http://www.nature.com/nature/journal/v535/n7613/full/535500a.html　

Solar-type dynamo behaviour in fully convective stars without a tachocline
• Nicholas J. Wright
• & Jeremy J. Drake
• Affiliations
• Contributions
• Corresponding author
Nature

535,

526–528

(28 July 2016)

doi:10.1038/nature18638
Received

25 April 2016
Accepted

02 June 2016
Published online

27 July 2016
Article tools
• PDF
• Citation
• Reprints
• Rights & permissions
• Article metrics
In solar-type stars (with radiative cores and convective envelopes like our Sun), the magnetic field powers star spots, flares and other solar phenomena, as well as chromospheric and coronal emission at ultraviolet to X-ray wavelengths. The dynamo responsible for generating the field depends on the shearing of internal magnetic fields by differential rotation1, 2. The shearing has long been thought to take place in a boundary layer known as the tachocline between the radiative core and the convective envelope3. Fully convective stars do not have a tachocline and their dynamo mechanism is expected to be very different4, although its exact form and physical dependencies are not known. Here we report observations of four fully convective stars whose X-ray emission correlates with their rotation periods in the same way as in solar-type stars. As the X-ray activity–rotation relationship is a well-established proxy for the behaviour of the magnetic dynamo, these results imply that fully convective stars also operate a solar-type dynamo. The lack of a tachocline in fully convective stars therefore suggests that this is not a critical ingredient in the solar dynamo and supports models in which the dynamo originates throughout the convection zone.
Subject terms:
• Stars
• Solar physics
Main
• Main•
• Methods•
• References•
• Acknowledgements•
• Author information
Stars across the Hertzsprung–Russell diagram are known to emit X-rays, with only a few exceptions. In main-sequence solar-type and low-mass stars, the X-rays arise from a magnetically confined plasma known as a corona that reaches temperatures of several million kelvin5. Coronal X-ray emission is ultimately powered by the dissipation of the magnetic fields generated by an interior magnetic dynamo. A close relation between the surface magnetic flux and X-ray radiance based on both solar and stellar observations that span several decades for both quantities6 indicates that X-ray emission is a reliable proxy for magnetic activity. The dynamo is thought to be driven in part by differential rotation in the interior of the star1, 2, which itself is generated by the action of the Coriolis force on the rotating convective envelope, but the detailed mechanism remains to be properly understood7. The relationship between stellar rotation and tracers of magnetic activity is therefore an important probe of the stellar dynamo.
In solar-type stars, X-ray emission is observed to increase monotonically with increasing stellar rotational velocity for periods exceeding a few days8, 9. This relationship is often quantified in terms of the dependence of the ratio of stellar luminosity expended in X-rays to bolometric luminosity LX/Lbol, on the Rossby number, which is defined as Ro = Prot/τ, the ratio of the stellar rotation period and the mass-dependent convective turnover time10. A recent study9 using the largest available sample of 824 solar- and late-type stars fitted the rotation–activity relation as LX/Lbol = 5.3 × 10−6Ro−2.7.
This relationship has been observed in stars from late F-type through to early M-type, that is, those with radiative cores and convective envelopes. The interface layer between these two regions, named the tachocline, is believed to play an important role in the generation of the magnetic field3. Shear between the rigidly rotating core and the differential rotation of the convective envelope with latitude is thought to amplify and store the magnetic field11, generating what is known as an α–Ω dynamo, named for the interplay between cyclonic eddies (the α effect) and the shearing of the field (the Ω effect).
Others have argued that the latitudinal and radial gradients of the angular velocity in the convection zone may be sufficient for global dynamo action12, 13. In the Sun, the extremely strong levels of radial shear just beneath the surface are actually higher than in the tachocline, making it plausible that the solar dynamo is distributed across the convection zone rather than confined to the tachocline12.
Despite this hypothesis it has been widely accepted that magnetic structures in the convection zone would be disrupted by magnetic buoyancy or turbulent pumping, preventing large-scale magnetic fields from being established there14. However, some recent three-dimensional magneto-hydrodynamic simulations without a tachocline have produced persistent magnetic wreaths in the convection zone15 and shown that it is possible to produce large-scale magnetic fields in stellar convection layers16, 17.
For very fast rotators, the rotation–activity relationship has been found to break down, with X-ray luminosity reaching a saturation level of approximately LX/Lbol ≈ 10−3, independent of the spectral type18. This saturation level is reached at a Rossby number of approximately 0.13 ± 0.02 (ref. 9; the error represents one standard deviation), corresponding to a rotation period that increases towards later spectral types, from 2 days for a star similar to the Sun to up to about 20 days for low-mass M dwarfs, and is also seen in both chromospheric emission and magnetic field measurements. It is unclear whether this is caused by a saturation of either the dynamo mechanism or the transport of the magnetic flux to the corona, a change in the type of dynamo at work within the star9 or because coronal X-ray emission itself becomes insensitive to the strength of the magnetic field, the energy of which is then dissipated in other ways.
Main-sequence stars later than spectral type M3–3.5 (M < 0.4M⊙, where M is the mass of the star and M⊙ is the mass of the Sun) are predicted to be fully convective and therefore do not possess a tachocline. If the tachocline is critical to the operation of a solar-type dynamo, fully convective stars should not be able to sustain such a dynamo. Instead, it is generally thought that they generate magnetic fields entirely by helical turbulence4. Nevertheless, observations indicate that stars throughout the M-type spectral range exhibit high magnetic field strengths19 and high fractional X-ray luminosities9. In fact, no discernible difference in magnetic activity properties has been identified on either side of the fully convective boundary.
One problem with existing studies is that nearly all fully convective stars that have so far been studied have saturated levels of X-ray emission9. It is therefore unclear whether these stars all exhibit saturated levels of X-ray emission (possibly hinting at some facet of their dynamo mechanism), or whether slower rotators follow a more solar-like rotation–activity relationship. Studies of magnetic activity in slowly rotating fully convective stars have so far been lacking, adding to the uncertainty about their dynamo mechanism. Despite this, fully convective stars are common and, given their spin-down times of a few billion years20, at least half of all such stars are expected to be slow rotators.
Understanding the dynamo mechanism in these slowly rotating stars is important for various astrophysical problems, including the dynamo-driven angular momentum loss rate of low-mass stars, the particle and photon radiation environment of exoplanets and the notorious period gap in the cataclysmic variables. The lack of cataclysmic variables with periods in the 2–3 h range is often attributed to a change in rotational spin-down for fully convective stars with diminished magnetic dynamos and stellar winds21. Recent numerical simulations of stellar winds suggest that increasing the complexity of the surface magnetic morphology can also suppress angular momentum loss and spin-down without requiring any change in the total surface magnetic flux22.
Four slowly rotating, fully convective M-type stars were observed by either NASA’s Chandra X-ray Observatory or the ROSAT satellite. These stars are of spectral type M4–5.5, well beyond the fully convective boundary (M3–3.5) and are therefore genuine fully convective stars. Figure 1 illustrates the traditional rotation–activity diagram showing the fractional X-ray luminosity as a function of the Rossby number for all of the stars from the most recent large-scale study of the rotation–activity relationship in solar-type and low-mass stars9. The positions of the four slowly rotating fully convective stars studied in this work are also shown, and are in excellent agreement with the rotation–activity relationship of partly convective stars.
Figure 1: Rotation–activity relationship diagram for partly and fully convective stars.

Fractional X-ray luminosity, LX/Lbol, plotted against the Rossby number, Ro = Prot/τ, for 824 partly (grey circles) and fully (dark red circles) convective stars from the most recent large compilation of stars with measured rotation periods and X-ray luminosities7. The best-fitting saturated (horizontal) and unsaturated (diagonal) rotation–activity relationships from that study are shown as black dashed lines. The four slowly rotating fully convective M dwarfs studied here are shown in light red (error bars indicate one standard deviation). The uncertainties for the other data points are not quantified, but will be comparable to the M dwarfs for the Rossby number and approximately twice as large for LX/Lbol.
• Full size image (169 KB)
• Download CSV source data (55 KB)
• Download PowerPoint slide (490 KB)
The observations presented here provide clear evidence for unsaturated X-ray emission in fully convective stars and quantification of their rotation–activity relationship. The results show that fully convective stars, at least when they have spun down sufficiently, operate a dynamo that exhibits a rotation–activity relationship that is indistinguishable from that of solar-type stars. As the dynamo action in fully convective stars is expected to be different from that in solar-type stars owing to their lack of a tachocline4, where the magnetic field is thought to be amplified by radial shear, this is a surprising finding.
The most direct conclusion from these observations is that both partly and fully convective stars operate very similar rotation-dependent dynamos in which the tachocline is not a vital ingredient and differential rotation combined with the action of the Coriolis force is sufficient7. This implies that current models for the solar dynamo that rely on the tachocline layer to amplify the magnetic field are incorrect, and lends weight to recent three-dimensional magneto-hydrodynamic simulations without a tachocline that produce large-scale magnetic fields entirely within the convective layers23, 24. Recent studies with mean-field dynamo models have also suggested that differential rotation in the convection zone may play a greater role in the generation of the magnetic field than does the tachocline25, 26.
Another alternative possibility is that fully convective stars are able to generate a purely turbulent dynamo that exhibits a rotation–activity relationship that is similar to that in partly convective stars, but by a different mechanism that does not rely on a shear layer. Existing dynamo simulations for fully convective stars succeed in generating magnetic fields, but are unable to predict their behaviour as a function of the rotation rate17. However, it seems unlikely that both partly and fully convective stars would have the same rotation–activity relationship (requiring both their dynamo efficiency and rotational dependence to behave in the same way) without their dynamo mechanisms sharing a major feature.
A third possibility is that convection in the cores of fully convective stars could be magnetically suppressed27, leading to the existence of a solar-like tachocline, although some studies suggest that convection would not be completely halted, only made less efficient28. Furthermore, the field strengths that are necessary for such a transition are 107–108 G (refs 28, 29), orders of magnitude larger than the fields thought to exist in the solar interior and at levels that simulations suggest are impossible to maintain30.
Methods
• Main•
• Methods•
• References•
• Acknowledgements•
• Author information
New observations
Two targets were chosen from the MEarth Transit survey20 of fully convective stars: G184-31 and GJ 3253, of spectral types M4.531 and M532 with rotation periods of 83.8 days (ref. 33) and 78.8 days (ref. 20), respectively.
The targets were observed with the Advanced CCD Imaging Spectrometer (ACIS)34 on the Chandra X-ray Observatory35 using the ACIS-S (spectroscopic) CCD array. Observations were performed in ‘very faint’ mode and were placed on the back-illuminated S3 (the 3rd CCD chip in the ACIS-S spectroscopic array) chip (owing to its higher sensitivity to soft X-rays. The observations were performed on 6 December 2012 and 25 September 2013 for G184-31 and GJ 3253 respectively, with exposure times of 8.0 ks and 21.0 ks.
Observations were processed using the CIAO 4.5 software tools36 and the CALDB 4.5.8 calibration files following standard procedures. The two sources were clearly identified and detected with significances of 5.0σ and 10.2σ at their expected positions. Point-source extraction was performed using CIAO 4.5. The total number of net counts was measured to be and for G184-31 and GJ 3253, respectively. Light curves were constructed for both sources to search for high levels of variability that might inflate the quiescent flux level measured, but no significant variability was detected.
Thermal plasma X-ray spectral models were fitted to the extracted spectra using XSPEC37 version 12.6.0 and compared to APEC38 single-temperature optically thin model spectra of an absorbed thermal plasma in collisional ionization equilibrium, allowing the plasma temperature (kBT, where kB is Boltzmann’s constant and T is temperature) and the hydrogen column density (NH) to vary freely. A grid of initial thermal plasma temperatures covering the range kBT = 0.1–3.0 keV was used to prevent fitting to local minima. The model with the lowest Cash statistic39 was selected as the best fit for each source (the Cash statistic is an application of the likelihood ratio test that is suitable for low-signal data). The best-fitting thermal plasma temperatures were found to be kBT = 0.78 ± 0.13 keV and 0.30 ± 0.05 keV for G184-31 and GJ 3253, respectively, consistent with the values found for other M-type dwarf stars. The fitted hydrogen column density is consistent with no absorption, as expected for the proximity of these sources.
Absorption-corrected broadband (0.5–8.0 keV) fluxes of FX = (2.08 ± 0.38) × 10−14 erg/s/cm2 and (4.35 ± 0.41) × 10−14 erg/s/cm2 were calculated from the model fits for G184-31 and GJ 3253 respectively. Combined with their known parallax distances31, 32 the fluxes were used to calculate X-ray luminosities in the ROSAT band (0.1–2.4 keV), for consistency and ease of comparison with previous studies7. Fractional X-ray luminosities were then calculated from the observed J-band magnitudes and the appropriate bolometric corrections40.
Literature data
We searched the literature for fully convective stars (spectral type M4 or later) with existing measured rotation periods and X-ray luminosities, excluding any with short rotation periods (Prot < 20 days), which would place the object in the saturated regime of Fig. 1. Two stars were found that met our criteria: GJ 699 (Barnard’s star, M4V; ref. 41) and GJ 551 (Proxima Centauri, M5.5V; ref. 42) with rotation periods43 and X-ray luminosity values44 existing in the literature. Rossby numbers and fractional X-ray luminosities for these stars were calculated as for the two newly observed stars.
We also uncovered a number of stars that are close to the convective boundary in the literature45, 46 (spectral types M3 or M3.5). The spectral types and colours9 of these stars imply that they are partly convective and so they are shown in Fig. 1as grey dots.
Sample size
No statistical methods were used to predetermine sample size.
Code availability
The CIAO code used to reduce the Chandra X-ray Observatory data are available at http://cxc.cfa.harvard.edu/ciao and the associated calibration database can be found at http://cxc.cfa.harvard.edu/caldb. The XSPEC code used to perform X-ray spectral fitting is available at https://heasarc.gsfc.nasa.gov/xanadu/xspec.
References
• Main•
• Methods•
• References•
• Acknowledgements•
• Author information
1. Parker, E. N. Hydrodynamic dynamo models. Astrophys. J. 122, 293–314(1955)
o Article
o Show context
2. Babcock, H. W. The Sun’s magnetic field. Annu. Rev. Astron. Astrophys. 1,41–58 (1963)
o Article
o Show context
3. Speigel, E. A. & Zahn, J.-P. The solar tachocline. Astron. Astrophys. 265,106–114 (1992)
o ISI
o Show context
4. Durney, B. R., De Young, D. S. & Roxburgh, I. W. On the generation of the large-scale and turbulent magnetic fields in solar-type stars. Sol. Phys. 145,207–225 (1993)
o ISI
o Article
o Show context
5. Vaiana, G. S. et al. Results from an extensive Einstein stellar survey.Astrophys. J. 244, 163–182 (1981)
o Article
o Show context
6. Pevtsov, A. A. et al. The relationship between X-ray radiance and magnetic flux. Astrophys. J. 598, 1387–1391 (2003)
o CAS
o Article
o Show context
7. Spruit, H. C. Theories of the Solar Cycle: a Critical View 39–54 (Springer,2011)
o Show context
8. Pallavicini, R. et al. Relations among stellar X-ray emission observed from Einstein, stellar rotation and bolometric luminosity. Astrophys. J. 248,279–290 (1981)
o Article
o Show context
9. Wright, N. J., Drake, J. J., Mamajek, E. E. & Henry, G. W. The stellar-activity-rotation relationship and the evolution of stellar dynamos.Astrophys. J. 743, 48 (2011)
o Article
o Show context
10. Noyes, R. W., Hartmann, L. W., Baliunas, S. L., Duncan, D. K. & Vaughan, A. H. Rotation, convection, and magnetic activity in lower main-sequence stars. Astrophys. J. 279, 763–777 (1984)
o CAS
o ISI
o Article
o Show context
11. Charbonneau, P. Solar dynamo theory. Annu. Rev. Astron. Astrophys. 52,251–290 (2014)
o Article
o Show context
12. Brandenburg, A. Location of the solar dynamo and near-surface shear. InASP Conf. Series Vol. 354 (eds Leibacher, J., Stein, R. F. & Uitenbroek, H.) 121–126 (Astronomical Society of the Pacific, 2006)
o Show context
13. Guerrero, G. & de Gouveia Dal Pino, E. M. How does the shape and thickness of the tachocline affect the distribution of the toroidal magnetic fields in the solar dynamo? Astron. Astrophys. 464, 341–349 (2007)
o Article
o Show context
14. Parker, E. N. The generation of magnetic fields in astrophysical bodies. X. Magnetic buoyancy and the solar dynamo. Astrophys. J. 198, 205–209(1975)
o ISI
o Article
o Show context
15. Brown, B. P., Browning, M. K., Brun, A. S., Miesch, M. S. & Toomre, J.Persistent magnetic wreaths in a rapidly rotating Sun. Astrophys. J. 711,424–438 (2010)
o Article
o Show context
16. Dobler, W., Stix, M. & Brandenburg, A. Magnetic field generation in fully convective rotating spheres. Astrophys. J. 638, 336–347 (2006)
o Article
o Show context
17. Browning, M. K. Simulations of dynamo action in fully convective stars.Astrophys. J. 676, 1262–1280 (2008)
o CAS
o Article
o Show context
18. Vilhu, O. The nature of magnetic activity in lower main sequence stars.Astron. Astrophys. 133, 117–126 (1984)
o Show context
19. Johns-Krull, C. M. & Valenti, J. A. Detection of strong magnetic fields on M dwarfs. Astrophys. J. 459, 95–99 (1996)
o Article
o Show context
20. Irwin, J. et al. On the angular momentum evolution of fully convective stars: rotation periods for field M-dwarfs from the MEarth transit survey.Astrophys. J. 727, 56 (2011)
o Article
o Show context
21. Spruit, H. C. & Ritter, H. Stellar activity and the period gap in cataclysmic variables. Astron. Astrophys. 124, 267–272 (1983)
o Show context
22. Garraffo, C., Drake, J. J. & Cohen, O. Magnetic complexity as an explanation for bimodal rotation populations among young stars. Astrophys. J. 807, L6 (2015)
o Article
o Show context
23. Nelson, N. J., Brown, B. P., Brun, A. S., Miesch, M. S. & Toomre, J.Magnetic wreaths and cycles in convective dynamos. Astrophys. J. 762, 73(2013)
o Article
o Show context
24. Fan, Y. & Fang, F. A simulation of convective dynamo in the solar convective envelope: maintenance of the solar-like differential rotation and emerging flux. Astrophys. J. 789, 35 (2014)
o Article
o Show context
25. Dikpati, M. & Gilman, P. A. Simulating and predicting solar cycles using a flux-transport dynamo. Astrophys. J. 649, 498–514 (2006)
o Article
o Show context
26. Muñoz-Jaramillo, A., Nandy, D. & Martens, P. C. H. Helioseismic data inclusion in solar dynamo models. Astrophys. J. 698, 461–478 (2009)
o Article
o Show context
27. Cox, A. N., Hodson, S. W. & Shaviv, G. On the ratio of mixing length to scale height in red dwarfs. Astrophys. J. 245, L37–L40 (1981)
o Article
o Show context
28. Moss, D. L. & Taylor, R. J. The influence of a poloidal magnetic field on convection in stellar cores. Mon. Not. R. Astron. Soc. 147, 133–138 (1970)
o Article
o Show context
29. Mullan, D. J. & MacDonald, J. Are magnetically active low-mass M dwarfs completely convective? Astrophys. J. 559, 353–371 (2001)
o Article
o Show context
30. Browning, M. K., Weber, M. A., Chabrier, G. & Massey, A. P. Theoretical limits on magnetic field strengths in low-mass stars. Astrophys. J. 818, 189(2016)
o Article
o Show context
31. Reid, I. N. et al. Meeting the cool neighbors. VIII. A preliminary 20 parsec census from the NLTT catalogue. Astron. J. 128, 463–483 (2004)
o Article
o Show context
32. Newton, E. R. et al. Near-infrared metallicities, radial velocities, and spectral types for 447 nearby M dwarfs. Astron. J. 147, 20 (2014)
o CAS
o Article
o Show context
33. West, A. A. et al. An activity-rotation relationship and kinematic analysis of nearby mid-to-late-type M dwarfs. Astrophys. J. 812, 3 (2015)
o Article
o Show context
34. Garmire, G. P., Bautz, M. W., Ford, P. G., Nousek, J. A. & Ricker, G. R. Jr.Advanced CCD Imaging Spectrometer (ACIS) instrument on the Chandra X-ray Observatory. Proc. SPIE 4851, 28–44 (2003)
o Show context
35. Weisskopf, M. C., Tananbaum, H. D., Van Speybroeck, L. P. & O’Dell, S. L.Chandra X-ray Observatory (CXO): overview. Proc. SPIE 4012, 2–16(2000)
o Show context
36. Fruscione, A. et al. CIAO: Chandra’s data analysis system. Proc. SPIE6270, 62701V (2006)
o Show context
37. Arnauld, K. A. XSPEC: the first ten years. ASP Conf. Series 101, 17–20(Astronomical Society of the Pacific, 1996)
o Show context
38. Smith, R. K., Brickhouse, N. S., Liedahl, D. A. & Raymond, J. C. Collisional plasma models with APEC/APED: emission-line diagnostics of hydrogen-like and helium-like ions. Astrophys. J. 556, L91–L95 (2001)
o CAS
o ISI
o Article
o Show context
39. Cash, W. Parameter estimation in astronomy through application of the likelihood ratio. Astrophys. J. 228, 939–947 (1979)
o Article
o Show context
40. Cruz, K. L. & Reid, I. N. Meeting the cool neighbors. III. Spectroscopy of northern NLTT stars. Astron. J. 123, 2828–2840 (2002)
o Article
o Show context
41. Kirkpatrick, J. D., Henry, T. J. & McCarthy, D. W. Jr. A standard stellar spectral sequence in the red/near-infrared – classes K5 to M9. Astrophys. J. Suppl. Ser. 77, 417–440 (1991)
o Article
o Show context
42. Bessell, M. S. The late-M dwarfs. Astron. J. 101, 662–676 (1991)
o CAS
o ISI
o Article
o Show context
43. Benedict, G. F. et al. Photometry of Proxima Centauri and Barnard’s Star using Hubble Space Telescope Fine Guidance Sensor 3: a search for periodic variations. Astron. J. 116, 429–439 (1998)
o Article
o Show context
44. Schmitt, J. H. M. M. & Liefke, C. NEXXUS: a comprehensive ROSAT survey of coronal X-ray emission among nearby solar-like stars. Astron. Astrophys. 417, 651–665 (2004)
o Article
o Show context
45. Kiraga, M. & Stepien, K. Age-rotation-activity relations for M dwarf stars based on ASAS photometric data. Acta Astron. 57, 149–168 (2007)
o Show context
46. Jeffries, R. D., Jackson, R. J., Briggs, K. R., Evans, P. A. & Pye, J. P.Investigating coronal saturation and supersaturation in fast-rotating M-dwarf stars. Mon. Not. R. Astron. Soc. 411, 2099–2112 (2011)
o Article
o Show context
Download references
Acknowledgements
• Main•
• Methods•
• References•
• Acknowledgements•
• Author information
N.J.W. acknowledges a Royal Astronomical Society Research Fellowship and an STFC Ernest Rutherford Fellowship. J.J.D. was supported by NASA contract NAS8-03060 to the Chandra X-ray Center. We thank J. Irwin, R. Jeffries and A. West for assistance and comments on an early draft of this paper. This research has made use of the Vizier (http://vizier.u-strasbg.fr/cgi-bin/VizieR) and SIMBAD (http://simbad.u-strasbg.fr/simbad/sim-fid) databases (operated at CDS, Strasbourg, France).
Author information
• Main•
• Methods•
• References•
• Acknowledgements•
• Author information
Affiliations
1. Astrophysics Group, Keele University, Keele ST5 5BG, UK
o Nicholas J. Wright
2. Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, Massachusetts 02138, USA
o Jeremy J. Drake
Contributions
N.J.W. reduced the Chandra observations, measured the X-ray fluxes and made the necessary calculations to plot the stars in Fig. 1. N.J.W. and J.J.D. wrote the interpretation and discussion of the results.
Competing financial interests
The authors declare no competing financial interests.
Corresponding author
Correspondence to:
• Nicholas J. Wright
Reviewer Information Nature thanks D. Moss and the other anonymous reviewer(s) for their contribution to the peer review of this work.