Surprising Variability of the Planetary Nebula IC 4997 = QV Sge
V.P. Arkhipova, M.A. Burlak, N.P. Ikonnikova, G.V. Komissarova, V.F. Esipov, V.I. Shenavrin
aa r X i v : . [ a s t r o - ph . S R ] F e b Surprising Variability of the Planetary NebulaIC 4997=QV Sge
V.P. Arkhipova, M.A. Burlak ∗ , N.P. Ikonnikova, G.V. Komissarova,V.F. Esipov, V.I. Shenavrin Sternberg Astronomical Institute,Moscow State University (SAI MSU), Universitetskii pr. 13, Moscow, 119992 Russia
We present the results of a new epoch (2009-2019) of a long-term (50 years) photometric moni-toring of the variable planetary nebula IC 4997 (QV Sge). The integral (star + nebula)
U BV lightcurves display a continuing brightening of 0. m in V , a slight rise (<0. m ) in B , and constancyin U . The B − V color has got redder from 0. m in 2000 to 0. m in 2019, whereas the U − B colorhas not changed significantly at that time. We carried out near infrared (IR) JHKL photometryin 2019, and comparing it to the data obtained in 1999-2006 we found the source to be fainter by0. m in L and bluer by 0. m in the K − L color. The long-term brightness variations in the opticaland IR regions are shown to be due mostly to the changing input of emission lines to the integrallight.Low-dispersion spectroscopic observations carried out in 2010-2019 revealed a continuing de-crease in the [OIII] λ γ intensity ratio: it decreased by a factor of ∼ in 30 years andreached the level of 1960-1970. We discovered that the absolute intensities of [OIII] λ λ > from 1990 to 2019, whereas the [OIII] λ β absolute intensity in 1960-2019 was shown to be similar to that of [OIII] λ λ β , [OIII] λ , lines and their ratios we propose a possible scenario describing the change of physicalconditions ( N e , T e ) in IC 4997 in 1970-2019. The main features of spectral variability of IC 4997could be explained by a variation of electron temperature in the nebula caused by not so muchthe change in ionizing flux from the central star as the variable stellar wind and related processes.The photometric and spectral changes observed for IC 4997 in 1960-2019 may be interpreted as anobservable consequence of a single episode of enhanced mass loss from the variable central star. Keywords: planetary nebulae, photometric and spectral variability, IC 4997, QV Sge, gas shelldiagnostics, electron temperature, electron density. ∗ E-mail: [email protected] NTRODUCTION
IC 4997 attracted particular attention as the first confirmed variable source of radiation among theyoung planetary nebulae (PNs). Liller and Aller (1957) reported a noticeable change in the λ γ intensity ratio, having compared their own spectral measurements of 1956 with those ofMenzel et al. (1941) made in 1938. Vorontsov-Velyaminov (1960) confirmed the variability of the ratiobasing on the spectra obtained in Crimea in 1959-1960.IC 4997 received a great deal of interest in the 1960-1970s: the spectrophotometric observationswere carried out by O’Dell (1963), Aller and Kaler (1964), Aller and Liller (1966), - and later byFerland (1979), Feibelman et al. (1979), Purgathofer (1981). The optical spectrum of IC 4997 wasthe most thoroughly investigated in Hyung et al. (1994), Hyung and Aller (1993), where the relativeintensities of more than 500 emission lines in the λλ − region were measured in the echellespectra obtained in 1990 and 1991. The authors also noted the change in relative fluxes for quite anumber of other emission lines besides [OIII] λ γ over a year of high-dispersion observations.A large infrared (IR) excess in the 11 µ m range was first detected for IC 4997 by Gillett et al.(1971). Natta and Panagia (1981), Pottasch et al. (1984), Lenzuni et al. (1989) thoroughly studiedthe IR spectrum of the nebula and its dust envelope. Basing on IR photometry carried out in 1999-2006 Taranova and Shenavrin (2007) found the variability of IC 4997 with a peak-to-peak amplitudeof 0. m m
25 in H and 0. m
05 in J .Radio observations carried out by Miranda et al. (1996), Miranda and Torrelles (1998) madeit possible to construct for IC 4997 a map of 3.6 cm and 2 cm continuum radiation with angularresolution better than . ′′ . The authors described new details in the structure of its outer and innershells and confirmed the double-shell model for the nebula suggested previously by Hyung et al. (1994).Investigating the variability of IC 4997 radio flux Miranda and Torrelles (1998) found short timescale( < year) morphological changes in the inner shell structure in the vicinity of the central star < . ′′ and explained them by the collimated stellar wind impinging on the outer shell of the nebula. Accordingto the archival data, the 5 GHz radio flux from the optically thin part of the nebula has decreasedfrom 100 mJy around 1989 (Acker et al., 1992) to 45 mJy in 1996 (G´omez et al., 2002). Subsequentobservations (Casassus et al., 2007, Pazderska et al., 2009) showed an increase in radio flux from 80to 108 mJy at a frequency of 30 GHz but the data was not enough to analyze the radio spectrum indetail.The central star of IC 4997, HD 193538=QV Sge, was classified in a number of papers as a weakemission line star – wels . It’s hard to divide the observed integral optical spectrum into the componentsbelonging to the nebula and the central star, but the CIV λ λ λ λ < ′′ from the central star and consists of two pairs of diametrically opposed lobes whose axes areperpendicular one to another.Regular U BV photometric and spectral observations of IC 4997 at the Crimean astronomicalstation (CAS) of SAI MSU were initiated by E.B. Kostyakova and started in 1968. The angular2isible size of the nebula is about ′′ . Both photometric and spectral observations realized as partof the program include the whole nebula together with the central star HD 193538. The results ofphotometric and spectral monitoring in different epochs were published in a quite number of papers:Vorontsov-Velyaminov et al. (1970), Kostyakova (1971), Kostyakova et al. (1973), Arkhipova et al.(1994), Kostyakova (1990, 1999), Kostyakova and Arkhipova (2009), Burlak and Esipov (2010). Wecontinue observing IC 4997 on the CAS telescopes and in this paper we analyze the measurementsobtained in 2009-2019 together with previous data. OBSERVATIONS
U BV photometry
Although there is a good deal of single integral brightness estimates for IC 4997 published, it wouldhardly make sense to compare them as there are many strong emissions and the photometric systemsare slightly different, which leads to a significant scatter in data. We succeeded to maintain a long-termmonitoring at the same telescope with permanent equipment.Our
U BV observations of IC 4997 have been carried out with the photon counting photometerconstructed by Lyutyi (Lyutyi, 1971) mounted at the Cassegrain focus of the Zeiss-600 telescope ofCAS since 1971. Previous results obtained mainly by Kostyakova E.B. can be found in Kostyakovaet al. (1973), Kostyakova (1991), Arkhipova et al. (1994), Kostyakova and Arkhipova (2009). In thispaper we present recent
U BV photometry obtained in 2009-2019. As usual, HD 355464 was used as areference star, and its magnitudes ( V = 9 . m , B = 10 . m , U = 10 . m ) were acquired via referencingto photometric standards in NGC 6633 and NGC 7063 (Hiltner et al., 1958). The standard errors inthe photometric magnitudes are σ V = 0 m . , σ B = 0 m . , σ U = 0 m . . The measurements werecarried out with an aperture of ′′ (sometimes ′′ ), so we measured the integral brightness of theobject – the PN and the central star.Here we present the magnitudes in the instrumental system which is very close to the standard oneof Johnson. Standard data reduction procedures were performed, and besides, the magnitudes werereduced to the common under-dome temperature ( t = +10 ◦ C) and a correction was introduced aftera slight change of instrumental system in 1989 when the photomultiplier was substituted.The magnitude – temperature relations for B and V can be expressed by the equations: ∆ V = 0 . − . t + 1 . × − t , ∆ B = − . − . t, (1)where t is the under-dome temperature in Celsius degree.To reduce new data to the previous photometric system we introduce the following corrections: ∆ V = 0 . m , ∆ B = − . m , ∆ U = − . m .In Table 1 we present U BV photometry for IC 4997 obtained in 2009-2019. To analyze its long-termvariability we calculated the annual average
U BV magnitudes and present them in Table 2 togetherwith standard deviation ( σ ) and number of observations ( N ). IR photometry
We carried out IR photometric monitoring at the 1.25-m telescope of CAS in 1999-2006 and resumedin 2019. The photometer with a liquid nitrogen cooled photovoltaic indium antimonide (InSn) detectorinstalled at the Cassegrain focus was used. The output aperture was ≈ ′′ , the spatial separation ofthe beams during modulation was ≈ ′′ in the east–west direction. The star BS 7635 from Johnsonet al. (1966) was used as a photometric standard. The results of IR monitoring in 1999-2006 werepublished in Taranova and Shenavrin (2007). New J HKL magnitudes obtained in 2019 are listed inTable 3. 3 pectral observations
Optical spectroscopy of IC 4997 was carried out at the 1.25-m telescope of CAS in 2010-2019. We used alow-dispersion spectrograph with a 600 lines mm − grating. The detector was a ST-402 CCD 765 × ′′ . The spectral resolution obtained wasnear 7.4 ˚A as measured from the fullwidth half maximum (FWHM). The spectrograph configurationwas different in August and October, 2019: the spectra were obtained with another camera objectivelens and a FLI PL-4022 CCD 2048 × × ρ Aql (Glushneva et al., 1998; Pickles, 1998). The standard stars were observed before or after IC 4997at close airmasses. We show the log of observations in Table 4.Due to the spectrograph design it’s possible to obtain simultaneously a part of spectrum of ∼ ˚Aor ∼ ˚A in the previous or new configuration respectively. To cover the whole observable wavelengthrange ( ∼ ∼ I ( H β ) = 100 , and the observed flux F ( H β ) inabsolute units. For the lines brighter than 1% of the H β line the estimated error is about 10%, andabout 25% for the weaker ones.In order to examine the physical conditions in the nebula, we had to correct the relative intensitiesfor the interstellar reddening. Burlak and Esipov (2010) reviewed the previous attempts to determinethe reddening coefficient for IC 4997, and using their spectral data obtained in 2003–2009 they derived c ( H β ) = 0 . taking into account self-absorption in the hydrogen lines. Our new spectral data agreewell enough with this value of c ( H β ) . It’s worth mentioning that the Balmer decrement was indicativeof self-absorption in 2010–2019, too, but the effect seems to have weakened by 2019. PHOTOMETRIC VARIABILITY
The integral brightness of IC 4997 varies with a typical seasonal range of less than 0. m U BV brightness, and U − B , B − V color indices in 1968–2019 basing on the results of Kostyakovaet al. (1973), Kostyakova (1991, 1999), Arkhipova et al. (1994), Kostyakova and Arkhipova (2009),unpublished measurements of E.B. Kostyakova and new data.The variation of the annual average magnitude appears to have an amplitude of . m in V , . m in B , and about . m in U . Note that in 50 years of our monitoring the integral U and B magnitudesdescribed a gradual asymmetric curve and returned to the initial state, while the V brightness continuedto increase after 2010, and the object is . m brighter at present than it was when we started observingit. The B − V and U − B color indices varied less gradually and appeared bluer and with larger scatterwhen the object was fainter that could be ascribed to the effect of variable emission lines.To estimate the input from emission into the U BV brightness we isolated the emission nebularcomponent of IC 4997 from the average
U BV magnitudes for the year 1990 using the most completeand reliable spectral data for this year from Hyung et al. (1994). We took into account the lines notfainter than 0.05 on the scale I (H β ) = 100 . Unfortunately, the given paper contains only relative lineintensities for IC 4997 in August, 1990. The data were flux-calibrated using the absolute measurement4f H β intensity in 1990 made by Kostyakova and Arkhipova (2009): F (H β ) = 2 . × − erg cm − s − .We calculated the averaged values of brightness and color over 12 estimates obtained in May-October,1990: V = 11 . m , B = 11 . m , U = 10 . m , B − V = +0 . , U − B = − . . The input fromemission lines was estimated using the passbands of our U BV instrumental system.Note that photometric and spectrophotometric observations were carried out at different nights,the averaged date of all measurements falls on the end of July, 1990.So, we estimated the emission lines input into the
U BV brightness, and calculated the integralcontinuum flux and magnitude for the object in 1990. F V (lines) = 4 . × − erg cm − s − , F V (cont) = 8 . × − erg cm − s − , V (cont) = 11 . m ; F B (lines) = 11 . × − erg cm − s − , F B (cont) = 4 . × − erg cm − s − , B (cont) = 12 . m ; F U (lines) = 2 . × − erg cm − s − , F U (cont) = 17 . × − erg cm − s − , U (cont) = 11 . m .The emission lines input into the U BV integral brightness of the variable PN IC 4997 in 1990 wasas large as 30 % in V , 70 % in B , and more than 13 % in U (since no emission lines were measuredbeyond the Balmer discontinuity in 1990).We’d like to highlight the input of the [OIII] nebular lines λ λ β into the B and V integral light. In Figure 2, we present the relationship between the summary flux from H β , [OIII] λ λ V and B brightness in 1972-2019. Itis clearly seen that the summary flux from these three lines lg( F ( Hβ ) + F ( λ F ( λ andthe V and B brightness are closely related: the correlation coefficients are 0.90 and 0.81 respectively.Therefore, the B and V photometric variability may be explained by the changing emission spectrumof the nebula. Similar analysis for the U band is complicated by the absence of data on the changesof nebular spectrum in the wavelength range bluer than λ B − V =+1 . m , U − B < ˘1 . m . Assuming E ( B − V ) = 0 . m for IC 4997, we got the reddening correctedcolor indices for the summary continuum: ( B − V ) = +1 . m , ( U − B ) < ˘1 . m .The emission lines input having been removed, the position of IC 4997 in the color-color diagramappears to be similar to that of a hot star with a temperature more than 35000 K and a rather opticallythick gas continuum. We estimate the position uncertainty associated with the mismatching of datesof photometric and spectral observations to be equal to ∼ . m .Note that IC 4997 was undoubtedly classified as a bipolar PN (Sahai et al., 2011). The bipolarPNs which have an hourglass-like morphology (i.e. a cylinder with a ’pinched-in’ shape in the regionaround the center) are now widely considered to have a binary central star. And in this regard, we wereinterested to find some traces of the possible central star binarity in the summary optical continuum.The color indices ( B − V ) (cont) = 1 . and ( U − B ) (cont) < − . resulted from subtracting theemission lines input and reddenning correction demonstrate a slight red excess in the ( B − V ) colorwhich is not reliable due to low accuracy.IC 4997 displays a variation in IR brightness, too. Taranova and Shenavrin (2007) basing on the J HKL photometry made in 1999-2006 showed that the amplitude of variation in
J HKL was nearly . m − . m and that in the J − H , H − K и K − L colors was up to ∼ . m within a characteristictime of 260–280 days. The authors suggested that the detected variation in H was associated withthe changing input of the hydrogen Brackett lines. In 2019 we carried out the observations of IC 4997with the same equipment as had been used previously. In Figure 3 we present the near IR light andcolor curves for IC 4997 compiled from the data published by Taranova and Shenavrin (2007) and thenew measurements obtained in 2019. As is seen from the figure, the H and K brightness in 2019 isat the level of the minimum values for the 1999-2006 interval, whereas the average J brightness hasdecreased by . m . The most prominent change is seen for the L band: the brightness has decreasedby . m . The color indices have barely changed on average since 1999 and are about J − H = − . m , J − K = 0 . m , H − K = 0 . m , only the K − L color has decreased from . m to . m .5hitelock (1985) reviewed the principal sources of near IR emission for PNs: they are the free-freeand free-bound radiation of hydrogen and helium plasma and the thermal emission from dust with T d >1000 K (if there is dust in the nebula). There is also a small input of the central star and asignificant one from the emission lines, the strongest ones being: J band: P β , P γ , He II λ . µ m, He I λ . µ m; H band: Brackett series from Br 10 λ . µ m to nearly the series limit at λ . µ m; K band: Br γ and He I λ . µ m.Whitelock (1985) proposed a classification scheme for PNs based on the major source of 1-2 µ memission and assigned IC 4997 to the N (Nebula) group due to the nebula radiation dominating thenear IR domain. Note that IC 4997 is a PN of low excitation, so the He II λ . µ m is not presentin the spectrum. Thus, trying to explain the J HK variability of IC 4997 we can ignore the impact ofthe dust component and consider the change in intensities of H I and, to a lesser degree, of He I linesresponsible for variation.Ohsawa et al. (2016) presented near IR ( . − . µ m) spectra for 72 PNs including IC 4997 obtainedwith IRC/AKARI. Due to this work, we have an idea of the most important emission lines of IC 4997falling into the L passband: they are Br α λ . µ m and Br β λ . µ m. Therefore, the L brightnessvariability of IC 4997 may relate to the change in the intensity of these lines. SPECTRAL VARIABILITY
The change in the intensity ratio of the lines [OIII] λ γ was the first evidence for the spectralvariability of IC 4997 due to the fact that it could be easily measured with confidence. In Figure 4we reconstruct the long-term (several decades) evolution of the F ( λ /F ( H γ ) ratio basing on thedata from this work and the previous estimates made by different researchers since 1938: Aller (1941),Struve and Swings (1941), Page (1942), White (1952), Liller and Aller (1963), Aller and Liller (1966),Vorontsov-Velyaminov et al. (1965), Aller and Kaler (1964), O’Dell (1963), Ahern (1978), Feibelmanet al. (1979), Purgathofer (1978), Purgathofer and Stoll (1981), Acker et al. (1989), Hyung et al.(1994), Kostyakova and Arkhipova (2009). Besides, we have estimated line intensities in the spectrumof IC 4997 published by Hajduk et al. (2015) and in another one found in the HASH PN database(Parker et al., 2016). Over the entire period of observations the ratio has described a wave with apeak-to-peak amplitude of 0.45 in the logarithmic scale and a characteristic time of 50-60 years. The λ / H γ ratio continued to decrease in 2010-2019 following the trend set around 1990, and by 2019the ratio has reached the value observed in 1960–1970.In addition to the variation of F ( λ /F ( H γ ) , the nebular [OIII] lines intensities relative to H β were found to vary, too. Since 1960–1970 their intensities at first decreased until 1985, then they grewto 2005, and since then they stay at the approximately constant level slightly higher than that of theearly sixties. Note that the relative intensities of the nebular and auroral lines have changed differentlywhich is well illustrated by Figure 5. There is also a hint of variability in He I lines but the dataare less reliable and available for a smaller time interval. In Figure 5 we present the evolution of theHe I λ intensity relative to H β : in general, the change was inverse to that of the nebular lines(Vorontsov-Velyaminov et al., 1965; O’Dell, 1963; Aller and Walker, 1970; Ahern, 1978; Acker et al.,1989; Hyung et al., 1994; Hajduk et al., 2015; Parker et al., 2016). The He I λ line varies in asimilar way to λ , at least since 1986 – we have not managed to find earlier measurements for the λ line.Figure 6 shows the observed absolute intensities for the [OIII] λ , λ lines measured bydifferent researchers over the last 60 years. Although there is a big scatter in estimates for these lines,the trend is well defined. Since 1960 the absolute intensity of the auroral line grew at an approximatelyconstant rate, reached maximum level about 2000–2005 having increased more than twice, and then6tarted to decrease faster than it had been increasing. Since 1960 the absolute intensities of nebularlines stayed quite constant with a possible slight tendency to decrease, but in the mid-1970s droppedsharply and reached minimum level around 1985-1990, then increased at a lower rate to the valuesobserved about 1960 or little higher, thereby having reproduced the optical light curve of IC 4997.It would be interesting to ascertain whether the absolute intensity of H β varies, but the absolutespectrophotometry is very difficult to perform, and there are inconsistencies between observers whichproduce large scatter in data constituting definite obstruction to the determination of the shape andcharacteristic parameters of the H β intensity variation. Nevertheless, one can see a period of higherabsolute intensity before 1980, then it faded by 30% and stayed more or less constant until 2000–2005,and after that started to grow.As regards the variability of other lines, it’s worth mentioning that mostly short wave region ofoptical spectrum was investigated in earlier epochs, while the long wave one – later. So, just severallines were measured over a long-term period, and only for a few the estimates are reliable. Figure 7shows the evolution of relative intensities for the [SIII] λ and [NII] λ auroral lines, and the[ArIII] λ nebular line. In 1986–2019 [SIII] λ reproduced the behavior of the [OIII] auroralline, its relative intensity decreased, whereas the variation of [NII] λ was inverse to that of [OIII] λ . We have not managed to ascertain the change in [ArIII] λ relative intensity over the giventime interval. DIAGNOSTICS AND PHYSICAL PARAMETERS OF THE IC 4997NEBULA "One has to be an optimist to attempt to constructa model for IC 4997..."(Huyng, Aller, Feibelman, 1994)After the variability of the F ( λ /F (H γ ) and F ( λ /F ( λ λ ratios was discov-ered and till about 1970, some suggestions were proposed to explain these spectral changes. Gurzadian(1958), Vorontsov-Velyaminov (1960) and Khromov (1962) considered the fall in the central star tem-perature to be the cause of the decrease in F ( λ /F ( H γ ) , though Gurzadian believed this change intemperature to be of an evolutionary character, Vorontsov-Velyaminov – to have a long periodic nature,Khromov – to be oscillatory fluctuations. According to Aller and Liller (1966), the spectral changesresulted from the expansion of the nebula with a corresponding decrease both in electron density andelectron temperature. In addition to long-term changes, fast oscillations of the F ( λ /F ( H γ ) ratiowithin one year were detected, and Ferland (1979) supposed that they were due to the variation of elec-tron temperature attributed to small changes in the number of ionizing photons emitted by the centralstar: since the electron density is high in the inner parts of the nebula ( N e ∼ см − ) where the[OIII] λ N e , T e values in the diagnostic diagram for IC 4997. To understand the observational data and toestimate the abundances it is necessary to adopt for the nebula at least a two-component structureconsisting of a more dense inner zone enclosed in a shell of lower density.7ur aim was to determine the physical conditions in the nebula, so we measured several diagnosticratios. The most actively used intensity ratio has always been the relation involving the auroral [OIII] λ and one or both of the nebular [OIII] λ , transitions. We did not always manage tomeasure the λ line intensity, therefore Figure 8 shows the change of the R = F ( λ /F ( λ value over a time period of about 80 years. In 2010–2019 the ratio decreased thus keeping the tendencystarted about 1990, and by 2019 it returned to the value observed in 1960-1970. Due to the availabilityof absolute intensities, we can schematically divide the evolution of R in 1970-2019 into three intervals.During the first phase ( ∼ R was growing due to the strengthening of the auroral transitionand weakening of the nebular one; at the second stage ( ∼ R was decreasing with thesimultaneous strengthening of both lines; during the third phase ( ∼ R was decreasing dueto the strengthening of the nebular transition and weakening of the auroral one. The big grey circles inFigures 6 и 8 indicate the averaged values of corresponding quantities at the points which delimit thehighlighted intervals: 1970, 1990, 2003 (the year when we started our spectroscopic monitoring with aCCD), 2019.Some other line ratios seem to be variable too, though their variation can be traced over a shorterperiod of time and is not evident (see Figure 9). So, the [ArIII] F ( λ /F ( λ ratio was observedto decrease from ∼ . in the early 1990s (Hyung et al.,1994) to ≤ . (this work) in 2019. The[ArIII] λ line, however, is very weak and forms a blend with two other lines of comparable intensity,[NI] λ и λ , and the low dispersion makes its measuring uncertain. Similar behavior is seenfor the [SIII] lines, though a large wavelength separation between them ( λ and λ ) leads toa significant calibration uncertainty. On the contrary, the [NII] F ( λ /F ( λ ratio does notshow any significant change. Besides, the scatter of estimates is big, possibly due to the difficulty inseparating the λ line from the H α wing.The diagnostic relations sensitive to electron density, [SII] F ( λ /F ( λ and [ClIII] F ( λ /F ( λ , stayed constant within the limits of accuracy over the whole period of our ob-servations since 2003. The F ( λ /F ( λ ratio is close to its critical value, so it indicates onlythe lower limit of log N e ∼ which corresponds to the outer envelope of lower density where thelow excitation species emit. The ratio of [ClIII] lines is also close to its critical value. This ratio isattributed to the zones of higher excitation and suggests a somewhat higher estimate for the densityof lg N e ∼ . . But the second ionization potential for chlorine is lower than that for oxygen, there-fore the zone where the λ , lines originate may match only partially with the [OIII] emittingregion. The constancy of the ratios mentioned above implies that even if the electron density in thelow excitation zones varies it does not fall much lower than the critical value ( lg N crite ∼ − . forsulphur и ∼ − . for chlorine).Let’s now consider the displacement of diagnostic curves representing the ions N + , S , O andAr in the N e − T e diagram. These ions have a similar structure of energetic levels but differ in theionization potential and critical density, and must emit in different regions of the nebula. Figure 10contains the N e − T e diagrams constructed by means of the PyNeb package (Luridiana et al., 2015)for the moments which limit the intervals of time highlighted according to the change of the [OIII]absolute intensities. Besides the fact that one pair of N e , T e values can not explain the observed data,the total appearance of the diagrams suggests a high value of the electron density ( lg N e ≥ ), at leastfor the strata where the [OIII] lines arise. The ratios used for the diagrams were corrected for theinterstellar reddening with c ( H β ) = 0 . . The diagram for the year 1990 was drawn basing on thedata from Hyung et al. (1994), and they derived c ( H β ) = 0 . . If we had corrected the ratios using thisvalue of c ( H β ) , the diagnostic curves would have been located in the area of larger values of N e , T e .Since 1990 the curves associated with O , S и Ar evolved in similar manner, they shifted tothe smaller values of N e , T e , while the location of the N + curve almost did not change. Probably, thevariation of N e , T e does not affect the [NII] emitting zone.8s we are not able to derive the absolute N e , T e values for the given epochs, let’s try to determinehow they change. The initial data for our further calculations are the changes of absolute intensitiesfor H β , [OIII] λ , , and also the changes of their relations corrected for reddening. We assumethat the zones where all these transitions arise are characterized by the same N e , T e values and thenumber of emitting hydrogen atoms is constant. Some definite value of temperature is considered asan initial one (in 1970). First, we find N e for this initial moment using R , and the relative abundanceof O using relative intensities. Then, as we know the relative change of F ( H β ) between the twoepochs and the R value at the second moment, we can estimate N e , T e at the second moment andalso the abundance of O using relative intensities. The same procedure is executed as we passfrom the second moment to the third and from the third to the fourth. Table 6 lists the results ofthe simulations performed by means of the package PyNeb for several initial values of the electrontemperature: T e = 8000 , , K. For all initial temperatures, N e varies only slightly, whereas T e first grows and then drops by several thousand degrees.A significant growth of temperature is necessary to interpret the decrease in the absolute intensityof H β since 1970. It сould have been explained by the decrease in N e , but it is in contrary to the growthof R . On the other hand, the intensity of collision excited lines must increase with temperature, butwe observe the nebular lines decreasing since 1970, and the auroral line increasing is less than expectedfor such change in temperature. To explain the observed data it’s necessary to reduce significantly thenumber of emitting O ions (by a factor of 10), which may be due to the subsequent ionization ofoxygen or to the diminution of the emitting zone.And the question arises: what did make T e to increase by several thousand degrees? Accordingto Ferland (1979), if the change in the electron temperature of the nebula is caused by a changein the flux of ionizing photons Q from the central star, we can relate them through an equation: ∆ T e ≃ × ∆ Q/Q . So, the 300 K temperature change proposed by Ferland to explain fast variabilityof the F ( λ /F ( H γ ) ratio requires an 8% change in the ionizing flux. But our simulations imply atemperature increase of several thousand degrees that would mean ionizing flux increasing by severaltimes. We find it hard to imagine such a process. A rise in T e must have been caused not so muchby the growth in the effective temperature of the star because the level of excitation of the spectrumof IC 4997 did not change over the observed period, but by some other processes, for instance, theinteraction of stellar winds.Table 6 also lists the relative intensities for the HeI λ line calculated using the obtained valuesof N e , T e and He + / H + =0.1. One can see that in the case of the initial T e = 8000 K the intensity changesonly slightly, whereas for T e ≥ K the simulated variation of intensity matches qualitatively withthe observed one, though the amplitude is smaller.The results presented in Table 6 can be considered as a kind of estimation. A larger value ofinterstellar reddening ( c ( H β ) > . ) will require a larger increase in electron temperature to explainthe observed spectral changes. If we take into account the fact that we measure the integral flux fromthe nebula, whereas the fluctuation of N e , T e may occur only in some strata, then the amplitude oftemperature change will be larger. DISCUSSION AND CONCLUSIONS
We have presented optical and near IR photometric and low-resolution spectroscopic data for IC 4997obtained in 2009–2019. New results are investigated together with the previously published data.Basing on the observational data obtained with our invariable
U BV photometric system we haveplotted an annual average light curve of IC 4997 from 1970 till 2019. We have found a long-termmarked dip in
U BV flux with a peak-to-peak range of . m in V . The source started fading after 1970,reached the minimum brightness near 1985, then recovered to the initial level in B and U , whereas9he V brightness continued growing up to 2019. In 2019 we obtained new near IR J HKL photometricdata. The nebula was found to be fainter in 2019 if compared to the epoch of 1999–2006 with theeffect most prominent in L . We have shown that the long-term optical and IR brightness variation isrelated mostly to the changing input from the nebular emission lines.After the emission contribution was subtracted from the B and V brightness, the B − V color ofthe integral continuum corrected for reddening appeared still too red for a sum of the stellar and gascontinua, which may be indicative of the presence of one more source of continuum radiation in thiswavelength range (a satellite of the central star?), although the evidence is very uncertain. It’s worthmentioning that the central star of IC 4997 is not only suspected to be binary but it was also includedin the list of PN possibly shaped by a triple stellar system (= maybe triple class) with a probability of0.33.Basing on new and previously published data we have traced the evolution of relative and absoluteintensities and their relations for some emission lines originated in the nebula over a period 1970–2019.In 2010–2018 the λ / H γ ratio decreased as it had been doing since about 1990 and by 2019 returnedto the value observed in 1960–1970. Over the whole period of observations the ratio drew a wave witha peak-to-peak amplitude of 0.45 in logarithmic scale and a characteristic time of 50-60 years.With the use of archival and new data we have traced the evolution of absolute intensity for H β : onecan note a period of higher values before 1980, a fading by a factor of 1.5 and subsequent maintenanceat the same level with a slight tendency to growth till 2000–2005, then a more pronounced increase.The variation of the absolute intensity of the nebular [OIII] λ line in 1960-2019 was roughly similarto that of H β but had a larger amplitude ( F max /F min ≥ ). Having recovered to the value observedbefore fading, the absolute intensity of [OIII] λ did not change significantly in 2010-2019. Theabsolute intensity of the auroral [OIII] λ line grew since 1960 at a nearly constant rate, reachedmaximum value about 2000-2005 having increased more than twice, then started to fade and is stillfading.We have also reconstructed the variations of intensity for some other transitions. In particular, wehave traced the variability of HeI λ ∼
2, then weakening and return to the former level.We have estimated the lower limit for the electron density in the outer shell of IC 4997 using thediagnostic ratios of [SII] и [ClIII]: lg N e ∼ и ∼ . respectively. The ratios stayed constant withinerrors and were close to critical values in 2010-2019.The location of diagnostic curves of N + , S , O and Ar on the N e , T e plane indicates that,first, there are zones of different electron temperature and density in the nebula, and, second, theparameters vary with time. A rise in one or both of these characteristics was observed at least in theinner part of the nebula in 1970-1990 with subsequent decrease which lasts until now. Using the dataon absolute intensity variations for H β , [OIII] λ , and assuming that N e , T e are the same forthe strata where these lines arise, we have estimated the scale of the variation. We think that thevariation of T e in the inner part of the nebula is responsible for the spectral variability of IC 4997, andthe change in T e is due not so much to the change in ionizing flux from the central star, but to thevariable stellar wind and the related processes. In general, the spectral changes observed in 1960-2019may be interpreted as the observable consequence of a single episode of enhanced mass loss from thevariable central star. There are still the questions of what has triggered the increase in mass loss rate,whether this episode was a unique one or could occur again in one form or another. ACKNOWLEDGEMENT
The authors dedicate this paper to the memory of Doctor of Science Elena B. Kostyakova, SeniorResearcher Fellow at the Sternberg Astronomical Institute (1924–2013).10his research has made use of the ADS, SIMBAD and VIZIER databases.
REFERENCES
1. A. Acker, J. Koppen, B. Stenholm, and G. Jasniewicz, Astron. Astrophys. Suppl. Ser. , 201(1989)2. A. Acker, J. Marcout, F. Ochsenbein, et al., The Strasbourg-ESO Catalogue of Galactic PlanetaryNebulae, Parts I, II, (1992)3. F.A. Ahern, Astrophys. J. , 901 (1978)4. L.H. Aller, Astrophys. J. , 236 (1941)5. L.H. Aller and J.B. Kaler, Astrophys. J. , 621 (1964)6. L.H. Aller and W. Liller, MNRAS , 337 (1966)7. L.H. Aller and M.F. Walker, Astrophys. J. , 917 (1970).8. V.P. Arkhipova, E.B. Kostyakova, R.I. Noskova, Astron. Lett. , 99 (1994)9. E. Bear, N. Soker, MNRAS , L10 (2017)10. M.A. Burlak, V.F. Esipov, Astron. Lett., , 752 (2010)11. S. Casassus, L.-˚A. Nyman, C. Dickinson, and T. J. Pearson, MNRAS , 1607 (2007)12. W.A. Feibelman, R.W. Hobbs, C.W. Mc Cracken, L.W. Brown, Astrophys. J. , 111 (1979)13. G.J. Ferland, MNRAS , 669 (1979)14. F.C. Gillett, R.F. Knacke, W.A. Stein, Astrophys. J. , L57-59, 197115. I.N. Glushneva, V.T. Doroshenko, T.S. Fetisova, T.S. Khruzina, E.A. Kolotilov, L.V. Mossakovskaya,S.L. Ovchinnikov, and I.B. Voloshina, VizieR Online Data Catalog III/208 (1998).16. Y. G´omez, L.F. Miranda, J.M. Torrelles, L.F. Rodr´ıguez, J.A. L´opez, MNRAS , 1139 (2002)17. G.A. Gurzadian G.A., Soviet Astronomy , 482 (1958)18. M.Hajduk, P.A.M. van Hoof, and A.A. Zijlstra, Astron. and Astrophys.J. , A65 (2015)19. W.A. Hiltner, B. Iriarte, M.L. Johnson, Astrophys. J. , 539 (1958)20. S. Hyung, L.H. Aller, Proc. Nat. Acad. Sci. USA , 413 (1993)21. S. Hyung, L.H. Aller, W.A. Feibelman, Astrophys. J. Suppl. Ser. , 465 (1994)22. H.L. Johnson, R.I. Mitchel, B. Iriarte, W.Z. Wisniewski, Comm. Lunar and Planet. Lab. , 99(1966).23. G.S. Khromov, Sov. Astron. , 619 (1962)24. E.B. Kostyakova, Sov. Astron. , 794 (1971)115. E.B. Kostyakova, Astron. Lett. , 465 (1990)26. E.B. Kostyakova, Trudy GAISh, , 143, (1991)27. E.B. Kostyakova, Astron. Lett. , 389 (1999)28. E.B. Kostyakova, V.P. Arkhipova, Astron. Rep. , 1155 (2009)29. Е.B. Kostyakova, V.P. Arkhipova, M.V. Savel’eva, Mem. Soc. Roy. Sci. Liege, 6 Ser., , 473(1973)30. P. Lenzuni, A. Natta, N. Panagia, Astrophys. J. , 306 (1989)31. W. Liller and L.H. Aller, Sky and Tel. , 222 (1957)32. W. Liller and L.H. Aller, Proc. Nat. Acad. Sci. , 695 (1963)33. V. Luridiana, C. Morisset, and R.A. Shaw, Astron. Astrophys. , 42 (2015)34. V. M. Lyutyi, Soobshch. GAISh , 30 (1971)35. W.L.F. Marcolino, and F.X. de Ara´ujo, Astron. J. , 887 (2003)36. W.L.F. Marcolino, F.X. de Ara´ujo, H.B.M. Junior, and E.S. Duarte, Astron. J., , 1380 (2007)37. D.H. Menzel, L.H. Aller, M.H. Hebb, Astrophys. J. , 230 (1941)38. L.F. Miranda, J.M. Torrelles, C. Eiroa, Astrophys. J. , L111 (1996)39. L.F. Miranda, J.M. Torrelles, Astrophys. J. , 274 (1998)40. A. Natta, N. Panagia, Astrophys. J. , 185 (1981)41. C.R. O’Dell, Astrophys. J. , 1018 (1963)42. R. Ohsawa, T. Onaka, I. Sakon, M. Matsuura, H. Kaneda, Astron. J. , 93 (2016)43. T.L. Page, Astrophys. J. , 78 (1942)44. Q.A. Parker, I.S. Bojiˇci´c, D.J. Frew, Journal of Physics: Conference Series , article id. 032008(2016)45. B.M. Pazderska, M.P. Gawro´nski, R. Feiler, M. Birkinshaw, I.W.A. Browne, R. Davis, A.J. Kus,K. Lancaster, S.R. Lowe, E. Pazderski, M. Peel and P.N. Wilkinson, Astron. Astrophys. ,463 (2009)46. A.J. Pickles, Publ. Astron. Soc. Pacific , 863 (1998).47. S.R. Pottasch, B. Baud, D. Beinteme, J. Emerson et al., Astron. Astrophys. , 10 (1984)48. A.T. Purgathofer, Circ. IAU (1978)49. A.T. Purgathofer and M. Stoll, Astron. Astrophys. , 218 (1981)50. R. Sahai, M.R. Morris, and G. Villar, Astron. J. , 134 (2011)51. O. Struve and P. Swings, Astrophys. J. , 356 (1941)12
970 1980 1990 2000 2010 202011.611.411.211.010.810.610.4 1970 1980 1990 2000 2010 20200.80.60.4-0.4-0.6
UBV Years m agn i t ude U-BB-V Years m agn i t ude Figure 1: The annual average
U BV light and U − B, B − V color curves for IC 4997 in 1968–2019.52. O.G. Taranova, V.I. Shenavrin, Astron. Lett. , 584 (2007)53. B.A. Vorontsov-Velyaminov, Sov. Astron. , 929 (1960)54. B.A. Vorontsov-Velyaminov, E.B. Kostyakova, O.D. Dokuchaeva, V.P. Arkhipova, Sov. Astron. , 364 (1965)55. B.A. Vorontsov-Velyaminov, E.B. Kostyakova, O.D. Dokuchaeva, V.P. Arkhipova, Trudy GAISh , 57 (1970)56. M.L. White, Astrophys. J. , 71 (1952)57. P.A. Whitelock, MNRAS , 59 (1985)Table 1: U BV photometry for IC 4997 in 2009-2019.JD
V B U
V B U
V B U
V B U s) V log(F( 5007+ 4959+H ), erg/cm s) B Figure 2: The relationship between lg( F ( Hβ ) + F ( λ F ( λ ) and the V and B brightnessin 1972–2019. The asterisks represent the data from Kostyakova and Arkhipova (2009), the points –this work. Years
JD 2400000+...
LKHJ m agn i t ude Years
JD 2400000+...
K-LH-KJ-H m agn i t ude Figure 3: The change in near IR brightness and color for IC 4997 in 1999–2019.17
930 1940 1950 1960 1970 1980 1990 2000 2010 2020-0.2-0.10.00.10.20.30.4 l g F ( ) / F ( H ) Years
Figure 4: The evolution of the observed intensity ratio F ( λ /F ( H γ ) based on data from differentstudies: asterisks – Kostyakova, Arkhipova (2009), filled circles – Burlak, Esipov (2010), open circles– this work, crosses – other sources, see the references in the text.Table 2: The annual average U BV magnitudes for IC 4997 in 2009-2019.Year
V σ V B σ B U σ U N .20.40.60.81.01.01.52.02.53.01950 1960 1970 1980 1990 2000 2010 20200.20.30.4 F ( ) / F ( H ) F ( ) / F ( H ) Years F ( ) / F ( H ) Figure 5: The evolution of the observed relative intensities of [OIII] λ , λ and HeI λ basedon data from various sources. The symbols are the same as for Figure 4.19 .01.52.0 1960 1970 1980 1990 2000 2010 20202342468 F ( ) Years F ( H ) F ( ) Figure 6: The evolution of the observed absolute intensities of [OIII] λ , λ and H β expressedin units of − erg cm − s − . The symbols are the same as for Figure 4. Big grey circles representaverage values at the points that delimit the time intervals distinguished by the peculiarities of [OIII] λ and λ variations. 20 l g F ( ) / F ( H ) l g F ( ) / F ( H ) l g F ( ) / F ( H ) Years
Figure 7: The evolution of the observed relative intensities of [ArIII] λ , [SIII] λ , [NII] λ based on data from various sources. The symbols are the same as for Figure 4.21
940 1960 1980 2000 20200.10.20.30.40.50.60.7 F ( ) / F ( ) Years
Figure 8: The evolution of the observed auroral to nebular line ratio F ( λ /F ( λ based ondata from various sources. The symbols are the same as for Figures 4 and 6.Table 3: J HKL photometry for IC 4997 in 2019.JD
J σ J H σ H K σ K L σ L .030.040.050.060.070.050.100.150.201985 1990 1995 2000 2005 2010 2015 20200.010.020.03 F ( ) / F ( ) F ( ) / F ( ) F ( ) / F ( ) Years
Figure 9: The evolution of the observed auroral to nebular line ratios [ArIII] F ( λ /F ( λ , [SIII] F ( λ /F ( λ , [NII] F ( λ /F ( λ based on data from various sources. The symbols arethe same as for Figure 4. 23 T e , K [ O III ] [ O III ] [ S III ] [ A r III ] [ N II ] e , cm −3 T e , K [ O III ] [ S III ] [ A r III ] [ N II ] e , cm −3 [ O III ] [ S III ] [ A r III ] [ N II ] Figure 10: Diagnostic diagrams for IC 4997 drawn with the use of averaged data from different sourcesfor the epochs of 1970, 1990, 2003 and 2019. All the intensity ratios were corrected for interstellarreddening with c ( H β ) = 0 . . Different lines indicate ions: solid for [SIII], dotted for [OIII], dashedfor [ArIII], and dotted-dashed for [NII]. 24able 4: Log of spectroscopic observations of IC 4997.Date JD Range, ˚A Standard04.08.2010 2455413 4000–7200 18 Vul31.07.2011 2455774 4000–7400 HD 19677526.08.2011 2455800 4000–9100 40 Cyg21.07.2012 2456130 4000-7400 18 Vul23.08.2012 2456163 4200–7400 ρ Aql09.07.2013 2456483 4000–7200 29 Vul08.08.2015 2457243 4000–7700 107 Her06.08.2016 2457607 4000–6700 ρ Aql09.08.2016 2457610 4000–7700 ρ Aql21.07.2017 2457956 4000–9700 ρ Aql19.10.2017 2458046 4000–9700 ρ Aql08.08.2018 2458339 4000–7000 29 Vul13.10.2018 2458405 4000–9400 29 Vul07.07.2019 2458672 4000–9400 29 Vul, 30 Vul07.08.2019 2458703 4000–9250 29 Vul, HD 196775, η Sge03.10.2019 2458760 4000–9250 HD 196775, η Sge25able 5: The observed relative intensities of emission lines for IC 4997 on the scale I ( H β ) = 100 andthe observed intensity of H β in units of 10 − erg cm − s − . λ , ˚A Ion 04.08.10 31.07.11 26.08.11 21.07.12 23.08.12 09.07.13 08.08.15 06.08.164102 H δ γ α
376 397 380 366 463 376 425 -6584 [NII] 24.7 26.2 28.1 25.4 33.3 27.4 30.1 38.86678 HeI 5.52 5.75 5.69 5.41 6.56 5.11 5.31 -6716 [SII] 0.86 0.96 0.92 0.88 1.14 0.77 1.04 -6731 [SII] 1.85 1.95 1.91 1.85 2.34 1.71 1.89 -7065 HeI 17.0 19.5 18.5 18.6 20.9 15.8 17.6 -7136 [ArIII] 12.6 15.1 13.7 14.9 16.7 12.4 14.4 -7170 [ArIV] 0.61 0.35 0.54 0.34 - - 0.62 -7237 [ArIV] - 0.26 0.32 0.28 - - 0.44 -7281 HeI - 1.63 1.39 1.31 1.18 - 1.70 -7751 [ArIII] - - 3.12 - - - - -9069 [SIII] - - 21.7 - - - - - F ( H β ) λ , ˚A Ion 09.08.16 21.07.17 19.10.17 08.08.18 13.10.18 07.07.19 07.08.19 03.10.194102 H δ γ α F ( H β ) β , [OIII] λ , for the 1970-2019period. 1970 1990 2003 2019 T e , K 8000 11900 9740 7650 N e , см − O / H + F ( λ /F ( H β ) T e , K 10000 16700 12800 9440 N e , см − O / H + F ( λ /F ( H β ) T e , K 12000 22600 16150 12220 N e , см − O / H + F ( λ /F ( H β ))