A spectral data release for 104 Type II Supernovae from the Tsinghua Supernova Group

We present 206 unpublished optical spectra of 104 type II supernovae obtained by the Xinglong 2.16m telescope and Lijiang 2.4m telescope during the period from 2011 to 2018, spanning the phases from about 1 to 200 days after the SN explosion. The spectral line identifications, evolution of line velocities and pseudo equivalent widths, as well as correlations between some important spectral parameters are presented. Our sample displays a large range in expansion velocities. For instance, the Fe ii 5169 velocities measured from spectra at t ∼ 50 days after the explosion vary from 2000 km s − 1 to 5500 km s − 1 , with an average value of 3872 ± 949 km s − 1 . Power-law functions can be used to fit the velocity evolution, with the power-law exponent quantifying the velocity decline rate. We found an anticorrelation existing between H β velocity at mid-plateau phase and its velocity decay exponent, SNe II with higher velocities tending to have smaller velocity decay rate. Moreover, we noticed that the velocity decay rate inferred from the Balmer lines (i.e., H α and H β ) have moderate correlations with the ratio of absorption to emission for H α (a/e). In our sample, two objects show possibly flash-ionized features at early phases. Besides, we noticed that multiple high-velocity components may exist on the blue side of hydrogen lines of SN 2013ab, possibly suggesting that these features arise from complex line forming region. All our spectra can be found in WISeREP and Zenodo.


INTRODUCTION
Supernovae (SNe) with prominent hydrogen lines in their optical spectra are classified as type II supernovae (SNe II) (Minkowski 1941;Filippenko 1997).It is generally accepted that SNe II are produced by explosion of massive stars with initial masses ≥ 8M⊙ (Van Dyk et al. 2003;Smartt et al. 2009;Smartt 2009).The characteristic of SNe II display a wide variety in photometric and spectroscopic features.Based on the spectral and photometric features, SNe II can be divided into subclasses of SNe IIP, IIL, IIb, and IIn etc.The subclasses of SNe IIP and IIL are primarily based on the shape of light curves.Those with almost constant luminosity (plateau) in light curves are called type IIP while those with a linear decline after maximun light are called type IIL (Barbon et al. 1979).Direct identifications of the progenitors of nearby SNe IIP, e.g.SN 2005cs (Maund et al. 2005;Li et al. 2006), SN 2012aw (Van Dyk et al. 2012;Fraser et al. 2012), SN 2017eaw (Kilpatrick & Foley 2018;Van Dyk et al. 2019;Rodríguez et al. 2020;Rui et al. 2019), et al. suggested that SNe IIP arise from red supergiant stars (RSG) with initial masses of 8-17 M⊙.Among SNe IIP, some have low expansion velocities and less amount of nickel synthesized in the explosion, which are thought to represent the low luminosity tail of a continuous distribution in the parameter space of SNe IIP and probably originate from intermediate-mass (10 − 15M⊙) stars (Spiro et al. 2014;Zhang et al. 2014).Although SNe IIP and SNe IIL have different lightcurve shapes, there is increasing studies showing that no significant distinction exists between the observational properties of these two subclasses (Anderson et al. 2014;Sanders et al. 2015;Valenti et al. 2016;Galbany et al. 2016;de Jaeger et al. 2019).Anderson et al. (2014) found that more luminous SNe II tend to have faster post-peak declines and this trend could be used to estimate distances to SNe II (i.e., the photometric color method, de Jaeger et al. 2015).Gutiérrez et al. (2014) also suggested that there is no definitive spectral distinction between SNe IIP and SNe IIL.More luminous SNe II tend to have larger Hα velocities and smaller ratios of absorption to emission (a/e) of Hα.Theoritical studies show that the plateaus in SNe IIP result from the combination of hydrogen envelope (Litvinova & Nadezhin 1983;Bartunov & Blinnikov 1992;Popov 1993;Morozova et al. 2015;Moriya et al. 2016), and the formation of SNe IIL can be due to that the exploding stars have less amount of hydrogen envelope (Anderson et al. 2014;Gutiérrez et al. 2017b).
Besides SNe IIP and SNe IIL, supernovae with long lasting narrow or/and intermediate-width emission lines of hydrogen in their optical spectra are classified as SNe IIn (Schlegel 1990).Different from 'normal' type IIP and type IIL SNe, circumstellar interaction (CSI) plays important roles in the observed properties of SNe IIn (Smith 2017).The spectra of some SNe II will evolve like SNe Ib a few weeks after the maximum light, which are called SNe IIb and are thought to be the transitional objects linking between SNe II with stripped-envelope SNe Ib (Filippenko et al. 1993).An even more rare subclass of SNe II is 1987A-like events, which shows unusually long rise to the peak (Hamuy et al. 1988;Xiang et al. 2023).We do not include the subclasses of SNe IIn, SNe IIb and 1987A-like events in this work.
With the wide-field, high-cadence transient surveys, the observed properties of both interesting individual objects and large samples of SNe II have been studied.Spectroscopic observations provide important information on the physical origins and even stellar winds blown from their progenitors when the spectra can be obtained in very early phases.For example, the flash-ionized signatures allow constraints on the density and velocity of circumstellar material (CSM), and therefore the mass loss history shortly before the SN explosion (Gal-Yam et al. 2014;Shivvers et al. 2015;Khazov et al. 2016;Yaron et al. 2017;Lin et al. 2021;Zhang et al. 2023).On the other hand, the spectra in the nebular phase can be used to estimate the progenitor mass (Maguire et al. 2012;Jerkstrand et al. 2012;Silverman et al. 2017).In this paper, we present data and analysis of 206 spectra for 104 SNe II obtained by Tsinghua Supernovae Group during the period from 2011 to 2018.We describe our sample in Section 2. The spectral evolution and line identifications are presented in Section 3. The evolution of expansion velocities and the pseudo equivalent width (pEW) of some important spectral lines are presented in Sections 4 and 5, respectively.The correlations between these spectrocopic parameters are given in Section 6, and we conclude in Section 7.

DATA SAMPLE
Our sample consists of 206 unpublished spectra for 104 SNe IIP/L observed between 2011 and 2018.The spectra were obtained using the Xinglong 2.16-m telescope (+BFOSC/OMR; Fan et al. 2016) of NAOC and the Lijiang 2.4-m telescope (+YFOSC; Fan et al. 2015) of Yunnan Observatories .All the spectra were reduced using standard IRAF pipelines (Tody 1986(Tody , 1993)), including bias and flat-field correction, cosmicray removal, wavelength and flux calibration.All spectra were corrected for atmospheric extinction using the extinction curves of the local observatories.Telluric absorptions were also removed for the spectra whenever possible.
For each SN, we used the same method following Gutiérrez et al. (2017a) to estimate their explosion dates.If highcadence observations are available for the sample near the discovery, we adopt the midpoints between the last nondetection and the first discovery as the estimated explosion date.The lower and upper limits of the estimated explosion date will cover the non-detection and first-discovery epochs.For the object with sparse or without pre-discovery images, the explosion dates were estimated by matching spectral templates via the Supernova Identification (SNID) code (Blondin & Tonry 2007).SNID automatically gives matches that satisfy the default rlap cut (rlapmin ≥ 5).In general, the best-fit templates are the first several templates that have the largest rlap values.The goodness of matching was also checked by eyes and we chose the best two from those rlap-ordered templates in most cases.The details of spectral matching and the plots with the best matches for each SN in our sample are shown in the supplementary materials.The explosion date is then taken as the average epoch of the two best-fit templates and the uncertainty is taken as the standard deviation of the average value.For those that the explosion dates can not be estimated from either the midpoint or the SNID fit, we have to treat the time of first discovery as the explosion dates1 .Among our 104 objects, 25 (24%) explosion epochs were obtained using midpoint, 35 (34%) were obtained using SNID template matching, while 25 (24%) have to use the time of first discovery as the estimated explosion time.It should be noted that some of our SN II samples have already been studied, and the explosion dates of this portion (about 18% of our sample) were taken from the literatures.The host galaxy redshifts of all our sample were taken from the NASA/IPAC extra galactic Database (NED) and the Galactic extinctions were from Schlafly & Finkbeiner (2011).The information of each SN II of our sample is listed in Table A1.
Figure 1 shows the statistical distribution of our sample, including the number of spectra, host-galaxy redshift, the phases of first spectrum and all spectra, respectively.A total of 206 spectra were collected for 104 objects.These spectra cover the phases from 1 days to 200 days after the explosion.Of the 104 SNe IIP/L, 37 objects have at least two spectra, while PS 15cwo has 11 spectra.A total of 20 objects (19%) were observed earlier than 5 days after the explosion and 46 object (44%) were observed earlier than 10 days after the explosion.The redshifts of these SNe II range from 0.00199 (SN 2012A) to 0.08 (SN 2017hxz), with a median value of 0.016, among which 78 objects (75%) are at hubble-flow distances (z > 0.01).

The spectral evolution of our sample is shown in
The evolution of two well-observed SNe IIL (PS15cwo and SN 2018aoq), and the spectral line identifications are shown in Figure 2. One can see that the spectra at early phases are characterized by P-Cgyni profile of Balmer series and He i lines.The He i 5876 gradually evolves into Na I D absorption as the temperature decreases.For example, in the spectra of SN 2018aoq, the P-Cgyni profile near 5850 Å disappeared at t ∼ 20.5 days after the explosion and it reappeared in the t ∼ 24.6 day spectrum, with the new emergent feature being due to the Na i D absorption.As the SN ejecta expands and gets cooler, features from the inner materials start to appear in the spectra.For instance, at t ∼ 1 month, the absorption feature of Fe ii lines near 4800 Å to 4900 Å, including Fe ii 5169, 5018, 4929, start to appear and get stronger with time.Later on, absorption features of other elements such as Sc ii 6247, 5663, Ba ii 6142, Ca ii near-infrared (NIR) triplet and O i 7774 become visible in the spectra, and absorption features of Ba, Ti and Sc are blended with Fe ii lines.
In the nebular phase, when the SN ejecta became transparent, the supernova is powered by the radioactive decay of 56 Co.At this late phase, the nebular spectra are dominated by emission lines due to recombination, collisional excitation and fluorescence (Branch et al. 2001).The prominent features include emission lines of Hα, [O i] 6300,6343, [Ca ii] 7291,7323 and Ca ii NIR triplet (see Figure 3).

High velocity features of hydrogen
A few of SNe II are found to exhibit an extra absorption component on the blue side of Ha lines.The notch could be the high velocity (HV) feature of Hα or absorption of other species.Gutiérrez et al. (2017a) suggested that this absorption feature observed before t∼35 days after explosion can be associated with Si ii 6355 line, while those observed after the middle of the plateau are HV features of hydrogen.There are two explanations proposed for the origin of the HV features.Baron et al. (2000) suggested that the blue side absorption can be due to the second P-Cgyni profile of Hα caused by non-LTE effects.They reproduced the second absorption of Hβ line in their synthetic spectra with the full non-LTE atmosphere code PHOENIX.On the other hand, the HV features were also regarded as signs of interaction between SN ejecta and the CSM (Chugai et al. 2007), which can be used to estimate the mass loss rate of the progenitor star.In theory, the X-rays from the reverse shock can ionize and excite the outer layers of the ejecta, producing a shallow depression in the blue wings of absorption part of P-Cgyni profile (Chugai et al. 2007).Moreover, the HV features can also result from cold dense shell, and such an absorption is usually deeper (Chugai et al. 2007).Owing to the observed diversity of absorption shape, Gutiérrez et al. (2017a) suggested that these HV features are most likely produced by interaction.
For our sample presented in this work, we examined the blue part of Hα in each spectrum and consider the notch as HV feature of Hα if there is a similar absorption in Hβ line profile.The spectra with clear HV features are shown in Figure 4, where one can also see that these HV features show significant diversities.The notches seen in SN 2012A and SN 2016B are shallower, while those seen in SN 2012ec, SN 2012fs and SN 2018aoq are deeper.As CSM interaction can produce different shapes of HV features, we suggested that CSM interaction be more likely to produce the HV features.
We noticed that SN 2013ab likely showed two HV features in t ∼ 76 day spectrum, with the absorption minma locating at 9000 km s −1 and 6500 km s −1 , respectively.We further examed its t ∼ 75 day spectrum from Silverman et al. (2012) and found that the Hβ did have two distinct absorptions, though the corresponding absorption in Hα is very weak at 6500 km s −1 and we were not sure if this absorption really existed.Nevertheless, multiple notches are found to exist in other individuals.For instance, SN 2001X seems to have two notches detected in the same spectra (see Figure 18 in Faran et al. 2014a).However, Faran et al. (2014a) noticed that such an identification of HV feature is dubious as the velocities of Hα and Hβ do not match.Another sample is SN 2009N, where two absorption features next to Hα (i.e., an absorption with a velocity of about 8000 km s −1 and another absorption with a lower velocity) were detected at the same time in its spectra ranging from 62 days to 77 days after the explosion (Takáts et al. 2014) (see their Fig 9).In SN 2009ip, four notches are found to embeded in the blue side of Hα and its polarization spectra reveal a possibly clumpy line forming region.It is possible that some notches shown in the spectra of these objects belongs to other species instead of hydrogen.Othewise, multiple HV features may indicates complex structure of line forming region which is possibly produced by interactions with asymmetric CSM.Khazov et al. (2016) estimated that ≥ 18% of SNe II tend to show ionization emission lines (flash ionized, FI) in their spectra if they are observed at sufficiently early times.Such FI features are formed by the recombination of the outermost CSM, which has been ionized by high-energy photons created during SN shock breakout (Gal-Yam et al. 2014).In our sample, narrow Hα emission is primarily used to identify the flash-ionized signatures as it is stronger and lasts for longer time.Ionization emission lines of other species, including helium, carbon, nitrogen and oxygen are also used whenever possible.As narrow emission lines of hydrogen can be easily contaminated by host-galaxy H II region, the objects with clear emission lines of host galaxy (e.g.[O iii] 5007, S ii 6717, 6730) but without emission lines of other species except hydrogen, are not assigned as SN II sample with FI features (e.g.SN 2016hvu, see Figure A-1).Finally, we identified possible FI features in two objects, i.e., SN 2013ac and SN 2016aqw, among the 49 objects with spectra taken within 11 days after the explosion, (see Figure 5).The narrow emission lines of He i 5876, 7065 were observed in SN 2013ac and SN 2016aqw.In addition, the He ii 4686 line (may blend with C iii/N iii) can be observed in the spectra of SN 2016aqw.Lorentzian function is used to measure the full width at half maximum (FWHM) of Hα line profile, with the FWHM being ∼ 1000 km s −1 and ∼ 800 km s −1 for SN 2013ac and SN 2016aqw, respectively (see Figure A2).Note that for SN 2013fs, the FI feature can not be detected in our 5.4d spectrum, however, the emission features of O vi 3811, O v 5597, O iv 3410, N v 4604 and He ii 4686 were detected in spectra that taken at earlier epochs than we observed (Yaron et al. 2017).Broad emission (with FWHM ∼ 5000 km s −1 ) has been detected at the location of Hα in SN 2016fmt.However, a narrow component seems to be superposed on the broad component (see Figure A2) and a possible emission signature of C iv 5801 seems to be detected, suggesting that SN 2016aqw possibly has FI features.

Flashed-ionization Features
Besides the object with FI features, 19 of all 49 objects only show blue featureless spectra at early times (≤ 11 d).The phases of these spectra range from 1.1 days to 9.3 days after explosion.As the FI features usually disappear quickly, detection of such features require observations of SNe II within a few days after their explosions.SNe II with only blue featureless continuum probably produce no FI features at all or show FI signatures in the spectra at an epoch earlier than we observed.

EXPANSION VELOCITY
We measured the velocities of Hα, Hβ and Fe ii 5169 from the absorption minima of P Cgyni profiles in every spectrum where they can be clearly detected.At around the absorption minima, we choose a few regions with lower spectral flux density to measure the velocity.The average value of these measurements is used as the velocity and the standard derivation is used as the uncertainty.The results are shown in Figure 6 and tabulated in Table A3.For comparison, we also included the average velocity evolution from the sample of Gutiérrez et al. (2017a), the typical IIP SN 1999em (Elmhamdi et al. 2003;Takáts & Vinkó 2012), the high-luminosity type IIP SN 2012aw (Bose et al. 2013) and the low-luminosity SN IIP 2005cs (LLSNe) (Pastorello et al. 2009).In Figure 6, we see that all of our SNe except for SN 2016cok have velocities larger than SN 2005cs, which means that SN 2016cok could be a low-luminosity SN II.Kochanek et al. (2017) identified a pre-explosion counterpart to SN 2016cok in archival images of Hubble Space Telescope and they found that the initial mass of progenitor was most likely in the mass range of 8 ∼ 12 M⊙.Such a relative lower mass range is consistent with that expected for the progenitor stars of low-luminosity SNe.
As the SN ejecta expand homologously, thus the expansion velocities decrease with time approximately in a power-law fashion (Faran et al. 2014a;de Jaeger et al. 2019).We thus fit the velocity evolution with a power-law function (i.e., vHα = v 50 Hα (t/50) n Hα , v Hβ = v 50 Hβ (t/50) n Hβ , vFe = v 50 Fe (t/50) n Fe ) to estimate both the velocity at t = 50 days after the explosion and the power-law exponent.Power-law fitting can only be used for those having at least two velocity data points.In our sample, however, many SNe have only one spectrum.If their spectra were taken between 45 to 55 days after the explosion, the velocity measured by absorption minimum can be treated as v50 and they are included in our analysis to increase the number of statistics.The parameters of v50 and the power-law exponent are shown in Table A4.The velocity evolution of a part of SNe II and the fitting results are shown in Figure A3, and the distribution of these parameters are shown in Figure 7.At t ∼ 50 days after the explosion, the velocities have large varieties.For instance, the velocity of Hα ranges from 4500 to 10,000 km s −1 , with a mean value of 6904 ± 1336 km s −1 , while the velocities of Hβ range from 3000 to 10,000 km s −1 and the velocity of Fe ii 5169 range from 2000 to 5500 km s −1 .As seen from Figure A3, for each SN, the velocity of hydrogen is higher than that of iron at similar epoch.This can be explained as that the Balmer lines are formed at larger radii than the iron lines.Gutiérrez et al. (2017a) introduced ∆v(Hβ) (the mean velocity decline rate in a fixed phase range) to quantify the decline rate of Hβ vecloity.They found that SNe II with larger decline rates at early times continue to show such behaviors at late times.In our analysis, we use the power-law exponent (e.g., nHα, see Section 4) to describe the velocity decline rate.Larger absolute values suggest faster decline rates.As seen from Figure 7, the exponent measured from Hα velocity evolution ranges from -0.45 to -0.15 and has a smaller range compared with those inferred from Hβ and Fe ii 5169 velocities.For example, the exponent inferred from the Hβ velocity evolution has a range of -1.1 ∼ -0.15.Among our sample, SN 2018bek, SN 2013gd and SN 2011az are the three showing the fastest decline rate of Hβ velocity.The exponent derived from Fe ii 5169 velocity is found to vary from -1.1 to -0.2, with SN 2015V and SN 2016jfu are the two showing the fastest decline rates and SN 2011bi and PS15cwo showing the slowest decline rates.However, it should be pointed out that the power-law exponent was derived from only a few data points, which may suffer large uncertainties.Sometimes we may overestimate or underestimate the decline rate.For example, the n Hβ of SN 2013gd, nFe of SN 2015V and nFe of SN 2016jfu may be overestimated as shown in Figure A3.

PSEUDO-EQUIVALENT WIDTHS
The pseudo equivalent width (pEW) is used to describe the strength of spectral lines.We measured the pEW of the absorption part of Hα, Hβ and Fe ii 5169 lines, the emission part of Hα and the pEW ratio of absorption to emission component (a/e) of Hα.The measurement results of our sample and the average values from sample presented in Gutiérrez et al. (2017a) are shown in Figure 8.One can see that the measurements of our sample are consistant with those derived by Gutiérrez et al. (2017a), although it is complicated to determine by the underlying continuum of each line feature. 2he pEW of absorption component can reach at ∼ 100 Å , ∼ 80 Å , and ∼ 70 Å for Hα, Hβ and Fe ii 5169, respectively, while the emission component of Hα can reach at ∼ 300 Å in the first ∼140 days.As noted by de Jaeger et al. ( 2019) and Gutiérrez et al. (2017a), generally speaking, the absorption components of Balmer lines and Fe II lines tend to increase their strengths with time in the first one or two months and  this increase behavior becomes slowly at later phases.However, the pEW evolution shows diversities among different objects.For example, the absorption part of Hα in PS15cwo, SN 2013gd and SN 2018aoq start to decrease in strength at t ∼ 60-70 days after explosion, while this happened in SN 2016cok since t ∼ 15 days after the explosion.Diversities are also observed in the evolution of Hα emission component.The evolution shows an overall increase with time but several objects show a decrease trend at some point.The general tendency of a/e shows an increasing tendency at early times and it remains constant or shows slight decrease at late times.
The large diversities in the strength of the above lines among different SNe II may reflect diverse temperature evolution and progenitor metallicity (Gutiérrez et al. 2017a).

Correlations between velocity parameters
For these velocity parameters measured for our sample, we use the Pearson correlation coefficient (ρ) to examine their correlations.According to Evans (1996), the coefficients in the range of 0-0.19, 0.2-0.39,0.4-0.59,0.6-0.89 and 0.8-1.0represent no, weak, moderate, strong, and very strong correlations, respectively.The results are shown in Figure 9.For the velocities measured at 50 days after the explosion, we found that any two of v 50 Hα , v 50 Hβ and v 50 FeII5169 parameters show a strong positive correlation, with the Pearson correlation coefficient being larger than 0.8 (see the left three panels of Figure 9).However, for the evolution exponent nHa, n Hβ and nFe, we find no correlation between them, as shown in the middle panels of Figure 9. Besides, we found that the Hβ velocity tends to show a slower decline for SNe II with higher velocities.(see panel (f) of Figure 9).After excluding SN 2018bek, SN 2013gd and SN 2011az, which are the three showing the fastest velocity declines3 , the Pearson coefficient increase froms 0.586 to 0.660.
To test whether the above negative tendency really exists for SNe II, we further include some well-studied individual SNe II from the literature and those of Gutiérrez et al. (2017a) in the statistical sample.The velocity evolution and power-law fit for these well studied objects are shown in Figure A4 and the relevant parameters of these well-studied sample are also given in Table A4.The power-law fitting and the parameters for the sample of Gutiérrez et al. (2017a) are shown in Figure A5 and Table A5, respectively.For each spectral species, Pearson correlation coefficients were calculated for all the samples and for only the gold sample, respectively.(Those with error of power-law exponent less than 20% are selected as the gold sample and they are represented by solid markers in Figure 10.)As can be seen from Figure 10, the tendency that high-velocity SNe II have slower velocity gradients still remains for Hβ when including more objects.Similar tendency probably exists in Hα, with the Pearson coefficient being 0.45 for all sample and 0.65 for the well-observed SNe II from literature (after removing SN 2016bkv).Faran et al. (2014b) found that the velocities inferred from hydrogen lines of SNe IIL, especially Hβ, evolve more slowly than those of SNe IIP.In general, the velocities of SNe IIL are higher than those of SNe IIP, which means that SNe II with higher Hβ velocities also evolve more slowly in the samples of Faran et al. (2014b).Therefore, what they found is consistent with the negative correlation in Hβ revealed in this work.In comparison, the velocity decline rate of hydrogen derived in this work shows a continuous distribution and no distinction is found between the whole SNe II smaple.
Besides, a negative correlation possibly exists in Fe ii velocity and its velocity decay .However, we caution, that the negative correlation is not evident in the well-studied sample and the parameter distribution of Fe seems to have a larger scatter than that of hydrogen.Moreover, both Faran et al. (2014b) andde Jaeger et al. (2015) suggested that the Fe ii velocity of SNe IIL shows a similar evolution as that of SNe IIP (but with a significant scatter), which is not consistent with the negative correlation.Therefore, we can not conclude the correlation between Fe ii velocity and its decay rate.
As shown in Figure 10, the velocity parameters of some objects deviate from others.In some cases, the power-law results seem unreasonable due to bad spectral sampling (e.g., SN 2015V and SN 2016jfu).In other cases, the velocities evolve in distinct ways.The velocities of both SN 2016bkv (a low- luminosity SN II) and SN 2018zd show an increase at early times, which may be related to CSM interaction (Zhang et al. 2020), as shown in FigureA4.Moreover, the velocity parameters of SN 2016bkv seem to show significant differences from others.Nakaoka et al. (2018) also noticed that SN 2016bkv showed a slower velocity evolution when compared with other LL SNe.It is not clear why some LL SNe II deviate from the "normal" SN II sample in the v 50 − n relation.The evolution of velocity is related to the physical properties of the progenitor star and expansion itself.Since SN 2016bkv could be the possible electron-capture supernovae candidate (Hosseinzadeh et al. 2018),its distinct velocity evolution may be partly related with the possible electron-capture origin.

Correlations between pEW and velocity
The ratio of absorption part to emission part (a/e) of Hα can be used to describe the diversity of SNe II.Patat et al. (1994) and Schlegel (1996) proposed that the subclass of SNe IIL have shallower P-Cgyni profiles, and hence smaller a/e values.Whereas Gutiérrez et al. (2014) proposed that there is no definitive spectral distinction between SNe IIP and SNe IIL, and SNe II with smaller a/e ratios of Hα tend to have higher Hα velocities, more rapid post-peak declines in light curve, higher peak brightness, and shorter optically thick phase duration (defined as time from the explosion epoch through the end of the plateau phase.).Our spectral dataset also indicate that SNe II with higher Hα velocities have smaller a/e ratios.Moreover, the exponents of velocity decay inferred from hydrogen lines seem to have a moderate correlation with the a/e values (see Fugure 11), suggesting that the velocity evolution of hydrogen lines may be related to the hydrogen envelope.Faran et al. (2014b) proposed that the velocities of SNe IIL evolve more slowly than those of SNe IIP is due to that the hydrogen lines are formed in the outer thin layers of the ejecta rather than in a thick, gradually-exposed hy-  drogen envelope.The moderate correlation in hydrogen lines revealed by our sample (i.e., SN II with slower velocity decay of hydrogen lines tends to have a smaller a/e ratio) favours the above explanation and also favours the traditional consensus that those traditionally classified SNe IIL likely have smaller hydrogen envelope mass at SN explosion.
If SNe IIL are formed by single stars, less hydrogen envelope mass (larger mass loss) could point to higher progenitor mass than SNe IIP (Smartt 2009) or high progenitor metallicity at which mass loss is strong enough to remove much of the hydrogen envelope (Heger et al. 2003).Direct observations on progenitor stars of SNe IIL are rare and the estimated masses have large uncertainties.Nevertheless, Valenti et al. (2016) proposed that SNe IIL appear not to come from more massive progenitors than SNe IIP.Moreover, single stars with enough metallicity will be so massive and they tend to form blackholes by fallback with weak supernovae (Heger et al. 2003).If part of SNe IIL do not have larger progenitor mass than SNe IIP, then it is possible that they originate from binary system and the mass transfer in binaries may play important roles in stripping part of the hydrogen shell.In future works, we intend to discuss the lightcurve properties and the correlations between spectral and photometric parameters, as well as the SN environment, aiming to understand the SN progenitor and its explosion physics.

CONCLUSIONS
In this work, we present a compilation of the optical spectra of SNe IIP/IIL observed over the past decade by the Tsinghua Supernova group.The sample consists 206 unreleased spectra for 104 SNe II, covering the phases from ∼1d to ∼200d after the explosion.Among 49 objects with spectra obtained within 10 days after explosion, two objects (SN 2013ac and SN 2016aqw) are found to show possble flash-ionized emission lines, including narrow lines of Hα, Hβ,HeI 5876, 7065, and HeII 4686.Prominent HV features of hydrogen are detected in six objects and they show different line profiles.Two are shallower and three are deeper.Two HV features are found to likely exist in the blue side of both Hα and Hβ in the t ∼ 76.2d spectrum of SN 2013ab.Based on the diversity of line profiles and multiple components, we suggested that CSM interactions could be at least a partial origin of HV features.
The pEW of absorption component of Hα in our sample can reach at ∼ 100 Å and the emission component can reach at ∼ 300 Å in the first ∼ 140 days.For individual object, the pEW of Balmer and metal lines show large scatter in the evolution, but we do not find definitive distinctions to separate SNe IIP from SNe IIL.A power-law function is used to fit the velocity evolution in order to estimate the velocity at 50 days after the explosion; moreover, the power-law exponent can be used to describe the velocity decay rate at the same time.For our sample, we found that SNe II with higher velocities during plateau phase show slower velocity evolution for Hβ.Moreover, the velocity decay rate of hydrogen (i.e., nHα and n Hβ ) have a moderate correlation with the a/e of Hα.SNe II with smaller velocity gradient of hydrogen lines tend to have smaller a/e ratios as well, suggesting the progenitors of those SNe II having less amount of hydorgen kept in the stellar envelope before explosion.Photometric parameters, as well as SN environment, will be discussed in future works, in order to understand the observational diversity of SNe II and the mechanisms responsible for these diversities.

Figure 1 .
Figure 1.Distribution of SN-level and spectrum-level parameters.The upper-left panel shows the number of spectra per SN.The upper-right panel shows the redshift distribution of our sample.The lower-left and lower-right panels show the distribution of the phase of the first spectrum and all spectra, respectively.

Figure 2 .
Figure 2. Spectral evolution and line identifications of PS 15cwo and SN 2018aoq.Shaded regions of different colors represent different spectral lines, with green representing hydrogen, yellow representing He I or Na I, purple representing iron, barium and other metal elements, respectively

Figure 4 .
Figure 4. High Velocity features of hydrogen identified in our sample.Blue lines represent line profile of Hα while red lines represent that of Hβ.The high velocity notches are marked by green shadow region.HV features have different shape: the notch of SN 2012A and SN 2016B are shallower, while SN 2012ec, SN 2012fs, SN 2018aoq are deeper.We found two notches in the t ∼ 76.2d spectrum of SN 2013ab at 76.2d, with the absorption minima locating at 9000 km s −1 and 6500 km s −1 , respectively.The overplotted t ∼ 75d spectrum of SN 2013ab is taken from Silverman et al. (2012).

Figure 5 .
Figure 5.The SNe II of our sample showing possible flash ionized features in the early phases.Left panel: the original spectra and the best-fit blackbody (gray line).Right panel: the blackbody continuum-subtracted spectral features.Vertical dashed lines represent emission lines of different species.

Figure 6 .
Figure 6.Velocity evolution of Hα, Hβ and Fe ii 5169 lines of our SN II sample and some well-studied SNe II.The blue circles represent our sample, while the black solid line and dashed lines represent the average velocity evolution and 1 − σ standard deviation from the sample of Gutiérrez et al. (2017a).

Figure 7 .
Figure 7. Distribution of velocity measured at t ∼50 days after the explosion and the power-law exponents of Hα, Hβ and Fe ii 5169.The vertical dashed lines mark the average values.The total number and the average velocity (the unit of average velocity is km s −1 ) of each parameter are listed in each panel.

Figure 8 .
Figure 8. Temporal evolution of pEW of Hα, Hβ and Fe ii 5169.The blue dots and the gray shadow represent the average pEW value and 1σ standard deviation from the sample given by Gutiérrez et al. (2017a)

Figure 9 .
Figure9.Correlations between velocity evolution parameters inferred from Hα, Hβ, and Fe ii 5169 lines.v 50 represents the velocity derived from t ∼ 50d spectra, while n Hα , n Hβ and n Fe represents the power-law index determined by applying the power-law fits to the velocity decay from the Hα, Hβ and Fe ii 5169 absortpion features, respectively.

Figure 10 .
Figure 10.Similar as Fig 9, with well-studied individuals in literature included.Blue circles represent the data presented in this paper, while other well-studied sample are denoted by different symbols, with red symbols representing "typical" SNe IIP, green symbols representing fast-declining objects and yellow ones representing low-luminosity SNe II, respectively.The solid markers represent the gold sample which satisfies the criteria that the uncertainty is less than 20%, while the open markers represent other objects..

Figure 11 .
Figure 11.Correlation between pEW ratios of absorption to emission component (a/e) of Hα and the velocity power-law exponents.In the upper-right, lower-left, and lower-right panels, the gold sample (The definition of a gold sample is the same as Figure 10) are represented by solid markers, while other objects are represented by open markers.

Figure A- 1 .
Figure A-1.Spectral evolution of our SN II sample presented in this work.

Figure A3 .
Figure A3.The velocity evolution inferred from Hα (blue dots), Hβ (green dots) and Fe ii 5169 (red dots) lines in the spectra of our sample of SNe II.Overplotted are the power-law fitting to the observed velocity evolution.We must notice that the power-law exponent of our sample has larger errors, and this is due to bad spectral sampling.

Figure A5 .
Figure A5.Same as Figure A3 and Figure A4, but for the SN II sample in Gutiérrez et al. (2017a).

Table A2 :
Journal of Spectroscopic Observations of our SN II sample.Phase relative to the estimated explosion date estimated in Section 2.

Table A3 :
Line velocities and pseudo equivalent widths of Hα, Hβ and Fe ii 5169.
Line velocities and pseudo equivalent widths of Hα, Hβ and Fe ii 5169.(continue)

Table A6 .
pEW of Hα at 50 days after explosion