The Connection between Different Tracers of the Diffuse Interstellar Medium: Kinematics

Seventeen sightlines were chosen by examining the SIMBAD database (Wenger et al. 2000); only GOT C+ pointings accessible to the Northern Hemisphere were considered. The background targets in our survey had to be relatively bright (B, V ≤ 10) A and B stars with amounts of extinction per kpc typical for directions with diffuse molecular gas (such as that seen toward ζ Oph). The lines of sight had to be within 30 arcminutes of the Herschel pointings. Visual extinctions were derived by assuming A_V = 3.1 E(B-V). Distances were obtained by spectroscopic parallax, as many stars did not have definitive Hipparcos or GAIA measurements and are known to approximately 20% based on average uncertainties in magnitudes (∼0.02).

The data at visible wavelengths were taken at the McDonald Observatory in 2012 December, 2014 July, and 2017 October with the echelle spectrograph (Tull et al. 1995). Table 1 lists the stellar data and observational details for the lines of sight. References for spectral types and (B-V)₀ are provided in the footnotes of Table 1. Ca ii λ3933, Ca i λ4226, CH⁺ λ4232, CH λ4300, and CN λ3874 absorption features were able to all be observed with a single exposure using the 56th order of the echelle grating centered at 4065 Å. This setting is denoted as "Blue" in Table 1. K i λ7699 was observed with the 31st order of the echelle grating centered on 7165 Å. This setting is denoted as "Red" in Table 1.

The data reduction was performed in IRAF using standard routines. Dark and bias frames were combined and applied to all other exposures prior to cosmic ray removal and other processing. Flat-fields were normalized and divided into the exposures. A wavelength scale was established using Th-Ar lamp exposures and applied to all stellar exposures before they were corrected for their V_lsr offset, combined, and subsequently normalized. It should be noted that K i was only observed for a subset of the sightlines and has a slightly lower spectral resolution than the rest of the McDonald data, 125,000 versus 135,000. After normalization, the ISMOD routine (Sheffer, unpublished), an rms minimization spectrum synthesis FORTRAN program, was used to model the Voigt absorption profiles via spectrum synthesis and automatic rms minimization of (data minus fit) residuals. For each synthesized absorption feature, ISMOD fits the total column density, the number of components, relative column density fractions, radial velocity, and their Doppler widths (b-values) for each component. The resulting fit was compared with fits for other visible species (Ca ii, Ca i, K i, CH⁺, CH, CN) from the same line of sight. If there was a disagreement in the overall component structure; all species were iteratively refit until the best overall agreement was reached.

Some species, such as Ca ii, have absorption components that extend well beyond those of other species observed at visible wavelengths. This high-velocity gas seen in Ca ii is a well recognized phenomenon (Pan et al. 2005). This makes the determination of the extended structure more difficult. By requiring realistic b-value ranges for each species seen in McDonald spectra (taken from the extensive survey conducted by Pan et al. 2005) and using the least number of components detected at or above 3σ, we can reduce the number of possible solutions to only secure reproducible ones. These fits were made independently of the fits to the species examined in the GOT C+ survey, and in the majority of cases, there is agreement in the structure of the components.

Table 2 gives the results of fitting the absorption from these species (ordered by Galactic coordinates given in Table 1). Although HD 35652 and HD 60146 were observed at McDonald, Herschel measurements were not obtained, nor were data for CO or H i emission. These two sightlines are not discussed further. Here, we focus on the velocities of components; the column densities are needed for the analyses in the subsequent paper.

The GOT C+ project is described in detail by Langer et al. (2010) and Langer et al. (2014). GOT C+ sources are labeled as GXXX.X+YY, which are the latitude and longitude rounded to one decimal; the actual coordinates are given in Table 3. The details of the data reduction are described in Pineda et al. (2013). The GOT C+ data sets are available as a Herschel User Provided Data Product.⁵ The GOT C+ data were fit independently of the results from the McDonald data. GOT C+ data were fitted with 1–3 Gaussians on top of a linear continuum, in IDL, using the MPFIT package from Markwardt et al. (2009). The initial fits for the Gaussians were determined by eye.

The [C ii] ²P_3/2 → ²P_1/2 observations from the HIFI (Pilbratt et al. 2010; de Graauw et al. 2010) instrument on board the Herschel Space Observatory have an angular resolution of 12''. The observations of the J = 0 → 1 transitions of ¹²CO, ¹³CO, and C¹⁸O from the ATNF Mopra Telescope have an angular resolution of 33''. The H i 21 cm observations from the VLA Galactic Plane Survey (VGPS; Stil et al. 2006) have an angular resolution of 1'. The species observed in emission C⁺, CO, and H i have a velocity resolution of 0.8 km s⁻¹, while C¹⁸O has a velocity resolution of 1.6 km s⁻¹. The species observed in absorption (Ca ii, Ca i, K i, CH⁺, CH, CN) have a velocity resolution of 2.2 km s⁻¹. Each detection of a component is required to be at least at the 3σ level. The uncertainty in column density for each component is between 1% and 30% of the column density of that component. Only the smallest components have uncertainties nearing the 3σ detection limit; typical total uncertainties are closer to 10%. The uncertainty in velocity is ∼0.3 km s⁻¹ for strong lines and ∼0.6 km s⁻¹ for weak lines, while the typical uncertainty in b-value is also ∼0.3 km s⁻¹.

Each sightline has a stacked plot showing the spectra of the species that were observed in absorption toward the star and in emission for the closest GOT C+ pointing. (Figure 1 is printed, while Figures 1.1 through 1.15 are available online in the figure set). Species not shown were not observed for that pointing, with the exception of C¹⁸O, which is not shown due to the low correspondence with absorption components. The V_lsr of components detected at a 3σ level are indicated for each species as a red tick above the spectra. Our results of the absorption and emission are combined into Tables 4 and 5. These two tables show the V_lsr for components with emission features and the V_lsr of any corresponding absorption features for the inner and outer Galaxy, respectively. These components, seen in emission and absorption, are discussed in Section 3. However, there are often components seen in absorption that do not correspond with any of the components seen in emission, or vice versa.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

View figure set (15 panels)

Tarball of figure set images

The following points were considered when seeking matches to component velocities. First, diffuse atomic and molecular gas, with kinetic temperatures of 50 to 80 K, have thermal widths of about 1 km s⁻¹. The extensive survey conducted by Pan et al. (2005) revealed typical b-values of 1 to 2 km s⁻² for absorption lines, depending on species. Since the species in our survey from the McDonald observations have atomic masses greater than 20 amu, thermal broadening makes a negligible contribution to the b-value, which mainly arises from turbulent motion after removing the instrumental width of 2.2 km s⁻¹. Second, the light path for the hollow cathode lamp used in wavelength calibration differs slightly from that of the stellar radiation. This difference causes 0.5 to 1.0 km s⁻¹ offsets from an absolute velocity scale. Third, most of the background stars in our survey have distances within 500 pc of the Sun, with only one significantly beyond 1000 pc. On Galactic scales, these stars are considered within the solar neighborhood. The majority of the gas along the lines of sight have velocities in the Local Standard of Rest less than 10 km s⁻¹. Dynamical phenomena like stellar winds and expanding supernova remnants create components with larger radial velocities, some of which are in excess of 100 km s⁻¹. An example involves directions that probe the supernova remnant IC 443 (Welsh & Sallmen 2003; Hirschauer et al. 2009). This leads to complications in associating components seen in absorption and emission in the inner Galaxy where large velocities may arise from dynamical phenomena, especially in atomic gas, or Galactic rotation. In light of these considerations, correspondences were noted when velocities associated with line centers agreed within the FWHM of the lines. Weighted averages based on column density were performed for some species when comparisons were made with the broadest lines (usually H I).

When comparing observations in absorption and in emission, there will frequently be components that appear shifted, blended, or even missing entirely. Over half (56%) of the emission components are not associated with an absorption component, and over half (53%) of the absorption components are not associated with an emission component. The majority of the unassociated components are only seen in species associated with low-density gas. Over half (52%) of the unassociated emission components are only seen in H i and 80% of the unassociated absorption components are only seen in Ca ii.

There are several sources of these inconsistencies. One is the fact that absorption lines only sample gas along the line of sight between the background source and the observer, while the emission lines sample all the gas. This is the reason components observed in emission at high V_lsr are sometimes absent in absorption. The V_lsr structure for two nearby absorption sightlines can exhibit similar differences when the background sources are at different distances.

Second, at low densities, almost all the gas seen in absorption is in the lowest states and the maximum column density is probed, while emission lines typically come from weakly excited states in diffuse gas and are sensitive to density. This difference results in the absorption measurements being more sensitive to detecting the gas than the emission lines, but is also restricted to lines of sight with background sources, while emission lines can be mapped. This is one of the sources for additional components seen in absorption that are not seen in emission. Over half (80%) of the unassociated absorption components are only seen in Ca ii, which indicates low-density gas.

Another source of inconsistency is from the blending of lines. Sometimes the relatively broad emission lines result in a component that corresponds to a blend of two absorption components with slightly different velocities. Whenever two components had to be combined to form a corresponding component, it is indicated by the use of the superscript "a" on the V_lsr in Tables 4 and 5, where an averaged velocity (weighted by column density) is quoted for the McDonald data.

There are occasional additional components seen in emission that are not seen as absorption components and fall within the overlapping V_lsr ranges. The most likely explanation is from variations in the small-scale structure between the different locations of the emission and absorption measurements. In general, whenever an observed species is associated with one or more of the unobserved species, those unobserved species are likely present at varying levels between the emission and absorption pointings. However, at the location of the measurements, the levels are too low to be confidently measured. Table 6 shows a summarized component-by-component interpretation for a pointing with two nearby sightlines. The majority of components for the two sightlines are the same, but there are also clear differences. The component with a V_lsr of 7 km s⁻¹ is seen in emission toward the GOT C+ pointing and in absorption toward HD 165783 but not toward the other nearby sightline, HD 165918. This small-scale structure can also cause components to be slightly shifted in velocity.

HD 168607 is at a distance of 1100 pc and is 12 arcmin away (≤3.8 pc at 1100 pc) from the pointing G014.8-1.0. Almost all the components seen in H i are also seen in Ca ii, indicating they are both probes of diffuse regions of the individual clouds. However, there are many components seen in Ca ii that are not seen in emission (at V_lsr of −45.5, −39.7, −33.6, −28.4, −18.2, −13.4, −5.3, −0.8, 3.4, 9.9, 59.2, 65.6 km s⁻¹). This difference is due to the increased sensitivity of absorption measurements relative to emission measurements. There are also some H i features that do not have corresponding absorption features; these features have large V_lsr (70.0 and 112.6 km s⁻¹) and are most likely probing the material behind HD 168607.

There are three components that are seen in Ca ii, CH⁺, and CH (near V_lsr of 6.7, 15.0, 23.3 km s⁻¹); these components only correspond with emission features seen in H i. The presence of CH and CH⁺ indicates that these components are probing denser regions than if only Ca ii or H i were observed. We would expect these types of regions, with CH or CH⁺ detections, to have some amount of CN or CO. The CH⁺ absorption indicates the presence of relatively low-density gas (Pan et al. 2005); otherwise, reactions with H₂ would destroy CH⁺. There are clear relationships involving CO, CH⁺, and CH with molecular hydrogen for N (H₂) ≥ 10¹⁹ cm⁻², as documented in Sheffer et al. (2008). These H₂ column densities correspond to molecular hydrogen fractions of 5% to 10%. This is the minimum molecular hydrogen fraction we use to define gas as molecular. In low-density gas, the CH⁺ + O → CO⁺ + H reaction leads to CO production. The CH and CO are also coupled through the neutral–neutral reaction route CH + O → CO + H. However, the abundances of these species are not high enough for a reliable detection in CO emission. Therefore, these components are most likely probing CO-dark H₂ regions, where CO is present but at a column density too low to be confidently measured in emission.

There is one component with detectable CN, CH, and Ca ii near a V_lsr of 18.7 km s⁻¹. This component corresponds to emission in C⁺, ¹²CO, ¹³CO, and C¹⁸O. The presences of ¹²CO,¹³CO, and C¹⁸O indicate that this component is probing much denser gas. Here, the associated data from absorption are revealing the diffuse outer regions of a molecular cloud.

There are several components seen in CO emission that are not seen as molecular absorption components (at V_lsr of −26.1, 30.0, 39.1, 46.3, 55.9, 75.6 km s⁻¹), despite the majority being associated with components in Ca ii. The presence of CO emission indicates that the density in the area being probed should be high enough for detectable amounts of other species, such as CH or CH⁺. There will be some differences that arise from variations in the small-scale structure between the different locations of the emission and absorption measurements, in this case 12 arcmin. Thus, the presence of only Ca ii indicates that the absorption measurements are likely probing the lower-density atomic region that surrounds the molecular gas associated with the CO emission. Table 7 provides a summarized component-by-component interpretation for this sightline. See Figure 1.3 for the stacked spectra.

Tables 4 and 5 show the V_lsr structure of components with emission features and the V_lsr of any corresponding absorption features for the inner and outer Galaxy. However, there are often components seen in absorption that do not correspond with any of the components seen in emission. These components are discussed here.

G010.4+0.0: HD 165783 is 4 arcmin away and HD 165918 is 6 arcmin away from G010.4+0.0. Assuming a distance of 500 pc, the separations between directions related to the absorption and emission measurements are ≤0.6 pc and ≤0.9 pc, respectively.

There are five components seen in emission and toward both HD 165783 and HD 165918 (near V_lsr of −21.2, −10.3, 0.0, 4.2, 10.7 km s⁻¹). Three of these are only associated with H i emission and are probing the lower-density regions (near V_lsr of −21.2, 0.0, 10.7 km s⁻¹). One component, seen in Ca ii at 4.2 km s⁻¹, is also associated with CH/CH⁺ absorption and C⁺ emission. The presence of CH and CH⁺ without CO indicates this is likely probing diffuse molecular gas and CO-dark H₂ gas.

There are six components from species seen in absorption without corresponding emission features. These additional components are seen toward HD 165783 near V_lsr (km s⁻¹) of −15.9 (Ca ii and CH⁺), −6.4 (Ca ii), −3.6 (Ca ii and CH⁺), and 13.5, 16.6, 39.0 (Ca ii). Only three of these additional components are seen toward HD 165918 near V_lsr (km s⁻¹) of −16.0, −3.5, 16.0 (Ca ii). These differences likely stem from small-scale variations in structure and weak Ca ii components. There are several nearby dark clouds including ZHS2005 C3C, C3B, C3A (Zhang et al. 2005), but all measured V_lsr are beyond the observed range spanned by the sightlines toward HD 165783 or HD 165918. The V_lsr given in the literature (Zhang et al. 2005) for C¹⁸O at 66 km s⁻¹ agrees with our measurements of a component seen in C¹⁸O, C⁺, and H i at 65 km s⁻¹. Table 6 shows the summarized component-by-component interpretation for this GOT C+ pointing for both of its two nearby stellar sightlines. See Figures 1.1 and 1.2 for the stacked spectra. The structure of each individual component can be seen in Figure 1.1 as dotted lines, with the center of each component indicated by a red tick mark. Other stacked spectra indicate the components with only red tick marks, as in Figure 1.2.

G015.7+1.0. HD 167498 is 9 arcmin away and HD 167812 is 15 arcmin away. Assuming a distance of 250 pc, the separations between the directions observed in absorption and emission are ≤0.7 pc and ≤1.1 pc, respectively. There is one component seen in emission from G015.7+1.0 with H i and toward both HD 167498 and HD 167812 in Ca ii and CH (near a V_lsr of ∼4.1 km s⁻¹). This component is likely probing CO-dark H₂ gas. There are two weak Ca ii components toward HD 167498 that have no emission features associated with them (near V_lsr of −14.7 and −10.8 km s⁻¹). There is also one component toward HD 167498 and HD 167812 observed in absorption from CH, CH⁺, and Ca ii that has no emission feature (near a V_lsr of 1.0 km s⁻¹). This component is likely probing CO-dark H₂ gas. See Figures 1.4 and 1.5 for the stacked spectra.

G020.0+0.0. HD 169754 is 14 arcmin away from G020+0.0. Assuming a distance of 2000 pc, the directions are ≤8.1 pc apart. Almost all the H i emission components are also seen in Ca ii. There are two components from species detected via absorption toward HD 169754 that are unassociated with components seen in emission near V_lsr (km s⁻¹) of −7.3 (Ca ii) and 12.2 (Ca ii and CH⁺). There are several nearby dark clouds including L413 (Lynds 1962) but only one with additional velocity information. BGPS G020.050+0.195 has a velocity component around 18 km s⁻¹ seen in HCO⁺, N₂H⁺, and NH₃ (Schlingman et al. 2011) that agrees with a component seen in Ca ii, H i and CH⁺ near 20 km s⁻¹, and is likely associated with the outer envelope of the molecular cloud. See Figure 1.6 for the stacked spectra.

G032.6+0.0. HD 174509 is 13 arcmin away from G032.6+0.0 and assuming a distance of 500 pc, the directions probe material ≤1.9 pc apart. The component with H i emission at −0.5 km s⁻¹ is associated with Ca ii, Ca i, K i, CH⁺, and CH absorption and is likely probing diffuse molecular gas. The component with H i and C⁺ emission near 14.4 km s⁻¹ is associated with Ca ii, Ca i indicating diffuse atomic gas. The other three absorption components all have detected Ca ii, Ca i, K i, CH, and CN and are associated with CO emission near V_lsr of 5.8, 9.7, and 12.1 km s⁻¹ and are likely probing the cloud envelope.

There are several dark clouds nearby with measurements in NH₃. BGPS G032.524-00.131 (79.3 km s⁻¹), BGPS G032.527-00.123 (79.4 km s⁻¹), BGPS G032.411-00.0099 (80.2 km s⁻¹), BGPS G032.436-00.17 (77.4 km s⁻¹) (Dunham et al. 2011); however, all the corresponding V_lsr measurements lie beyond the range measured for species in absorption. There is agreement with a component seen in emission in ¹²CO, ¹³CO,C¹⁸O, C⁺, and H i near 75 km s⁻¹, indicating this molecular cloud lies behind HD 174509. See Figure 1.7 for the stacked spectra.

G091.7+1.0. BD+49° 3482 is 9 arcmin away and BD+49° 3484 is 4 arcmin away from G091.7+1.0. For a distance of 250 pc, the directions are ≤0.7 pc and ≤0.3 pc apart, respectively. It should be noted that the pointing G091.7+1.0 did not include CO measurements. There is one component seen in H i emission and in Ca ii and CH⁺ for both BD+49° 3482 and BD+49° 3484 near V_lsr of 3.5 km s⁻¹. There is also a component near 1.6 km s⁻¹ seen toward BD+49° 3482 that is associated with K i and CH⁺ absorption and H i emission.

The weak Ca ii component from BD+49° 3482 (at V_lsr of −4.4 km s⁻¹) has no emission features associated with it and three Ca ii components toward BD+49° 3484 also have no associated emission features (at V_lsr of 8.5, 10.3, and 13.0 km s⁻¹). Two of the components from BD+49° 3484 also have a CH component (near 8.5 and 10.3 km s⁻¹). One of these components also has CN absorption (near 10.3 km s⁻¹). There are two dust clouds along adjacent directions, Dobashi 2934 (Dobashi 2011) and TGU H541 P30 (Dobashi et al. 2005). There is also a molecular cloud, [DBY 94] 091.7+00.9, observed in ¹³CO with a V_lsr of 10.9 km s⁻¹ (Dobashi et al. 1994). This agrees nicely with the absorption seen in CN (10.2 km s⁻¹), CH (10.6 km s⁻¹), and Ca ii (10.3 km s⁻¹) and C⁺ emission, indicating the presence of a diffuse molecular envelope. The nearby sightline (BD+49° 3482), although farther from the Herschel pointing, mainly shows the lower-velocity material. See Figures 1.8 and 1.9 for the stacked spectra.

G109.8+0.0. HD 240179 and HD 240183 are both 7 arcmin away from G109.8+0.0 and have a separation of ≤1.4 pc, assuming a distance of 700 pc. It should be noted that the pointing G109.8+0.0 did not include CO measurements. There are two components seen in H i emission and Ca ii absorption toward both HD 240179 and HD 240183 (near V_lsr of −10.3 and 0.2 km s⁻¹). Spectra of HD 240179 also reveal this component in CH⁺, as well as an additional component near 4.0 km s⁻¹ that is not seen toward HD 240183. This component near 4.0 km s⁻¹ is detected in H i emission and Ca ii and K i absorption and is likely sampling a diffuse molecular cloud edge.

There are several components seen in absorption that have no corresponding emission features (near V_lsr of −7.0, −3.5, −1.0, 7.6, 11.9 and 17.5 km s⁻¹), several of which are only weak Ca ii components (near V_lsr of 7.6, 11.9, and 17.5 km s⁻¹). However, there are three components with Ca ii, CH, CH⁺, and K i with no corresponding emission features (near V_lsr of −7.0, −3.5, −1.0 km s⁻¹) and might point to the presence of CO-dark H₂ gas, but without CO emission data for this pointing it is difficult to confirm. There is a dense core nearby, IRAS 23033+5951, that has been studied extensively. NH₃ observations by Wouterloot et al. (1988) show emission at V_lsr of −52.8 km s⁻¹, while ¹²CO and ¹³CO measurements by Wouterloot & Brand (1989) indicate respective V_lsr of −51.8 and −53.1 km s⁻¹. This component is clearly associated with the emission seen in H i and C⁺. See Figures 1.10 and 1.11 for the stacked spectra.

G207.2-1.0. HD 47073 and HD 260737 are both 7 arcmin away from G207.2-1.0, and assuming a distance of 200 pc, the directions are both separated by ≤0.4 pc. There is one component seen in H i emission and in Ca ii absorption for both HD 47073 and HD 260737 near a V_lsr of 3.8 km s⁻¹ and is likely only probing diffuse atomic gas. There is one additional component toward HD 260737 that is seen in both emission (H i and ¹²CO) and absorption (Ca ii and K i) near a V_lsr of 9.4 km s⁻¹ and is likely probing the envelope of a molecular cloud. Four additional weak Ca ii components toward HD 260737 are not seen in emission (−15.6, −8.6, −2.9, 1.8, and 5.3 km s⁻¹). See Figures 1.12 and 1.13 for the stacked spectra.

G225.3+0.0. HD 55469 is 26 arcmin away and HD 55981 is 8 arcmin away from G225.3+0.0. Assuming a distance of 400 pc, the separations are ≤3.0 pc and ≤0.9 pc, respectively. There is no H i data for this pointing. Emission components are detected toward G225.3+0.0 near V_lsr of 17.8 (C⁺ and ¹²CO) and 51.9 (C⁺) km s⁻¹. There are absorption components detected toward HD 55981 near V_lsr of −7.3, −2.3, 2.8, 5.7, 9.2 (Ca ii), 14.8 (Ca ii, K i, and CH⁺), and 20.9 km s⁻¹ (Ca ii). There are also absorption components detected toward HD 55469 near V_lsr of 2.0 (Ca ii and K i) and 3.8, 10.3 km s⁻¹ (Ca ii). While there are components seen in emission and in absorption, no clear correspondence is found between the emission and absorption components. The closest correspondence is seen for the absorption from CH⁺, K i, and Ca ii around a V_lsr of 14.8 km s⁻¹ and the emission from C⁺ and ¹²CO around a V_lsr of 17.5 km s⁻¹. Kim et al. (2004) studied L1658 in ¹³CO at coordinates G224.3-1.1 (at a V_lsr of 15.5 km s⁻¹) and G226.1-0.04 (at a V_lsr of 16.2 km s⁻¹). These velocities both fall between those of the emission at 17.5 km s⁻¹ and the absorption at 14.8 km s⁻¹, suggesting that the components are related. See Figures 1.14 and 1.15 for the stacked spectra.

The majority of components seen in H i are also seen in Ca ii, which are both probes of the more diffuse regions of individual clouds. Components only seen in Ca ii and H i gas are likely probing diffuse atomic gas, where densities are too low to excite [C ii] ²P_3/2 →²P_1/2 efficiently or react rapidly enough to produce CO. This accounts for approximately 1/4 of the associated components. Components seen in Ca ii and H i gas that are also associated with any of the additional species considered here typically probe denser gas than the diffuse atomic regime.

The majority of components seen in C⁺, Ca i, K i, CH⁺, CH, or CN also tend to be associated with probes of the very diffuse gas (H i and Ca ii). The presence of C⁺, Ca i, K i, CH⁺, or CH indicates that the component is probing denser material than if only Ca ii or H i were observed (Pan et al. 2005). The additional presence of CN indicates that the component is probing even denser material than if only C⁺, Ca i, K i, CH⁺, or CH were observed.

The species that probe very diffuse gas in emission seem to be associated with species seen in absorption that are seen in diffuse molecular clouds. The presences of C⁺, Ca i, K i, CH⁺, or CH all suggest that there should be some CO being produced, as discussed in Section 3.1, and the presence of CN suggests even higher levels of CO should be present. Thus, if there are no detectable amounts of CO at the same V_lsr, it does not mean CO is not present, but that the conditions and/or density are not suitable for detecting the emission. Therefore, these components with C⁺, Ca i, K i, CH⁺, or CH and no detectable CO are most likely probing regions of CO-dark H₂ gas and components associated with CN and no detectable CO are probably probing the slightly denser inner regions of CO-dark H₂ gas.

These regions of CO-dark H₂ gas can be further probed by looking at the pure rotational emission lines of H₂. These lines have been detected by ISO and then by Spitzer several times, e.g., the Taurus molecular cloud (Goldsmith et al. 2010) and in translucent clouds (Ingalls et al. 2011). The James Webb Space Telescope will likely build on these results using the rotational emission lines of H₂ to probe CO-dark H₂ gas at the edges of molecular clouds.

Detectable amounts of CO emission often correspond to components seen in C⁺, K i, CH⁺, CH, or CN. This accounts for almost half of the associated CO components. The other halves of the associated CO emission components are only associated with Ca ii components. About 70% of the CO emission components also have ¹³CO or C¹⁸O emission. The additional presence of ¹³CO or C¹⁸O indicates a component that is associated with much denser gas, specifically the inner portion of the associated molecular cloud. Detectable amounts of C¹⁸O are only present in the most dense gas regions considered here.