On the structure of the Pleiades cluster

Abstract

The data available on the distances (both from Gaia DR2 and those determined from interstellar Ca ii lines) to the brightest stars in the Pleiades (‘Seven Sisters’ and their ‘parents’) suggest that these stars highly likely do not form a compact cluster but are localized in front of the main (Pleiades) cluster – at a distance close to that given by Hipparcos. Spectrophotometric distances agree with those from Gaia and Ca ii for the brighter stars, but exhibit a large scatter for the stars at the cluster core. According to our study, this effect can be explained by additional (‘grey’) interstellar extinction due to the presence of larger dust grains within the Pleiades cluster core. Also, we confirm the suggestion that the spatial association of the Pleiades stars and interstellar gas is a result of a chance encounter between the cluster and one or more (at least two) approaching clouds. Indeed, we found that the relatively strong CH⁺ lines observed in the spectra of the Pleiades stars can be divided into two sets: one with very strong lines of the radical and another with relatively faint ones. Both groups also differ in radial velocities.

1 INTRODUCTION

The nine brightest stars of the Pleiades are named for the ‘Seven Sisters’ of Greek mythology: Asterope, Merope, Electra, Maia, Taygeta, Celaeno, and Alcyone, along with their parents Atlas and Pleione. In common astronomical opinion, this compact group of bright stars in the Northern hemisphere forms a model open cluster. It is one of the closest clusters to Earth that is considered to be young and rich in stars. It is thus a cornerstone for understanding the physical properties and evolution of young stars. Moreover, the ease with which one may determine the spectral types and visual magnitudes of the members of this cluster makes it a basic spectrophotometric standard for constructing cosmic distance scales.

The basic role of this cluster in so many astronomical problems stimulated many efforts to determine the distance to this group. The latter is slightly over 100 pc and thus – too large for ground-based parallaxes. The first sufficiently accurate distance estimates were derived by main-sequence fitting, i.e. comparing the main sequence of the Pleiades with that constructed from nearby stars with known trigonometric parallaxes (after correcting for the difference in evolution and metallicity of the compared groups). The obtained main-sequence fitting distance modulus was equal to 5.60 ± 0.04 mag (Pinsonneault et al. 1998). The Hipparcos mission launched in 1994 measured accurate positions, proper motions, and parallaxes for nearly all stars brighter than V > 9 mag. The result obtained by Hipparcos for the Pleiades (van Leeuwen 2009) was, however, a complete surprise, yielding a distance modulus of m − M  = 5.40 ± 0.03 mag. The Hipparcos result suggests that the stars in the Pleiades are about 0.25 mag fainter than other, similar stars of the solar neighbourhood. This large discrepancy prompted a careful reexamination of the assumptions and input parameters of the stellar models, but also a study of potential errors of the Hipparcos data. It was found that the changes in physics or input parameters needed to account for the Hipparcos distances are not reasonable, and the Hipparcos measurements were declared controversial. However, the situation was not so simple. The Hipparcos mission provided a way of obtaining the cluster distance that is free of physical stellar models, and combines distance measurements for more than 50 cluster members, which gives an average distance with an error of only about 1 per cent.

The controversy following the results from the Hipparcos data was a reason to pick up many efforts to determine parallax or distance modulus to the Pleiades. The results of distance modulus determinations, done over 20 yr after Hipparcos mission, are listed in the table 3 of the paper by van Leeuwen (2009). Taking only the determinations, published after the publication of the Hipparcos data in 1997, the weighted mean and standard deviation lead to values 5.63 and 0.58, respectively, much larger than the Hipparcos data-based distance modulus. This was the main reason for more detail analysis of the Hipparcos measurements. The new reduction of the Hipparcos data was made by van Leeuwen (2007) and provides some 27 000 stars with parallaxes known to better than 10 per cent, and over 10 000 known to the precision better than 5 per cent. Van Leeuwen (2009) has used the new reduction of the astrometric Hipparcos data to derive mean parallax and proper motion estimates for 20 open clusters, including Pleiades. The distance moduli of these clusters were derived through superimposition of the HR diagrams of such clusters on the HR diagram of the nearby stars, using parallax-based distance moduli rather than isochrone fitting. The new distance modulus obtained by van Leeuwen (2009) for the Pleiades is equal to 5.40 ± 0.03 and the distance 120.2 pc which does not however change significantly in earlier estimates. The cluster parallax and the problem of the Pleiades distance remains a controversy. Although most of the efforts to determine distance to the Pleiades were based on the main-sequence fitting, there were also attempts to determine it astrometrically.

Soderblom et al. (2005) have re-observed some stars in the Pleiades using the traditional method of parallax astrometry – precise measurements of stellar positions relative to nearby reference stars with the Fine Guidance Sensor onboard the Hubble Space Telescope. Their average distance was equal to 134.6 ± 3.1 pc, which ‘in concert with other recent independent measurements of the Pleiades distance clearly and unambiguously shows that the Hipparcos parallax is wrong and that traditional main-sequence fitting results in reliable estimates’. Their results were also supported by two other recent determinations of visual orbits for binary members of the Pleiades. Pan, Shao & Kulkarni (2004), using the Palomar Testbed Interferometer, have determined very precise relative positions of the two stars comprising Atlas. They concluded that the distance to the Pleiades cannot be less than 127 pc and that the most likely distance lies between 133 and 137 pc. However, without a radial velocity orbit they could not determine reliably all the orbital parameters, so the obtained result cannot be treated as very credible. Munari et al. (2004) analysed the light curves and radial velocity curves of HD 23642, an eclipsing binary in the Pleiades, and determined the distance to this star as 132 ± 2 pc. Similar results were obtained earlier for the Pleiades distance from main-sequence fitting: 132 ± 4 pc by Stello & Nissen (2001) and 132 ± 2 pc by Pinsonneault et al. (1998). Gatewood et al. (2000) have determined the Pleiades distance to be 131 ± 7 pc. Melis et al. (2014) used radio astrometry, referencing essentially stationary quasi-stellar objects, giving a precision better than 0.0001 arcsec in stellar position measurements. Their very long baseline interferometry (VLBI) was used for four Pleiades star systems observed over a period of 1.5 yr. The resulting distances and errors for the four considered systems were: 134.8 ± 0.5 pc for H ii 174, 138.4 ± 1.1 pc for H ii 625, 135.5 ± 0.6 pc for H ii 1136, and 136.6 ± 0.6 pc for H ii 2147. The accuracy of VLBI distance measurements for individual objects was better than 1 per cent. To derive the absolute parallax of the cluster, for each individual star measurement they included additional uncertainties of each star’s position with respect to the centre of the cluster. They adopted the approach of Soderblom et al. (2005) of using the angular dispersion of the cluster as the systematic cluster depth uncertainty. For the assumed Pleiades distance of 130 pc and a projected cluster size of 1 |$\deg$|⁠, they estimated the cluster depth uncertainty to be 2.3 pc. The resulting distance calculated for the Pleiades cluster from the VLBI measurements was 136.2 ± 1.2 pc. Next, this value was combined with all the parallaxes available in the literature for 17 individual Pleiades star systems to form a single non-Hipparcos cluster distance of 136.1 ± 1.0 pc.

A distance to the Pleiades equal to 136 pc was obtained quite recently by Guillermo Abramson (2018) from the Gaia DR2 data, again in full agreement with the most precise recent astrometric measurements and in significant controversy to the Hipparcos equal to 120.2 pc. The Hipparcos revised distance for the Pleiades (van Leeuwen 2009) based on Hipparcos data equal to 120.2 pc (see Figs 1 and 2). It is seriously different from the above-mentioned determinations and requires interpretation. Astrometry (also including that of Hipparcos) is arguably the one branch of astronomy in which accurate and precise knowledge of uncertainties cannot be overlooked. This created a serious problem of the possible source of this discrepancy.

Stello & Nissen (2001) have suggested that ‘the Hipparcos distance could be reconciled with traditional (main-sequence fitting) measures if the bright stars – the Seven Sisters – that dominate the Hipparcos result happen to lie at the near end of an elongated cluster.’ Such an explanation indeed seems possible. The main-sequence fitting method and precise astrometric parallaxes rely mostly on fainter stars (A to G type), which could be located at different distances than the brightest early-type ‘Seven Sisters‘. The authors have considered a possible non-spherical shape of the cluster. They proceed from the fact that the angular size of the Pleiades cluster in the sky can be approximated by the size of the region covered by the member stars used in the investigation; its radius is 6° according to Narayanan & Gould (1999). On the other hand, Stello and Nissen suggested that ‘it could be that the cluster has a more oblong shape in the direction of the line of sight, say with a length that is twice the projected diameter. This shape could be a result of a process of star formation in a scenario in which the first born bright stars (O and B type) form in one part of the gas cloud, and they start to blow the gas cloud in one initial direction. In effect these stars will end up at one end of this deformed shape and the fainter stars (F and G type) will (as an overall trend) form a tail’. So if we see this shape head-on, there will be a trend that the brightest B-type stars are closer to us, and the later classes are further away. Because the calculation of the Hipparcos mean parallax gives the largest weight to the brighter stars, the result will be a slightly shorter distance than the current mean (main body) cluster distance. The idea of a deformed Pleiades cluster seems to be a possible answer to the Pleiades distance problem. In such a situation, it would be of interest to check precisely the possible distance separation between the brighter Pleiades members and the fainter ones.

For this purpose, we have considered all the suggested Pleiades members with individual distance measurements available in the literature (including those of Gaia DR2). In addition to the astrometric estimates, we also used distances determined spectrophotometrically and those obtained from interstellar Ca ii lines (Megier et al. 2009).

2 THE CHOICE OF PLEIADES MEMBERS AND DETERMINATION OF THEIR DISTANCES

Essentially, the distances to chosen stars can be obtained using:

• Trigonometric parallaxes: To this end, we can use the data from Hipparcos and Gaia DR2. It is important to check whether the measurements based on the data of these two satellites are in agreement, or, possibly, systematically different. As was mentioned in the previous paragraph, the Pleiades distance measurements based on the data from these two satellites indeed seem to be systematically different. However, this result is relevant to the Pleiades region as a whole, not to particular stars, especially the brightest ones.

• Spectrophotometric parallaxes: Here the precision depends significantly on the accuracy of spectral classification, the accuracy of absolute magnitude calibration, on the recognition of binarity and/or variability of the chosen targets and, which also seems important, on proper account of the interstellar extinction effects. In any case, spectrophotometric distances are so overwhelmingly and commonly used in astronomy that it seems very important to check their applicability for this cluster, which is considered to be a ‘spectrophotometric standard’. It is very difficult to estimate properly the uncertainties of the spectrophotometric distances. The Sp/L estimates (though checked using our spectra) are uncertain inside one spectral subclass and inside one luminosity class. This leads to serious differences in the calibrated absolute stellar magnitudes. Another source of uncertainty is the extinction that is usually very difficult to be calculated reliably. It is impossible to estimate numerically the resulting errors. Thus, we leave the spectrophotometric distances as they are – as a sort of suggestion.

• Finally, to determine the distances, we can also use the intensities of interstellar Ca ii lines (Megier et al. 2009). This elaborate method, calibrated using Hipparcos trigonometric parallaxes, is proven to be in agreement to those of Gaia DR2 (see fig. 3 of the paper by Krełowski et al. 2018). The Ca ii distances are imprecise inside the errors depending on the S/N ratio of the investigated spectra and limitations of the method (see details in Megier et al. 2009). For the best ones (e.g. HD23480), the uncertainty is about 5 per cent; in the worst case (HD23338) the error may be as big as 26 per cent.

For our purpose of verifying the spatial positions of the brightest Pleiades stars (the ‘Seven Sisters’) relative to the hypothetical main body of the Pleiades cluster, we used all the above-mentioned methods.

Our Ca ii measurements and spectrophotometric determinations (based on the estimated Sp/L’s) were based on the spectra acquired with three echelle spectrographs: UVES, Feros, and MAESTRO (Terskol):

• the UVES spectrograph is fed with the 8 m Kueyen VLT mirror. The spectral resolution is up to R  = 80 000 in the blue range and |$R = 110\, 000$| in the red. The telescope size allows one to get high S/N spectra of even relatively faint stars.

• the FEROS spectrograph is fed with the 2.2 m ESO LaSilla telescope. It allows recording in a single exposure the spectral range of 3600 to 9200 Å  divided into 39 echelle orders. The resolution of Feros spectra is |$R = 48\, 000$|⁠. Feros spectral orders cover sufficiently broad wavelength intervals, making the spectrograph a very useful tool for checking the spectral types and luminosity classes of the observed targets.

• MAESTRO (MAtrix Echelle SpecTROgraph) is attached to the 2 m telescope at the Terskol Observatory (North Caucasus). It is a three branch cross-dispersed echelle spectrograph installed at the coude focus (F/36) of the telescope. It was designed for stellar spectroscopy, with high resolutions ranging from 45 000 to 190 000 in the 3500–10 000 Å spectral region. The lowest resolution mode (sufficient for our programme) allows one to obtain spectra of objects as faint as ∼10^m with a sufficient (∼100) signal-to-noise (S/N) ratio

• Two spectra were acquired using the MIKE spectrograph (Bernstein et al. 2003) fed by 6.5 m telescope at Las Campanas observatory (Chile). Spectra were observed with a 0.35 × 5 arcs slit. We estimated the resolving power using the solitary Thorium lines. It is ∼56 000 (Δv ∼ 5.4 km s⁻¹) on the blue branch (3600–5000 Å) of the spectrograph. The recorded spectra are averages (in the pixel space) of 10 (S/N ∼ 300–400) individual exposures.

All spectra were processed in a standard way using the dech¹ software package

Determining spectrophotometric distances requires precise spectral/luminosity class estimates, their absolute-magnitude calibrations, and estimates of true interstellar extinction, which are most uncertain. The necessary absolute magnitudes of the stars were taken from the calibration of M_V to Sp/L as given by Schmidt-Kaler (1982).

Pleiades members exhibit a rather small observed colour excess – E(B−V). This almost eliminates the need for accurate knowledge of the extinction law: the product of the unknown total-to-selective extinction ratio (R_V) and E(B − V) will be negligibly small. However, the possibility of the ‘grey’ extinction effect cannot be ruled out. This specific problem is discussed below.

We have also limited our Ca  ii method distance measurements to early-type stars to avoid the contamination of interstellar Ca ii lines by stellar ones. In most of cases (e.g. for fast rotators where stellar lines are broadened) confident measurements of interstellar Ca  ii lines are possible for late B-spectral types (like the Pleiades members).

The sample of Pleiades stars, chosen for analysis is listed in the Table 1 (parallaxes from Gaia DR2) and Table 2 (distances derived from spectrophotometric parallaxes and interstellar Ca ii lines).

The reason for dividing the data into two separate sets is the different applicability of particular methods – spectrophotometric and Ca ii – to stars with available high-quality spectra. The tables contain information on star identifications, Sp/L estimates, photometric data, Gaia trigonometric parallaxes, interstellar Ca ii H and K line intensity measurements, and the resulting Ca ii and spectrophotometric distance measurements.

3 DATA ANALYSIS

Fig. 1 shows the apparent brightness of the stars listed in Table 1 versus their Gaia DR2 parallaxes. The distribution of the points in the figure suggests a dependence of the parallaxes on the brightness of the stars and appears to indicate closer distances of the brightest Pleiades stars in comparison with the ‘main’ cluster. To verify this effect, in Fig. 2 we compare the distances to the brightest stars as obtained by GaiaDR2 and the distances inferred from our Ca  ii method. As one can see:

• the two methods of distance determination are in a very distinct agreement (see also Table 2)

• the presented relation suggests that the brightest Pleiades are indeed distributed in front of the ‘main’ cluster – closer than the cluster’s main body. This fact could be a result of a specific process of cluster formation (as suggested by, e.g. Stello & Nissen 2001), or could indicate a more complex spatial structure of the region – e.g. a presence of more separate groups

• Another interesting feature in the figure is the qualitative agreement between the DR2 distances to the brightest stars and those of Hipparcos (120 pc). This fact can diminish the distance controversy.

It is of interest to investigate whether the brightest Pleiades stars form an independent spatially and/or kinematically distinct group. To this end, we plotted in Fig. 3 all the stars listed in Tables 1 and 2 in the Galactic coordinate plane.

The same stars are shown on the μ_α versus μ_δ plane in Fig. 4. The brightest stars are denoted by black dots. There is no distinct difference in the distributions of the brightest (‘Seven Sisters’) and fainter Pleiades members in both (Figs 3 and 4). The tight distribution of both groups under comparison in both figures is an argument that both groups belong to the same cluster.

4 SPECTROPHOTOMETRIC DISTANCES

The analysis presented above shows that the ‘Seven Sisters’ and the Pleiades open cluster main body are located close in the sky. However, as was shown, the most reliable distance measurements place the ‘Seven Sisters’ in the front of the Pleiades open cluster. This fact, if true, should also be confirmed by spectrophotometrically determined distances. The use of spectrophotometric distances for the Pleiades stars is particularly justified as this cluster is a potential spectrophotometric standard and the spectrophotometric method should work especially well for its members. The spectral/luminosity class, necessary for spectrophotometric distance determination, was established according to the adopted spectral criteria – the Mg ii, He i, and H i line ratios. The relation between the two used criteria is presented in Fig. 5 together with the spectral class identifications according to Walborn & Fitzpatrick (1990).

Fig. 6 shows a comparison between spectrophotometrically determined distances (using the intrinsic colours of Papaj; Krełowski & Wegner 1993 and the absolute magnitude calibration of Schmidt-Kaler 1982) to the bright Pleiades and their distances derived from Gaia DR2.

The picture is somewhat confusing. Contrary to the perfect agreement between the Gaia and the Ca ii distance measurements, the spectrophotometric distances to Pleiades do not generally agree with either of the former estimations. Some stars in Fig. 6 deviate significantly from the identity line towards larger spectrophotometric distances. It is interesting that this effect is concentrated in the main cluster region, but not along the line of sight to the cluster itself. A careful examination of spectral and luminosity classes of the considered stars (exhibiting the effect) based on available high-quality spectra, practically excludes any significant improvement to the Sp/L estimates adopted in this paper. Applying the Schmidt–Kaler absolute-magnitude calibration to spectrophotometric distance measurements seems justified – this calibration was proven to be correct in many other cases. There is also no real indication of an extraordinary selective interstellar extinction in the direction of the deviating stars. In this situation, the only logical explanation for the larger than expected spectrophotometric distances for some considered stars seems to be a possible presence of ‘grey’ interstellar extinction. Such an effect, if present, would effectively extend the spectrophotometric distance without any indication of selective extinction. This may also point to the possible peculiar properties of the interstellar matter within the Pleiades cluster.

4.1 Properties of interstellar matter within the Pleiades cluster

As is evident from Fig. 6, stars with ‘peculiar’ (larger than expected) spectrophotometric distances seem to be limited to the cluster main body. In Fig. 7 this effect of ‘enlarged’ spectrophotometric distances (supposedly the effect of ‘grey’ extinction, discovered by us: Skórzyński, Strobel & Galazutdinov 2003; Krełowski et al. 2016) is presented as a function of distance from the cluster centre. The ‘effect’ is measured simply by the difference between the spectrophotometric and Gaia distances. The distance of the considered star from the cluster centre was calculated adopting its Gaia distance (136 pc; Abramson 2018) and position of the cluster centre in the sky: α = 56.00° and δ  = +23.96° taken from the Simbad data base). The figure proves that the ‘effect’ under consideration systematically diminishes with distance from the cluster centre.

It is of interest to investigate whether this supposedly ‘grey extinction’ effect, as presented in Figs 6 and 7, is related to any known interstellar feature in the cluster.

The distribution and properties of diffuse matter in the Pleiades region were studied by White (2003, and in a series of previous papers) and by Ritchey et al. (2006). White (2003) has made a comprehensive interpretation of the complex structure of the ISM in and near the Pleiades star cluster, focusing primarily on the constraints provided by the Na i absorption-line data. He recognized that the gas observed in absorption lies close to the Pleiades and that more than one cloud contributes to several known anomalies in the relative strengths of the interstellar absorption lines. Among the anomalies in the interstellar absorption-line spectra of the Pleiades, he pointed to the exceptionally strong CH⁺ lines, exceptionally high rotational excitation of H₂, and the weakness of Na i, K i, and CH absorptions relative to C i and H₂. According to the author, the combination of shocks and the presence of an enhanced radiation field provides a full explanation for the extraordinary H₂ excitation. The author interpreted the unique strength of CH⁺ lines towards the Pleiades as being due to the inflow of molecular gas towards the luminous stars in the cluster. Weaker CH⁺ lines occur in the direction of most of the cluster stars, supporting the suggestion that the radiative or thermal environment of the cluster plays a role in the formation of this molecular ion. The relative weakness of Na i, K i, and CH absorptions compared to C i and H₂ appears to be a direct consequence of the gas lying in the immediate vicinity of the Pleiades, where the interstellar ionizing and dissociating radiation has a softer spectrum than the Galactic background.

As a result of the performed analysis, White concluded that the interstellar medium in the vicinity of the Pleiades shows the influence of the interactions between stellar radiation and the gas and dust in interstellar clouds. According to the author, the Pleiades stars were not formed from the surrounding material visible as a reflection nebulosity. Rather, the spatial association of the stars and the interstellar gas is the result of a chance encounter between the cluster and one or more approaching clouds. The paper also presents maps showing the general picture of the distribution and properties of interstellar matter around Pleiades.

Ritchey et al. (2006) have come to similar conclusions about the properties of interstellar matter within the Pleiades, explained by the interaction between cluster stars and the surrounding interstellar clouds of gas. They based their analysis on higher resolution (R  = 175 000) and high signal-to-noise observations of interstellar lines along 20 lines of sight towards members of the Pleiades. The spectra they used allowed the detection of absorption features from CN near 3874 Å, Ca ii, K at 3933 Å, Ca i at 4226 Å, CH⁺ at 4232 Å, and CH at 4300 Å. However, only Ca ii and CH⁺ lines are evident.

The authors have found molecular components with radial velocities relative to the local standard of rest of either about +7 or about  +9.5 km s⁻¹, with the higher velocity components being associated with the strongest absorptions. Atomic gas, traced by Ca ii, shows a strong central component at the LSR velocity of about  +7 km s⁻¹, exhibiting velocity gradients indicative of cloud–cluster interactions. According to the authors, they were able to clean the Ca ii ‘K’ profile from stellar contaminations owing to the strength of the interstellar lines and their location away from the core of the stellar profile, but for CN, Ca i , CH⁺, and CH, they were unable to remove the stellar features. Figs 2 to 4 of their paper show the final normalized spectra of Ca ii ‘K’, CH⁺, and CH, respectively. Weaker velocity components, seen in these particular figures, are also marked (though hardly visible). These are however not seen in spectra of common stars, but with a much larger spectral resolution (Hobbs 1969, 1973). These components may partly result from remnants of stellar lines, not completely removed, and causing a wrong continuum. Ritchey et al. have found that the total equivalent widths of the considered interstellar structures are consistent with the results of White (2003), except for the weaker features they detected.

Looking for a possible explanation of the mentioned ‘grey extinction’ effect, we have compared the distribution of this effect in the sky with the maps of interstellar properties in the Pleiades region from the paper by White (2003). The comparison shows no distinct coincidence of the considered effect with spatial interstellar structures on White’s maps. The only spatial characteristic of this effect seems to be its concentration towards the cluster core (Fig. 7).

Since for most targets considered by Ritchey et al. (2006), we also have spectra covering the spectral range of interstellar Ca ii, CH, and CH⁺ structures, we have measured their intensities. However, the spectra at our disposal allow us to measure the intensities of only the main (strongest) line components.

Fig. 8 shows CH and CH⁺ structures in the spectra of two Pleiades, HD 23441 and HD 23850, observed with a high S/N ratio with the Las Campanas Magellan/Clay telescope equipped with the MIKE spectrograph by one of us (GAG).

The CH structure in the spectrum of HD 23850 is stronger than that in HD 23441 distinctly opposite to the behaviour of CH⁺ lines. Moreover, the CH⁺ structures in both spectra are much stronger than the CH ones. Taking into account the intensities of CH⁺ lines, the brightest Pleiades can be divided into two distinct groups: with strong and weak CH⁺ lines, respectively. This division is in full agreement with a similar division made in the paper by Ritchey et al. (2006). These groups are shown in Fig. 9 in the plane of CH⁺ intensity versus that of the K i line.

However, the sky-projected distribution of both CH⁺ groups, presented in Fig. 10, does not show a significant difference (although some of the ‘CH⁺ strong’ stars seem to form a compact and isolated group north-west of the cluster centre). The ‘CH⁺ weak’ stars are distributed over the entire field but mainly at the west side of the cluster. This is in full accordance with the conclusions of Ritchey et al. (2006). Low statistics do not allow us to make any more definite conclusions concerning the spatial distribution of both CH⁺ groups.

There is, however, another, although slight, difference between the groups. Stars with strong CH⁺ lines exhibit a little higher heliocentric radial velocities (+17.8 km s⁻¹) in comparison with the CH⁺ weak ones (+16.4 km s⁻¹), as shown in Fig. 11.

This effect may possibly reflect the differences in the kinematics of gas clouds in front of particular group members, as suggested by White (2003) and Ritchey et al. (2006). The presented division of stars according to CH⁺ intensities does not correlate with the effect of ‘grey extinction’. This may indicate that the latter effect is indeed a purely circumstellar one.

Hobbs (1973) already mentioned that the CH⁺ lines, observed in Pleiades, are slightly redshifted in comparison to Ca ii lines. We have checked whether this kind of shift can be observed in our spectra as well. Fig. 12 demonstrates the same effect in the spectrum of Asterope (HD23432) acquired using the Maestro spectrograph of the Terskol Observatory. In this object, the intensities of the Ca ii ‘K’ line and the neighbouring CH⁺ 3957 Å feature are nearly the same.

Looking for another possible peculiarity of the properties of interstellar matter within the Pleiades region, we found that the intensities of the diffuse interstellar lines towards the Pleiades stars do not differ from those in the general field. Fig. 13 shows fragments of spectra of two Pleiades stars with interstellar lines Na i D₁ and D₂ and diffuse interstellar features at 5780, 5797, 6196, 6205, and 6614 Å. As is clearly seen, the spectra of both stars, which exhibit very different CH and CH⁺ (Fig. 8) features, show no differences either in interstellar diffuse features or the Na i lines.

All this suggests that the interstellar matter within the Pleiades region does not deviate from the interstellar matter in the general field, judging from their diffuse lines.

5 CONCLUSIONS

The above considerations led us to the following conclusions:

• the Pleiades cluster seemingly is an elongated structure where the brightest members are situated closest to the Sun

• the central parts of the cluster are likely filled with diffuse matter populated by relatively large grains; these grains likely produce the ‘grey extinction’ which influences spectrophotometric distances to the Pleiades

• Gaia DR2 and Ca ii distances to the Pleiades agree very well

• spectra of certain Pleiades’ members exhibit extraordinarily strong CH⁺ lines while other molecular features and atomic lines are either very weak or below the detection level

• the intensity of CH⁺ lines allows one to divide the Pleiades objects into two distinct groups; not only intensities but also radial velocities of the two groups are evidently different

• the radial velocities of Ca ii lines and CH⁺ features are slightly different, as reported by Hobbs (1973); this effect can also be found in Ritchey et al. (2006) and is fully confirmed here

• other interstellar features (diffuse bands) are observed in Pleiades objects, though they are very weak and thus – seen only in high S/N ratio spectra.

ACKNOWLEDGEMENTS

JK acknowledges the financial support of the Polish National Science Centre during the period 2015–2019 (grant 2015/17/B/ST9/03397) and the grant UMO-2017/25/B/ST9/01524 for the period 2018–2021. GAG, JK, and GGV acknowledge the Chilean fund CONICYT grant REDES180136 for financial support of their international collaboration.

This paper includes data gathered with the VLT (UVES) ESO telescope and the 3.6 m (ESpADoNs) CFH telescope. This work has made use of data from the European Space Agency (ESA) mission Gaia (https://www.cosmos.esa.int/gaia), processed by the Gaia Data Processing and Analysis Consortium (DPAC, https://www.cosmos.esa.int/web/gaia/dpac/consortium). Funding for the DPAC has been provided by national institutions, in particular the institutions participating in the Gaia Multilateral Agreement..

Footnotes

REFERENCES