Parsec-scale Variations in the ⁷Li i/⁶Li i Isotope Ratio Toward IC 348 and the Perseus OB 2 Association^{* †}

In the seminal study of stellar nucleosynthesis by Burbidge et al. (1957), the origin of the two stable lithium isotopes was ascribed to an ill-defined process called the x-process. A step toward greater understanding came with the proposal that lithium was a product of Galactic cosmic-ray (GCR) spallation and α–α fusion reactions in interstellar space (Reeves et al. 1970; Meneguzzi et al. 1971, hereafter MAR). Unless the low-energy spectrum of cosmic rays is specifically designed to favor ⁷Li production, GCR spallation fails to account for the solar system value of the ⁷Li/⁶Li ratio (12.2; Lodders 2003), producing instead a ratio of about two. MAR's alternative to crafting the low-energy spectrum was to invoke "thermonuclear ⁷Li from giant stars in 'dirty' regions of our Galaxy."

Today, sources of "thermonuclear" ⁷Li are known. These sources utilize the Cameron–Fowler mechanism (Cameron & Fowler 1971) in which the reaction ³He(α,γ)⁷Be is followed by ⁷Be(e⁻, ν)⁷Li operating in a convective region such that ³He is burnt to ⁷Be at "high" temperatures, but the fragile ⁷Be and ⁷Li are convected out to "low" temperatures and so avoid destruction by protons. Ejection of ⁷Li from stars experiencing the Cameron–Fowler mechanism will raise the interstellar isotopic ratio of lithium from two to a higher value. Known ⁷Li stellar synthesizers include AGB stars (Smith & Lambert 1989), red giant branch stars (Smith et al. 1995; Uttenthaler et al. 2012), novae (Izzo et al. 2015; Tajitsu et al. 2015), and core-collapse supernovae (Woosely & Weaver 1995). In addition, Big Bang nucleosynthesis (BBN) is now recognized as a source of ⁷Li (not ⁶Li), but there remains serious disagreement between the observed and predicted primordial ⁷Li abundance (Fields 2011; Cyburt et al. 2016). This discrepancy influences the need for post-BBN sources of ⁷Li.

Given the various sources of ⁷Li, it is of interest to measure the isotopic ratio in regions where cosmic rays are prevalent. Observations show that both o Per and IC 348 reside within a supernova remnant (SNR) surrounding the Per OB 2 association (Bally et al. 2008). From measured OH (Snow 1975) and HD column densities (Snow 1976), Hartquist & Morfill (1983) conclude that there is a factor of 10 higher cosmic-ray ionization rate toward o Per than toward ζ Per, indicating that the SNR is the source of cosmic rays. These low-energy (≤ a few MeV) cosmic rays are produced when ionized material within the SNR is accelerated and focused by the magnetic turbulence in the direction of o Per. Based on the radio observations of interstellar OH, Sancisi et al. (1974) find all the OH resides on the near side of o Per and conclude that the cosmic rays must originate within the SNR making this star-forming cloud an ideal astrophysical laboratory for further investigations of GCR spallation reactions.

Knauth et al. (2000, 2003) presented moderate- and high-resolution spectroscopic measurements of the ⁷Li/⁶Li isotope ratios for the two interstellar clouds observed toward o Per in IC 348. The weighted averages of these measurements are ⁷Li/⁶Li = 1.9 ± 0.3 and 5.9 ± 3.2 in the 4 and 7 km s⁻¹ interstellar clouds, respectively, These low isotope ratios are significantly different than the solar system value of 12.2 (Lodders 2003) and are consistent with the ratio expected from GCR spallation reactions (MAR).

To determine the extent of these cosmic-ray interactions, we acquired data toward additional stars within several parsecs of o Per in the active star-forming region IC 348. If the superbubble surrounding IC 348 is the source of the GCRs contributing to the lower isotope ratio toward o Per, other nearby sight lines should yield lower than average isotope ratios. Here, we report lithium isotope ratios toward three stars within 15 pc of o Per: HD 23478, HD 24534 (X Per), and HD 281159 (BD+31 643).

We obtained high signal-to-noise, high-resolution (R ≥ 100,000) spectra of the Li i doublet at 6708 Å and the K i doublet at 4044 and 4047 Å toward our three target stars. Table 1 shows the relevant stellar data, and Figure 1 shows the detections of interstellar Li i and K i toward our target stars.

Zoom In Zoom Out Reset image size

2.1. Subaru

2.2. HET

2.3. Data Reduction

For a typical, single, optically thin interstellar cloud that exhibits a solar system lithium isotope ratio of 12.2, the two fine structure lines of the Li i doublet exhibit relative strengths of 2:1 and a separation of ∼0.15 Å. Coincidentally, the isotope shift and fine structure separation of the lithium atom are such that the stronger ⁶Li absorption line is superimposed on the weaker ⁷Li absorption line. This confusion requires high-resolution and high signal-to-noise spectroscopy (S/N ∼ 1000) to extract meaningful ⁷Li/⁶Li isotope ratios.

Additional complications arise when multiple clouds are present along the line of sight, which is typical for the interstellar medium (e.g., Welty & Hobbs 2001). Thus, the first step in our analysis was to derive templates for the line-of-sight component structure using the strong ⁷Li i λ6708, K i λ4044, and CH λ4300 lines because the three species are expected to co-exist in an interstellar cloud (Pan et al. 2005). The weaker K i line at 4044 Å provides a better template for Li i λ6708 than CH λ4300 since both K i and Li i are alkali elements with similar ionization potentials.

Using a C⁺⁺ code, we performed an independent profile synthesis (Knauth et al. 2003) of the Li i, CH, and K i lines, and these results are presented in Table 2. In the analysis, the component velocities, b-values, and column densities of the species are left as free parameters while minimizing the rms deviations in the residuals of the fit. Hyperfine structure is included in our synthetic profiles for Li i and K i (Knauth et al. 2003), but lambda doubling for CH was not considered for these optically thin lines. The LSR velocity and Doppler broadening parameter for ⁷Li and ⁶Li are assumed to be the same, with the column density of each left as a freely varying parameter. The results from our fits to K i and CH were used to determine the number of velocity components to include in the synthesis for Li i. Our best two component fits to the Li i and K i profiles are presented in Figure 1. The column densities, velocities, and b-values resulting from all fits to a given star are in good agreement with each other and the corresponding isotope ratios toward each star are presented in Table 2. Additionally, we performed a consistency check on the Li i profiles using the independently derived CH λ4300 and K i λ4044 velocity structures that yield similar ⁷Li/⁶Li isotope ratios for all three lines of sight in this study.

For HD 23478 and HD 24534, a single component contributes more than 90% of the total column density of Li i, despite the presence of two velocity components detected in both K i and CH at our resolution. However, due to the lower abundance and the inherent uncertainties, only lower limits are presented for the isotope ratio in the second velocity component toward these stars. For HD 281159, the two velocity components observed in both K i and CH are stronger and, as a result, two components in Li i are clearly evident. In addition, the Li i profile exhibits the fine structure lines of ⁷Li of almost equal depth and not the 2:1 relative strength expected indicating a clear strong presence of ⁶Li along this line of sight. For the stars HD 24534 and HD 281159, the ⁷Li/⁶Li ratios derived (from K i, CH, and directly fitting Li i) are indistinguishable; however, a larger scatter is observed toward HD 23478, which is attributed to the lower S/N of that data (see Table 2). For consistency, we adopt the ratios derived from directly fitting the Li i profile for further analysis.

To gain further insight, we determined the N(K i)/N(Li i) and elemental K/Li abundance ratios, which are presented in Table 3. We find N(K i)/N(Li i) = 181 ± 24 toward HD 23478 and slightly higher values of N(K i)/N(Li i) = 290 ± 56 and N(K i)/N(Li i) = 408 ± 133 toward HD 24534 and HD 281159, respectively. Following the prescription laid out in Knauth et al. (2003) for HD 24534 and using the extinction curves for HD 23478 (Papaj et al. 1991) and HD 281159 (Snow et al. 1994), we derived the following elemental K/Li ratios: 50 ± 7 toward HD 23478, 81 ± 16 toward HD 24534, and 113 ± 37 toward HD 281159. For comparison, we include our earlier results at the bottom of Table 3 for both o Per (Knauth et al. 2003) and ζ Per with updated photoionization rates (Knauth et al. 2000, 2003).

The Per OB 2 association has been the subject of numerous investigations due to its close proximity of 316 ± 22 pc (Herbig 1998) and bright stars. These studies reveal that there is an SNR surrounding Per OB 2 with o Per inside that shell on the near side of the dense B5 molecular cloud and ζ Per lying on the far side of the shell (Hartquist & Morfill 1983). The stars o Per and HD 281159 are separated by 8', which corresponds to ∼0.7 pc at a distance of 300 pc (Luhman et al. 2016). The line of sight to HD 281159 resides near the center of the star-forming region IC 348 (Herbig 1998) with o Per just to the north and HD 23478 about 3 pc beyond IC 348 (Krelowski et al. 1996). ζ Per and HD 24534 lie more than 10 pc away from o Per.

Analyses of OH and HD observations toward o Per and ζ Per (Snow 1975, 1976; Hartquist et al. 1978; Federman et al. 1996) reveal an order of magnitude higher cosmic-ray ionization rate (ζ_p = 2.5 × 10⁻¹⁶ s⁻¹) toward o Per than toward ζ Per (ζ_p = 2.2 × 10⁻¹⁷ s⁻¹). Hartquist & Morfill (1983) determine the magnetic field structure of the SNR and claim that the magnetic field focuses the cosmic rays on the dense molecular cloud near o Per, thereby enhancing the cosmic-ray ionization rate. The ionization rate toward ζ Per is lower because the magnetic field enters the region at small angles due to the shock enhancing the perpendicular component of the field.

Our results clearly demonstrate that ⁶Li is enhanced within a few parsecs of IC 348. We find a ⁷Li/⁶Li isotope ratio toward HD 23478 of ⁷Li/⁶Li = 3.7 ± 0.6 and toward HD 281159, the central star of IC 348, reveal low isotope ratios in both velocity components of ⁷Li/⁶Li = 1.8 ± 0.8 and 5.1 ± 1.6, which are remarkably similar to the weighted average toward o Per (Knauth et al. 2000, 2003). The isotope ratios toward the center of IC 348 (HD 23478, HD 281159, and o Per) are lower than the solar system by factors of 2–4 (see Table 3). While for stars further from IC 348 but still within the Per OB 2 Association, our results for ⁷Li/⁶Li of 10.6 ± 2.9 toward ζ Per (Knauth et al. 2000) and of 11.2 ± 4.9 toward HD 24534 are consistent with the solar system value of 12.2 (Lodders 2003). To date, all results are in excellent agreement with the model presented by Hartquist & Morfill (1983).

Indriolo & McCall (2012) measured ${{\rm{H}}}_{3}^{+}$ abundances toward stars in this region and determined the cosmic-ray ionization rates of ζ₂ = (5.85 ± 3.54) × 10⁻¹⁶ s⁻¹ toward HD 24534 and (5.55 ± 3.18) × 10⁻¹⁶ s⁻¹ toward ζ Per. Above, the primary cosmic-ray ionization rate ζ_p = 0.5 ζ₂, given in Indriolo & McCall (2012). However, toward o Per and HD 281159, Indriolo & McCall (2012) report only upper limits of ζ₂ ≤ 2.55 × 10⁻¹⁶ and ζ₂ = ≤5.01 × 10⁻¹⁶, respectively. Although, these upper limits apparently conflict with our findings, it is important to note that the results of Indriolo & McCall (2012) rely on a number of assumptions concerning the physical conditions in the clouds containing H₃⁺. More importantly, the abundance of ${{\rm{H}}}_{3}^{+}$ traces the current cosmic-ray ionization rate, while the Li isotope ratio probes the integrated cosmic-ray flux. Thus, some differences in these two observables might be expected (Taylor et al. 2012).

If the molecular gas toward o Per and HD 281159 has been subjected to an enhanced flux of cosmic rays, in accordance with the Hartquist & Morfill (1983) model, then an increase in the relative abundance of ⁶Li would be a natural consequence. Indeed, the ⁷Li/⁶Li ratio toward both o Per (Knauth et al. 2000, 2003) and HD 281159 approaches the value predicted by models of GCR nucleosynthesis (MAR; Ramaty et al. 1997; Lemoine et al. 1998). A decrease in the ⁷Li/⁶Li ratio due to enhanced cosmic-ray activity (resulting in newly synthesized ⁶Li and ⁷Li) should be accompanied by an increase in the elemental Li abundance. For example, if the clouds toward HD 281159 initially possessed a ⁷Li/⁶Li ratio of 12.2 (the solar system value), then the observed ratio of 5.1 implies a 40% increase in the elemental Li abundance.

Our measurement of the N(K i)/N(Li i) ratio toward HD 23478 is similar to our earlier results toward o Per and ζ Per (Knauth et al. 2000, 2003) and to the average value N(K i)/N(Li i) ≃ 185 found by Welty & Hobbs (2001), who reported a linear relationship between these two species in diffuse interstellar gas toward 54 galactic stars. While the higher values toward HD 24534 and HD 281159 are unexpected, it may not be appropriate to compare these lines of sight to the diffuse interstellar medium. HD 24534 is an X-ray binary with a circumstellar disk (e.g., Li et al. 2014) and the physical and chemical state of the gas has been altered by processes within IC 348 due to the recent star formation (e.g., Snow et al. 1994). Our determinations of the elemental K/Li abundance toward these lines of sight are similar (see Table 3) to the solar system average value of 67.6 (Lodders 2003). An alternative interpretation could be that there is either enhanced depletion (or ionization) of Li i and K i in this region.

The discovery of a second low ⁷Li/⁶Li ratio toward HD 281159, the central star in IC 348, supports the picture that the SNR associated with the Per OB 2 Association is the source of the cosmic-ray activity (Knauth et al. 2000, 2003). The observed ⁷Li/⁶Li ∼ 2 as theoretically predicted (MAR) occurs through both alpha–alpha fusion reactions with interstellar He (Fields & Prodanovic 2005) and cosmic rays bombarding C, N, and O nuclei in the interstellar medium surrounding the IC 348 region. Cosmic-ray bombardment of the interstellar medium also creates Be and B. It is interesting to note that Ritchey et al. (2011) reported a 50% enhancement in the B ii abundance toward o Per compared to three other stars in this region (ζ Per, 40 Per, and HD 24534), which provides further support for cosmic-ray spallation reactions occurring in this region. However, there remains the unsatisfying result that there is no evidence of enhanced Li i with respect to K i or in the elemental K/Li ratio. Further high-quality observations of the ⁷Li/⁶Li and ¹¹B/¹⁰B isotope ratios toward additional stars in IC 348 and other OB associations (e.g., Sco OB2 and Cyg OB2) are necessary to further quantify light element production via the GCR spallation mechanism.

We thank Nick Indriolo for useful discussions concerning the cosmic-ray ionization rates toward the Perseus stars. Alanna Garay and Akito Tajitsu, the dedicated telescope operators at the Subaru telescope, and the telescope operators at the Hobby-Eberly telescope for their invaluable assistance with collecting these data. D.L.L thanks the Robert A. Welch Foundation of Houston, Texas, through grant F-634. We would like to thank the anonymous referee for their worthwhile comments on our Letter.

This Letter includes data taken at The McDonald Observatory of The University of Texas at Austin. The Hobby-Eberly Telescope (HET) is a joint project of the University of Texas at Austin, the Pennsylvania State University, Ludwig-Maximilians-Universität München, and Georg-August-Universität Göttingen. The HET is named in honor of its principal benefactors, William P. Hobby Jr. and Robert E. Eberly.