A SPHEREX PIPELINE AND SPECTRAL LIBRARY FOR ULTRACOOL DWARFS Journal: ApJ Jonathan Gagné, Jacqueline K. Faherty, Azul Ruiz Diaz, Louis-Philippe Coulombe, Thomas P. Bickle, Adam C. Schneider, J. Davy Kirkpatrick, Marc J. Kuchner, Aaron M. Meisner, Dan Caselden, Adam J. Burgasser, Sarah Casewell, Easton J. Honaker, Frank Kiwy, Federico Marocco, Nikolaj Stevnbak Andersen, Lizzeth Ruiz Arroyo, Bruce Baller, Paul Beaulieu, John Bell, Martin Bilsing, Troy K. Bohling, Guillaume Colin, Giovanni Colombo, Sam Deen, Alexandru Dereveanco, Kevin Dixon, Hugo A. Durantini Luca, Deiby Flores, Christoph Franck, Christopher Fulvi, Michael Gallmann, Jean Marc Gantier, Konstantin Glebov, Léopold Gramaize, Leslie K. Hamlet, Ken Hinckley, Kevin Jablonski, Peter A. Jałowiczor, Martin Kabatnik, Peter Kasprowitz, K Ly, David W. Martin, Naoufel Marzak, Alexander McColgan, Neil J. McEwan, Marianne N. Michaels, William Pendrill, Stéphane Perlin, Ben Pumphrey, James Rabe, Henry Raway, Walter Ruben Robledo, David Roser, Animesh Roy, Arttu Sainio, Vincent Schindler, Manfred Schonau, Jörg Schümann, Karl Selg-Mann, Andrea Serio, Patrick Smith, Andres Stenner, Christopher Tanner, Melina Thévenot, Vinod Thakur, Mayahuel Torres Guerrero, Maurizio Ventura, Nikita V. Voloshin, Jim Walla, Zbigniew Wȩdracki, Bailey Weyandt, Breck Wilhite, Spartacus Zitouni Abstract We present a Python spectrophotometry extraction tool tailored for fast-moving point sources detected in the SPHEREx mission, and use it to construct a set of 0.75–5.0 μ \mu m low-resolution ( λ / Δ ​ λ ≈ 50 \lambda/\Delta\lambda\approx 50 ) spectrophotometry data products based on the SPHEREx Quick Release 2 (QR2) for a set of 6003 L0–Y1 ultracool dwarfs: 2050 known ultracool dwarfs, 3008 known photometric ultracool dwarf candidates, and 947 newly identified ultracool dwarfs. This work more than doubles the number of ultracool dwarfs with spectroscopy, from 3449 to 7402. We provide SPHEREx templates for each spectral subtype and a set of tools to assign automated spectral types. The QR2 data release generates spectrophotometry with an average signal-to-noise per spectral channel above ≈ \approx 10 for most objects with WISE W2 magnitudes of 14.0 mag and brighter. The compiled data set is made available publicly at https://mocadb.ca , where new spectral compilations from future data releases will also be made available as they are published. These new data provide a significant increase in the number of substellar objects for which the 2.4–5.0 μ \mu m window is now accessible, making it possible to probe important molecular chemistry of key CNOS-bearing species for the coolest brown dwarfs. We flag 2668 ultracool dwarfs as candidate young brown dwarfs, 250 as candidate subdwarfs, and 865 as possibly otherwise peculiar for future investigation. The SPIFF library presented here opens the doors to efficient confirmation of candidate substellar objects and follow-up studies of population-level atmospheric properties of cold brown dwarfs. I INTRODUCTION The recent advent of the Spectro-Photometer for the History of the Universe, Epoch of Reionization and Ices Explorer (SPHEREx) mission ( J. J. Bock et al. 2026 ; B. P. Crill et al. 2020 ; O. Doré et al. 2014 ; P. M. Korngut et al. 2018 ) provides unique benefits to the study of ultracool stars and brown dwarfs, even though this was not part of its core mission. SPHEREx launched into low-Earth orbit on 2025, March 11, and began an all sky scan obtaining images with a 3.5 \deg angular size (consisting of 2048 \times 2048 pixels with angular sizes \approx 6\farcs 15 ). The imaging mission began in May of 2025 using a set of 102 narrow spectral channels spanning \approx 0.75-5.0 \mu m, with a nominal mission duration of 25 months. One of the core advantages of SPHEREx for the study of brown dwarfs is the possibility of reconstructing spectrophotometry with properties akin to a low-resolution \lambda/\Delta\lambda\approx 35–125 spectrum ( H. Hui et al. 2024 ; H. Hui et al. 2026 ) , with a spectral coverage that includes the 2.4–5.0 \mu m segment, which is difficult to observe from the ground. This wavelength range is particularly important to measure the bolometric luminosities of the coolest brown dwarfs ( J. C. Filippazzo et al. 2015 ) , and to constrain the atmospheric chemistry of cold, T and Y dwarfs ( M. C. Cushing et al. 2006 ; C. Visscher et al. 2006 ) , because it contains key molecular bands of CO 2 ( I. Yamamura et al. 2010 ; S. Sorahana & I. Yamamura 2012 ; J. K. Faherty et al. 2024 ) and potentially PH 3 and SiH 4 ( C. Visscher et al. 2006 ; A. J. Burgasser et al. 2025 ; J. K. Faherty et al. 2025 ) that are hard or impossible to constrain at < 2.4 \mu m alone. The full SPHEREx data also probes molecules such as CO, CH 4 and H 2 O over a wider range of atmospheric pressures. Furthermore, the faint Y dwarfs emit a significant portion of their energy past 2.4 \mu m ( M. C. Cushing et al. 2011 ; S. K. Leggett et al. 2010 ; S. K. Leggett et al. 2013 ) . However, the data sets produced by the SPHEREx mission are not immediately amenable to this goal of large-scale brown dwarf spectroscopic characterization, because they consist of individual image files that are expensive to download, store, and analyze in order to reconstruct the reduced ‘spectra’ in question. To address this challenge, we present the publicly available ‘SPHEREx Photometry and Image Fitting Framework’ (SPIFF) Python library to reconstruct the spectrophotometry of fast-moving point sources (Section II ), and use it to build a spectral library of known ultracool dwarfs (Section III ). We construct hybrid SPHEREx spectral templates across the 1.0–5.0 \mu m range along with an automated spectral typing scheme, and use it to confirm the substellar nature of 3953 candidate ultracool dwarfs (Section IV ). We conclude in Section V with possible areas of research that are unlocked by this new spectral library. II The SPIFF tool (a) PSF reconstruction (b) scipy.optimize PSF fitting (c) Median of ultranest samples Figure 1: Diagnostic figures generated by the SPIFF pipeline. Upper row: PSF reconstruction from the provided over-sampled PSF (left column), placed on the detector grid using a box sum (second column), and mapped onto the observed image (third column). A comparison with an observed PSF for SIMP J013656.5+093347.3 is shown on the right column. Middle row: PSF fitting using scipy.optimize . Bottom row: Final PSF fitting parameters obtained from the median samples of ultranest . See Section II for more details. The ‘SPHEREx Photometry and Image Fitting Framework’ (SPIFF) Python tool is analogous to the spectrophotometry tool ( R. Akeson et al. 2025 ) provided by the NASA IPAC InfraRed Science Archive (IRSA) 1 1 1 Available at https://irsa.ipac.caltech.edu/applications/spherex/tool-spectrophotometry . , with a few different characteristics. The IRSA tool takes as input several coordinates on which to perform forced photometry using the SPHEREx instrumental point spread function (PSF) appropriate for the detector position for a given visit, as provided in the SPHEREx calibration files. Background subtraction is performed locally by subtracting the median of the flux in a 15-pixel-wide box, after masking bad pixels and known stars. The spectrophotometry of the target sources can then be extracted, accounting for crowding effects. The IRSA spectrophotometry tool requires manual input of the target coordinates, and does not allow for large batch queries, nor does it account for the proper motion of objects that move significantly between SPHEREx epochs. The latter constraint can be problematic for fast-proper motion objects, especially given that the tool performs forced photometry without allowing the exact PSF position to move during the fitting procedure. This makes it particularly important to provide precise coordinates for proper results. In order to extract the spectrophotometry of a large number of fast-moving substellar objects efficiently, and to do so routinely as the SPHEREx data get released in increments, we built a custom Python SPHEREx spectrophotometry tool, the ‘SPHEREx Photometry and Image Fitting Framework’ (SPIFF) 2 2 2 Available at https://github.com/jgagneastro/spherex_spiff_pipeline . . This tool works similarly to the IRSA spectrophotometry tool, with the following notable differences: • The proper motion of the target can be provided such that the initial detector position of the PSF is adjusted using the epoch of the individual SPHEREx image 3 3 3 We do not attempt to model parallax motion because it is negligible compared to a SPHEREx pixel even for the nearest brown dwarf. ; • only a single PSF can be provided; • the detector position of the PSF is allowed to vary during the PSF fitting step; • the PSF fitting step is performed with scipy.optimize , and the Bayesian likelihood corresponding to measurement uncertainties is explored with the ultranest nested sampling package. Examples of PSF fitting steps are provided in Figure 1 . Similar to the IRSA spectrophotometry tool, SPIFF uses the instrumental SPHEREx PSF provided in the calibration files, which are oversampled with respect to detector pixels, and performs a lateral shift using linear interpolation before downgrading the spatial resolution of the PSF onto the spatial resolution of the observed image using a box sum (after shifting the PSF). An example of this step is provided in Figure 1(a) . The SPIFF pipeline avoids downloading the full 2048 \times 2048-pixels FITS image or each data point to improve efficiency. Instead, it first uses the IPAC SIA2 service to identify the individual URLs of FITS images in a 5″ cone search around the target position projected to epoch 2025.5. For each individual FITS image, the SPIFF pipeline then requests a small 20 \times 20-pixels cutout of the FITS image stored on Amazon S3, around the coordinates of the target projected to the epoch of the specific FITS image observation. This substantially reduces the network transfers required for analysis, but downloading the headers of the FITS images can still be the bottleneck in terms of total computing time required to construct a SPIFF data product, depending on internet bandwidth. In order to run SPIFF efficiently on large samples, we utilized the Digital Research Alliance resources 4 4 4 alliancecan.ca , which allow for both fast computing and high internet bandwidth. The data products of the SPIFF pipeline consist of quality flags, extracted central wavelength, approximate wavelength bin width (obtained from the SPHEREx calibration files at the best-fitting PSF detector position), as well as the flux and measurement error provided in three flavors: • using aperture photometry with a set of 18 aperture radii (1.0-5.0 pixels, 6\farcs 5 – 30\farcs 8 ), preserving the measurement that maximizes S/N, • based on the scipy optimization of the PSF fitting, • and based on the median of the ultranest samples (for the measurement) and the median absolute deviation (for the measurement error) when exploring the PSF fitting likelihood. The full columns of the SPIFF outputs are described in Table 1 . In practice, the most reliable flux measurements are usually obtained from the median ultranest samples, because it tends to explore the likelihood more thoroughly than scipy.optimize , and it is much less sensitive to background contaminants compared with aperture photometry, especially given the large 6\farcs 15 pixel size of the SPHEREx detectors. In this work, we present a set of 6003 high-quality spectrophotometry products compiled with SPIFF, and stored in the Montreal Open Clusters and Associations (MOCA) database 5 5 5 Available at https://mocadb.ca . A persistent and versioned copy of the SPIFF spectrophotometric library is also available on Zenodo at doi:10.5281/zenodo.19051216 . ( J. Gagné 2024 ; J. Gagné et al. 2026 ) . They are stored as raw SPIFF outputs in the table data_spherex_spectra_spiff . However, SPHEREx revisits each sky location multiple times, and every time the target falls on a slightly different region of the detector, resulting in small wavelength shifts. Because every such measurement is actually a photometric measurement with a relatively wide 0.02–0.08 \mu m bandpass ( H. Hui et al. 2026 ) , the true spectral resolving power is not as high as these repeated measurements with small wavelength shifts may suggest. Instead, this results in an oversampling of the spectral resolution element. We also provide a simpler data product in the MOCAdb data_spectra table, in which we compile the error-weighted mean of all available flux measurements in each of the 102 SPHEREx spectral channels. The current version of SPIFF relies on preliminary response functions of each SPHEREx spectral channel, detailed in SPHEREx mission’s Quick Release (QR) explanatory supplements 6 6 6 Available at https://irsa.ipac.caltech.edu/data/SPHEREx/docs/SPHEREx_Expsupp_QR.pdf . , listed in Table 2 , although refined response functions will eventually be released ( H. Hui et al. 2026 ) and implemented in future versions of SPIFF. It is worth noting that each individual pixel of the SPHEREx detectors have their own specific response functions, which can vary across the detector even with a fixed spectral channel, meaning that the reconstructed spectrophotometry stored in data_spherex_spectra_spiff are spectrally over-sampled in a complex way, and the binned spectra provided in the data_spectra table are an approximate reconstruction. Example MySQL queries to recover the raw measurements or the compiled spectra are provided in the Appendix. III SAMPLE We applied the SPIFF pipeline on the set of all known ultracool dwarfs (UCDs), consisting mostly of brown dwarfs with some older very low-mass stars (spectral types L0 to \approx L2, J. E. Gizis et al. 2000 ; S. B. Dieterich et al. 2014 ), maintained as part of the MOCA database ( J. Gagné 2024 ; J. Gagné et al. 2026 ) , which benefited from the Ultracoolsheet ( W. M. J. Best et al. 2024 ) 7 7 7 A persistent and versioned copy of the Ultracoolsheet is available on Zenodo at doi:10.5281/zenodo.4169084 , and the SIMPLE archive K. Cruz et al. 2025 8 8 8 A persistent and versioned copy of the SIMPLE archive is available on Zenodo at doi:10.5281/zenodo.13937301 , and the current version is available at https://simple-bd-archive.org . . In this work, we define ultracool dwarfs as all objects with spectral types L0 and later. We considered all spectroscopically confirmed ultracool dwarfs and selected the best-available reference astrometry and proper motion from the literature. This set of literature ultracool dwarfs is available in the individual tables of J. Gagné et al. 2026 9 9 9 A persistent and versioned copy of the MOCAdb is available on Zenodo at doi:10.5281/zenodo.18166117 . The spectral type and W2 magnitude distribution of this sample are shown in Figure 2 . (a) Spectral type distribution of our sample (b) WISE W2 magnitudes of our sample Figure 2: Left panel: Histogram of MOCAdb spectral types for known confirmed and candidate ultracool dwarfs. Right panel: WISE W2 magnitudes as a function of spectral types for known confirmed and candidate ultracool dwarfs. Known UCD candidates are concentrated at early L spectral types and magnitudes brighter than W2 \approx 15 because they were mostly uncovered using the individual-epoch detections in the WISE mission (which reaches a 5 \sigma depth of 15.5 in W2 band E. L. Wright et al. 2010 ), which samples a much larger volume in space for warmer and intrinsically brighter early-L dwarfs. See Section III for more details. In addition to the set of known ultracool dwarfs, we compiled photometrically-selected ultracool dwarf candidates from the literature and custom searches in wide-area red and infrared surveys, to attempt a confirmation of their ultracool nature using SPHEREx spectrophotometry. The ultracool dwarf candidates were compiled from 4 different samples: • All SIMBAD ( M. Wenger et al. 2000 ) objects flagged as brown dwarf candidates, • the ultracool dwarf candidates identified through the Backyard Worlds: Planet 9 citizen science project ( M. J. Kuchner et al. 2017 ; A. C. Schneider et al. 2026 ), • the BASS ( J. Gagné et al. 2015 ) and BASS-Ultracool ( J. Gagné et al. 2015 ) searches for young ultracool dwarfs, • a search based on colors and proper motions (when available) in CatWISE ( F. Marocco et al. 2021 ) , UKIDSS LAS DR11 + ( A. Lawrence et al. 2007 ) , UHS DR3 ( A. C. Schneider et al. 2025 ) ,Gaia DR3 ( G. Collaboration et al. 2023 ) and Pan-STARRS1 DR2 ( K. C. Chambers et al. 2016 ) with colors consistent with L0 or later in each detected spectral band 10 10 10 See J. D. Kirkpatrick et al. 2011 ; W. M. J. Best et al. 2018 ; B. Burningham et al. 2010 for average ultracool dwarf colors in each respective catalogs. The UHS DR3 selection criteria were designed in a way to recover fast-moving point sources in a single and efficient ADQL query: • \sigma_{\mu_{\alpha}} and \sigma_{\mu_{\alpha}}<50 \mathrm{mas}\,\mathrm{yr}^{-1} , • |\mu_{\alpha}| or |\mu_{\delta}|\geq 50 \mathrm{mas}\,\mathrm{yr}^{-1} , • at least one proper motion component is significant at the 5\sigma level, • P_{\mathrm{saturated}}<0.1 , • P_{\mathrm{star}}>0.9 , • No Gaia EDR3 ( G. Collaboration et al. 2021 ) counterpart exists within 8^{\prime\prime} , based on the uhsSourceXGEDR3gaia_source table at http://wsa.roe.ac.uk . The criteria we used for CatWISE were the following: • \mu_{\rm tot}\geq 20 \mathrm{mas}\,\mathrm{yr}^{-1} , • \mu_{\rm tot} is detected at 5 \sigma , • -0.5<\mathrm{w1sky}<0.5 , • -0.5<\mathrm{w2sky}<1 , • W1-W2\geq 0.5 , • \sigma_{W2} must be detected with \sigma_{W2}\leq 0.15 . The criteria for UKIDSS LAS DR11 + were: • |\mu_{\alpha}|+|\mu_{\delta}|\geq 100 \mathrm{mas}\,\mathrm{yr}^{-1} , • y-J\geq 1.7 if y is detected with \sigma_{y}<0.3 , • J msut be detected with \sigma_{J}<0.3 , • P_{\mathrm{star}}\geq 0.8 . The criteria for Gaia DR3 were: • The parallax is \varpi\geq 10 mas, • \varpi/\sigma_{\varpi}\geq 10 , • visibility_periods_used \geq 10 , • ruwe <1.4 , • duplicated_source = false , • phot_rp_mean_flux_over_error >10 , • phot_g_mean_flux_over_error >10 , • G-G_{\mathrm{RP}}\geq 1.55 , • Approximate absolute G -band magnitude G+5\log_{10}(\varpi)-10\geq 14.0 . The criteria for Pan-STARRS1 DR2 were: • y and z must be detected with \sigma_{y}<0.1 , • The source must lie outside the Galactic plane: |b|>15^{\circ} • g-z\geq 5 if \sigma_{g}<0.5 , • r-z\geq 3.5 if \sigma_{r}<0.5 , • i-z\geq 1.5 if \sigma_{i}<0.5 , • z-y\geq 1 . IV DISCUSSION The application of the SPIFF pipeline to the samples described in Section III resulted in 6003 spectrophotometry products, which include every available epochs from the SPHEREx mission as of 2026 March 11. We refer to this data set as the SPIFF spectral library 11 11 11 A persistent and versioned copy of the SPIFF spectrophotometric library is available on Zenodo at doi:10.5281/zenodo.19051216 . (a) Differences between the SPIFF and IRSA data reduction pipelines (b) Differences between SPIFF and JWST-based synthetic SPHEREx spectrophotometry (c) Differences between IRSA and JWST-based synthetic SPHEREx spectrophotometry Figure 3: Differences in extracted absolute flux between the SPIFF and IRSA spectrophotometry tools for the T2.5 substellar object SIMP J013656.5+093347.3. Synthetic spectrophotometry was also reconstructed based on the JWST NIRSpec prism observations of GO program 3548, where the median of all observed epochs is shown for each wavelength bin. The SPIFF pipeline recovers absolute flux values slightly more consistent with those of JWST (the median of residuals indicate a 2.2% smaller flux in the SPIFF reduction, with median absolute deviations of 7.6%, whereas the residuals of the IRSA spectrum are 4.9% fainter with a median absolute deviation of 8.6%). Splines are shown as solid lines to guide the eye only. See Section IV.1 for more details. IV.1 Pipeline Validation We used the JWST GO program 3548 and the IRSA spectrophotometry tool to assess the absolute flux calibration of the SPIFF library (see Figure 3 ) based on the nearby, bright T2.5 dwarf SIMP J013656.5+093347.3 (SIMP J0136; É. Artigau et al. 2006 ), observed with JWST/NIRSpec using the PRISM/CLEAR configuration with the S1600A1 slit and SUB512 subarray in Bright Object Time-Series mode ( R. Akhmetshyn et al. 2025 ) . We choose this particular data set for comparison because SIMP J0136 is a well-characterized, bright source representative of our sample for which we obtain high-S/N SPHEREx spectrophotometry, with a well-characterized astrometry (crucial for the IRSA spectrophotometry tool) and variability (1.0–2.5% in the 1–5 \mu m range; A. M. McCarthy et al. 2025 ). The GO3548 observations also cover the same wavelength range over a range of epochs. The current JWST NIRSpec documentation 12 12 12 See https://jwst-docs.stsci.edu/jwst-calibration-status/nirspec-calibration-status/nirspec-fixed-slit-calibration-status . notes: Current fixed slit absolute flux calibration is generally better than 5% absolute for sources well centered in the aperture, and should be good to \approx 3% for most disperser-filter combinations (2% relative). An exception is for observations in S1600A1 [which applies to our data set] at longer than 3.5 \mu m, where, with the current F-flat reference files ( T. D. Rawle et al. 2016 ) , the current absolute flux accuracy can be 5%–6%. New F-Flats are expected to be delivered in the near future that should provide absolute flux calibration good to 1%–2%.’ Given these flux calibration requirements, the photometric variability of SIMP J0136 should not affect our assessment of flux calibration. The JWST data 13 13 13 Obtained from MAST at https://archive.stsci.edu/missions-and-data/jwst , with the provided reduced spectra based on SDP_VER=2025_4a, PRD_VER=PRDOPSSOC-072, OSS_VER=008.007.000.000, GSC_VER=GSC31, CAL_VER=1.20.2, CAL_VCS=RELEASE, CRDS_VER=13.0.6, CRDS_CTX=jwst_1464.pmap . were convolved on the SPHEREx bandpasses listed in Table 2 . This particular data set covers a range of epochs, and we adopted the median of all epochs along with the median absolute deviation as our reference flux and flux error. The SPHEREx mission design aims for an absolute flux calibration of 2% within each spectral channel ( E. McCallum et al. 2024 ; E. McDonald et al. 2024 ) , although the empirical in-flight performance has not yet been published ( H. Hui et al. 2024 ; Ashby et al., in prep.). We performed our comparison using the IRSA and SPIFF spectrophotometry products grouped in the 102 spectral channels, as described in Section II . We find only slight differences in absolute flux calibration: SPIFF recovers a median absolute flux 2.2% fainter than the JWST spectrum, but 3.4% brighter than the IRSA spectrophotometry, with respective median absolute deviations of the residuals of 7.6% and 5.8% when comparing SPIFF and JWST or IRSA and JWST, respectively. The differences between the SPIFF and IRSA results are based on the same data products and SPHEREx calibration, and must therefore be related either to different performances in the PSF-fitting algorithm, or to a slight mismatch between the provided astrometry and the true location of the PSF on the SPHEREx detectors. It is plausible that the thorough likelihood exploration provided by ultranest , combined with the flexible PSF detector position, may allow slightly more flux to be recovered than the IRSA pipeline. The differences between SPIFF and JWST are close to the expected NIRSpec absolute flux calibration performance, and may also be related to the current calibration of the SPHEREx data products that are still being improved (especially background subtraction) and to our use of preliminary SPHEREx bandpasses to convolve the JWST spectrum (we expect these differences to decrease as future SPHEREx data releases occur). Based on these comparisons, we recommend adopting a conservative error floor of \approx 8% on the absolute flux calibration measurements provided by SPIFF, until a more thorough characterization of SPHEREx absolute flux calibration is available. We do not apply this floor directly in the SPIFF data products, because measurement errors smaller than this floor may still be a valid representation of relative flux errors. (a) All high-quality SPHEREx spectra presented here (b) Average SPHEREx S/N versus WISE W2 magnitude Figure 4: Left panel: Histogram of high-quality SPHEREx spectrophotometry products presented here. The thick black line represents the current population of ultracool dwarfs with spectral type determinations based on spectroscopy, including the SPIFF library (which is shown independently as a thin gray line), as well as every other spectrometer. A significant fraction of known and newly-discovered ultracool dwarfs with spectral types T5 and earlier have spectrophotometry data products available in the SPIFF library. New spectroscopic confirmations of literature ultracool dwarf candidates are shown with a dashed blue line. New discoveries of confirmed or candidate ultracool dwarfs are shown with red and dotted orange lines, respectively. Right panel: Average spectrophotometry S/N as a function of WISE W2 magnitude. Objects with W2 =14 or brighter usually result in high-quality SPHEREx spectrophotometry with an average S/N of 10 or above, when free from background contamination. See Section IV.1 for more details. In Figure 4 , we show the average signal-to-noise ratio (S/N) for the SPIFF spectrophotometry as a function of WISE W2 magnitudes across the sample of known ultracool dwarfs, along with a histogram of the spectral types for the high-quality SPHEREx spectra. We find that data products with an average S/N of 10 per spectral channel correspond to a WISE W2 magnitude of 14.0 mag, on average. We observed only a very small dependency of the average S/N on the ecliptic latitude of the targets. SPHEREx schedules more frequent visits near the ecliptic poles which increases the number of data points per spectral channel, but the impact on the average S/N of the data binned onto the 102 SPHEREx spectral channels is much smaller than that of the W2 magnitude of the target. IV.2 Hybrid Spectral Templates We built hybrid spectral templates for each spectral subtype in the range M9–Y1 by combining the available SPHEREx spectra (binned into the 102 SPHEREx spectral channels) in our library for each known ultracool dwarf of the corresponding spectral type in the literature, while ignoring those with mentions of peculiarities (including low gravity, subdwarfs and spectral binaries). Each spectrum was scaled to its median before constructing a first template estimate based on a per-spectral channel median of each spectrum, and then the scaling that optimizes the \chi^{2} between each individual spectrum and the template estimate was computed to scale each spectrum individually in an optimal manner. After this was completed, the individual spectra were combined again, using a median. The median absolute deviation was also recorded in each spectral channel to represent the typical spread within a spectral subtype while ignoring large outliers. A spectral standard from the literature (see Table 3 ) was then convolved onto the set of SPHEREx bandpasses to create a synthetic SPHEREx spectrum at wavelengths shorter than \approx 2.4 \mu m. A linear error propagation formula was used to determine the measurement error within each SPHEREx spectrophotometric band. We then determined the scaling factor that minimized \chi^{2} with our template, and replaced all channels blue than 2.4 \mu m in our raw SPHEREx template with the synthetic standard, which results in a higher S/N, especially for the few spectral classes with only a small sample of high-quality SPHEREx spectra. We refer to these templates as hybrid in the remainder of this work, and we also make the raw SPHEREx templates available in the online-only material of this paper. Figure 5 shows example hybrid template construction steps; we note that the displayed error bars do not represent the error on the mean, but the median absolute deviation of all observed SPHEREx spectra, dominated by the limited S/N of the individual SPHEREx spectrophotometry products. For spectral types T6–Y1, we used JWST NIRSpec PRISM spectra from programs GO 2302, GTO 1189, and GTO 4539 for the brown dwarfs SDSSp J162414.37+002915.6 (T6, A. J. Burgasser et al. 2025 ), 2MASS J03480772-6022270 (T7, A. J. Burgasser et al. 2006 ), WISEPC J225540.74-311841.8 (T8, J. D. Kirkpatrick et al. 2011 ), WISE J210200.15-442919.5 (T9, J. D. Kirkpatrick et al. 2012 ), WISEPC J205628.90+145953.3 (Y0, J. D. Kirkpatrick et al. 2011 ), and WISEPA J154151.66-225025.2 (Y1, A. C. Schneider et al. 2015 ) as they cover the full range of SPHEREx bandpasses and the available SPHEREx spectrophotometry are of a much lower S/N. The final set of field hybrid templates is shown in Figures 6 and 7 . These steps were repeated to construct hybrid templates for low-gravity and subdwarf standards available in the literature (also in Table 3 ) when either the literature data or available SPHEREx data products allowed a full wavelength coverage from 1 \mu m to 5 \mu m. (a) L0 SPHEREx template based on 607 objects (b) L5 SPHEREx template based on 112 objects (c) T0 SPHEREx template based on 32 objects (d) T5 SPHEREx template based on 84 objects Figure 5: Examples of hybrid SPHEREx template construction. Synthetic spectrophotometry from a known spectral standard is shown in blue (top), and a raw SPHEREx template constructed from the median of all high-quality SPHEREx data with the appropriate spectral type are shown in red (bottom). The hybrid template, constructed by extending the synthetic standard-based spectrophotometry with the raw SPHEREx template beyond 2.4 \mu m (vertical, dashed line) is shown in black (middle row). For some templates, this limit was moved to 2.1–2.3 \mu m when the quality of the ground-based data was insufficient in this region. The gray error bars do not represent the error on the mean, but the median absolute deviation of all observed SPHEREx spectra to represent visually the typical spread in observed SPHEREx fluxes, which are mostly due to the limited S/N of the spectrophotometry. Splines are shown as solid lines to guide the eye only. The complete figure set (47 images) is available in the online journal. See Section IV.2 for more details. Figure 6: Sequence of field L-type SPHEREx field hybrid templates constructed in this work, along with relevant chemical species. Splines are shown as solid lines to guide the eye only. See Section IV.2 for more details. Figure 7: Sequence of field T- and Y-type SPHEREx field hybrid templates constructed in this work, along with relevant chemical species. Splines are shown as solid lines to guide the eye only. See Section IV.2 for more details. IV.3 Spectral Typing Automated spectral types were determined using \chi^{2} fitting of each hybrid spectral template applied to every spectrum of the SPIFF library. When doing this, the worst 5 data points are always ignored to make spectral typing more robust against outlier data points, and the optimal scaling factor that minimizes \chi^{2} is determined analytically, independently for each hybrid template. Additionnally, non-field hybrid templates (young brown dwarfs or subdwarfs) are assigned a penalty in their total \chi^{2} score of 12.4, and extremely low-gravity or extreme subdwarfs are assigned a larger penalty of 26.2, to avoid classifying low-S/N spectra as potential subdwarfs or young objects. Figures comparing the SPIFF spectrophotometry to the best three hybrid templates were generated for each literature ultracool dwarf and were then visually vetted to flag contaminated or otherwise bad-quality spectra. The high-quality spectra are shown in Figure 8 , and the four typical failure modes or resulting bad-quality spectrophotometry are shown in Figure 9 . Figure 8: Template-fitting validation figures for the high-quality spectra of literature ultracool dwarfs. The complete figure set (2050 images) is available in the online journal. See Section IV.3 for more details. (a) A contaminating background star (b) A false-positive star reddened by interstellar dust (c) An ambiguous case due to low-quality data (d) Confusion between two close sources Figure 9: The four common failure modes resulting in poor-quality SPIFF spectrophotometry. Upper left: an unresolved contaminating background star dominates the signal (occurred for 8.6% of known ultracool dwarfs). Upper right: A false-positive brown dwarf corresponding to a reddened background star (occurred for 0.3% of known ultracool dwarfs). Lower left: An ambiguous spectrum due to low S/N data (occurred for 0.3% of known ultracool dwarfs with W2 \leq 14.0 mag). Lower right: Two or more resolved but closely separating sources causing confusion, where the PSF model can converge to either source depending on the epoch (occurred for 2.6% of known ultracool dwarfs). See Section IV.4 for more details. Figure 10 shows the comparison in automated spectral types with literature spectral types, showing good agreement with a median absolute deviation of 1.5 subtypes. We caution that SPIFF data products that are contaminated by background stars can be significantly more discrepant. Figure 11 shows the fitting reduced \chi^{2} as a function of spectral types and W2 magnitudes. Figure 10: Validation of the automated SPHEREx spectral types compared with literature spectral types for known ultracool dwarfs (blue circles or field objects, rightward purple triangles for known peculiar UCDs), and for photometric ultracool dwarf candidates with spectral type estimates in the literature (upward green triangles). The thick, dashed line represents the 1:1 relation and the thinner dashed lines represent the median absolute deviation (1.5 subtypes). A random jitter of 0.3 subtypes was added to both spectral type dimensions to improve visibility. See Section IV.3 for more details. (a) Spectral type versus reduced \chi^{2} (b) WISE W2 magnitude versus reduced \chi^{2} Figure 11: Autotyping reduced \chi^{2} as a function of spectral type (left) and W2 magnitude (right). We included a floor of 10% on the minimum SPIFF error bars to compute this \chi^{2} (as a guard against high-S/N outliers) and adjust a robust linear fit as a function of spectral type (red line in the left panel) to compensate for the systematic trend observed here. All objects with \chi^{2}> 3 were flagged as potentially peculiar targets. See Section IV.4 for more details. All template-fitting figures were inspected visually to assess the quality of the resulting SPHEREx spectrophotometry data products output by SPIFF. We used the set of known ultracool dwarfs with previous spectroscopic confirmation and a WISE W2 magnitude 14.0 or brighter as a test sample to establish the different failure modes and their respective occurrence rates. Out of 1327 known brown dwarfs with previous spectroscopic confirmation that have W2 \leq 14.0, we classified 1078 (81.2%) has having high-quality SPHEREx spectrophotometry clearly matching an ultracool hybrid template, 114 (8.6%) as having a slightly earlier spectral type (M8–M9), 14 (1.1%) as a blended ultracool dwarf and background star spectra, 20 (1.5%) as ambiguous cases, and 101 (7.6%) as a low-quality spectrum. Out of the 101 bad cases, 4 (0.3%) were classified as background, reddened sources, 4 (0.3%) were classified as having a S/N too low for proper classification, 35 (2.6%) were classified as a clear mismatch where the signal is dominated by a background star with a much earlier spectral type, and 58 (2.6%) as blended spectra from more than one source. We note that these problematic cases likely indicate a contaminated SPIFF data product, and not necessarily a rejection of the ultracool status of the targets, given that contamination from background stars can play an important role with the large pixel size of SPHEREx. The final library of SPIFF spectrophotometry products that were visually vetted cover a total of 6003 UCDs. Their average signal-to-noise ratio as a function of spectral type are presented in Figure 12 . Figure 12: Average signal-to-noise ratio of the spectrophotometry products compiled here as a function of spectral type. A random jitter of 0.3 subtypes was added for clarity. See Section IV.3 for more details. IV.4 Confirmation of Ultracool Dwarf Candidates We constructed SPHEREx spectrophotometry for a set of ultracool dwarf candidates described in Section III , applied the automated spectral typing method of Section IV.3 , and visually inspected all resulting template-fitting figures to confirm the nature of ultracool dwarfs in an attempt to demonstrate the power of SPHEREx in distinguishing photometrically selected ultracool dwarfs from contaminants such as reddened background stars and galaxies, especially in the absence of reliable proper motion detection. We visually verified the hybrid template matches for all SPHEREx spectrophotometry with flux measurements available in at least 50/102 of the SPHEREx channels. This allowed us to 3953 ultracool dwarfs with spectral types L0–Y1, 1937 of which are from the Backyard Worlds set of high-proper motion candidates ( A. C. Schneider et al. 2026 ) . This represents more than a doubling of the number of ultracool dwarfs with spectroscopic confirmation. All newly confirmed ultracool dwarfs are shown in Figure 13 . There are 1916 low-S/N SPHEREx data products that we classify as ‘New UCD candidates’ and hint at water, carbon dioxide and/or methane features but that we do not consider of high-enough quality to warrant a confirmation of the UCD nature of the target (see Figure 4(a) ). We categorize these cases as new UCD candidates, and further follow-up with higher-S/N spectroscopy will be required to confirm them. Figure 13: Confirmation of new and known ultracool dwarf candidates. The complete figure set (3953 images) is available in the online journal. See Section IV.4 for more details. IV.5 Peculiar Spectra Our hybrid template-based spectral classification of known and newly confirmed ultracool dwarfs allowed us to flag a set of 3739 spectra with potentially peculiar features, because they were best-matched to a subdwarf or young hybrid template, or because of the high \chi^{2} of the best hybrid template match. These fall in three categories: low-gravity brown dwarfs (125 previously known young brown dwarfs, 2668 new young candidates, although some in the later category may have implicitly been called young in the literature from their membership in young associations); subdwarfs (25 previously known subdwarfs, 250 new subdwarf candidates); and otherwise peculiar spectra (10 previously flagged as peculiar, and 865 newly flagged as such). These peculiar spectra are shown in Figure 14 and listed in Table 4 sorted by dataset and from late to early spectral types, and would be interesting candidates for future spectroscopic characterization. We caution that some subdwarf hybrid templates (e.g., esdL8, sdT4) resemble a blend between a T dwarf and an early-type star across 1–5 \mu m, and therefore these hybrid templates are often attributed to cases with a background star contaminating a SPHEREx pixel. These cases should therefore be treated with care; we generally adopted only those with a high S/N and clear detections of water or methane that were not within 10″ of a background star. Figure 14: Peculiar spectra identified by a visual comparison of best-matching hybrid templates with high-quality SPHEREx spectra. The example object shown is a known T8 ( M. C. Cushing et al. 2014 ) with an extreme \approx 4.2 \mu m CO 2 feature for its spectral type. The complete figure set (3739 images) is available in the online journal. See Section IV.5 for more details. IV.6 Access to Data Products We provide both our final set of SPHEREx hybrid and raw templates in the online Data Behind Figures product of Figures 6 and 7 , along with the current snapshot of high-quality SPHEREx spectra, both as a package of csv spectra, and as template-matching figures in the online-only material associated with this work. All data products are also available on the MOCA database 14 14 14 Available at https://mocadb.ca . ( J. Gagné 2024 ; J. Gagné et al. 2026 ) , and will be updated there, as well as in a persistent and versioned Zenodo repository at doi:10.5281/zenodo.19051216 , as new SPHEREx data products become available. V CONCLUSIONS We present a new data reduction package for SPHEREx spectrophotometry, designed to recover high-quality data sets for fast-moving objects such as nearby brown dwarfs. We compiled a set of 6003 high-quality 0.75–5.0 \mu m spectra for 2050 known ultracool dwarfs and 3953 newly confirmed ones, along with tools to automatically determine their spectral types based on hybrid spectral templates. These data sets already more than double the number of ultracool dwarfs with spectroscopic confirmation and will be refined further as the SPHEREx mission progresses and will facilitate several follow-up studies benefiting from both the wavelength and temporal coverages provided by SPHEREx. Such examples include: the empirical determination of bolometric luminosities for a large number of brown dwarfs; the search for unresolved Y dwarf companions; detailed atmospheric properties across a wide range of atmospheric layers with spectral retrieval methods; brown dwarf variability studies; the identification of brown dwarfs with peculiar properties (e.g., subdwarfs, young brown dwarfs or chemically peculiar objects); the search or confirmation of protoplanetary disks around brown dwarfs; and the spectroscopic confirmation of new ultracool dwarf candidates. We thank the anonymous reviewer for their helpful comments that improved the quality of this work. We would like to thank the IPAC helpdesk for their generous help with this project. We thank Jean-Maxime Couillard for his technical support with this project. J.G. and A.R.D. acknowledge the support of the Canadian Space Agency (CSA) [ 25JWGO4B10 ], and J.G. acknowledges the support of the Natural Sciences and Engineering Research Council of Canada (NSERC), funding reference number RGPIN-2021-03121. L.-P.C. acknowledges support from Mitacs through the Mitacs Accelerate program, in partnership with the Montreal Planetarium. J.F acknowledges the support of NSF CAREER award number 2238468. This research was enabled in part by support provided by Université de Montréal ( umontreal.ca ) and the Digital Research Alliance of Canada ( alliancecan.ca ). This publication makes use of data products from the Spectro-Photometer for the History of the Universe, Epoch of Reionization and Ices Explorer (SPHEREx), which is a joint project of the Jet Propulsion Laboratory and the California Institute of Technology, and is funded by the National Aeronautics and Space Administration. This paper uses the SPHEREx Quick Release Spectral Images – QR2 http://dx.doi.org/10.26131/IRSA652 (catalog 10.26131/IRSA652) . This research made use of the Montreal Open Clusters and Associations (MOCA) database, operated at the Montréal Planétarium ( J. Gagné 2024 ; J. Gagné et al. 2026 ) . This work has benefited from The UltracoolSheet ( W. M. J. Best et al. 2024 ) at http://bit.ly/UltracoolSheet , maintained by Will Best, Trent Dupuy, Michael Liu, Rob Siverd, and Zhoujian Zhang, and developed from compilations by T. J. Dupuy & M. C. Liu 2012 , T. J. Dupuy & A. L. Kraus 2013 , M. C. Liu et al. 2016 , W. M. J. Best et al. 2018 , and W. M. J. Best et al. 2021 . This work is based in part on observations made with the NASA/ESA/CSA James Webb Space Telescope. The data were obtained from the Mikulski Archive for Space Telescopes at the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS 5-03127 for JWST. These observations are associated with programs GO 3548, GO 2302, GTO 1189, and GTO 4539. This research made use of: the SIMBAD database and VizieR catalog access tool, operated at the Centre de Données astronomiques de Strasbourg, France ( M. Wenger et al. 2000 ) ; data products from the Two Micron All Sky Survey ( 2MASS ; M. F. Skrutskie et al. 2006 ), which is a joint project of the University of Massachusetts and the Infrared Processing and Analysis Center (IPAC)/California Institute of Technology (Caltech), funded by the National Aeronautics and Space Administration (NASA) and the National Science Foundation ( M. F. Skrutskie et al. 2006 ) ; data products from the Wide-field Infrared Survey Explorer ( WISE ; and E. L. Wright et al. 2010 ), which is a joint project of the University of California, Los Angeles, and the Jet Propulsion Laboratory (JPL)/Caltech, funded by NASA. This work has made use of the SIMPLE Archive of low-mass stars, brown dwarfs, and directly imaged exoplanets: 10.5281/zenodo.13937301 . All the JWST data used in this paper can be found in MAST: http://dx.doi.org/10.17909/edj9-e737 (catalog 10.17909/edj9-e737) , http://dx.doi.org/10.17909/4wg9-8196 (catalog 10.17909/4wg9-8196) . In this Appendix, we provide example MySQL queries which can be used in the MOCAdb to retrieve distinct SPHEREx SPIFF data products. The main data products are spectrophotometry in F_{\lambda} binned over the 102 SPHEREx spectral channels, and can be acquired for a single object using: ⬇ -- Obtain one SPHEREx data product for SIMP J0136 -- (binned per 102 SPHEREx channel) SELECT -- Convert wavelengths to microns ds . wavelength_angstrom /1 e4 AS wv_microns , -- Convert absolute F_lambda (W/m^2/Angstroms) to F_lambda (W/m^2/microns) ds . flux_flambda /1 e4 AS flux , -- Same with errors ds . flux_flambda_unc /1 e4 AS flux_err -- Grab the MOCAdb object ID based on designation FROM mechanics_all_designations mad -- Grab spectra headers by matching unique MOCAdb object ID JOIN moca_spectra ms USING ( moca_oid ) -- Grab spectra data points by matching unique MOCAdb spectrum ID JOIN data_spectra ds USING ( moca_specid ) WHERE -- Specify object designation mad . designation = ’SIMP ␣ J013656.5+093347.3’ -- Only select latest (ignored=0) SPHEREx SPIFF -- reduction (moca_specpackid=55). -- Use moca_specpackid=54 for the (rare) IRSA-pipeline -- SPHEREx data products stored in MOCAdb AND ms . ignored =0 AND ms . moca_specpackid =55; The original SPIFF outputs based on ultranest samples in F_{\nu} , often with repeated wavelengths and individual observing epochs, can be obtained with: ⬇ -- Obtain one SPHEREx data product for SIMP J0136 -- (time-resolved, not binned per channel, based on ultranest) SELECT -- Grab the exact bandpass wavelength center and width -- at that specific detector location ds . psf_un_wv_um AS wv_um , ds . psf_un_wv_width_um AS wv_wid_um , -- Grab F_nu flux (microjanskies) and error ds . psf_un_flux_uJy AS flux , ds . psf_un_flux_uJy_err AS flux_err , -- Grab the exact MJD epoch of this SPHEREx FITS -- image (useful for time-resolved analysis) ds . mjd_avg -- Grab the MOCAdb object ID based on designation FROM mechanics_all_designations mad -- Grab the raw SPIFF data points by matching unique MOCAdb object ID JOIN data_spherex_spectra_spiff ds USING ( moca_oid ) WHERE -- Specify object designation mad . designation = ’SIMP ␣ J013656.5+093347.3’ -- Reject any data points flagged as bad AND ds . ignored =0; Spectrophotometry based on the scipy optimizer without ultranest sampling can be obtained with: ⬇ -- Obtain one SPHEREx data product for SIMP J0136 -- (time-resolved, not binned per channel, based on scipy) SELECT -- Grab the exact bandpass wavelength center and width -- at that specific detector location ds . psf_scipy_wv_um AS wv_um , ds . psf_scipy_wv_width_um AS wv_wid_um , -- Grab F_nu flux (microjanskies) and error ds . psf_scipy_flux_uJy AS flux , ds . psf_scipy_flux_uJy_err AS flux_err , -- Grab the exact MJD epoch of this SPHEREx FITS -- image (useful for time-resolved analysis) ds . mjd_avg -- Grab the MOCAdb object ID based on designation FROM mechanics_all_designations mad -- Grab the raw SPIFF data points by matching unique MOCAdb object ID JOIN data_spherex_spectra_spiff ds USING ( moca_oid ) WHERE -- Specify object designation mad . designation = ’SIMP ␣ J013656.5+093347.3’ -- Reject any data points flagged as bad AND ds . ignored =0; Spectrophotometry based on aperture photometry (which requires extreme care for background contamination) can be obtained with: ⬇ -- Obtain one SPHEREx data product for SIMP J0136 -- (time-resolved, not binned per channel, based on aperture photometry) SELECT -- Grab the exact bandpass wavelength center and width -- at the detector location corresponding to the -- photometric center-of-mass ds . com_wv_um AS wv_um , ds . com_wv_width_um AS wv_wid_um , -- Grab F_nu flux (microjanskies) and error ds . ap_flux_uJy AS flux , ds . ap_flux_uJy_err AS flux_err , -- Grab the exact MJD epoch of this SPHEREx FITS -- image (useful for time-resolved analysis) ds . mjd_avg -- Grab the MOCAdb object ID based on designation FROM mechanics_all_designations mad -- Grab the raw SPIFF data points by matching unique MOCAdb object ID JOIN data_spherex_spectra_spiff ds USING ( moca_oid ) WHERE -- Specify object designation mad . designation = ’SIMP ␣ J013656.5+093347.3’ -- Reject any data points flagged as bad AND ds . ignored =0; Table 1: Preview of columns in data_spherex_spectra_spiff Column Datatype Description id bigint(20) unsigned Primary identifier for this row. md5_uid varchar(32) Deterministic MD5 UID built from pipeline versio(…) moca_oid int(10) unsigned MOCA object identifier (foreign key to moca_obje(…) moca_specid int(10) unsigned MOCA spectrum identifier linked to this SPIFF ro(…) target_name varchar(255) Human-readable target name passed to LV2 for thi(…) reference_ra_deg double Reference right ascension in degrees used for PM(…) reference_dec_deg double Reference declination in degrees used for PM pro(…) reference_crd_epoch_yr double Epoch (year) of the reference sky coordinates. reference_pmra_masyr double Reference proper motion in RA*cos(Dec), in mas/yr. reference_pmdec_masyr double Reference proper motion in Dec, in mas/yr. input_ra_deg double RA in degrees actually sent to the image analyze(…) input_dec_deg double Dec in degrees actually sent to the image analyz(…) Note. — This table is available in its entirety online. The full table contains 112 rows. Table 2: Preliminary SPHEREx channel characteristics (Preview) Index \lambda_{\mathrm{center}} ( \mu m) \lambda_{\mathrm{width}} ( \mu m) 1 0.744 0.022 2 0.766 0.022 3 0.788 0.022 4 0.810 0.022 5 0.832 0.022 6 0.853 0.022 7 0.875 0.022 8 0.897 0.022 9 0.919 0.022 10 0.941 0.022 11 0.963 0.022 12 0.985 0.022 Note. — This table is available in its entirety online. The full table contains 102 rows. Table 3: SPHEREx Template Grids (Full Table) Grid Type Spectral Instrument N_{\mathrm{spec}} a a Number of SPHEREx spectra used for the construction of this template. Standard Designation b b Designation of the object we used as a spectral standard for wavelengths below 2.4 \mu m. References specify which work established their spectral type. Ref. Type Field M8 SpeX IRTF 76 VB 10 1 Field M9 SpeX IRTF 71 LHS 2924 1 Field L0 SpeX IRTF 544 2MASP J0345432+254023 1 Field L1 SpeX IRTF 399 2MASS J21304464-0845205 1 Field L2 SpeX IRTF 382 2MASSI J0408290-145033 2 Field L3 SpeX IRTF 178 2MASSW J1506544+132106 3 Field L4 SpeX IRTF 132 2MASS J21580457-1550098 1 Field L5 SpeX IRTF 112 2MASSI J0652307+471034 3 Field L6 SpeX IRTF 56 2MASSI J1010148-040649 1 Field L7 SpeX IRTF 57 2MASS J02052940-1159296 4 Field L8 SpeX IRTF 35 2MASSW J1632291+190441 1 Field L9 SpeX IRTF 51 2MASS J02550357-4700509 1 Field T0 SpeX IRTF 31 SIMP J0150+3827 5 Field T1 Nires Keck 38 SDSS J083717.21-000018.0 1 Field T2 SpeX IRTF 56 SDSS J125453.90-012247.4 1 Field T3 SpeX IRTF 47 2MASS J12095613-1004008 6 Field T4 SpeX IRTF 58 2MASSI J2254188+312349 6 Field T5 SpeX IRTF 85 2MASS J15031961+2525196 6 Field T6 SpeX IRTF 0 SDSSp J162414.37+002915.6 6 Field T7 SpeX IRTF 0 2MASSI J0727182+171001 6 Field T8 SpeX IRTF 0 2MASSI J0415195-093506 6 Field d/sdT8 SpeX IRTF 0 2MASSI J0415195-093506 6 Field T9 Nires Keck 0 UGPS J072227.51-054031.2 6 Field Y0 NIRSpec JWST 0 WISEA J120604.25+840110.5 7 Field Y1 NIRSpec JWST 0 WISE J053516.80-750024.9 8 Intermediate gravity M8 \beta SpeX IRTF 22 2MASSI J2323134-024435 9 Intermediate gravity M9 \beta SpeX IRTF 7 2MASSI J1411213-211950 10 Intermediate gravity L0 \beta SpeX IRTF 15 2MASS J11544223-3400390 2 Intermediate gravity L1 \beta SpeX IRTF 22 2MASS J10224821+5825453 2 Intermediate gravity L2 \beta SpeX IRTF 19 LSPM J0602+3910 11 Intermediate gravity L3 \beta SpeX IRTF 19 2MASS J22495345+0044046 11 Intermediate gravity L4 \beta SpeX IRTF 11 2MASSW J0030300-145033 9 Very low gravity M9 \gamma SpeX IRTF 6 TWA 28 9 Very low gravity L0 \gamma SpeX IRTF 25 2MASS J01415823-4633574 2 Very low gravity L1 \gamma SpeX IRTF 21 2MASSI J0518461-275645 2 Very low gravity L2 \gamma SpeX IRTF 9 2MASSI J0536199-192039 2 Very low gravity L3 \gamma Flamingos-2 Gemini South 13 2MASS J04185879-4507413 9 Very low gravity L4 \gamma SpeX IRTF 16 2MASS J16154255+4953211 2 Very low gravity L5 \beta / \gamma SpeX IRTF 9 2MASS J03264225-2102057 9 Very low gravity L7 \gamma SpeX IRTF 8 PSO J318.5338-22.8603 12 Extremely low gravity L0 \delta SpeX IRTF 7 2MASS J06085283-2753583 9 Slight subdwarfs d/sdL8 SpeX IRTF 28 2MASS J11582077+0435014 6 Slight subdwarfs d/sdT1 Fire Magellan 2 WISEA J030119.39-231921.1 6 Subdwarfs sdM9.5 SpeX IRTF 4 2MASS J10130734-1356204 13 Subdwarfs sdL3.5 SpeX IRTF 4 2MASS J12563716-0224522 1 Subdwarfs sdT4 Nires Keck 3 WISE J155349.98+693355.2 6 Extreme subdwarfs esdL8 Nires Keck 2 2MASS J05325346+8246465 6 Note. — References: (1) J. D. Kirkpatrick et al. 2010 ; (2) K. L. Cruz et al. 2018 ; (3) K. L. Cruz et al. 2009 ; (4) J. D. Kirkpatrick et al. 1999 ; (5) A. Bravo et al. 2023 ; (6) A. J. Burgasser et al. 2025 ; (7) A. C. Schneider et al. 2015 ; (8) J. D. Kirkpatrick et al. 2012 ; (9) J. Gagné et al. 2015 ; (10) K. L. Cruz et al. 2003 ; (11) K. N. Allers & M. C. Liu 2013 ; (12) M. C. Liu et al. 2013 ; (13) N. Lodieu et al. 2019 . Table 4: SPHEREx Spectral Classifications for High-Quality SPIFF Spectrophotometry Products (Preview) Designation RA Dec Dataset SPIFF \chi^{2}_{\mathrm{r}} Literature Ref. J2000 J2000 Autotypes Spectral Types CWISE J101438.37-272622.7 10:14:38.37 -27:26:22.8 New Discoveries T9 1.6 \cdots \cdots CWISE J232834.16+183438.1 23:28:34.17 +18:34:38.1 New Discoveries T9 1.1 \cdots \cdots CWISE J233232.63+232559.0 23:32:32.63 +23:25:59.0 New Discoveries T9 1.8 \cdots \cdots UHS DR3 459624484576 01:51:26.04 +42:27:03.4 New Discoveries T9 1.3 \cdots \cdots UHS DR3 459702022403 23:28:46.63 +31:39:17.4 New Discoveries T9 1.4 \cdots \cdots UHS DR3 459758596359 22:40:51.02 +36:25:31.4 New Discoveries T9 1.0 \cdots \cdots UHS DR3 459761069871 00:52:58.95 +47:56:06.0 New Discoveries T9 1.0 \cdots \cdots UHS DR3 460104839527 23:38:02.12 +38:53:04.9 New Discoveries T9 1.3 \cdots \cdots CWISE J043424.24+093228.3 04:34:24.24 +09:32:28.5 New Discoveries T8 1.3 \cdots \cdots CWISE J055506.02+394701.6 05:55:06.03 +39:47:01.7 New Discoveries T8 1.5 \cdots \cdots CWISE J080858.88-125605.6 08:08:58.89 -12:56:05.6 New Discoveries T8 1.3 \cdots \cdots CWISE J094449.87+265451.6 09:44:49.87 +26:54:51.6 New Discoveries T8 1.6 \cdots \cdots CWISE J094615.56+351434.0 09:46:15.56 +35:14:34.0 New Discoveries T8 1.7 \cdots \cdots CWISE J010134.74+403832.8 01:01:34.74 +40:38:32.8 New Discoveries T8 1.4 \cdots \cdots CWISE J221355.02-392301.2 22:13:55.02 -39:23:01.2 New Discoveries T8 1.5 \cdots \cdots UHS DR3 459568471419 17:37:07.04 +44:58:46.0 New Discoveries T8 1.9 \cdots \cdots UHS DR3 459659536509 23:12:41.24 +39:26:53.4 New Discoveries T8 1.8 \cdots \cdots UHS DR3 459687123729 02:24:20.41 +36:33:53.0 New Discoveries T8 1.4 \cdots \cdots UHS DR3 459775185717 07:59:31.48 +57:59:25.5 New Discoveries T8 1.2 \cdots \cdots UHS DR3 459785050391 22:36:09.71 +38:14:01.6 New Discoveries T8 1.6 \cdots \cdots UHS DR3 459856516289 14:27:38.14 +31:57:54.3 New Discoveries T8 1.8 \cdots \cdots UHS DR3 459900436213 17:08:30.09 +44:24:57.3 New Discoveries T8 1.1 \cdots \cdots UHS DR3 459985723337 02:28:18.01 +45:25:49.8 New Discoveries T8 1.5 \cdots \cdots CWISE J044853.70-193543.6 04:48:53.63 -19:35:44.5 New Discoveries T7 1.5 \cdots \cdots CWISE J081921.15+740017.4 08:19:21.17 +74:00:17.5 New Discoveries T7 1.4 \cdots \cdots UGCS J122914.34+221448.6 12:29:14.34 +22:14:48.6 New Discoveries T7 1.5 \cdots \cdots CWISE J012055.42+752710.5 01:20:55.43 +75:27:10.5 New Discoveries T7 1.4 \cdots \cdots CWISE J111858.30-242304.1 11:18:58.31 -24:23:04.1 New Discoveries T7 1.9 \cdots \cdots CWISE J131932.56-075412.6 13:19:32.56 -07:54:12.7 New Discoveries T7 1.6 \cdots \cdots CWISE J220948.71+235913.2 22:09:48.71 +23:59:13.2 New Discoveries T7 1.5 \cdots \cdots UHS DR3 459778426590 20:37:46.99 +14:23:57.3 New Discoveries T7 pec 4.7 \cdots \cdots UHS DR3 459800102948 12:22:42.76 +53:34:15.2 New Discoveries T7 1.2 \cdots \cdots CWISE J060305.29+543836.8 06:03:05.30 +54:38:36.9 New Discoveries T6 1.2 \cdots \cdots CWISE J165709.68+261334.6 16:57:09.68 +26:13:34.7 New Discoveries T6 1.3 \cdots \cdots UHS DR3 459597400384 20:06:50.46 +12:52:21.2 New Discoveries T6 1.2 \cdots \cdots UHS DR3 459661307265 10:07:50.45 +47:14:15.0 New Discoveries T6 1.4 \cdots \cdots UHS DR3 459686287978 06:08:52.54 +46:07:53.0 New Discoveries T6 1.4 \cdots \cdots UHS DR3 459765697995 23:39:17.64 +50:41:15.8 New Discoveries T6 1.8 \cdots \cdots UHS DR3 459988753959 20:51:52.03 +10:38:06.9 New Discoveries T6 1.4 \cdots \cdots CWISE J022243.38+323137.2 02:22:43.39 +32:31:37.4 New Discoveries T5 1.9 \cdots \cdots Note. — References: (1) C. G. Tinney et al. 2018 ; (2) This paper; (3) C. V. Cardoso et al. 2015 ; (4) J. Gagné et al. 2023 ; (5) J. D. Kirkpatrick et al. 2024 ; (6) J. D. Kirkpatrick et al. 2012 ; (7) C. G. Tinney et al. 2012 ; (8) G. N. Mace et al. 2013 ; (9) E. C. Martin et al. 2018 ; (10) J. D. Kirkpatrick et al. 2011 ; (11) N. Lodieu et al. 2009 ; (12) A. J. Burgasser et al. 2025 ; (13) M. C. Cushing et al. 2014 ; (14) E. L. Wright et al. 2013 ; (15) B. Burningham et al. 2008 ; (16) A. J. Burgasser et al. 2003 ; (17) A. J. Burgasser et al. 2011 ; (18) A. M. Meisner et al. 2020a ; (19) J. D. Kirkpatrick et al. 2021 ; (20) G. Robbins et al. 2023 ; (21) A. M. Meisner et al. 2020b ; (22) L. Albert et al. 2011 ; (23) B. Burningham et al. 2013 ; (24) D. L. Looper et al. 2007 ; (25) A. C. Schneider et al. 2015 ; (26) D. J. Pinfield et al. 2008 ; (27) N. Lodieu et al. 2007 ; (28) M. A. Thompson et al. 2013 ; (29) N. Lodieu et al. 2012 ; (30) A. J. Burgasser et al. 2006 ; (31) B. Burningham et al. 2010 ; (32) W. M. J. Best et al. 2015 ; (33) K. Chiu et al. 2006 ; (34) J. S. Pineda et al. 2016 ; (35) J. J. Greco et al. 2019 ; (36) N. R. Deacon et al. 2012 ; (37) A. J. Burgasser et al. 2010 ; (38) M. Aberasturi et al. 2014 ; (39) M. J. Kuchner et al. 2017 ; (40) G. Bihain et al. 2013 ; (41) É. Artigau et al. 2010 ; (42) D. Stern et al. 2007 ; (43) C. Reylé et al. 2010 ; (44) N. R. Deacon et al. 2011 ; (45) Z. H. Zhang et al. 2019 ; (46) Z. H. Zhang et al. 2025 ; (47) F. Marocco et al. 2024 ; (48) A. C. Schneider et al. 2022 ; (49) J. M. Vos et al. 2022 ; (50) K. Kellogg et al. 2017 ; (51) S. S. Sheppard & M. C. Cushing 2009 ; (52) F. Marocco et al. 2015 ; (53) M. C. Cushing et al. 2011 ; (54) T. R. Kendall et al. 2007b ; (55) F. Marocco et al. 2013 ; (56) W. M. J. Best et al. 2013 ; (57) A. C. Day-Jones et al. 2013 ; (58) A. Sanghi et al. 2023 ; (59) L. Smith et al. 2014 ; (60) Y. Matsuoka et al. 2019 ; (61) M. C. Liu et al. 2016 ; (62) J. Radigan et al. 2013 ; (63) J. D. Kirkpatrick et al. 2008 ; (64) S. A. Metchev et al. 2008 ; (65) É. Artigau et al. 2011 ; (66) K. L. Luhman & S. S. Sheppard 2014 ; (67) A. C. Schneider et al. 2016a ; (68) J. Y. Zhang et al. 2024 ; (69) J. Robert et al. 2016 ; (70) Y. Matsuoka et al. 2018 ; (71) É. Artigau et al. 2006 ; (72) J. D. Kirkpatrick et al. 2010 ; (73) A. C. Schneider et al. 2023a ; (74) J. C. Beamín et al. 2015 ; (75) W. M. J. Best et al. 2020 ; (76) A. C. Schneider et al. 2024 ; (77) S. L. Hawley et al. 2002 ; (78) D. L. Looper et al. 2008a ; (79) T. J. Dupuy & M. C. Liu 2012 ; (80) K. Kellogg et al. 2015 ; (81) K. Chiu et al. 2008 ; (82) T. R. Kendall et al. 2004 ; (83) Z. H. Zhang et al. 2018 ; (84) J. Gagné et al. 2015 ; (85) A. Bravo et al. 2023 ; (86) J. D. Kirkpatrick et al. 1999 ; (87) A. C. Schneider et al. 2014 ; (88) R. D. Scholz et al. 2014 ; (89) J. D. Kirkpatrick et al. 2014 ; (90) K. L. Cruz et al. 2007 ; (91) J. D. Kirkpatrick et al. 2000 ; (92) K. L. Cruz et al. 2003 ; (93) K. Mužić et al. 2012 ; (94) G. R. Knapp et al. 2004 ; (95) P. J. Castro et al. 2013 ; (96) N. R. Deacon et al. 2014 ; (97) A. C. Schneider et al. 2017 ; (98) A. J. Burgasser et al. 2013 ; (99) E. Pons et al. 2019 ; (100) R. D. Scholz et al. 2009 ; (101) Z. H. Zhang et al. 2009 ; (102) A. J. Burgasser et al. 2008 ; (103) J. D. Kirkpatrick et al. 2016 ; (104) K. Geißler et al. 2011 ; (105) J. E. Gizis et al. 2015 ; (106) A. Rothermich et al. 2024 ; (107) D. C. Bardalez Gagliuffi et al. 2019 ; (108) M.-E. Naud et al. 2014 ; (109) W. M. J. Best et al. 2017a ; (110) J. Gagné et al. 2017 ; (111) C. Alves de Oliveira et al. 2012 ; (112) J. K. Faherty et al. 2016 ; (113) M. C. Liu et al. 2013 ; (114) A. C. Schneider et al. 2023b ; (115) I. N. Reid et al. 2008 ; (116) A. Pérez-Garrido et al. 2018 ; (117) E. E. Mamajek et al. 2018 ; (118) D. C. Bardalez Gagliuffi et al. 2014 ; (119) S. J. Schmidt et al. 2010 ; (120) B. Burningham et al. 2021 ; (121) R. Kiman et al. 2019 ; (122) A. Pérez-Garrido et al. 2017 ; (123) K. L. Cruz et al. 2009 ; (124) J. K. Faherty et al. 2013 ; (125) D. L. Looper et al. 2008b ; (126) Z. Zhang et al. 2022 ; (127) J. I. Gomes et al. 2013 ; (128) P. J. Castro & J. E. Gizis 2016 ; (129) K. L. Luhman 2024 ; (130) K. L. Cruz et al. 2018 ; (131) K. L. Luhman 2023 ; (132) T. R. Kendall et al. 2003 ; (133) J. K. Faherty et al. 2010 ; (134) N. Phan-Bao et al. 2008 ; (135) E. Softich et al. 2022 ; (136) A. C. Schneider et al. 2016b ; (137) K. L. Luhman 2025 ; (138) S. A. Hurt et al. 2024 ; (139) K. M. Aller et al. 2016 ; (140) K. L. Luhman et al. 2018 ; (141) G. Bihain et al. 2010 ; (142) T. L. Esplin et al. 2017 ; (143) J. D. Kirkpatrick et al. 2001 ; (144) J. K. Faherty et al. 2009 ; (145) N. Lodieu et al. 2014 ; (146) X. Fan et al. 2000 ; (147) J. R. Thorstensen & J. D. Kirkpatrick 2003 ; (148) H. Bouy et al. 2022 ; (149) J. E. Gizis 2002 ; (150) M. B. Stumpf et al. 2009 ; (151) A. J. Burgasser et al. 2009 ; (152) M. C. Cushing et al. 2018 ; (153) K. M. Aller 2016 ; (154) Z. H. Zhang et al. 2010 ; (155) K. L. Luhman 2006 ; (156) E. L. Martín et al. 2020 ; (157) J. I. Canty et al. 2013 ; (158) N. Lodieu et al. 2011 ; (159) P. Chinchilla et al. 2020 ; (160) C. Alves de Oliveira et al. 2013 ; (161) E. L. Martín et al. 2000 ; (162) M. R. Zapatero Osorio et al. 2000 ; (163) S. L. Folkes et al. 2012 ; (164) E. L. Martín et al. 2010 ; (165) F. Baron et al. 2015 ; (166) T. R. Geballe et al. 2002 ; (167) A. A. West et al. 2008 ; (168) W. M. J. Best et al. 2017b ; (169) T. R. Kendall et al. 2007a ; (170) F. Marocco et al. 2020 ; (171) T. Ravinet et al. 2024 ; (172) C. A. Theissen et al. 2017 ; (173) V. Almendros-Abad et al. 2022 ; (174) N. Lodieu et al. 2021 ; (175) A. J. Burgasser et al. 2016 ; (176) S. Salim et al. 2003 ; (177) E. L. Martín et al. 1999 ; (178) A. C. Schneider et al. 2019 ; (179) J. K. Faherty et al. 2012 ; (180) N. Lodieu et al. 2005 ; (181) J. C. Wilson et al. 2003 ; (182) D. P. Schneider et al. 2002 ; (183) A. J. Burgasser et al. 2007 ; (184) J. E. Gizis et al. 2000 ; (185) K. L. Luhman et al. 2012 ; (186) S. O. Kepler et al. 2021 ; (187) R. Ahumada et al. 2020 ; (188) W. J. Cooper et al. 2024 ; (189) S. J. Schmidt et al. 2015 ; (190) C. Cabello et al. 2019 ; (191) N. Skrzypek et al. 2016 ; (192) C. Reylé 2018 ; (193) K. L. Luhman et al. 2017 ; (194) Z. Zhang et al. 2018 ; (195) P. Dawson et al. 2014 ; (196) K. L. Luhman & T. L. Esplin 2020 ; (197) K. N. Allers & M. C. Liu 2020 ; (198) Z. H. Zhang et al. 2017 ; (199) Z. Zhang 2019 ; (200) A. A. West et al. 2011 ; (201) A. J. Burgasser et al. 2004 ; (202) J. D. Kirkpatrick et al. 1995 ; (203) M. R. McGovern et al. 2004 ; (204) M. T. Ruiz et al. 1991 ; (205) N. Phan-Bao et al. 2006 ; (206) J. E. Gizis et al. 2013 ; (207) R. D. Scholz et al. 2004 ; (208) J. D. Kirkpatrick et al. 1997 ; (209) A. S. Rajpurohit et al. 2013 ; (210) A. G. Stahl et al. 2022 ; This table is available in its entirety online; the full table contains 16917 rows.