We present light curves and spectra of the tidal disruption event (TDE) ASASSN-18pg / AT 2018dyb spanning a period of seven months. The event shows a plethora of strong emission lines, including the Balmer series, He II, He I and metal lines of O III $\lambda$3760 and N III $\lambda\lambda$ 4100 and 4640. The latter lines are consistent with originating from the Bowen fluorescence mechanism. By analyzing literature spectra of past events, we conclude that these lines are common in TDEs. The spectral diversity of optical TDEs is thus larger than previously thought and includes N-rich besides H- and He-rich events. We study how the spectral lines evolve with time, by means of their width, relative strength and velocity offsets. The velocity width of the lines starts at ~ 12,000 km s$^{-1}$ and decreases with time. The ratio of H$\alpha$ to H$\beta$ remains close to three, while the ratio of He II over N III increases with time. The same is true for ASASSN-14li, which has a very similar spectrum to AT 2018dyb but its lines are narrower by a factor of $>$2. High-resolution spectroscopy at maximum light does not reveal any narrow features that can be attributed to the TDE. By fitting the light curves of AT 2018dyb we estimate a mass of $4^{+5}_{-2}\times 10^6 M_{\odot}$ for the black hole and of $0.7^{+4}_{-0.6} M_{\odot}$ for the disrupted star. The detection of strong Bowen lines in the optical spectrum is an indirect proof for extreme ultraviolet and (re-processed) X-ray radiation and favors an accretion origin for the TDE optical luminosity. A model where photons escape after multiple scatterings through a super-Eddington thick disk and its optically-thick wind, viewed at an angle close to the disk plane, is consistent with the observations. (Abridged)

© 2017. The American Astronomical Society. All rights reserved. We present the light curves of the hydrogen-poor superluminous supernovae (SLSNe I) PTF 12dam and iPTF 13dcc, discovered by the (intermediate) Palomar Transient Factory. Both show excess emission at early times and a slowly declining light curve at late times. The early bump in PTF 12dam is very similar in duration (∼10 days) and brightness relative to the main peak (2-3 mag fainter) compared to that observed in other SLSNe I. In contrast, the long-duration (>30 days) early excess emission in iPTF 13dcc, whose brightness competes with that of the main peak, appears to be of a different nature. We construct bolometric light curves for both targets, and fit a variety of light-curve models to both the early bump and main peak in an attempt to understand the nature of these explosions. Even though the slope of the late-time decline in the light curves of both SLSNe is suggestively close to that expected from the radioactive decay of 56Ni and 56Co, the amount of nickel required to power the full light curves is too large considering the estimated ejecta mass. The magnetar model including an increasing escape fraction provides a reasonable description of the PTF 12dam observations. However, neither the basic nor the double-peaked magnetar model is capable of reproducing the light curve of iPTF 13dcc. A model combining a shock breakout in an extended envelope with late-time magnetar energy injection provides a reasonable fit to the iPTF 13dcc observations. Finally, we find that the light curves of both PTF 12dam and iPTF 13dcc can be adequately fit with the model involving interaction with the circumstellar medium.