EDP Sciences logo
Subscriber Authentication Point
Search
Menu
Home All issues Volume 687 (July 2024) A&A, 687 (2024) A102 Full HTML
Open Access
Issue
A&A
Volume 687, July 2024
Article Number A102
Number of page(s) 12
Section Interstellar and circumstellar matter
DOI https://doi.org/10.1051/0004-6361/202348953
Published online 02 July 2024

© The Authors 2024

Licence Creative CommonsOpen Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This article is published in open access under the Subscribe to Open model. Subscribe to A&A to support open access publication.

1 Introduction

New telescopes, instruments, and sky surveys in low-frequency radio astronomy (Wayth et al. 2015; Shimwell et al. 2017, 2019, 2022; Wolleben et al. 2019) are poised to greatly expand studies of the diffuse interstellar medium (ISM). Rotation measure (RM) synthesis emphasizes the information encoded in the observations by distinguishing polarized emission along the line of sight (LoS) based on the Faraday rotation it has undergone (Brentjens & de Bruyn 2005; Van Eck 2018). This data analysis technique, which provides data cubes of polarized intensity versus Faraday depth, is referred to as Faraday tomography. The data cubes reveal an unexpected structural richness in the polarized Galactic synchrotron emission that offers information on the 3D structure of interstellar magnetic fields and electrons (Iacobelli et al. 2013; Jelić et al. 2014, 2015; Lenc et al. 2016; Van Eck et al. 2017, 2019; Thomson et al. 2019; Turić et al. 2021; Erceg et al. 2022). When interpreting the data, we are faced with our limited understanding of the ionization of the diffuse ISM and, in particular, with the difficulty in identifying the ISM component to which the Faraday structures relate.

Correlation with ISM tracers provides clues on the nature of Faraday features. Comparison of LOFAR observations with Planck maps have revealed a tight correlation between the orientation of structures in Faraday data and the magnetic field component in the plane of the sky traced by dust polarization in a few mid-latitude Galactic fields located in the surroundings of the extragalactic point source 3C 196 (Zaroubi et al. 2015; Jelić et al. 2018; Turić et al. 2021). LOFAR structures have also been observed to correlate with H I filaments associated with the cold neutral medium (Kalberla & Kerp 2016; Kalberla et al. 2017; Jelić et al. 2018) and, more generally, with maps of ISM phases (cold, unstable and warm gas) derived from the analysis of H I spectral data cubes (Bracco et al. 2020). The observational results are limited because the published data comparisons focus on a few fields close together in the sky1. However, they point to a common physical framework in which the LOFAR structures trace the neutral dusty ISM, and not the warm ionized medium (WIM). To test this hypothesis, we present this work in which we estimate the RM associated with the warm neutral medium (WNM) in the local ISM.

The UV spectroscopic data in absorption toward background stars indicate that the ionization fraction of the WNM is about 0.1, in the thick Galactic disk (Spitzer & Fitzpatrick 1993), the local ISM (Jenkins 2013) and small clouds of warm gas embedded within the Local Bubble2 (LB, Redfield & Falcon 2008), in particular the local interstellar cloud encompassing the Sun (Slavin & Frisch 2008; Gry & Jenkins 2017; Bzowski et al. 2019)). The WNM, rather than the WIM (Reynolds et al. 1998), could be the specific ISM component traced by low-frequency Faraday observations at high Galactic latitude, as proposed by Heiles & Haverkorn (2012) and Van Eck et al. (2017), and supported by the numerical simulations presented by Bracco et al. (2022).

This work brings together the analysis of the LOFAR Two-meter Sky Survey (LoTSS, Shimwell et al. 2022), the mosaic at high latitudes toward the outer Galaxy analyzed by Erceg et al. (2022), with UV spectroscopy and dust polarization observations. We combine electron column densities derived from UV spectroscopic observations toward stars with a model of the local magnetic field fitted on Planck dust polarization maps to estimate the RM of the WNM in the local ISM. The comparison with Faraday spectra on the same LoS allows us to assess the WNM contribution to the Faraday rotation of the polarized synchrotron emission measured by LOFAR.

The paper is organized as follows. We introduce stellar targets and present the UV observational data in Sec. 2. The model of the local magnetic field that we use to estimate the RM from electrons in the WNM is presented in Sec. 3 and Appendix A. The comparison with LOFAR observations is presented in Sec. 4. Our work brings a new perspective to the interpretation of LOFAR Faraday data: the origin of the structures and the contribution from the WIM, which we discuss in Sec. 5. The main results are summarized in Sec. 6.

thumbnail Fig. 1

Stars in our sample overlaid on a LOFAR image of the LoTSS field from Erceg et al. (2022). The image displays the first moment of the Faraday spectra in units of rad m−2. Celestial and Galactic coordinates are plotted with solid and dashed lines, respectively.

2 Stellar targets and data

We introduce our sample of stellar targets with spectroscopic UV observations in Sec. 2.1. Our analysis of the UV spectra is presented in Sec. 2.2. In Sec. 2.3, the results of our data fitting are used to derive column densities of electrons in the WNM.

2.1 Spectroscopic UV observations

To build this project, we searched scientific publications for UV spectroscopic observations, at high Galactic latitudes over the sky area covered by the LOFAR Faraday tomography mosaic (hereafter the LoTSS field) analyzed by Erceg et al. (2022). We obtained a sample of nine stellar sources. This sample includes eight stars from the study of Jenkins (2013) plus the halo star HD 93521 (Spitzer & Fitzpatrick 1993). To this set, we added the star HZ43A (Kruk et al. 2002), which lies close to the edge of the LoTSS field. This star, located 60 pc from the Sun, allows us to probe the contribution of the very nearby gas in this region of the sky.

In Fig. 1, the stellar positions are overlaid on an image of the first moment of the Faraday data cube from Erceg et al. (2022), representing the polarized intensity-weighted mean of Faraday spectra in units of rad m−2. All stars have been observed with the Far Ultraviolet Spectroscopic Explorer (FUSE) with a spectral resolution R ≡ ν/Δν ≃ 2 × 104.

Stellar coordinates and data relevant to this study are listed in Table 1. The parallaxes and corresponding distances are from the Gaia data release eDR3 (Lindegren et al. 2021). The column densities of Ar I and O I are inferred from the FUSE UV spectra. We follow Jenkins (2013) using the [Ar I/O I] column density ratio, normalized to Solar abundances, to estimate the ionization fraction of the WNM. The ionization potentials of Ar I and H I are close, but the photo-ionization cross section of Ar I is much larger. The ionization fraction of O I is strongly locked to that of H I through a strong charge-exchange reaction. These two factors make [Ar I/O I] a sensitive tracer of H I ionization in the WNM (Sofia & Jenkins 1998). Since neutral forms of argon and oxygen are virtually absent in fully ionized regions (either the prominent H II regions around hot stars or the much lower density but more pervasive WIM), our probes sample only regions that have appreciable concentrations of H I. This sets our measurements apart from conventional determinations of average electron densities (e.g. pulsar dispersion measures, Hα line intensity, C II fine-structure excitation), which are strongly influenced by contributions from fully ionized gas.

Table 1

Stellar targets and results of the data fit.

2.2 Data analysis

The column densities of OI and Ar I (number of absorbing atoms in a column of unit cross-section along the LoS) are determined by a curve of growth analysis (Strömgren 1948; Draine 2011). The curve of growth relates the gas column densities of the absorbers, N, to the strength of absorption lines measured by their equivalent width Wλ (normalized area of the line expressed in wavelength units). The function Wλ(N) is computed assuming that the absorbers have a Gaussian velocity distribution characterized by the Doppler width , where συ is the velocity dispersion. By fitting the equivalent widths of multiple lines of different strength one can infer the two parameters N and b from the data.

We used the equivalent widths of multiple OI lines and a single Ar I line, measured by Jenkins (2013), and by Kruk et al. (2002) for HZ43A. The oscillator strengths of the lines are listed in these two papers. The fit of the curve of growth is performed with three free parameters: the column density N(O I) of O I atoms, the abundance ratio [Ar I/O I] normalized to the Solar value, and the Doppler parameter, b, of the line profiles, assumed to be the same for Ar I and O I. The spontaneous decay rate that determines the Lorentzian broadening is set to A = 3 × 107 s−1 ignoring variations among O I states (Morton 2003). This simplification has no significant impact on the results of our data analysis.

We performed a Monte-Carlo Markov chain (MCMC) analysis to sample the joint posterior distribution of the three model parameters. We utilize the EMCEE python package as described by Foreman-Mackey et al. (2013) in accordance with their recommendations. We use a Gaussian likelihood and broad priors on the model parameters. We restricted b to values greater than 2 km s−1 because we consider that in our LoS the O I lines are dominated by warm gas (the partially ionized WNM), which implies a temperature of T ≥ 6000 K and, thus, a line broadening b > 2 km s−1. To confirm this hypothesis we looked at the H I 21 cm spectra in direction of each of our stars using the HI4PI data from HI4PI Collaboration (2016) and their Gaussian decomposition by Kalberla & Haud (2018). These data indicate that some of the gas along our LoS is cold but its contribution to the OI absorption lines is minor due to line saturation. The cold neutral medium (CNM) could contribute to very weak transitions but the OI lines that have the most influence in our analysis have equivalent widths that are not much different from that of the Ar I line. In this case, comparisons of Ar I to OI remain valid, and the very narrow components from the CNM have no influence. Moreover, fits to the data for small values of b suppose that the lines are highly saturated and give O I column densities that are incompatible with H I column densities inferred from H I 21 cm observations. Figure 2 presents the posterior distributions of the model parameters for two stars Feige 34 and PG1544+488 chosen as examples of low and high column densities of O I. The corresponding data fits with the curves of growth are shown in Fig. 3.

The model parameters with their error bars are listed in Table 1. The best estimate is the median value of the posterior distribution, and the negative and positive error bars are computed from the 16 and 84 percentiles. We added to Table 1 data for the star HD 93521. Interstellar matter foreground to HD 93521 has been analyzed in detail by Spitzer & Fitzpatrick (1993). Here, we report the OI column density and the [Ar I/O I] abundance ratio that we measured in the FUSE spectra for the low-velocity component. We checked that our estimates of [Ar I/O I] are consistent with the values reported by Jenkins (2013) but for PG1610+5193. For this parameter of the model, the added value of our data analysis is the posterior distribution that quantifies the interdependence with N(O I) and b.

2.3 Column densities of electrons

The column densities of electrons for the WNM, foreground to the stars are estimated from the results of our fits of FUSE spectra. We follow the approach introduced by Jenkins (2013), to derive the ionization fraction of HI, xH ≡ n(H+)/n(H I), from the [Ar I/H I] abundance ratio. We use formula (7) in this paper: (1)

where P′Ar is a unit-less parameter, which accounts for various forms of photoionization and recombination with free electrons, as well as additional processes (Jenkins 2013). [Ar I/H I] is deduced from [Ar I/O I] for the Solar abundance of oxygen log[O/H] + 12 = 8.76. We assume that there is no depletion of oxygen on dust grains, which should be safe for the low-density conditions we are sampling here (Jenkins 2009). We use P′Ar = 22.8, the value from Table 4 in Jenkins (2013) for a model including ionization by soft X-ray emission from cooling supernova remnants. The statistical distribution of xH is derived from the posterior distribution of [Ar I/O I] using Eq. (1), and that of the HI column density N(H I) from N(O I).

We combine xH and N(H I) values to compute the electron column density: (2)

where the factor 1.2 accounts for electrons from ionized helium, based on the ionization model of the WNM in Jenkins (2013).

The column densities of neutral hydrogen and electrons are listed in Table 2 with their error bars and plotted versus distance in Fig. 4. and N(H I) seem to be correlated: <N(H I) / > = 10.9 with a small dispersion of 3.6. We observe a tendency for N(H I) to increase with distance but our star sample is too small to draw some firm conclusion about the distance distribution of WNM electrons.

The probability distribution functions (PDFs) of the mean column density of electrons and the ionization fraction averaged over our set of stars are presented in Fig. 5. The PDFs are computed by averaging the samples of the posterior distributions, inferred from our data analysis of UV spectra, over the nine stars in Table 2 more distant than HZ43A. The nearest star in this set is located 140 pc from the Sun. The median values for the electron column density and the ionization fraction are and , where the lower and upper error bars represent the 16 and 84 percentiles of the distributions. Our median value is consistent with that inferred by Jenkins (2013) from the complete sample of his study. For = 13.4, the model value without the contribution of cooling supernovae (Jenkins 2013), we find ⟨xH⟩ = 0.17. The difference quantifies the dependence of our ⟨xH⟩ estimates on the choice of ionization model.

The various sources and ionization processes that contribute to the ionization fraction of the WNM are detailed by Jenkins (2013). These include cosmic rays as well as supernova-driven shocks (Sutherland & Dopita 2017) and soft X-ray emission from cooling supernova remnants (SNRs, Slavin et al. 2000). The degree of ionization of the WNM could also be locally out of equilibrium as an aftereffect of the supernova explosions that have created the LB (Fuchs et al. 2006; Zucker et al. 2022). Within this hypothesis, the WNM ionization could be enhanced around the LB. Analysis of UV data for a larger sample of stars is required to test this hypothesis.

thumbnail Fig. 2

Corner plots of MCMC data fit for two stars in our sample Feige 34 (left) and PG1544+488 (right), chosen as examples of low and high column densities. The histograms show the probability distribution of the three model parameters, and the contours display their interdependence. The dashed lines plotted over the histograms represent the median values of the model parameters.
thumbnail Fig. 3

Curves of growth for the two stars Feige 34 (left) and PG1544+488 (right), chosen as examples of low and high column densities. For each graph, the y-axis represents the ratio between the equivalent width, Wλ, and the wavelength, λ, of the line, and the x-axis the product of the column density N expressed in cm−2, the strength of the transition (unit-less f-value), and λ in Å. The blue circles represent the equivalent widths of the O Ilines used to determine the column density of OI and the Doppler parameter, b, and the red star the equivalent width of the Ar I line used to determine the [Ar I/O I] abundance ratio.
Table 2

Column densities and ionization fractions.

thumbnail Fig. 4

Column densities of HI and WNM electrons. The HI column densities are plotted with blue circles and the electron column densities in the WNM with magenta squares. The two triangles pointing down represent upper limits.

3 Model of the local magnetic field from Planck observations

To estimate the RM from the partially ionized WNM foreground to the stars, we need a model of the magnetic field in the local ISM. The tight correlation between the orientation of Faraday structures in LOFAR data with that of dust polarization (Zaroubi et al. 2015; Jelić et al. 2018) leads us to use the phenomenological model introduced by Planck Collaboration Int. XLIV (2016) and Vansyngel et al. (2017) to statistically model dust polarization data at high Galactic latitudes as measured by Planck.

The magnetic field B is expressed as the sum of its mean (ordered) B0 and random (turbulent) Bt components: (3)

where and are the unit vectors associated with B0 and Bt, and fM a model parameter that sets the relative strength of the random component of the field. The direction of B0 is assumed to be uniform over the sky area used to fit the model parameters. The model also makes the simplifying assumption that the direction of B0 does not vary along the LoS. At high Galactic latitudes (|b| ≥ 60°), comparison with stellar polarization data indicates that dust polarization measured by Planck originates mainly from the surroundings of the LB, within about 300 pc from the Sun (Skalidis & Pelgrims 2019). Even if the LOFAR field in Fig. 1 extends to lower latitudes, it is not necessary to account for the structure of the magnetic field across the Galactic disk and into the halo to model Planck dust polarization data over this sky area.

The computational steps taken to determine the magnetic field model are detailed in Appendix A. The direction of is defined by the Galactic coordinates l0 and b0, which are derived from a fit of the Planck Stokes Q and U maps at 353 GHz over the sky regions toward the northern Galactic cap comprising the LoTSS field. The random component Bt is statistically described using the model of Planck Collaboration Int. XLIV (2016) with parameters derived from a fit of the Planck power spectra of dust polarization by Vansyngel et al. (2017). The orientation of Bt varies in discrete steps along the LoS over a small number of layers, which schematically represent the correlation length of the turbulent component of the magnetic field. The ratio between |Bt| and |B0|, fM = 0.9, is determined fitting Planck power spectra of dust polarization (Vansyngel et al. 2017).

Because the modeling of dust polarization only determines the orientation of the mean magnetic field, we need to use Faraday observations to determine the field strength and also to discriminate between the two opposite directions associated with the orientation4. Comparison of the rotation and dispersion measures of pulsars in the northern sky indicates that the mean field strength in diffuse ionized gas in the Galactic disk near the Sun is about 3–4 μG (see Table 3 in Sobey et al. 2019). The magnetic field strength in the warm interstellar medium can also be estimated from synchrotron emission. As discussed in Beck et al. (2003), the two estimates differ if fluctuations in magnetic field and electron density are not statistically independent. For the specific objective of the paper, the strength of the mean field estimated from pulsar RMs is a directly relevant reference. We chose this strength |B0| to be 3.5 μG. The total field strength that includes the turbulent component is 5 μG.

To estimate the RM from electrons in the WNM, we compute the mean magnetic field component along the LoS, BLoS: (4)

where Bt is averaged over the model layers and is the unit vector defining the LoS. The minus sign makes BLoS positive when it points toward the Sun. Figure 6 presents our calculation results in the LOFAR field. The top image shows the mean value of BLoS. This map represents BLoS computed for the mean magnetic field B0, ignoring the random component. The bottom image shows BLoS for one statistical realization of Bt. The mean value of BLoS and the dispersion associated with Bt are listed in Table 3.

thumbnail Fig. 5

Probability distribution of the mean value of the electron column density (left) and ionization fraction (right). The histograms are built averaging the samples of the posterior distributions over the 9 stars in our sample more distant than 100 pc (i.e. excluding HZ43A). The solid black lines indicate the median value and the dashed lines the 16% and 84% percentiles.
Table 3

BLoS and WNM RMs.

thumbnail Fig. 6

Maps of the magnetic field component along the LoS, BLoS, over the LOFAR field. The top image shows BLoS for the mean field B0 and the bottom image that computed for B including one realization of the turbulent component Bt. BLoS is positive when it points toward the Sun (see Eq. (4)).

4 The warm neutral medium and LOFAR Faraday observations

In this section, we assess whether the partially ionized WNM foreground to the stars may account for the LOFAR Faraday spectra. The RMs associated with electrons within the WNM foreground to the stars are determined in Sec. 4.1, and compared with the LOFAR data in Sec. 4.2. Electrons in the WIM are ignored in this section.

4.1 Rotation measures from WNM electrons

The RM to a source at a distance d is defined as (5)

where the integral along the LoS goes from the source at l = 0 to the observer at l = d, ne is the electron density and B∥ the magnetic field component along the LoS5. Using BLoS, the mean value of B∥ introduced in Eq. (4), we reduce the integral in Eq. (5) to (6)

where RMWNM(i) is the RM associated with the column density of electrons (i) to the star i. Within our statistical model of the magnetic field (Sec. 3), BLoS for a given LoS has a normal distribution. The mean values and standard deviations of BLoS are listed for the LoS to the stars in Table 3. The PDF of BLoS is combined with the posterior distribution of , inferred from our analysis of the FUSE spectra, to obtain the statistical distribution of RMWNM. The median values of RMWNM with 1σ uncertainties are listed in Table 3. These uncertainties are large for all stars because they include the statistical dispersion of the turbulent component of BLoS.

thumbnail Fig. 7

PDFs of the mean RM and the mean LOFAR M1 moment. The plot displays the PDFs of ⟨RMWNM × 0.5⟩ (red) and ⟨M 1LOFAR⟩ (blue). The PDF of ⟨RMWNM⟩ is also shown as a black line. The mean values are computed over the eight stars with both UV and LOFAR data.

4.2 Comparison with LOFAR data

We now proceed to the comparison between RMWNM and the LOFAR data.

The third column of Table 3 lists the first moment of the Faraday spectra for each LoS to the stars. The first moment M1LOFAR is the mean in Faraday depth of the spectra, weighted by the polarized intensity. The M1LOFAR values were computed by Erceg et al. (2022) for a polarized intensity threshold of 460 μJy per angular and Faraday resolution elements. No values could be measured for HD 93521 because the polarized emission at this position is too weak to be detected. HZ43A is just outside the field of the LoTSS field in Fig. 1.

To reduce the uncertainties in the observational data and the statistical dispersion of the turbulent component of the magnetic field, we compared the mean values of RMWNM and M1LOFAR, computed for the eight stars with the LOFAR values in Table 3. The PDFs of ⟨RMWNM⟩ are computed by combining the statistical distributions of and BLoS. The PDF of ⟨M1LOFAR⟩, is calculated from the LOFAR data points using bootstrap statistics (Efron 1979). We drew thousands of samples of eight values from the data points. Each value in a sample is one of the data points drawn at random, and a given sample may contain the same data point several times. We compute the mean value for each sample. The PDF that we obtain accounts for the data scatter and the sample variance.

In Fig. 7, the PDFs of the mean RM and M1LOFAR are compared. We discuss this graph in the reference framework introduced by Burn (1966) and consider two idealized cases: (1) a Faraday foreground screen of WNM electrons that rotates the polarization angle of a background synchrotron emission, and (2) a Faraday slab where the polarized synchrotron emission is distributed along the LoS together with the WNM electrons. In the first case, we expect the mean values of RMWNM and M1LOFAR to be the same. In the second case, for a symmetric distribution of thermal electrons and polarized synchrotron emission between the near and far sides of the slab, we expect the mean of RMWNM to be twice that of M1LOFAR (Sokoloff et al. 1998).

We find a good match between ⟨RMWNM × 0.5⟩ and ⟨M1LOFAR⟩, which favors case (2). Individual data points for each star are compared and plotted versus distance in Fig. 8. The large error bars on the data points prevent a detailed assessment. We only point out that this plot does not show a systematic trend with distance. The data comparison does not include the systematic uncertainty on electron column densities associated with our choice for in Sec. 2.3. We note that within the ionization models quantified by Jenkins (2013) the alternative choice for would produce a mean difference between the WNM RMs and the first moment of the LOFAR spectra greater than a factor of 2.

thumbnail Fig. 8

Comparison of the WNM RMs with the first moments of the LOFAR Faraday spectra for each star individually. The values of RMWNM × 0.5 are plotted with blue squares, and the intensity weighted Faraday depths from LOFAR, M1LOFAR, with red circles.

5 Insight into the interpretation of LOFAR Faraday data

Our work contributes to a new perspective on the interpretation of LOFAR Galactic polarization data. We discuss two specific questions: the origin of the LOFAR Faraday structures in Sec. 5.1 and the contribution of the WIM in Sec. 5.2.

thumbnail Fig. 9

PDFs of the mean RM and the mean LOFAR M1 moments for synthetic observations derived from a numerical simulation by Bracco et al. (2022). The mean values are computed on random sets of eight LoS. The PDFs of ⟨RMWNM × 0.5⟩ and ⟨M1LOFAR⟩ are plotted in red and blue. The PDF of ⟨RMWNM⟩ is also shown as a dashed line.

5.1 Origin of LOFAR Faraday structures

Comparison of the UV and LOFAR data in Fig. 7 suggests that the relativistic electrons producing the synchrotron-polarized emission detected by LOFAR and the thermal electrons in the WNM contributing to Faraday rotation are mixed within the same ISM slab. We take the median value of ⟨RMWNM⟩ as the mean RM across this slab, RM0 ~ 0.7 rad m−2. Because RM0 is smaller than the resolution in Faraday depth of the LoTSS data (1.2 rad m−2), the synchrotron emission from the near and far sides of the slab are not separated through the LOFAR RM-synthesis.

In this context, it is relevant to mention the spatial correlation between polarized dust emission measured by Planckat 353 GHz and the polarized synchrotron emission measured by Planck and WMAP at 30 and 23 GHz (Planck Collaboration Int. XXII 2015; Choi & Page 2015; Planck Collaboration XI 2020). The correlation coefficient, measured to be approximately 20%, provides an estimate of the fraction of polarized synchrotron emission that comes from the same ISM slab as the WNM, given the tight correlation between dust polarization and HI at high Galactic latitudes (Ghosh et al. 2017; Clark & Hensley 2019). The remaining 80% of the polarized synchrotron emission would be background to the WNM electrons. We hypothesize that the background emission at LOFAR frequencies undergoes depolarization due to differential Faraday rotation along the LoS. This idea was previously proposed by Van Eck et al. (2017) and examined in a separate context by Hill (2018). If this view holds, deep LOFAR observations should reveal weak diffuse polarized emission at higher absolute Faraday depths than the M1LOFAR values reported by Erceg et al. (2022). Analysis of the LoTSS ELAIS-N1 Deep Field at high Galactic latitude, which combines 150 h of LOFAR observations, does support this hypothesis. Šnidarić et al. (2023) successfully detected faint Galactic polarized emission at higher absolute Faraday depths, probing the full range of RMs expected from the Galaxy.

Faraday rotation and depolarization produced by a slab of magnetized ISM have been studied analytically by Sokoloff et al. (1998). The polarized emission depends on the relative distribution of the thermal electrons that produce the Faraday rotation and the relativistic electrons that produce the synchrotron emission. Together with the differential Faraday rotation along the LoS, these factors contribute to the variations of the observed polarized intensity. However, this formal framework ignores the bandwidth of the LOFAR observations.

To put the data comparison presented in Fig. 7 into an astrophysical context, we confront our results with those obtained from MHD simulations. To this end, we perform the same comparison between WNM RMs and first moments of Faraday spectra using synthetic observations built from numerical simulations of two colliding shells as input (Ntormousi et al. 2017). We use the synthesized LOFAR maps produced and analyzed by Bracco et al. (2022). Specifically, we use their Case A data where the shells collide in a direction perpendicular to the initial magnetic field for their high value of the cosmic-ray ionization rate ζH = 2.6 × 10−16 s−1. In these simulations, most of the Faraday rotation occurs within partially ionized WNM gas. The synthesized LOFAR maps have morphological features similar to those revealed by LoTSS data, including depolarization features.

In Fig. 9, we compare the PDFs of the mean RM and the mean first moment, computed on these synthesized LOFAR maps, for a large number of random sets of eight LoS. The similarity to the corresponding plot in Fig. 7 is striking because the simulations were not designed to model our observational work. Additional work is required to investigate the link between Faraday structures and the dynamics of the magnetized multiphase ISM. In particular, it would be interesting to take into account our specific view point within the LB as done for the dust polarization by Maconi et al. (2023), as well as the impact of the supernovae at the origin of the LB (Fuchs et al. 2006) on the ionization of the local WNM gas (de Avillez et al. 2020) and the associated synchrotron emission.

5.2 Contribution from the warm ionized medium

We broaden the context of our work by comparing the column density of electrons in the WNM with that in the WIM. To this purpose, we collected dispersion and rotation measures of pulsars located within the LoTSS field, from the pulsar catalog presented by Manchester et al. (2005)6. The data are listed in Table 4. The catalog does not provide error bars, but those are specified for several of these pulsars by Sobey et al. (2019). These error bars are not significant for our analysis because they are much smaller than the scatter between individual values.

The mean column density of electrons computed from the dispersion measures is = 4.2 ± 0.6 × 1019 cm−2, where the error bar is estimated using the bootstrap method (Efron 1979). is much larger than (Sec. 2.3). The difference highlights the contribution of the WIM to pulsar dispersion measurements. It indicates that the total column density of electrons within the LoTSS field is dominated by the contribution from the WIM.

Pulsar RMs that range from −15 to 10 rad m−2 (Table 4) have a wider spread than the LOFAR first moments in the LoTSS field. This observational fact suggests that the WIM is the main source of the total RM of the Galaxy within the LoTSS fields, although this contribution is not easily detectable at LOFAR frequencies. Erceg et al. (2022) compared the first LOFAR Faraday moment with those reported by Dickey et al. (2019) using data at higher frequencies (1270–1750 MHz) from the Dominion Radio Astrophysical Observatory (DRAO) and the Galactic RM sky map, derived from observations of polarized extragalactic sources (Hutschenreuter et al. 2022). They find that the range of values for the LoTSS first Faraday moment is much smaller than that reported for the DRAO data and the total Galactic RM (see their Fig. 14).

To account for these results within our astrophysical framework, we propose that the WNM electrons are foreground to the bulk of the WIM electrons. To show that this condition is met, we estimate the distance LWNM over which the WNM electrons are distributed. Combining the electron column density and ionization fraction derived from our data analysis, we obtained: (7)

where and fWNM are the hydrogen density and volume filling factor of the WNM. Using standard values for these two quantities (Wolfire et al. 1995), we find: (8)

This estimate of LWNM is much smaller than the scale-height of the WIM ~1.8 kpc (Gaensler et al. 2008). We also note that it is consistent with the distance estimate for LOFAR structures derived from stellar extinction data by Turić et al. (2021) in the 3C196 field, which is close to the LoTSS field in the sky.

The brightness of diffuse Hα emission toward the stars could give some guidance on the WIM in the LoTSS field, but these observations combine foreground and background contributions. Appendix B presents an attempt to estimate the emission measure (EM) of ionized gas foreground to the stars, using absorption lines in FUSE spectra of ionized nitrogen in its second excited fine-structure state. The FUSE spectra provide upper limits and two detections for HD 113001 and Ton102, which are all significantly higher than the EM of the full LoS derived from the Hα emission. Thus, these EM estimates prove to be of very little significance for our study.

Table 4

Data for pulsars within the LoTSS field.

6 Summary

We have analyzed stellar UV spectroscopic observations that provide us with estimates of electron column densities in the WNM foreground to nine stars within the sky area of the LoTSS mosaic presented by Erceg et al. (2022). The stellar distances range from 140 to 1350 pc. The UV data are combined with a model of the local magnetic field fitted to Planck maps of dust polarization. We obtain estimates of the RM to the stars from electrons in the WNM, which we compare to LOFAR Faraday spectra. We draw the following results from the data analysis and comparison:

  • The mean electron column density in the partially ionized WNM foreground to the stars is . The mean ionization fraction of hydrogen is ;

  • The first moment of the LOFAR spectra at the positions of our stars are on average half the WNM RMs. This is the result expected for a slab of magnetized ISM where the thermal electrons that produce the Faraday rotation and the relativistic electrons that radiate synchrotron emission are mixed with a roughly symmetric distribution along the LoS;

  • The mean RM across the WNM slab is RM0 ~ 0.7 rad m−2. As RM0 is smaller than the resolution in Faraday depth of the LoTSS data, the synchrotron emission from the near and far sides of the slab are not separated through the LOFAR RM-synthesis;

  • We broadened the context of our work by comparing the column density of electrons in the WNM with dispersion measures of pulsars located within the same LoTSS field. The mean electron column density computed from the dispersion measures is about one order of magnitude larger than . This difference highlights the contribution of the WIM to pulsar dispersion measures.

  • Although the WIM constitutes the majority of the total electron column density in the LoTSS field, its impact on the RM at LOFAR frequencies is elusive. To account for this observational fact, we propose that the WNM electrons and the associated polarized synchrotron emission are foreground to the bulk of the WIM.

Our work sheds new light on the interpretation of the LOFAR Galactic polarization data. This suggests that LOFAR Faraday structures are associated with neutral, partially ionized, WNM. Pending confirmation with a larger sample of stars when the full LOFAR northern sky maps are published, this preliminary finding lays an astrophysics framework for exploring the relationship between Faraday structures and the dynamics of the magnetized multiphase interstellar medium. Independently of the LOFAR data, the analysis of UV FUSE data for a larger sample of stars over the whole sky can provide valuable information on the distance distribution of WNM electrons in the local ISM.

Acknowledgements

This research was supported by the COGITO No. 42816 exchange program between France and Croatia. AB and FB acknowledge the support of the European Research Council, under the Seventh Framework Program of the European Community, through the Advanced Grant MIST (FP7/2017-2022, No. 787813). AB also acknowledges financial support from the INAF initiative “IAF Astronomy Fellowships in Italy” (grant name MEGASKAT). We thank M. Alves, K. Ferrière and P. Lesaffre for helpful discussions, and the anonymous referee whose report helped us significantly improve the original version of the article.

Appendix A Magnetic field model

This appendix details the derivation of the magnetic field model that we use to estimate the Faraday depth from the WNM in Sec. 4. The magnetic field model in Eq. 3 is written as a sum of the mean magnetic field B0 and a random component Bt. We describe how we determine the direction of B0 in Sect. A.1 and how we compute statistical realizations of Bt in Sect. A.2. The model presented here was introduced by Planck Collaboration Int. XLIV (2016) and Vansyngel et al. (2017). This appendix explains how the specific model we use in this paper was obtained.

A.1 Mean magnetic field

A.1.1 Dust polarization on large angular scales

The direction of B0 is assumed to be uniform. It is defined by the Galactic coordinates l0 and b0 of the unit vector . Introducing the LoS unit vector , we derive the component of along the LoS, , and in the plane of the sky, : (A.1)

To compute the Stokes parameters I, Q, and U, we start from the integral equations in for instance Sect. 2 of Regaldo-Saint Blancard et al. (2020). Within the assumption of a uniform , we get: (A.2)

where P ≡ (Q2 + U2)0.5 is the polarized intensity, p0 is a parameter quantifying the intrinsic fraction of dust polarization, γ is the angle that B0 makes with the plane of the sky, and ψ the polarization angle. We note that Planck Collaboration Int. XLIV (2016) used a simplified version of these equations without P. The angles γ and ψ are computed from the two following equations: (A.3)

where is the unit vector perpendicular to within the plane ( is the unit vector pointing toward the North Galactic Pole).

A.1.2 Model fit

Our model of dust polarization on large angular scales defined by Eqs. A.2 and A.3 has three parameters: l0, b0 and p0. We introduce the Planck data and explain how the fit is performed to determine the parameter values with their error bars.

We use the Planck polarization maps at 353 GHz produced by the SRoll2 software7 (Delouis et al. 2019). This data release improves the PR3 Legacy polarization maps by correcting systematic effects on large angular scales. We use Healpix all-sky maps at a resolution Nside = 32 corresponding to a pixel size of 1.8° (Górski et al. 2005).

Eqs. A.2 and A.3 may be combined to compute model maps of the Q/(I + P) and U/(I + P) ratios on a full-sky Healpix grid. This model has three parameters p0, l0 and b0.

At Nside = 32, the structure associated with the turbulent component of the magnetic field dominates the data noise and the Cosmic Microwave Background. As this uncertainty scales with (I+P), we compute and fit Q/(I + P) and U/(I + P) maps giving equal weights to all pixels within the unmasked sky area.

To estimate the error bars of the parameters, we use mock data. The input Stokes maps of dust polarization are those introduced in Appendix A of Planck Collaboration III (2020). We build a set of simulated maps by adding the CMB signal and independent realizations of the data noise from end-to-end simulations of SRoll2 data processing (Delouis et al. 2019) to these input maps. The dispersion of the parameter values provides the error bars on l0 and b0. The error bar in p0 is dominated by the uncertainty in the offset correction of the total intensity map, which we take to be 40 μK as in Planck Collaboration XII (2020). We perform the fit over three high Galactic latitude areas in the northern hemisphere L42N, L52N and L62N introduced by Planck Collaboration XI (2020). The three areas encompass the LOFAR field in Fig. 1. The sky fraction is fsky = 0.24, 0.28 and 0.33 for LR42N, LR52N and LR62N, respectively. The results of the fit are listed in Table A.1. The model parameters are very close for the three sky regions. To compute the Faraday depths, we use the parameters obtained for the L52N mask. As dust polarization only constrains the orientation of , the opposite direction to the values of l0 and b0 in Table A.1 is an equivalent fit of the Planck data. Among the two possible directions, we choose the one closest to the direction derived from Faraday RMs of nearby pulsars, which show that the mean Galactic magnetic field in the Solar neighborhood points toward longitudes around l = 90° (Xu & Han 2019). Dust-polarization data do not constrain the field strength either. Based on the data on pulsars in the northern sky presented by Sobey et al. (2019), we choose the total field strength to be 3 μG.

Table A.1

Planck model fit

A.2 Statistical model of random component

To model the random component of the magnetic field Bt, we use the model introduced by Planck Collaboration Int. XLIV (2016) and Vansyngel et al. (2017). Then, Bt in the sky direction is computed from (A.4)

where the integration along the LoS is approximated by a sum over a finite number, N, of independent Gaussian random fields, Gi, generated on the sphere from an angular power spectrum Cℓ scaling as a power-law ℓ−αM for ℓ ≥ 2. The mean value and dipole of each Gi are set to zero. The parameter fM sets the ratio between the standard deviation of |Bt| and |B0|.

The model has three parameters: fM, N, and αM. We use values: fM = 0.9, N = 7, and αM = 2.5, determined by Vansyngel et al. (2017) from a fit of Planck power spectra of dust polarization at high Galactic latitudes. These parameter values also provide a good fit of the PDF of the dust polarization fraction p, the local dispersion of polarization angles, S, and the anti-correlation between p and S (Planck Collaboration XII 2020).

Appendix B Fine structure levels of N II

This appendix presents an attempt to estimate the EM of ionized gas foreground to the stars using FUSE spectra. Our approach makes use of the collisional excitation of the two fine-structure levels of ionized nitrogen. As the electron density in the WIM is a few orders of magnitude smaller than the critical densities (Goldsmith et al. 2015), the column density of ionized nitrogen N(n II**) at its second excited fine-structure level 3P2 is directly related to the gas emission measure.

N(N II**) is estimated from the two absorption lines around 1085 Å that are within the wavelength range of FUSE spectra. The results are listed in Table B.1. To derive the emission measure from N(n II**), we used the collisional cross-sections with electrons computed by Tayal (2011) for a kinetic temperature of 8000 K, and the spontaneous radiative decay rates listed by Goldsmith et al. (2015). We use an interstellar nitrogen abundance of log(N/H)ISM = −4.2. Then we find that EMNII** (pc cm−6) = N(n II**)/1.8 1012 cm−2.

For comparison, we also list in Table B.1 the Hα line intensities I(Hα) at the position of the stars that we measured using the Wisconsin H-alpha Mapper (WHAM) sky survey (Haffner et al. 2003). The total emission measure in the direction of the stars was derived from I(Hα) for an electron temperature of 8000 K.

For most LoSs, we derive upper limits on N(n II**) of the order of 1013 cm−2, which translates into upper limits for EMNII** that are significantly larger than the total emission measure derived from I(Hα). For the two cases, HD113001 and Ton 102, which yielded actual detections of N(n II**) absorption, the inferred EM values far exceeded those determined from the Hα emission. It is possible that these large contributions of N(n II**) arise from the H II regions surrounding the stars that subtend an angle much smaller than the 1° beam width of the WHAM survey.

Table B.1

Column densities of N II** and emission measures

References

  1. Bailer-Jones, C. A. L., Rybizki, J., Fouesneau, M., Demleitner, M., & Andrae, R. 2021, AJ, 161, 147 [Google Scholar]
  2. Beck, R., Shukurov, A., Sokoloff, D., & Wielebinski, R. 2003, A&A, 411, 99 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  3. Bracco, A., Jelić, V., Marchal, A., et al. 2020, A&A, 644, A3 [Google Scholar]
  4. Bracco, A., Ntormousi, E., Jelić, V., et al. 2022, A&A, 663, A37 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  5. Brentjens, M. A., & de Bruyn, A. G. 2005, A&A, 441, 1217 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  6. Burn, B. J. 1966, MNRAS, 133, 67 [Google Scholar]
  7. Bzowski, M., Czechowski, A., Frisch, P. C., et al. 2019, ApJ, 882, 60 [Google Scholar]
  8. Choi, S. K., & Page, L. A. 2015, J. Cosmology Astropart. Phys., 2015, 020 [CrossRef] [Google Scholar]
  9. Clark, S. E., & Hensley, B. S. 2019, ApJ, 887, 136 [NASA ADS] [CrossRef] [Google Scholar]
  10. de Avillez, M. A., Anela, G. J., Asgekar, A., Breitschwerdt, D., & Schnitzeler, D. H. F. M. 2020, A&A, 644, A156 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  11. Delouis, J. M., Pagano, L., Mottet, S., Puget, J. L., & Vibert, L. 2019, A&A, 629, A38 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  12. Dickey, J. M., Landecker, T. L., Thomson, A. J. M., et al. 2019, ApJ, 871, 106 [NASA ADS] [CrossRef] [Google Scholar]
  13. Draine, B. T. 2011, Physics of the Interstellar and Intergalactic Medium (Chapter 9) [CrossRef] [Google Scholar]
  14. Efron, B. 1979, Ann. Statist., 7, 1 [Google Scholar]
  15. Erceg, A., Jelić, V., Haverkorn, M., et al. 2022, A&A, 663, A7 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  16. Foreman-Mackey, D., Hogg, D. W., Lang, D., & Goodman, J. 2013, PASP, 125, 306 [Google Scholar]
  17. Fuchs, B., Breitschwerdt, D., de Avillez, M. A., Dettbarn, C., & Flynn, C. 2006, MNRAS, 373, 993 [NASA ADS] [CrossRef] [Google Scholar]
  18. Gaensler, B. M., Madsen, G. J., Chatterjee, S., & Mao, S. A. 2008, PASA, 25, 184 [NASA ADS] [CrossRef] [Google Scholar]
  19. Ghosh, T., Boulanger, F., Martin, P. G., et al. 2017, A&A, 601, A71 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  20. Goldsmith, P. F., Yildiz, U. A., Langer, W. D., & Pineda, J. L. 2015, ApJ, 814, 133 [NASA ADS] [CrossRef] [Google Scholar]
  21. Górski, K. M., Hivon, E., Banday, A. J., et al. 2005, ApJ, 622, 759 [Google Scholar]
  22. Gry, C., & Jenkins, E. B. 2017, A&A, 598, A31 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  23. Haffner, L. M., Reynolds, R. J., Tufte, S. L., et al. 2003, ApJS, 149, 405 [NASA ADS] [CrossRef] [Google Scholar]
  24. Heiles, C., & Haverkorn, M. 2012, Space Sci. Rev., 166, 293 [NASA ADS] [CrossRef] [Google Scholar]
  25. HI4PI Collaboration (Ben Bekhti, N., et al.) 2016, A&A, 594, A116 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  26. Hill, A. 2018, Galaxies, 6, 129 [NASA ADS] [CrossRef] [Google Scholar]
  27. Hutschenreuter, S., Anderson, C. S., Betti, S., et al. 2022, A&A, 657, A43 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  28. Iacobelli, M., Haverkorn, M., & Katgert, P. 2013, A&A, 549, A56 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  29. Jelić, V., de Bruyn, A. G., Mevius, M., et al. 2014, A&A, 568, A101 [Google Scholar]
  30. Jelić, V., de Bruyn, A. G., Pandey, V. N., et al. 2015, A&A, 583, A137 [Google Scholar]
  31. Jelić, V., Prelogović, D., Haverkorn, M., Remeijn, J., & Klindžić, D. 2018, A&A, 615, A3 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  32. Jenkins, E. B. 2009, ApJ, 700, 1299 [Google Scholar]
  33. Jenkins, E. B. 2013, ApJ, 764, 25 [NASA ADS] [CrossRef] [Google Scholar]
  34. Kalberla, P. M. W., & Haud, U. 2018, A&A, 619, A58 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  35. Kalberla, P. M. W., & Kerp, J. 2016, A&A, 595, A37 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  36. Kalberla, P. M. W., Kerp, J., Haud, U., & Haverkorn, M. 2017, A&A, 607, A15 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  37. Kruk, J. W., Howk, J. C., André, M., et al. 2002, ApJS, 140, 19 [Google Scholar]
  38. Lallement, R., Vergely, J. L., Valette, B., et al. 2014, A&A, 561, A91 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  39. Lenc, E., Gaensler, B. M., Sun, X. H., et al. 2016, ApJ, 830, 38 [Google Scholar]
  40. Lindegren, L., Klioner, S. A., Hernández, J., et al. 2021, A&A, 649, A2 [EDP Sciences] [Google Scholar]
  41. Maconi, E., Soler, J. D., Reissl, S., et al. 2023, MNRAS, 523, 5995 [NASA ADS] [CrossRef] [Google Scholar]
  42. Manchester, R. N., Hobbs, G. B., Teoh, A., & Hobbs, M. 2005, AJ, 129, 1993 [Google Scholar]
  43. Morton, D. C. 2003, ApJS, 149, 205 [Google Scholar]
  44. Ntormousi, E., Dawson, J. R., Hennebelle, P., & Fierlinger, K. 2017, A&A, 599, A94 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  45. Planck Collaboration III. 2020, A&A, 641, A3 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  46. Planck Collaboration XI. 2020, A&A, 641, A11 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  47. Planck Collaboration XII. 2020, A&A, 641, A12 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  48. Planck Collaboration Int. XXII. 2015, A&A, 576, A107 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  49. Planck Collaboration Int. XLIV. 2016, A&A, 596, A105 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  50. Redfield, S., & Falcon, R. E. 2008, ApJ, 683, 207 [Google Scholar]
  51. Regaldo-Saint Blancard, B., Levrier, F., Allys, E., Bellomi, E., & Boulanger, F. 2020, A&A, 642, A217 [EDP Sciences] [Google Scholar]
  52. Reynolds, R. J., Tufte, S. L., Haffner, L. M., Jaehnig, K., & Percival, J. W. 1998, PASA, 15, 14 [NASA ADS] [CrossRef] [Google Scholar]
  53. Shimwell, T. W., Röttgering, H. J. A., Best, P. N., et al. 2017, A&A, 598, A104 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  54. Shimwell, T. W., Tasse, C., Hardcastle, M. J., et al. 2019, A&A, 622, A1 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  55. Shimwell, T. W., Hardcastle, M. J., Tasse, C., et al. 2022, A&A, 659, A1 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  56. Skalidis, R., & Pelgrims, V. 2019, A&A, 631, A11 [Google Scholar]
  57. Slavin, J. D., & Frisch, P. C. 2008, A&A, 491, 53 [Google Scholar]
  58. Slavin, J. D., McKee, C. F., & Hollenbach, D. J. 2000, ApJ, 541, 218 [Google Scholar]
  59. Sobey, C., Bilous, A. V., Grießmeier, J. M., et al. 2019, MNRAS, 484, 3646 [Google Scholar]
  60. Šnidarić, I., Jelić, V., Mevius, M., et al. 2023, A&A, 674, A119 [Google Scholar]
  61. Sofia, U. J., & Jenkins, E. B. 1998, ApJ, 499, 951 [NASA ADS] [CrossRef] [Google Scholar]
  62. Sokoloff, D. D., Bykov, A. A., Shukurov, A., et al. 1998, MNRAS, 299, 189 [Google Scholar]
  63. Spitzer, Lyman, J., & Fitzpatrick, E. L. 1993, ApJ, 409, 299 [Google Scholar]
  64. Strömgren, B. 1948, ApJ, 108, 242 [Google Scholar]
  65. Sutherland, R. S., & Dopita, M. A. 2017, ApJS, 229, 34 [Google Scholar]
  66. Tayal, S. S. 2011, ApJS, 195, 12 [Google Scholar]
  67. Thomson, A. J. M., Landecker, T. L., Dickey, J. M., et al. 2019, MNRAS, 487, 4751 [NASA ADS] [CrossRef] [Google Scholar]
  68. Turić, L., Jelić, V., Jaspers, R., et al. 2021, A&A, 654, A5 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  69. Van Eck, C. 2018, Galaxies, 6, 112 [NASA ADS] [CrossRef] [Google Scholar]
  70. Van Eck, C. L., Haverkorn, M., Alves, M. I. R., et al. 2017, A&A, 597, A98 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  71. Van Eck, C. L., Haverkorn, M., Alves, M. I. R., et al. 2019, A&A, 623, A71 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  72. Vansyngel, F., Boulanger, F., Ghosh, T., et al. 2017, A&A, 603, A62 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  73. Wayth, R. B., Lenc, E., Bell, M. E., et al. 2015, PASA, 32, e025 [Google Scholar]
  74. Wolfire, M. G., Hollenbach, D., McKee, C. F., Tielens, A. G. G. M., & Bakes, E. L. O. 1995, ApJ, 443, 152 [Google Scholar]
  75. Wolleben, M., Landecker, T. L., Carretti, E., et al. 2019, AJ, 158, 44 [NASA ADS] [CrossRef] [Google Scholar]
  76. Xu, J., & Han, J. L. 2019, MNRAS, 486, 4275 [NASA ADS] [CrossRef] [Google Scholar]
  77. Zaroubi, S., Jelić, V., de Bruyn, A. G., et al. 2015, MNRAS, 454, L46 [NASA ADS] [Google Scholar]
  78. Zucker, C., Goodman, A. A., Alves, J., et al. 2022, Nature, 601, 334 [NASA ADS] [CrossRef] [Google Scholar]
1

The correlation between dust polarization and Faraday rotation should at least depend on the average orientation of the magnetic field, as seen in the study by Bracco et al. (2022) using synthetic data.

2

The low density interstellar region surrounding the Sun to distances of 100 to 300 pc (Lallement et al. 2014; Zucker et al. 2022).

3

For this star, the linear fit used by Jenkins (2013) departs from the curve of growth because only a few O I lines were measured.

4

We select the l0 values in Table A.1 to be between 0 and 90°.

5

B∥ is positive when it points toward the Sun.

7

The SRoll2 maps are available here.

All Tables

Table 1

Stellar targets and results of the data fit.

In the text
Table 2

Column densities and ionization fractions.

In the text
Table 3

BLoS and WNM RMs.

In the text
Table 4

Data for pulsars within the LoTSS field.

In the text
Table A.1

Planck model fit

In the text
Table B.1

Column densities of N II** and emission measures

In the text

All Figures

thumbnail Fig. 1

Stars in our sample overlaid on a LOFAR image of the LoTSS field from Erceg et al. (2022). The image displays the first moment of the Faraday spectra in units of rad m−2. Celestial and Galactic coordinates are plotted with solid and dashed lines, respectively.
In the text
thumbnail Fig. 2

Corner plots of MCMC data fit for two stars in our sample Feige 34 (left) and PG1544+488 (right), chosen as examples of low and high column densities. The histograms show the probability distribution of the three model parameters, and the contours display their interdependence. The dashed lines plotted over the histograms represent the median values of the model parameters.
In the text
thumbnail Fig. 3

Curves of growth for the two stars Feige 34 (left) and PG1544+488 (right), chosen as examples of low and high column densities. For each graph, the y-axis represents the ratio between the equivalent width, Wλ, and the wavelength, λ, of the line, and the x-axis the product of the column density N expressed in cm−2, the strength of the transition (unit-less f-value), and λ in Å. The blue circles represent the equivalent widths of the O Ilines used to determine the column density of OI and the Doppler parameter, b, and the red star the equivalent width of the Ar I line used to determine the [Ar I/O I] abundance ratio.
In the text
thumbnail Fig. 4

Column densities of HI and WNM electrons. The HI column densities are plotted with blue circles and the electron column densities in the WNM with magenta squares. The two triangles pointing down represent upper limits.
In the text
thumbnail Fig. 5

Probability distribution of the mean value of the electron column density (left) and ionization fraction (right). The histograms are built averaging the samples of the posterior distributions over the 9 stars in our sample more distant than 100 pc (i.e. excluding HZ43A). The solid black lines indicate the median value and the dashed lines the 16% and 84% percentiles.
In the text
thumbnail Fig. 6

Maps of the magnetic field component along the LoS, BLoS, over the LOFAR field. The top image shows BLoS for the mean field B0 and the bottom image that computed for B including one realization of the turbulent component Bt. BLoS is positive when it points toward the Sun (see Eq. (4)).
In the text
thumbnail Fig. 7

PDFs of the mean RM and the mean LOFAR M1 moment. The plot displays the PDFs of ⟨RMWNM × 0.5⟩ (red) and ⟨M 1LOFAR⟩ (blue). The PDF of ⟨RMWNM⟩ is also shown as a black line. The mean values are computed over the eight stars with both UV and LOFAR data.
In the text
thumbnail Fig. 8

Comparison of the WNM RMs with the first moments of the LOFAR Faraday spectra for each star individually. The values of RMWNM × 0.5 are plotted with blue squares, and the intensity weighted Faraday depths from LOFAR, M1LOFAR, with red circles.
In the text
thumbnail Fig. 9

PDFs of the mean RM and the mean LOFAR M1 moments for synthetic observations derived from a numerical simulation by Bracco et al. (2022). The mean values are computed on random sets of eight LoS. The PDFs of ⟨RMWNM × 0.5⟩ and ⟨M1LOFAR⟩ are plotted in red and blue. The PDF of ⟨RMWNM⟩ is also shown as a dashed line.
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.