MINDS: MIRI Mid-Infrared Disk Survey

• MINDS is a JWST Guaranteed Time Observations program that inventories volatile chemistry and dust mineralogy in the inner regions of protoplanetary disks.
• It employs MIRI’s Medium Resolution Spectrometer and complementary NIR instruments to detect water, CO2, and organic molecules with high spectral resolution.
• The survey refines our understanding of disk structure and volatile transport by linking molecular abundance patterns to evolving planet formation processes.

Searching arXiv for recent MINDS papers and survey overviews. The MIRI mid-Infrared Disk Survey (MINDS) is a JWST Guaranteed Time Observations program that uses the Mid-Infrared Instrument, primarily the Medium Resolution Spectrometer (MRS), to inventory the volatile chemistry, dust mineralogy, and physical conditions of the warm inner few astronomical units of planet-forming disks. MINDS spans disks around Herbig Ae stars, T Tauri stars, very low-mass stars, brown dwarfs, and a small young debris-disk sample, and combines mid-infrared spectroscopy with selected NIRSpec, NIRCam, and MIRI coronagraph observations to connect inner-disk chemistry, gas dispersal, disk structure, and planet-formation signposts (Henning et al., 2024, Dishoeck et al., 2024).

1. Program definition, scope, and scientific rationale

MINDS is JWST GTO program 1282, led by PI Th. Henning with co-PI I. Kamp. The survey comprises 52 targets spanning $0.1$–$3\,M_\odot$, and its core aims are to investigate the chemical inventory in the terrestrial planet-forming zone across spectral type, follow gas evolution into the disk-dispersal stage, and study protoplanetary and debris-disk structure in the thermal mid-IR (Henning et al., 2024).

Its scientific framing is tightly connected to pebble accretion and the inward drift of volatile-rich solids across icelines. In the MINDS picture, radial transport, sublimation, dust trapping, and settling can reshape the inner-disk elemental budget, especially the volatile C/O and N/O ratios, and therefore alter the material available to terrestrial planets and the inner envelopes of giant planets. The survey was designed explicitly to test how these processes imprint themselves on mid-infrared molecular emission, dust features, and isotopologue diagnostics (Kamp et al., 2023).

The target selection was intentionally heterogeneous. It includes compact and extended disks, smooth and substructured disks, cavity and transition disks, very low-mass and brown-dwarf disks, and a few benchmark systems with known protoplanets or prominent outer-disk substructure. This breadth is central to the survey’s logic: MINDS is not only an inventory program, but also a comparative framework for linking inner-disk chemistry to stellar mass, disk architecture, accretion, irradiation, and transport.

2. Observational architecture and instrumental capabilities

The core MINDS dataset is obtained with JWST/MIRI-MRS, an integral-field spectrograph covering approximately $4.9$–$27.9\,\mu\mathrm{m}$ in four channels with three sub-bands each. The resolving power is wavelength-dependent, reaching $R \approx 3500$ at the shortest wavelengths and decreasing to $R \approx 1500$ near $28\,\mu\mathrm{m}$; around the $15\,\mu\mathrm{m}$ CO$_2$ bending mode, $R \approx 2500$ is typical (Henning et al., 2024). This regime contains the principal mid-IR bands of H$3\,M_\odot$0O, OH, CO, CO$3\,M_\odot$1, HCN, C$3\,M_\odot$2H$3\,M_\odot$3, CH$3\,M_\odot$4, larger hydrocarbons, H$3\,M_\odot$5 rotational lines, hydrogen recombination lines, and fine-structure tracers such as [Ne II], [Ne III], and [Ar II].

Compared with Spitzer/IRS, MIRI combines substantially higher sensitivity with medium spectral resolution sufficient to resolve Q-branch structure, separate adjacent P/R-branch lines, and identify isotopologues such as $3\,M_\odot$6CO$3\,M_\odot$7, CO$3\,M_\odot$8O, and $3\,M_\odot$9CCH$4.9$0. This is particularly important for molecules without permanent dipole moments, including CO$4.9$1, C$4.9$2H$4.9$3, CH$4.9$4, and benzene, whose mid-IR bands were only partially accessible at Spitzer resolution (Dishoeck et al., 2024).

MINDS also includes selected complementary observations. NIRSpec/IFU high-resolution spectroscopy in the $4.9$5–$4.9$6 range targets CO rovibrational bands and near-IR diagnostics of inner gas. MIRI coronagraphy is used for a few benchmark disks such as TW Hya and HD 169142. NIRCam imaging within MINDS provides high-angular-resolution context in systems such as PDS 70, where Pa-$4.9$7 and $4.9$8 imaging constrain protoplanets, circumplanetary environments, and disk substructure (Christiaens et al., 2024).

Operationally, MRS observations typically use four-point dithers, with target acquisition employed selectively. For compact sources, spectra are usually extracted with apertures scaled to the PSF FWHM, while extended sources and close multiples require customized extraction, PSF forward modeling, or dedicated background strategies. The IFU format is essential not only for point-source spectroscopy but also for mapping jets, winds, extended H$4.9$9, and spatially structured continuum emission.

3. Reduction pipelines and inference methodology

Across MINDS papers, data reduction commonly combines the JWST Calibration Pipeline with survey-specific routines from VIP for bad-pixel mitigation, image alignment, centroiding, and extraction. Residual fringing is treated at multiple stages, and continuum subtraction is performed with spline-based or Savitzky–Golay/PyBaselines procedures, depending on source morphology and spectral crowding. Representative workflows include sigma-filtering of bad pixels, Gaussian-kernel interpolation, sub-band alignment by image cross-correlation, aperture photometry with annular background subtraction, and tailored handling of Q-branch regions to avoid baseline oversubtraction (Vlasblom et al., 2024).

The default gas-analysis framework in MINDS is the $27.9\,\mu\mathrm{m}$0D LTE slab model. Each molecular component is parameterized by gas temperature $27.9\,\mu\mathrm{m}$1, column density $27.9\,\mu\mathrm{m}$2, and emitting area $27.9\,\mu\mathrm{m}$3, often reported through an equivalent radius $27.9\,\mu\mathrm{m}$4 with $27.9\,\mu\mathrm{m}$5. The emergent intensity is written as

$27.9\,\mu\mathrm{m}$6

with optical depth expressed as

$27.9\,\mu\mathrm{m}$7

or equivalently $27.9\,\mu\mathrm{m}$8. LTE level populations follow

$27.9\,\mu\mathrm{m}$9

In optically thin regimes, line fluxes scale directly with $R \approx 3500$0; in optically thick regimes, temperature and emitting area are usually better constrained than column density (Vlasblom et al., 2024).

These slab models are augmented by species-specific refinements. Mutual shielding within dense Q-branches is included for CO$R \approx 3500$1, HCN, and C$R \approx 3500$2H$R \approx 3500$3. H$R \approx 3500$4 analyses use rotational diagrams under optically thin LTE assumptions when appropriate. Atomic hydrogen recombination lines are compared to Case B and Kwan-Fischer local line-excitation models in order to estimate accretion luminosities and gas densities. For dust, MINDS employs tools such as DuCK and DuCKLiNG to fit silicate mineralogy, grain sizes, and pseudo-continuum structure. At the interpretation stage, these empirical retrievals are compared with thermo-chemical and radiative-transfer calculations using ProDiMo, DALI, MCMAX-3D, RADMC-3D, FLiTs, RADLite, RADEX, and CLIcK-like multi-component frameworks (Henning et al., 2024).

A recurring methodological theme is that isotopologues are not treated as secondary details, but as primary opacity diagnostics. Because the main $R \approx 3500$5CO$R \approx 3500$6 band is often optically thick, $R \approx 3500$7CO$R \approx 3500$8 and, in favorable cases, CO$R \approx 3500$9O provide a more reliable route to the total CO$R \approx 1500$0 column than raw $R \approx 1500$1CO$R \approx 1500$2 band strengths. This has become one of the distinctive technical signatures of the survey.

4. Molecular inventory and chemical diversity across the MINDS sample

MINDS established early that inner-disk mid-IR spectra are chemically diverse rather than convergent. Among T Tauri disks, some systems are strongly H$R \approx 1500$3O-rich, whereas others are CO$R \approx 1500$4-dominated or display only modest water despite signatures of inward volatile delivery. The GW Lup spectrum was a landmark result: it provided the first detection of gas-phase $R \approx 1500$5CO$R \approx 1500$6 in a protoplanetary disk, demonstrated that the $R \approx 1500$7CO$R \approx 1500$8 Q-branch is optically thick, and yielded $R \approx 1500$9, far above values inferred previously from optically thin Spitzer analyses (Grant et al., 2022).

CX Tau sharpened this picture by showing that a compact, drift-dominated disk need not be water-bright. In that source, JWST/MIRI detects H$28\,\mu\mathrm{m}$0O, $28\,\mu\mathrm{m}$1CO$28\,\mu\mathrm{m}$2, $28\,\mu\mathrm{m}$3CO$28\,\mu\mathrm{m}$4, C$28\,\mu\mathrm{m}$5H$28\,\mu\mathrm{m}$6, HCN, OH, and a potential CO$28\,\mu\mathrm{m}$7O feature. The $28\,\mu\mathrm{m}$8CO$28\,\mu\mathrm{m}$9 emission traces $15\,\mu\mathrm{m}$0 K gas, $15\,\mu\mathrm{m}$1CO$15\,\mu\mathrm{m}$2 traces a significantly colder $15\,\mu\mathrm{m}$3 K component, warm H$15\,\mu\mathrm{m}$4O traces $15\,\mu\mathrm{m}$5–$15\,\mu\mathrm{m}$6 K gas, and line ratios beyond $15\,\mu\mathrm{m}$7 indicate an additional $15\,\mu\mathrm{m}$8–$15\,\mu\mathrm{m}$9 K H$_2$0O reservoir. The combination of cold isotopologue emission and comparatively weak warm H$_2$1O was interpreted as evidence that inner-disk chemistry can pass through a CO$_2$2-rich drift phase, with accretion luminosity and dust opacity modulating the observed water brightness (Vlasblom et al., 2024).

At the opposite chemical extreme, Sz 98 is a water-rich T Tauri disk with CO, H$_2$3O, OH, CO$_2$4, and HCN, but no C$_2$5H$_2$6, CH$_2$7, or NH$_2$8. Its H$_2$9O emission spans a large radial surface zone, with shorter-wavelength lines tracing hotter, more compact gas and longer-wavelength rotational lines tracing cooler, larger emitting areas. The inferred composition points to a sub-solar inner-disk C/O ratio, in strong contrast to the outer disk, where ALMA data imply volatile gas C/O $R \approx 2500$0 (Gasman et al., 2023).

Very low-mass stars and brown dwarfs occupy yet another regime. In the 10-source VLMS/brown-dwarf sample, C$R \approx 2500$1H$R \approx 2500$2 and HCN are detected in all sources, C$R \approx 2500$3H$R \approx 2500$4 in $R \approx 2500$5, C$R \approx 2500$6H$R \approx 2500$7 in $R \approx 2500$8, CO$R \approx 2500$9 in $3\,M_\odot$00, H$3\,M_\odot$01O in $3\,M_\odot$02, and CO in $3\,M_\odot$03. Detection rates of organic molecules correlate with other organic molecules and anti-correlate with inorganic molecules, while hydrocarbon-rich sources show weaker $3\,M_\odot$04 dust strength and lower outer-disk dust mass than oxygen-rich sources (Arabhavi et al., 3 Jun 2025). Sz 28 is exemplary: JWST reveals CH$3\,M_\odot$05, CH$3\,M_\odot$06, C$3\,M_\odot$07H$3\,M_\odot$08, $3\,M_\odot$09CCH$3\,M_\odot$10, C$3\,M_\odot$11H$3\,M_\odot$12, C$3\,M_\odot$13H$3\,M_\odot$14, C$3\,M_\odot$15H$3\,M_\odot$16, C$3\,M_\odot$17H$3\,M_\odot$18, CO$3\,M_\odot$19, $3\,M_\odot$20CO$3\,M_\odot$21, HCN, and HC$3\,M_\odot$22N, while H$3\,M_\odot$23O and OH are absent, consistent with a gas-phase C/O $3\,M_\odot$24 (Kanwar et al., 2024).

Brown-dwarf disks extend this chemical reach further. Cha H$3\,M_\odot$25 1 shows strong silicate emission plus a dense hydrocarbon inventory including C$3\,M_\odot$26H$3\,M_\odot$27, $3\,M_\odot$28CCH$3\,M_\odot$29, CH$3\,M_\odot$30, CH$3\,M_\odot$31, C$3\,M_\odot$32H$3\,M_\odot$33, C$3\,M_\odot$34H$3\,M_\odot$35, C$3\,M_\odot$36H$3\,M_\odot$37, C$3\,M_\odot$38H$3\,M_\odot$39, and C$3\,M_\odot$40H$3\,M_\odot$41, together with CO$3\,M_\odot$42, $3\,M_\odot$43CO$3\,M_\odot$44, HCN, H$3\,M_\odot$45, and H$3\,M_\odot$46O. The study describes it as the most chemically diverse brown-dwarf disk observed to date in the mid-IR (Morales-Calderón et al., 7 Aug 2025). A plausible implication is that MINDS has moved VLMS/BD disks from being a chemically marginal population to a central testbed for high-C/O inner-disk chemistry.

5. Disk structure, volatile transport, winds, and planet-forming environments

A central MINDS result is that inner-disk chemistry is inseparable from disk structure. Compact dust disks were initially hypothesized to show systematically enhanced cold H$3\,M_\odot$47O because strong radial drift should deliver icy pebbles efficiently into the inner disk. A dedicated MINDS analysis of eight compact disks showed that the outcome is more diverse. The sample separates into three observational classes: Type N (“Normal”) disks with no strong cold-water enhancement, Type E disks with enhanced cold H$3\,M_\odot$48O, and Type P H$3\,M_\odot$49O-poor disks. Only two of the eight objects display the strong cold-water enhancement expected from the simplest drift scenario, implying that compactness alone is not a sufficient predictor of inner-disk water brightness (Temmink et al., 21 May 2025).

CX Tau is the archetypal drift-dominated case in structural terms. ALMA measures a $3\,M_\odot$50 dust radius of $3\,M_\odot$51 au, whereas $3\,M_\odot$52CO gas extends to $3\,M_\odot$53 au, giving $3\,M_\odot$54. In MINDS, that same source shows cold $3\,M_\odot$55CO$3\,M_\odot$56 and cold H$3\,M_\odot$57O components near the snow-line temperature regime, supporting a picture in which inward-drifting pebbles feed volatile-rich gas into the inner few au, but with the relative visibility of H$3\,M_\odot$58O and CO$3\,M_\odot$59 modulated by accretion luminosity, dust pile-up, and time-dependent drift phase (Vlasblom et al., 2024).

The survey also revised the interpretation of dust traps and cavities. Transitional and gapped disks were often expected to be chemically dry inside their cavities, but MINDS has repeatedly detected inner-disk water and other molecules in such systems. PDS 70 and SY Cha both retain molecular gas in their inner regions, indicating that traps are “leaky” rather than absolute barriers to inward transport. In SY Cha, MIRI further showed that a large millimeter cavity can coexist with a dynamic, gas-rich inner disk and substantial continuum variability (Dishoeck et al., 2024, Schwarz et al., 2023).

Because MRS is an IFU, MINDS has also expanded from chemistry into resolved gas dynamics. In SY Cha, extended H$3\,M_\odot$60 emission is detected in the S(3)–S(7) lines, with semi-opening angles from $3\,M_\odot$61 to $3\,M_\odot$62, high excitation temperature, and low column density consistent with a molecular disk wind. The same dataset resolves a highly collimated [Ne II] jet; the blue lobe has a semi-opening angle of $3\,M_\odot$63. The inferred molecular wind mass-loss rate, $3\,M_\odot$64, exceeds the previously derived stellar accretion rate of $3\,M_\odot$65, highlighting either strong accretion variability or partially decoupled wind-driving mechanisms (Schwarz et al., 2024).

MINDS has also probed how multiplicity and embedded planets alter the inner disk. In three Class II binaries, primary disks are water-rich while secondaries are mostly line-poor at MIRI sensitivity, and all systems show [Ne II] jet emission, with several also showing extended H$3\,M_\odot$66 structures associated with outflows or dynamical interactions (Kurtovic et al., 4 Aug 2025). In PDS 70, NIRCam imaging within MINDS re-detected both protoplanets and revealed a spiral-like feature interpreted near planet c as an accretion stream feeding its circumplanetary disk, while the outer portion may reflect asymmetric illumination of the outer disk (Christiaens et al., 2024). This suggests that MINDS is not merely surveying quiescent chemistry, but observing chemistry within actively structured and dynamically evolving planet-forming environments.

6. Dust mineralogy, survey synthesis, and evolving directions

MINDS is equally a dust survey. In a 26-disk T Tauri subset, Mg-rich and Fe-poor silicates reproduce the MIRI spectra with residuals typically within $3\,M_\odot$67. Grain sizes are skewed toward sizes larger than $3\,M_\odot$68, the average dust composition is dominated by Mg$3\,M_\odot$69SiO$3\,M_\odot$70-stoichiometry grains at $3\,M_\odot$71, followed by MgSiO$3\,M_\odot$72 at $3\,M_\odot$73 and SiO$3\,M_\odot$74 at $3\,M_\odot$75, and crystalline mass fractions lie in the $3\,M_\odot$76–$3\,M_\odot$77 range with a mean of $3\,M_\odot$78. Annealed silica is robustly detected in nine objects, with cristobalite identified as the main polymorph (Varga et al., 9 Jun 2026).

These dust results are chemically coupled to the gas. Disks with strong annealed-silica features show stronger CO$3\,M_\odot$79 emission, whereas forsterite-rich disks show stronger H$3\,M_\odot$80O emission; silica-rich systems may also have elevated gas-phase C/O (Varga et al., 9 Jun 2026). In the VLMS/BD sample, an independent MINDS analysis classifies disks into less-settled, more-settled, and fully-settled geometries based on mid-IR spectral slopes and silicate-band strengths, and finds an overall trend of increasing gas column density with decreasing spectral slope. This suggests that as dust settling lowers mid-IR opacity, deeper and denser molecular layers become visible (Jang et al., 19 Sep 2025).

The survey’s broader significance is therefore synthetic rather than merely catalog-based. MINDS has shown that inner-disk chemistry depends simultaneously on stellar mass, disk structure, drift efficiency, accretion luminosity, irradiation, dust opacity, and time variability. It has also established isotopologue spectroscopy as a practical route to robust mid-IR abundance constraints, particularly for CO$3\,M_\odot$81. Current survey directions include routine use of $3\,M_\odot$82CO$3\,M_\odot$83 and CO$3\,M_\odot$84O to overcome optical-depth effects, multi-component and radial-gradient modeling of H$3\,M_\odot$85O and CO$3\,M_\odot$86, coordinated ALMA and high-resolution near-IR follow-up to test for small cavities and dust pile-ups, and time-domain monitoring of accretion and molecular-band variability (Vlasblom et al., 2024, Dishoeck et al., 2024).

Taken together, MINDS has redefined the observational study of the inner few au of protoplanetary disks. It established that mid-infrared disk spectroscopy with JWST can recover not just a chemical inventory, but a coupled picture of volatile delivery, dust evolution, wind launching, and planet-forming substructure. Its most consequential result may be methodological as much as empirical: the inner disk is now accessible as a chemically and physically stratified system rather than a blended unresolved continuum with a few superposed bands.