Comment on the paper “ Calorimetric Dark Matter Detection with Galactic Center Gas Clouds ”

In a recent Letter, Bhoonah et al. [1] (hereafter B18) attempted to derive limits on dark matter interactions with ordinary matter, by demanding that DM heating of gas clouds not exceed the known astrophysical cooling rate based on the temperature, density and metallicity of observed clouds. In B18, the cloud G1.4−1.8+87 from [2] (hereafter McG13) was singled out as most suitable by virtue of its exceptionally low temperature and relatively low density. In this Comment, we point out that (i) the cloud G1.4−1.8+87 cannot be assumed to be at / 22K as done by B18, (ii) some other parameters quoted in B18 for G1.4−1.8+87 are incorrect (apparently read from neighboring rows in the McG13 online data table [2]) and B18 confuses VLSR (which is the velocity with respect to the local standard of rest) with the velocity of the cloud relative to the DM, and (iii) use of the gas clouds reported in McG13, which are tracers of the high velocity nuclear outflow (HVNO) discovered in McG13, is not appropriate for an analysis relying on the assumption of stability over a long cooling time. McG13 detected a population of gas clouds in the HVNO, in the 21cm Hi line using the Australia Telescope Compact Array (ATCA). The velocity resolution of that survey was 1 km/s, and a criterion for detection was that there be a significant Hi signal over at least 5 km/s [2]. Fig. 1 shows the Hi spectrum at the location of the cloud G1.4−1.8+87, from the public online data [3]. One sees detectable emission over velocities 70 100 km/s, thus meeting the criterion for a real cloud, however the summary table of McG13 reports only the single-channel peak at 87 km/s with a narrow FWHM of 1 km/s. Single-channel fluctuations of this amplitude relative to neighboring bins appear elsewhere in this spectrum and other spectra, so the quoted extremely-narrow line width of the cloud must be assumed to be spurious noise, unless established by further observations [9]. As seen in Fig. 1, most of the Hi emission for this cloud is characterized by a line with a FWHM of 26.6 km/s (red line). For comparison, the spectrum of the cloud G33.4-8.0 [4] used in [5], is also shown. B18 calculated the temperature of G1.4−1.8+87 to be Tg / 22K taking the velocity dispersion to be 1 km/s. Using the correct value, 26.6 km/s, gives Tg above 15,000 K. In addition to the erroneous temperature inference, some other parameters given in B18 for the cloud 60 80 100 120 140 160 180 200 VLSR (kms −1 ) 2 1 0 1 2 3 4 5

In a recent Letter, Bhoonah et al. [1] (hereafter B18) attempted to derive limits on dark matter interactions with ordinary matter, by demanding that DM heating of gas clouds not exceed the known astrophysical cooling rate based on the temperature, density and metallicity of observed clouds. In B18, the cloud G1.4−1.8+87 from [2] (hereafter McG13) was singled out as most suitable by virtue of its exceptionally low temperature and relatively low density. In this Comment, we point out that (i) the cloud G1.4−1.8+87 cannot be assumed to be at 22K as done by B18, (ii) some other parameters quoted in B18 for G1.4−1.8+87 are incorrect (apparently read from neighboring rows in the McG13 online data table [2]) and B18 confuses V LSR (which is the velocity with respect to the local standard of rest) with the velocity of the cloud relative to the DM, and (iii) use of the gas clouds reported in McG13, which are tracers of the high velocity nuclear outflow (HVNO) discovered in McG13, is not appropriate for an analysis relying on the assumption of stability over a long cooling time.
McG13 detected a population of gas clouds in the HVNO, in the 21cm Hi line using the Australia Telescope Compact Array (ATCA). The velocity resolution of that survey was 1 km/s, and a criterion for detection was that there be a significant Hi signal over at least 5 km/s [2]. Fig. 1 shows the Hi spectrum at the location of the cloud G1.4−1.8+87, from the public online data [3]. One sees detectable emission over velocities 70 -100 km/s, thus meeting the criterion for a real cloud, however the summary table of McG13 reports only the single-channel peak at 87 km/s with a narrow FWHM of 1 km/s. Single-channel fluctuations of this amplitude relative to neighboring bins appear elsewhere in this spectrum and other spectra, so the quoted extremely-narrow line width of the cloud must be assumed to be spurious noise, unless established by further observations [9]. As seen in Fig. 1, most of the Hi emission for this cloud is characterized by a line with a FWHM of 26.6 km/s (red line). For comparison, the spectrum of the cloud G33.4-8.0 [4] used in [5], is also shown.
B18 calculated the temperature of G1.4−1.8+87 to be T g 22K taking the velocity dispersion to be 1 km/s. Using the correct value, 26.6 km/s, gives T g above 15,000 K. In addition to the erroneous temperature inference, some other parameters given in B18 for the cloud ter of mass. V LSR is defined to be an object's line-of-sight velocity relative to a frame of reference in a circular orbit around the Galactic Center at the position of the solar system. Instead, the velocity of the cloud relative to the Galaxy is to a good approximation the outflow velocity of the Hi clouds entrained in the nuclear wind, which is ∼ 330 km/s according to the interpretation of the data via comparison to simulations by [6]. Constraints on the milli-charged DM cross-section are sensitive to the relative velocity between DM and gas; for milli-charged DM the cross-section σ ∼ v −4 rel and the energy transfer rate ∼ σ v KE ∼ v −4 vv 2 ∼ v −1 , so correcting v rel weakens the limit by an additional factor ≈ 4.
Furthermore, the DM velocity dispersion is taken by B18 to be v 0 = 180 km/s at the location of the clouds, quoting [7], but that reference actually gives the rotation velocity, not the DM velocity dispersion. Additionally, the Burkert density profile [8] for a DM halo is written ]. This may be a typo, but a clearly significant issue with the B18 analysis is that they use ρ b = 14 GeV/cm 3 & r b = 3 kpc (without citing a source) whereas the latest fit from [10] has ρ b = 1.57 GeV/cm 3 & r b = 9.26 kpc; this farfrom-conservative assumption about the DM density in the HVNO clouds exaggerates the limit by a factor 9.
The most serious problem with B18 is a conceptual one: the clouds they use are in the hot, high-velocity wind (10 6−7 K, 330 km/s) emanating from the Galactic Center. Therefore these clouds are likely transient objects, making them inappropriate for an analysis requiring equilibrium conditions. The subsequent more detailed analysis in Bhoonah et al. [11] relies on additional unjustified assumptions about G1.4−1.8+87, such as the absence of shocks in the gas. In a separate paper [5] two of us correct the B18 limits, by using gas clouds like G33.4-8.0 which are not in the HVNO and are plausibly in equilibrium for an adequate timescale. The limits from robust clouds are 2-4 orders of magnitude less stringent for the millicharge coupling and velocity-independent cross section, respectively, than obtained by B18 using G1.4−1.8+87.
We thank C. McKee for helpful comments. GRF ac-knowledges support of NSF-1517319. The Green Bank Observatory is a facility of the National Science Foundation operated under a cooperative agreement by Associated Universities, Inc.