betaFeedback

WorksPhysics and Astronomy

All-sky spectroscopic survey of bright δ Scuti stars from TESS Projected rotational velocities and pulsation characteristics

Abstract

Context. The second most numerous class of pulsating stars – δδ Scuti (δδ Sct) – lack a homogeneous, large-scale spectroscopic sample in the Kepler and TESS era, limiting our understanding of how rotation shapes stellar pulsations. Aims. We analyse projected rotational velocities (vsin⁡iv \sin i) from high-resolution spectroscopy alongside 2-minute TESS photometry to establish empirical links between stellar rotation and pulsation parameters. Methods. Combining observations from three high-resolution échelle spectrographs with two spectral libraries, we assemble an all-sky sample of 262 δδ Sct stars (R=25 000R = 25\,000 to 115 000115\,000). We derive vsin⁡iv \sin i via three independent methods and extract pulsation frequencies and amplitudes from TESS light curves. Results. The vsin⁡iv \sin i distribution is unimodal, with a mean of 120.3 km s−1120.3\text{ km s}^{-1} and peak probability at 113.8 km s−1113.8\text{ km s}^{-1}. Minimum (νminν_{\rm min}), dominant (νdomν_{\rm dom}), amplitude-weighted (νAWν_{\rm AW}), and power-weighted (νPWν_{\rm PW}) frequencies decrease with vsin⁡i/vcritv \sin i / v_{\rm crit}, whereas maximum frequency (νmaxν_{\rm max}) does not. Fast rotators show a narrower range of photometric and pulsation amplitudes than slow rotators, and their lower mean stellar density (ρˉ\barρ) cannot fully explain the frequency decrease. Calibrating Gaia's vbroadv_{\rm broad} against our measurements removes its systematic offset, enabling use across larger δδ Sct samples. Multi-spectrum comparisons reveal method discrepancies mainly for fast rotators (vsin⁡i>200 km s−1v \sin i > 200\text{ km s}^{-1}) and inter-spectrograph offsets up to 20 km s−120\text{ km s}^{-1}. Conclusions. We present the largest homogeneous high-resolution vsin⁡iv \sin i sample for bright δδ Sct stars in the TESS era, showing that vsin⁡iv \sin i measurements in pulsating stars require careful treatment across instruments and velocity regimes.

Introduction

The pulsating stars of δ\delta Scuti (δ\delta Sct) are situated where the classical cepheid instability strip crosses the main sequence; their cores are therefore in the stage of burning hydrogen. This population of pulsating stars of early A- to late F-type has masses in a range of 1.5 M⊙1.5\,M_\odot to 2.3 M⊙2.3\,M_\odot [76], temperatures in a range of 6000 K to 9000 K (Ytterhoeven et al. 2011; Kurtz 2022), luminosities of 2 to 50 L⊙L_\odot, and frequencies between 4 d−1^{-1} [27, 2] and 70 d−1^{-1} [21]. The frequency of 4 d−1^{-1} is commonly used as a distinction between δ\delta Sct (>4> 4 d−1^{-1}) and γ\gamma Doradus (γ\gamma Dor) stars (<4< 4 d−1^{-1}), although many authors adopt 5 d−1^{-1} as a dividing threshold (Handler & Shobbrook 2002; Grigahcène et al. 2010; [15, 112, 100, 99].

Generally, both radial and non-radial pressure (p) mode pulsations exist among δ\delta Sct stars. Some of these stars can have hundreds of frequencies that require p, gravity (g), and Rossby (r) modes [95, 55], rotational splitting, nonlinear combinations of frequencies, as well as additional complexities such as variable amplitudes on longer timescales [61]. Alternatively, they can have ridge-like frequency clusters, as studied by Antoci et al. [7]. The complexity of their pulsation spectra makes the identification and interpretation of individual modes challenging, particularly in hybrid pulsators [53, 62, 116]. A large fraction of δ\delta Sct stars have frequencies also in the γ\gamma Dor range [112, 12, 14, 100]. Pulsation modes are primarily excited by the classical κ\kappa mechanism operating in He ii ionisation zones [35, 32, 2]. Antoci et al. [5] showed that the opacity mechanism alone cannot explain oscillations, and turbulent pressure in the hydrogen ionisation zone also contributes to the excitation of pulsations, mainly moderate radial p modes.

The last two decades have seen enormous developments in asteroseismology, thanks to the launch of space missions CoRoT [11], the Kepler space telescope [25], and most recently, the Transiting Exoplanet Survey Satellite (TESS, Ricker et al. [88]). Space photometry provides precision about 100 times better than ground-based photometry, with nearly continuous, evenly sampled data [61]. This precision and the sheer number of objects observed have increased the number of known pulsating stars and allowed us to probe their interiors. Building on the Kepler era, TESS brought the possibility of ensemble studies of δ\delta Sct stars across the whole sky, see e. g. Antoci et al. [6]; Cunha et al. [36]; Bedding et al. [21]; Barac et al. (2022); Gootkin et al. [50]. Another advantage of TESS is its long-term observational coverage, with observations spanning multiple sectors and, for some targets, several years. Two-minute cadence data are also available for many stars. Additionally, TESS all-sky coverage includes many early-type hotter stars not observed by Kepler, along with bright stars that can be observed by smaller telescopes.

Because of the overabundance of space photometry compared to spectroscopy, most of the recent studies of δ\delta Sct stars rely purely on space photometry, and only a handful of stars have any spectroscopic information. Before the advent of space photometry, Solano & Fernley [107] conducted a survey of the rotational velocities of 68 δ\delta Sct stars and found that stars with low vsin⁡iv \sin i tend to have higher pulsation amplitudes, often being part of a distinct high-amplitude δ\delta Sct (HADS) subclass. The fact that fast rotators pulsate with lower amplitudes was already pointed out by Danziger & Faber [37]. No other dependence on projected rotation has been found [58]. Using more than 4000 δ\delta Sct and γ\gamma Dor stars observed with the medium-resolution LAMOST data (R≈7500R \approx7500), it was discovered that both δ\delta Sct and γ\gamma Dor stars have similar vsin⁡iv \sin i distributions [114].

Recently, multiple studies have looked into the role of stellar rotation in the populations of A- and F-type stars [50, 75, 77, 76] (Berry et al. 2026). Even with the abundance of the Gaia data [45, 46, 43], the effect of projected rotation on pulsation characteristics cannot be fully understood without proper spectroscopic follow-up, as the Gaia broad spectral line broadening parameters are systematically biased [50, 75, 114].

In light of these recent discoveries and the huge amount of precise photometric data, we decided to carry out a comprehensive study of bright, manually vetted δ\delta Sct stars with high-resolution spectroscopy. We aim to understand the influence of projected rotation vsin⁡iv \sin i on stellar oscillations and excited pulsation modes. We also revisit the results of previous studies in the context of high precision photometry from short cadence 2-minute TESS data and search for scaling relations. We observed 262 bright (VV mag <7< 7) stars and gathered 307 high-resolution échelle spectra.

The article is structured as follows: we define the sample and observations in Sect. 2; in Sect. 3, we measure projected rotational velocities vsin⁡iv \sin i using 3 methods; in Sect. 4, we analyse TESS data and extract pulsation frequencies; and we conclude in Sect. 5.

Spectroscopic observations

Sample definition

Our studied sample of δ\delta Sct stars is based on the TESS Input Catalog (TIC, Stassun et al. [109]), which incorporates more than 1.5 billion stars from Gaia DR2 (Gaia Collaboration et al. 2018), available publicly at the Mikulski Archive for Space Telescopes2. We have selected stars with VV mag <7< 7 and a temperature range of 6500 K to 9000 K. Then, we manually inspected the selected objects using a Python script for visual representation based on Skarka et al. [100]; Skarka & Henzl [99]. We visually inspected the candidates and selected stars showing pulsation in the frequency range above 5 d−1^{-1}, while ignoring γ\gamma Dor/δ\delta Sct hybrids with significant peaks below 5 d−1^{-1}. We have also excluded stars with temperatures above 9000 K that show δ\delta Sct-like pulsations above >5> 5 d−1^{-1} to avoid crossing the blue boundary of the instability strip, even though, for example, Balona & Ozuyar [16] use 10 000 K as a boundary between δ\delta Sct and Maia variables.

After visual inspection, we have obtained 435 objects that form our initial sample. Out of this sample, 262 were observed by at least one of the spectrographs at our disposal. Altogether, we observed 307 high-resolution échelle spectra. The distribution of VV magnitudes and effective temperatures TeffT_{\rm eff} of the observed sample can be seen in Figure 1 and the Gaia Hertzsprung–Russell diagram in Figure 2. The sky map of all observed objects can be seen in Figure 3.

Figure 1. Distribution of V magnitudes and effective temperatures TeffT_{\rm eff} from TIC of observed δ\delta Scuti stars.

Figure 2. Gaia Hertzsprung–Russell diagram of the initial sample (gray) and observed sample (colored, 262 stars) of δ\delta Scuti stars. The dashed line is the ZAMS, the black lines are the evolutionary tracks for stars with different masses, the blue and red line show the empirical boundaries of the instability strip all derived by Murphy et al. [78].

Figure 3. δ Scuti stars for which at least one spectrum was obtained.

Instruments

We used three échelle spectrographs for our observations: MUSICOS, OES, and PLATOSpec, together with MELCHIORS and HARPS/ESO spectral libraries.

MUSICOS (Pribulla et al. [83]) is an échelle spectrograph at the 1.3 m telescope at Skalnaté Pleso Observatory (20°14′02′′ E, 49°11′22′′ N), 1786 m above sea level. It has a resolving power of R≈25 000−38 500R \approx25\,000 - 38\,500 and a spectral range of 4250 to 7375 Å covering 65 échelle orders.

Ondřejov Echelle Spectrograph (OES) (Koubský et al. [60]; Kabáth et al. [57]) is an échelle spectrograph installed on a 2 m Perek telescope in Ondřejov (14°47′01.1′′ E, 49°54′38′′ N), 528 m above sea level, 30 km southeast of Prague. It has a resolving power of R≈50 000R \approx50\,000 and covers a spectral range 3750 to 9200 Å.

PLATOSpec (Kabáth et al. [56]) is a high-resolution, blue sensitive spectrograph with a resolving power of R≈70 000R \approx70\,000 and a spectral range of 3800 to 7000 Å, installed on the refurbished ESO 1.52 m telescope (E152)2^{2} at La Silla observatory (70°44′00′′ W, 29°15′00′′ S), 2400 m above sea level. It was built to screen and validate exoplanetary candidates and study stellar variability as a support facility for the PLATO mission (Rauer et al. [85]).

We have also searched the Mercator Library of High Resolution Stellar Spectroscopy (MELCHIORS, Royer et al. [93]) and HARPS spectral databases. The list of spectra is summarised in Table 1.

InstrumentResolution RRRange (Å)No. of spectra
MUSICOS25000–385004250–737572
OES500003750–920075
PLATOSpec700003800–700071
HERMES850003770–900052
HARPS80000/1150003800–680037

Table 1. Journal of spectra.

HERMES (Raskin et al. [84]) is a high-resolution spectrograph installed on 1.2 m Mercator telescope at the Roque de Los Muchachos Observatory on the island of La Palma, Spain (17°52′43′′ W, 28°45′45′′ N), 2331 m above sea level. It has a resolving power of R≈85 000R \approx85\,000 and a spectral range of 3770 to 9000 Å.

HARPS (Pepe et al. [79]) is a high-resolution spectrograph on the ESO 3.6 m telescope at the La Silla observatory, with a primary aim of high precision spectroscopy measurements of exoplanets. It has a spectral range of 3800 to 6800 Å. We have searched the ESO Science Archive and found that 37 objects were also observed by HARPS. We have selected and downloaded one spectrum with the highest S/NS/N ratio if multiple spectra were present. All spectra have a high resolving power of either R≈80 000R \approx80\,000 or 115 000.

Spectra from MUSICOS and OES were reduced and normalised by the standard custom-made observatory pipelines. Observations from PLATOSpec were reduced by the Ceres pipeline (Brahm et al. [26]). Spectra from HERMES were downloaded in an already normalised form, and for HARPS spectra normalisation, we have used an automated neural network normalisation tool SUPPNet (Różański et al. [94]) that is trained to predict a pseudo-continuum, which should lead to results similar to careful manual normalisation.

Measurements of vsin⁡iv \sin i

There are multiple methods to determine projected rotational velocities vsin⁡iv \sin i of stars, and none of these, as pointed out by Gray [51], are standard. Most simple methods measure the direct rotational broadening of a spectral line [101, 102, 103]. Slettebak et al. [104] developed a calibration of rotational velocity based on the full width at half depth of the absorption lines in stellar spectra. In the first study of a homogeneous sample of δ\delta Sct stars, Solano & Fernley [107] determined the rotational velocities of 68 stars using the standard rotational profile by Gray [51]. The introduction of the Fourier transform (FT) on line profiles allows for much deeper analysis. There have been multiple spectroscopic surveys of A and F spectral type stars by Royer et al. (2002a,b, 2007) that used FT of single spectral lines, as it allows for better identification of the rotation profile. All of the previous studies rely heavily on the calibration of single strong lines. Identification of suitable spectral lines is a difficult task, as summarised by Díaz et al. [40], since one has to find

  1. lines that have no blends from surrounding lines,

  2. lines broadened primarily by the effect of rotation, e.g. no Balmer lines as they show a significant Stark effect,

  3. lines that are present for stars from low to critical rotators,

to which we add

  1. lines that are available across a wide temperature range,

  2. lines that are not in close proximity to telluric lines.

Because δ\delta Sct stars are often characterised as mid-to-rapid rotators (Breger [27]; Aerts et al. [2]), estimating their rotational velocities is a challenging task. With high vsin⁡iv \sin i, it is increasingly difficult to find multiple lines that satisfy the conditions mentioned above. We have therefore carried out a threefold measurement of vsin⁡iv \sin i using the following methods

  • method 1 fitting of the single line Fe i 4957.613 Å by the convolution of the Voigt profile and the standard rotational kernel by Gray [51],

  • method 2 fitting the spectral region using iSpec ([24, 23]),

  • method 3 using the Fourier transform of the cross correlation function (CCF) of the observed and non-rotation template spectrum developed by Díaz et al. [40].

Examples of the results for all spectrographs and all three methods can be seen in Appendix A. For method 2 and method 3 we use the same spectral range of 5000 Å to 5800 Å as this region is well covered among all used spectrographs. Results are presented in tbl:2 only available at the Strasbourg Astronomical Data Centre (CDS).

Method 1 - Single spectral line

We fitted the spectral line Fe i 4957.613 Å with the convolution of a Voigt profile and a rotational kernel in the standard form by Gray [51]. A similar method was employed by Solano & Fernley [107], who fitted only the rotational kernel but used multiple spectral lines. Given the different origins of the spectra, we decided to use a single spectral line that is consistently present in all spectra and provides stable results compared to other surveys or different spectral lines. We have used the standard linear limb darkening law (Schwarzschild & Villiger [97])

I(μ)=I(1)(1−ϵ(1−μ))(1)I(\mu) = I(1)(1 - \epsilon(1 - \mu)) \tag*{(1)}

where I(1)I(1) is the stellar intensity at the centre of the disc, μ=cos⁡θ\mu= \cos\theta and θ\theta being the angle between the line of sight and the direction of the emerging flux. We have used ϵ=0.6\epsilon= 0.6 to be consistent with previous studies, e.g. Solano & Fernley [107]. After the fit, we inspected each fitted line and discarded those that clearly did not fit. The advantage of method 1 compared to method 2 and method 3 is that it allows us to derive vsin⁡iv \sin i even for very fast rotators, while the others can sometimes fail for vsin⁡i>250v \sin i > 250 km s−1^{-1}.

Method 2 - Fitting spectral regions

The quality of our spectra allows us to derive projected rotational velocities by fitting large segments of the spectra. We have used iSpec ([24, 23]), which compares observed and synthetic spectra, employing χ2\chi^2 minimisation by executing a nonlinear least-squares method (Markwardt [72]). The code searches the parameter space and performs multiple iterations until χ2\chi^2 is lower than a given threshold. We have processed and corrected the spectrum for systemic velocity using the built-in iSpec templates and renormalised it to one. After this, we used the effective temperature TeffT_{\rm eff} and log⁡g\log g from Gaia DR3 (Gaia Collaboration et al. [46]) and iterated for all the available parameters except for log⁡g\log g, which was kept fixed. We have decided not to iterate log⁡g\log g as it often leads to degeneracies in the model of the spectra, since the spectrum is not strongly dependent on log⁡g\log g compared to other parameters, and it speeds up our computations. Synthetic spectra were computed with the LTE radiative transfer code MOOG (Sneden [106]; Sneden et al. [105]) using ATLAS9 model atmospheres (Castelli & Kurucz [31]) and solar abundances from Grevesse & Sauval [52]. The linear limb darkening coefficient was fixed at ϵ=0.6\epsilon= 0.6 to ensure consistency with method 1 and previous studies. This procedure was performed individually for each spectrum over the wavelength range of 5000 to 5800 Å, as the automatic processing of the spectra would not be reliable for both slow and fast rotators. Examples of the spectra for different vsin⁡iv \sin i with their models are in Figs. A.2 and A.3. Derived values of the vsin⁡iv \sin i and the effective temperature TeffT_{\rm eff} can be found in tbl:2, which is only available at the Strasbourg Astronomical Data Centre (CDS). Obtaining the error of the iSpec model requires uncertainties of spectral flux due to χ2\chi^2 minimisation, which are often either not present or are underestimated or overestimated. As the largest contribution to the error is a coarse grid of synthetic models that do not include gravity darkening, we determined the error using a Monte Carlo method by perturbing basic atmospheric and stellar parameters according to a Gaussian distribution from their expected values. The errors are then estimated as the standard deviation from the statistical dispersion across 100 iterations. We have performed a Monte Carlo analysis for a small subset of spectra from each spectrograph and calculated the mean uncertainty. This statistical error is close to 100 K, and we use it for all spectra. However, as seen in the bottom of Figure 4, different spectrographs can give different values for the effective temperature. The standard deviation for different instruments is ≈ 300 K, which might be closer to the actual estimate of the temperature error. The difference in temperature from different spectrographs may come from poor normalisation, but the actual change in the stellar photosphere may also play a role since we are dealing with pulsating stars. The biggest outlier, TIC 405410684, is also a fast rotator, where finding a best fit may be uncertain.

Figure 4. Top: Comparison of the effective temperature Teff,TICT_{\mathrm{eff,TIC}} from TIC catalogue and values derived by iSpec Teff,specT_{\mathrm{eff,spec}}. Dashed gray is the 1:1 line. Right: histogram of differences with mean μ=−41\mu= -41 K (red line) and standard deviation σ=278\sigma= 278 K (dashed gray lines). Bottom: Comparison of effective temperatures TeffT_{\mathrm{eff}} derived by different spectrographs. Dashed gray line is the 1:1 line.

Method 3 - Fourier method for spectral region

Díaz et al. [40] developed a method for finding vsin⁡iv \sin i in Fourier space using the CCF of observed spectra and a non-rotation template, which has the advantage of not requiring single spectral lines. We derived vsin⁡iv \sin i using a version of this method implemented in Python, which requires an object and a template spectrum. We used template spectra that have been specifically created on a star-by-star basis by using the derived parameters in iSpec (Blanco-Cuaresma et al. [24]; Blanco-Cuaresma [23]) from Sect. 3.2, while assuming slow rotation of 1 km s−1^{-1}. The whole method can be summarised in the following stages

  • calculation of the no rotation template spectrum for each object,

  • calculation of the object×template and template×template CCF,

  • removal of the CCF maximum from the template×template CCF by linear interpolation,

  • decontamination of the object×template CCF to obtain the cleaned rotational broadening profile,

  • measurement of the first zero of the FT,

  • converting the first zero σ1\sigma_1 of the FT into vsin⁡iv \sin i using the formula vsin⁡i=kn(ϵ)σnv \sin i = \frac{k_n(\epsilon)}{\sigma_n}, where kn(ϵ)k_n(\epsilon) can be approximated by the formula kn(ϵ)=0.60975+0.0639ϵ+0.0205ϵ2+0.021ϵ3k_n(\epsilon) = 0.60975 + 0.0639\epsilon+ 0.0205\epsilon^2 + 0.021\epsilon^3, and σn=σ1\sigma_n = \sigma_1 is the first zero of the FT (Díaz et al. [40]).

This was repeated for multiple overlapping wavelength regions with a length of about 220 Å in the interest region of 5000 to 5800 Å. The final value of vsin⁡iv \sin i is the mean of the vsin⁡iv \sin i of selected regions. The estimate of its error is taken from the standard deviation of vsin⁡iv \sin i of selected regions. We have also assumed a fixed limb darkening coefficient ϵ=0.6\epsilon= 0.6, consistent with method 1 and method 2, Solano & Fernley [107] and Royer et al. [90, 92] and excluded the region 516 to 519 nm, due to blended lines and the Mg I b triplet. We have also assumed that the rotational profile is symmetric, which might not always be the case and neglected the impact of non-radial pulsations on the spectral line profiles (Aerts et al. [3]).

Comparisons

The comparison of all three methods is shown in Figure 5, subplots (a), (b) and (c). Methods are compatible in the low to medium range, but they start to differ at high vsin⁡iv \sin i (>200> 200 km s−1^{-1}). Deviation at high vsin⁡iv \sin i cannot be explained by the spectral resolution, as both high and medium-resolution spectrographs deviate in the same direction in subplots (a) and (b). The differences between spectrographs can also be seen, for example, OES (yellow points) drifts the most in all 3 subplots. In subplots (b) and (c), OES drifts in the opposite direction Δ=0\Delta= 0 compared to HARPS, probably caused by different resolution, instrumental profile, continuum normalisation, or the handling of microturbulence. HARPS, which has the highest resolution, deviates already in the range of 100 to 150 km s−1^{-1} in subplot (c) while in subplot (a) it seems stable compared to other spectrographs.

Figure 5. Left: Comparison of vsin⁡iv \sin i measured with Fourier method and derived by fitting spectra in iSpec. Middle: Comparison of vsin⁡iv \sin i measured with Fourier method and by the standard rotational profile. Right: Comparison of vsin⁡iv \sin i derived by fitting spectra in iSpec and single line fitting of the standard rotation profile. Dashed gray is the 1:1 line. Error bars in the y axes are for the first method in the difference.

It has been a long established fact that gravity darkening can underestimate the value of vsin⁡iv \sin i, e.g. Frémat et al. [44]; Zorec et al. [118, 119], and has been also studied in detail for the rapidly rotating Be stars (Townsend et al. [111]). In some cases, due to the quality of the spectra, poor normalisation, and high effective stellar temperature, high or low stellar rotation, not all methods could be used simultaneously, which is why we decided to use three different methods.

Best method for vsin⁡iv \sin i

Even though the different methods used for deriving vsin⁡iv \sin i are well established, it is always important to keep in mind that stellar properties and instrumental effects play a huge role, see Figure 5. Method 1 and method 3 can often fail if multiple stellar components are present and clearly visible in the spectra, while method 2 can still be used in a limited way. It is difficult to decide which method is preferable and which one should be used for the final value. Our methodology can lead to up to 9 values in some cases, as multiple stars were observed with up to three different spectrographs. All measurements per spectrograph and per method are summarised in Table 2 only available at the Strasbourg Astronomical Data Centre (CDS). We also estimate one value of vsin⁡iv \sin i per star, as the mean of values from all available instruments and all three methods. The uncertainty of the final value vsin⁡iv \sin i is taken as the standard deviation of the available values. This final value of vsin⁡iv \sin i can also be found in the last column of Table 2. This mean value is later used in our combined analysis of the TESS photometry.

A different option was a weighted mean of different methods; however, this comes with varying reliabilities of errorbars, as they are almost always underestimated compared to the actual error. Using different methods allows us to quantify the spread in the values for each star. The final measured distribution of vsin⁡iv \sin i is shown in Figure 6, with a mean value of 120.3 km s−1^{-1}. The probability density distribution is unimodal and positively skewed [120, 114].

Figure 6. Distribution of measured vsin⁡iv \sin i for δ\delta Sct stars with bin size 20 km s−1^{-1}, mean value 120.3 km s−1^{-1} and peak probability density value 113.8 km s−1^{-1}.

We note that micro and macroturbulence play an important role in determining the true vsin⁡iv \sin i, mainly in slow rotators, e.g. <20< 20 km s−1^{-1}. The rotational profile in method 1 does not distinguish vsin⁡iv \sin i from macroturbulence, and we obtain an upper broadening limit. When using method 2 with iSpec, all 3 parameters can be fitted, which in the χ2\chi^2 minimisation may cause degeneracies, mainly for the slowest rotators. It is possible that some derived uncertainties for slower rotators are underestimated for different methods; because of that, we prefer to use the mean value from all measurements. As expected, different values at higher vsin⁡iv \sin i can be attributed to gravity darkening, which creates a different line profile than that of a simple non-rotating star. While it is possible to disentangle all three: vsin⁡iv \sin i, microturbulence, and macroturbulence [98], we do not expect statistically different results. We just note that treating stars with low and very high vsin⁡iv \sin i should be approached with caution.

Comparison to previous studies

A fraction of the stars from our observed sample already had published values of vsin⁡iv \sin i, [1, 107, 90, 92] (Zorec et al. 2011), which we compare to our derived values in Figure 7, subplots (A), (B) and (C). We do not show Kahraman Aliçavuş et al. (2017) in the comparison, as only two of the stars, TIC 17992601 with vsin⁡i=145(6)v \sin i = 145(6) km s−1^{-1} (our value 142(6)142(6) km s−1^{-1}) and TIC 66693422 with vsin⁡i=75(3)v \sin i = 75(3) km s−1^{-1} (our value 75(1)75(1) km s−1^{-1}), were observed in both our sample and theirs. None of the stars from Wang et al. (2024) is in our sample; therefore, we cannot make a comparison to our derived values.

Comparison of $v \sin i$ measurements from three methods

Figure 7. Left: Comparison of vsin⁡iv \sin i from this work compared to previous studies. Left to right: Method 1, 2 and 3. Dashed gray is the 1:1 line. Error bars in the yy axes are for our derived values.

Sometimes, multiple studies yield values that are off by more than 30 km s−1^{-1}; therefore, we plot available values from all different sources. In the comparison of our values to previous studies, there is a much greater scatter than in the comparison of different methods (See Figure 5 and Figure 7), which should be expected. The only visible trend is in subplot (C), which is probably the range where gravity darkening becomes prominent. Interestingly, the Fourier method (C) also deviates from Díaz et al. [40], who used the same Fourier method, suggesting that other factors, such as the continuum normalisation of the spectra or the changes to the spectral line profiles due to non-radial pulsations, can play a role.

Comparison of instruments

A subsample of δ\delta Sct stars was observed using up to three different instruments to verify their compatibility. The comparison of vsin⁡iv \sin i measured by different methods can be seen in Figure 8, subplots (α\alpha), (β\beta) and (γ\gamma). The largest spread is for iSpec (method 2) with ≈12\approx12 km s−1^{-1}, while Fourier and Voigt*Rot have ≈9\approx9 km s−1^{-1} and ≈8\approx8 km s−1^{-1} respectively. A larger spread in subplot (β\beta) is likely caused by the dependence on the spectrograph’s resolving power, as lower resolution MUSICOS and OES have a smaller spread than HERMES-MUSICOS or HERMES-OES. The biggest outliers in the subplot (γ\gamma) are the HERMES-OES and HERMES-MUSICOS pairs. Similarly to the bottom part of

Figure 8. Comparison of vsin⁡iv\sin i measured by various spectrographs with method 1, 2 and 3. In total 48, 63 and 45 star pairs were simultaneously observed by more instruments, sometimes not all methods for all spectra could be used. Dashed gray is the 1:1 line. Error bars in the y axes are for the instrument 2.

Figure 4, we can see that the difference between spectrographs is often far greater than the error of separate measurements.

Gaia vbroad broadening parameter

We have also compared our measurement of vsin⁡iv \sin i to the Gaia DR3 vbroad spectral line broadening parameter [43], which combines both the effects of rotation and macroturbulence. It is well established that vbroad is under/overestimated compared to the projected rotational velocity at low/high vsin⁡iv \sin i, and it also deviates for stars with Teff>7500T_{\mathrm{eff}} > 7500 K [43, 50, 75, 76]. The comparison of our measurement of vsin⁡iv \sin i to available vbroad is shown in Figure 9 and confirms that vbroad in the low/high vsin⁡iv \sin i range is under/overestimated. If divided into 20 km s−1^{-1} bins, the mean difference smaller than 4 km s−1^{-1} is only in the range 60 km s−1^{-1} to 120 km s−1^{-1}, while above 200 km s−1^{-1}, the mean difference is ≈45\approx45 km s−1^{-1}. We also see that the biggest outliers are stars with Teff>8000T_{\mathrm{eff}} > 8000 K and vsin⁡iv \sin i above 200 km s−1^{-1}. Using the second-order polynomial Δ=vbroad−vsin⁡i=a(vsin⁡i)2+b(vsin⁡i)+c\Delta= \mathrm{vbroad} - v \sin i = a(v \sin i)^2 + b(v \sin i) + c (red line in Figure 9), we obtain the best-fit parameters a=(1.0±0.2)×10−3a = (1.0 \pm0.2) \times10^{-3} km−1^{-1} s, b=(−4.9±2.9)×10−2b = (-4.9 \pm2.9) \times10^{-2}, and c=−5.3±1.2c = -5.3 \pm1.2 km s−1^{-1} for calibrating between vsin⁡iv \sin i and vbroad for the δ\delta Sct stellar population.

Comparison of measured $v\sin i$ and Gaia DR3 vbroad parameter

Figure 9. Comparison of measured vsin⁡iv\sin i and Gaia DR3 vbroad parameter. Dashed gray is the 1:1 line. Red line is a second-order polynomial fit. Effective temperature TeffT_{\mathrm{eff}} is obtained by fitting spectra in iSpec.

Effect of binaries

Binary stars, especially eclipsing binaries, are one of the best laboratories for testing stellar structure and evolution [108]. The binary fraction in A- and F-type stars can be as high as 50 % [41, 74], and separating them from single stars can be challenging. While our sample selection avoids close eclipsing binaries and ellipsoidal variables, many non-eclipsing or even long-period binaries are present in our sample. We have checked for binarity using the SIMBAD Astronomical Database4 [115] and the Gaia DR3 for multiplicity [46]. We used the Liakos & Niarchos [66] catalogue of δ\delta Sct stars in binaries and crossmatched it with our sample. We have also crossmatched our sample with the catalogue of binaries detected from phase modulation of the pulsations in TESS photometry [39] with only one star, TIC 235709086, being in both samples.

Out of a total of 81 binaries, we have found 5 new binary candidates. For every observed spectrum, we performed a cross correlation with the template spectrum in iSpec, and if more than one dominant peak is present, we assume possible binarity. We list new binary candidates in Appendix B. One star, TIC 331893210, has 3 visible components in the spectra (see Fig. B.1). Some of the binaries in our sample probably remain hidden and might cause low amplitude pulsations to come from the secondary, while our measured vsin⁡iv \sin i is from the primary.

Critically rotating stars

For stars with mass concentrated in the centre, the Roche model relates the critical rotational velocity to mass and radius as

vcrit=GM∗Re=2GM∗3Rp,(2)v_{\mathrm{crit}} = \sqrt{\frac{GM_*}{R_{\mathrm{e}}}} = \sqrt{\frac{2GM_*}{3R_{\mathrm{p}}}}, \tag*{(2)}

where GG is the universal gravitational constant, M∗M_* is the mass of the star, and ReR_{\mathrm{e}} and RpR_{\mathrm{p}} are the equatorial and polar radii, respectively [89, 70, 49]. We have used the mass M∗=MTICM_* = M_{\mathrm{TIC}} and the polar radius value Rp=RTICR_{\mathrm{p}} = R_{\mathrm{TIC}} [33, 34, 49, 77] from the TIC catalogue [109] to calculate the vcritv_{\mathrm{crit}}. In the next section, it will be useful to work with the ratio vsin⁡i/vcritv \sin i/v_{\mathrm{crit}} and not purely vsin⁡iv \sin i as the critical rotation can range from 200 km s−1^{-1} to 400 km s−1^{-1}. In Figure 10, we plot the HR diagram, colour-coded by vsin⁡iv \sin i (left) and vsin⁡i/vcritv \sin i/v_{\mathrm{crit}} (right).

Figure 10. Same as Figure 2 but with vsin⁡iv \sin i (left) and vsin⁡i/vcritv \sin i/v_{\mathrm{crit}} ratio (right) for observed δ\delta Scuti sample.

Inclination and equatorial velocity

It is almost impossible to obtain an inclination ii and therefore true equatorial rotational velocity veqv_{\mathrm{eq}} without major assumptions and very specific cases.

Mode splitting - Asteroseismology provides an independent constraint on the inclination ii through the rotational splitting, which lifts the degeneracy in the azimuthal order mm, splitting each non-radial mode (ℓ≠0\ell\ne0) into a multiplet whose spacing is proportional to the rotation frequency [63]. Combined with the stellar radius, the measured frequency splitting gives equatorial velocity veqv_{\mathrm{eq}} and hence sin⁡i=vsin⁡i/veq\sin i = v \sin i/v_{\mathrm{eq}} [2]. However, this perturbative approach is valid only for slow rotators and breaks down at the equatorial velocities of 50–200 km s−1^{-1} typical of δ\delta Sct stars, where the multiplets become asymmetric and overlap [68, 86].

Gravity darkening

Several studies in the past have demonstrated that vsin⁡iv\sin i is underestimated due to gravity darkening [111, 44, 118, 119], and Barceló Forteza et al. [18] have even proposed utilising gravity darkening in the context of δ Sct stars.

Because the morphology of the spectral line profile of a fast pole-on star (i→0∘i \rightarrow0^\circ) and a slowly rotating star differs, Reiners [87] proposed that by assuming rigid stellar rotation, we can find its true rotational velocity for fast rotators (veq>200 km s−1v_{\mathrm{eq}} > 200\ \mathrm{km\ s}^{-1}) utilising the zeros of the Fourier transform of the spectra. We have tried to apply the method by Reiners [87] to our best available spectra from PLATOSpec, HERMES, and HARPS, but the change in spectral lines due to gravity darkening can also be caused by pulsations [3], which makes this method impractical without a substantial number of high-resolution observations at different pulsation phases.

Barceló Forteza et al. [18] proposed a method for using the scaling relation between the frequency at maximum power and temperature, where one could use the scatter from dTdT, which was supposed to be caused by gravity darkening, to obtain inclination. However, as newer studies have pointed out, [17] frequency at maximum power carries enough ambiguity, and it is hard to break the degeneracy in inclination. Murphy et al. [75] pointed out that frequency at maximum power does not have a strong correlation with effective temperature across a homogeneous δ Sct population. We can conclude that all methods involving gravity darkening cannot be readily applied to δ Sct, as only a small fraction of stars are actually very fast rotators, in which the effects of gravity darkening would be observable in the spectra [92, 120].

Surface modulation - Due to the photometric modulations caused by rotating starspots, it is possible to extract stellar rotation. If this rotational period ProtP_{\mathrm{rot}} is identified properly, we can then calculate stellar rotation as veq=2πR∗Protv_{\mathrm{eq}} = \frac{2\pi R_*}{P_{\mathrm{rot}}}. We have manually inspected all lightcurves for rotational modulation; however, this is complicated by the gg mode frequencies excited in the γ\gamma Dor range (0.3 to 3 d−1^{-1}) [15, 47, 100] together with instrumental artefacts from the spacecraft. We also require the presence of the rotational frequency νrot\nu_{\mathrm{rot}} and its harmonic 2νrot2\nu_{\mathrm{rot}} where ν2=2ν1\nu_2 = 2\nu_1 is within formal analytical error. We have not found any clear candidates, which again points to the fact that none of these methods can be used for a statistical sample, but only for very specific stars.

TESS photometry

At the beginning, we downloaded all available sectors from TESS using the lightkurve Python package [67]. We used Pre-search Data Conditioned Simple Aperture Photometry flux (PDCSAP) in the 2-minute cadence mode. If SPOC lightcurves were not available, we used Quick-Look Pipeline (QLP) data. The data were normalised to their median value per sector, and no further detrending was done. Available sectors were stitched into contiguous light curves without removing the gaps [20]. The sectors can be even years apart, and this can lead to a spectral window contaminated with aliases [19]; however, in the statistical sense of this study, this does not induce substantial error. The longest lightcurve spans 12 sectors, while the median spans 3 sectors. An example light curve can be seen in Figure 11. For every TIC number in our sample, we have checked the contratio, which identifies neighbouring sources and estimates their contribution to the target flux within a nominal TESS aperture using a PSF-based contamination model. For targets with contratio >0.01> 0.01 we have manually inspected TESS target pixel files using tpfplotter (Aller et al. [4]) and discarded strongly contaminated targets.

Figure 11. Top: Example of relative semi-amplitude determination. 5th (P5P_{5}) and 95th (P95P_{95}) percentile are shown using red dashed lines, semi-amplitude ATESSA_{\mathrm{TESS}} using red arrow. Bottom: periodogram with extracted frequencies with S/N>4S/N > 4 (red), dominant frequency νdom\nu_{\mathrm{dom}} (green) and amplitude weighted frequency νAW\nu_{\mathrm{AW}} (blue).

Pulsation amplitudes

Because δ\delta Sct stars can pulsate at hundreds of frequencies, with different amplitudes evolving over time, no single photometric amplitude is defined. Because of this, we have decided to use a 5th (P5P_5) and 95th (P95P_{95}) percentile approach of the normalised TESS flux to estimate the normalised semi-amplitude ATESSA_{\mathrm{TESS}}. The semi-amplitude ATESSA_{\mathrm{TESS}} was calculated as half of the relative flux distance from the P5P_5 and P95P_{95} percentile lines, see Figure 11. This simple approach allows us to define amplitude even for light curves with multiple pulsation modes and complicated behaviour.

The relation between the TESS semi-amplitude ATESSA_{\mathrm{TESS}} and the projected rotational velocity vsin⁡iv \sin i can be seen in Figure 14 subplot (VII). Although this relation between photometric amplitude and vsin⁡iv \sin i is well established, measured by Danziger & Faber [37] and Solano & Fernley [107] in VV magnitude, their sample did not adequately cover all values of the projected rotational velocity vsin⁡iv \sin i. In addition, we find that the upper envelope drops and the lower envelope rises; therefore, the range of photometric amplitudes is very limited for higher vsin⁡iv \sin i; see the underpopulated regions in Subsect. 4.3.

Plots of pulsation amplitudes against projected rotational velocity

Figure 14. Same as Fig. 13 but for amplitudes. Left column: measured vsin⁡iv\sin i vs dominant amplitude AdomA_{\mathrm{dom}}, mean amplitude AmeanA_{\mathrm{mean}}, mean-square amplitude ⟨A2⟩\langle A^2\rangle and TESS pulsation semi-amplitude ATESSA_{\mathrm{TESS}}. Right column: same as left but for vsin⁡i/vcritv\sin i/v_{\mathrm{crit}}.

High-amplitude δ\delta Sct (HADS) McNamara [73] defined HADS as a class of radial pulsators having peak-to-peak photometric amplitudes ΔV>0.3\Delta V > 0.3 mag. They pulsate in fundamental and first-overtone radial modes and exhibit small rotational velocities (vsin⁡i<40v \sin i < 40 km s−1^{-1}); however, not all slow rotators have high amplitudes. Some HADS may even pulsate in low-amplitude non-radial pulsations (Poretti [81]; Poretti et al. [82]). The nature of HADS remains poorly understood, and it is still an open question how they differ from other δ\delta Sct stars (Balona [13]).

Pulsation amplitudes in the TESS passband (600–1000 nm, λeff≈786\lambda_{\mathrm{eff}} \approx786 nm) are suppressed compared to Johnson VV. Because of this, we define HADS in the TESS passband by calibrating 59 stars from Solano & Fernley [107] using our semi-amplitude determination ATESSA_{\mathrm{TESS}} to their peak-to-peak amplitude ΔV=2AV\Delta V = 2A_V, which is twice the semi-amplitude we use (see Figure 12). We empirically derive the semi-amplitude ratio AV/ATESS=1.49±0.05A_V/A_{\mathrm{TESS}} = 1.49 \pm0.05, which is consistent with theoretical non-adiabatic pulsation models (Daszyńska-Daszkiewicz et al. [38]; Dupret et al. [42]) and with the empirical ratio ATESS/AKepler=0.74±0.01A_{\mathrm{TESS}}/A_{\mathrm{Kepler}} = 0.74 \pm0.01 (Antoci et al. [6]). This would put HADS at ATESS>0.1A_{\mathrm{TESS}} > 0.1 in our relative TESS semi-amplitude. HADS make up only a small fraction of the general δ\delta Sct population (McNamara [73]; Breger [27]), and due to our sample definition, we have not observed any HADS. However, one star in our sample, TIC 306889170, with vsin⁡i=7(1)v \sin i = 7(1) km s−1^{-1}, has HADS characteristics with much lower amplitude AV=0.035A_V = 0.035. TIC 306889170 has two dominant modes with a period ratio P1/P0=0.79P_1/P_0 = 0.79, typical of the radial first overtone and the fundamental mode (Petersen & Christensen-Dalsgaard [80]). This again confirms that not all slowly rotating, double-mode radial pulsators are necessarily HADS (Zima et al. [117]; Lenz et al. [65]) and that the distinction between HADS and classical δ\delta Sct stars is still not understood.

Photometric TESS semi-amplitude versus projected rotation

Figure 12. Photometric TESS semi-amplitude ATESSA_{\mathrm{TESS}} vs vsin⁡iv\sin i for our sample complemented with stars from Solano & Fernley [107]. Determined ratio of TESS and VV amplitudes is in the top-right corner. Red boundary is the HADS region defined by McNamara [73].

Frequency and amplitude analysis

A common characteristic of δ\delta Sct stars is that they pulsate in a variety of frequencies, combining both p and g modes (Breger [27]; Breger et al. [29, 28]; Aerts et al. [2]; Gautam et al. [48]). Given the size of our sample, individual mode identification for each star is beyond the scope of this study; therefore, we adopt a simplified approach for the frequency analysis. For each object, we calculated Lomb-Scargle periodograms normalised to Fourier semi-amplitude (Lomb [69]; Scargle [96]) using the Astropy (Astropy Collaboration et al. [10, 9, 8]) Python package and extracted all frequencies with S/N>4S/N > 4 (Breger et al. [30]) through iterative prewhitening, with all frequencies fitted and optimised using the multi-sinusoidal least-squares model (Lenz & Breger [64]). We searched the frequency range from 0 to 150 d−1^{-1}, allowing a maximum of 50 extracted frequencies, however, this limit was never reached. We also require that each extracted frequency must lie at least one Rayleigh resolution (1/T1/T, where TT is the length of the data set) away from every frequency already found (Antoci et al. [6]), and its amplitude must be between 0.8 and 1.25 times the amplitude in the original periodogram at the same frequency [6, 113]. An example periodogram with extracted frequencies is shown in Figure 11.

Using the extracted frequencies, we define the following metrics [59, 18, 17]: the lowest extracted frequency νmin⁡\nu_{\min} in a range >5 d−1> 5\ \mathrm{d}^{-1}, the highest extracted frequency νmax⁡\nu_{\max}, the dominant frequency νdom\nu_{\mathrm{dom}} as the frequency with the highest amplitude, the dominant amplitude AdomA_{\mathrm{dom}} as the highest amplitude, the amplitude weighted frequency νAW\nu_{\mathrm{AW}}, and the mean amplitude AmeanA_{\mathrm{mean}}

νAW=∑iNνiAi∑iNAi,Amean=∑iNAiN.(3)\nu_{\mathrm{AW}} = \frac{\sum_{i}^{N}\nu_i A_i}{\sum_{i}^{N} A_i}, \qquad A_{\mathrm{mean}} = \frac{\sum_{i}^{N} A_i}{N}. \tag*{(3)}

power weighted frequency νPW\nu_{\mathrm{PW}} and mean-square amplitude ⟨A2⟩\langle A^2\rangle

νPW=∑iNνiAi2∑iNAi2,⟨A2⟩=∑iNAi2N.(4)\nu_{\mathrm{PW}} = \frac{\sum_{i}^{N}\nu_i A_i^2}{\sum_{i}^{N} A_i^2}, \qquad\langle A^2\rangle= \frac{\sum_{i}^{N} A_i^2}{N}. \tag*{(4)}

Distributions of defined frequency metrics are shown in Figure 13, subplots (i), (iii), (v), (vii), (ix), (xi), (xiii), (xv), (xvii), (xix) and amplitude metrics in Figure 14, subplots (I) to (VIII).

Measured frequencies versus rotational velocity

Figure 13. Column 1: Measured vsin⁡iv\sin i vs lowest extracted frequency νmin⁡\nu_{\min} in the range >5 d−1> 5\ \mathrm{d}^{-1}, largest extracted frequency νmax⁡\nu_{\max}, dominant frequency νdom\nu_{\mathrm{dom}}, amplitude weighted frequency νAW\nu_{\mathrm{AW}}, and power weighted frequency νPW\nu_{\mathrm{PW}}. Column 2: Same as column 1 but for mean density normalised frequencies. Column 3: Same as column 1 but for vsin⁡i/vcritv\sin i/v_{\mathrm{crit}}. Column 4: Same as column 2 but for vsin⁡i/vcritv\sin i/v_{\mathrm{crit}}. In columns 1 and 3 red regions are the upper underpopulated regions. Permutation pupp_{\mathrm{up}} value is printed together with σ\sigma in the middle red squares. Uncertainty weighted straight line fits for bins with 20 stars are shown in dashed red line (statistically significant) and dashed gray line (no significance). Spearman rank correlation ρ\rho and pp are shown in upper right corner. Horizontal dashed lines in columns 2 and 4 is the radial fundamental mode (F), and the first, second, and third radial overtones (1O–3O). Blue and purple points are the binned data (bin size = 20 stars) and lines are the uncertainty weighted straight line fits.

Statistical analysis

Underpopulated regions - Visual inspection of the subplots in Figure 13 and Figure 14 reveals the possibility that the frequencies decrease with increasing projected rotation vsin⁡iv\sin i and that certain regions of the plots are underpopulated. To test this hypothesis for all frequency metrics, we define the underpopulated region using the linear boundary envelope that encloses all of the observed data (y=ax+by=ax+b, Figure 13), and for amplitude metrics, the boundaries are log-linear (log⁡A=ax+b\log A=ax+b, Figure 14). This is done for the upper and lower boundaries, represented by the red and blue dashed lines. After that, we evaluate the statistical significance of these underpopulated regions using a non-parametric Monte Carlo permutation test. We permute parameters on the yy axes 10 000 times, while keeping vsin⁡iv\sin i or vsin⁡i/vcritv\sin i/v_{\mathrm{crit}} fixed. In each permutation, we refit the boundary envelope to the permuted data and record the area of the resulting empty region, so that the test accounts for the boundary itself being defined from the observed data. The pp value is the fraction of permutations whose empty region is at least as large as the observed one, and we express it as σ\sigma, the number of standard deviations from the mean of a normal distribution that would give the same probability (e.g. p=0.05p=0.05 corresponds to 2σ2\sigma). In Figure 13 and Figure 14, we only show regions with p<0.05p<0.05.

For vsin⁡iv\sin i, the most significant underpopulated regions are found for the minimum frequency (i) at 3.1σ3.1\sigma and, among the amplitude metrics, for ⟨A2⟩\langle A^2\rangle (V) at 3.4σ3.4\sigma and the TESS semi-amplitude ATESSA_{\mathrm{TESS}} (VII) at 3.3σ3.3\sigma, followed by AmeanA_{\mathrm{mean}} (III) at 2.8σ2.8\sigma and AdomA_{\mathrm{dom}} (I) at 2.3σ2.3\sigma. When the frequency metrics are plotted against vsin⁡i/vcritv\sin i/v_{\mathrm{crit}}, the trends become even more significant; in subplots (iii) and (xi), the underpopulated regions reach p≤10−4p\leq10^{-4} (≥3.7σ\geq3.7\sigma), followed by (xix) at 2.8σ2.8\sigma and (xv) at 2.3σ2.3\sigma, while the maximum extracted frequency νmax⁡\nu_{\max} (vii) remains unchanged. When pulsation amplitudes are plotted against vsin⁡i/vcritv\sin i/v_{\mathrm{crit}}, the upper underpopulated region disappears and the lower becomes more prominent, most notably for the TESS amplitude (VIII) at 2.9σ2.9\sigma. Because the effective temperature TeffT_{\mathrm{eff}} and vsin⁡iv\sin i are intertwined, we also performed random permutations in the 250 K temperature bins. With this temperature permutation, we can obtain even stronger statistical significance.

General trends - Minimum and maximum extracted frequencies can be heavily dependent on the S/NS/N ratio in the periodograms; therefore, we also investigate the general trends using a straight line fit. We have divided the sample into bins with 20 stars in the vsin⁡iv\sin i and vsin⁡i/vcritv\sin i/v_{\mathrm{crit}}, performed an uncertainty weighted straight line fit, and calculated the nonparametric Spearman correlation test for all unbinned data. For vsin⁡iv\sin i, the only statistically significant yet modest decrease is νmin⁡\nu_{\min} (ρ=−0.245\rho=-0.245 and p=0.0001p=0.0001), while νmax⁡\nu_{\max} (ρ=+0.150\rho=+0.150 and p=0.0202p=0.0202) shows an even smaller increase that is less statistically significant. All the other parameters remain invariant (p>0.05p>0.05). When we evaluated pulsation characteristics against vsin⁡i/vcritv\sin i/v_{\mathrm{crit}}, we find that the only statistically significant results are νmin⁡\nu_{\min} (ρ=−0.429\rho=-0.429, p<10−4p<10^{-4}), νdom\nu_{\mathrm{dom}} (ρ=−0.315\rho=-0.315, p<10−4p<10^{-4}), νAW\nu_{\mathrm{AW}} (ρ=−0.284\rho=-0.284, p<10−4p<10^{-4}), and νPW\nu_{\mathrm{PW}} (ρ=−0.302\rho=-0.302, p<10−4p<10^{-4}), while all other metrics remain invariant.

Frequency metrics νmin⁡\nu_{\min}, νdom\nu_{\mathrm{dom}}, νAW\nu_{\mathrm{AW}}, and νPW\nu_{\mathrm{PW}} are decreasing with vsin⁡i/vcritv\sin i/v_{\mathrm{crit}}, which can be explained by the fact that the frequencies of p modes scale with mean stellar density ν∝ρˉ\nu\propto\sqrt{\bar{\rho}} [27, 2], and the centrifugal force in fast rotators causes the equatorial radius to inflate while conserving mass, which lowers the mean stellar density ρˉ\bar{\rho}.

To investigate whether the inflation of the fast rotating star is solely responsible for this monotonic decrease in frequencies, we have normalised all frequency metrics by the square root of the mean stellar density ρˉ/ρˉ⊙\sqrt{\bar{\rho}/\bar{\rho}_{\odot}}, which is calculated from the MTICM_{\mathrm{TIC}} and RTICR_{\mathrm{TIC}} values. This can be done in the first approximation even for equator-on fast rotators (which is our case for high vsin⁡iv\sin i), and the error in density would not exceed 5–10%. As seen in Figure 13 subplots (ii), (iv), (x), (xii), (xiv), (xvi), (xviii), (xx) the decreasing trends persist, which suggests that density is not the only driver of lower frequencies in fast rotators. Horizontal dashed lines in Figure 13 columns 2 and 4 mark the radial fundamental mode (F), derived from the empirical pulsation constant QF=0.0315Q_{\mathrm{F}}=0.0315 d (Barac et al. 2022; Mani et al. 2025), and the first, second, and third radial overtones (1O–3O), obtained from period ratios Pn/P0P_n/P_0 predicted by pulsation models [110].

Conclusions

In this paper, we present the relationships between the rotation of δ\delta Sct stars and their frequencies and amplitudes. Our findings show the general trends among this stellar population. We have derived projected rotational velocities vsin⁡iv\sin i for 262 bright δ\delta Sct stars from an all-sky survey using 5 high-resolution spectrographs. Together, we have obtained 307 high-resolution spectra, with some of the stars being observed with multiple instruments. We have derived values of vsin⁡iv\sin i using three different methods and compared them to each other. We conclude that the different methods provide consistent results at low to moderate rotational velocities but begin to diverge at higher velocities, which we attribute, at least partly, to gravity darkening in rapidly rotating stars. We show that the statistical error from deriving vsin⁡iv\sin i and effective temperature TeffT_{\mathrm{eff}} is smaller than the spread of the values derived using different spectrographs. We conclude that deriving per-star values of vsin⁡iv\sin i and effective temperature TeffT_{\mathrm{eff}} is not only sensitive to the method but also to the instrument itself and the quality of its spectra, mainly the normalisation of the continuum and S/N. We demonstrate that instrumental effects and drifts can appear between different spectrographs (see Figure 5, subplot (b)), highlighting the need for cross-checking to ensure homogeneous parameter estimates. We discussed the different possibilities of finding true veqv_{\mathrm{eq}} and showed that the methods are either highly impractical or almost impossible for δ\delta Sct stars.

We have also compared our derived vsin⁡iv\sin i to the literature values (Figure 7), and different spectrographs with each other (Figure 8), which shows that lower resolution and higher resolution spectra yield different values depending on the methods used. This behaviour can reflect the fact that some methods are more effective in decoupling the intrinsic rotational broadening from the spectrograph’s instrumental profile.

The measured vsin⁡iv\sin i is compared to the Gaia vbroad spectral line broadening parameter (Figure 9), which confirms that for low/high vsin⁡iv\sin i, the vbroad parameter is under/overestimated. Using the second-order polynomial, we obtain the calibration relation Δ[km s−1]=vbroad−vsin⁡i=(1.0±0.2)×10−3(vsin⁡i)2−(4.9±2.9)×10−2(vsin⁡i)−(5.3±1.2)\Delta[\mathrm{km\,s}^{-1}]=v_{\mathrm{broad}}-v\sin i=(1.0\pm0.2)\times10^{-3}(v\sin i)^2-(4.9\pm2.9)\times10^{-2}(v\sin i)-(5.3\pm1.2). This indicates that additional spectroscopic information is needed to reliably determine rotational velocities across the full range.

The distribution of vsin⁡iv\sin i is unimodal, with a mean value of 120.3 km s−1^{-1} and a peak probability at 113.8 km s−1^{-1} (Figure 6).

Bedding et al. [21] showed that fast rotators viewed pole-on (i→0∘i\rightarrow0^{\circ}) have a simplified pulsation spectrum, as the disc-integrated visibility arises primarily from zonal (m=0m=0) and pure radial (ℓ=0\ell=0) modes. Gravity darkening in fast rotators causes shifts in temperature and apparent luminosity, and moves them in the HR diagram. Naturally, we will have some pole-on pulsators located in the low vsin⁡iv\sin i range, but the probability distribution of randomly orientated inclinations is P(i) di=sin⁡i diP(i)\,di=\sin i\,di, so intermediate-to-high inclinations are statistically much more likely than pole-on configurations. This allows us to use vsin⁡iv\sin i not only as a firm lower limit but also as a robust proxy for the intrinsic equatorial rotation velocity veqv_{\mathrm{eq}}. Using all available TESS photometry, we have derived photometric amplitudes, extracted pulsation frequencies and amplitudes, and compared them to our defined pseudo metrics (Figure 13 and Figure 14). We confirm that the photometric amplitude decreases with vsin⁡iv\sin i and shows a modest decrease even for stars that are not high-amplitude δ\delta Sct. We find that the value of photometric amplitude is constrained from the lower region with high vsin⁡iv\sin i. We find that stars with high vsin⁡iv\sin i have a much more limited range of allowed amplitudes than slow rotators. We calibrate the VV magnitude with TESS flux, obtaining the semi-amplitude ratio AV/ATESS=1.49±0.05A_V/A_{\mathrm{TESS}} = 1.49 \pm0.05. Using the definition by McNamara [73], we can define HADS in the TESS passband as stars with a relative semi-amplitude ATESS>0.10A_{\mathrm{TESS}} > 0.10.

We find that, apart from the maximum frequency νmax⁡\nu_{\max}, all frequency metrics decrease with rotation. The minimum extracted frequency νmin⁡\nu_{\min} decreases with vsin⁡iv\sin i, and the empty region of fast rotators with high frequencies has a significance of 3.1σ3.1\sigma, while the largest frequency νmax⁡\nu_{\max} does not decrease with any statistical significance. The dominant frequency νdom\nu_{\mathrm{dom}}, amplitude weighted frequency νAW\nu_{\mathrm{AW}}, and power weighted frequency νPW\nu_{\mathrm{PW}} decrease with vsin⁡i/vcritv\sin i/v_{\mathrm{crit}}. We show that the amplitudes of the extracted frequencies and their following metrics (AdomA_{\mathrm{dom}}, AmeanA_{\mathrm{mean}}, ⟨A2⟩\langle A^2\rangle) are also constrained in the region of fast rotators, and that pulsation amplitudes have a more limited range with higher rotation. At high vsin⁡iv\sin i, the TESS photometric amplitude is bounded from above and, more weakly, from below.

Frequencies of p modes scale with the average density ν∝ρˉ\nu\propto\sqrt{\bar{\rho}}. Centrifugal force in fast rotators causes the equatorial radius to inflate while conserving mass, which results in lower densities and, therefore, lower frequencies. We show that the change in stellar density cannot be solely responsible for the decrease in observed frequencies. Dividing frequencies by the ρˉ/ρˉ⊙\sqrt{\bar{\rho}/\bar{\rho}_{\odot}}, we remove the first-order effect of rotational inflation. This property then measures the pulsation constant ν/ρˉ/ρˉ⊙=1/Q\nu/\sqrt{\bar{\rho}/\bar{\rho}_{\odot}} = 1/Q, which is nearly the same for all δ\delta Sct stars for a given radial mode.

We suggest that rapid rotation lowers the characteristic radial order of excited p modes, which, together with a decrease in mean stellar density, causes lower observed frequencies. This agrees with Murphy et al. [75], who found that rapid rotators in the Cep–Her complex populate the low-order ridge of the νmax⁡\nu_{\max}–TeffT_{\mathrm{eff}} diagram, which is dominated by modes near the radial fundamental.

Even though the frequency normalisation relies on TIC masses and radii, the individual stars can be uncertain by half a radial order, whereas the shift to lower frequencies is significant for faster rotators. The highest extracted frequency lies well above the third overtone and is independent of vsin⁡iv\sin i. These cannot be low-order radial p modes, but they may be higher-order radial overtones, non-radial p modes, or combination frequencies; distinguishing between these possibilities requires mode identification, which is beyond the scope of this study. The lowest extracted frequencies have values below the expected fundamental radial mode (≈32 d−1\approx32\,\mathrm{d}^{-1}), so they cannot be low-order radial modes. We attribute these peaks to non-radial p modes, mainly low-degree retrograde components shifted to lower frequencies in the inertial frame by rotational splitting, and suggest that fast rotators excite a larger fraction of non-radial p modes at low radial orders. The excited frequencies of fast rotators tend to be more complex than those of slow rotators, often caused by rotational splitting and the coupling of modes of different degrees ([2, 21, 75]).

In this study, we have demonstrated that the pulsation frequencies of δ\delta Sct stars depend on stellar rotation. Our findings demonstrate the need for asteroseismic modelling that takes rotation into account. We show the need for large spectroscopic samples in the era of high-precision space photometry. Relations between vsin⁡iv\sin i and pulsation characteristics established in this study provide a spectroscopic reference which can be used and compared for a much larger population of δ\delta Sct stars observed by the upcoming PLATO mission ([85]).

Data availability

tbl:2 is available only in electronic form at the CDS via anonymous ftp to cdsarc.u-strasbg.fr (130.79.128.5) or via http://cdsweb.u-strasbg.fr/cgi-bin/qcat?/J/A+A/.

Acknowledgements. We are grateful to all observers for their work. MZ acknowledges the support from EU MSCA EXOWORLD project ID:101086149. MS and JZ acknowledge support by PRODEX agreement no. 400012791. MV and TP acknowledge support from the Slovak Research and Development Agency supported this work under contract No. APVV-24-0160 and the VEGA grant of the Slovak Academy of Sciences No. 2/0033/26. We acknowledge a bilateral mobility project between Slovak Academy of Sciences and Czech Academy of Sciences SAV-25-10 (CAS-SAS-2024-01). The research of PG was supported by the Slovak Research and Development Agency under contract No. APVV-24-0160. This study is based on new spectroscopic observations with the MUSICOS, OES, and PLATOSpEC echele spectrographs and the ESO and MELCHIORSS spectral library archives. The study uses space photometry from the Transiting Exoplanet Survey Satellite (TESS), which is publicly available from the Mikulski Archive for Space Telescopes (MAST). Funding for the TESS mission is provided by NASA’s Science Mission Directorate. This work made use of tpfplotter by J. Lillo-Box (publicly available in www.github.com/jlillo/tpfplotter), which also made use of the Python packages astropy, lightkurve, matplotlib and numpy. Finally, we acknowledge the use of the electronic database from the CDS, Strasbourg, and the electronic bibliography maintained by the NASA/ADS system.

References

Examples of derived vsin⁡iv \sin i

In Fig. A.1 we present the examples of fitted spectral line Fe i 4957.613 Å for various objects (method 1). In Figs. A.2 and A.3 we provide examples of spectra fitted with iSpec (method 2). In Fig. A.4 we show Fourier transforms of spectral regions 500 to 580 nm (left) and extracted rotational profiles (right) (method 3).

Figure A.1. Examples of method 1 on line Fe i 4957.613 Å fitted with the convolution of the Voigt profile and rotational kernel.

Figure A.2. Examples of method 2. Spectra from different instruments fitted with iSpec are ordered in increasing vsin⁡iv\sin i.

Figure A.3. Fig. A.3: Fig. A.2 continued.

Figure A.4. Examples of method 3 ordered with increasing vsin⁡iv \sin i. Left: Fourier transforms of spectral regions. First zeros of the Fourier transforms, from which the vsin⁡iv \sin i is calculated is shown using vertical dashed lines. Right: Extracted rotational profiles of spectral regions.

Binary candidates

TICnote
331893210triple system
409736562
6029735
278535185
229284388

Table B.1. List of new binary candidates in our sample of δ\delta Sct stars.

Figure B.1. Examples of new binary candidates. Top to bottom: Triple system TIC 331893210, TIC 409736562, TIC 6029735, TIC 278535185 and TIC 229284388. Left to right: cross-correlation function of the entire observed spectrum with a slow-rotation template, spectrum, spectral lines Mg II 4481 Å, Fe II 5018 Å, Hα\alpha 6563 Å.

References

  1. [1]Abt, H. A. & Morrell, N. I. 1995, ApJS, 99, 135
  2. [2]Aerts, C., Christensen-Dalsgaard, J., & Kurtz, D. W. 2010, Astroseismology
  3. [3]Aerts, C., Simón-Díaz, S., Groot, P. J., & Degroote, P. 2014, A&A, 569, A118
  4. [4]Aller, A., Lillo-Box, J., Jones, D., Miranda, L. F., & Barceló Forteza, S. 2020, A&A, 635, A128
  5. [5]Antoci, V., Cunha, M., Houdek, G., et al. 2014, ApJ, 796, 118
  6. [6]Antoci, V., Cunha, M. S., Bowman, D. M., et al. 2019, MNRAS, 490, 4040
  7. [7]Antoci, V., Labadie-Bartz, J., Święch, M., et al. 2026, A&A, 712, A220
  8. [8]Astropy Collaboration, Price-Whelan, A. M., Lim, P. L., et al. 2022, ApJ, 935, 167
  9. [9]Astropy Collaboration, Price-Whelan, A. M., Sipőcz, B. M., et al. 2018, AJ, 156, 123
  10. [10]Astropy Collaboration, Robitaille, T. P., Tollerud, E. J., et al. 2013, A&A, 558, A33
  11. [11]Auvergne, M., Bodin, P., Boisnard, L., et al. 2009, A&A, 506, 411
  12. [12]Balona, L. A. 2014, MNRAS, 437, 1476
  13. [13]Balona, L. A. 2016, MNRAS, 459, 1097
  14. [14]Balona, L. A., Daszyńska-Daszkiewicz, J., & Pamyatnykh, A. A. 2015, MNRAS, 452, 3073
  15. [15]Balona, L. A., Guzik, J. A., Uytterhoeven, K., et al. 2011, MNRAS, 415, 3531
  16. [16]Balona, L. A. & Ozuyar, D. 2020, MNRAS, 493, 5871
  17. [17]Barceló Forteza, S., Moyá, A., Barrado, D., et al. 2020, A&A, 638, A59
  18. [18]Barceló Forteza, S., Roca Cortés, T., & García, R. A. 2018, A&A, 614, A46
  19. [19]Barceló Forteza, S., Pascual-Granado, J., Suárez, J. C., et al. 2024, MNRAS, 535, 2189
  20. [20]Bedding, T. R. & Kjeldsen, H. 2022, Research Notes of the AAS, 6, 202
  21. [21]Bedding, T. R., Murphy, S. J., Hey, D. R., et al. 2020, Nature, 581, 147
  22. [23]Blanco-Cuaresma, S. 2019, MNRAS, 486, 2075
  23. [24]Blanco-Cuaresma, S., Soubiran, C., Heiter, U., & Jofré, P. 2014, A&A, 569, A111
  24. [25]Borucki, W. J., Koch, D., Basri, G., et al. 2010, Science, 327, 977
  25. [26]Brahm, R., Jordán, A., & Espinoza, N. 2017, PASP, 129, 034002
  26. [27]Breger, M. 2000, in Astronomical Society of the Pacific Conference Series, Vol. 210, Delta Scuti and Related Stars, ed. M. Breger & M. Montgomery, 3
  27. [28]Breger, M., Balona, L., Lenz, P., et al. 2011, MNRAS, 414, 1721
  28. [29]Breger, M., Lenz, P., Antoci, V., et al. 2005, A&A, 435, 955
  29. [30]Breger, M., Stich, J., Garrido, R., et al. 1993, A&A, 271, 482
  30. [31]Castelli, F. & Kurucz, R. L. 2003, in IAU Symposium, Vol. 210, Modelling of Stellar Atmospheres, ed. N. Piskunov, W. W. Weiss, & D. F. Gray, A20
  31. [32]Chevalier, C. 1971, A&A, 14, 24
  32. [33]Collins, George W., I. 1963, ApJ, 138, 1134
  33. [34]Collins, George W., I. 1965, ApJ, 142, 265
  34. [35]Cox, J. P. 1963, ApJ, 138, 487
  35. [36]Cunha, M. S., Antoci, V., Holdsworth, D. L., et al. 2019, MNRAS, 487, 3523
  36. [37]Danziger, I. J. & Faber, S. M. 1972, A&A, 18, 428
  37. [38]Daszyńska-Daszkiewicz, J., Dziembowski, W. A., & Pamyatnykh, A. A. 2003, A&A, 407, 999
  38. [39]Dholakia, S., Murphy, S. J., Huang, C. X., Venner, A., & Wright, D. 2025, MNRAS, 536, 2313
  39. [40]Díaz, C. G., González, J. F., Levato, H., & Grosso, M. 2011, A&A, 531, A143
  40. [41]Duchêne, G. & Kraus, A. 2013, A&A, 51, 269
  41. [42]Dupret, M.-A., Grigahcène, A., Garrido, R., Gabriel, M., & Scuflaire, R. 2005, A&A, 435, 927
  42. [43]Frémat, Y., Royer, F., Marchal, O., et al. 2023, A&A, 674, A8
  43. [44]Frémat, Y., Zorec, J., Hubert, A.-M., & Floquet, M. 2005, A&A, 440, 305
  44. [45]Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. 2018, A&A, 616, A1
  45. [46]Gaia Collaboration, Vallenari, A., Brown, A. G. A., et al. 2023, A&A, 674, A1
  46. [47]García, S., Van Reeth, T., De Ridder, J., et al. 2022, A&A, 662, A82
  47. [48]Gautam, A., Murphy, S. J., & Bedding, T. R. 2026, MNRAS, 545, staf2001
  48. [49]Georgy, C., Ekström, S., Granada, A., et al. 2013, A&A, 553, A24
  49. [50]Gootkin, K., Hon, M., Huber, D., et al. 2024, ApJ, 972, 137
  50. [51]Gray, D. F. 1992, The observation and analysis of stellar photospheres., Vol. 20
  51. [52]Grevesse, N. & Sauval, A. J. 1998, Space Science Reviews, 85, 161
  52. [53]Grigahcène, A., Antoci, V., Balona, L., et al. 2010, ApJ, 713, L192
  53. [55]Henriksen, A. I., Antoci, V., Saio, H., et al. 2023, MNRAS, 524, 4196
  54. [56]Kabáth, P., Skarka, M., Hatzes, A., et al. 2026, MNRAS, 545, staf1972
  55. [57]Kabáth, P., Skarka, M., Sabotta, S., et al. 2020, PASP, 132, 035002
  56. [58]Kahraman Aliçavuş, F., Nieczmura, E., Polińska, M., et al. 2017, MNRAS, 470, 4408
  57. [59]Kallinger, T., Weiss, W. W., Barban, C., et al. 2010, A&A, 509, A77
  58. [60]Koubský, P., Mayer, P., Čáp, J., et al. 2004, Publ. Astron. Inst. Czech. Acad. Sci., 92, 37
  59. [61]Kurtz, D. W. 2022, Annual Review of Astronomy and Astrophysics, 60, 31
  60. [62]Kurtz, D. W., Saio, H., Takata, M., et al. 2014, MNRAS, 444, 102
  61. [63]Ledoux, P. 1951, ApJ, 114, 373
  62. [64]Lenz, P. & Breger, M. 2005, Communications in Asteroseismology, 146, 53
  63. [65]Lenz, P., Pamyatnykh, A. A., Breger, M., & Antoci, V. 2008, A&A, 478, 855
  64. [66]Liakos, A. & Niarchos, P. 2017, MNRAS, 465, 1181
  65. [67]Lightkurve Collaboration, Cardoso, J. V. d. M., Hedges, C., et al. 2018, Lightkurve: Kepler and TESS time series analysis in Python, Astrophysics Source Code Library
  66. [68]Lignières, F., Rieutord, M., & Reese, D. 2006, A&A, 455, 607
  67. [69]Lomb, N. R. 1976, Ap&SS, 39, 447
  68. [70]Maeder, A. & Meynet, G. 2000, A&A, 361, 159
  69. [72]Markwardt, C. B. 2009, in Astronomical Society of the Pacific Conference Series, Vol. 411, Astronomical Data Analysis Software and Systems XVIII, ed. D. A. Bohlender, D. Durand, & P. Dowler, 251
  70. [73]McNamara, D. H. 2000, in Astronomical Society of the Pacific Conference Series, Vol. 210, Delta Scuti and Related Stars, ed. M. Breger & M. Montgomery, 373
  71. [74]Moe, M. & Di Stefano, R. 2017, ApJS, 230, 15
  72. [75]Murphy, S. J., Bedding, T. R., Gautam, A., Kerr, R. P., & Mani, P. 2024, MNRAS, 534, 3022
  73. [76]Murphy, S. J. & Gautam, A. 2026, MNRAS, stag1825
  74. [77]Murphy, S. J., Gautam, A., & Claytor, Z. R. 2026, MNRAS, 550, stag1237
  75. [78]Murphy, S. J., Hey, D., Van Reeth, T., & Bedding, T. R. 2019, MNRAS, 485, 2380
  76. [79]Pepe, F., Mayor, M., Delabre, B., et al. 2000, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, Vol. 4008, Optical and IR Telescope Instrumentation and Detectors, ed. M. Iye & A. F. Moorwood, 582–592
  77. [80]Petersen, J. O. & Christensen-Dalsgaard, J. 1996, A&A, 312, 463
  78. [81]Poretti, E. 2003, A&A, 409, 1031
  79. [82]Poretti, E., Suárez, J. C., Niarchos, P. G., et al. 2005, A&A, 440, 1097
  80. [83]Pribulla, T., Vaňko, M., Komžík, R., & Sivanič, P. 2024, Contrib. Astron. Observ. Skalnate Pleso, 54, 43
  81. [84]Raskin, G., Van Winckel, H., Hensberge, H., et al. 2011, A&A, 526, A69
  82. [85]Rauer, H., Aerts, C., Cabrera, J., et al. 2025, Experimental Astronomy, 59, 26
  83. [86]Reese, D. R., MacGregor, K. B., Jackson, S., Skumanich, A., & Metcalfe, T. S. 2009, A&A, 506, 189
  84. [87]Reiners, A. 2003, A&A, 408, 707
  85. [88]Ricker, G. R., Winn, J. N., Vanderspek, R., et al. 2015, Journal of Astronomical Telescopes, Instruments, and Systems, 1, 014003
  86. [89]Roché, É. 1854, Mémoires de l’Académie des Sciences et Lettres de Montpellier, Section des Sciences, 2, 399
  87. [90]Royer, F., Gerbaldi, M., Farraggiana, R., & Gómez, A. E. 2002a, A&A, 381, 105
  88. [92]Royer, F., Zorec, J., & Gómez, A. E. 2007, A&A, 463, 671
  89. [93]Royer, P., Merle, T., Dsilva, K., et al. 2024, A&A, 681, A107
  90. [94]Różański, T., Niemczura, E., Lemiesz, J., Posilek, N., & Różański, P. 2022, A&A, 659, A199
  91. [95]Saio, H., Kurtz, D. W., Murphy, S. J., Antoci, V. L., & Lee, U. 2018, MNRAS, 474, 2774
  92. [96]Scargle, J. D. 1982, ApJ, 263, 835
  93. [97]Schwarzschild, K. & Villiger, W. 1906, ApJ, 23, 284
  94. [98]Simón-Díaz, S. & Herrero, A. 2014, A&A, 562, A135
  95. [99]Skarka, M. & Henzl, Z. 2024, A&A, 688, A25
  96. [100]Skarka, M., Žák, J., Fedurco, M., et al. 2022, A&A, 666, A142
  97. [101]Slettebak, A. 1954, ApJ, 119, 146
  98. [102]Slettebak, A. 1955, ApJ, 121, 653
  99. [103]Slettebak, A. 1956, ApJ, 124, 173
  100. [104]Slettebak, A., Collins, II, G. W., Boyce, P. B., White, N. M., & Parkinson, T. D. 1975, ApJS, 29, 137
  101. [105]Sneden, C., Bean, J., Ivans, I., Lucatello, S., & Sobeck, J. 2012, MOOG: LTE line analysis and spectrum synthesis, Astrophysics Source Code Library, record ascl:1202.009
  102. [106]Sneden, C. A. 1973, PhD thesis, The University of Texas at Austin
  103. [107]Solano, E. & Fernley, J. 1997, A&AS, 122, 131
  104. [108]Southworth, J. & Bowman, D. M. 2026, Annual Review of Astronomy and Astrophysics, 64, 223
  105. [109]Stassun, K. G., Oelkers, R. J., Paegert, M., et al. 2019, AJ, 158, 138
  106. [110]Stellingwerf, R. F. 1979, ApJ, 227, 935
  107. [111]Townsend, R. H. D., Owocki, S. P., & Howarth, I. D. 2004, MNRAS, 350, 189
  108. [112]Uytterhoeven, K., Moya, A., Grigahcène, A., et al. 2011, A&A, 534, A125
  109. [113]Van Reeth, T., Tkachenko, A., Aerts, C., et al. 2015, A&A, 574, A17
  110. [114]Wang, J., Chen, X., Deng, L., Zhang, J., & Sun, W. 2024, ApJ, 978, 53
  111. [115]Wenger, M., Ochsenbein, F., Egret, D., et al. 2000, A&AS., 143, 9
  112. [116]Yang, T.-Z., Zuo, Z.-Y., Li, G., et al. 2021, A&A, 655, A63
  113. [117]Zima, W., Lehmann, H., Stütz, C., Ilyin, I. V., & Breger, M. 2007, A&A, 471, 237
  114. [118]Zorec, J., Frémat, Y., Domiciano de Souza, A., et al. 2016, A&A, 595, A132
  115. [119]Zorec, J., Frémat, Y., Domiciano de Souza, A., et al. 2017, A&A, 602, A83
  116. [120]Zorec, J. & Royer, F. 2012, A&A, 537, A120

Paper details

vlnc:W_82dhgvdg
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents