The CaFe Project: Optical Fe II and Near-Infrared Ca II triplet emission in active galaxies -- simulated EWs, the co-dependence of cloud sizes and metal content
DDraft version April 29, 2020
Typeset using L A TEX twocolumn style in AASTeX63
Optical
Fe II and Near-Infrared
Ca II triplet emission in active galaxies(II) radial sizes from photoionization modelling
Swayamtrupta Panda
1, 2 Center for Theoretical Physics (PAS), Al. Lotników 32/46, 02-668 Warsaw, Poland Nicolaus Copernicus Astronomical Center (PAS), ul. Bartycka 18, 00-716 Warsaw, Poland (Received April 29, 2020; Revised XX XX, XXXX; Accepted XX XX, XXXX)
Submitted to ApJABSTRACTI analyse the emitting regions for the optical Fe II and near-infrared Ca II emission pertaining tothe broad line region in active galaxies, using the CLOUDY photoionisation modelling to ascertain thetight correlation shown between these species. I explicitly show the connection between two physicalquantities, i.e. metallicity in the BLR cloud, and, the cloud column density (N H ) highlighting the co-dependence between them suggesting that even strong Fe II emitters, such as I Zw 1 can be modelledwith metallicities that do not require values as high as shown from previous studies. The studysuggests that the bulk of the Ca II emitting region is located farther in the BLR by a factor ∼ Keywords: galaxies: active, quasars: emission lines; accretion disks; radiative transfer; scaling relations INTRODUCTIONThe complexity in the Fe ii emission, with its ori-gin from the inner parsec scales in active galactic nuclei(AGNs), is yet to be solved completely (see Collin & Joly2000, for an overview). This complexity is majorly dueto the numerous transition lines this first ionized stateof Fe has, spreading across the near infrared to ultravi-olet wavelengths (Boroson & Green 1992; Bruhweiler &Verner 2008; Garcia-Rissmann et al. 2012) which makesit quite complicated to be modelled. Since its inception(Greenstein & Schmidt 1964), the study of this complexionic species has seen significant development, from thepoint of view of the spectral quality of the data with im-proved telescope technologies (Laor et al. 1997; Kovače-vić et al. 2010; Kovačević-Dojčinović & Popović 2015;Marinello et al. 2016) including long-term reverberationmapping campaigns (Hu et al. 2015; Zhang et al. 2019,and references therein), to the spectral fitting routines(Kriss 1994; Calderone et al. 2017; Guo et al. 2019)and empirical templates (Boroson & Green 1992; Vester- Corresponding author: Swayamtrupta [email protected], [email protected] gaard & Wilkes 2001) for
I Zw 1 , a prototypical narrow-line Seyfert galaxy. Simultaneously, there has been no-table stride in understanding the excitation mechanismof this species in AGNs and corresponding templateshave been proposed strictly from the theoretical stand-point (Verner et al. 1999; Sigut & Pradhan 2003). Thecurrent consensus is shifted towards the use of the semi-empirical templates (Véron-Cetty et al. 2004; Kovačevićet al. 2010; Garcia-Rissmann et al. 2012) that solves theproblem to a great extent, although not all.Fe ii emission also bears extreme importance in thecontext of the main sequence of quasars. Several note-worthy works have established the prominence of thestrength of the optical Fe ii emission (4434-4684 Å ) withrespect to the broad H β line width (henceforth R FeII )and it’s relevance to the Eigenvector 1 sequence linkingprimarily to the Eddington ratio (Sulentic et al. 2000,2001; Shen & Ho 2014; Marziani et al. 2018). Recentstudies have addressed the importance of the Fe ii emis-sion and its connection with the Eddington ratio, and tothe black hole mass, cloud density, metallicity and tur-bulence (Panda et al. 2018), to the shape of the ionizingcontinuum (Panda et al. 2019a), and to the orientationeffect (Panda et al. 2019b, 2020b). a r X i v : . [ a s t r o - ph . H E ] A p r Panda
The difficulty in understanding the Fe ii emission hasled us in search of other reliable, simpler ionic speciessuch as Ca ii and O I (Martínez-Aldama et al. 2015,and references therein) which would originate from thesame part of the BLR and could play a similar role inquasar main sequence studies. Here, the Ca ii emissionrefers to the Ca triplet (CaT), i.e., the IR triplet emit-ting at λ λ λ ii emission.In P20, we compile an up-to-date catalogue of quasarswhich have spectral measurements of the strength of theoptical Fe ii and NIR CaT emission (with respect to H β )and re-estimate the existing tight correlation (Martínez-Aldama et al. 2015) between them. We also perform asuite of CLOUDY photoionisation models to derive the cor-relation from the theoretical standpoint with emphasison the ionization parameter and the local cloud density.We touch upon the effect of metallicity and cloud columndensity and show their contribution, from a qualitativepoint of view.While P20 was devoted to justify the connection be-tween the optical Fe ii and NIR CaT, the main goalof this paper is to constrain the relative location of Fe ii and CaT, and to determine the metallicity requiredto optimize the emission strengths of these two species.Additionally, I investigate the effect of the cloud columndensities (N H ) on the net emission strengths of the afore-mentioned species, which, for a given local mean densityof the BLR cloud, estimates the size of the BLR cloud.In Section 2, I layout the photoionisation modellingsetup that also takes in to account the dust sublima-tion. This prescription is identical to P20 although thenovelty of this work lies in the systematic treatment ofthe metallicity, unlike P20 where we assumed only tworepresentative cases, i.e. 0.2Z (cid:12) and 5Z (cid:12) . In Section 3, Iillustrate the zone of emission for the two species in the log U − log n H parameter space which constrains them asa function of the metallicity and N H . I further show theco-dependence between the aforementioned key parame-ters (metallicity and N H ) suggesting that even strong Fe ii emitters, such as I Zw 1 can be modelled with metal-licities that do not require values as high as shown fromprevious studies. Additionally this re-affirms the bulkemitting region of CaT to lie behind the optical Fe ii .These analyses open up new frontiers in the BLR physicsand emission, some of which are discussed in Section 4.The key findings from this study are summarized in Sec-tion 5. METHODS AND ANALYSIS I perform a suite of
CLOUDY models by varying thecloud particle density, . ≤ n H ≤ (cm − ), theionization parameter, − . ≤ log U ≤ − . , the metal-licity, 0.1Z (cid:12) ≤ Z ≤ (cid:12) , and cloud column density, ≤ N H ≤ cm − . The log U − log n H range isconstrained with respect to the dust sublimation radiusprescription from Nenkova et al. (2008). The currentanalysis considers the classical view of the dustless BLR(Adhikari et al. 2018) . The basis of this separation isbased on the sublimation of the dust grain. I employthe approach from our recent work (P20), wherein weconsider a characteristic dust sublimation temperature, T sub = 1500 K. Assuming a singular dust grain size, a =0.05 µ m, we simplify the sublimation radius scaling de-pendence (Barvainis 1987; Koshida et al. 2014) on onlythe source bolometric luminosity: R sub = 0 . (cid:112) L/ parsecs (Nenkova et al. 2008). The details of the post-photoionization implementation of the dust-line to sepa-rate the dusty and non-dusty BLR emission is shown inP20. The model assumes a distribution of cloud densi-ties at various radii from the central illuminating sourceto mimic the gas distribution around the close vicinity ofthe active nuclei. The range of metallicity incorporatedhere is inspired by the works on quasar main sequence,containing distribution of quasars ranging from the low-R FeII “normal” Seyfert galaxies which can be modelledwith sub-solar assumption, and the Narrow-line Seyfertgalaxies (NLS1s), especially the extreme Fe ii emittersthat have super-solar metallicities (Laor et al. 1997; Ne-grete et al. 2012; Marziani et al. 2019a, Śniegowska etal. in prep.). Also, the range of cloud column den-sity used is in agreement with previous works, mainly inFerland & Persson (1989); Matsuoka et al. (2007, 2008);Negrete et al. (2012) and further extension shown inP20. I utilize the spectral energy distribution (SED) forthe nearby (z= 0.061) NLS1, I Zw 1 . The R FeII andCaT/H β estimates are extracted from these simulations. RESULTSThe formalism described in the previous section allowsus to portray the emitting regions for the two species (Fe ii and CaT) in terms of the their emission strengths withrespect to broad H β emission, i.e. R FeII and CaT/H β ,respectively. In the following sub-sections, I discuss the N( U ) × N( n H ) × N(Z) × N( N H ) = 8 × × × Alternative views of the BLR (Czerny & Hryniewicz 2011; Baskinet al. 2014) are not considered in this work. The
I Zw 1 ionizing continuum shape is obtainedfrom NASA/IPAC Extragalactic Database. Seehttps://github.com/Swayamtrupta/CaT-FeII-emission forthe final SED used in this work. e II and
Ca II emission in AGNs: Paper II
Parametrization of emitting regions for Fe ii andCaTFigure 1 illustrates the log U − log n H parameter space( − . ≤ log U ≤ − . , . ≤ n H [cm − ] ≤ ) atcolumn density, N H = 10 cm − . The color-axis repre-sents the strength of the Fe ii emission with respect toH β , i.e. R FeII . The first five panels (top three and bot-tom two from left) depict the change in the metallicity,-1 ≤ log Z [Z (cid:12) ] ≤
1, with a step size of 0.5 in log-scale.The extent of the colorbar in each of these five plots iskept fixed to appreciate the effect of the change in metal-licity. The last panel (bottom right) represents the com-bined contribution from all the previous five plots. Thisapproach of plotting is kept consistent in the subsequentplots in this section.Notice the concentration of the higher R
FeII for thehighest ionization parameters (-2.0 (cid:46) log U (cid:46) -1.5) andmoderately high cloud density ( . (cid:46) n H [cm − ] (cid:46) . ). This region of maximum R FeII remains un-changed with change in the metallicity, although the re-covered R
FeII estimates increase with increasing metal-licity (from R
FeII ≈ (cid:12) to R FeII ≈ (cid:12) ) which has been shown in earlier works(Panda et al. 2018, 2019a). This maximum value forthe R FeII is obtained for the same exact value of ioniza-tion parameter, log U = -1.75, and cloud density, log n H = 12.0 ( cm − ) for these two extreme cases .Higher values of column densities (N H ∼ cm − )have been used in previous studies (Ferland & Persson1989; Negrete et al. 2012). In those studies, it is pointedthat such high values were needed to show that theStrömgren depth is less than the size of the clouds for theionic states that are considered, i.e., clouds are radiationbounded (Negrete et al. 2012). I test this dependenceof the increase in the column density in the subsequentFigures 7, 8, 9 and 10, going upto N H = 10 cm − .Figure 2 illustrates similarly the log U − log n H param-eter space wherein the color-axis represents the strengthof the CaT emission with respect to H β , i.e. CaT/H β , atidentical column density, N H = 10 cm − . Comparingthis with Figure 1, the plots clearly show a shift in theposition of the emitting region towards lower ionizationparameter, i.e. log U ∼ -3.25. The cloud density remains for the intermediate metallicities, log Z [Z (cid:12) ] = -0.5, 0, 0.5, thevalue of the cloud density for which I retrieve the maximum R FeII is slightly lower, log n H = 11.75 ( cm − ). Although the case withlog n H = 12.0 ( cm − ) in these cases has R FeII values very closeto this maximum, within 1-3%. For more details see Table 1. almost unchanged, i.e. ( . (cid:46) n H [cm − ] (cid:46) . ),suggesting similar emitting regions for these two species,in radial scales, CaT region following Fe ii region.The maximum estimates for R FeII and CaT/H β forthe different models (changing metallicities and col-umn densities) are reported in Table 1. Increasing thecolumn density has a similar effect to injecting moremetal species in the BLR cloud on the overall intensi-ties of these strengths, For example, a BLR cloud witha N H = 10 cm − and Z ∼ (cid:12) , recovers compa-rable R FeII estimates to that of a BLR cloud with aN H = 10 cm − and Z ∼ (cid:12) . Notice the slightchange in the log U − log n H values that are requiredto produce these values for the R FeII in both the cases(see footnote in Table 1).3.2.
Extracting the information on the BLR size
I investigate the coupled distribution between theionization parameter and local cloud density. As hasbeen previously explored in Negrete et al. (2012, 2014);Marziani et al. (2019b), I take the product of the ion-ization parameter and the local cloud density ( U · n H ),i.e. this entity bears resemblance to ionizing flux, andfor a given number of ionising photons emitted by theradiating source, this can be used to estimate the sizeof the BLR ( R BLR ). In this paper, I use a constantshape for the ionizing continuum apt for the nearbyNLS1,
I Zw 1 . The bolometric luminosity of
I Zw 1 is L bol ∼ . × erg s − . This is obtained by apply-ing the bolometric correction prescription from Netzer(2019) on I Zw 1 ’s L ∼ . × erg s − (Persson1988). Hence, putting this all together, I have R BLR [ cm ] = (cid:115) Q ( H )4 πU n H c ≡ (cid:114) L bol πhν U n H c (cid:104) . × √ U n H (1)where, R BLR is the distance of the emitting cloud fromthe ionizing source which has a mean local density n H and receives an ionizing flux that is quantified by theionization parameter, U. Q ( H ) is the number of ionizingphotons, which can be equivalently expressed in terms ofthe bolometric luminosity of the source per unit energyof a single photon, i.e. h ν . Here, I consider the averagephoton energy, h ν = 1 Rydberg (Wandel et al. 1999;Marziani et al. 2015).In Figures 3 and 4, I extract the information from theresults obtained from the CLOUDY models and plot theR
FeII and CaT/H β estimates versus this indicator of the R BLR , i.e. log Un H . The color-coding in the left panelsin Figures 3 and 4 are with respect to log U . On the rightpanels the color axis is with respect to log n H . Here, thecloud column density is assumed to identical to the case Panda log U l o g n log Z [Z ] = -1, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = -0.5, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = 0, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = 0.5, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = 1, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] [-1,1], log N H = 10 cm l o g R F e II Figure 1. log U − log n H
2D histograms color-weighted by R
FeII (in log-scale) with column density, N H = 10 cm − . Eachof the first 5 panels correspond to a case of metallicity (in log-scale, in units of Z (cid:12) ). The plot on the bottom right combinestogether the contribution from all the five panels shown before. log U l o g n log Z [Z ] = -1, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = -0.5, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = 0, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = 0.5, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = 1, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] [-1,1], log N H = 10 cm l o g C a T / H Figure 2.
Similar to Figure 1. Plots are color-weighted by log CaT/H β . shown in the previous section, i.e. N H = 10 cm − . Iconsider three cases for the metallicities, i.e. log Z [Z (cid:12) ]= -0.5, 0 and 0.5, to represent a sub-solar, at solar andsuper-solar compositions, respectively.Looking carefully, first at Figure 3, one finds that eachof the trend (with respect to the color-axis) shows a clearpeak which corresponds to the maximum R FeII estimatecorresponding to the log U (or equivalently the log n H ).Similar to the panels in Figure 1, I find that the over-all maximum R FeII estimate is recovered for a log U = -1.75 and log n H = 11.75, regardless of the increase inthe metallicity in the BLR cloud. If one tries to followthe location of the maximum of the peak for the range ofthe ionization parameter considered, will see the gradualdecline in the recovered R FeII estimates in either direc-tion. The same non-monotonic trend holds true for thecases considered with respect to the cloud local density.For the low metallicity case, log Z [Z (cid:12) ] = -0.5, themaximum R
FeII obtained is ∼ FeII estimate nearly by a fac- e II and
Ca II emission in AGNs: Paper II
FeII ∼ (cid:12) ] = 0. And there is a furtherincrease by a factor 2 (R FeII ∼ (cid:12) ] = 0.5. Hence, from this base model analysis, I findthat we can indeed recover the R FeII estimates that areconsistent with the highest Fe ii emitters if I increasethe metallicity to be ∼ (cid:12) .If I increase the column density by an order, i.e. N H =10 cm − (see Figure 15), the maximum R FeII jumps to1.114 for the low metallicity case (factor 1.36 increase).Similar rise in the maximum R
FeII estimate is seen inthe other two cases of metallicities (at solar and super-solar). There’s a further increase in R
FeII for all thecases in metallicity when the column density is furtherincreased to N H = 10 cm − (see Figure 16).Next, in Figure 4, on a first look, I find subtle differ-ences in the trends as compared to the previous figurewith respect to R FeII . Here, as was also shown in thedensity histogram plots for CaT/H β , I find that peakstrength is recovered for the lowest ionization parame-ter, log U = -3.25, which is smaller by 1.5 (log-scale)than that required for the corresponding maximum inR FeII . While the corresponding cloud density in thiscase is log n H = 12.25 which is comparable to that ob-tained for the R FeII . As is seen for the R
FeII cases, herealso there is an increase in the CaT/H β strengths bynearly a factor 1.5 when the column density is increasedto N H = 10 cm − (see Figure 17). There’s a furtherincrease by a factor 1.3 when the column density is as-sumed to be N H = 10 cm − (see Figure 18). The max-imum estimates for the R FeII and CaT/H β are reportedin the Table 1 for all the considered cases in terms ofthe metallicities and column densities.Notice the shift in the x-position for the CaT/H β to-wards a smaller value of log Un H ( ∼ FeII case, this is smaller by an order of magni-tude (the x-position for the R
FeII has a log Un H ∼ R BLR , means thatthe emitting region for CaT is formed ∼ ii emitting region .3.3. Co-dependence of metallicity and cloud columndensity
In the previous sections, I have shown how the R
FeII and CaT/H β estimates can be maximized with respectto increase in the metallicity and cloud column densi-ties independently. There is a clear hint that the real From Equation 1, in radial scales, the R BLR corresponding tothe maximum R
FeII is at ≈ × cm. Likewise, for theCaT/H β maximum, this value of R BLR moves to ≈ × cm. scenario perhaps points towards a collective increase inboth these quantities. This might counter the argumentstowards the use of the very high metallicities (Z (cid:38) (cid:12) )to recover the R FeII estimates for the strong Fe ii emit-ters (Nagao et al. 2006; Negrete et al. 2012, Śniegowskaet al. in prep.) which has strong implications for theBLR cloud properties, especially their density distribu-tion function and their radial distribution. Additionally,(Nagao et al. 2006) comment that the 5Z (cid:12) estimates canbe partly due to lack of spectral resolution and that thetypical metallicities required in BLR is atleast super-solar if not higher. We now show in this paper thatsuch high R FeII can be obtained with a modest increasein the cloud size without requesting such high metallicityvalues. In this section, I explicitly test this connectionbetween the two aforementioned parameters in terms ofthe R
FeII and CaT/H β estimates they recover.From the analyses in the previous section, the pairsof ionization parameter and local cloud density, i.e. log Un H , that reproduce the maximum R FeII and CaT/H β estimates are (a) [-1.75, 11.75] and (b) [-3.25, 12.25],respectively. Below, I consider each of these two casesto highlight the co-dependence of the metallicity andthe column density in terms of the R FeII and CaT/H β estimates.In Figure 5, I utilize the pair (a) log Un H = [-1.75,11.75], and considering three representative cases for themetallicity, i.e. Z = Z (cid:12) , 2Z (cid:12) and 3Z (cid:12) , I have shown thetrends of increasing R FeII as a function of the columndensity, ≤ N H [cm − ] ≤ (left panel). On theright panel of this figure, I have the corresponding casefor the CaT/H β for reference. The best-fit parametersfor these trends are shown by the black dashed lineson the figure and reported in Table 2 along with therespective Pearson’s correlation coefficients and p-valueswhich have been estimated using the lm routine in theR Statistical language (R Core Team 2019).From prior spectroscopic observations for I Zw 1 ,the R
FeII and CaT/H β estimates have been reported:(a) R FeII and CaT/H β estimates from Persson (1988):1.778 ± ± FeII and CaT/H β estimates from Marinello et al. (2016):2.320 ± ± Un H = [-3.25, 12.25] and report the best-fitparameters in Table 3.Subsequently, I can make a deduction about the pos-sible metallicity and column density that is able to re-produce the observed value of the R FeII and CaT/H β .As these two pairs (for log Un H ) of models, i.e. (a) Panda log Un R F e II log Z [Z ] = -0.5, log N H = 10 cm l o g U log Un R F e II log Z [Z ] = -0.5, log N H = 10 cm l o g n H log Un R F e II log Z [Z ] = 0, log N H = 10 cm l o g U log Un R F e II log Z [Z ] = 0, log N H = 10 cm l o g n H log Un R F e II log Z [Z ] = 0.5, log N H = 10 cm l o g U log Un R F e II log Z [Z ] = 0.5, log N H = 10 cm l o g n H Figure 3.
Non-monotonic behaviour of R
FeII versus log Un H color-coded with respect to (a) log U (left panels); and (b) log n H (right panels). The three sets represent metallicity cases: log Z [Z (cid:12) ]: -0.5 (top), 0 (middle) and 0.5 (bottom). Column density,N H = 10 cm − is assumed. e II and Ca II emission in AGNs: Paper II log Un C a T / H log Z [Z ] = -0.5, log N H = 10 cm l o g U log Un C a T / H log Z [Z ] = -0.5, log N H = 10 cm l o g n H log Un C a T / H log Z [Z ] = 0, log N H = 10 cm l o g U log Un C a T / H log Z [Z ] = 0, log N H = 10 cm l o g n H log Un C a T / H log Z [Z ] = 0.5, log N H = 10 cm l o g U log Un C a T / H log Z [Z ] = 0.5, log N H = 10 cm l o g n H Figure 4.
Non-monotonic behaviour of CaT/H β versus log Un H color-coded with respect to (a) log U (left panels); and (b)log n H (right panels). The three sets represent metallicity cases: log Z [Z (cid:12) ]: -0.5 (top), 0 (middle) and 0.5 (bottom). Columndensity, N H = 10 cm − is assumed. Panda [-1.75, 11.75] and (b) [-3.25, 12.25], provide the maxi-mum R
FeII and CaT/H β estimates, I mark the R FeII -based plot from the Figure 5 and CaT/H β -based plotfrom the Figure 6 as the best models, respectively. Forthe R FeII case (Figure 5), one can recover the observedvalue of R
FeII = 2.320 ± ≈ (cid:12) with N H ≈ cm − . This isalso possible with a metallicity Z ≈ Z (cid:12) , although witha much higher column density, i.e. N H ≈ . cm − .For the older estimate of R FeII = 1.778 ± ≈ Z (cid:12) with N H ≈ . cm − .On the other hand, for the CaT/H β estimates, dueto the large dispersion in the quoted values from theobservations (Persson 1988; Marinello et al. 2016), thepossibilities to reproduce the CaT/H β is higher. In Fig-ure 6, one can recover the observed value of CaT/H β = 0.564 ± ≈ (cid:12) with N H ≈ cm − . This is also pos-sible with a metallicity Z ≈ (cid:12) with N H ≈ . − cm − , and, with a metallicity Z ≈ Z (cid:12) with N H ≈ . − cm − . While the estimates for CaT/H β from Persson (1988) don’t differ much from the newerestimates from Marinello et al. (2016), the correspond-ing value (= 0.513 ± H ≈ cm − , for closer to solar metallicities (Z =1-2Z (cid:12) ).Additional constraints from high signal-to-noise rest-frame UV spectrum for I Zw 1 can help to narrowdown the possibilities with respect to the metallicity.There are quite a few metallicity indicators such asAl iii λ ii λ iv λ iv] λ iv λ I Zw 1 ’s HST-FOS spectrum and reportedthe various spectral parameters in their paper. TheAl iii /He ii flux ratio from their analysis is ≈ iv +O iv] /C iv gives ≈ v λ ii flux ratio suggests a metallicity of ∼ (cid:12) , although this ratio is quite sensitive to changein ionization parameter (Wang et al. 2012). Other ra-tios, such as C iv /He ii and Si iv +O iv] /He ii point to-wards even higher metallicities (Z (cid:38) (cid:12) ), althoughthey are not so reliable due to issues related to blendingwith other species which becomes cumbersome unless a better quality spectra is available. Hence, utilizing theAl iii /He ii flux ratio, coupled with the photoionization-based estimates in this work, puts the column den-sity required for R FeII to be ∼ cm − and for theCaT/H β to be slightly above this limit, i.e. N H ≈ − cm − . DISCUSSIONIn this paper, I analyse and ascertain the emittingregions required to maximize the strengths of the op-tical Fe ii (R FeII ) and NIR Ca ii (CaT/H β ) emission,the latter of which is touted to be used as a proxy forthe estimation of the former attributed to its complex-ity studied in detail in prior studies (see P20 for anoverview). The log U − log n H parametrization with re-spect to the strengths of these two species allowed toconstrain the ionization parameters and the cloud den-sities that put the CaT emitting region embedded deeperin BLR cloud by a factor ∼ ii emitting region. This paper also looks more carefullyinto the co-dependence of the metallicity and columndensity that plays an important role in recovering themaximum R FeII and CaT/H β estimates. Applying theconstraints from the observational measurements putsthe metallicity slightly above solar abundances with thecolumn density ∼ cm − for R FeII and by almost anorder higher for the CaT/H β .The radial scales for the bulk of the Fe ii and CaTemitting regions are found to be apart by a factor ∼ ∼ ii region. When translated to velocity dis-persions, this suggests that the optical Fe ii should have ≈ ii and the CaT are linearlycorrelated from measurements for 13 sources. The diffi-culty in measuring individual Fe ii lines in the optical isa long standing issue, the presence of numerous overlap-ping transitions make it really difficult to identify andisolate them. This is possible for extreme NLS1s wherethe Fe ii lines in the NIR have FWHM of ∼
700 km s − ,such that the optical Fe ii lines may be resolved (Al-berto Rodríguez-Ardila, priv. comm.). A systematic more recent works suggest a slightly higher value of theses lineratios, for example, Al iii /He ii = 5.35 ± ∼ (cid:12) themodels predict a column density ∼ cm − for R FeII and (cid:38) . cm − for CaT/H β . e II and Ca II emission in AGNs: Paper II log N H [cm ] R F e II best modelR FeII (Z ) R
FeII (2Z ) R
FeII (3Z ) 24.0 24.5 25.0 25.5 26.0 log N H [cm ] C a T / H CaT/H (Z ) CaT/H (2Z ) CaT/H (3Z ) log U = -1.75, log n H = 11.75 [cm ] Figure 5. R FeII versus column density (N H ) (left-panel) for three cases of metallicity, Z = Z (cid:12) , 2Z (cid:12) and 3Z (cid:12) . These estimatesare from model with log U = -1.75 and log n H = 11.75 ( cm − ). The best-fit relations are shown with black dashed lines(see Table 2 for corresponding values). Patches in orange (Persson 1988) and cyan (Marinello et al. 2016) mark the observedestimates for I Zw 1 with uncertainties. The corresponding case for the CaT/H β is shown on the right panel. log N H [cm ] R F e II R FeII (Z ) R
FeII (2Z ) R
FeII (3Z ) 24.0 24.5 25.0 25.5 26.0 log N H [cm ] C a T / H best modelCaT/H (Z ) CaT/H (2Z ) CaT/H (3Z ) log U = -3.25, log n H = 12.25 [cm ] Figure 6.
Same as in Figure 5. These estimates are from model with log U = -3.25 and log n H = 12.25 ( cm − ). study of the optical Fe ii lines and their NIR counter-parts is required to disentangle this mystery. Neverthe-less, in Marinello et al. (2016), the authors found thatthe FWHM(Fe ii opt ) < FWHM(H β broad ). This sets anupper limit on the optical Fe ii FWHMs which is alsoconfirmed from the studies by Kovačević et al. (2010)where they analyzed 302 AGNs and found the averageFWHMs for optical Fe ii to be ≈ − . In Ko-vačević et al. (2010), the authors found a strong correla- tion between the FWHMs for the optical Fe ii and H β ,where the average FWHM for the intermediate compo-nent of H β was ≈ − and ≈ − forthe very-broad component. From the theoretical point,line recombination is more efficient for optical lines thanNIR. If the bulk of the optical Fe ii comes from recom-bination, e.g. Ly α fluorescence accounts for ∼
20% ofthe observed Fe ii (Sigut & Pradhan 2003), that wouldexplain a broader optical Fe ii FWHM. Collisional exci-0
Panda tation and other mechanisms could be dominant in theouter parts of the BLR (Marinello et al. 2016, and ref-erences therein).With increase in the column density effectively in-creasing the cloud size (as the inner radius gets fixed inthis prescription for a given U , n H and bolometric lumi-nosity), there is a possibility that, for quite high valuesof the column density, the cloud effectively extends be-yond the dust sublimation cut. This cut assumes thatthe dust sublimates at T=1500 K and follows the scal-ing with respect to the source luminosity (Nenkova et al.2008), which can be safely approximated to its bolomet-ric luminosity. The sublimation radius then is R sub ≈ . Based on the formalism in this paper, after thedust cut the lowest local density achievable is 10 . andthe lowest ionization is ∼ -3.25. Using the Equation 1,gives the R BLR (or the inner radius) to be ≈ × cm. For the highest column density ( = 10 cm − ) anda considerably low cloud local density, (= 10 . cm − ),gives a cloud size, d ≈ × cm. This sets the outerradius for the BLR at R out ≈ × cm (= 0.557pc). This is the largest possible outer radius within thissetup, where, when we consider the lowest cloud densityand ionization parameter along with maximum columndensity to maximize the cloud size. Although, the cloudsize is fairly small as compared to the R BLR by almost 4orders of magnitude, hence quite negligible. Hence, theemitting regions for bot the optical Fe ii and the CaT isprecisely within the non-dusty BLR region.On the other hand, H β reverberation mapping givesthe R BLR for
I Zw 1 at 9.636 × cm (Huang et al.2019). It has been previously proposed that the Fe ii time-lags are about a factor 2-3 times longer comparedto that of H β (Barth et al. 2013). This puts the innerradius slightly smaller (factor of 0.53) than what is pre-dicted from the photoionisation modelling for the maxi-mum R FeII (see Section 3.2). This also brings the outerradius below the sublimation radius assumed in this pa-per (i.e. 0.83 pc). Additionally, the maximum time-lagreported from reverberation mapping campaigns is forthe source PG 1700+518 (Bentz et al. 2013; Martínez-Aldama et al. 2019) with measured lag of ∼ . This puts the cloud’s outer edge at 251.922 lightdays, which is 3.92 times smaller than the sublimationradius ( 0.83 pc ≈ ≈ × cm. ≈ × cm. within optically thin media , such that, for both thesespecies, with a slightly above solar metallicities, we canestimate the R FeII and CaT/H β values in agreement toobserved measurements.The physical driver behind the tight R FeII -CaT/H β correlation obtained from P20 from the point of theobservations will be confirmed in an upcoming paper(Panda et al. in prep.). CONCLUSIONSIn this paper, I analyze the emitting regions for the op-tical Fe ii and NIR CaT emission originating from theBLR in AGNs using the photoionization code CLOUDY and ascertained the tight correlation shown betweenthese two species. I explicitly show the connection be-tween the two physical quantities, metallicity in the BLRcloud, and the cloud column density (N H ). This high-lights the co-dependence between these parameters sug-gesting that even strong Fe II emitters, such as I Zw 1 can be modelled with metallicities that do not requirevalues as high as shown from previous studies. I summa-rize the important conclusions derived from this work asfollowing: • The log U − log n H based parametrization allowedus to visualize the emitting regions for the Fe ii and CaT strengths, i.e. R FeII and the CaT/H β .The zones of emission for maximizing the R FeII isobtained with the high ionization parameters, i.e.-2.0 (cid:46) log U (cid:46) -1.5) and moderately high clouddensity ( . (cid:46) n H [cm − ] (cid:46) . ). While,for the case of the CaT/H β , much lower ionizationparameters are requested ( log U ∼ -3.25) althoughthe cloud density remains almost unchanged, i.e.( . (cid:46) n H [cm − ] (cid:46) . ). • With the log U − log n H based parametrization andwith the knowledge of the bolometric luminosity ofthe considered source in this study, the prototypi-cal NLS1 galaxy I Zw 1 , I’m able to establish theemitting regions in radial scales as well. I find, forthe parameters apt for the maximum R
FeII andCaT/H β , when referred to in terms of the BLRsize ( R BLR ), ∼ ii . • I extend the parametrization to include also the ef-fects of the metallicity and the cloud column den-sity in retrieving the maximum R
FeII and CaT/H β i.e., optical depth, τ = σ T · N H . Here, σ T is the Thompson’sscattering cross-section and N H is the cloud column density. e II and Ca II emission in AGNs: Paper II • I establish the co-dependence between the metal-licity and the cloud column density. I find thatthere exists a coupling between these two quanti-ties which put very high metallicity assumptions,i.e. Z (cid:38) (cid:12) ), in existing photoionisation modelsto question. • With additional constraints from high S/N spec-tral measurements for the R
FeII and CaT/H β andobserved Al iii /He ii / flux ratios suggesting metal-licity close to 2Z (cid:12) ( ≈ FeII to be ∼ cm − andfor the CaT/H β to be slightly above this limit, i.e.N H ≈ − cm − . ACKNOWLEDGMENTSThe project was partially supported by the Pol-ish Funding Agency National Science Centre, project2017/26/A/ST9/00756 (MAESTRO 9) and MNiSWgrant DIR/WK/2018/12. I would like to thank PaolaMarziani for performing a detailed spectral fitting toestimate the metallicity for I Zw 1 . I’d like to thankBożena Czerny, Mary Loli Martínez-Aldama, DeepikaAnanda Bollimpalli and Marzena Śniegowska for fruit-ful discussions leading to the current state of the paper.I thank Bożena Czerny and Sushanta Kumar Panda forproof-reading the manuscript and suggesting correctionsto improve the overall readability.
Software:
CLOUDY v17.01 (Ferland et al. 2017);
MATPLOTLIB (Hunter 2007);
NUMPY (Oliphant 2015); R (R Core Team 2019)APPENDIX Table 1 . Maximum estimates from
CLOUDY modelslog Z [Z (cid:12) ] N H = 10 N H = 10 . N H = 10 N H = 10 . N H = 10 R FeII
CaT/H β R FeII
CaT/H β R FeII
CaT/H β R FeII
CaT/H β R FeII
CaT/H β -1 0.491 (a) (c , d) (e) (c) (e) (c) (e) (h) (e) (g) -0.5 0.822 (b) (c) (b , e) (c) (a) (h) (a) (g) (a) (g) (b) (d) (b) (c) (b) (h) (b) (g) (b) (g) (b) (d) (b) (c) (b) (g) (b) (g) (b) (g) (a) (c) (f) (g) (a) (g) (b) (g) (a) (g) Notes.
The column densities (N H ) are in the units of cm − . Corresponding to [log U , log n H ] values: (a) [-1.75, 12.0]. (b) [-1.75, 11.75].(c) [-3.25, 12.0]. (d) [-3.25, 12.25]. (e) [-1.5, 12.0]. (f) [-1.5, 11.75]. (g) [-3.25, 11.5]. (h) [-3.25, 11.75]. Table 2 . Figure 5 best-fit parameters for the relation: f(log N H ) =m*(log N H ) + cZ = Z (cid:12) Z = 2Z (cid:12)
Z = 3Z (cid:12) R FeII
CaT/H β R FeII
CaT/H β R FeII
CaT/H β m 0.489 ± ± ± ± ± ± ± ± ± ± ± ± × − × − × − × − × − × − Notes.
Rows are as follows: (1) Slope. (2) Intercept. (3) Pearson’s correlation coefficient. (4) p-value. The best-fit parameters arecomputed using the lm routine in R Statistical language (R Core Team 2019). Panda log U l o g n log Z [Z ] = -1, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = -0.5, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = 0, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = 0.5, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = 1, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] [-1,1], log N H = 10 cm l o g R F e II Figure 7.
Similar to Figure 1 with column density, N H = 10 . cm − . Table 3 . Figure 6 best-fit parameters for the relation: f(log N H ) =m*(log N H ) + cZ = Z (cid:12) Z = 2Z (cid:12)
Z = 3Z (cid:12) R FeII
CaT/H β R FeII
CaT/H β R FeII
CaT/H β m 0.357 ± ± ± ± ± ± ± ± ± ± ± ± × − × − × − × − × − × − Notes.
Rows are as follows: (1) Slope. (2) Intercept. (3) Pearson’s correlation coefficient. (4) p-value. The best-fit parameters arecomputed using the lm routine in R Statistical language (R Core Team 2019). REFERENCES
Adhikari, T. P., Hryniewicz, K., Różańska, A., Czerny, B.,& Ferland, G. J. 2018, ArXiv e-prints.https://arxiv.org/abs/1803.00090Barth, A. J., Pancoast, A., Bennert, V. N., et al. 2013,ApJ, 769, 128, doi: 10.1088/0004-637X/769/2/128Barvainis, R. 1987, ApJ, 320, 537, doi: 10.1086/165571Baskin, A., Laor, A., & Stern, J. 2014, MNRAS, 438, 604,doi: 10.1093/mnras/stt2230Bentz, M. C., Denney, K. D., Grier, C. J., et al. 2013, ApJ,767, 149, doi: 10.1088/0004-637X/767/2/149Boroson, T. A., & Green, R. F. 1992, ApJS, 80, 109,doi: 10.1086/191661Bruhweiler, F., & Verner, E. 2008, ApJ, 675, 83,doi: 10.1086/525557 Calderone, G., Nicastro, L., Ghisellini, G., et al. 2017,MNRAS, 472, 4051, doi: 10.1093/mnras/stx2239Collin, S., & Joly, M. 2000, New Astronomy, 44, 531,doi: 10.1016/S1387-6473(00)00093-2Czerny, B., & Hryniewicz, K. 2011, A&A, 525, L8,doi: 10.1051/0004-6361/201016025Ferland, G. J., & Persson, S. E. 1989, ApJ, 347, 656,doi: 10.1086/168156Ferland, G. J., Chatzikos, M., Guzmán, F., et al. 2017,RMxAA, 53, 385. https://arxiv.org/abs/1705.10877Garcia-Rissmann, A., Rodríguez-Ardila, A., Sigut,T. A. A., & Pradhan, A. K. 2012, ApJ, 751, 7,doi: 10.1088/0004-637X/751/1/7Greenstein, J. L., & Schmidt, M. 1964, ApJ, 140, 1,doi: 10.1086/147889 e II and
Ca II emission in AGNs: Paper II log U l o g n log Z [Z ] = -1, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = -0.5, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = 0, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = 0.5, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = 1, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] [-1,1], log N H = 10 cm l o g R F e II Figure 8.
Similar to Figure 1 with column density, N H = 10 cm − . log U l o g n log Z [Z ] = -1, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = -0.5, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = 0, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = 0.5, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = 1, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] [-1,1], log N H = 10 cm l o g R F e II Figure 9.
Similar to Figure 1 with column density, N H = 10 . cm − .Guo, H., Liu, X., Shen, Y., et al. 2019, MNRAS, 482, 3288,doi: 10.1093/mnras/sty2920Hamann, F., & Ferland, G. 1999, ARA&A, 37, 487,doi: 10.1146/annurev.astro.37.1.487Hu, C., Du, P., Lu, K.-X., et al. 2015, ApJ, 804, 138,doi: 10.1088/0004-637X/804/2/138Huang, Y.-K., Hu, C., Zhao, Y.-L., et al. 2019, ApJ, 876,102, doi: 10.3847/1538-4357/ab16efHunter, J. D. 2007, Computing in Science and Engineering,9, 90, doi: 10.1109/MCSE.2007.55 Koshida, S., Minezaki, T., Yoshii, Y., et al. 2014, ApJ, 788,159, doi: 10.1088/0004-637X/788/2/159Kovačević-Dojčinović, J., & Popović, L. Č. 2015, ApJS,221, 35, doi: 10.1088/0067-0049/221/2/35Kovačević, J., Popović, L. Č., & Dimitrijević, M. S. 2010,ApJS, 189, 15, doi: 10.1088/0067-0049/189/1/15Kriss, G. 1994, Astronomical Society of the PacificConference Series, Vol. 61, Fitting Models to UV andOptical Spectral Data, ed. D. R. Crabtree, R. J. Hanisch,& J. Barnes, 437 Panda log U l o g n log Z [Z ] = -1, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = -0.5, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = 0, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = 0.5, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] = 1, log N H = 10 cm l o g R F e II log U l o g n log Z [Z ] [-1,1], log N H = 10 cm l o g R F e II Figure 10.
Similar to Figure 1 with column density, N H = 10 cm − . log U l o g n log Z [Z ] = -1, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = -0.5, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = 0, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = 0.5, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = 1, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] [-1,1], log N H = 10 cm l o g C a T / H Figure 11.
Similar to Figure 2 with column density, N H = 10 . cm − .Laor, A., Jannuzi, B. T., Green, R. F., & Boroson, T. A.1997, ApJ, 489, 656, doi: 10.1086/304816Marinello, M., Rodríguez-Ardila, A., Garcia-Rissmann, A.,Sigut, T. A. A., & Pradhan, A. K. 2016, ApJ, 820, 116,doi: 10.3847/0004-637X/820/2/116Martínez-Aldama, M. L., Czerny, B., Kawka, D., et al.2019, ApJ, 883, 170, doi: 10.3847/1538-4357/ab3728Martínez-Aldama, M. L., Dultzin, D., Marziani, P., et al.2015, ApJS, 217, 3, doi: 10.1088/0067-0049/217/1/3Marziani, P., Sulentic, J. W., Negrete, C. A., et al. 2015,ApSS, 356, 339, doi: 10.1007/s10509-014-2136-z Marziani, P., Dultzin, D., Sulentic, J. W., et al. 2018,Frontiers in Astronomy and Space Sciences, 5, 6,doi: 10.3389/fspas.2018.00006Marziani, P., del Olmo, A., Martínez-Carballo, M. A., et al.2019a, A&A, 627, A88,doi: 10.1051/0004-6361/201935265Marziani, P., Bon, E., Bon, N., et al. 2019b, Atoms, 7, 18,doi: 10.3390/atoms7010018Matsuoka, Y., Kawara, K., & Oyabu, S. 2008, ApJ, 673, 62,doi: 10.1086/524193 e II and Ca II emission in AGNs: Paper II log U l o g n log Z [Z ] = -1, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = -0.5, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = 0, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = 0.5, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = 1, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] [-1,1], log N H = 10 cm l o g C a T / H Figure 12.
Similar to Figure 2 with column density, N H = 10 cm − . log U l o g n log Z [Z ] = -1, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = -0.5, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = 0, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = 0.5, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = 1, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] [-1,1], log N H = 10 cm l o g C a T / H Figure 13.
Similar to Figure 2 with column density, N H = 10 . cm − Panda log U l o g n log Z [Z ] = -1, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = -0.5, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = 0, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = 0.5, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] = 1, log N H = 10 cm l o g C a T / H log U l o g n log Z [Z ] [-1,1], log N H = 10 cm l o g C a T / H Figure 14.
Similar to Figure 2 with column density, N H = 10 cm − e II and Ca II emission in AGNs: Paper II log Un R F e II log Z [Z ] = -0.5, log N H = 10 cm l o g U log Un R F e II log Z [Z ] = -0.5, log N H = 10 cm l o g n H log Un R F e II log Z [Z ] = 0, log N H = 10 cm l o g U log Un R F e II log Z [Z ] = 0, log N H = 10 cm l o g n H log Un R F e II log Z [Z ] = 0.5, log N H = 10 cm l o g U log Un R F e II log Z [Z ] = 0.5, log N H = 10 cm l o g n H Figure 15. log Z [Z (cid:12) ]: -0.5 (top), 0 (middle) and 0.5 (bottom) at N H = 10 cm − Panda log Un R F e II log Z [Z ] = -0.5, log N H = 10 cm l o g U log Un R F e II log Z [Z ] = -0.5, log N H = 10 cm l o g n H log Un R F e II log Z [Z ] = 0, log N H = 10 cm l o g U log Un R F e II log Z [Z ] = 0, log N H = 10 cm l o g n H log Un R F e II log Z [Z ] = 0.5, log N H = 10 cm l o g U log Un R F e II log Z [Z ] = 0.5, log N H = 10 cm l o g n H Figure 16. log Z [Z (cid:12) ]: -0.5 (top), 0 (middle) and 0.5 (bottom) at N H = 10 cm − e II and Ca II emission in AGNs: Paper II log Un C a T / H log Z [Z ] = -0.5, log N H = 10 cm l o g U log Un C a T / H log Z [Z ] = -0.5, log N H = 10 cm l o g n H log Un C a T / H log Z [Z ] = 0, log N H = 10 cm l o g U log Un C a T / H log Z [Z ] = 0, log N H = 10 cm l o g n H log Un C a T / H log Z [Z ] = 0.5, log N H = 10 cm l o g U log Un C a T / H log Z [Z ] = 0.5, log N H = 10 cm l o g n H Figure 17. log Z [Z (cid:12) ]: -0.5 (top), 0 (middle) and 0.5 (bottom) at N H = 10 cm − Panda log Un C a T / H log Z [Z ] = -0.5, log N H = 10 cm l o g U log Un C a T / H log Z [Z ] = -0.5, log N H = 10 cm l o g n H log Un C a T / H log Z [Z ] = 0, log N H = 10 cm l o g U log Un C a T / H log Z [Z ] = 0, log N H = 10 cm l o g n H log Un C a T / H log Z [Z ] = 0.5, log N H = 10 cm l o g U log Un C a T / H log Z [Z ] = 0.5, log N H = 10 cm l o g n H Figure 18. log Z [Z (cid:12) ]: -0.5 (top), 0 (middle) and 0.5 (bottom) at N H = 10 cm −−