DICER: Diffractive Interfero Coronagraph Exoplanet Resolver

• DICER is a mid-IR space observatory concept that uses diffractive interferometry and nulling coronagraphy to resolve exoplanets with high angular resolution.
• It employs two primary objective gratings and a dispersion leverage coronagraph to mitigate on-axis starlight while preserving off-axis companion signals.
• The design targets habitable-zone exoplanets near 10 µm by breaking single-aperture resolution limits and enhancing spectral contrast.

The Diffractive Interfero Coronagraph Exoplanet Resolver (DICER) is a notional 20 m class infrared space observatory concept that uses two very large primary objective gratings (POGs) to achieve enhanced one-dimensional angular resolution at mid-infrared wavelengths, especially near $10~\mu\mathrm{m}$, for the discovery and characterization of nearby, habitable exoplanets. Its enabling coronagraph, the Dispersion Leverage Coronagraph (DLC), is a variation of Achromatic Interfero Coronagraphy (AIC) that uses a field-flipping periscope to null an on-axis source across the focal plane while preserving off-axis companions through spectral non-overlap. In this formulation, DICER sits at the intersection of exoplanet interferometry, nulling coronagraphy, and dispersive primary optics, and operationalizes the broader interferometric program advocated for exoplanet direct detection beyond the single-aperture limit of “a few $\lambda/D$” (Swordy et al., 15 Jul 2025, Monnier et al., 2019).

1. Scientific motivation and observational regime

The scientific motivation for DICER follows from a central limitation in direct exoplanet imaging: diffraction fundamentally limits single-aperture coronagraphs to inner working angles of “a few $\lambda/D$.” In the interferometric white paper literature, this limitation is identified as a major obstacle for crucial topics such as demographics of exoplanets within the first $50$ Myr and the infrared characterization of terrestrial planets. Practical examples given for single apertures are a “typical $0.2''$ inner working angle” for today’s $8$–$10$ m telescopes and “$\sim 0.07''$” for ELTs, whereas “interferometers (inner working angles $<0.01''$) will explore the full region of interest from $0.1$ au to $\lambda/D$0 au” (Monnier et al., 2019).

DICER is explicitly formulated for the nearby habitable-planet problem in the thermal infrared. The benchmark use case is the discovery and characterization of planets within $\lambda/D$1 pc, including measurements of small spectral segments such as the strong ozone absorption band near $\lambda/D$2. In the broader exoplanet-interferometry context, this aligns with the argument that infrared nulling from space offers unrivaled potential to detect terrestrial planets and search for atmospheric biomarkers, while long baselines decouple angular resolution from telescope diameter and permit much smaller effective working angles than conventional coronagraphs (Swordy et al., 15 Jul 2025, Monnier et al., 2019).

A plausible implication is that DICER should be understood not as a scaled-up monolithic coronagraph, but as a deliberately anisotropic observatory optimized for the thermal-IR regime where Earth–Sun contrast is substantially less severe than in reflected light. The data block frames this contrast regime as $\lambda/D$3 to $\lambda/D$4 at $\lambda/D$5, in contrast to the $\lambda/D$6 reflected-light requirement around $\lambda/D$7 AU for a Sun analog. Within that logic, DICER concentrates on the part of parameter space where nulling interferometry, dispersion, and high-resolution narrowband spectroscopy are mutually reinforcing.

2. Optical architecture: primary objective gratings and the DLC

DICER uses two POGs that are long in the dispersion direction, with benchmark dimensions $\lambda/D$8 m and width $\lambda/D$9 m, observed at grazing exodus with a secondary telescope. The benchmark grating pitch is $\lambda/D$0, and the exodus angle is $\lambda/D$1. Each POG feeds a secondary telescope of approximately $\lambda/D$2 m $\lambda/D$3 $\lambda/D$4 m aperture; square pupils are natural to the rectangular grating footprint. The benchmark secondary optics use focal length $\lambda/D$5 m, back focal length $\lambda/D$6 m, and primary-secondary separation $\lambda/D$7 m. Mid-IR operation is assumed near $\lambda/D$8 at L2 with cryogenic optics and end-to-end throughput of approximately $\lambda/D$9 (Swordy et al., 15 Jul 2025).

The DLC is the beam-combination scheme that makes this architecture useful for exoplanet suppression. It is described as a variation of AIC that uses a field-flipping periscope to impose an achromatic pupil inversion and a $50$0 phase flip in one arm before interferometric recombination. A beam splitter sends half of the light to destructive interference, the dark port used for coronagraphy, and half to constructive interference, effectively a light dump. Because the two gratings are pointed in opposite directions but view the same star, the periscope inversion aligns the dispersed stellar spectra for all wavelengths across the focal plane. In the ideal case, this produces broadband nulling of an on-axis star everywhere in the focal plane, expressed as

$50$1

This is the defining distinction between the DLC and ordinary focal-plane nullers built for non-dispersive pupils (Swordy et al., 15 Jul 2025).

The grating-based angular leverage is equally central. In the dispersed axis, the effective on-sky resolution is set by the grating length,

$50$2

while in the cross-dispersed axis it is set by the secondary aperture width,

$50$3

The white-paper synthesis motivating DICER had already argued for baselines and interferometric configurations that break the single-aperture angular-resolution ceiling; DICER’s distinctive move is to realize that leverage through primary gratings and a dispersion-aware nulling coronagraph rather than through a conventional long-baseline imaging interferometer alone. The 2019 white paper does not explicitly reference diffractive primary elements; this suggests that DICER is a concrete architectural instantiation of its broader interferometric program rather than a direct restatement of that paper’s instrument design (Monnier et al., 2019).

3. Nulling formalism, transmission maps, and leakage terms

For a single square aperture of side $50$4 and focal length $50$5, the DLC paper writes the image-plane field as

$50$6

with

$50$7

and intensity

$50$8

After combining the two beams with a $50$9 phase shift in the second beam and a pupil inversion in $0.2''$0, the total intensity is

$0.2''$1

When the PSFs fully overlap, the dark-port intensity vanishes everywhere in the focal plane for the on-axis star (Swordy et al., 15 Jul 2025).

The corresponding integrated focal-plane transfer function is

$0.2''$2

Using the grating geometry, the focal-plane offsets are

$0.2''$3

The dispersed-axis PSF separation is then

$0.2''$4

and the single-wavelength dark-port transmission map becomes

$0.2''$5

Under small $0.2''$6, the paper states that $0.2''$7 scaled by the grating dispersive mapping, so the single-wavelength PSF separation exceeds the true star–planet separation by roughly a factor $0.2''$8, helping preservation of the companion at small $0.2''$9 (Swordy et al., 15 Jul 2025).

The null is not unlimited. Three leakage terms are treated explicitly. First, finite stellar angular size produces stellar-disk leakage,

$8$0

For a Sun-like star at $8$1 pc with $8$2 m, $8$3, and $8$4, the estimate given is $8$5, limiting suppression to $8$6. Second, residual optical path difference $8$7 gives

$8$8

Targeting $8$9 at $10$0 requires $10$1, and $10$2 requires $10$3. Third, random pointing jitter in the dispersed and cross-dispersed axes yields

$10$4

This produces asymmetrical fine-guidance tolerances, tighter in $10$5 than in $10$6. A common misconception is that broadband nulling implies insensitivity to spacecraft stability; the DLC analysis instead shows that broadband nulling and stringent OPD/jitter control are simultaneous requirements, not substitutes (Swordy et al., 15 Jul 2025).

4. Benchmark design, background suppression, and sensitivity model

The benchmark DICER implementation is narrowband by design. The central wavelength is $10$7 and the bandpass is $10$8 nm, with secondary field of view $10$9 to capture that band. The paper gives a bandwidth-invariance relation for fixed target resolution $\sim 0.07''$0,

$\sim 0.07''$1

independent of $\sim 0.07''$2, $\sim 0.07''$3, or $\sim 0.07''$4. The single-wavelength PSF size at $\sim 0.07''$5 is quoted as $\sim 0.07''$6 in dispersion by $\sim 0.07''$7 in cross-dispersion for $\sim 0.07''$8 m, and the coronagraphic null region is approximately $\sim 0.07''$9 by $<0.01''$0. Planets are resolvable when the separation exceeds half the single-wavelength PSF size in each axis: at least $<0.01''$1 along $<0.01''$2 or at least $<0.01''$3 along $<0.01''$4 (Swordy et al., 15 Jul 2025).

A second disperser is required because the focal-plane coordinate $<0.01''$5 mixes sky angle and wavelength. Without it, diffuse zodiacal light from a very long extent along the grating’s dispersion direction overlays the planet signal. The paper states that, even with the narrow $<0.01''$6 nm filter around $<0.01''$7, background from roughly $<0.01''$8 is integrated when summing the planet signal if no second disperser is used. The adopted total local zodiacal plus exozodiacal brightness is $<0.01''$9. The proposed remedy is five $0.1$0 ZnSe immersion gratings, reducing the effective background collection to approximately $0.1$1 per companion spectrum and improving background-limited signal-to-noise by approximately $0.1$2 as a qualitative estimate (Swordy et al., 15 Jul 2025).

The example photon budget is correspondingly severe. For an Earth analog at $0.1$3 pc modeled as a $0.1$4 K blackbody with Earth radius, the estimated planet signal is approximately $0.1$5 photons per hour. The zodiacal plus exozodiacal background is approximately $0.1$6 photons per hour, and stellar leakage is approximately $0.1$7 photons per hour for a G star at the same distance. In the background-limited approximation with negligible dark counts and read noise,

$0.1$8

giving

$0.1$9

The quoted consequence is that reaching $\lambda/D$00 requires approximately $\lambda/D$01 hours, or about $\lambda/D$02 days, if the planet remains in the high-transmission region. This makes the second disperser not an optional refinement but a structural component of the observatory’s sensitivity model (Swordy et al., 15 Jul 2025).

5. Mission concept, survey cadence, and atmospheric characterization

The mission simulation in the DLC paper targets $\lambda/D$03 main-sequence Sun-like stars within $\lambda/D$04 pc, restricted to F/G/early K systems that are single or have separations of at least $\lambda/D$05, with $\lambda/D$06. The planet-occurrence model is Bryson et al. (2021) implemented via P-pop, and $\lambda/D$07 realizations produce $\lambda/D$08 simulated habitable-zone planets, with an expectation of approximately $\lambda/D$09 habitable planets around the $\lambda/D$10 hosts. The discovery survey observes each star $\lambda/D$11 times for $\lambda/D$12 days exposure per epoch, totaling $\lambda/D$13 days per star and approximately $\lambda/D$14 years of exposure time for discovery; the duty cycle is approximately $\lambda/D$15, and the full mission duration is $\lambda/D$16 years (Swordy et al., 15 Jul 2025).

The resulting detection yields are reported cumulatively. After $\lambda/D$17 days per star, approximately $\lambda/D$18 habitable planets are detected at $\lambda/D$19; after $\lambda/D$20 days, approximately $\lambda/D$21; after $\lambda/D$22, approximately $\lambda/D$23; after $\lambda/D$24, approximately $\lambda/D$25; and after $\lambda/D$26, approximately $\lambda/D$27. The paper therefore concludes that the benchmark DICER design could plausibly find and characterize approximately $\lambda/D$28 nearby, habitable exoplanets around Sun-like stars in a seven year mission, corresponding to about $\lambda/D$29 of the habitable exoplanets within $\lambda/D$30 pc in the simulation (Swordy et al., 15 Jul 2025).

Atmospheric characterization is framed around the ozone band. Follow-up ozone spectroscopy requires approximately $\lambda/D$31 days per planet at three wavelengths, $\lambda/D$32, $\lambda/D$33, and $\lambda/D$34, for roughly $\lambda/D$35 years to characterize four objects. The gratings must tilt by approximately $\lambda/D$36 total, using $\lambda/D$37, $\lambda/D$38, and $\lambda/D$39 settings for those wavelengths at $\lambda/D$40. In the wider exoplanet-direct-detection context, this places DICER squarely in the space-based mid-IR nulling lineage identified as promising for rocky-planet detection and atmospheric biomarker searches, and complementary to visible-light mission concepts such as HabEx and LUVOIR rather than a substitute for them (Monnier et al., 2019).

Rotation of the observatory is operationally important because it modulates the planet signal as the source traces a circle in $\lambda/D$41-space through the transmission function $\lambda/D$42. This increases detectability and enables estimation of separation and position angle. A plausible implication is that DICER’s observing mode is inherently dynamic: discovery depends on time-variable transmission geometry and narrowband spectral sampling, not on a single static high-contrast image.

6. Technical risks, multiple-star extensions, and broader applicability

The principal technical risks identified for DICER concern grating fabrication, path-length control, pointing stability, thermal stability, and background management. Meter-scale grating segments must maintain highly regular line spacing and integer groove phasing over the full $\lambda/D$43 m scale. The tolerances for rejection at $\lambda/D$44 are OPD stability of roughly $\lambda/D$45–$\lambda/D$46 nm for $\lambda/D$47 to $\lambda/D$48 suppression and anisotropic pointing jitter requirements of approximately $\lambda/D$49 mas RMS in the dispersed axis and $\lambda/D$50 mas RMS in the cross-dispersed axis for $\lambda/D$51 nulling. JWST-like jitters of approximately $\lambda/D$52–$\lambda/D$53 mas yield $\lambda/D$54 nulling. Thermal gradients are also consequential: at $\lambda/D$55–$\lambda/D$56 K, Be coefficient of thermal expansion values of approximately $\lambda/D$57–$\lambda/D$58 imply that a $\lambda/D$59 K gradient would produce about $\lambda/D$60 length differences over JWST-like segments (Swordy et al., 15 Jul 2025).

The DICER synthesis also admits a natural extension to multiple-star systems through aliased wavefront control (AWC). In that method, a weak diffractive grid in a pupil plane replicates the off-axis companion’s speckle field into the deformable mirror’s controllable region, so that Electric Field Conjugation can suppress the aliased copy and thereby suppress the true super-Nyquist contamination. For a representative $\lambda/D$61 DM and an $\lambda/D$62 Cen-like separation of approximately $\lambda/D$63, the simulated performance improves median dark-hole contrast from approximately $\lambda/D$64 to approximately $\lambda/D$65 in the monochromatic no-aberration case, and from approximately $\lambda/D$66 to approximately $\lambda/D$67 in a $\lambda/D$68 band without aberrations. With $\lambda/D$69–$\lambda/D$70 nm RMS aberrations, the reported results are approximately $\lambda/D$71 monochromatic and approximately $\lambda/D$72 over a $\lambda/D$73 band (Thomas et al., 2014).

In a DICER context, AWC is described as directly compatible and synergistic: the interferometric coronagraph or nuller suppresses the on-axis target star, while diffractive aliasing plus DM control addresses a bright off-axis companion without requiring additional telescopes or a starshade. This suggests that DICER’s “diffractive” character is not limited to its primary gratings. Diffractive structures can also serve as control-enabling elements in the wavefront architecture, especially for recovering high-contrast discovery space in binary fields (Thomas et al., 2014).

Beyond the habitable-exoplanet survey use case, the DLC paper states that the method may be useful for any application requiring extremely high resolution, close-companion spectroscopy, including binary-star characterization, protoplanetary and AGN disks, spatially resolved ultra-high-resolution stellar spectroscopy, and general high-resolution spectroscopy with bright sources in narrow bands. The broad significance of DICER is therefore twofold: it is a specific exoplanet mission concept with quantified yield and control requirements, and it is also a prototype for a broader class of dispersive interfero-coronagraphic observatories in which nulling, spectral leverage, and anisotropic angular resolution are designed as a single optical system rather than as separable subsystems.