Discovery of CWISE J052306.42−015355.4, an Extreme T Subdwarf Candidate

The CatWISE2020 W1 and W2 magnitudes for J0523−0153 are 17.27 ± 0.06 mag and 15.91 ± 0.06 mag, respectively, resulting in a color of W1 – W2 = 1.36 ± 0.08 mag. This object was not detected in the 2MASS survey (Skrutskie et al. 2006). We then searched the VISTA Science Archive (Cross et al. 2012) and WFCAM Science Archive (Hambly et al. 2008). A deep J-band detection was found on the VISTA Hemisphere Survey as shown in Figure 2, with J_MKO = 19.75 ± 0.186 (mag).

Zoom In Zoom Out Reset image size

With the W1 − W2 color and J − W2 color, we can plot J0523−0153 on a color–color plot like that of Figure 8 of Kirkpatrick et al. (2016). This is illustrated in Figure 3, showing that the J − W2 color is extremely red for its W1 − W2 color, compared to the location of normal brown dwarfs. Such colors imply an extreme T subdwarf as defined by Meisner et al. (2021). As a result of there being no spectral type versus color relations for esdTs yet, we approximate a mid-T type using the relation for normal T dwarfs, as done in Meisner et al. (2021) and Schneider et al. (2020). We provide more context with other known and suspected T subdwarfs in Section 5.1.

Zoom In Zoom Out Reset image size

However, when we look at Figure 8(i) from Zhang et al. (2019) at a fixed spectral type, sdTs are redder in W1 − W2 compared to normal T dwarfs. If we assume a similar behavior for esdTs then we would expect J0523−0153 to be an early esdT. This is also supported by WISEA J041451.67−585456.7 (hereafter, J0414−5854; Schneider et al. 2020), which has similar colors (Table 2 and Figure 3) to J0523−0513 and has a tentative spectroscopic classification of esdT0 ± 1. Thus, a slightly preferred estimation for the spectral type of J0523−0512 is ∼esdT0.

As there is no confirmed esdT that has a measured trigonometric parallax, we have no clear way to estimate the distance. In Figure 7 of Meisner et al. (2021) the LOWZ models show that W1 − W2 is not as temperature sensitive as J − W2 at colors similar to esdTs. A better indicator of temperature, or spectral type, would be obtained from the J − W2 color. Comparing our value of J − W2 = 3.84 ± 0.195 mag to Figure 20(c) of Kirkpatrick et al. (2021a), we obtain M_W2 ≈ 14.0 using Figure 15 of the same paper, where they equate Spitzer ch2 photometry with WISE W2 photometry, which results in a distance of ≈24 pc.

If we use the standard W1 − W2 versus spectral type relation of Kirkpatrick et al. (2021a), which applies only to solar metallicity brown dwarfs, we estimate a T5 spectral type. From Figure 16(d) of Kirkpatrick et al. (2021a) we find that a regular T5 has an M_W2 ≈12.5, which results in a distance of ≈48 pc. However, it is much more likely that J0523−0153 is an esdT0; therefore, using the same process as before we get M_W2 ≈ 11.75, which results in a distance of ∼68 pc.

Another way to estimate the distance of J0523−0153 is by looking at Figure 20(c) of Kirkpatrick et al. (2021a). We see that sdTs tend to be brighter in M_ch2. If we plot J0523−0153 brighter, at about the distance that sdTs are removed from the regular sequence, we estimate M_ch2 ≈ 13. This results in a distance of ≈38 pc.

We can also estimate other distance limits using the measured proper motion. As Kirkpatrick et al. (2021b) shows, the highest velocity of any brown dwarf in the 20 pc sample is ∼200 km s⁻¹. If we instead use this value as a reasonable upper bound, then we find that J0523−0153 most likely has a distance <84 pc. Even though this is a very loose constraint, it is in agreement with our previous estimates.

As a result of the total velocity distance estimate being a rough upper bound, we revert to our previous estimate for esdT0's of ≤68 pc for J0523−0153. A measured trigonometric parallax is needed to better understand this, and other, esdT candidates.

J0523−0153 is shown in Figure 1 to have a significant proper motion, and this is measured in the CatWISE2020 Catalog as μ_R.A. = 0.52 ± 0.05'' yr⁻¹, μ_decl. = −0.04 ± 0.06'' yr⁻¹, and μ_tot = 0.52 ± 0.08'' yr⁻¹. Using our poorly constrained distance estimate of ≤68 pc, we find a tangential velocity of ≤167 km s⁻¹.

Based on the location of J0523−0153 in the color–color plot of Figure 3, we can attempt to deduce its metallicity. If we plot it on Figure 7 of Meisner et al. (2021), which is similar to Figure 3 here, we find that it falls along the LOWZ model with metallicities of [M/H] = −2. However, WISEA J153429.75−104303.3 (hereafter, J1534−1043; Meisner et al. 2020 and Kirkpatrick et al. 2021b) falls in a far more unusual spot on the color–color diagram and is believed to have a metallicity [M/H] ≈ −1.5 (Kirkpatrick et al. 2021b), more metal-rich than the LOWZ model predicts.

For additional guidance, we can examine the location of WISE J200520.38+542433.9 (hereafter, J2005+5424; Mace et al. 2013) on Figure 3, since this object is a co-moving companion to a star with a known subsolar metallicity of [Fe/H] = −0.64 ± 0.17 dex. If we use the LOWZ model, it predicts a subsolar metallicity of [M/H] = −1.0. However, the actual measured metallicity from the primary of the system is higher than the prediction adopting the LOWZ model. Thus, the models are making predictions that are systematically too metal-poor. The location of J2005+5424 on Figure 3 is closer to the normal T dwarf sequence than that of J0523−0153, so this means that this object would have a higher metallicity. But J1534−1043 is further removed from the normal sequence than both of these objects. Since we believe that J0523−0153 is less metal-poor than J1534−1043 and more than J2005+5424, then it most likely has a metallicity in the range of −1.5 < [M/H] < −0.5.

The first extreme T subdwarf candidates were discussed in Schneider et al. (2020). Both J0414−5854 and J1810−1010 have spectroscopic signatures indicative of subsolar metallicity, which identified them as T subdwarfs. The authors came to this conclusion using their Figure 2, which compares the spectra of these objects to the spectra of known dTs and sdTs. Accordingly, their Figure 4 compares the spectra to model predictions in both temperature and metallicity.

J0414−5854 and J1810−1010 were both shown to have similar unusual colors compared to known brown dwarfs, as shown in Table 2. As shown in Figure 3 of Schneider et al. (2020), middle to late esdTs are bluer in W1 − W2 and redder in J − W2, J − K, and J − H when compared to regular dTs.

Both CWISEP J050521.29−591311.7 and CWISEP J090536.35+740009.1 (hereafter, J0505−5913 and J0905+7400) were discussed in Meisner et al. (2020) as mid-T dwarf candidates, even possibly T subdwarfs. We would now list both objects as mid-esdT candidates based on their colors. These magnitudes, found in Table 2, are very similar to other esdTs that are found in the esdT box in Meisner et al. (2021), shown in Figure 3. They estimated that J0505−5913 is a T6.5 and J0905+7400 is a T6, assuming both are normal T type dwarfs. With ch1−ch2 data from Spitzer, they find distance estimates for J0505−5913 and J0905+7400 of 34.9−46.3 pc and 41.6−55.1 pc, respectively.

Meisner et al. (2021) discussed two new esdT candidates, CWISE J073844.52−664334.6 (hereafter, J0738−6643) and CWISE J221706.28−145437.6 (hereafter, J2217−1454). J0738−6643 was found to have characteristics of a T5 ± 1.2. J2217−1454 has a spectral type estimate of T5.5 ± 1.2, these estimates came from the W1 − W2 versus brown dwarf spectral type, which has been shown to be problematic. This is similar to J0738−6643 and comparable to J0523−0153, a reminder that these values are poor estimates given that these objects are likely not normal T type dwarfs.

The most exotic of the extreme subdwarf candidates is WISEA J153429.75−104303.3 (hereafter, J1534−1043; Meisner et al. 2020 and Kirkpatrick et al. 2021b). J1534−1043 was so faint in the near-infrared that those authors needed to use the Hubble Space Telescope to obtain a detection of it. Even with this hurdle, those authors show that J1534−1043 is the reddest of all esdTs in J − W2, including candidates, and is possibly an esdY. Even though it has a red J − W2, it still falls within the W1 − W2 color constraint of the esdT box of Figure 3. In Section 4.5 we discussed how this object compares to J0523−0153 in metallicity. There is a temperature estimate of T_eff < 500 K using the theoretical LOWZ model, but these predictions have not been tested yet.

When looking at J0523−0153 we can see that it has many neighbors in Figure 3. Peering at confirmed/candidate esdTs that have similar colors to J0523−0153, we can conclude that they likely have similar characteristics. For example, as both J0738−6653 and J0523−0153 have similar colors, we can assume they have comparable spectral types. Likewise, the metallicities that we estimated are comparable to similarly colored esdTs that have spectra, with which a more accurate metallicity can be estimated.