Are High-Σ1 Massive Blue Spiral Galaxies Rejuvenated Systems?

Quiescent galaxies generally possess denser cores than star-forming galaxies with similar mass. As a measurement of the core density, the central stellar mass surface density within a radius of 1 kpc (Σ1) was thus suggested to be closely related to galaxy quenching. Massive star-forming galaxies with high Σ1 do not fit into this picture. To understand the origin of such galaxies, we compare the spatially resolved stellar population and star formation properties of massive (>1010.5 M ⊙) blue spiral galaxies with high and low Σ1, divided by Σ1 = 109.4 M ⊙ kpc−2, based on the final release of MaNGA integral field unit data. We find that both high-Σ1 and low-Σ1 blue spirals show large diversities in stellar population and star formation properties. Despite the diversities, high-Σ1 blue spirals are statistically different from the low-Σ1 ones. Specifically, the radial profiles of the luminosity-weighted age and Mgb/〈Fe〉 show that high-Σ1 blue spirals consist of a larger fraction of galaxies with younger and less α-element-enhanced centers than their low-Σ1 counterparts, ∼55% versus ∼30%. The galaxies with younger centers mostly have higher central specific star formation rates, which still follow the spaxel-based star formation main-sequence relation. Examinations of the Hα velocity field and the optical structures suggest that galactic bars or galaxy interactions should be responsible for the rejuvenation of these galaxies. The remaining ∼45% of high-Σ1 blue spirals are consistent with the inside-out growth scenario.


Introduction
How a galaxy quenches is a key question to understand in the context of galaxy evolution.Both observational and theoretical studies suggest that the buildup of a central dense core is a requisite for galaxy quenching (e.g., Cheung et al. 2012;Fang et al. 2013;Zolotov et al. 2015;Barro et al. 2017;Dekel et al. 2019).Nowadays, a widely used probe for a central dense core is the stellar mass surface density within a radius of 1 kpc, denoted as Σ 1 .The connection between the presence of a dense core and galaxy quenching is manifested by the tight correlation between Σ 1 and the total stellar mass for quiescent galaxies (Cheung et al. 2012;Fang et al. 2013).A similar correlation also exists for star-forming galaxies, but it shows a steeper slope, a smaller normalization, and a much larger scatter.Such correlations have existed since z ∼ 3, and hence it was suggested that star-forming galaxies evolve along the relations (Barro et al. 2017).As soon as they experience a phase of significant core growth, known as the compaction process, leading to a rapid increase in Σ 1 and depletion of cold gas, they would migrate up to the relation of the quiescent population and start quenching (Zolotov et al. 2015;Barro et al. 2017;Dekel et al. 2019).Therefore, galaxies with high Σ 1 are expected to be quenched systems.
Interestingly, it has been noted that some massive (>10 10.5 M e ) blue galaxies lie on the relation for quiescent galaxies (Fang et al. 2013;Guo et al. 2020), named high-Σ 1 blue spirals hereafter.They have dense cores but are still forming stars.Although they only account for a small fraction of blue spirals (∼10%; Guo et al. 2020), they are important for our understanding of the whole picture of galaxy formation and evolution.Fang et al. (2013) suggested that high-Σ 1 blue spirals are candidates for rejuvenation, and they pointed out that the rejuvenated star formation mainly carries on in the outer parts.The more recent study by Guo et al. (2020) found that high-Σ 1 blue spirals possess massive bulges, large host dark matter halo masses, and a high bar/ring or interacting/ merger incidence rates, similar to massive red (i.e., passive) spirals.On the basis of these results, the authors conjectured that high-Σ 1 blue spirals may be rejuvenated red spirals.The rejuvenation scenario is consistent with studies based on cosmological simulations (Tacchella et al. 2016) and observations (Tacchella et al. 2022), which revealed that star-forming galaxies oscillate about the star formation main-sequence relation due to the gas depletion and replenishment.A galaxy that had obtained its dense core via the compaction process and quenched would form stars again as long as new gas fed it.The galaxy with re-ignited star formation would appear as a high-Σ 1 blue spiral, which is a natural result of such a process.
However, Woo & Ellison (2019) proposed another possibility.Based on the Mapping Nearby Galaxies at the Apache Point Observatory (MaNGA) survey (Bundy et al. 2015) integral field unit (IFU) spectra, they found that the centers of galaxies with higher Σ 1 have younger ages, enhanced specific star formation rates (sSFRs), and less metals compared to their outer parts, which is well in agreement with the compactionlike core-building scenario.The mechanisms for core building are less clear though.
In this work, we aim to deepen our understanding of the possible formation channels for high-Σ 1 blue spirals by focusing on their spatially resolved stellar population and star formation properties derived from MaNGA data.Low-Σ 1 blue spirals will be used as a comparison sample.Throughout this paper, a flat ΛCDM cosmology with Ω m = 0.3, Ω Λ = 0.7, and H 0 = 70 km s −1 Mpc −1 is adopted.

Sample and Parameters
The samples used in this work are similar to the sample of blue spirals in Hao et al. (2019), but updated with the final data release of the MaNGA survey, which is part of the 17th data release of the Sloan Digital Sky Survey (SDSS; Abdurro'uf et al. 2022).They consist of 109 blue spirals within a stellar mass range of 10 10.5 -10 11 M e , under the assumption of a Chabrier initial mass function (IMF; Chabrier 2003), and a redshift range of 0.02-0.05.These galaxies were drawn from a parent sample in Mendel et al. (2014) with 0.02 < z < 0.05, which is included in the legacy area of SDSS DR7.The morphological information was retrieved from the Galaxy Zoo 1 (Lintott et al. 2008(Lintott et al. , 2011)), and the galaxies were separated in color by their locations in the dust-corrected u − r color versus stellar mass diagram.<-+ was used to single out the blue galaxies.Readers can find more details of the sample selection in Hao et al. (2019) and Guo et al. (2020).In this work, to further distinguish blue spirals with dense cores from those with less dense cores, we use Σ 1 = 10 9.4 M e kpc −2 as a dividing point. 7This yields 33 high-Σ 1 and 76 low-Σ 1 blue spirals.The stellar mass and Σ 1 distributions for the two types of blue spirals are shown in Figure 1.To evaluate the reasonability of our adopted threshold for separating high-Σ 1 from low-Σ 1 galaxies, we examined our results by dividing the blue spirals evenly into three subsamples, which correspond to a log Σ 1 range of <9.2, 9.2-9.4 and >9.4.Interestingly, the statistical properties of the galaxies within the middle bin are also in between those of the first and last bins, but are much more similar to the low-Σ 1 bin in most cases.Therefore, the results based on the samples selected using our adopted dividing point should be robust.
To investigate the global star formation properties of our sample galaxies, we drew the total stellar masses and star formation rates (SFRs) from Mendel et al. (2014) and Salim et al. (2018), respectively.Both parameters were derived using spectral energy distribution fitting.The stellar masses are based on photometry in the five SDSS u, g, r, i, z bands, while the SFRs are based on the combined UV to IR photometry from the Galaxy Evolution Explorer, SDSS, and the Wide-field Infrared Survey Explorer.Both stellar masses and SFRs were derived under the assumption of a Chabrier IMF.
For the purpose of investigating the possible formation channels of high-Σ 1 blue spirals, we derive azimuthally averaged radial profiles of stellar population and star formation properties and the kinematic asymmetry of the gas content from the two-dimensional maps.To measure the azimuthally averaged radial profiles for each individual galaxy, the r-band elliptical Petrosian 50% light radius (R e ), the axis ratio (b/a), and the position angle used for elliptical apertures were extracted from the NASA-Sloan Atlas catalog (Blanton et al. 2011).A radial bin size of 0.15 R e was adopted.
Following Hao et al. (2019), we derived stellar population properties, measured by mass-weighted and luminosityweighted age and metallicity, from the MaNGA-Pipe3D value-added catalog (VAC) for DR17 (Sánchez et al. 2022).Compared to DR15 (Sánchez et al. 2016(Sánchez et al. , 2018)), the authors adopted an improved version of the fitting code (Lacerda et al. 2022) and a new simple stellar population (SSP) spectral library based on the MaNGA stellar library (MaStar; Yan et al. 2019) in DR17.Both versions of the Pipe3D VACs adopted a Salpeter IMF (Salpeter 1955).For common objects in the 15th and 17th data releases, we carefully compared the results from these two versions and found that not only the absolute values, but also the radial profiles of the stellar population properties of blue spirals changed from DR15 to DR17.Only the luminosityweighted age is almost immune to the changes in different versions of Pipe3D VAC.Therefore, we only use the luminosity-weighted age and the model-independent parameter Mgb/〈Fe〉 (=Mgb/(0.5* Fe5270+0.5 * Fe5335)) to probe the stellar populations in this paper.Mgb/〈Fe〉 is an α-element enhancement indicator, which is widely used to measure the star formation timescale.The measurements of the metal absorption lines were taken from the MaNGA data analysis pipeline (DAP; Westfall et al. 2019).By examination, we found that there are 11 high-Σ 1 and 24 low-Σ 1 blue spirals without reliable Fe measurements, because of the contamination of strong night-sky emission lines.Hence, they will not be included in the analysis of Mgb/〈Fe〉 profiles.
To study the status of the interstellar medium (ISM) of our sample galaxies, we distinguished star-forming regions from composites and active galactic nuclei  (Belfiore et al. 2019;Westfall et al. 2019).Only the emission lines with signal-to-noise ratio (S/N) greater than 3 were used.We corrected the Hα emission for internal dust attenuation based on the Balmer decrement Hα/Hβ and the O'Donnell (1994) reddening law.An intrinsic value of 2.86 was adopted for Hα/Hβ.For spaxels with Hα/Hβ below the intrinsic value, no dust attenuation correction was applied.The SFRs were calculated from the dust-corrected Hα luminosities, using a conversion factor of 8.77 × 10 −42 , calibrated using STARBURST99 (Leitherer et al. 1999(Leitherer et al. , 2010(Leitherer et al. , 2014;;Vázquez & Leitherer 2005) and under the assumption of a Salpeter IMF, following Kennicutt et al. (2009).To obtain the radial profile of the sSFR (=SFR/M * ), we drew the stellar mass surface density from the Pipe3D VAC, and used the ratio of the total SFR to the total stellar mass contained within each elliptical annulus as a measure of the sSFR at the corresponding radius.For the spaxel-based deprojected stellar mass and SFR surface density, we followed Barrera-Ballesteros et al. (2016) to correct for the inclination effect (Σ=Σ obs × (b/a)).We note that we adopted a Chabrier IMF for the sample selection and the study of the global properties inherited from our previous work, and we assumed a Salpeter IMF for the analyses of the spatially resolved stellar population and star formation properties for the sake of using the Pipe3D VAC.But this will not affect our results and conclusions, considering that only a constant conversion factor needs to be applied to convert the stellar mass and SFR from one IMF to the other.
The kinematic asymmetry was obtained by fitting the velocity field of the ionized gas using the KINEMETRY package (Krajnović et al. 2006), following Feng et al. (2020).Specifically, the average kinematic asymmetry within 1 R e (v asym ¯) was used to characterize the entire galaxy.Because of the low S/N of their Hα emission lines, we cannot obtain reliable estimates of v asym ¯for 10 high-Σ 1 and 17 low-Σ 1 blue spirals, as judged by their larger relative uncertainties, v 0.01 asym d > (Feng et al. 2022).Therefore, they were removed from the sample in the statistics of v asym ¯.

Results
Before diving into the spatially resolved properties, we first examine the global star formation properties of high-and low-Σ 1 blue spirals.The left panel of Figure 2 shows the locations of the blue spirals and their parent sample in the SFR versus M * diagram.To better define the star formation main sequence, we only include galaxies with star-forming centers in the parent sample in Figure 2, as classified using their central 3″ diameter SDSS fiber spectra.We do not fit a relation to the data points, since the best fit depends both on the sample selection and on the fitting method.Instead, we look at the distributions directly.It is obvious that both low-and high-Σ 1 blue spirals occupy the massive end of the distribution of their parent sample, and they do not deviate from the sequence.This is further confirmed by the right panel of Figure 2, which shows the histograms of sSFR for the two types of blue spirals and the parent sample with star-forming centers and 10 10.5 < M * < 10 11 M e .Although the high-Σ 1 blue spirals consist of a larger fraction of galaxies with higher sSFR than the low-Σ 1 blue spirals and the parent sample, they are not too different from them, as indicated by the Kolmogorov-Smirnov (KS) test.The probability that high-and low-Σ 1 blue spirals are drawn from the same distribution is 3.5%, while the KS test yields a probability of 7.9% for high-Σ 1 blue spirals and the parent sample.The standard deviations of sSFR for high-Σ 1 and low-Σ 1 blue spirals and the parent sample are 0.25, 0.27, and 0.27, respectively.
To shed light on the possible origins of high-Σ 1 blue spirals, we need to look into the spatially resolved properties.

Stellar Populations
We first investigate the luminosity-weighted age, which is sensitive to newly formed stars.Adopting the widely used method in studies using IFU data, we plot the median radial profiles and their 16% and 84% percentiles of the luminosityweighted age for the two types of galaxies in Figure 3.Both the median and the 16% and 84% percentiles indicate that the two types of blue spirals are similar in the inner age profiles, but high-Σ 1 blue spirals show older stellar populations than low-Σ 1 blue spirals toward the outer parts of galaxies.More strikingly, both high-and low-Σ 1 blue spirals show a broad age distribution across the probed radius, as indicated by the spread of the shadows in Figure 3.This suggests a large diversity in the age profile within each population.Therefore, we examine the age profiles one by one to both understand the whole population and reveal the true difference between highand low-Σ 1 blue spirals.It turns out that the age profiles can be roughly divided into three categories for both types of blue spirals, as illustrated by the three examples in Figure 4. From left to right, the panels show a declining profile, an increasing profile, and a profile with an age peak at some intermediate radius, respectively.The galaxies with the third type of age profiles are actually the "turnover" galaxies, as identified and named by Lin et al. (2017).They pointed out that the turnover feature is closely linked with the bar structure (Lin et al. 2017(Lin et al. , 2020)).We also find that the majority of our "turnover" galaxies are barred.Evidently, galaxies with the latter two kinds of profiles, i.e., increasing and "turnover" radial profiles,  + /76) for the sample of low-Σ 1 blue spirals.The errors represent the 1σ binomial confidence limits, derived using the method of Cameron (2011).It is obvious that low-Σ 1 blue spirals consist of a larger fraction of objects with older centers than high-Σ 1 galaxies.
Gradients of physical parameters are often used to quantify their radial profiles.In general, the gradients are derived by fitting a straight line to the radial profile of the parameter.In the case of luminosity-weighted age, it would be done by a linear fit to the logarithm age versus radius.However, as can be seen from Figure 4, a linear model is not a proper representation of the data here, especially when considering the whole radius range, which is essential for our purpose of identifying possible rejuvenated galaxies.By examining the individual age profiles, we find that a linear fit to the entire radial profile until 1.5 Re for the increasing and decreasing profiles and a linear fit to the inner 0.5 Re profile for the "turnover" galaxies can separate the galaxies with old centers from those with young centers in most cases.Hence, we perform error-weighted regressions for each individual age profile.During the fitting, the standard deviations of the luminosity-weighted age within the elliptical annuli are taken as the errors of the age.A fit to the inner 0.6 Re profile for the "turnover" galaxies does not change the sign of the gradient.Figure 5 shows the derived age gradients.It is  clear that high-Σ 1 blue spirals tend to show more positive slopes than the low-Σ 1 blue spirals, which results from the larger fraction of high-Σ 1 blue spirals with younger centers compared to low-Σ 1 ones.According to the age gradients, the numbers of galaxies with negative (positive) inner profiles are 17 (16) for high-Σ 1 and 55 (21) for low-Σ 1 blue spirals, respectively.They agree with the visual classification based on the whole radial profiles presented above, confirming the population difference between high-and low-Σ 1 blue spirals.
The model-independent parameter Mgb/〈Fe〉 is a powerful probe of the star formation timescale (Thomas et al. 2005).It is generally believed that the earlier a galaxy formed, the shorter the star formation timescale was.A shorter (longer) star formation timescale leads to a larger (smaller) Mgb/〈Fe〉.Therefore, a positive correlation between age and Mgb/〈Fe〉 is expected.Because of the large scatters, we opt not to show the radial profiles of Mgb/〈Fe〉.Instead, we use the same method to derive the gradients of the Mgb/〈Fe〉 profiles as for the age gradients.Figure 6 plots the gradient of the Mgb/〈Fe〉 profile as a function of the gradient of luminosity-weighted age.The horizontal and vertical dashed lines separate galaxies with negative slopes from those with positive slopes.It is obvious that the one-to-one correspondence between old (young) age and high (low) Mgb/〈Fe〉 is not perfect, with the consistency percentages of ∼ 60% (13/22) and ∼70% (36/52) for highand low-Σ 1 blue spirals, respectively.Interestingly, the remaining ∼40% of the high-Σ 1 blue spirals and ∼30% of the low-Σ 1 blue spirals without the age-Mgb/〈Fe〉 correspondence mostly include galaxies that host old centers with low Mgb/〈Fe〉, i.e., those with negative age gradient and positive Mgb/〈Fe〉 gradient.In other words, for almost all galaxies with younger centers, their central Mgb/〈Fe〉 is low.This may also apply to the galaxies without reliable Fe measurements.The expected fractions of high-and low-Σ 1 blue spirals with positive Mgb/〈Fe〉 gradient would be about 16 33 , respectively, as deduced from the age gradients.The association between positive age gradient and positive Mgb/ 〈Fe〉 gradient implies that there is new star formation in the inner regions of these galaxies.The newly formed stars from the iron-enriched gas by Type Ia supernovae could lower the Mgb/〈Fe〉.Alternatively, if the galaxies have experienced multiple rejuvenations, the lower Mgb/〈Fe〉 could have been produced by previous rejuvenations, which took place late enough to allow the Type Ia supernova explosion.The galaxies with old and low Mgb/〈Fe〉 centers probably had experienced rejuvenations at earlier times.

Star Formation Properties
By selection, both high-Σ 1 and low-Σ 1 blue spirals belong to the "blue cloud" in the dust-corrected u − r versus M * diagram, and they show similar global sSFR distributions, as shown in Figure 2. In this subsection, we study their ISM status and evaluate the star formation intensity in a spatially resolved way.
We use the fractional area of star formation spaxels for our samples of blue spirals to quantify their ISM status.In the left panel of Figure 7, we present the histogram distributions of the total fractional area of star formation spaxels for the two samples of blue spirals.It is clear that more than 70% of low-Σ 1 blue spirals are forming stars throughout most (>70%) areas of the galaxies, whereas high-Σ 1 blue spirals show very extended and relatively flat distributions.We find that the spaxels that are not identified as star-forming regions are mostly composite regions, except for a small fraction of spaxels with low S/N.The total areal fraction of composition regions is quantified and plotted in the right panel of Figure 7.It shows clearly that compared to low-Σ 1 blue spirals, a much larger fraction of high-Σ 1 blue spirals have more than 20% and higher Figure 6.Mgb/〈Fe〉 gradient as a function of luminosity-weighted age gradient.Low-and high-Σ 1 blue spirals are represented by blue triangles and cyan squares, respectively.The horizontal and vertical dashed lines separate galaxies with negative gradients from those with positive gradients.Galaxies with "turnover" age profiles are denoted with filled symbols.The error bars shown in the bottom right corner represent the median measurement errors for low-Σ 1 (blue) and high-Σ 1 (cyan) blue spirals, respectively.fractions of areas dominated by composite regions.To explore whether there is any association between the age profile and the ISM status, we plot the age gradients versus the areal fraction of the composite regions in Figure 8. "Turnover" galaxies were pointed out by solid points.As stated above, most of them are barred galaxies.Generally, composite regions are either in galaxy centers or at the interface between central AGNs and outer star-forming regions.To estimate the influence of such cases, we label galaxies with composite or AGN centers using little green and red dots, respectively.It is interesting to see that most galaxies with positive age gradients, i.e., young centers, have larger fractions of composite regions regardless of the central compactness of blue spirals, and more than half of them do not have a composite or AGN center.By contrast, for the galaxies with old centers, i.e., negative age gradients, only a small fraction of low-Σ 1 blue spirals and about half of the high-Σ 1 blue spirals have larger fractions of composite regions, and they mostly occur in galaxies with composite or AGN centers.Therefore, apart from the general cases, composite regions are mainly associated with barred galaxies and galaxies with positive age gradients.The two-dimensional BPT diagrams reveal that composite regions mostly appear around Figure 8.Total fractional area of spaxels with "composite" spectral features as a function of luminosity-weighted age gradient.The blue triangles and cyan squares represent low-and high-Σ 1 blue spirals, respectively.The vertical dashed line separates galaxies with positive age gradients from those with negative age gradients.Galaxies with "turnover" age profiles are denoted with filled symbols.The small red and green dots in the middle of the symbols indicate galaxies hosting central AGN and composite spectral features, respectively.The error bars in the bottom right corner are the median measurement errors for low-Σ 1 (blue) and high-Σ 1 (cyan) blue spirals.
bar regions in barred galaxies or in the outer regions of galaxies with rising age profiles.This suggests that composite regions are linked with old stellar populations, and the star formation is more concentrated toward the central regions.In comparison with low-Σ 1 blue spirals, the higher fraction of high-Σ 1 blue spirals with larger areal fractions of composite regions is a result of its greater fraction of members with younger centers.
We adopt the most commonly used quantity, sSFR, to evaluate the star formation activity, which is also a simple tracer of the star formation history.From the lessons that we learned from the radial profiles of the luminosity-weighted age and Mgb/〈Fe〉, we expect a large diversity in the sSFR profile as well.Nonetheless, to gauge the overall star formation intensity, we plot the 50%, 16%, and 84% of the sSFR profiles for both types of blue spirals in Figure 9.It can be seen that for both types of galaxies, the sSFR is less than ∼10 −10 yr −1 , which indicates that they are not forming stars in a burst mode.A one-by-one examination of the sSFR and luminosityweighted age profiles shows that they have very good correspondence, in the sense that a higher sSFR usually corresponds to a younger age.Corresponding to the three types of age profiles shown in Figure 4, there are also three types of sSFR profiles, i.e., a profile with low sSFR in the center, a profile with high sSFR in the center, and a profile with an sSFR valley at some intermediate radius (i.e., "turnover" galaxies).Specifically, the percentages of galaxies with these types of sSFR profiles are 61.8 % 5.8 5.2 -+ , 7.9 % for high-Σ 1 blue spirals, respectively.These fractions are not exactly the same as those derived from the age profiles, because the correspondence between age and sSFR profiles is not perfect, which is also demonstrated by the anticorrelation between age gradients and sSFR gradients presented in Figure 10.For high-Σ 1 blue spirals, ∼88% (29/33) of the sample galaxies show consistent age and sSFR profiles, and for low-Σ 1 blue spirals, the fraction is ∼79% (60/76).
We note that in the measurement of the sSFR profile, the sSFR at each radius is represented by the ratio of the total SFR to the total stellar mass within the corresponding elliptical annulus.In essence, it is an average over the elliptical annulus.Therefore, the sSFR profiles shown in Figure 9 represent an overall star formation property across the galaxies, which probably conceal some regions with starburst features.Therefore, we derive the spatially resolved relation between the SFR surface density and the stellar mass surface density to quantify the star formation intensity on each spaxel.
Figure 11 shows the spatially resolved star formation mainsequence relations, i.e., the relations between the stellar mass surface density and the SFR surface density.The contours represent the distribution of the star-forming spaxels in low-Σ 1 blue spirals.The outermost contour includes 95% of the spaxels.The star-forming spaxels of high-Σ 1 blue spirals are represented by the data points color-coded by their spatial locations in radius.It is obvious that the majority of the spaxels in high-Σ 1 blue spirals follow a similar distribution to the low-Σ 1 blue spirals, except the high-density end.The colorcoded data points show a stratified distribution, with the inner spaxels hosting higher stellar mass surface density and SFR surface density.In spite of the higher values of SFR surface densities in the inner regions, the data points do not deviate from the constant sSFR of 10 −10 yr −1 , which roughly serves as a valid estimate for the low-Σ 1 blue spirals represented by the contours.Therefore, the spatially resolved measurements of the star formation intensity provide a consistent result with the azimuthally averaged measures shown in Figure 9, in the sense that the star formation mode in high-Σ 1 blue spirals is not significantly different from low-Σ 1 blue spirals.
Figure 9. Radial profiles of sSFR for high-Σ 1 (cyan) and low-Σ 1 blue spirals (blue).The sSFR within each elliptical annulus was derived from the total total stellar mass within the respective elliptical annulus.The solid curves and the shaded regions indicate the median, 16%, and 84% of the distributions, respectively.

Discussion
In this paper, we compare the stellar population and star formation properties of high-Σ 1 blue spirals with those of low-Σ 1 blue spirals.The purpose is to understand the origin of the high-Σ 1 blue spirals.As mentioned in Section 1, Guo et al. (2020) found that high-Σ 1 blue spirals are similar to red spirals in many aspects, e.g., bulge-to-total-mass ratios, dark matter halo masses, and detailed morphological features.Accordingly, Figure 10.sSFR gradient as a function of luminosity-weighted age gradient.Low-and high-Σ 1 blue spirals are represented by blue triangles and cyan squares, respectively.The horizontal and vertical dashed lines separate galaxies with negative gradients from those with positive gradients.Galaxies with "turnover" age profiles are denoted with filled symbols.The error bars shown in the bottom right corner represent the median measurement errors for low-Σ 1 (blue) and high-Σ 1 (cyan) blue spirals, respectively.they suggested that high-Σ 1 blue spirals are rejuvenated systems from quenched red spirals.Fang et al. (2013) also proposed a rejuvenation origin for high-Σ 1 blue spirals, and they speculated that the rejuvenated star formation mostly occurs in the outer regions.On the other hand, Woo & Ellison (2019) suggested that a compaction-like process may contribute to the formation of the dense core.
The results presented in Section 3 reveal that both low-and high-Σ 1 blue spirals show large diversities in stellar populations and star formation properties.But they can be broadly divided into two types: one type includes galaxies with decreasing age profiles, i.e., old centers, and the other comprises galaxies with younger centers.Compared to low-Σ 1 blue spirals, high-Σ 1 blue spirals consist of a larger fraction of galaxies with a younger, less α-element-enhanced center, and an older, more α-element-enhanced disk (∼55% versus ∼29%).Furthermore, such galaxies generally show higher sSFR in the central regions and lower sSFR in the outer disks.However, both the azimuthally averaged sSFR profiles and the spatially resolved star formation main-sequence relation show that these central star formation activities are not starbursts.The relatively low sSFRs (∼10 −10 yr −1 ) may imply that the current star formation activities are not responsible for the buildup of the central dense cores, which formed at an earlier epoch.A rejuvenation scenario is able to explain the properties of these high-Σ 1 blue spirals with younger centers.In such a scenario, new gas inflow and star formation triggered by some mechanisms take place in the quenched galaxies that formed via a fast process at an early epoch.The gas inflow and star formation have been carrying on across the galaxies, but are more concentrated on the central regions, which lowered the luminosity-weighted age and Mgb/ 〈Fe〉 of the stellar populations.
We next look for possible triggering mechanisms to verify the rejuvenation scenario.Rejuvenation must be associated with some physical processes that can cause gas inflow and the re-ignition of star formation.It is straightforward to search for signs of such processes using asymmetry in morphologies and gas velocity fields.
We retrieve morphological information from the Galaxy Zoo 2 (GZ2; Willett et al. 2013;Hart et al. 2016) and adopt the suggested mean vote fraction of 0.5 as the threshold (Hart et al. 2016).We also obtain the information on bar and tidal features from the MaNGA visual morphology catalog (A.Vazquez-Mata et al. 2022, 2024, in preparation).This catalog contains a visual morphological classification based on the inspection of image mosaics created using new digital processing of SDSS and Dark Energy Legacy Survey (Dey et al. 2019) images.This new processing enables the identification of internal structures and low-surface-brightness features.After comparing with the GZ2 results, we find that for the classification of barred galaxies, if we adopt 0.75 as the threshold for A. Vazquez-Mata et al. (2024, in preparation), i.e., count galaxies with clear conspicuous bars, the two catalogs produce consistent results.Furthermore, the identification of tidal features in A. Vazquez-Mata et al. (2024, in preparation) is also in good agreement with the GZ2 classification for disturbed morphologies.Since GZ2 includes a much larger sample than the MaNGA visual morphology catalog, we adopt the GZ2 results for the following analysis to facilitate future comparison studies.For our samples of galaxies, the bar fractions for high-and low-Σ 1 blue spirals are ∼50% and ∼30%, respectively.The fractions of galaxies with tidal features are similar in the two types of galaxies, ∼15%.These fractions are more or less consistent with those based on the parent samples, as presented in Guo et al. (2020).The association between bar structures and age profile types is shown in Figure 12.It is clearly seen that Figure 12.Fraction of barred galaxies as a function of the type of luminosity-weighted age profile for low-Σ 1 (blue) and high-Σ 1 (cyan) blue spirals.The x-axis represents the type of the luminosity-weighted age profile, number-coded by 1, 2, and 3. 1 stands for a descending age profile, 2 stands for an increasing profile, and 3 stands for the age profile with a peak (i.e., "turnover" feature).The error bars represent the 1σ binomial confidence limits, based on the method of Cameron (2011).galaxies with a "turnover" age profile are mostly barred galaxies for both types of spirals, which was also mentioned in Section 3.1.Half of the high-Σ 1 blue spirals with an increasing age profile are unbarred, though.We examined these galaxies and found that they show either tidal features or rings,8 which are probably the causes of the rejuvenation.
The kinematic asymmetry in the Hα velocity field can probe the perturbance suffered by a galaxy directly.We adopt the three intervals of the kinematic asymmetry parameter used in Feng et al. (2020) to characterize our samples of galaxies.Specifically, the three intervals of v 0.007 0.027 asym < < , v 0.027 0.041 asym < < , and v 0.041 0.316 asym < < correspond to low, medium, and high asymmetry in the Hα velocity field, respectively.We then evaluate the association between kinematic asymmetry and the luminosity-weighted age profile.Among the 12 high-Σ 1 blue spiral galaxies with medium to high asymmetry, excepting the two galaxies with the lowest asymmetry, the other 10 show either increasing age profiles or "turnover" features in the age profiles.The other way around, we also examine the fraction of galaxies with medium to high asymmetry as a function of the age profile type, as shown in Figure 13.It is obvious that the galaxies with decreasing age profiles have the lowest fractions of galaxies with kinematic asymmetries, and more than 75% (60%) of the high-(low-) Σ 1 blue spirals with increasing age profiles or "turnover" features show kinematic asymmetries.This suggests that the asymmetries are closely linked with the age profiles for both types of blue spirals.By examination, we find that barred galaxies show more kinematic asymmetry in the Hα velocity field, consistent with the finding of Feng et al. (2022).But a large fraction of the kinematic asymmetries are also produced by nonbarred galaxies, some of which show morphological disturbances or are involved in pair or group systems.In summary, optical morphologies and gas kinematics provide consistent results that blue spirals with young centers show much more asymmetry than those with old centers.Specifically, bars and galaxy interactions that disturbed the gas velocity fields play an important role in transporting gas inward for the blue spirals with younger centers.The effect of bars on gas inflow has been demonstrated in many studies (Yu et al. 2022 and references therein).We notice that these previous studies mainly proved the essential role of bars in the growth of pseudo-bulges, i.e., the secular evolution of disk galaxies.In this work, we find that bars are also able to drive gas inflow and trigger new star formation in massive disk galaxies with dense cores.
In such a gas inflow scenario, a depressed central gas-phase metallicity may be expected, which is often seen in galaxy pairs or mergers (e.g., Kewley et al. 2006;Ellison et al. 2008Ellison et al. , 2013;;Rupke et al. 2008;Peeples et al. 2009;Scudder et al. 2012;Guo et al. 2016).It has long been known that galaxy luminosity or stellar mass is positively correlated with the gas-phase metallicity on a global scale (e.g., Lequeux et al. 1979;Tremonti et al. 2004).In recent years, the development of IFU observations has enabled the confirmation of a similar correlation between stellar mass surface density and gas-phase metallicity on local scales (e.g., Rosales-Ortega et al. 2012;Sánchez et al. 2013;Barrera-Ballesteros et al. 2016).
Following the widely used method, we search for possible metallicity dilution in the blue spirals with younger centers, by studying the spatially resolved stellar mass surface density versus gas-phase metallicity relations in Figure 14.To obtain reliable oxygen abundance, the most popular probe for the gasphase metallicity, only [S II]/Hα-selected star-forming spaxels , as a function of the type of luminosity-weighted age profile for low-Σ 1 (blue) and high-Σ 1 (cyan) blue spirals.The x-axis represents the type of the age profile, number-coded by 1, 2, and 3. 1 stands for a decreasing age profile, 2 stands for an increasing profile, and 3 stands for the age profile with a peak (i.e., "turnover" feature).The error bars represent the 1σ binomial confidence limits, based on the method of Cameron (2011).
are used.The oxygen abundance was calculated using the O3N2 calibration presented by Equation (2) in Marino et al. (2013).The data points in the left and right panels of Figure 14 represent star-forming spaxels in high-Σ 1 and low-Σ 1 blue spirals, respectively.Both panels show that the data points in the central regions, represented by the blue color, have the highest-mass surface densities, but the majority of them do not show significantly lower gas-phase metallicities compared to the blue contours that represent the distribution of the entire population of low-Σ 1 blue spirals.In fact, the gas-phase metallicity is the result of the competition between gas inflow/ outflow, star formation, and supernova feedback.It is only when the metal-poor gas inflow is strong and the star formation does not last long enough to enrich the ISM via supernova explosion that the galaxy could show central metallicity deficiency (Montuori et al. 2010).Such metallicity deficiency is often seen in galaxy merging systems and evolves as the merger proceeds.Actually, even in galaxy pairs, both metallicity dilution and enrichment are found (Omori & Takeuchi 2022).In a study of CALIFA galaxies, the authors found that the oxygen abundances are only slightly lower in tidally perturbed galaxies than the control galaxies (Morales-Vargas et al. 2021).These studies confirm the complexity of the physical processes imprinted on the gas-phase metallicity.In the case of gas inflow triggered by galactic bars, the gasphase metallicity should be determined by similar physical processes, i.e., gas migration, star formation, and the related supernova explosion, etc., which deserves a separate detailed study.Therefore, the normal oxygen abundance shown in the central regions of high-Σ 1 blue spirals does not provide us with additional supportive evidence for gas inflow, but it does not conflict with the gas inflow scenario either.
On the other hand, we find that the properties of high-Σ 1 blue spirals with younger centers are also consistent with the fading counterparts of Luminous Infrared Galaxies (LIRGs) with spiral morphologies (R. Guo et al., 2024, in preparation).In the parallel study, we find that spiral LIRGs also show dense (i.e., high-Σ 1 ), younger, and less α-element-enhanced centers, but they host much more intensive star formation.A more detailed comparison between high-Σ 1 blue spirals and spiral LIRGs will be performed in a forthcoming paper.
For the remaining high-Σ 1 blue spirals, which show old centers and low sSFR in the center, we cannot distinguish their origins.Their properties are seemingly consistent with the inside-out growth scenario.Furthermore, they mostly do not show kinematic asymmetries in the Hα velocity field, as illustrated in Figure 13, although some of them have bars.However, we could not rule out a rejuvenation origin if some minor disturbances caused by a flyby event trigger a small amount of gas inflow and star formation on a galactic scale.

Summary and Conclusions
The stellar mass-Σ 1 relation followed by quiescent galaxies suggests a close connection between galaxy quenching and the buildup of a dense core.However, high-Σ 1 blue spirals do not seem to fit the picture of galaxy evolution drawn from this relation.They have assembled dense cores, but are still forming new stars.Based on the final data release of the SDSS-IV MaNGA IFU data, we make an effort to shed light on the origins of massive (>10 10.5 M e ) high-Σ 1 blue spirals at 0.02 < z < 0.05, by comparing the spatially resolved stellar populations and star formation properties between high-Σ 1 and low-Σ 1 blue spirals with similar mass and redshift.The main results are summarized below.
1.Both low-Σ 1 and high-Σ 1 blue spirals show large diversities in stellar population and star formation properties.The luminosity-weighted age profiles of blue spirals can be roughly divided into three categories, including a descending profile (i.e., an older center and younger disk), an increasing profile (i.e., a younger center and older disk), and a profile with a "turnover" feature (a younger center and younger disk with an age peak in between).Compared to their low-Σ 1 counterparts, high-Σ 1 blue spirals contain a larger fraction of galaxies with younger centers, including galaxies with either an increasing profile or a profile with "turnover" features (∼55% versus ∼30%).These galaxies generally show smaller Mgb/〈Fe〉 in the centers and larger Mgb/〈Fe〉 in the outer parts.2. Compared with their low-Σ 1 counterparts, high-Σ 1 blue spirals possess a greater portion of galaxies with relatively large areal fractions of [N II]/Hα-identified composite regions.The composite regions prefer surrounding areas of bars or the outer disks of galaxies with rising age profiles.This suggests that the star formation in these galaxies with "younger" centers is carrying on across the galaxy, but "pure star formation" is more concentrated in the central regions.3. The sSFR profiles show a good correspondence with the age profiles.Therefore, high-Σ 1 blue spirals are composed of a larger fraction of galaxies with higher central sSFRs than low-Σ 1 blue spirals (∼70% versus ∼40%).However, both the azimuthally averaged sSFR profiles and the spaxel-based star formation main-sequence relations show that the star formation activities in the central regions of high-Σ 1 blue spirals are not in a starburst mode.4. Both the optical morphologies and the kinematic asymmetries in the Hα velocity field are closely linked with the shapes of the luminosity-weighted age profiles.More than 75% of high-Σ 1 blue spirals with increasing age profiles or "turnover" features show kinematic asymmetries.Bars or galaxy interactions should be responsible for the asymmetries shown in the Hα velocity field.5. High-Σ 1 blue spirals that possess old centers and young disks do not show kinematic asymmetries, although some of them have bars.
The properties summarized above suggest that more than half of the high-Σ 1 blue spirals are likely to be rejuvenated systems.Bars or galaxy interactions drove the gas to flow into the galactic centers and triggered new star formation mostly in the centers and also in the disks.However, we caution that a direct identification of the rejuvenation from the stellar populations is challenging.Although Zhang et al. (2023) proposed a method, its applicability to real observational data still suffers from some difficulties.The remaining high-Σ 1 blue spirals have properties consistent with the inside-out growth scenario.

Figure 1 .
Figure 1.Stellar mass and Σ 1 distributions for our sample of blue spirals.Low-and high-Σ 1 blue spirals are represented by blue triangles and cyan squares, respectively.
(AGNs) in the [O III]/Hβ versus [N II]/Hα Baldwin-Phillips-Terlevich (BPT) diagram (Baldwin et al. 1981) developed by Kauffmann et al. (2003) and Kewley et al. (2001) on a spaxel basis.For the measurement of the SFR, we used the [O III]/Hβ versus [S II]/Hα diagram calibrated by Kewley et al. (2006) to single out star-forming regions.SFRs based on the [S II]/Hα- identified star-forming spaxels are similar to those based on a combination of star-forming regions and composites selected by the [O III]/Hβ versus [N II]/Hα diagram.The Milky Way dust-extinction-corrected emission-line fluxes were drawn from the MaNGA DAP

Figure 2 .
Figure 2. The global star formation properties of the two samples of blue spirals in terms of the star formation main-sequence relation (left) and the distribution of the sSFR (right).In the left panel, the two-dimensional histogram and the gray data points represent the parent sample with star-forming centers, while the blue triangles and cyan squares represent the low-and high-Σ 1 blue spirals, respectively.In the right panel, the colors of the histograms have the same meaning as those in the left panel, but the gray histogram only includes galaxies with 10 10.5 < M * < 10 11 M e in the parent sample for a fair comparison.

Figure 3 .
Figure3.Radial profiles of luminosity-weighted age for high-Σ 1 (cyan) and low-Σ 1 (blue) blue spirals, respectively.The solid lines show the median of the respective samples, and the shaded regions represent the 16% and 84% of the distributions.

Figure 4 .
Figure 4. Luminosity-weighted age profiles for three examples of high-Σ 1 blue spirals.The solid circles and the associated error bars are the median and standard deviation of the data within each elliptical annulus.From left to right, the three panels show three types of age profiles that are representative of blue spirals.

Figure 5 .
Figure 5. Distributions of the radial gradient of luminosity-weighted age for high-Σ 1 (top) and low-Σ 1 (bottom) blue spirals, respectively.The error bars represent the 1σ binomial confidence limits based on the method of Cameron (2011).

Figure 7 .
Figure7.Histograms of the total fractional area of star formation spaxels (left) and spaxels with "composite" spectral features (right) for the high-Σ 1 (top) and low-Σ 1 (bottom) blue spirals.The error bars represent the 1σ binomial confidence limits, following the method ofCameron (2011).

Figure 13 .
Figure13.Fraction of galaxies with medium to high asymmetry in the Hα velocity field, i.e., v 0.027 0.316 asym < < , as a function of the type of luminosity-weighted age profile for low-Σ 1 (blue) and high-Σ 1 (cyan) blue spirals.The x-axis represents the type of the age profile, number-coded by 1, 2, and 3. 1 stands for a decreasing age profile, 2 stands for an increasing profile, and 3 stands for the age profile with a peak (i.e., "turnover" feature).The error bars represent the 1σ binomial confidence limits, based on the method ofCameron (2011).

Figure 14 .
Figure14.Spatially resolved stellar mass surface density vs. gas-phase metallicity relations of star-forming spaxels for high-(left) and low-(right) Σ 1 blue spirals with younger centers, i.e., increasing or "turnover" age profiles.The data points are color-coded according to radial distance to galactic center.The blue contours represent the number density distribution drawn from all the star-forming spaxels in low-Σ 1 blue spirals.The contours enclose 5%, 30%, 68%, and 95% of the sample, respectively.