Shocks and Ejecta Mass: Radio Observations of Nova V1723 Aql
Jennifer H. S. Weston, Jennifer L. Sokoloski, Yong Zheng, Laura Chomiuk, Amy Mioduszewski, Koji Mukai, Michael P. Rupen, Miriam I. Krauss, Nirupam Roy, Thomas Nelson
aa r X i v : . [ a s t r o - ph . S R ] J u l STELLA NOVAE: FUTURE AND PAST DECADESASP Conference Series, Vol. **Volume Number**P. A. Woudt and V. A. R. M. Ribeiro, eds c (cid:13) Shocks and Ejecta Mass: Radio Observations of Nova V1723 Aql
Jennifer H. S. Weston, Jennifer L. Sokoloski, Yong Zheng, LauraChomiuk, , Amy Mioduszewski, Koji Mukai, , Michael P. Rupen , MiriamI. Krauss, Nirupam Roy, and Thomas Nelson. Columbia Astrophysics Laboratory, Columbia University, New York, NY10027, USA Department of Physics and Astronomy, Michigan State University, EastLansing, MI 48824, USA National Radio Astronomy Observatory, P.O. Box O, Socorro, NM 87801,USA CRESST and X-ray Astrophysics Laboratory, NASA / GSFC, Greenbelt, MD20771, USA Department of Physics, University of Maryland, Baltimore County, 1000Hilltop Circle, Baltimore, MD 21250, USA School of Physics and Astronomy, University of Minnesota, 116 ChurchStreet SE, Minneapolis, MN 55455, USA
Abstract.
The radio light curves of novae rise and fall over the course of months toyears, allowing for detailed observations of the evolution of the nova shell. However,the main parameter determined by radio models of nova explosions — the mass ofthe ejecta — often seems to exceed theoretical expectations by an order of magnitude.With the recent technological improvements on the Karl G. Jansky Very Large Array(VLA), new observations can test the assumptions upon which ejecta mass estimatesare based. Early observations of the classical nova V1723 Aql showed an unexpectedlyrapid rise in radio flux density and a distinct bump in the radio light curve on the riseto radio maximum, which is inconsistent with the simple model of spherical ejectaexpelled in a single discrete event. This initial bump appears to indicate the presenceof shocked material in the outer region of the ejected shell, with the emission fromthe shocks fading over time. We explore possible origins for this emission and itsrelation to the mass loss history of the nova. The evolution of the radio spectrum alsoreveals the density profile, the mass of the ejected shell, and other properties of theejecta. These observations comprise one of the most complete, longterm set of multi-wavelength radio observations for any classical nova to date.
Introduction
The nova V1723 Aql went into outburst on September 11th, 2010 (Yamanaka et al.2010; Balam et al. 2010) . Based solely on optical observations, V1723 Aql is a mem-ber of the “fast nova” speed class (Gaposchkin 1957), fading by 2 magnitudes over 20days, with emission lines implying an expansion velocity of ∼ / s on the day of1 Weston, et al.discovery (Yamanaka et al. 2010). It is neither embedded in the wind of an M giantcompanion nor a recurrent nova, making it representative of the majority of novae. Infact, V1723 Aql was the first northern classical nova bright enough for detailed obser-vations in radio since the recent expansion of the VLA radio array.Given a few reasonable assumptions, radio observations can be used to trace thebulk of the ejected mass of a nova simply and accurately. When we observe a nova inthe radio regime, the spectrum reveals whether the emission is thermal bremsstrahlungradiation. A nova shell emitting thermal free-free emission starts as optically thick atall radio frequencies. During this stage of the nova’s evolution, the radio flux densitiesare proportional to the surface area of the shell projected on the sky, and as the shellexpands it becomes brighter at all radio frequencies. The flux density during this timeshould depend only on the distance to the nova, the temperature of the emitting material,and the maximum velocity of the ejecta. However, as the density drops, radio frequencyemission starts to penetrate through part of shell, resulting in a mix of optically thickand thin emission. During this time, the light curve peaks and starts to turn over. Oncethe photospheres have transitioned through the entire shell, we see purely optically thinemission. The shell transitions towards being optically thin at higher frequencies first,followed by the lower frequencies, until eventually the entire shell is optically thin atall wavelengths. The timing of the period when the light curve peaks and starts to turnover and the resulting rate of decline of the flux density light curve depends critically onthe mass of the ejecta and on the density profile (Bode & Evans 2008; Hjellming et al.1979). Therefore, the evolution of the radio opacity at di ff erent wavelengths can beused to infer density profiles of the ejected shell, and trace mass and density changeswithin the shell.When modeling the behavior of a free-free emitting nova shell, we begin by mak-ing several additional assumptions. We assume a spherically symmetric nova shell, witha constant temperature of 10 K throughout its evolution. We use a velocity distributionsuch that velocity is proportional to the distance to the white dwarf, known as the “Hub-ble flow” velocity distribution, fitting the ratio between innermost and outermost veloc-ity to the data (Hjellming et al. 1979). Finally, for the distribution of the ejected mass,we assume a smooth density profile that goes as 1 / r . This simple model has been rela-tively e ff ective in explaining radio light curves of classical novae in the past, e.g. V1500Cyg, QU Vul, and V723 Cas (Hjellming et al. 1979; Taylor et al. 1988; Heywood et al.2005). However, only a small handful of classical novae have well-sampled radio lightcurves spanning the evolution from optically-thick rise to optically-thin decay. Observations and Results
We have now been monitoring this source with the VLA for over 2 years at multiplefrequencies, starting 2 weeks after the initial discovery, and have produced one of thebest, most detailed radio light curves of a classical nova to date. Krauss et al. (2011)presented the initial report of the early part of the VLA observations. To determine themass of the nova shell for V1723 Aql, we first consider the changes in the radio spec-trum (Figure 1). On days 325 and 350 there is clearly one break in the spectrum, whichhas two parts: the rising optically thick portion at low frequencies, and a shallowerspectrum revealing the region where material is transitioning between being opticallythick and thin at higher frequencies. However, by day 426 there are two breaks visible.The lowest frequencies exhibit a spectrum of an optically thick photosphere, the middleadio Observations ofV1723 Aql 3
Figure 1. Spectrum of V1723 Aql during transition from optically thick to thin.The breaks in the spectrum reveal the transition of the radio photospheres from opti-cally thick to optically thin. frequencies show a mix of optically thin and thick emission, and the highest frequen-cies produce purely optically thin emission, demonstrating that the highest frequencyphotospheres have completely penetrated the nova shell.The low frequency break shows the frequency and time the emitting shell hasan optical depth of one ( τ ν = α ∼ . / r (Bode & Evans 2008). A 1 / r density profile is at least initially consistent with ourmeasurements, as we get a spectral index of 0.6 on day 325 between 20.1 and 36.5GHz. The high frequency break shows that there is a distinct inner boundary to theshell, which radio emission can fully penetrate. We can find the outer radius at thisepoch by using the expansion velocity and the time from outburst, and the inner radiusby using the flux density to find the angular size of the τ = E M = R R o R i n ( r ) dl , we integratedensity over the volume of the shell and get an independent mass estimate at eachepoch we see the second break. Using the epochs where we see multiple breaks, wefind an ejecta shell mass of 1 . ± . × − M ⊙ . Modeling our late time light curve,we find our best fit for an ejecta mass of M ∼ × − M ⊙ , an ejecta temperature of T ∼ K, a maximum velocity v ∼ / s, a distance of d ∼ . / v ∼ . t as the brightness increases proportionally to the surface area of the shell during the Weston, et al. Figure 2. Full set of observations of V1723 Aql. Model indicates light curve of anexpanding thermal shell of mass M ∼ × − M ⊙ , ejecta temperature of T ∼ K,maximum velocity v ∼ / s, distance of d ∼ . / v ∼ . first six months of a thermal shell’s expansion. Contrasting this to our early observa-tions of V1723 Aql, we instead see a distinct bump that is inexplicable by an expandingthermal shell alone (Figure 2), as discussed in Krauss et al. (2011). The radio flux den-sity initially rises rapidly, proportional to t . , then decreases and rises again, causinga dip in the light curve. In addition, the spectrum during the bump and the dip is notconsistent with optically thick emission. During the initial rise the spectrum is shal-lower than is consistent with a purely optical thick expansion. Over the course of thebump, the source produces an optically thin flash and the spectrum flattens further, be-fore returning to an optically thick spectrum when the flux density dips. These spectralchanges are the opposite of what one would expect for an expanding optically thicksource.While the light curve and spectrum are not consistent with an expanding shellalone, they could be explained by having multiple emission regions. If there werea shocked outside region then the increase in temperature in the shock-heated plasmawould cause the post-shock ejecta to become optically thin at higher densities, resultingin the optically thin flash. This optically thin emission would fade as the plasma cooled.The expansion of the bulk of the ejecta would then cause the second peak in the lightcurve as the underlying optically thick material expanded. Thus, instead of a singlecomponent model, we instead have a two component spectrum: the bulk of the materialcomprising of an expanding shell exhibiting optically thick free-free emission, and out-side of that, a region that has been heated by shocks producing optically thin emission.adio Observations ofV1723 Aql 5Observations with Swift during the radio bump period detected V1723 Aql as an X-raysource (Krauss et al. 2011); this X-ray emission also suggests strong shocks, supportingthe two component model. Any secondary emission component must be outside of theoptically thick shell, since any internal shock would not be visible through the opticallythick material. Therefore, either the nova shell expanded into some outside material,or the ejecta itself may have collided with slower material that was ejected prior tothis point. V1723 Aql is not embedded within the wind of a red giant companion, sothere is no signature of a surrounding dense environment which might be the cause ofstrong shocks and synchrotron emission, unlike the cases of RS Oph and V407 Cyg(O’Brien et al. 2006; Rupen et al. 2008; Sokoloski et al. 2008; Chomiuk et al. 2012a).This fact could indicate that the unusual light curve of V1723 Aql is a result of multipleperiods of ejection, similar to the case of T Pyx in 2011 (Nelson et al. 2012). Final Thoughts
Figure 3. Radio Observations of Nova Sco 2012
Prior to spring 2012, the bump in the light curve of V1723 Aql was a uniqueoccurrence in radio observations of a classical novae, although hints of a similar phe-nomenon might have been precent in V1974 Cyg (Lloyd et al. 1996). However, VLAobservations of Nova Sco 2012 (Chomiuk et al. 2012b) show a similar steep rise in thelight curve, followed by a period where the flux density falls, then rises again (figure3). While V1723 Aql appears to be an otherwise typical classical nova, Nova Sco 2012is not – it was detected in γ -rays over the course of its outburst (Cheung et al. 2012).One can speculate that whatever caused γ -rays in Nova Sco may be related to the causeof the bump in V1723 Aql. V1723 Aql is by all accounts a typical classical nova inthe optical spectrum — it is only in the radio regime that this irregularity reveals itself. Weston, et al.Without early and frequent observations in the radio such a bump could be missed en-tirely, resulting in an incorrectly modeled light curve, which in turn a ff ects distance andmass estimates. In fact, in many historical novae observed in radio prior to this the timeand frequency coverage was insu ffi cient to catch any hypothetical bump during earlytimes, though this is not the case in particularly well sampled novae such as V1974 Cygor FH Ser 1970 (Hjellming 1996). It is possible that these early bumps are much morecommon than we realize. Acknowledgments.
We are grateful to H. Uthas and J. Patterson for their adviceand expertise. We thank NRAO for its generous allocation of time which made thiswork possible. The National Radio Astronomy Observatory is a facility of the NationalScience Foundation operated under cooperative agreement by Associated Universities,Inc. J. Weston and J.L. Sokoloski acknowledge support from NSF award AST-1211778.