The IGRINS YSO Survey III: Stellar parameters of pre-main sequence stars in Ophiuchus and Upper Scorpius

Our temperatures follow the same trend with spectral type as we found for single Taurus YSOs. To improve our previous IGRINS temperature scale, we incorporated the $T_{\rm eff}$ values determined for Ophiuchus and Upper Scorpius YSOs. To construct the IGRINS temperature scale, we computed the mean $T_{\rm eff}$ and the standard deviation in bins of $\pm$0.5 spectral type subclasses. We included in our IGRINS temperature scale spectral type bins with more than one object.

In Figure 3 we compared the updated IGRINS temperature scale, which is in Table LABEL:tab:scale, to the published scales of pecaut13, herczeg14 and luhman03b. In general, the shape of our temperature scale is similar to the published ones. Our $T_{\rm eff}$ values for M0-M3 stars agree with the three temperature scales. However, for late spectral, our scale agrees better with the luhman03b. For the K stars, we found that our temperature values are cooler than the temperature scales of pecaut13 and herczeg14.

The differences between our $T_{\rm eff}$ values and the K stars temperatures scales of pecaut13 and herczeg14, could be explained as a cumulative result of using different atmospheric models or line lists, different methodologies (spectroscopy vs. photometry), and different wavelength intervals (optical vs. infrared).

Regarding the $T_{\rm eff}$ distributions of the Ophiuchus and Upper Scorpius, we found a Kolmogorov-Smirnov⁴⁴4The Kolmogorov-Smirnov (KS) test computes the probability that two populations come from the same parent distribution. (KS) probability of $\sim$10%, meaning that at this level, both samples are statistically the same in terms of $T_{\rm eff}$. Is this KS too low? Why? Likely because we have just 13 USco objects?

Based on the findings of hussaini20 and the IGRINS spectral resolution (see Paper I for more details), we define those B $\geq$ $v\sin i$/8, and no lower than 1.0 kG, as a B determination. We considered those B values that do not meet these criteria as upper limits. We found that 7 and 20 B field values meet the detection threshold in the Upper Scorpius and Ophiuchus star-forming region, respectively. Both determinations and limits are reported in Table 1. The KS probability computed between the two samples of B field determinations is 50%.

Regarding our $v\sin i$ determinations, we found a mean value and a standard deviation of the mean of 14.8$\pm$1.7 and 14.1$\pm$ 0.9 km s^-1 for the members of Upper Scorpius and Ophiuchus, respectively. We also found a KS probability of 70% that these two $v\sin i$ populations come from the same parent distribution. This probability suggests no evidence of a rotational evolution between the Ophiuchus and Upper Scorpius members. However, our Upper Scorpius sample has only thirteen objects, preventing us from giving a more robust conclusion.

rebull18 analyzed light curves of K2 Campaign 2 of candidate members of Ophiuchus and Upper Scorpius to determine rotation periods ($P_{\rm rot}$). We have 18 and 9 Ophiuchus and Upper Scorpius YSOs in our sample with values of $P_{\rm rot}$ determined by rebull18.

Combining our $v\sin i$ values with Rebull et al.’s $P_{\rm rot}$ allows establishing lower limits to the stellar radius as:

where $R\sin i$ is in solar radii, $v\sin i$ in km s^-1, and $P_{\rm rot}$ in days.

We computed the $R\sin i$ for the 27 objects in our sample with $P_{\rm rot}$ from rebull18 and we propagated the $v\sin i$ uncertainty to obtain an error on $R\sin i$. The $R\sin i$ as well as $P_{\rm rot}$ are reported Table 1.

To look for trends between the stellar radii and the stellar parameters, we compared in Figure 4 these two values as well as calculated the Pearson correlation coefficient.

We did not find any trend with $\log$ g  but we found weak trends with $v\sin i$ and K-band veiling as their Pearson correlation coefficient of 0.24 and 0.26 showed. On the other hand, we found a moderate correlation with $T_{\rm eff}$ as we computed a coefficient of 0.36. Finally, we found a Person’s coefficients of -0.47 when compared stellar radii B-field. The moderate correlations we found between the stellar radii and $T_{\rm eff}$ and B are identifiable in Figure 4.

The spectroscopic Hertzsprung-Russell diagram (sHR; langer14), which is the $\log$ g–$T_{\rm eff}$ plane, helps estimate stellar properties such as age and mass.

In Figure 5 we construct the sHR of our Ophiuchus and Upper Scorpius samples. We also included the Paper I results for the TW Hydrae association ($\sim$ 7–10 Myr; ducourant14; herczeg15; sokal18), and Taurus-Auriga ($\sim$1-5 Myr; kraus09; gennaro12).

The Ophiuchus sample has a mean $\log$ g (and error on the mean) of 3.83$\pm$0.03 dex, while we found 4.13$\pm$0.08 dex for the Upper Scorpius YSOs. On the other hand, the Taurus and TWA YSOs have a mean $\log$ g of 3.87$\pm$0.03 dex and 4.22$\pm$0.03 dex, respectively.

Our mean $\log$ g values show that, in terms of $\log$ g, Ophiuchus is similar to Taurus, while Upper Scorpius is similar to TWA. The KS probabilities computations also confirm the latter. We found KS probabilities of 78% and 35% when we compared the $\log$ g distributions of Taurus and Ophiuchus, and Upper Scorpius and TWA, respectively. Then the probability drops to values smaller than 0.5% when we compared Ophiuchus with TWA and Upper Scorpius with Taurus. Even Ophiuchus and Upper Scorpius are different in terms of $\log$ g as the KS probability of 0.1% shows.

As in Paper I, we make a rough estimate of the age of Ophiuchus and Upper Scorpius, using their mean value of $\log$ g and $T_{\rm eff}$, and the set of isochrones from baraffe15. Our calculations lead to a Ophiuchus mean age of $\sim$2.5 Myr and a Upper Scorpius mean age of $\sim$6.3 Myr.

Previous studies (greene95; luhman99; wilking05; erickson11; esplin20) have determined ages for the Ophiuchus L1688 cloud ranging from 0.1 Myr to 4 Myr. The mean age we found for the Ophiuchus sample, which conforms mostly by members of the L1688 cloud, agrees with the age range previously published.

On the other hand, the age of $\sim$6.3 Myr we found for the Upper Scorpius objects also agrees with the $\sim$5 Myr derived by preibisch02 and slesnick08, and the $\sim$10$\pm$3 Myr of pecaut16.

Although the mean age values that we determined for the Upper Scorpius and Ophiuchus YSOs agree with previous determinations, our simple age pretends to quantify, in a relative way, the age of both samples and not to provide an accurate age determination for each region.

The homogeneous stellar parameter determination that we perform for the Ophiuchus, Upper Scorpius, Taurus, and TWA YSOs allows us to analyze the similarities and differences of these samples from an evolutionary point of view.

Based on the shape of their spectral energy distributions, lada84 divided the YSOs into three different morphological classes (i, ii, and iii), by means of the spectral index $\alpha$:

The YSOs class i (protostars) are objects embedded in an infalling cold dust envelope. Such envelope dominates their emission, making them almost undetectable at optical wavelengths. Then, the initial infall stage ends, leaving a residual envelope and an accretion disk around the YSOs (class ii), which at this stage become visible at optical wavelengths but with infrared excesses. Finally, the accretion disk and the residual envelope dissipate, revealing the class iii YSOs isolate pre-main-sequence stars, possibly with planetary or stellar companions. Class iii YSOs have not (or barely have) infrared excesses and are considerably similar to the main-sequence stars. With the years, three more classes have been added to the latter scheme: Class 0 YSOs (objects that are detectable just at sub-millimeter wavelengths andre93), flat spectrum YSOs (greene94) and objects with optically thick disks (transitional disks, lada06).

The classes scheme represents an evolutionary sequence of low-mass stars, and it is helpful to understand the star formation process. Although most of our YSOs have a previous classification, we computed $\alpha$ indices of all our YSOs to investigate trends between the evolutionary stage of the stars and their stellar parameters homogeneously.

We first construct the infrared spectral energy distribution (SED) to compute the spectral indices. We gathered from the ALLWISE catalog (cutri13) the 2MASS ($J$, $H$, and $K_{s}$) and WISE ($W1$, $W2$, $W3$, and $W4$) magnitudes. We included those magnitudes with a flux signal-to-noise ratio greater than two and ignore the upper limits.

We first obtained the visual line-of-sight extinction ($A_{v}$) toward each YSO position from The Galactic Dust Reddening and Extinction service⁵⁵5https://irsa.ipac.caltech.edu/applications/DUST/. This service gives the Galactic dust reddening for a line of sight, and an estimation of the total Galactic visual extinction from schlegel98 and schlafly11. In this work, we used the schlafly11 optical extinction. We then corrected all the SEDs for extinction using the extinction laws reported by groschedl19.

We converted the obtained $A_{V}$ to the extinction in the $K_{\rm s}$ band ($A_{K}{}_{\rm s}$) by means of $A_{K}{}_{\rm s}$/A_V = 0.112 (rieke85; groschedl19) and we employed the zero magnitude flux densities and extinction laws reported in Table 2 to obtained the de-reddened fluxes. Using the de-reddened fluxes and the central wavelengths of each band, we created the infrared SED of each YSO in our sample.

Finally, we computed two different de-reddened $\alpha$ indices from the SEDs. The first one is a linear fit between the de-reddened $K_{\rm s}$ and $W4$ magnitudes ($\alpha_{K_{s}-W4}$), while the second linear fit includes only the WISE de-reddened magnitudes ($\alpha_{\tt WISE}$). Both $\alpha$ indices should classify our sample in the same way. To check this, we compared these two indices in Figure 6.

In general, we have good agreement between low values of $\alpha_{K-W4}$ and $\alpha_{\tt WISE}$. This agreement slightly worsens toward $\alpha$ indices greater than $\sim-1.5$. Still, this difference, which in the worst case is about 0.5, does not affect the YSO classification, as with both spectral indices, we recover the same Class designation.

According to groschedl19, the Class iii YSOs are those with $\alpha\leq-1.6$, the Class ii have $\alpha$ values between -1.6 and -0.3, the Flat spectrum category comprises YSOs with -0.3 $\leq\alpha\leq$ 0.3, and finally the Class 0 + i presents $\alpha>0.3$. The computed spectral indices and their corresponding classification are reported in Table 1.

Additionally, we compared, in Figure 7, our spectral indices with those determined by dunham15 and luhman10 for YSOs in the Ophiuchus and Taurus star-forming region, respectively. dunham15 determined the spectral index $\alpha$ through a linear least-squares fit of all available 2MASS and Spitzer photometry between 2 and 24 $\mu$m. On the other hand, luhman10 computed the spectral index between four pairs of photometric bands, including the pair $K_{s}$ and Spitzer 24 $\mu$m. These spectral indices are compatible with those determined here, as we obtained them from the same wavelength range (2 - 24 $\mu$m), and they were de-reddened. The only difference between the works of dunham15 and luhman10, and ours is that we used WISE photometry instead of Spitzer.

In general, our classifications are very consistent with the literature. For the Taurus $\alpha$ indices, we found that our $\alpha_{K_{s}-W4}$ values are around the one-to-one line for indices greater than -1.6, but this agreement worsens for indices lower than this value. This discrepancy might be because the luhman10 indices piled up around $\sim$-2.8. why Luhman values piled up?. On the other hand, most of our Ophiuchus $\alpha$ indices are larger than those determined by dunham15, on average, $\sim$0.12. not sure why. Offset due to use WISE instead of Spitzer? Different Av?

Once we homogeneously computed the infrared spectral $\alpha$ indices, we looked for trends between them and the determined stellar parameters. In Figure 8, we present our stellar parameter determinations as function of their respective $\alpha_{K_{s}-W4}$. The values of $\alpha_{K_{s}-W4}$ and $\alpha_{\rm WISE}$ are very similar, as seen in Figure 6, therefore using one over another does not affect our findings.

We did not find any correlation between $\alpha_{K_{s}-W4}$ and the $T_{\rm eff}$ values determined here. The dispersion of the $T_{\rm eff}$ values in the four populations are very similar as well as between Class ii and Class iii YSOs. Regarding $\log$ g, we found that, on average, $\log$ g values of Class iii are higher than their Class ii counterparts. We found a mean and standard deviation of the mean for Class ii YSOs of 3.87$\pm$0.03 dex, while for Class iii we found 4.04$\pm$0.04 dex.

Similarly, we found that $v\sin i$ values of Class iii are slightly higher than the values of Class ii. We found a mean value of 15.2$\pm$0.7 km s^-1 and 18.0$\pm$1.2 km s^-1 for the Class ii and Class iii YSOs, respectively.

We consider actual determinations for the B field analysis. We have 74 determinations in total, of which 20, 7, 8, and 39 are from Ophiuchus, Upper Scorpius, TWA, and Taurus members, respectively. We did not find any trend with the $\alpha_{K_{s}-W4}$ values. Both Class ii and Class iii YSOs have B fields very similar between them. We found a mean value of 1.97$\pm$0.06 kG for Class ii and 2.11$\pm$0.13 kG for Class iii YSOs, respectively. I should expect younger objects being more active

The most notorious and expected trend of $\alpha_{K_{s}-W4}$ arises with the K-band veiling values. More evolved objects have lower K-band veiling values. Interestingly, the K-band veiling values show a large dispersion for Class ii YSOs. Such dispersion could be related to the variable nature of the Class ii YSOs or with intrinsic differences between the YSOs, such as stellar mass.

To assess the effect of age on the $\log$ g  we divided our star-forming regions into three different temperature (mass) bins.

In this $T_{\rm eff}$ bin, there are 41 YSOs, from which 18, 2, 13, 8 are located in Ophiuchus, Upper Scorpius, Taurus, and TWA, respectively. As we have two stars within this $T_{\rm eff}$ bin located in Upper Scorpius, we excluded Upper Scorpius from the following analysis. We found a mean $\log$ g and standard error on the mean of 3.71$\pm$0.05, 3.69$\pm$0.08, and 4.23$\pm$0.05 dex for Ophiuchus, Taurus, and TWA YSOs. We encountered that Ophiuchus and Taurus YSOs have a KS probability of coming from the same parent distribution of 45%, while this probability decreases to $\ll$ 1% when comparing TWA to Ophiuchus.

This $T_{\rm eff}$ bin is the most populated with 90 YSOs. We have 15, 4, 62, and 9 in Ophiuchus, Upper Scorpius, Taurus, and TWA, respectively. We found a mean $\log$ g of 3.90$\pm$0.04, 4.16$\pm$0.02, 3.84$\pm$0.03, and 4.17$\pm$0.02 dex for Ophiuchus, Upper Scorpius, Taurus, and TWA. We found a KS probability of 51% when we compared Ophiuchus and Taurus $\log$ g distributions. This probability value is similar to the KS probability found when comparing TWA and Upper Scorpius, which is of 59%.

This $T_{\rm eff}$ bin is the most populated with 50 YSOs. We have 14, 4, 31, and 1 in Ophiuchus, Upper Scorpius, Taurus, and TWA, respectively. We found a mean $\log$ g of 3.90$\pm$0.08, 3.93$\pm$0.13, and 3.94$\pm$0.04 dex for Ophiuchus, Upper Scorpius, and Taurus. We found a KS probability of 98% when we compared Ophiuchus and Taurus $\log$ g distributions. This probability is about $\sim$ 50% when comparing Upper Scorpius to Ophiuchus or Taurus.