bender et al. Pre-Phase A Report, MPQ233 (1998) 45 LISA: “Unveiling a hidden Universe", ESA publication # ESA/SRE (2011) 46 LIGO: Caltech 47 VIRGO: Pisa MFB: Mid-Frequency-Band Space Gravitational Wave Observer for the 2020 Decade 12 48 NASA Astrophysics Visionary Roadmap (2013) 49 NASA Science Plan (2014) 50...

work page Pith review arXiv 1984

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

Optimal orbits for eLISA science

1 R. Weiss, Quarterly Progress Report of RLE, MIT 105, 54 (1972) 2 LIGO Caltech 3 VIRGO Pisa 4 O. Jennrich, et al. NGO assessment study report (Yellow Book) 〈hal-00730260〉(2012) 5 P.F. Michelson ASTRO 2020 Science White Paper (2019) 6 B. P. Abbott, et al. Astrophys. J. Lett. 848(2), L12 (2017) 7 C. M. Will, Living Rev. Relativity, 9 (2006) 8 I. Mandel, A....

arXiv 1972

[2] [2]

A Geostationary Gravitational Wave Interferometer (GEOGRAWI)

Office, Washington, D.C., 689-690 (1984) 44 P. Bender et al. Pre-Phase A Report, MPQ233 (1998) 45 LISA: “Unveiling a hidden Universe", ESA publication # ESA/SRE (2011) 46 LIGO: Caltech 47 VIRGO: Pisa MFB: Mid-Frequency-Band Space Gravitational Wave Observer for the 2020 Decade 12 48 NASA Astrophysics Visionary Roadmap (2013) 49 NASA Science Plan (2014) 50...

work page Pith review arXiv 1984