CHEMICAL ABUNDANCES OF PLANETARY NEBULAE IN THE SUBSTRUCTURES OF M31^*

We selected seven PNe located in or associated with the Northern Spur and the extension of the Giant Stream. Of the seven PNe in our sample, six were selected from the catalog of Merrett et al. (2006), and the other one was identified by the Large Sky Area Multi-Object Fiber Spectroscopic Telescope (LAMOST; Su et al. 1998) during its early commissioning phase (Yuan et al. 2010). Three PNe of our sample have been identified as being in the Northern Spur by Merrett et al. (2006), and the other four are kinematically associated with the Giant Stream. Their projected galactocentric distances range from 0.7 to 50.2 kpc. The coordinates (right ascension—R.A., and declination—decl.), apparent magnitudes m(λ5007), radial velocities, and sky-projected galactocentric distances (in kpc) are presented in Table 1. The six PNe selected from Merrett et al. (2006) are relatively bright, with m(λ5007) ∼20.4–21.3. If the bright Mercury line (Hg i λ4358) exists, the heliocentric velocity of M31 (∼−300 km s⁻¹) could make the [O iii] λ4363 auroral line in a PN spectrum difficult to resolve (e.g., Kwitter et al. 2012; Paper I), and thus a loss of important temperature diagnostic for this target. Given the excellent observing condition at ORM in La Palma, where the Mercury line is faint, target selection was not confined by radial velocities but mainly based on the brightness. The PN newly discovered by LAMOST, which is located ∼365 from the center of M31 (Figure 1), is of particular interest because it is both spatially and kinematically related to the Giant Stream, and is also the first PN discovered in the outer extension of this stream. These targets are hereafter referred to as PN1–7 (Table 1).

Zoom In Zoom Out Reset image size

Table 1.  Properties and Observing Log of the Seven PNe

PN IDa	R.A.	Decl.	m(λ5007)b	${v}_{{\rm{helio}}}$	${R}_{{\rm{gal}}}{}^{{\rm{c}}}$	GTC Exp.
	(J2000.0)	(J2000.0)		(km s−1)	(kpc)	(s)
PN1 (M2445)	0:45:52.8	42:36:52.9	20.65	−149.1	20.1	4 × 1250
PN2 (M2451)	0:46:59.3	42:37:58.2	20.81	−80.8	21.6	4 × 1250
PN3 (M2427)	0:46:26.5	43:00:43.0	20.48	−420.3	25.7	4 × 1250
PN4 (M77)	0:43:32.3	41:58:42.5	20.99	−579.0	9.93	4 × 1200
PN5 (M586)	0:43:07.9	41:28:45.5	20.82	−713.5	3.05	4 × 1200
PN6 (M2750)	0:42:40.3	41:19:07.4	21.23	−766.5	0.70	4 × 1200
PN7 (LAMOST)	0:49:29.0	37:50:45.4	21.51	−203.0	50.2	4 × 2400
						1 × 1200

Notes.

^aNumber in the bracket that follows the PN ID is ID number from Merrett et al. (2006) except PN7, which was newly discovered by LAMOST. ^bm(λ5007) = −2.5 $\mathrm{log}F(\lambda 5007)$ − 13.74. ^cSky-projected galactocentric distance.

Download table as:  ASCIITypeset image

Spectroscopy of the seven PNe was carried out with Optical System for Imaging and low-intermediate-Resolution Integrated Spectroscopy (OSIRIS) on the 10.4 m GTC at ORM in La Palma, Spain. The observations were obtained on seven nights, from 2014 August 23 to 29 in service mode. The excellent observing conditions at ORM allow photometric and clear nights with seeing 05–10. The 1''wide long slit and the grism R1000B were used. The R1000B is a 1000 lines mm⁻¹ grism centered at 5455 Å. The detector of OSIRIS consists of a mosaic of two Marconi CCDs (named CCD1 and CCD2) with 2048 × 4096 pixels each. The pixel physical size is 15 μm, which corresponds to a scale of 0127 on the sky. The OSIRIS standard observing mode uses 0254 binned pixels. This instrument setup enabled us to perform spectroscopy over the wavelength range ∼3630–7760 Å with a spectral resolution of ∼6 Å (full width at half-maximum, FWHM) at 2.06 Å pixel⁻¹.

Except PN6, which is close to the center of M31 and suffers from the bright background emission of the galaxy, direct acquisition imaging of all PNe was made with an exposure of 20–30 s. Figure 2 is the GTC OSIRIS CCD2 acquisition image of PN7, the faintest target in our sample (Table 1), with an exposure of 30 s. The GTC exposures are summarized in Table 1. In order to avoid light losses due to atmospheric diffraction, the long slit was placed along the parallactic angle during exposures. During the observations of each night, exposures of the spectrophotometric standard stars G191-B2B, Ross 640 and G158-100 (Oke 1974, 1990) were made. Arc lines of the neon and HgAr lamps were obtained for wavelength calibration. Exposures of the spectrophotometric standards and the arc lamps were made right after observations of each PN target.

Zoom In Zoom Out Reset image size

All data were reduced with standard procedures for long-slit spectra using iraf.⁶ The raw two-dimensional (2D) spectra were first combined, and then bias-subtracted, flat-fielded, cosmic-ray removed, and wavelength calibrated using exposures of the HgAr lamp. Geometry rectification, which was used to correct for geometry distortion along the slit, was also made during wavelength calibration. The sky background was then fitted along the slit direction using spline functions, and subtracted from the 2D spectra. The 1D spectrum of each target PN was extracted from the background subtracted 2D spectrum. The 1D spectra were then corrected for atmospheric extinction, and flux calibrated using observations of the standard stars. As an example, Figure 4 shows the fully calibrated 1D spectra of PN3 and PN7, the brightest and faintest targets in our sample, respectively.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Much effort was made to subtract the sky background. The excellent observing condition at ORM enables detection of the [O iii] λ4363 auroral line even in the faintest target PN of our sample (Figure 3, top panel). However, the sky background in the red part (>6000 Å) is relatively strong. Besides, the intensity of the sky lines is inhomogeneous along the long slit, and this inhomogeneity becomes significant for the strong sky lines such as [O i] $\lambda \lambda 5577,6300,6363.$ We used multiple cubic-spline functions to carefully fit the background emission along the slit direction, and then subtract it from the wavelength-calibrated (also geometry-rectified) 2D spectrum. Subtraction of the background emission is generally satisfactory (e.g., Figure 3, bottom panel) except PN6, which is close to the center of M31 and thus its spectrum is affected by the strong galactic background emission. Despite much effort, the signal-to-noise ratio (S/N) of the 1D spectrum of PN6 is lower than those of the other six targets.

The [O iii] λ4363 auroral line is seen in the reduced 2D spectra of all seven PNe (e.g., Figure 3, bottom panel). The [N ii] λ5755 auroral line is also detected in the spectra of relatively bright PNe. We also managed to discern the nebular emission of the [O i] $\lambda \lambda 6300,6363$ lines in the spectra of all targets except PN6. The [S iii] λ6312 auroral line, which is close to [O i] λ6300, was detected in PN1, PN2, PN3 and PN7 (Figure 4). The former three PNe are bright and located in the Northern Spur outskirts, and the latter one in the outer stream of M31, far (∼365; see Figure 1) from the galactic center. These four PNe also suffer much less from the galactic background emission than the other three.

We noticed the presence of second-order contamination in the final 1D spectra of the targets (Figure 4), although the R1000B grism was claimed to be free of such effects. Our derived response curve shows that the second-order contamination is obvious above 6300 Å. We corrected the red part (6200–7760 Å) of the response curve by carrying out global fitting using a five-order polynomial. The efficiency curve of R1000B given by the OSIRIS User Manual⁷ was used as a template. The final corrected R1000B response curve is safe for flux calibration (A. Cabrera-Lavers, GTC 2015, private communications).

The logarithmic extinction parameter at Hβ, c(Hβ), was derived for the seven M31 PNe by comparing the observed H i Balmer line ratios with the predicted Case B values. It is worth noting that some low-order H i Balmer lines are blended with He ii lines: Hγ is blended with He ii λ4338 (4f ²F^o–10g ²G), Hβ is blended with He iiλ4859 (4f ²F^o–8g ²G), and Hα with He ii λ6560 (4f ²F^o–6g ²G). Flux contributions of these He ii lines are negligible (<2%). The other low-order Balmer lines were not used for extinction correction because H i λ3889 (n = 2–8) is blended with the He i λ3888 (2s ³S–3p ³P^o) line, H i λ3970 (n = 2–7) is blended with [Ne iii] λ3967 (2p^{4 3}P₁–¹D₂), and Hδ (λ4101) is probably blended with the N iii λλ4097, 4103 (3s ²S–3p ²P^o) lines, although the flux contribution of the latter is expected to be negligible.

The average c(Hβ) value derived from the Hα/Hβ and Hγ/Hβ ratios, as adopted for all PNe, are presented in Table 2. Here the theoretical H i Balmer line ratios were adopted from Storey & Hummer (1995), assuming an electron temperature of 10,000 K and a density of 10⁴ cm⁻³. Given that the H i recombination line ratios are insensitive to the electron temperature and essentially independent on the density, this assumption of T_e and N_e is reasonable. The observed line fluxes were dereddened by

$$\begin{eqnarray}&&I(\lambda )={10}^{c({\rm{H}}\beta ){\text{}}f(\lambda )}F(\lambda ),\end{eqnarray} \tag{ 1 }$$

where $f(\lambda )$ is the extinction curve adopted from Cardelli et al. (1989), with a total-to-selective extinction ratio ${R}_{{\rm{V}}}$ = 3.1. The effects on the extinction-corrected line intensities caused by the use of different reddening curves (e.g., Savage & Mathis 1979; Cardelli et al. 1989; Clayton et al. 2015) was assessed. They resulted in very small differences in the dereddened intensities of emission lines in the wavelength range 3630–7760 Å covered by the R1000B grism of GTC OSIRIS. The observed fluxes and the extinction-corrected intensities of the emission lines detected in all seven M31 PNe, both normalized to Hβ = 100, are presented in Table 2. The total Hβ fluxes of the targets, as integrated in the extracted spectra, are also presented. Given that the slit width (10) we used is larger than the typical seeing (06–08) and that the slit is always placed along the parallactic angle during exposures, no significant light loss in Hβ is expected.

Table 2.  Fluxes and Intensities^a

Ion	λ	Transition	PN1		PN2		PN3		PN4
	(Å)		F(λ)	I(λ)	F(λ)	I(λ)	F(λ)	I(λ)	F(λ)	I(λ)
[O ii]	3727b	2p3 4So–2p3 2Do	47.2	54.6 ± 5.4	79.3	91.8 ± 9.1	58.2	66.4 ± 6.6	56.2	67.6 ± 6.7
H i	3798	2p 2Po–10d 2D	4.28	4.92 ± 1.35	8.42	9.69 ± 2.60	3.01	3.42 ± 0.94	4.51	5.38 ± 1.48
H i	3835	2p 2Po–9d 2D	5.84	6.68 ± 1.70	6.06	6.94 ± 1.74	4.88	5.52 ± 1.38	3.22	3.82 ± 1.00
[Ne iii]	3868	2p4 3P2–2p4 1D2	74.4	84.8 ± 4.2	112	128 ± 6	104	117 ± 6	69.1	81.7 ± 4.0
H i	3889c	2p 2Po–8d 2D	16.2	18.4 ± 2.0	16.0	18.2 ± 2.0	16.8	19.0 ± 2.0	14.5	17.1 ± 1.9
[Ne iii]	3967d	2p4 3P1–2p4 1D2	41.2	46.5 ± 3.6	52.3	59.1 ± 4.6	53.8	60.0 ± 4.6	38.3	44.6 ± 3.5
He i	4026	2p 3Po–5d 3D	2.21	2.47 ± 0.50	3.57	4.00 ± 0.83	1.89	2.10 ± 0.42	2.57	2.97 ± 0.60
[S ii]	4068e	3p3 4S${}_{3/2}^{{\rm{o}}}$–3p3 2P${}_{3/2}^{{\rm{o}}}$	4.56	5.10 ± 1.78	7.78	8.67 ± 2.20	5.99	6.60 ± 1.60	3.13	3.60 ± 1.10
H i	4101	2p 2Po–6d 2D	26.8	30.0 ± 3.3	29.1	32.4 ± 3.6	29.5	32.4 ± 3.6	28.9	33.0 ± 3.6
He i	4144	2p 1P${}_{1}^{{\rm{o}}}$–6d 1D2	2.24	2.50 ± 0.74	⋯	⋯	1.15	1.26 ± 0.38	⋯	⋯
C ii	4267	3d 2D–4f 2Fo	⋯	⋯	1.60	1.74 ± 0.43	⋯	⋯	⋯	⋯
H i	4340f	2p 2Po–5d 2D	44.4	47.7 ± 2.1	45.7	49.1 ± 2.2	43.6	46.5 ± 2.0	42.3	46.3 ± 2.0
[O iii]	4363	2p2 1D2–2p2 1S0	8.28	8.9 ± 1.0	11.9	12.8 ± 1.5	18.1	19.2 ± 2.2	8.63	9.42 ± 1.13
He i	4471	2p 3Po–4d 3D	5.58	5.88 ± 0.94	6.18	6.52 ± 1.04	5.10	5.35 ± 0.85	6.11	6.53 ± 1.05
N iii	4641g	3p 2P${}_{3/2}^{{\rm{o}}}$–3d 2D${}_{5/2}$	0.83	0.9 ± 0.5	2.03	2.09 ± 0.52	3.57	3.67 ± 0.90	⋯	⋯
C iii	4665h	3s${}^{\prime }$ 3P${}_{2}^{{\rm{o}}}$–3p${}^{\prime }$ 3P2	⋯	⋯	2.77	2.84 ± 0.71	⋯	⋯	⋯	⋯
He ii	4686	3d 2D–4f 2Fo	0.74	0.76 ± 0.30	4.00	4.10 ± 0.60	18.6	19.0 ± 1.7	1.36	1.40 ± 0.43
[Ar iv]	4711i	3p3 4S${}_{3/2}^{{\rm{o}}}$–3p3 2D${}_{5/2}^{{\rm{o}}}$	1.13	1.15 ± 0.35	1.21	1.24 ± 0.20	4.41	4.49 ± 0.67	2.18	2.24 ± 0.68
[Ar iv]	4740	3p3 4S${}_{3/2}^{{\rm{o}}}$–3p3 2D${}_{3/2}^{{\rm{o}}}$	1.03	1.04 ± 0.34	1.01	1.02 ± 0.20	5.19	5.27 ± 0.80	1.97	2.01 ± 0.60
H i	4861	2p 2Po–4d 2D	100	100	100	100	100	100	100	100
He i	4922	2p 1P${}_{1}^{{\rm{o}}}$–4d 1D2	2.0	2.0 ± 0.8	1.97	1.96 ± 0.77	1.20	1.20 ± 0.47	1.90	1.89 ± 0.75
[O iii]	4959	2p2 1P1–2p2 1D2	345	342 ± 17	434	429 ± 21	482	478 ± 23	321	317 ± 15
[O iii]	5007	2p2 1P2–2p2 1D2	1042	1025 ± 31	1297	1275 ± 38	1414	1393 ± 42	956	937 ± 28
[N i]	5198j	2p3 4S${}_{3/2}^{{\rm{o}}}$–2p3 2D${}_{3/2}^{{\rm{o}}}$	⋯	⋯	1.80	1.73 ± 0.50	0.67	0.65 ± 0.21	⋯	⋯
He ii	5411	4f 2Fo–7g 2G	⋯	⋯	⋯	⋯	1.81	1.72 ± 0.34	⋯	⋯
[Cl iii]	5517	3p3 4S${}_{3/2}^{{\rm{o}}}$–3p3 2D${}_{5/2}^{{\rm{o}}}$	⋯	⋯	⋯	⋯	0.48	0.45 ± 0.20	⋯	⋯
[Cl iii]	5537	3p3 4S${}_{3/2}^{{\rm{o}}}$–3p3 2D${}_{3/2}^{{\rm{o}}}$	⋯	⋯	⋯	⋯	0.72	0.68 ± 0.21	⋯	⋯
[N ii]	5755	2p2 1D2–2p2 1S0	1.90	1.76 ± 0.50	2.06	1.89 ± 0.60	2.64	2.45 ± 0.50	1.53	1.38 ± 0.42
C iv	5805	3s 2S–3p 2Po	⋯	⋯	11.6	10.7 ± 0.8	⋯	⋯	⋯	⋯
He i	5876	2p 3Po–3d 3D	19.1	17.4 ± 1.9	20.8	18.9 ± 2.1	16.3	15.0 ± 1.6	17.5	15.6 ± 1.7
[O i]	6300	2p4 3P2–2p4 1D2	4.10	3.64 ± 0.94	10.8	9.60 ± 2.40	8.82	7.93 ± 2.00	5.42	4.68 ± 1.20
[S iii]	6312k	3p2 1D2–3p2 1S0	1.37	1.21 ± 0.37	2.55	2.26 ± 0.69	1.41	1.26 ± 0.25	⋯	⋯
[O i]	6363	2p4 3P1–2p4 1D2	1.84	1.63 ± 0.73	4.52	4.00 ± 1.79	3.32	2.97 ± 1.33	2.30	1.98 ± 0.88
[N ii]	6548	2p2 3P1–2p2 1D2	21.0	18.4 ± 2.2	38.0	33.3 ± 4.0	34.0	30.2 ± 2.3	22.6	19.1 ± 2.3
H i	6563	2p 2Po–2d 2D	334	293 ± 1	356	310 ± 1	328	291 ± 1	341	288 ± 1
[N ii]	6583	2p2 3P2–2p2 1D2	64.7	56.5 ± 6.4	112	97.9 ± 10.7	99.6	88.2 ± 5.0	68.2	57.5 ± 6.3
He i	6678l	2p 1P${}_{1}^{{\rm{o}}}$–3d 1D2	5.76	5.01 ± 0.75	6.88	5.97 ± 0.90	4.74	4.17 ± 0.62	5.46	4.58 ± 0.68
[S ii]	6716	3p3 4S${}_{3/2}^{{\rm{o}}}$–3p3 2D${}_{5/2}^{{\rm{o}}}$	2.17	1.88 ± 0.20	5.21	4.51 ± 0.50	4.34	3.82 ± 0.43	7.24	6.05 ± 0.67
[S ii]	6731	3p3 4S${}_{3/2}^{{\rm{o}}}$–3p3 2D${}_{3/2}^{{\rm{o}}}$	4.00	3.45 ± 0.42	9.24	8.00 ± 1.01	8.39	7.37 ± 0.90	9.97	8.32 ± 0.92
[Ar v]	7005	3p2 3P2–3p2 1D2	⋯	⋯	0.70	0.60 ± 0.12	0.77	0.67 ± 0.13	⋯	⋯
He i	7065	2p 3Po–3s 3S	13.9	11.8 ± 1.5	15.7	13.3 ± 1.7	13.2	11.3 ± 1.4	9.24	7.52 ± 1.00
[Ar iii]	7136	3p4 3P2–3p4 1D2	18.4	15.5 ± 2.0	19.1	16.1 ± 2.1	24.7	21.2 ± 2.7	18.8	15.2 ± 2.0
He i	7281	2p 1P${}_{1}^{{\rm{o}}}$–3s 1S0	0.55	0.46 ± 0.12	⋯	⋯	0.62	0.52 ± 0.13	0.55	0.44 ± 0.11
[O ii]	7320	2p3 2D${}_{5/2}^{{\rm{o}}}$–2p3 2P${}_{3/2}^{{\rm{o}}}$	8.40	7.02 ± 1.14	9.79	8.17 ± 1.22	10.4	8.82 ± 1.32	5.62	4.49 ± 0.67
[O ii]	7330m	2p3 2D${}_{3/2}^{{\rm{o}}}$–2p3 2P${}_{3/2}^{{\rm{o}}}$	8.08	6.74 ± 1.00	10.9	9.13 ± 1.37	7.13	6.06 ± 0.90	4.21	3.36 ± 0.50
[Ar iii]	7751	3p4 3P1–3p4 1D2	3.86	3.14 ± 1.16	1.67	1.35 ± 0.47	7.28	6.05 ± 2.45	5.31	4.10 ± 1.44
c(Hβ)			0.196		0.198		0.177		0.247
log F(Hβ)n			−14.70		−14.96		−14.73		−14.88
Ion	λ	Transition	PN5		PN6		PN7
	(Å)		F(λ)	I(λ)	F(λ)	I(λ)	F(λ)	I(λ)
[O ii]	3727b	2p3 4So–2p3 2Do	42.4	59.4 ± 7.1	33.2	46.4 ± 9.3	73.4	84.3 ± 7.6
H i	3798	2p 2Po–10d 2D	⋯	⋯	⋯	⋯	5.23	5.97 ± 1.64
H i	3835	2p 2Po–9d 2D	⋯	⋯	10.2	13.9 ± 3.5	4.17	4.74 ± 1.20
[Ne iii]	3868	2p4 3P2–2p4 1D2	87.0	118 ± 6	63.0	85.4 ± 4.2	75.6	85.6 ± 4.2
H i	3889c	2p 2Po–8d 2D	13.0	17.6 ± 2.0	16.0	21.6 ± 2.3	18.3	20.8 ± 2.3
[Ne iii]	3967d	2p4 3P1–2p4 1D2	43.2	57.0 ± 5.0	32.3	42.7 ± 3.3	42.2	47.3 ± 3.7
[S ii]	4068e	3p3 4S${}_{3/2}^{{\rm{o}}}$–3p3 2P${}_{3/2}^{{\rm{o}}}$	6.36	8.2 ± 3.2	⋯	⋯	4.37	4.84 ± 1.16
H i	4101	2p 2Po–6d 2D	27.8	35.0 ± 4.1	30.4	38.7 ± 4.3	25.5	28.2 ± 3.1
H i	4340f	2p 2Po–5d 2D	37.8	44.7 ± 2.0	38.8	45.8 ± 2.1	44.0	47.1 ± 2.1
[O iii]	4363	2p2 1D2–2p2 1S0	8.22	9.6 ± 1.2	4.84	5.68 ± 0.70	6.79	7.25 ± 0.80
He i	4471	2p 3Po–4d 3D	7.82	8.84 ± 1.40	4.50	5.10 ± 0.80	6.54	6.87 ± 0.76
N iii	4641	3p 2P${}_{3/2}^{{\rm{o}}}$–3d 2D${}_{5/2}$	⋯	⋯	⋯	⋯	2.10	2.15 ± 0.43
He ii	4686	3d 2D–4f 2Fo	1.85	1.95 ± 0.30	1.76	1.85 ± :	2.23	2.28 ± 0.34
[Ar iv]	4711i	3p3 4S${}_{3/2}^{{\rm{o}}}$–3p3 2D${}_{5/2}^{{\rm{o}}}$	2.23	2.33 ± 0.28	⋯	⋯	1.55	1.58 ± 0.20
[Ar iv]	4740	3p3 4S${}_{3/2}^{{\rm{o}}}$–3p3 2D${}_{3/2}^{{\rm{o}}}$	3.30	3.43 ± 0.40	⋯	⋯	1.67	1.70 ± 0.22
H i	4861	2p 2Po–4d 2D	100	100	100	100	100	100
He i	4922	2p 1P${}_{1}^{{\rm{o}}}$–4d 1D2	3.03	3.00 ± 1.24	⋯	⋯	2.13	2.12 ± 0.84
[O iii]	4959	2p2 1P1–2p2 1D2	422	413 ± 20	376	368 ± 18	342	339 ± 16
[O iii]	5007	2p2 1P2–2p2 1D2	1260	1216 ± 37	1106	1068 ± 32	1033	1017 ± 31
[N i]	5198j	2p3 4S${}_{3/2}^{{\rm{o}}}$–2p3 2D${}_{3/2}^{{\rm{o}}}$	⋯	⋯	⋯	⋯	1.62	1.57 ± 0.55
[N ii]	5755	2p2 1D2–2p2 1S0	2.57	2.12 ± 0.64	⋯	⋯	1.10	1.02 ± 0.30
He i	5876	2p 3Po–3d 3D	21.2	17.2 ± 1.60	16.4	13.3 ± 1.5	17.9	16.4 ± 0.8
[O i]	6300	2p4 3P2–2p4 1D2	6.87	5.24 ± 1.30	⋯	⋯	7.88	7.04 ± 1.76
[S iii]	6312k	3p2 1D2–3p2 1S0	⋯	⋯	⋯	⋯	2.03	1.81 ± 0.56
[O i]	6363	2p4 3P1–2p4 1D2	3.73	2.82 ± 1.26	⋯	⋯	2.33	2.08 ± 0.80
[N ii]	6548	2p2 3P1–2p2 1D2	33.1	24.4 ± 2.9	5.91	4.36 ± 0.52	28.1	24.8 ± 3.0
H i	6563	2p 2Po–2d 2D	357	263 ± 1	338	249 ± 1	341	300 ± 1
[N ii]	6583	2p2 3P2–2p2 1D2	96.5	71.0 ± 7.5	18.0	13.2 ± 1.5	87.3	76.8 ± 8.4
He i	6678l	2p 1P${}_{1}^{{\rm{o}}}$–3d 1D2	5.39	3.90 ± 0.58	⋯	⋯	5.11	4.47 ± 0.67
[S ii]	6716	3p3 4S${}_{3/2}^{{\rm{o}}}$–3p3 2D${}_{5/2}^{{\rm{o}}}$	10.3	7.45 ± 0.83	⋯	⋯	10.1	8.8 ± 1.0
[S ii]	6731	3p3 4S${}_{3/2}^{{\rm{o}}}$–3p3 2D${}_{3/2}^{{\rm{o}}}$	13.6	9.8 ± 1.0	⋯	⋯	15.4	13.5 ± 1.0
[Ar v]	7005	3p2 3P2–3p2 1D2	⋯	⋯	⋯	⋯	0.50	0.42 ± 0.11
He i	7065	2p 3Po–3s 3S	9.90	6.80 ± 0.86	⋯	⋯	11.6	10.0 ± 1.3
[Ar iii]	7136	3p4 3P2–3p4 1D2	34.5	23.4 ± 3.0	⋯	⋯	20.9	17.8 ± 2.13
[O ii]	7320	2p3 2D${}_{5/2}^{{\rm{o}}}$–2p3 2P${}_{3/2}^{{\rm{o}}}$	⋯	⋯	⋯	⋯	5.77	4.87 ± 0.73
[O ii]	7330m	2p3 2D${}_{3/2}^{{\rm{o}}}$–2p3 2P${}_{3/2}^{{\rm{o}}}$	⋯	⋯	⋯	⋯	5.94	5.01 ± 0.75
[Ar iii]	7751	3p4 3P1–3p4 1D2	11.1	6.93 ± 2.43	⋯	⋯	4.29	3.53 ± 1.24
c(Hβ)			0.451		0.450		0.186
logF(Hβ)n			−14.94		−15.13		−15.10

Notes.

^aFluxes (also the intensities) are normalized such that Hβ = 100. Emission lines that were not detected in the spectra of PN5, PN6, or PN7 are not presented in the table for the three targets. ^bA blend of the O ii λ3726 (2p^{3 4}S${}_{3/2}^{{\rm{o}}}$–2p^{3 2}D${}_{3/2}^{{\rm{o}}}$) and λ3729 (2p^{3 4}S${}_{3/2}^{{\rm{o}}}$–2p^{3 2}D${}_{5/2}^{{\rm{o}}}$) lines. ^cBlended with the He i λ3888 (2s ³S–3p ³P^o) line. ^dBlended with the H i λ3970 (2p ²P^o–7d ²D) and He i λ3965 (2s ¹S₀–4p ¹P${}_{1}^{{\rm{o}}}$) lines. ^eBlended with the O ii M10 3p ⁴D^o–3d ⁴F and C iii M16 4f ³F^o–5g ³G, and [S ii] 4076 (3p^{3 4}S${}_{3/2}^{{\rm{o}}}$–3p^{3 2}P${}_{1/2}^{{\rm{o}}}$) lines. ^fCorrected for the flux contribution from the blended He ii λ4338 (4f ²F^o–10g ²G) line. Fluxes of Hβ and Hα have also been corrected for the blended He ii λ4859 (4f ²F^o–8g ²G) and λ6560 (4f ²F^o–6g ²G) lines, respectively. ^gBlended with the N iii λ4634 and λ4642 lines of multiplet M2 3p 2P^o–3d ²D and the O ii λ4642 and λ4639 lines of multiplet M1 3s ⁴P–3p ⁴D^o. ^hProbably blended with the C iv λ4658 (M8 5g ²G–6h ²H^o) line for the case of PN2. ⁱCorrected for the flux contribution from the blended He iλ4713 (2p ³P^o–4s ³S) line. ^jBlended with [N i] λ5200 (2p^{3 4}S${}_{3/2}^{{\rm{o}}}$–2p^{3 2}D${}_{5/2}^{{\rm{o}}}$. ^kCorrected for the flux contribution from the blended He ii λ6311 (5g ²G–16h ²H^o) line. ^lCorrected for the flux contribution from the blended He ii λ6683 (5g ²G–13h ²H^o) line. ^mBlended with the [Ar iv] λ7331 (3p^{3 2}D${}_{5/2}^{{\rm{o}}}$ –²P${}_{1/2}^{{\rm{o}}}$) line, whose flux contribution is negligible (<1%). ⁿIn unit of erg cm⁻² s⁻¹ in the extracted spectra.

Download table as:  ASCIITypeset images: 1 2

We present in Table 2 the normalized (to Hβ = 100) intensities and measurement errors of the emission lines. For brighter PNe, the typical uncertainties in fluxes of the [O iii] λ4363 auroral line are less than 10%. For the PN1, PN2, and PN3, the relatively bright targets in our sample, the S/N of the λ4363 line are ∼30, 21 and >50, respectively. For the faintest target PN7, S/N([O iii] λ4363) ∼10. Electron temperature diagnostics using the [O iii] (λ4959 + λ5007)/$\lambda 4363$ line ratio are presented in Section 3.3.

Some emission lines detected in our spectra have blending issues.We corrected the flux of the [Ne iii] λ3967 line for the blended H i λ3970 line, using the theoretical hydrogen line ratio of Storey & Hummer (1995). However, for all targets, the corrected [Ne iii] λ3967 (dereddened) line flux still yields higher Ne²⁺/H⁺ abundance ratios than that derived from [Ne iii] λ3868, which is free of line blending. The [Ar iv] λ4711 line is blended with He i λ4713 (2p ³P^o–4s ³S), which contributes 30%–40% to the total flux of the [Ar iv] line for all seven targets. The He i line flux was estimated using the theoretical Case B He i λ4713/λ4471 ratio calculated by Porter et al. (2012), whose calculation for this ratio only differs from that of Benjamin et al. (1999) by less than 0.5%. The [S iii] λ6312 line is blended with He iiλ6311 (5g ²G–16h ²H^o), which typically contributes less than 1% to the total flux. For PN3, whose He ii λ4686 is the strongest among the targets, flux contribution of the He ii λ6311 line is close to 5%.

Some permitted lines of heavy elements were detected (Table 2). These lines are faint, with a typical intensity <5% of Hβ. The N iii λ4641 (M2 3p ²P${}_{3/2}^{{\rm{o}}}$–3d ²D${}_{5/2}$) line was well detected in the spectrum of PN3 (Figure 4). This line is blended with N iii λ4634, 4642 of the same multiplet, and is probably blended with the O ii M1 3s ⁴P–3p ⁴D^o lines (see for example, Figure 1 in Liu et al. 1995 or Figure 32 in Fang & Liu 2013). The dominant excitation mechanism of the [N iii] λ4641 line is the starlight and/or Bowen fluorescence (e.g., Bowen 1935; Grandi 1976). Thus this line is unsuitable for the ionic abundance (N³⁺/H⁺) calculation. Carbon emission lines were detected in the spectrum of PN2, as discussed in Section 3.2 below.

The uncertainties of the emission line intensities were derived by quadratically adding the uncertainties associated with the Poisson noise of the line and the CCD and background noise. The former scales with the electron counts of the line, and the latter scales with the local noise and the full line width. The error contributions from systematic effects, such as flux calibration, subtraction of the background level, determination of reddening, or line blending, are expected to be much smaller and were not taken into account. Our flux calibration is reliable as the extinction parameter c(Hβ) derived from the Hγ/Hβ ratio generally agrees with that derived from Hα/Hβ. The important nebular emission lines of our PNe are generally not much affected by the sky lines (e.g., the Mercury line Hg i λ4358 is faint at ORM, La Palma, and the detection of the [O iii] λ4363 auroral is not much affected). Moreover, the background subtraction in the 2D spectra seemed flawless (e.g., Figure 3; see also the description of data reduction in Section 2.3). The errors introduced by reddening are also small, less than 5% as discussed in Section 2.4. Finally, line blending was also considered when analyzing the emission lines. Fortunately, errors due to this effect are mostly negligible for the lines that are critical for plasma diagnostics and abundance determinations (Table 2), except for the [Ar iv] λ4711 line.

Very broad C iv λ5805 (M1 3s ²S–3p ²P${}^{{\rm{o}}}$) line, with FWHM ∼43 Å, was detected in the spectrum of PN2 (Figure 5, right). This line is a blend of two fine-structure components λλ5801.33, 5811.98⁸ , and can be attributed to the emission of a Wolf–Rayet (WR) type central star. The same feature has been detected in the spectra of another two PNe in the outer disk of M31 by (Balick et al. 2013, Figure 2). The extinction-corrected flux of the C iv λ5805 line in our PN2 is 1.20 × 10⁻¹⁶ erg cm⁻² s⁻¹, with an uncertainty of ∼10%. This flux is generally in line with those observed by Balick et al. (2013).

Zoom In Zoom Out Reset image size

Also detected in the spectrum of PN2 are the C iii λ4665 (M5 3s${}^{\prime }$ ³P${}_{2}^{{\rm{o}}}$–3p${}^{\prime }$ ³P₂) and C ii λ4267 (M6 3d ²D–4f ²F^o) lines. The high-excitation C iii λ4665 line might be due to emission of the WR central star. A Gaussian profile with FWHM ∼11 Å seems to well fit this line (Figure 5, middle), but its accurate profile is difficult to discern due to noise and line blending. C iii λ4665 is probably blended with the C iv λ4658 (M8 5g ²G–6h ²H^o) and C iii λ4649 (M1 3s ³S–3p ³P^o) recombination lines.

The C ii λ4267 optical recombination line (ORL) is produced by radiative recombination only, and thus is due to nebular emission. Its line width, although difficult to discern due to noise and possible blending (Figure 5, left), seems narrower than that of the C iv λ5805 line. The extinction-corrected flux of the λ4267 line is ∼2.1 × 10⁻¹⁷ erg cm⁻² s⁻¹, with an uncertainty over 50%. This flux value is close to that observed in one of the two PNe of Balick et al. (2013), which is ∼1.8 × 10⁻¹⁷ erg cm⁻² s⁻¹. The C ii λ4267 line was used to calculate the ${{\rm{C}}}^{2+}$/H⁺ abundance ratio in Section 3.4.

Plasma diagnostics were determined using the collisionally excited lines⁹ (CELs) of heavy elements in conventional manner (e.g., Osterbrock & Ferland 2006). The [O iii] (λ4959 + λ5007)/λ4363 auroral-to-nebular line ratio was used to derive the electron temperature. The [N ii] λ5755 auroral line, whose flux is typically ∼5%–15% of that of the [O iii] λ4363 line, was detected in all targets except PN6, and the [N ii] (λ6548 + λ6583)/λ5755 ratio was also used as a temperature diagnostic. The [Ar iv] λ4711/λ4740 and [S ii] λ6716/λ6731 nebular line ratios were used for density diagnostics. The [O ii] λ3727 line is a blend of the λλ3726, 3729 doublet. Since the [O ii] λλ7320, 7330 auroral lines were detected in the spectra of PN1–4 and PN7, we also used the [O ii] λ3727/(λ7320 + λ7330) line ratio to estimate electron temperatures for these five objects. The weak [S ii] λ4068 line is blended with the even weaker [S ii] λ4076 line as well as with the O ii M10 3p ⁴D^o–3d ⁴F and C iii M16 4f ³F^o–5g ³G lines. The electron temperatures derived from the [S ii] (λ6716 + λ6731)/λ4072 ratio are highly uncertain.

Figure 6 shows the plasma diagnostic diagrams for six of our PNe. The diagram of PN6 is not presented because only the [O iii] λ4363 diagnostic line was detected in the spectrum of this object. These diagnostic diagrams are based on CELs and were created by solving the level population equations for five-level atomic models using equib, a fortran code originally developed by Howarth & Adams (1981) to solve the equations of statistical equilibrium in multi-level atoms and yield level populations and line emissivities for a specified physical condition, appropriate to the zones where the ions are expected to exist. The atomic data used by the equib code were updated manually. References for the atomic data set used for plasma diagnostics based on CELs, as well as for ionic abundance determinations (Section 3.4.1), are presented in Table 3. Results of our plasma diagnostics are presented in Table 4.

Zoom In Zoom Out Reset image size

The uncertainties in the electron temperatures and densities (Table 4) were estimated through the propagation of uncertainties in the emission line intensities, and do not include the contribution from systematic effects. As discussed in Section 3.1, the error contributions from the systematic effects, such as reddening, flux calibration, and line blending, are expected to be small. The atomic data used for plasma diagnostics are mostly updated (Table 3). It is difficult to quantitatively estimate the systematic uncertainties introduced by the atomic data. We expect that the atomic data adopted are reliable, since these calculations were improved over the previous work (see the discussion in Section 7 of Kwitter et al. 2014 contributed by X.-W. Liu and X. Fang).

Figure 6 shows that the [O iii] line ratio is an excellent temperature diagnostic, and variation of the electron density does not change the resultant electron temperature much. Uncertainties in the electron temperatures derived from the [N ii] line ratio are much larger than those of the [O iii] temperatures, mainly due to weakness of the [N ii] λ5755 auroral line. Besides, the [N ii] nebular-to-auroral line ratio also has slight density dependence (Figure 6). The [O ii] nebular-to-auroral line ratio is sensitive to both temperature and density, and is thus not an ideal temperature indicator. For some targets, the [N ii] temperature is higher than the [O iii] temperature, which is likely due to the overestimated intensity of the very weak λ5755 line. In only one object (PN7) does the [N ii] temperature seem to be meaningfully different from the [O iii] temperature (Table 4). The [O iii] temperature is thus adopted in the abundance calculations for all PNe. We are aware that in a PN there are different ionization zones, as represented by the ${{\rm{O}}}^{2+}$ and ${{\rm{N}}}^{+}$ ions, and using only the [O iii] temperature may introduce extra errors in the ionic abundances for the singly ionized species. In this respect, the actual uncertainties in the ${{\rm{N}}}^{+}$/H⁺, ${{\rm{O}}}^{+}$/H⁺, and ${{\rm{S}}}^{+}$/H⁺ abundance ratios might be slightly larger than those given in Table 5. Recently, in a spectroscopic study of Galactic PNe, Dufour et al. (2015) adopted a constant [N ii] temperature, which was carefully derived by Kaler (1986) for the case where there is no detection of the [N ii] λ5755 line but He ii λ4686 is present.

Table 5.  Ionic Abundances

Ion	Line	Abundance (X${}^{i+}$/H+)
	(Å)	PN1	PN2	PN3	PN4	PN5	PN6	PN7
He+	3888	0.059 ± 0.027	0.057 ± 0.026	0.062 ± 0.018	0.049 ± 0.023	0.053 ± 0.024	0.082 ± 0.031	0.076 ± 0.025
	4471	0.112 ± 0.018	0.124 ± 0.020	0.102 ± 0.012	0.124 ± 0.020	0.168 ± 0.034	0.097 ± 0.024	0.131 ± 0.014
	5876	0.113 ± 0.012	0.123 ± 0.014	0.097 ± 0.010	0.101 ± 0.011	0.112 ± 0.015	0.087 ± 0.015	0.106 ± 0.006
	6678	0.122 ± 0.019	0.145 ± 0.023	0.099 ± 0.014	0.112 ± 0.017	0.096 ± 0.013	⋯	0.109 ± 0.016
Adopteda		0.113 ± 0.012	0.123 ± 0.014	0.097 ± 0.010	0.101 ± 0.011	0.112 ± 0.015	0.087 ± 0.015	0.106 ± 0.006
He2+	4686	0.63( ± 0.26)×10−3	3.40( ± 0.49)×10−3	1.58( ± 0.14)×10−2	1.17( ± 0.36)×10−3	1.63( ± 0.39)×10−3	1.53( ± :)×10−3	1.89( ± 0.28)×10−3
C2+	4267	⋯	1.7( ± 0.8)×10−3	⋯	⋯	⋯	⋯	⋯
N+	5755	1.46( ± 0.60)×10−5	1.27( ± 0.39)×10−5	9.23( ± 2.10)×10−6	9.12( ± 2.45)×10−6	1.76( ± 0.74)×10−5	⋯	1.08( ± 0.32)×10−5
	6548	8.51( ± 1.02)×10−6	1.34( ± 0.20)×10−5	1.01( ± 0.13)×10−5	7.84( ± 1.28)×10−6	1.39( ± 0.34)×10−5	2.95( ± 1.32)×10−6	1.39( ± 0.17)×10−5
	6583	8.91( ± 0.97)×10−6	1.35( ± 0.21)×10−5	1.01( ± 0.15)×10−5	8.02( ± 1.30)×10−6	1.37( ± 0.35)×10−5	3.03( ± 1.30)×10−6	1.47( ± 0.20)×10−5
Adoptedb		8.91( ± 0.97)×10−6	1.35( ± 0.21)×10−5	1.01( ± 0.15)×10−5	8.02( ± 1.30)×10−6	1.37( ± 0.35)×10−5	3.03( ± 1.30)×10−6	1.47( ± 0.20)×10−5
O+	3727	2.05( ± 0.36)×10−5	2.52( ± 0.61)×10−5	1.98( ± 0.40)×10−5	2.06( ± 0.49)×10−5	2.35( ± 0.66)×10−5	2.96( ± 0.78)×10−5	5.1( ± 0.9)×10−5
	7320	6.13( ± 1.54)×10−5	6.17( ± 1.65)×10−5	2.54( ± 0.70)×10−5	3.02( ± 0.96)×10−5	⋯	⋯	5.1( ± 1.2)×10−5
	7330	7.24( ± 1.82)×10−5	8.50( ± 2.27)×10−5	2.14( ± 0.58)×10−5	2.78( ± 0.91)×10−5	⋯	⋯	6.42( ± 1.51)×10−5
Adoptedc		2.05( ± 0.36)×10−5	2.52( ± 0.61)×10−5	1.98( ± 0.40)×10−5	2.06( ± 0.49)×10−5	2.35( ± 0.66)×10−5	2.96( ± 0.78)×10−5	5.1( ± 0.9)×10−5
O2+	4363	2.58( ± 0.60)×10−4	2.73( ± 0.57)×10−4	2.19( ± 0.43)×10−4	2.02( ± 0.48)×10−4	3.61( ± 1.05)×10−4	4.45( ± 1.52)×10−4	3.17( ± 0.64)×10−4
	4959	2.50( ± 0.24)×10−4	2.68( ± 0.31)×10−4	2.18( ± 0.24)×10−4	1.99( ± 0.32)×10−4	3.44( ± 0.56)×10−4	4.47( ± 0.84)×10−4	3.08( ± 0.33)×10−4
	5007	2.60( ± 0.26)×10−4	2.75( ± 0.32)×10−4	2.20( ± 0.25)×10−4	2.03( ± 0.37)×10−4	3.51( ± 0.57)×10−4	4.49( ± 0.86)×10−4	3.20( ± 0.34)×10−4
Adoptedd		2.60( ± 0.26)×10−4	2.75( ± 0.32)×10−4	2.20( ± 0.25)×10−4	2.03( ± 0.37)×10−4	3.57( ± 0.57)×10−4	4.49( ± 0.86)×10−4	3.20( ± 0.34)×10−4
Ne2+	3868	5.85( ± 0.66)×10−5	7.31( ± 0.97)×10−5	4.62( ± 0.59)×10−5	4.69( ± 0.76)×10−5	9.48( ± 1.78)×10−5	1.07( ± 0.23)×10−4	7.61( ± 0.96)×10−5
	3967	7.06( ± 1.17)×10−5	8.22( ± 1.40)×10−5	5.82( ± 1.00)×10−5	5.46( ± 0.88)×10−5	1.09( ± 0.20)×10−4	1.12( ± 0.25)×10−4	9.36( ± 1.22)×10−5
Adoptede		5.85( ± 0.66)×10−5	7.31( ± 0.97)×10−5	4.62( ± 0.59)×10−5	4.69( ± 0.76)×10−5	9.54( ± 1.78)×10−5	1.07( ± 0.23)×10−4	7.61( ± 0.96)×10−5
S+	6716	1.34( ± 0.28)×10−7	2.37( ± 0.45)×10−7	3.30( ± 0.36)×10−7	3.77( ± 0.64)×10−7	4.84( ± 0.74)×10−7	⋯	1.01( ± 0.11)×10−6
	6731	1.60( ± 0.30)×10−7	3.08( ± 0.52)×10−7	3.44( ± 0.38)×10−7	3.43( ± 0.60)×10−7	4.61( ± 0.67)×10−7	⋯	8.75( ± 1.05)×10−7
Adoptedf		1.60( ± 0.30)×10−7	3.08( ± 0.52)×10−7	3.44( ± 0.38)×10−7	3.43( ± 0.60)×10−7	4.61( ± 0.67)×10−7	⋯	8.75( ± 1.05)×10−7
S2+	6312	1.92( ± 0.33)×10−6	2.93( ± 0.58)×10−6	1.09( ± 0.18)×10−6	⋯	⋯	⋯	3.7( ± 0.7)×10−6
Ar2+	7136	1.14( ± 0.20)×10−6	1.06( ± 0.23)×10−6	1.12( ± 0.21)×10−6	1.01( ± 0.28)×10−6	1.89( ± 0.47)×10−6	⋯	1.5( ± 0.4)×10−6
	7751	9.65( ± 3.38)×10−7	3.72( ± 1.41)×10−7	1.33( ± 0.52)×10−6	1.13( ± 0.51)×10−6	2.33( ± 0.88)×10−6	⋯	1.3( ± 0.7)×10−6
Adoptedg		1.14( ± 0.20)×10−6	1.06( ± 0.23)×10−6	1.12( ± 0.21)×10−6	1.01( ± 0.28)×10−6	1.89( ± 0.47)×10−6	⋯	1.51( ± 0.36)×10−6
Ar3+	4711	1.52( ± 0.35)×10−7	1.36( ± 0.31)×10−7	4.67( ± 0.80)×10−7	2.57( ± 0.67)×10−7	4.83( ± 1.35)×10−7	⋯	2.8( ± 0.6)×10−7
	4740	1.52( ± 0.27)×10−7	1.34( ± 0.30)×10−7	4.68( ± 0.75)×10−7	2.58( ± 0.65)×10−7	4.84( ± 1.30)×10−7	⋯	2.8( ± 0.5)×10−7
Adoptedh		1.52( ± 0.30)×10−7	1.35( ± 0.30)×10−7	4.68( ± 0.77)×10−7	2.58( ± 0.66)×10−7	4.84( ± 1.33)×10−7	⋯	2.8( ± 0.5)×10−7
Ar4+	7005	⋯	8.00( ± 2.83)×10−8	6.82( ± 2.25)×10−8	⋯	⋯	⋯	7.6( ± 2.3)×10−8

Notes.

^aThe He⁺/H⁺ abundance derived from the He i λ5876 line is adopted for all PNe. ^bThe ${{\rm{N}}}^{+}$/H⁺ abundance derived from the [N ii] λ6583 line is adopted for all PNe. ^cThe ${{\rm{O}}}^{+}$/H⁺ abundance derived from the [O ii] λ3727 doublet is adopted for all PNe. ^dThe ${{\rm{O}}}^{2+}$/H⁺ abundance derived from the [O iii] λ5007 line was adopted for all PNe. ^eThe Ne²⁺/H⁺ abundance derived from the [Ne iii] λ3868 line is adopted. [Ne iii] λ3967 is blended with H i λ3970. ^fThe ${{\rm{S}}}^{+}$/H⁺ abundance derived from the [S ii] λ6731 line is adopted, because the critical density of λ6731 is much higher than that of the λ6716 line. ^gThe Ar²⁺/H⁺ abundance derived from the [Ar iii] λ7136 line is adopted. Measurements of λ7136 are more reliable than those of the λ7751 line at the red end of the spectra. ^hThe Ar³⁺/H⁺ abundance derived from the total intensity of [Ar iv] λλ4711,4740 is adopted for all PNe.

Download table as:  ASCIITypeset image

In a few cases of our PN sample, there is large difference between the electron density derived from the [S ii] line ratio and that derived from [Ar iv], as shown in Table 4. It might be that in our PNe there are low-ionization regions where the [S ii] density is more appropriate, whereas the [Ar iv] density applies in the higher ionization regions. Thus we adopted the [S ii] densities for the ionic-abundance calculations of the low ionization species and adopted the [Ar iv] densities for the high ionization species (Section 3.4.1). Measurements of the relatively faint [Ar iv] lines in PN5 are of large uncertainty due to the noise, and for this object we adopted the [S ii] density.

An electron density of ∼2000 cm⁻³, typical for most of our targets, was assumed for PN6 because neither the [S ii] nor the [Ar iv] doublet were detected in this PN. In order to check how much the assumed electron density would affect the resultant ionic abundances, we calculated ionic abundances for two density cases, 10³ and 10⁴ cm⁻³, at 9500 K, the [O iii] temperature derived for PN6. The ${{\rm{N}}}^{+}$/H⁺ ionic abundance ratios derived at the two density cases differ by ∼10%, while both the ${{\rm{O}}}^{2+}$/H⁺ and the Ne²⁺/H⁺ ratios differ by <4%; the ${{\rm{O}}}^{+}$/H⁺ abundance ratios derived from the [O ii] λ3727 line differ by nearly 50% at the two densities. These differences in the CEL ionic abundances at the two density cases are due to the large differences in critical densities, as discussed in Paper I. For the helium recombination lines, differences in ionic abundances derived at the two density cases are negligible (<6%; Porter et al. 2012).

The relatively high electron temperature and density of PN3 (Table 4) indicate that this target may be young (compact) compared to other PNe. The [O iii] λ4363 auroral line in PN3 is stronger than other PNe. PN3 also has strong He iiλ4686 line (∼0.2 Hβ), while this line is much fainter in the spectra of the other targets (Table 2). The above information points to the high-excitation nature of PN3. This high excitation could be due to either a very hot central star or the low metallicity, or both. The relatively high density in PN3 may also cause collisional de-excitation of the heavy element ions, which makes the radiation cooling inefficient and as a consequence, produces relatively higher electron temperatures.

The intensity ratio of two He i optical recombination lines that originate from two atomic levels with different excitation energies is sensitive to the electron temperature (although not so acutely sensitive to T_e as the CELs), and thus can also be used as a temperature diagnostic. The temperature diagnostics for gaseous nebulae based on the He i recombination lines were constructed by Zhang et al. (2005) using the theoretical He i line emissivities calculated by Benjamin et al. (1999). We have determined the He i line temperatures for our PNe following the diagnostic method of Zhang et al. (2005) and the results are presented in Table 4. Generally the electron temperatures derived from the He i line ratios, designated T_e(He i), are lower than those derived from the heavy element CELs, T_e(CEL), except PN4 and PN7, whose He i λ6678/λ4471 ratio yields higher temperatures. This is in line with the temperature sequence, T_e(CEL) ≳ T_e(H i) ≳ T_e(He i), so far observed in more than 100 Galactic PNe through deep spectroscopy (e.g., McNabb et al. 2013; see also recent reviews by Liu 2006; Liu et al. 2012). Zhang et al. (2005) studied 48 Galactic PNe and found that the T_e(He i) values are significantly lower than the electron temperatures derived from the H i Balmer jump, T_e(H i), with an average difference of 4000 K. The significant difference between the electron temperatures of PNe derived from the CELs and the H i (also He i) recombination spectrum is the renowned "temperature discrepancy" in nebular astrophysics (e.g., Peimbert 1971; Liu & Danziger 1993). Investigation of this subject has never been extended to extragalactic PNe due to limited S/N. The spectral quality of our PN targets also prevented further study of the problem.

3.4.1. Ionic Abundances

3.4.2. Elemental Abundances

Of all our targets, including the three PNe studied in Paper I, only PN4 (M77) was observed by Sanders et al. (2012), who obtained spectra for 253 H ii regions and 407 PNe in M31 using the Hectospec multi-fiber spectrograph attached to the 6.5 m MMT. Oxygen abundance was derived for 51 PNe in the sample of Sanders et al. (2012) based on the electron temperatures estimated from the [O iii] λ4363 auroral line. However, the λ4363 line was not detected in the MMT spectrum of M77, and thus the O/H ratio was not estimated. For the most prominent nebular emission lines, fluxes measured in our GTC spectrum of M77 only differ slightly from those of the MMT spectrum: 2% for the [O iii] λλ4959, 5007 lines, 7% for the [N ii] λλ6548, 6583 lines, 12% for the [O ii] λ3727 lines, and 20%–40% for the [S ii] λλ6716, 6731 lines.

Modern large-area surveys have revealed that M31 has an extended stellar disk and a huge halo (≥300 kpc; Ibata et al. 2005, 2007, 2014). The outer regions of M31 are extremely complex. Numerous structures such as streams, loops, and overdensity regions were discovered throughout the halo (e.g., McConnachie et al. 2009; Lewis et al. 2013; Ibata et al. 2014). These features, with the Northern Spur and the Giant Stream being the most prominent, are mostly associated with the accretion/interaction of satellite dwarf galaxies. Although the possible connection between the Northern Spur and the Giant Stream had already been inferred by Ferguson et al. (2002) and McConnachie et al. (2003), Merrett et al. (2003) were the first to explicitly propose this connection. Based on a kinematic study of ∼20 PNe in the disk of M31, Merrett et al. (2003) presented a possible orbit for the stellar stream in M31, which connects the Giant Stream to the Northern Spur. The Northern Spur is located at the turning point of this model orbit, which is strongly warped as the stellar stream passes nearby the center of M31 (Figure 7, upper-left panel).

Zoom In Zoom Out Reset image size

Figure 7 shows the positions of PNe observed by Merrett et al. (2006) in the X–Y coordinate system in an M31-based reference frame, where X lies along the major axis of M31 and increases toward the southwest, and Y lies along the minor axis and increases toward the northwest. Both coordinates were calculated following the geometric transformations of Huchra et al. (1991). The orbit proposed by Merrett et al. (2003) is presented along with the PNe in Figure 7. The two side (bottom and right) panels in Figure 7 show the projection of this stellar orbit in the line of sight velocity with respect to M31 (${v}_{{\rm{los}}}$) versus distance along the major and minor axes. PNe in the Northern Spur and those associated with the Giant Stream are highlighted. The seven PNe in this study and the three observed in Paper I are also highlighted (see explanation of symbols in the caption of Figure 7).

The PNe in the two substructures are generally associated with the stellar orbit of Merrett et al. (2003), and this association is particularly consistent for those located on the Giant Stream. The three Northern Spur PNe observed in Paper I (see the discussion therein) are located at the turning point of the orbit. The positions of the seven objects targeted by the current work show excellent agreement with the orbit, both spatially and kinematically. The six PNe, named PN1–PN6, as carefully selected from the catalog of Merrett et al. (2003), were identified to be located in the two substructures by Merrett et al. (2006). It is worth noting that the spatial extent of the stellar orbit was originally confined to within ∼4° × 4° centered on M31 (see Figure 2 of Merrett et al. 2003). The newly discovered PN7, with ${v}_{{\rm{los}}}$ measured by the LAMOST spectroscopy, was located 365 from M31's center, beyond the orbit's extent of Merrett et al. (2003). However, the spatial position of PN7 matches well the extrapolation of the projected orbit in the tangential direction (the thick green dashed line in Figure 7).

Although slight deviation between PN7 and the extrapolated orbit exists in the ${v}_{{\rm{los}}}$ versus X and ${v}_{{\rm{los}}}$ versus Y diagrams (Figure 7, bottom and right panels), the agreement is still reasonable within the possible dispersion of the orbit, as Merrett et al. (2003) claimed that the stellar stream is probably not in reality a single orbit, but a family of adjacent orbits, and the singular nature of the potential of their simplified model also amplifies the dispersion in the orbit and increases the spread in the observed ${v}_{{\rm{los}}}$ to ∼100 km s⁻¹. This model orbit is only a generic representative of the stellar streams that connect the two photometric features.

The kinematics of the 27 outer-disk/halo PNe observed by Kwitter et al. (2012, Kwitter12; 16 PNe), Balick et al. (2013, Balick13; 2 PNe), and Corradi et al. (2015, Corradi15; 9 PNe) generally resemble the rotation pattern of the classical disk of M31 (Figure 7, bottom and right panels), and can be kinematically distinguished from our sample, which belong to the substructures and are mostly well located on the stellar stream of Merrett et al. (2003). Even for the PNe at very large radii, e.g., those along the major axis of M31 observed by Corradi et al. (2015), the kinematics still follow the extended disk. The kinematics of PNe in the Northern Spur (shown as open circles in Figure 7) are indistinguishable from those of the disk, although this substructure is located some distance from the plane.

PNe probes the past chemical composition of the interstellar medium (ISM). Abundances of the α-elements such as O, Ne, S, and Ar, of a PN represent the chemistry of the ISM in the era when the progenitor star just formed. On the other hand, H ii regions trace the current abundances of the ISM. Giant extragalactic H ii regions, where massive star formation activities occur, are among the most prominent features seen in a gas-rich, star-forming galaxy. They provide probes of relatively homogeneous interstellar material that generally show little evidence of abundance enhancements produced by the evolved stars embedded in them (e.g., Wofford 2009). Comparison of the abundances of PNe and H ii regions at the same site (e.g., on the galaxy disk) is useful for studying the enrichment history of galaxies (e.g., Stanghellini et al. 2010). Investigation of relations between the abundances of different α-elements helps to constrain the stellar evolution models and quantify the relative yields of each element in the asymptotic giant branch (AGB) stars, the immediate progenitors of PNe (e.g., Kwitter & Henry 2001; Milingo et al. 2002; Kwitter et al. 2003; Henry et al. 2004). The latest review on the status of modeling the evolution and nucleosynthesis of AGB stars was presented by Herwig (2005).

Figures 8–12 show the abundance correlations of our sample as well as those for the M31 disk and bulge PNe from the literature. Overplotted in these figures are the M31 outer-disk PNe observed by Kwitter et al. (2012), Balick et al. (2013), and Corradi et al. (2015), as well as the M31 disk and bulge sample observed by Jacoby & Ciardullo (1999). Also presented in the figures (except Figure 8) are the extragalactic H ii region data, which set a baseline for the assumed tight linear behavior (in logarithmic scale) between different α elements: the M101 H ii regions of Kennicutt et al. (2003, blue dots), the H ii regions in low-metallicity blue compact galaxies observed by Izotov & Thuan (1999) and Izotov et al. (2012, green dots), and the M31 H ii regions observed by Zurita & Bresolin (2012, magenta dots). These plots also include the solar values (Asplund et al. 2009) and the Orion nebular abundances (Esteban et al. 2004).

Zoom In Zoom Out Reset image size

Figure 8 shows the N/O versus He/H (left) and N/O versus O/H (right) correlations. The Galactic disk PNe (mostly Type I) observed by Milingo et al. (2010) are also presented in this figure for purpose of comparison. The N/O and He/H ratios of our sample are generally lower than those of the Galactic Type I PNe (Figure 8, left). Our PNe have N/O ratios much lower than 0.8, a criterion to distinguish the Galactic Type I and II PNe (Kingsburgh & Barlow 1994). The seven PNe in this study, together with the three Northern Spur PNe in Paper I, are qualified to be classified as Type II, i.e., their progenitor stars have relatively low masses. The sample of Milingo et al. (2010) shows a positive correlation between N/O and He/H and a decrease in N/O with oxygen, indicating that more massive progenitors produce nitrogen at the expense of oxygen. Neither our targets nor the samples of Kwitter et al. (2012), Balick et al. (2013), and Corradi et al. (2015) show such trends as clearly as the Galactic PNe.

There is tight positive correlation between log(Ne/H) and log(O/H), as established based on the H ii region data (e.g., Milingo et al. 2010). The neon–oxygen abundance distribution of our M31 PNe as well as those of Kwitter et al. (2012), Balick et al. (2013), and Corradi et al. (2015) generally agrees with this correlation within the errors, as shown in Figure 9 (left). Figure 9 (right) shows the flatness of Ne/O in our sample. Observations of the Galactic PNe have revealed that destruction/production of the α elements, particularly in the case of neon, does occur (Milingo et al. 2010). Figure 9 shows that there is at least one outlier in the sample of Jacoby & Ciardullo (1999), indicating possible destruction of Ne. However, the uncertainties of Jacoby & Ciardullo (1999) were not presented. Argon of our sample is generally correlated with oxygen, following the pattern of H ii regions; however, the M31 outer-disk PNe of Kwitter et al. (2012), Balick et al. (2013), and Corradi et al. (2015) have systematically lower argon abundances (Figure 10). The exact cause of this systematic difference in the argon abundances is unclear, although the use of different ICFs could not be totally ruled out. The sample of Jacoby & Ciardullo (1999), as well as the Type I PNe of Milingo et al. (2010), have large scatter.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The sulfur anomaly, first noticed in Galactic PNe by Henry et al. (2004) and confirmed by Milingo et al. (2010), is also present in the M31 sample (Figure 11): the sulfur abundances of the M31 PNe are significantly lower than those of H ii regions at similar oxygen abundance. Besides, the scatter in the PN sulfur abundances for a given metallicity is much larger than that in H ii regions. The sulfur anomaly is also evident in the diagrams of S/H versus Ne/H (Figure 12, left) and S/H versus Ar/H (Figure 12, right). Henry et al. (2004) originally proposed that the apparently low sulfur abundances were likely due to ICF problems. This argument was later confirmed by Henry et al. (2012). Under the physical conditions of PNe, a significant amount of sulfur is in the triply ionized stage ${{\rm{S}}}^{3+}$, which cannot be observed in the optical. The presence of ${{\rm{S}}}^{3+}$ is also indicated by the detection of He²⁺ (e.g., the strong He ii λ4686 line of PN3), because the ionization potential of He⁺ (54.4 eV) is higher than that of ${{\rm{S}}}^{2+}$ (34.8 eV). However, (Milingo et al. 2010, also Henry et al. 2012) found that the sulfur anomaly persists even after including the ${{\rm{S}}}^{3+}$/H⁺ abundances derived from the infrared Infrared Space Observatory data. It has been suggested that the sulfur anomaly could be due to the formation of dust such as MgS and FeS (Pottasch & Bernard-Salas 2006), but the exact amount of sulfur deficit caused by this mechanism is difficult to estimate.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The He/H and N/O ratios as well as the abundance ratios of the α-elements of our PNe generally span in the ranges within those of the disk PNe observed by Kwitter et al. (2012), Balick et al. (2013), and Corradi et al. (2015). This indicates that our sample probably has similar properties as those of Kwitter et al. (2012). Adopting the linear least-squares fit to the extinction parameter c(Hβ) versus PN core–mass relation presented by (Kwitter et al. 2012, Figure 4), we estimate that the core masses of our sample are in the range ∼0.59–0.65 M_⊙. The extinction versus core mass plot of Kwitter et al. (2012) seems to indicate that this relation is insensitive to core mass below ∼0.59 M_⊙ when the extinction parameter c(Hβ) ≲0.20. Five objects in our sample are in this range. Here our extinction parameters c(Hβ) have been corrected for the Galactic average foreground extinction toward M31 (E(B − V) = 0.062; Schlegel et al. 1998), assuming the ratio c(Hβ)/E(B − V) = 1.4 (Kaler & Lutz 1985). The initial masses of our target PNe are in the range 1.81–2.26 M_⊙, as estimated using the initial-final mass relationship of white dwarfs given by (Catalán et al. 2008, Equation (1)). The progenitors' main-sequence lifetimes of our sample PNe are in the range 0.85–1.8 Gyr, which was estimated using the stellar evolution model results of Schaller et al. (1992, Table 45). Although the model tracks of Vassiliadis & Wood (1994) for the post-AGB evolution were not used to constrain the core masses of our target PNe, the range of our core masses (also the initial masses) may be narrower than those of the sample studied by Kwitter et al. (2012).

Our estimate of the PN core masses is not robust, although the extinction versus PN core mass correlation presented in Kwitter et al. (2012) can be physically explained, and has already been validated by Jacoby & Ciardullo (1999) and discussed in more detail in Ciardullo & Jacoby (1999). This correlation is a simple consequence of the greater mass loss and faster evolution times of relatively high-mass stars (Ciardullo & Jacoby 1999). The mass-loss processes through the AGB wind are complex, and the circumstellar extinction depends largely on these ejection processes as well as on the central star evolution. Thus the linear fit of Kwitter et al. (2012) may still be crude, and it is inappropriate to simply use this relation to derive the core masses. In this regard, our core masses are of large uncertainty. With the future availability of deep GTC spectra obtained for more PNe in the substructures of M31¹⁰ , we will construct detailed photoionization models for our PN sample, and more robust core masses will be derived. This will be presented in a separate paper.

Previous deep Hubble Space Telescope Advanced Camera for Surveys (ACS) imaging revealed that the stellar fields at the Giant Stream are older than those on the disk (Bernard et al. 2015), suggestive of an early-type progenitor. The stream fields are generally consistent with the direction of target PN7 (Figure 1). The stellar field of the Northern Spur targeted by Bernard et al. (2015) also coincides with the position of target PN3 (Figure 1). In all ACS fields studied, Bernard et al. (2015) found a significant burst of star formation 2 Gyr ago, in rough agreement with the main-sequence ages we estimated for our target PNe.

The radial distribution of the oxygen abundances of our sample (the red open and filled circles) is presented in Figure 13, where we also plot the M31 outer-disk/halo samples observed by Kwitter et al. (2012, black filled circles), Balick et al. (2013) and Corradi et al. (2015, black filled diamonds). The PN samples of Sanders et al. (2012, black open circles) and Jacoby & Ciardullo (1999, black open triangles) are also plotted. The M31 H ii regions observed by Zurita & Bresolin (2012) and Esteban et al. (2009) are also presented in the figure for the purpose of comparison. Galactocentric distances (in kpc) have been rectified for the effects of projection on the plane of sky assuming that all objects are located on the disk, except PN7, which is located on the extension of the Giant Stream and whose distance to the center of M31 is probably $\geqslant$ 100 kpc if considering the three-dimensional structure of the Giant Stream (McConnachie et al. 2003). We assumed an inclination angle of 777 for the M31 disk and a position angle of 377 for the M31 main axis (de Vaucouleurs 1958) in the projection rectification.

Zoom In Zoom Out Reset image size

The M31 disk sample observed by Kwitter et al. (2012) yielded an oxygen gradient of −0.011 ± 0.004 dex kpc⁻¹ (the target PN16, whose PN ID in Merrett et al. 2006 is 1074, was not included in gradient fit because of absence of the [O ii] λ3727 line in the spectrum). The two distant (halo) PNe observed by Balick et al. (2013) have oxygen abundances comparable to that of the Sun (Figure 13), indicating a possible flattening in the outer disk. Our sample, including the three Northern Spur PNe in Paper I, spans a wide range in galactocentric distances and has rather homogeneous oxygen abundances. The oxygen abundance of PN7, the most distant PN in our sample, is consistent with those of the two outer-disk PNe of Balick et al. (2013) within the uncertainties. However, they belong to different groups: PN7 is associated with the Giant Stream both spatially and kinematically, while the kinematics of Balick's sample follow the rotation pattern of the classical disk of M31 (Figure 7). PN5 and PN6 are close to the galaxy center and both have oxygen abundances close to that of PN7. These three PNe are well associated with the extension of the Giant Stream of M31 (Figure 7).

PN1, PN2, and PN3 are located in the Northern Spur and their oxygen abundances are consistent with those in Paper I, which belong to the same substructure. The average 12 + log(O/H) of our sample (10 PNe) is 8.56 ± 0.10, about 0.1 dex below the solar value (8.69, Asplund et al. 2009), while the mean value of the 18 outer-disk PNe of Kwitter et al. (2012) and Balick et al. (2013) is 8.62 ± 0.14, which is consistent with ours within dispersion. The O/H ratios of our 10 PNe are also generally consistent within the errors. It is also worth noting that PN5 and PN6 have oxygen abundances that are higher than the PN samples of Jacoby & Ciardullo (1999) and Sanders et al. (2012) at similar galactocentric distances, although the latter two both show large scatter. The O/H ratios of these two PNe are also higher than those of the M31 H ii regions observed by Esteban et al. (2009) and Zurita & Bresolin (2012).

Of the carefully selected targets in this work, three belong to the Northern Spur and four are associated with the Giant Stream substructure, and have homogeneous oxygen abundances. Although they cannot be clearly distinguished from the outer-disk sample of Balick et al. (2013) by the abundances, their kinematics generally follow the stellar orbit that connects the Northern Spur and the Giant Stream. The most intriguing object in our sample is PN7, a faint PN in the extension of the Giant Stream with a sky-projected angular distance of 365. This PN is located probably >100 kpc from the center of M31, according to the three-dimensional structure of the Giant Stream (McConnachie et al. 2003). The similarity in its oxygen abundance with those of PN5 and PN6, which are both close to the bulge of M31 (Figure 13), confirms that these three Stream PNe probably originated from the same stellar population. The spatial, kinematical, and chemical information points to the possibility that these PNe have the same origin, and provides observational support for the stellar orbit, which was proposed by Merrett et al. (2003) based on the kinematic study of ∼20 PNe spatially located on the M31 disk.

Ibata et al. (2001a) initially suggested that either or both M32 and NGC 205 could be responsible for the Giant Stream, based on the geometrical alignment of the two satellites. The panoramic survey of RGB stars in the M31 halo by Ferguson et al. (2002) reveals that M32 seems to be a more promising candidate for the origin of the stream. These authors also inferred a possible connection between the Giant Stream and the extent of the Northern Spur, and even pointed out a strong warp of this connection orbit (if it does exist) near M31's nucleus. Based on a kinematic study of ∼20 disk PNe, Merrett et al. (2003) constructed a model orbit that connects the Northern Spur and the Giant Stream, and proposed that the two substructures are the tidal tails (one is the leading part, and the other the trailing part) of M32, given that this dwarf elliptical is close to the orbit both in position (projected on the X–Y plane) and velocity. Merrett et al. (2003) also provided two extra arguments to support this hypothesis: (1) the RGB in the Northern Spur and the Giant Stream are both particularly red compared with the rest of the halo of M31 (Ferguson et al. 2002), and (2) M32 appears to be highly tidally stripped. However, the exact position of M32 with respect to the disk of M31 remains yet to be determined (Mateo 1998).

The galaxy And viii is a tidally disrupted satellite of M31, and it was located very close to M32 in the X–Y plane (Figure 7) and might also be a candidate for the origin of the stellar stream (Merrett et al. 2003). This satellite was found to be strongly distorted, with a length of ∼10 kpc and a width of a few kiloparsecs, and contains 5–12 PNe and has a systematic velocity of −204 km s⁻¹ with respect to M31 (Morrison et al. 2003). And viii occupies the position where the Giant Stream meets the disk of M31, and the PNe in this satellite kinematically coincide with the model orbit (Figure 7, bottom). However, the kinematics of And viii is opposite to that of M32, casting doubts on its possible origin in the stellar stream. Given the large scatter of the stellar stream both in positions and line of sight velocities, as discussed in Merrett et al. (2003), M32 is associated with the orbit. Although spectroscopic observations of PNe in M32 have been carried out (Richer et al. 1999), accurate abundance determination is scarce. Future spectroscopic study of M32 PNe using the 8–10 m class telescopes will help to confirm their true nature.