- The paper completes 360° azimuthal coverage using deep GBT observations and channel-resolved WSRT cross-calibration, finding diffuse H I at column densities near 10¹⁹ cm⁻².
- Both galaxies contain 3–7 times more neutral circumgalactic gas off-axis than along their major or minor axes, contradicting azimuthal symmetry and suggesting a butterfly-like distribution.
- Off-axis gas contributes roughly half of the inner-CGM mass, revises NGC 891’s estimate upward by 2–3 times, and shows kinematics consistent with a lagged, co-rotating extended H I disk.
This paper completes a 360° characterization of the diffuse neutral hydrogen in the inner circumgalactic medium (CGM) of the two nearby edge-on spirals NGC 891 and NGC 4565, by adding four off-axis GBT pointings at 45° azimuth between the principal axes to the major- and minor-axis measurements of Das et al. (2020, 2024). The central methodological result is that 30–38% of the H I emission detected by the Green Bank Telescope toward NGC 891 and 18–28% toward NGC 4565 cannot be accounted for by the WSRT HALOGAS interferometric maps, implying genuine detection of diffuse circumgalactic gas at column densities of order 1019cm−2.
Observations and cross-instrument methodology
The off-axis pointings (NO 1E/W, SO 1E/W) were observed with the GBT L-band receiver using position-switched deep stares of roughly one hour effective integration per pointing, achieving a 5σ column density sensitivity of $1.1$–1.2×1017cm−2 over a 20 km s−1 line width — roughly an order of magnitude deeper than the earlier Pingel et al. (2018) mapping campaign. Because each CGM pointing lies only one GBT FWHM ($9.1'$) from the edge of the H I disk, the measured spectra are dominated by disk contamination; isolating the true CGM signal therefore requires subtracting a disk model derived from the HALOGAS cubes convolved with the circularized GBT beam.
Two improvements over the previous papers in this series are introduced here. First, a channel-wise brightness-temperature scaling factor between the GBT central-disk spectrum and the convolved WSRT spectrum replaces the single integrated scaling factor used previously, preserving velocity-dependent calibration differences. Second, velocity-edge corrections align the rising and falling edges of the disk emission before computing the scaling factor; these corrections are applied only where justified (a 7 km s−1 offset for NGC 4565, none for NGC 891), since the differing behavior between the two galaxies argues against a common instrumental artifact such as a mischaracterized GBT beam response. Spurious channels in the scaling factor are mitigated by a two-step sigma-clipping procedure combining nearest-neighbor difference clipping on the spectral edges with global clipping over a converged plateau region.
The residual CGM column density is then N(HI)CGM=N(HI)GBT−N(HI)WSRT, with masses computed under the optically thin assumption within the main GBT beam.
Azimuthal distribution of the neutral CGM
The headline result is that the residual CGM column density is systematically higher along the off-axes than along either principal axis in both galaxies. Averaged over pointings:
| Galaxy |
Off-axis |
Major axis |
Minor axis |
| NGC 891 |
(1.96±0.91)×1019 cm−2 |
5σ0 cm5σ1 |
5σ2 cm5σ3 |
| NGC 4565 |
5σ4 cm5σ5 |
5σ6 cm5σ7 |
5σ8 cm5σ9 |
There is thus 4–6× more H I along the off-axes than along the major axes and 3–7× more than along the minor axes. This directly contradicts the commonly assumed azimuthal symmetry of the neutral CGM. Notably, because the minor-axis and off-axis pointings are approximately equidistant from the galaxy while the major-axis pointings lie farther out, there is no correlation of $1.1$0 with impact parameter. The authors propose several explanations for this "butterfly-like" morphology: azimuthally varying ionization (with biconical ionization along the minor axes from stellar/nuclear radiation), smeared-out cold-mode accretion filaments preferentially located off-axis, scalloping or warping of an extended low-column-density disk, or a transitional boundary between active and inactive accretion sectors. They explicitly caution that the data cannot distinguish beam-filling diffuse gas from unresolved clouds below the HALOGAS sensitivity limit; high-resolution interferometry would be required to settle the morphology.
Mass budget and depletion times
The off-axis detections substantially revise the inner-CGM mass budgets. For NGC 891, the off-axis mass is $1.1$1, compared to $1.1$2 along the principal axes, giving a total of $1.1$3 — meaning that linear interpolation between principal axes underestimated the total by a factor of 2–3. The inner CGM constitutes $1.1$4 of the $1.1$5 disk, and the updated depletion time remains $1.1$6 Gyr, leaving NGC 891 short of fuel unless additional ionized reservoirs or in-situ fueling contribute. For NGC 4565, the off-axis mass is $1.1$7 against $1.1$8 on the principal axes, totaling $1.1$9 (about 5.1% of its larger disk) and yielding a comfortable 1.2×1017cm−20 Gyr.
A striking comparative finding is that the off-axis H I mass is statistically indistinguishable between the two galaxies despite their different star formation rates and different principal-axis properties. This suggests the off-axis reservoir may be a ubiquitous feature of spiral galaxies that does not track the star-forming state of the host.
Kinematics and momentum
The mean line-of-sight velocities of the residual CGM trace the shape of the disk rotation curve out to 20–30 kpc, consistent with a lagged co-rotating extended H I disk. In NGC 891 the CGM velocities are generally equal to or below the local disk values (consistent with the known extraplanar lag), whereas in NGC 4565 they are generally equal to or above them, with the exception of the major-axis pointing UP 1J where a known warp and flare complicate the comparison. The super-disk velocities in NGC 4565 are tentatively attributed to energy injection by its LINER AGN, though this interpretation is offered as a possibility rather than a demonstrated mechanism. Line-of-sight momentum follows the same hierarchy as mass — largest off-axis, intermediate on the major axes, smallest on the minor axes — with NGC 4565 showing a larger azimuthal spread (~2× the mean versus ~1.6× for NGC 891), again possibly linked to its AGN.
Systematics checks
The paper devotes considerable effort to establishing that the excess is physical rather than instrumental. Comparison with independently reduced GBT maps from Pingel et al. (2018) reproduces the same 2–3× off-axis excess, ruling out calibration systematics in the deep-stare data. Alternate treatments of the GBT beam (rotation through extreme position angles; convolution with median/16th/84th-percentile beam maps) all yield results within 1σ of the fiducial circularized-beam analysis. Convolving unmasked HALOGAS cubes shows no loss of true emission in the masking step. A deconvolution analysis treating each pointing as a point source confirms that sidelobe overlap between adjacent pointings cannot produce the off-axis excess, although it also yields unphysical negative values for some NGC 4565 pointings, indicating the point-source approximation itself breaks down for extended emission.
Limitations and open questions
Several caveats bear directly on the interpretation. The residual CGM measurement depends on the assumption that the WSRT spectrum fully characterizes the disk contamination within the GBT beam; any large-scale emission filtered by the interferometer's finite maximum resolvable scale would be misattributed to the CGM. The optically thin, beam-filling assumption underlying the mass estimates may not hold if the emission arises from unresolved clumps. The morphological interpretation ("butterfly" versus smeared clouds versus warped disk) remains undetermined by single-dish data alone. Finally, the sample comprises only two galaxies, so whether the off-axis excess is truly ubiquitous among spirals — and how it scales with mass, star formation rate, and AGN activity — is left open, as is the question of what physical process sets the azimuthal asymmetry.
Conclusion
By completing the azimuthal coverage of the inner neutral CGM of NGC 891 and NGC 4565, this work demonstrates that the low-column-density circumgalactic H I distribution is not azimuthally symmetric: off-axis regions contain several times more gas than the principal axes, contributing roughly half of the total inner-CGM mass and revising prior mass estimates upward by a factor of about two. The kinematics indicate a co-rotating, possibly lagged extended disk, and the consistency of the off-axis mass between the two galaxies hints at a common structural feature of spiral halos. The paper establishes deep single-dish stare observations combined with careful, channel-resolved cross-calibration against interferometric maps as a viable technique for detecting diffuse CGM gas, while leaving its morphology and generality to future high-resolution mapping.