A stellar census of the nearby, young 32 Orionis group

TL;DR

This paper presents the first large-scale spectroscopic survey of the 32 Orionis group, identifying new members, determining its age as approximately 24 million years, and estimating its debris disc fraction, thereby enhancing understanding of this nearby young stellar group.

Contribution

It provides the first comprehensive membership catalog, age estimation, and debris disc fraction for the 32 Orionis group, using combined kinematic, photometric, and spectroscopic data.

Findings

Identified 30 new group members, tripling the known population.

Estimated the group's age as 24±4 million years.

Determined a debris disc fraction of approximately 32%.

Abstract

The 32 Orionis group was discovered almost a decade ago and despite the fact that it represents the first northern, young (age ~ 25 Myr) stellar aggregate within 100 pc of the Sun ($d \simeq 93$ pc), a comprehensive survey for members and detailed characterisation of the group has yet to be performed. We present the first large-scale spectroscopic survey for new (predominantly M-type) members of the group after combining kinematic and photometric data to select candidates with Galactic space motion and positions in colour-magnitude space consistent with membership. We identify 30 new members, increasing the number of known 32 Ori group members by a factor of three and bringing the total number of identified members to 46, spanning spectral types B5 to L1. We also identify the lithium depletion boundary (LDB) of the group, i.e. the luminosity at which lithium remains unburnt in a coeval population. We estimate the age of the 32 Ori group independently using both isochronal fitting and LDB analyses and find it is essentially coeval with the {\beta} Pictoris moving group, with an age of $24\pm4$ Myr. Finally, we have also searched for circumstellar disc hosts utilising the AllWISE catalogue. Although we find no evidence for warm, dusty discs, we identify several stars with excess emission in the WISE W4-band at 22 {\mu}m. Based on the limited number of W4 detections we estimate a debris disc fraction of $32^{+12}_{-8}$ per cent for the 32 Ori group.

Tables

Table 1. Table 1: Properties of the 15 known stellar members of the 32 Ori group.

Name	$\alpha$ (J2000.0)	$\delta$ (J2000.0)	SpT	Ref.	${\mu }_{\alpha }\cos \left(\delta \right)$	${\mu }_{\delta }$	Ref.	RV	Ref.
	(hh mm ss.ss)	(hh mm ss.s)			($\mathrm{mas}{\mathrm{yr}}^{-1}$)	($\mathrm{mas}{\mathrm{yr}}^{-1}$)		($\mathrm{km}{s}^{-1}$)
32 Ori	05 30 47.05	+05 56 53.3	B5IV+B7V	1	$5.10\pm 0.67$	$-33.30\pm 0.35$	2	$18.6\pm 1.2$	3
HR 1807	05 26 38.83	+06 52 07.2	B9.5V	4	$9.22\pm 0.56$	$-33.15\pm 0.31$	2	$13.1\pm 2.5$	5
HD 35714	05 26 59.99	+07 10 13.0	A3	6	$9.34\pm 0.04$	$-32.97\pm 0.03$	7	$35.4\pm 1.0$	8
HD 36823	05 34 38.42	+06 07 36.7	A7.5	9	$9.07\pm 0.05$	$-31.67\pm 0.02$	7	–	–
HD 35499	05 25 14.56	+04 11 48.2	F4	6	$8.29\pm 0.54$	$-29.02\pm 0.56$	7	–	–
HD 36338	05 31 15.70	+05 39 46.4	F4.5	6	$9.48\pm 0.48$	$-32.76\pm 0.54$	7	–	–
HD 35695	05 26 52.03	+06 28 22.7	F9	6	$9.20\pm 0.80$	$-36.10\pm 1.20$	10	–	–
HD 245567	05 37 18.43	+13 34 52.5	G5	6	$7.20\pm 0.90$	$-33.50\pm 0.80$	10	$14.9\pm 0.8$	11
HD 245059	05 34 34.91	+10 07 06.4	G7	6	$10.10\pm 1.30$	$-35.00\pm 1.20$	10	$19.8\pm 1.0$	12
TYC 112-1486-1	05 20 31.82	+06 16 11.6	K3	6	$9.50\pm 1.80$	$-32.80\pm 2.10$	10	$18.5\pm 0.2$	13
TYC 112-917-1	05 20 00.29	+06 13 03.7	K4	6	$9.40\pm 1.90$	$-34.70\pm 2.10$	10	$18.8\pm 0.1$	13
2MASS J05234246+0651581	05 23 42.46	+06 51 58.2	K6.5	6	$7.80\pm 2.30$	$-36.60\pm 2.80$	10	$18.4\pm 1.0$	12
V1874 Ori†	05 29 19.00	+12 09 29.6	K6.5	6	$2.90\pm 2.10$	$-26.70\pm 3.00$	10	$18.4\pm 0.3$	14
2MASS J05253253+0625336	05 25 32.54	+06 25 33.7	M3	6	$8.00\pm 5.80$	$-28.20\pm 6.10$	10	–	–
2MASS J05194398+0535021	05 19 43.98	+05 35 02.2	M3	6	$5.80\pm 4.00$	$-27.80\pm 4.00$	15	–	–

Table 2. Table 2: Mean properties of the 32 Ori group (Mamajek 3).

$\alpha$ (J2000.0)	$\delta$ (J2000.0)	$\pi$†	${\mu }_{\alpha }\cos \left(\delta \right)$	${\mu }_{\delta }$	RV‡	$U$	$V$	$W$
(hh mm ss.ss)	(hh mm ss.s)	(mas)	($\mathrm{mas}{\mathrm{yr}}^{-1}$)	($\mathrm{mas}{\mathrm{yr}}^{-1}$)	($\mathrm{km}{s}^{-1}$)	($\mathrm{km}{s}^{-1}$)	($\mathrm{km}{s}^{-1}$)	($\mathrm{km}{s}^{-1}$)
05 27 16.32	+06 40 37.2	$10.49\pm 0.22$	$8.6\pm 0.5$	$-32.6\pm 0.5$	$18.6\pm 0.2$	$-11.8\pm 0.3$	$-18.9\pm 0.3$	$-8.9\pm 0.3$

Table 3. Table 3: Spectroscopic parameters and membership assignments for the candidate members and 11 known members of the 32 Ori group observed with WiFeS. Four additional early-type stars and a brown dwarf confirmed as members but not observed in this work are also included. HD 36002 and SCR 0522 − - 0606 were not in the original candidate list but were added and confirmed during observations. The full table, which also includes possible and non-members as well as derived E ​ ( B − V ) 𝐸 𝐵 𝑉 E(B-V) values and object-specific comments (omitted in the print version due to space restrictions), is available as Supporting Information with the online version of the paper.

THOR	2MASS J	Other name	$V$	Ref.	$K_s$	${\mu }_{\alpha }\cos \left(\delta \right)$	${\mu }_{\delta }$	Ref.	$d_{\mathrm{kin}}$	$d_{\mathrm{trig}}$	Ref.	RV	${\Delta }_{\mathrm{RV}}$	SpT	EW[H$\alpha$]	EW[Li]	Membership
#	designation		(mag)		(mag)	(mas yr-1)	(mas yr-1)		(pc)	(pc)		($\mathrm{km}{s}^{-1}$)	($\mathrm{km}{s}^{-1}$)		($\AA$)	($m\AA$)	RV,H$\alpha$,Li	Final
1	05304706+0556536	32 Ori	4.20	1	4.61	$12.9\pm 1.0$	$-28.1\pm 1.0$	2	102	${93}_{-5}^{+6}$	3	…	…	B5+B7	…	…	…	Y
2	05263883+0652071	HR 1807	6.41	4	6.41	$9.2\pm 1.0$	$-33.2\pm 1.0$	2	92	$92\pm 4$	3	…	…	B9.5	…	…	…	Y
3A	05265999+0710131	HD 35714	7.03	5	6.75	$8.4\pm 1.0$	$-34.2\pm 1.0$	2	91	${94}_{-4}^{+3}$	6	…	…	A3	…	…	…	Y
3B	05343842+0607367	HD 36823	7.73	5	7.14	$9.0\pm 0.6$	$-31.1\pm 0.5$	2	95	$102\pm 5$	6	…	…	A7.5	…	…	…	Y
4A	05284209+0113369	HD 36002	7.46	5	6.71	$6.9\pm 1.0$	$-25.6\pm 1.0$	2	105	${109}_{-5}^{+6}$	6	SB2	…	A7	14	$\sim$70	?,?,?	Y
4B	05284050+0113333		14.74	7	10.19	$8.8\pm 5.8$	$-26.6\pm 5.8$	8	99	…	…	$18.4\pm 1.0$	$-$1.5	M3.5	$-$4.8	50	Y,?,?	Y
5	05251457+0411482	HD 35499	8.66	5	7.50	$6.1\pm 0.6$	$-27.5\pm 0.8$	2	106	${105}_{-6}^{+8}$	9	$20.6\pm {2.6}^{\star }$	$+$1.4	F4	8.0	160	Y,?,Y	Y
6	05311570+0539461	HD 36338	8.52	5	7.40	$9.3\pm 0.6$	$-31.1\pm 0.9$	2	95	$96\pm 5$	9	$21.3\pm 0.2$	$+$2.4	F4.5	7.0	110	Y,?,Y	Y
7	05265202+0628227	HD 35695	9.27	5	7.71	$9.2\pm 0.8$	$-36.1\pm 1.2$	2	84	…	…	$20.0\pm {0.3}^{\star }$	$+$1.4	F9	5.0	280	Y,?,Y	Y
8A	05371844+1334525	HD 245567	9.54	5	7.59	$7.2\pm 0.9$	$-33.5\pm 0.8$	2	104	…	…	$16.0\pm 1.3$	$-$0.8	G5	$-$0.2§	265	Y,?,Y	Y
8B	05372061+1335310		14.96	7	10.07	$6.9\pm 6.2$	$-30.8\pm 5.7$	2	113	…	…	$15.7\pm 0.8$	$-$1.1	M3	$-$4.8	30	Y,?,?	Y
9	05343491+1007062	HD 245059	9.96	5	7.41	$10.1\pm 1.3$	$-35.0\pm 1.2$	2	92	…	…	$19.4\pm 1.1$	$+$1.6	G7	$-$0.2	285	Y,?,Y	Y
10	05203182+0616115	TYC 112-1486-1	11.87	7	8.57	$9.5\pm 1.8$	$-32.8\pm 2.1$	2	92	${99}_{-11}^{+14}$	9	$21.1\pm 0.3$	$+$2.5	K3	$-$1.2	500	Y,?,Y	Y
11	05200029+0613036	TYC 112-917-1	11.66	7	8.57	$9.4\pm 1.9$	$-34.7\pm 2.1$	2	86	…	…	$21.0\pm 0.3$	$+$2.4	K4	$-$0.3	470	Y,?,Y	Y
12	05234246+0651581	1RXS J052342.7+065156	12.81	7	9.03	$7.8\pm 2.3$	$-36.6\pm 2.8$	2	85	…	…	$18.3\pm {4.2}^{\star }$	$-$0.2	K6.5	$-$5.0	580	Y,?,Y	Y
13	05291899+1209295	V1874 Ori	13.22	7	9.19	$2.9\pm 2.1$	$-26.7\pm 3.0$	2	131	…	…	$22.4\pm 0.3$	$+$5.3	K6.5	$-$1.7	380	Y,?,Y	Y
14A	05351761+1354180		13.53	7	9.15	$7.0\pm 4.4$	$-30.2\pm 9.4$	2	116	…	…	$18.1\pm {2.8}^{\star }$	$+$1.4	M1.5	$-$1.7	180	Y,?,Y	Y
14B	05351625+1353594	1RXS J053516.6+135404	14.95	10	10.39	$7.6\pm 4.3$	$-31.8\pm 4.6$	2	110	…	…	$19.2\pm 1.0$	$+$2.5	M3.5	$-$6.1	40	Y,?,?	Y
15	05330574+1400365	1RXS J053306.7+140011	13.80	7	9.35	$9.4\pm 2.4$	$-37.4\pm 2.2$	2	94	…	…	$19.2\pm {2.9}^{\star }$	$+$2.5	M2	$-$2.0	160	Y,?,Y	Y
16	05430354+0606340	1RXS J054304.3+060646	14.98	7	10.29	$11.5\pm 5.4$	$-25.2\pm 5.9$	2	114	…	…	$21.0\pm 0.9$	$+$2.1	M2.5	$-$2.9	50	Y,?,?	Y
17A	05274313+1446121	1RXS J052743.4+144609	14.13	7	9.10	$8.6\pm 3.8$	$-41.1\pm 3.8$	2	87	…	…	$15.8\pm {1.1}^{\star }$	$-$0.6	M3	$-$10.3§	100	Y,?,?	Y
17B	05274404+1445584		16.26	10	10.97	$6.5\pm 6.0$	$-44.4\pm 6.7$	2	82	…	…	$15.4\pm 0.8$	$-$1.0	M5	$-$5.9	580	Y,?,Y	Y
18	05224069–0606238	SCR 0522–0606	14.27	7	9.13	$17.0\pm 3.2$	$-21.1\pm 3.3$	2	88	…	…	$25.6\pm {1.9}^{\star }$	$+$4.3	M3	$-$7.0	20	Y,?,?	Y
19	05253253+0625336	1RXS J052532.3+062534	14.51	7	9.78	$8.0\pm 5.8$	$-28.2\pm 6.1$	2	107	…	…	$19.6\pm {0.4}^{\star }$	$+$1.0	M3	$-$6.6	20	Y,?,?	Y
20	05192941+1038081	1RXS J051930.4+103812	14.66	7	9.77	$8.5\pm 7.1$	$-32.9\pm 5.1$	2	101	…	…	$17.7\pm {0.8}^{\star }$	$+$0.3	M3	$-$7.1	80	Y,?,?	Y
21	05251517+0030232	1RXS J052515.0+003027	14.85	7	9.93	$10.3\pm 5.8$	$-24.7\pm 6.0$	2	103	…	…	$19.9\pm 0.7$	$-$0.1	M3	$-$9.7§	20	Y,?,?	Y
22	05492632+0405379	1RXS J054926.3+040541	14.96	7	10.07	$11.2\pm 5.5$	$-33.3\pm 5.9$	2	84	…	…	$20.9\pm 1.0$	$+$1.4	M3	$-$2.5	50	Y,?,?	Y
23	05442447+0502114	1RXS J054424.7+050153	15.08	7	10.24	$15.8\pm 5.4$	$-28.4\pm 5.8$	2	98	…	…	$24.4\pm {2.6}^{\star }$	$+$5.2	M3	$-$4.4	30	Y,?,?	Y
24	05194398+0535021		15.13	7	10.00	$-4.2\pm 16.4$	$-29.3\pm 13.3$	2	116	…	…	$15.1\pm 0.8$	$-$3.6	M3	$-$15.0	20	Y,?,?	Y
25	05132631+1057439		15.25	7	10.22	$7.3\pm 5.9$	$-30.2\pm 5.9$	8	113	…	…	$17.0\pm 0.7$	$-$0.2	M3	$-$5.2	50	Y,?,?	Y
26	05302546–0256255		15.65	7	10.66	$12.1\pm 2.7$	$-25.7\pm 3.2$	2	88	…	…	$25.1\pm 1.2$	$+$4.3	M3	$-$10.0§	90	Y,?,?	Y
27	05274855+0645459	1RXS J052748.7+064544	14.75	7	9.47	$10.9\pm 5.9$	$-29.1\pm 6.3$	2	102	…	…	$18.8\pm 0.8$	$+$0.2	M4	$-$6.6	50	Y,?,?	Y
28	05264886+0928055		15.64	11	10.47	$6.7\pm 5.3$	$-32.7\pm 6.1$	2	100	…	…	$22.4\pm 0.8$	$+$4.6	M4	$-$3.9	70	Y,?,?	Y
29	05231438+0643531	1RXS J052315.0+064412	15.86	10	10.76	$3.2\pm 5.9$	$-28.4\pm 5.9$	8	113	…	…	$21.0\pm {0.6}^{\star }$	$+$2.4	M4	$-$3.3	50	Y,?,?	Y
30	05313290+0556597	1RXS J053132.6+055639	15.90	7	10.53	$-0.8\pm 5.7$	$-37.4\pm 5.7$	8	85	…	…	$24.7\pm 1.1$	$+$5.9	M4.5	$-$6.4	80	Y,?,?	Y
31	05363692+1300369		16.10	11	10.39	$4.6\pm 4.4$	$-36.3\pm 4.6$	2	96	…	…	$18.0\pm {6.6}^{\star }$	$+$1.0	M4.5	$-$17	350–500	Y,?,Y	Y
32	05264073+0712255		16.32	7	10.76	$1.9\pm 4.1$	$-32.8\pm 4.4$	2	99	…	…	$18.8\pm 0.5$	$+$0.4	M4.5	$-$4.6	600	Y,?,Y	Y
33	05315786–0303367		13.85	7	8.54	$4.2\pm 6.6$	$-34.8\pm 4.9$	2	71	…	…	$23.1\pm 1.0$	$+$2.3	M5‡	$-$9.0	150	Y,?,?	Y
34	05320596–0301159		15.61	7	9.70	$9.2\pm 4.9$	$-27.4\pm 4.9$	2	85	…	…	$23.6\pm {2.7}^{\star }$	$+$2.8	M5	$-$11.0	650	Y,?,Y	Y
35	05174962+0958221		16.63	7	10.84	$22.3\pm 5.9$	$-33.3\pm 5.9$	8	89	…	…	$18.1\pm 1.6$	$+$0.6	M5	$-$6.9	530	Y,?,Y	Y
36	05235565+1101027	1RXS J052355.2+110110	16.65	11	10.78	$7.4\pm 6.3$	$-30.3\pm 6.3$	8	110	…	…	$18.1\pm 1.2$	$+$0.7	M5	$-$5.8	680	Y,?,Y	Y
37	05350092+1125423		16.84	11	11.02	$7.1\pm 6.6$	$-43.0\pm 6.7$	2	79	…	…	$17.8\pm {4.0}^{\star }$	$+$0.4	M5	$-$6.7§	700	Y,?,Y	Y
38	05243009+0640349		16.97	10	11.13	$8.8\pm 6.0$	$-36.8\pm 6.0$	8	84	…	…	$16.1\pm 1.0$	$-$2.4	M5	$-$14.0	750	Y,?,Y	Y
39	05270634+0650377		17.60	11	11.64	$10.1\pm 5.9$	$-35.8\pm 5.9$	8	85	…	…	$20.8\pm 1.0$	$+$2.3	M5	$-$8.5	650	Y,?,Y	Y
40	05373000+1329344		16.51	10	10.78	$3.2\pm 6.4$	$-32.1\pm 6.4$	8	110	…	…	$18.5\pm 1.2$	$+$1.6	M5.5	$-$13.2	630	Y,?,Y	Y
41	…	WISE J052857.68+090104.4	…	…	14.97†	$-11.0\pm 10.0$	$-39.0\pm 12.0$	12	93	…	…	…	…	L1	…	…	…	Y

Table 4. Table 4: Individual WiFeS radial velocities of candidates observed at more than one epoch. Stars with a THOR number (Column 1) are confirmed members of the 32 Ori group. The final column denotes whether the star is a confirmed or suspected spectroscopic binary. Three additional resolved SB systems are listed with their component velocities in Table  5 .

THOR	Name	Epoch	RV	SB?
#		(UT)	(km s-1)
5	HD 35499	2015 Oct 25	$23.2\pm 0.8$	N
		2016 Oct 13	$18.0\pm 0.6$
7	HD 35695	2015 Oct 25	$20.3\pm 0.5$	N
		2016 Oct 13	$19.7\pm 1.3$
12	2M0523+0651	2015 Oct 23	$15.0\pm 2.2$	SB2?
		2016 Oct 13	$24.2\pm 1.4$
		2017 Jan 07	$15.6\pm 1.9$
14A	2M0535+1354	2016 Jan 23	$21.5\pm 0.7$	SB1?
		2017 Jan 07	$18.1\pm 0.9$
		2017 Jan 07	$14.7\pm 1.2$
15	2M0533+1400	2015 Oct 19	$23.1\pm 0.6$	SB1?
		2015 Oct 20	$18.5\pm 0.7$
		2017 Jan 07	$16.1\pm 1.2$
17A	2M0527+1446	2015 Oct 20	$14.4\pm 0.8$	N
		2016 Jan 26	$17.0\pm 1.0$
		2016 Jan 30	$16.0\pm 1.0$
18	SCR 0522$-$0606	2016 Oct 14	$23.7\pm 1.7$	SB2
		2017 Jan 07	$27.4\pm 1.7$
19	2M0525+0625	2015 Oct 23	$19.2\pm 0.7$	N
		2015 Oct 25	$20.0\pm 0.6$
20	2M0519+1038	2015 Oct 19	$18.4\pm 1.1$	N
		2015 Oct 19	$16.9\pm 0.9$
23	2M0544+0502	2016 Jan 23	$21.8\pm 1.8$	N
		2016 Oct 14	$26.9\pm 1.6$
29	2M0523+0643	2016 Jan 25	$20.4\pm 1.0$	N
		2016 Jan 29	$21.5\pm 0.9$
31	2M0536+1300	2016 Jan 25	$16.3\pm 1.5$	SB2
		2016 Oct 13	$28.4\pm 2.1$
		2016 Oct 14	$17.4\pm 2.4$
		2017 Jan 07	$10.0\pm 3.0$
34	2M0532$-$0301	2016 Jan 29	$26.2\pm 1.3$	N
		2017 Jan 08	$20.9\pm 2.6$
37	2M0535+1125	2016 Feb 18	$19.6\pm 1.9$	N?
		2016 Feb 21	$10.9\pm 1.4$
		2016 Oct 14	$21.0\pm 1.7$
		2017 Jan 08	$19.5\pm 0.9$
…	BD+08 900A	2015 Sep 28	$19.3\pm 1.1$	N
		2015 Sep 28	$20.6\pm 1.0$
		2016 Oct 13	$18.9\pm 0.7$
		2017 Jan 07	$22.3\pm {4.2}^{†}$
		2017 Jan 07	$19.3\pm {4.5}^{†}$
…	BD+08 900B	2015 Sep 28	$20.8\pm 1.3$	SB2?
		2015 Sep 28	$17.9\pm 0.9$
		2016 Oct 13	$20.7\pm 0.7$
		2017 Jan 07	$21.3\pm {5.9}^{†}$
		2017 Jan 07	$20.9\pm {6.9}^{†}$
…	HD 243086	2016 Jan 24	$23.7\pm 1.7$	N
		2016 Feb 21	$21.9\pm 1.0$
…	2M0556+0803	2016 Feb 21	$28.3\pm 1.3$	SB2
		2017 Jan 09	$23.9\pm 2.7$

Table 5. Table 5: Individual radial velocities for the three systems which could be resolved into two components. HD 37825 is a clear SB3 system (see Fig.  6 ) with a weaker third component not resolved in the 2017 Jan CCF. 2M0552 − - 0044 shows a single-peaked CCF but has a variable, double-peaked H α 𝛼 \alpha emission line from which a velocity offset was estimated and added to the primary velocity.

Name	Epoch	RV1	RV2
	(UT)	(km s-1)	(km s-1)
HD 36002 (SB2)	2017 Jan 11	$+69.8\pm {3.3}^{†}$	$-32.8\pm {3.0}^{†}$
	2017 Jan 11	$+58.2\pm {2.0}^{†}$	$-27.0\pm {3.0}^{†}$
HD 37825 (SB3)	2016 Jan 24	${25.3}^{‡}$	…
	2016 Feb 21	$+28.1\pm 0.8$	…
	2016 Oct 13	$-42.5\pm 2.5$	$+95.3\pm 1.9$
	2017 Jan 07	$-42.1\pm {3.0}^{†}$	$+101.9\pm {1.8}^{†}$
2M0552$-$0044	2015 Oct 23	$+65.5\pm 0.7$	$-99.5\pm 5.4$
(SB1+H$\alpha$)	2016 Oct 14	$-39.3\pm 1.5$	$+160.7\pm 5.6$
	2017 Jan 07	$+18.4\pm 1.6$	…
	2017 Jan 07	$-34.0\pm 1.0$	$+136.0\pm 5.5$
	2017 Jan 08	$-26.4\pm 0.8$	$+163.6\pm 5.4$
	2017 Jan 08	$-16.6\pm 1.0$	$+133.4\pm 5.5$
	2017 Jan 09	$-22.9\pm 0.7$	$+157.1\pm 5.4$

Table 6. Table 6: Wide binaries discovered in this study. Column 1 lists the THOR number for each component as assigned in this study, whereas Column 3 provides the Washington Double Star (WDS) catalogue designation assigned based on the position of the primary.

THOR	Component	WDS	Epoch (Ref.)	PA	Sep.	$V$
#	designation	designation	(yr)	(°)	(arcsec)	(mag)
4A	HD 36002	05287+0113	2000.08 (2MASS)	261.46	24.01	7.46
4B	2MASS J05284050+0113333		2015.00 (Gaia)	261.16	24.01	14.74
8A†	HD 245567	05373+1335	1998.74 (2MASS)	39.46	49.79	9.54
8B‡	2MASS J05372061+1335310		2015.00 (Gaia)	39.47	49.82	14.96
14A	2MASS J05351761+1354180	05353+1354	1998.74 (2MASS)	226.71	27.12	13.53
14B	2MASS J05351625+1353594		2015.00 (Gaia)	226.07	27.27	14.95
17A	2MASS J05274313+1446121	05277+1446	1998.73 (2MASS)	135.95	18.98	14.13
17B	2MASS J05274404+1445584		2015.00 (Gaia)	135.05	18.95	16.26

Table 7. Table 7: Preliminary orbital and physical parameters of 2MASS J05525572 − - 0044266, derived from a least squares Keplerian orbit fit to the primary and secondary velocities from Table  5 . Uncertainties were propagated from the covariance matrix of the fit. Similar orbital parameters are obtained fitting only the primary velocities.

Period, $P$	0.8589884 d (fixed)
Eccentricity, $e$	$0.10\pm 0.11$
Primary velocity semi-amplitude, $K_1$	$51.6\pm 3.0$ km s-1
Secondary velocity semi-amplitude, $K_2$	$137.6\pm 7.1$ km s-1
Systemic velocity, $\gamma$	$20.9\pm 2.3$ km s-1
Epoch of periastron passage (MJD), $\tau$	$57319.228\pm 0.064$
Orientation of periastron, $\omega$	$211\pm {33}^{\circ }$
Mass ratio, $q=M_2/M_1$	$0.375\pm 0.029$
Total mass, $\left(M_1+M_2\right){\sin }^{3}i$	$0.602\pm 0.074M_{\odot }$
Semi-major axis, $a\sin i$	$0.01494\pm 0.00061$ au
	$3.21\pm 0.13$ R⊙
Assuming inclination $i={90}^{\circ }$:
Primary mass, $M_1$	$0.438\pm 0.058M_{\odot }$
Secondary mass, $M_2$	$0.164\pm 0.019M_{\odot }$

Table 8. Table 8: Reddening estimates for the known B-, A- and F-type members of the 32 Ori group.

THOR	Name	$E\left(B-V\right)$	Method
#	(mag)
1AB	32 Ori AB	0.01	Q-method ($UBV$)
1A	32 Ori A	0.01	$B-V$, SpT
1B	32 Ori B	0.06	$B-V$, SpT
2	HR 1807	0.00	Q-method ($UBV$)
2	HR 1807	$0.01\pm 0.01$	SED ($UBVJH{K}_s$, SpT)
3A	HD 35714	$0.05\pm 0.02$	SED ($BVJH{K}_s$, SpT)
3B	HD 36823	0.02	$B-V$, SpT
4A	HD 36002	0.03	$B-V$, SpT
5	HD 35499	$0.04\pm 0.02$	SED ($BVJH{K}_s$, SpT)
6	HD 36338	$0.03\pm 0.02$	SED ($BVJH{K}_s$, SpT)
7	HD 35695	$0.06\pm 0.02$	SED ($BVJH{K}_s$, SpT)
	Median	$0.03\pm 0.01$

Table 9. Table 9: Isochronal and LDB ages for the 32 Ori group.

Method	Model	Age (Myr)	Mean age (Myr)
Isochrone	Dartmouth	${23}_{-3}^{+7}$
	BHAC15	${28}_{-4}^{+3}$	$25\pm 5$ ($1\sigma$)
	Pisa	${20}_{-1}^{+4}$	[$\pm 4$ (statistical), $\pm 3$ (systematic)]
	PARSEC	$29\pm 3$
LDB	Dartmouth	${22}_{-3}^{+4}$
	BHAC15	${22}_{-3}^{+4}$	$23\pm 4$ ($1\sigma$)
	Pisa	$21\pm 3$	[$\pm 4$ (statistical), $\pm 1$ (systematic)]
	Dartmouth ($⟨Bf⟩=2.5\mathrm{kG}$)	$26\pm 4$
Final adopted age	$\mathrm{𝟐𝟒}\pm \mathrm{𝟒}$ ($\mathrm{𝟏}𝝈$)
	[$\pm \mathrm{𝟒}$ (statistical), $\pm \mathrm{𝟐}$ (systematic)]

Table 10. Table 10: Mid-infrared photometry and excesses for 32 Ori group debris disc candidates.

THOR	WISE J	$W1$	$W2$	$W3$	$W4$	$E\left(K_s-W1\right)$	$E\left(K_s-W2\right)$	$E\left(K_s-W3\right)$	$E\left(K_s-W4\right)$
#	designation	(mag)	(mag)	(mag)	(mag)	(mag)	(mag)	(mag)	(mag)
2	052638.83+065206.9	$6.449\pm 0.079$	$6.435\pm 0.022$	$6.312\pm 0.016$	$4.766\pm 0.031$	$-0.019\pm 0.082$	$0.000\pm 0.032$	$0.132\pm 0.028$	$1.715\pm 0.039$
4B	052840.51+011333.1	$10.047\pm 0.023$	$9.879\pm 0.020$	$9.705\pm 0.042$	$8.370\pm 0.304$	$-0.031\pm 0.031$	$-0.003\pm 0.029$	$0.091\pm 0.047$	$1.286\pm 0.305$
6	053115.70+053946.1	$7.374\pm 0.035$	$7.405\pm 0.020$	$7.373\pm 0.018$	$6.606\pm 0.067$	$-0.019\pm 0.039$	$-0.055\pm 0.026$	$0.012\pm 0.025$	$0.694\pm 0.069$
17A	052743.14+144611.7	$8.899\pm 0.023$	$8.728\pm 0.021$	$8.604\pm 0.028$	$7.894\pm 0.216$	$0.037\pm 0.032$	$0.088\pm 0.030$	$0.132\pm 0.036$	$0.692\pm 0.217$
27	052748.55+064545.6	$9.306\pm 0.023$	$9.111\pm 0.020$	$8.918\pm 0.034$	$7.911\pm 0.238$	$-0.030\pm 0.030$	$0.005\pm 0.028$	$0.118\pm 0.039$	$0.995\pm 0.239$
38	052430.10+064034.6	$10.933\pm 0.023$	$10.684\pm 0.020$	$10.157\pm 0.072$	$8.390\pm 0.324$	$0.020\pm 0.032$	$0.019\pm 0.030$	$0.456\pm 0.075$	$2.123\pm 0.325$

Equations

$${\Delta }_{\mathrm{PM}}={\left[({\mu }_{\alpha }\cos {\delta }_{\mathrm{obs}}-{\mu }_{\alpha }\cos {\delta }_{\mathrm{expt}}{)}^2+({\mu }_{\delta \mathrm{obs}}-{\mu }_{\delta \mathrm{expt}}{)}^2\right]}^{1/2},$$

Full text

††pubyear: 2017

1 Introduction

The region surrounding the Sun out to a distance of $\sim$ 100 pc is often referred to as the ‘Local Bubble’ on account of the relatively low density of the interstellar medium and the accompanying lack of active star-forming regions. Hence the discovery just over three decades ago of young, seemingly isolated T-Tauri stars in close proximity to the Sun was a watershed moment and precipitated a massive observational effort to characterise the young population of the Local Bubble (see e.g. Rucinski & Krautter, 1983; de la Reza et al., 1989; Gregorio-Hetem et al., 1992; Webb et al., 1999). Our understanding of this young population has increased dramatically since these early discoveries with the advent of all-sky astrometric and X-ray/UV surveys. Not only have hundreds of additional young stars been identified and spectroscopically characterised within this region, but more importantly, it has been demonstrated that many of these stars comprise kinematically distinct, yet spatially dispersed groups within which the members share a common motion through space (see e.g. Zuckerman & Webb, 2000; Torres et al., 2000). To date approximately one dozen such groups have been identified within 100 pc with ages ranging from $\sim$ 10 to 200 Myr (see reviews by Zuckerman & Song, 2004, Torres et al., 2008 and Mamajek, 2016).

The study of nearby young moving groups plays an important role in constraining theories of star and planet formation. Given their youth and proximity to Earth, these moving groups provide the best available samples to investigate the early evolution of low- to intermediate-mass stars (see e.g. Mamajek & Bell, 2014). In addition to their obvious importance regarding stellar astrophysics, and in particular the pre-main-sequence (pre-MS) phase, these groups also represent the most readily accessible targets for direct imaging (and other measurements/studies) of dusty circumstellar discs; especially during the epoch of terrestrial planet formation (see e.g. Canup, 2004). Finally, constituent members of these groups are ideal candidates for the discovery and characterisation of young, substellar objects and, of course, extrasolar planets (see e.g. Lagrange et al., 2010).

1.1 The 32 Orionis Group

Unlike other nearby young moving groups and associations, the 32 Ori group has received little attention in the literature despite its discovery almost a decade ago, and as such its stellar population remains poorly characterised. Mamajek (2007) was the first to present evidence of a young stellar aggregate (age $\sim$25 Myr; designated Mamajek 3) associated with the B5IV+B7V binary 32 Ori based on a concentration of co-moving stars in a proper motion diagram. Fig. 1 shows the proper motions of stars within $10\degr$ of 32 Ori, for which the O-, B- and A-type stars (including Bellatrix and 32 Ori itself) are taken from the revised Hipparcos reduction of van Leeuwen (2007). Also plotted are pre-MS stars from the catalogue of Ducourant et al. (2005) and the 15 known stellar members of the 32 Ori group, compiled from a combination of Hipparcos, PPMXL (Röser, Demleitner & Schilbach, 2010) and the Fourth U.S. Naval Observatory CCD Astrograph Catalog (UCAC4; Zacharias et al., 2013).

While the majority of stars in Fig. 1 have small proper motions ($\lesssim$ 10 mas yr*-1*) and are associated with the more distant ($\sim$ 400 pc) Orion OB1 and $\lambda$ Orionis associations, there is a clear concentration of stars in proper motion space in the vicinity of 32 Ori. Within this concentration there are three A-type stars with Hipparcos entries (HR 1807 [HIP 25453], HD 36823 [HIP 26161] and HD 35714 [HIP 25483]) which are all approximately co-distant with 32 Ori, and which together provide a weighted mean group distance of $92.9\pm 2.3\,\rm{pc}$. Also appearing to be co-moving with 32 Ori are several X-ray bright late-type stars discovered by the ROSAT All-Sky Survey (Voges et al., 1999). Spectroscopic follow-up of a subset of these was performed by Alcalá et al. (1996, 2000) who measured strong lithium (Li) absorption in addition to deriving radial velocities which are consistent with that of 32 Ori ($18.6\pm 1.2\,\rm{km\,s^{-1}}$; Barbier-Brossat & Figon, 2000).

To date, a total of 15 stars have been identified as members of the 32 Ori group, in addition to a potential planetary-mass object (Burgasser et al., 2016; see Section 4.2). The stellar members have been compiled primarily by Mamajek and collaborators (Mamajek, 2007; Shvonski et al., 2016) – but see also additions by Mace et al. (2009) and Franciosini & Sacco (2011) – over the last decade, with new members being assigned on the basis of various diagnostics including common proper motion, radial velocities, H$\alpha$ emission and Li absorption. Note that the recent study of Bell, Mamajek & Naylor (2015) only included 14 stellar members of the group, having neglected HD 36823. This star has been shown to be a wide, common proper motion companion to HD 35714 (Shaya & Olling, 2011), and so we include it here in our list of known group members. Table 1 provides positions, spectral types, proper motions and radial velocities for the 15 known stellar members of the 32 Ori group, while Table 2 lists the mean group position, parallax, proper motion, radial velocity and Galactic $UVW$ velocity. The spectral types listed in Table 1 for all stars other than 32 Ori and HR 1807 have been determined through visual comparison of flux-calibrated spectra acquired over the past decade against a dense grid of MK spectral standards (see Shvonski et al., 2016). Fig. 2 depicts the $UVW$ velocity of the 32 Ori group relative to other young groups and associations within 100 pc. Velocities for the other young groups were taken from the recent compilation of Mamajek (2016), except for Argus which is from Gagné et al. (2014). Interestingly, not only does the 32 Ori group appear very close in velocity to the $\beta$ Pictoris moving group (BPMG) in all three $UVW$ planes, but the two groups also have very similar ages (see Bell et al., 2015 and the discussion in Section 5).

Two of the 15 stars listed in Table 1 (TYC 112-1486-1 and TYC 112-917-1) have previously been classified as members of the BPMG by Elliott et al. (2014). Using the Bayesian Analysis for Nearby Young AssociatioNs (BANYAN) II web tool (Gagné et al., 2014; Malo et al., 2013), we calculate that both stars have BPMG membership probabilities of $\lesssim$ 2 per cent, whereas the probabilities that they are associated with the young field are much higher ($\simeq$ 90 per cent). Furthermore, their kinematic distances (see Section 2.1 for details) are $\sim$ 80 pc for BPMG (adopting the $UVW$ velocity from Mamajek, 2016). Hence, if both stars are genuine BPMG members they would be two of the most distant members of the group, at approximately twice the $\sim$ 40 pc median distance of the ‘classic’ membership list of Zuckerman & Song (2004, see also , ). Based on this evidence we retain both TYC 112-1486-1 and TYC 112-917-1 as 32 Ori group members for this study and include them in our determination of the mean group properties listed in Table 2.

Shvonski et al. (2016) describes a Spitzer IRAC and MIPS survey of the 32 Ori group which was performed during 2007/08, covering all bandpasses from 3.6 to $70\,\rm{\mu m}$ (see also Shvonski et al., 2010). In this study the authors combined the Spitzer photometry with optical and near-infrared (IR) data to quantify excess emission arising from circumstellar material. Shvonski et al. report that 4/14 members exhibit excess $24\,\rm{\mu m}$ emission; HR 1807 ($f_{24}$ = 88.45 $\pm$ 0.36 mJy), HD 35499 ($f_{24}$ = 8.45 $\pm$ 0.10 mJy), HD 36338 ($f_{24}$ = 14.79 $\pm$ 0.15 mJy) and TYC 112-1486-1 ($f_{24}$ = 3.87 $\pm$ 0.09 mJy). HR 1807 also exhibits a $70\,\rm{\mu m}$ excess ($f_{70}$ = 91.0 $\pm$ 4.2 mJy). Note that the quoted fluxes represent total fluxes and correspond to excess emission more than 4$\sigma$ above typical photospheric levels. Modelling the excess emission, Shvonski et al. determined that the dust temperatures associated with these debris disc candidates were $\lesssim$ 200 K.

Recently, Bouy & Alves (2015) argued that the 32 Ori group should in fact be termed the ‘Bellatrix cluster’ on the basis that the sky position, distance ($77^{+4}_{-3}\,\rm{pc}$; van Leeuwen, 2007) and age ($20^{+2}_{-4}\,\rm{Myr}$; Janson et al., 2011) of the B2V star Bellatrix ($\gamma$ Ori) are similar to those of the 32 Ori group. There are three additional remarkable coincidences among its stellar observables:

Bellatrix’s radial velocity ($18.2\pm 0.8\,\rm{km\,s^{-1}}$; Barbier-Brossat & Figon, 2000) is consistent with the mean group radial velocity of the 32 Ori group ($18.6\pm 0.2\,\rm{km\,s^{-1}}$). 2. 2.

Bellatrix’s reddening [$E(B-V)=0.02$ mag, Friedemann, 1992; Zorec et al., 2009; Bhatt & Cami, 2015] is a good match for the mean of the 32 Ori group [($E(B-V)=0.03\pm 0.01$ mag; see Table 8]111Using the revised Johnson Q-method calibration of Pecaut & Mamajek (2013), and the $U-B$ and $B-V$ colours from Mermilliod (2006), we independently estimate the reddening of Bellatrix to be $E(B-V)=0.017$ mag.. 3. 3.

Bellatrix is the only star in the northern hemisphere within 100 pc of Earth of spectral type B2 or earlier.

Despite these coincidences, however, we find that Bellatrix’s deviant proper motion (see Fig. 1) translates to a 3D velocity of ($U,V,W)=(-14.7\pm 0.7,-7.1\pm 0.4,-9.8\pm 0.3)\,\rm{km\,s^{-1}}$ which is inconsistent ($>$10$\sigma$ in the $V$ component) with the mean velocity of the 32 Ori group shown in Table 2. In light of this kinematic inconsistency we do not believe there is sufficient evidence at present to include Bellatrix in the group, and do not recommend use of the name Bellatrix cluster at this time. Despite this inconsistency, however, we believe further astrometric, spectroscopic, and high-contrast imaging of Bellatrix is warranted to see whether it could harbour a dark companion which could be responsible for its deviant velocity222Bellatrix has a mass of $\sim$ $8.7\,\rm{M_{\odot}}$ (Hohle, Neuhäuser & Schutz, 2010), and its projected tangential velocity is peculiar compared to the 32 Ori group velocity by $\Delta\mu_{\alpha},\Delta\mu_{\delta}\simeq-16,+20$ mas yr*-1*($\sim$ 12 km s*-1*  if the star is actually at $d\simeq$ 93 pc), with negligible difference in radial velocity. A normal dwarf, white dwarf, or neutron star companion can not reconcile this velocity offset and the consistency of Bellatrix’s proper motion over the past decades. The star is too bright for Gaia DR1, but the revised Hipparcos catalogue (van Leeuwen, 2007) reports a 7-parameter solution for Bellatrix with a 2.5$\sigma$ significance acceleration in $\mu_{\alpha}$ (4.85 $\pm$ 1.94 mas yr*-2*) and negligible acceleration in $\mu_{\delta}$. Bellatrix’s velocity and properties can be reconciled with 32 Ori group membership if it is in a face-on orbit perturbed by a distant (a $\sim$ 102 au, $P$ $\sim$ centuries) black hole companion. We are currently investigating Bellatrix further to test this idea..

We hereby present the first large-scale spectroscopic survey for new low-mass members of the 32 Ori group to better understand its stellar population and properties. In Section 2 we discuss our candidate selection process. Section 3 details the medium-resolution optical spectroscopic observations in addition to describing how the spectroscopic properties were determined. In Section 4 we combine various indicators of stellar youth and group membership to identify new bona fide members, and place these in context by comparing them against the findings of previous surveys. Finally, Section 5 synthesises our stellar census results to discuss the global properties of the 32 Ori group, including its age, circumstellar disc frequency and spatial structure.

2 Candidate selection

Members of a moving group typically cover both a large range of distances and a large area on the sky; unlike compact clusters whose members essentially share the same proper motion, the proper motions and tangential velocities of individual group members can vary systematically and significantly. To identify potential kinematic members of a group one must therefore project its fixed Galactic $UVW$ velocity onto the sky over a range of distances and search for objects sharing a common motion. Furthermore, given the young age of the 32 Ori group, any genuine low-mass members will not only share a common space motion, but will also be over-luminous with respect to older main sequence stars of the same spectral type. In the absence of trigonometric parallaxes (soon to be rectified by Gaia), the combination of proper motions and photometry provides a robust method with which to identify potential group members, whilst also efficiently removing a substantial number of field interlopers which are naturally included in large area searches (see Kraus et al., 2014; Murphy & Lawson, 2015).

2.1 Input catalogues and search criteria

We adopt UCAC4 as our primary input catalogue, which provides positions, absolute proper motions and instrumental magnitudes (in a single, non-standard bandpass similar to $R$, hereafter termed $R_{\rm{UCAC}}$) complete to $\simeq$ 16 mag across the entire sky. In addition to these instrumental magnitudes, the catalogue also includes APASS (Henden et al., 2012) DR6 $BVgri$ and 2MASS Point Source Catalog (Cutri et al., 2003) $JHK_{\rm{s}}$ photometry. Note that as a consequence of the APASS limiting magnitude ($V\simeq 16\,\rm{mag}$), standard $BVgri$ photometry is only available for approximately 50 per cent of the sources in UCAC4.

Given the sky positions of the known group members (see Fig. 3), a 10$\degr$ search radius around 32 Ori itself represents a reasonable compromise between survey area and the telescope time required for spectroscopic follow-up. Within this region, UCAC4 returned a total of $\sim$ $7.1\times 10^{5}$ sources. To identify potential kinematic members from this sample we follow a formalism similar to that described in Murphy & Lawson (2015). Adopting the mean $UVW$ velocity for the 32 Ori group333Our selection of candidate members preceded Gaia DR1 and so the actual $UVW$ velocity we adopted was slightly different to that stated in Table 2, which we present as the current best estimate for the mean group velocity. The omission of the Gaia DR1 astrometry modifies the mean group parallax by $+0.27\,\rm{mas}$ and the mean group proper motion by $-0.4$ and $0.0\,\rm{mas\,yr^{-1}}$ in $\mu_{\alpha}\mathrm{cos}(\delta)$ and $\mu_{\delta}$, respectively, which results in a mean space motion of $(U,V,W)=(11.9\pm 0.3,-18.6\pm 0.4,-9.0\pm 0.3)\,\rm{km\,s^{-1}}$. We note, however, that this subtle difference in the adopted $UVW$ velocity does not have a significant impact on the calculated $\Delta_{\rm{PM}}$ and kinematic distances, on average affecting these at the $<1\,\rm{mas\,yr^{-1}}$ and $<$ 2 pc level, respectively., we project this over a range of distances ($50\leq d\leq 150\,\rm{pc}$, in 1 pc increments) for each object to calculate the expected proper motion and radial velocity, retaining only those objects whose proper motions and kinematic distances satisfied both of the following criteria:

The lowest total difference between the expected and observed proper motion,

[TABLE]

must be $\leq$10 mas yr*-1*. 2. 2.

The ‘best’ kinematic distance corresponding to this proper motion must be $70\leq d_{\rm kin}\leq 110\,\rm{pc}$.

These thresholds returned 5349 potential kinematic members of the 32 Ori group, each of which had an associated $\Delta_{\rm PM}$, $d_{\rm kin}$ and expected radial velocity. The $d_{\rm kin}$ and $\Delta_{\mathrm{PM}}$ limits are somewhat arbitrary, but are motivated by the bona fide members as found in UCAC4 (14/15 with $\Delta_{\rm PM}<10$ mas yr*-1* and 13/15 with $70<d_{\rm kin}<110$ pc). During the spectroscopic follow-up we also observed stars whose distance and $\Delta_{\mathrm{PM}}$ values fell outside these limits but otherwise resembled strong candidates.

2.2 Colour-magnitude selection

The kinematic distances determined in Section 2.1 must also be consistent with the colour-magnitude diagram (CMD) positions of young, low-mass stars above the main-sequence. Our primary photometric selection was performed using the UCAC4/APASS DR6 $M_{V},V-K_{\rm{s}}$ CMD. Fig. 4 shows the 3730 (out of 5349) potential kinematic members with APASS $V$- and 2MASS $K_{\rm{s}}$-band photometry, in addition to known members of the 32 Ori group which satisfy the same kinematic selection criteria. Also plotted is the empirical fit to the $\sim$ 130 Myr-old Pleiades single-star sequence by Stauffer et al. (2007), corrected for the effects of both distance and extinction ($d=136$ pc and $A_{V}=0.12$ mag from Melis et al., 2014 and Stauffer et al., 1998, respectively). As expected, the vast majority of sources lie below the Pleiades sequence, indicating that they are field stars with coincidental proper motions and not photometric members of the 32 Ori group. However, there are a significant number of candidates above the sequence whose CMD positions are consistent with known members, i.e. their kinematic distances agree with photometric distances for a putative $\sim$ 25 Myr population. These stars are potential new 32 Ori group members and we selected $\sim$100 of them for spectroscopic follow-up, with a particular emphasis on late-type stars ($V-K_{\rm{s}}\geq 3$ mag; spectral type $\gtrsim$ K4).

Of the 15 known stellar members listed in Table 1, three do not have counterparts in Fig. 4. While 32 Ori itself satisfies both criteria in Section 2.1, it is saturated in APASS $V$. In contrast, the best-fit distances for both V1874 Ori and 2MASS J05194398+0535021 fall outside of the allowed range (131 and 116 pc, respectively). The latter member also has $\Delta_{\rm PM}=12.5$ mas yr*-1*, but with large ($>$10 mas yr*-1*) errors on its UCAC4 proper motion444If we instead adopt proper motions from URAT1 for V1874 Ori and PPMXL for 2MASS J05194398+0535021 (as found in Table 1 and which has a smaller uncertainty than UCAC4), we calculate distances of 88 and 110 pc, respectively. Both stars would then satisfy the selection criteria..

Given the incompleteness of APASS $BVgri$ photometry in UCAC4, we also searched for additional low-mass members in the $M_{R_{\rm{UCAC}}},R_{\rm{UCAC}}-K_{\rm{s}}$ CMD using the stars selected in Fig. 4 as a reference, and identified $\sim$30 additional candidates with consistent CMD positions not present in the $M_{V},V-K_{\rm{s}}$ diagram. Furthermore, we also looked for potential members in the First U.S. Naval Observatory Robotic Astrometric Telescope Catalog (URAT1; Zacharias et al., 2015), which extends approximately 2 mag fainter than UCAC4 in an almost identical non-standard pseudo-$R$ bandpass. We used the same method as above, but allowed a $\Delta_{\mathrm{PM}}<15$ mas yr*-1* limit due to the two-epoch astrometry having larger proper motion uncertainties. From a similar kinematic and CMD analysis we identified another $\sim$20 candidates only present in URAT1 down to $R\approx 16\,\rm{mag}$, giving a total of $\sim$150 potential members from the combined searches.

The fainter $R_{\rm{UCAC}}$ and URAT1 samples are crucial for better estimating the age of the group. The limiting magnitude of APASS is $V\simeq 16~{}\rm{mag}$, which at a distance of 90 pc, corresponds to the expected location of the lithium depletion boundary (LDB) in a $\sim$25 Myr population. Only by confirming fainter Li-rich members can we can identify the precise position of the 32 Ori group LDB and calculate a semi-fundamental (Soderblom et al., 2014) age for the group (see Section 5.3).

3 Spectroscopic observations and data analysis

To unambiguously differentiate between genuine 32 Ori group members and field interlopers or other contaminants, additional spectroscopic diagnostics are required. These typically include measuring radial velocities to ascertain whether candidates share similar three-dimensional space motions, as well as identifying spectroscopic features associated with stellar youth (e.g. Li i $6708\,\rm{\AA}$ absorption and H$\alpha$ emission).

We observed 124 candidate members (plus 11 of the known members listed in Bell et al., 2015) in four runs between 2015 September and 2016 February using the Wide Field Spectrograph (WiFeS; Dopita et al., 2007) on the ANU 2.3-m telescope at Siding Spring Observatory (SSO). Poorer than expected conditions meant we were unable to observe all of the faintest ($V>15\,\rm{mag}$) candidates and so we prioritised objects with small $\Delta_{\rm PM}$ values, CMD or sky positions similar to known members and candidates coincident with ROSAT X-ray sources (Voges et al., 1999). We later took observations in 2016 October and 2017 January to revisit several possible spectroscopic binaries identified in earlier runs.

WiFeS is an image-slicing integral field spectrograph with a nominal 25$\times$38 arcsec field-of-view and 0.5 arcsec sampling along twenty-five 38$\times$1 arcsec slitlets. Observations were made in half-field (12$\times$38 arcsec) ‘stellar’ mode with 2$\times$ spatial binning (1 arcsec spaxels; well-matched to typical 1.5–2.5 arcsec SSO seeing). The $R$7000 grating and $RT$480 dichroic gave a resolution of $\lambda/\Delta\lambda\approx 7000$ and wavelength coverage from 5300–7000 Å. The field-of-view was aligned to the parallactic angle prior to each exposure, with exposure times up to 3$\times$1200 sec. In addition to science targets, we also obtained spectra for 4–10 FGKM radial velocity standards each night from the list of Nidever et al. (2002) and white dwarf flux calibrators following Bessell (1999).

After basic image processing, we used iraf, figaro and python routines to rectify, extract, wavelength-calibrate and combine the 3–6 image slices (depending on seeing) that contained the majority of the stellar flux. The slices were treated like long-slit spectra and individually extracted and wavelength calibrated against NeAr arc frames taken following each exposure. The rms of the final wavelength solution in all cases was better than 0.02 Å. Typical signal-to-noise ratios were 50–100 around H$\alpha$ and Li i, decreasing to $\sim$ 20 for the faintest candidates. Table 3 (the full version of which is available as Supporting Information with the online version of the paper) details our spectroscopic measurements for the observed candidates and known members listed in Bell et al. (2015). Typical WiFeS/$R7000$ spectra of several new M dwarf 32 Ori group members confirmed in this work are shown in Fig. 5.

3.1 Spectral types

Spectral types for each candidate were first estimated by visual comparison against the Pickles (1998) stellar spectral flux library and the Sloan Digital Sky Survey (SDSS) average M dwarf templates of Bochanski et al. (2007), with the WiFeS spectra Gaussian-smoothed to the approximate resolution of the templates prior to comparison. This analysis showed a substantial fraction of the candidates were clearly reddened, motivating us to fit their spectra to the templates with interstellar reddening as a free parameter. For this we adopted the Cardelli, Clayton & Mathis (1989) reddening law and a total-to-selective extinction ratio of $R_{V}=3.1$ and calculated the best-fit reddening [in the range $0<E(B-V)<3$ mag, in 0.05 mag increments], after resampling the template onto the same wavelength array as the star. The unreduced $\chi^{2}$ statistic was used as the goodness of fit.

The resulting spectral types for all candidates are listed in Table 3, with the estimated reddenings provided in the electronic version of the table (see also Table 8 for the early type members). Given the modest wavelength range of the WiFeS data, spectral type coverage of the Pickles library and degeneracies with reddening, we estimate these types are accurate to $\pm$1–2 subtypes for A-, F-, G- and K-type candidates. The homogeneity and high quality of the SDSS M dwarf templates permit spectral types for these candidates to approximately half a subtype.

3.2 Radial velocities

Radial velocities for each candidate were measured by cross-correlation over 5500–6500 Å against standards of similar spectral type observed that run using a python implementation of the fxcor algorithm (Tonry & Davis, 1979). Following the procedure described in Murphy & Lawson (2015), each spectrum was first resampled onto a log-linear wavelength scale and normalised by subtracting a boxcar-smoothed copy and dividing by the standard deviation. The waveforms were then cross-correlated and a Gaussian fitted to the cross-correlation function (CCF) peak, before transforming to a heliocentric frame. The velocities reported in Table 3 are the mean and standard deviation against standards observed that run. Repeat and standard star observations demonstrate that the instrument and this technique provide an external velocity precision of $\lesssim$ 1 km s*-1* for bright stars. For the faintest candidates in Table 3 this falls to 2–3 km s*-1* per exposure.

3.2.1 Spectroscopic binaries and fast rotators

Six candidates (HD 37825, BD+08 900B, 2MASS J05442447+0502114, 2MASS J05363692+1300369, 2MASS J05320596$-$0301159, 2MASS J05561307+0803034) and the previously known 32 Ori group members HD 35499 and HD 35695 showed average CCF widths significantly larger than other stars ($\textrm{FWHM}\gtrsim 2.5$ px; see Murphy & Lawson, 2015), with BD+08 900A and the known member 2MASS J05234246+0651581 borderline cases. This broadening can be attributed to either fast rotation or unresolved spectroscopic binarity at the modest ($c\Delta\lambda/\lambda\approx 45$ km s*-1*) velocity resolution of WiFeS (see discussion in Murphy & Lawson, 2015). Three of these stars – HD 37825, 2M0536+1300 and 2M0556+0803 – showed double line cores and broad, asymmetric (though unresolved) CCFs indicative of double-lined spectroscopic binary systems (SB2). Notably, the 2016 October and 2017 January CCFs of HD 37825 were clearly resolved into two near-equal amplitude peaks (the latter spectrum showing the system to be an SB3 with a weak tertiary component; see Fig. 6), whilst the 2016 February 21 spectrum had much narrower, single profiles, ostensibly close to the systemic velocity ($\textrm{RV}\approx 28$ km s*-1*).

In the absence of resolved double lines, the origin of the broadening seen in the other stars is less clear. Given their smoother spectral features and symmetric CCFs, it is likely HD 35499, HD 35695, 2M0532$-$0301 and 2M0544+0502 are fast rotators ($v\sin i\gtrsim 45$ km s*-1*), while BD+08 900B may be a binary (with BD+08 900A more likely a fast rotator). Higher resolution observations are necessary to confirm these predictions.

Of the 21 candidates or previously known members observed more than once with WiFeS (see Tables 4 and 5), eight non-SB2s have maximum radial velocity differences larger than 5 km s*-1* and therefore may be single-lined (SB1) spectroscopic binaries. One of these stars (2MASS J05525572$-$0044266) is a known $\sim$0.86 d period eclipsing binary (Drake et al., 2014, see Section 4.4.1), while three more have broad CCFs (and thus less precise velocities), yielding differences marginally greater than 5 km s*-1* (HD 35499, 2M0532$-$0301 and 2M0544+0502). The Li-rich candidate 2MASS J05350092+1125423 (RV${}_{\textrm{expt}}=17.4$ km s*-1*) was observed twice in 2016 February (19.6 and 10.9 km s*-1*, $\Delta t=3$ d), with further observations in 2016 October (21.0 km s*-1*) and 2017 January (19.5 km s*-1*). Given the agreement of the three other epochs, with no other indications of binarity it is likely that the second February velocity was erroneous. The known K6.5 member 2M0523+0651 (RV${}_{\textrm{expt}}=18.5$ km s*-1*) has two WiFeS velocities separated by 442 days which agree and a third which is discrepant by $\sim$9 km s*-1*, well outside the expected errors. Alcalá et al. (2000) reported a $18.4\pm 1.0$ km s*-1* velocity from a high resolution ($R\approx 30\,000$) spectrum which agrees with the expected velocity. Interestingly, that study noted 2M0523+0651 may be a spectroscopic binary based on an asymmetric CCF and we too find a somewhat broadened CCF in our observations. Finally, the Li-rich, early M-type candidates 2MASS J05351761+1354180 and 2MASS J05330574+1400365 both have velocity differences of $\sim$7 km s*-1* over three epochs. They do not show broadened CCFs and their mean velocities are close to those expected of bona fide members.

3.3 H$\alpha$ and Li equivalent widths

Young stars are predominantly active and one of the most common diagnostics of activity is the Balmer H$\alpha$ line at 6563 Å. For late-type stars (spectral type $\gtrsim$ M0) H$\alpha$ is typically seen in emission to ages of several Gyr (e.g. Gizis, Reid & Hawley, 2002), whereas in early and solar-type stars it can be seen in various levels of emission or absorption depending on activity. We measured the H$\alpha$ equivalent width (EW) of our candidates by fitting a single Gaussian or Voigt profile with linear continuum to the emission/absorption feature. For stars where the line could not be well fitted analytically, EW[H$\alpha$] was calculated by direct integration of the line profile. Due to uncertainties in the placement of the continuum and integration limits, we estimate uncertainties of up to 1 Å in the measured EW[H$\alpha$] values. Several of the M-type stars in our sample also showed strong He I emission at 5876 and $6678\,\rm{\AA}$. This is chromospheric in nature and is observed in active M-dwarfs with ages of up to several Gyr (Gizis et al., 2002).

Another key indicator of stellar youth is the resonant Li i $6708\,\rm{\AA}$ absorption feature. As a coeval group of young stars contracts toward the zero-age main-sequence their core temperatures increase until at $\sim$ $3\times 10^{6}$ K Li fuses. Such temperatures can be reached in either fully convective mid- to late-M dwarfs or at the base of the convective zone in late-K/early-M dwarfs. For stars between these luminosities, rapid Li depletion ensues and the Li i feature is no longer visible. At an age of $\sim$ 25 Myr we expect Li to be fully depleted in stars with spectral types between $\sim$ M1.5 and M4.5, however it should still be present in both earlier and later spectral types (see e.g. Mentuch et al., 2008; Jeffries et al., 2013). As in the case of H$\alpha$, we determined EW[Li] by Gaussian profile fitting if the line was present and estimated an upper limit from the local pseudo-continuum if it was not. No attempt was made to deblend Li i from the nearby Fe i feature at $6707.4\,\rm{\AA}$, which is unresolved at our resolution and is typically much weaker (EW[Fe] $\lesssim 30\,\rm{m\AA}$; Soderblom et al., 1993) compared to the expected EW[Li] of 32 Ori group members. WiFeS EW[H$\alpha$] and EW[Li] values for all observed candidate and known members of the 32 Ori group are listed in Table 3.

4 The low-mass population of the 32 Ori group

4.1 Newly identified members

Whilst any one of the diagnostics discussed in Section 3 could be used to assign membership of a candidate to the 32 Ori group, this increases the risk of erroneously including a member, which when assessed using an alternative diagnostic, is an obvious non-member. The most discriminating diagnostic to differentiate between young and older stars is the presence of Li, however this is only valid over a small range of spectral types (see e.g. Jeffries, 2006). In addition, stellar activity as traced by H$\alpha$ emission is not solely restricted to the youngest stars, with older K- and M-type field stars also exhibiting elevated levels of activity (see e.g. West et al., 2011) and so whilst the presence of H$\alpha$ is necessary, it alone does not provide a reliable membership diagnostic. Finally, although one would expect stars with the same space motion as the 32 Ori group to be likely members, within a given sample there will always be a small fraction of older field stars which are co-moving purely by coincidence. In contrast, radial velocities of genuine short-period binary members based on single-epoch spectra will, in the majority of cases, suggest that these are not co-moving.

To minimise the number of potential interlopers in our final membership, we combine all three spectroscopic diagnostics (Li, H$\alpha$, RV) in conjunction with the proper motion match ($\Delta_{\rm PM}$) and kinematic distance determined in Section 2.1. We adopt a threshold of $\mathrm{EW[Li]}\geq 100\,\rm{m\AA}$ across the entire spectral type range of our sample as an indicator of youth, whereas those stars with measured $\mathrm{EW[Li]}<100\,\rm{m\AA}$ for spectral types $\leq$ K5 and $\geq$ M5 are deemed to be older field interlopers. Although there is a range of observed H$\alpha$ emission levels in both young coeval populations and older field stars, young stars do exhibit a lower envelope of H$\alpha$ emission (see e.g. Kraus et al., 2014 in the slightly older Tucana-Horologium association; hereafter Tuc-Hor) and so we classify any star with emission below this envelope as an older field star. To illustrate these criteria, the EW[Li] and EW[H$\alpha$] values of observed candidates and members of the 32 Ori group are plotted as a function of $V-K_{\rm{s}}$ colour in Figs. 7 and 8.

Fig. 7 shows the Li depletion pattern of the 32 Ori group compared to other nearby young moving groups. Whilst the overall trend for all such groups is very similar for G-type stars and earlier, as one moves into the K- and M-type regime the older groups clearly show evidence of significant Li depletion at earlier spectral types compared to younger groups. The Li depletion pattern of the 32 Ori group closely mirrors that of the BPMG and is bracketed by those of the TW Hydrae association (TWA) and Tuc-Hor (ages $\sim$10 and 45 Myr, respectively). In TWA there are a significant number of Li-rich early-M dwarfs which are clearly absent in the 32 Ori group, for which the Li depletion pattern appears to turn over at a spectral type of $\sim$M0. Likewise, there are several Li-rich mid to late K-type stars in the 32 Ori group which are not observed in the older Tuc-Hor (turn over at $\sim$K4). This already provides a relative age ranking suggesting that the 32 Ori group is older than TWA, younger than Tuc-Hor and approximately coeval with the BPMG (age $\sim$25 Myr).

Given that young moving groups typically have very small intrinsic velocity dispersions ($\sigma_{\rm 1D}\lesssim 1.5$ km s*-1*; see Mamajek, 2016), we would expect the observed radial velocity of a genuine 32 Ori group member to be similar to the line-of-sight projection of the group space velocity at that position (see Section 2.1). We therefore retained only those candidates for which $|\Delta_{\mathrm{RV}}|=|\mathrm{RV}-\mathrm{RV_{expt}}|\leq 5$ km s*-1*, after allowing for the uncertainty on the WiFeS velocity (see Binks & Jeffries, 2016b). When more than one velocity measurement was made we adopted the unweighted average and standard deviation (also see discussion in Section 3.2.1). Fig. 9 shows the difference in radial velocity $\Delta_{\mathrm{RV}}$ for all observed candidates and known members of the 32 Ori group as a function of $V-K_{\rm{s}}$ colour. Only three stars have WiFeS velocities which place them just outside the $\Delta_{\rm{RV}}$ limit (all have $|\Delta_{\rm{RV}}|<6$ km s*-1*; see Table 3). After considering their velocity uncertainties, however, they can all be considered genuine members. One is the known member and SB2 system V1874 Ori ($\Delta_{\rm RV}=5.3$ km s*-1*) which has a high-resolution systemic velocity of $18.4\pm 0.3$ km s*-1* (see Table 1), giving $\Delta_{\rm RV}=1.3$ km s*-1*. The remaining two stars are both new members, one of which (2M0522+0502) is a possible fast rotator with a broad CCF.

Combining the Li absorption, H$\alpha$ emission and radial velocity criteria we identify 28 new members of the 32 Ori group. These new members, all with spectral types between M1.5 and M5.5, are listed in Table 3 along with the 90 non-members and six interesting systems requiring further study (see Section 4.4). The penultimate set of columns in Table 3 list the results of the membership tests described above, with the final membership decision including the proper motion match $\Delta_{\textrm{PM}}$ and (kinematic) distance. For convenience, each member is identified by both its 2MASS designation and a shorter 32 Ori group membership number prefixed by the letters THOR, similar to the TWA (TW Hydrae association) and RECX ($\eta$ Chamaeleontis cluster) nomenclatures commonly used in the literature for other young groups.

4.2 Recently proposed members from the literature

Burgasser et al. (2016) recently proposed the free-floating low surface gravity L1 brown dwarf WISE J052857.68+090104.4 as the first substellar member of the 32 Ori group. It is only $3\degr$ from 32 Ori itself and its estimated distance, proper motion, radial velocity and spectral characteristics are all consistent with group membership. At an age of $\sim$25 Myr, its 1880 K effective temperature implies a mass ($M=14^{+4}_{-3}\,\rm{M_{Jup}}$ ) which straddles the brown dwarf/planetary-mass boundary.

Riedel et al. (2016) also proposed the Li-poor M2.5 star SCR 0522$-$0606 (2MASS J05224069$-$0606238) as a potential 32 Ori group member on the basis of its proper motion, spatial position and low surface gravity. Their SALT/RSS radial velocity of $-1.5\pm 5.0$ km s*-1*, however, is approximately 4$\sigma$ from the $\sim$21 km s*-1* expected of a genuine member at that position. To test for spectroscopic binarity we obtained WiFeS/$R7000$ spectra on 2016 October 14 and 2017 January 7, finding a mean radial velocity of $25.6\pm 1.9$ km s*-1* and a spectral type of M3. The spectra showed broad, unresolved CCFs ($\textrm{FWHM}=3.6$ px) and double-peaked He i $5876\,\rm{\AA}$ emission lines, both suggestive of binarity. The star is outside our $10\degr$ survey radius, but based on its UCAC4 proper motion, we calculate $\Delta_{\rm PM}=7.9$ mas yr*-1* at $d_{\rm kin}=88$ pc. At this distance SCR 0522$-$0606 has $M_{V}=9.55\,\rm{mag}$ and an elevated CMD position consistent with an equal-mass binary. Given its low gravity, distance, radial velocity and reasonable proper motion, we confirm membership of SCR 0522$-$0606 in the 32 Ori group.

Including these two objects (as well as HD 36002, see Section 4.3.1), there are currently 46 known 32 Ori group members. Note that given the limited areal coverage and depth of the current study, the true stellar census of the group is almost certainly incomplete (especially at lower masses) and this number should be treated as a lower limit.

4.3 Notable systems

4.3.1 2MASS J05284050+0113333 (THOR 4B)

2M0528+0113 (M3.5, $d_{\rm kin}=99$ pc), is only 24 arcsec from and co-moving with the early-type star HD 36002 (see Fig. 10), which was in our final candidate list ($d_{\rm kin}=105$ pc, $\Delta_{\textrm{PM}}=0.2$ mas yr*-1*) but not initially observed. It was, however, observed by both Hipparcos ($d=103_{-8}^{+9}$ pc; van Leeuwen, 2007) and Gaia ($112_{-7}^{+8}$ pc; including the additional $\pm 0.3$ mas systematic uncertainty on the parallax as described in Gaia Collaboration et al., 2016). Its Gaia DR1 proper motion is $<$ 2 mas yr*-1* from that expected of a 32 Ori group member at the weighted mean distance of 109${}_{-5}^{+6}$ pc and 2M0528+0113 is $<$ 3 mas yr*-1* discrepant, well within its 5.8 mas yr*-1* URAT1 uncertainties. We obtained two WiFeS R7000 spectra of HD 36002 on 2017 January 11 and find it to be an SB2 system (see Table 5) with an estimated spectral type of A7. Given their close separation, congruent distances and proper motions, we propose HD 36002 and 2M0528+0113 are a genuine co-moving pair separated by $\sim$2600 au and thus HD 36002 is also a member of the 32 Ori group (THOR 4A; see Table 3).

4.3.2 2MASS J05372061+1335310 (THOR 8B)

This M3 star is 50 arcsec north-east of the previously known G5 member HD 245567 (THOR 8A; see Fig. 10). At a kinematic distance of 113 pc, 2M0537+1335 has a space motion only 1.5 km s*-1* from the mean 32 Ori group velocity. HD 245567 has a kinematic distance of 104 pc, which agrees with that of 2M0537+1335 (physical separation of $\sim$5500 au) to within the uncertainties propagated from their respective proper motions. Neither star has a Hipparcos or Gaia DR1 parallax. A wider possible co-moving companion, the M5.5 2MASS J05373000+1329344, is 6 arcmin from HD 245567 at a distance of 110 pc, corresponding to a separation of $\sim$40 kau.

4.3.3 2MASS J05351761+1354180 (THOR 14A)

This new Li-rich M1.5 member is only 27 arcsec from 2MASS J05351625+1353594 (THOR 14B, M3.5; see Fig. 10). The ROSAT source 1RXS J053516.6+135404 is more likely associated with the lower mass component. Despite observed radial velocities 3–5 km s*-1* larger than expected (noting that 2M0535+1354 is a suspected SB1; see Table 4), both stars are excellent proper motion matches to the group ($\Delta_{\rm PM}<2$ mas yr*-1*) at distances of 116 and 110 pc, respectively, and are unequivocal members of 32 Ori with a separation of $\sim$3000 au.

4.3.4 2MASS J05274313+1446121 (=DIL7; THOR 17A)

2M0527+1446 (spectral type M3) is associated with the ROSAT source 1RXS J052743.4+144609 and was catalogued as an H$\alpha$ emission line object (DIL7) in the north-western outskirts of the $\lambda$ Orionis star-forming region by Duerr, Imhoff & Lada (1982). It forms a 19 arcsec wide pair with the M5 member 2MASS J05274404+1445584 (THOR 17B; see Fig. 10). Both stars are excellent kinematic matches to the mean 32 Ori group space motion ($\Delta_{UVW}<2$ km s*-1*) at inferred distances of 87 and 82 pc, respectively. At such distances their angular separation corresponds to $\sim$1600 au. The pair are unlikely to be $\lambda$ Orionis members, given both the much larger distance to the association and its significantly younger age ($\sim$400 pc and $\sim$10 Myr, respectively; see Bell et al., 2013) and thus we assign them as members of the 32 Ori group. Interestingly, both this pair and THOR 14AB have components either side of the 32 Ori group lithium depletion zone but in the opposite sense; THOR 17A (M3) and THOR 14B (M3.5) lie in the Li-poor region, while THOR 17B (M5; $\textrm{EW[Li]}=580$ mÅ) and THOR 14A (M1.5; 180 mÅ) are Li-rich. Component identifications and Washington Double Star (WDS) catalogue parameters for these four new wide systems are listed in Table 6.

4.3.5 2MASS J05363692+1300369 (THOR 31)

The 2016 January 25 spectrum of this Li-rich SB2 member (combined spectral type M4.5) showed a strong ($\textrm{EW}\approx-5$ Å) emission line $\sim$ 400 km s*-1* redward of H$\alpha$, near the rest wavelength of 6573 $\rm{\AA}$ Ca i (see Fig. 11). Inspection of the raw image shows that the emission is associated with a single source and not the sky background or extended nebular emission. The feature is clearly variable as only Ca i absorption is seen in two 2016 October observations a night apart, but a weaker emission ‘bump’ at +200 km s*-1* is visible in the 2017 January 7 spectrum. At all times the shape and strength of H$\alpha$ remained approximately constant. The emission features cannot be red-shifted H$\alpha$ from the companion, whose velocity offset is only $\sim$ 50 km s*-1* based on the broad but unresolved CCF and double-peaked H$\alpha$ line.

Given the lack of excess H$\alpha$, other Ca emission lines (e.g. Ca i 6717 Å) or any emission lines in the January spectrum typically associated with flare activity (especially He i 5876 and 6678 Å), we do not believe this line is associated with flare-driven Ca i 6573 Å. Instead, we propose it is red-shifted H$\alpha$ emission arising from an eruptive prominence or coronal mass ejection (CME) moving away from the system at close to its escape velocity (c.f. $\sim$ 350 km s*-1* for $M=0.1\,\rm{M_{\odot}}$ and $R=0.3\,\rm{R_{\odot}}$, appropriate for a 25 Myr star of $T_{\rm eff}=3000$ K; Baraffe et al., 2015). Balmer line emission from CMEs has been detected in a handful of active M dwarfs (e.g. Houdebine, Foing & Rodono, 1990; Guenther & Emerson, 1997; Fuhrmeister & Schmitt, 2004) but is typically associated with contemporaneous flare activity and a blue-shifted asymmetry in the line profile. That we observed red-shifted emission with a projected velocity of 400 km s*-1* outside of an obvious flare event suggests the ejected material was long-lived and large enough to not be completely occulted by the stellar disc. The role of the binary companion in this scenario also remains unclear. High-cadence monitoring of 2M0536+1300 would be useful to firmly establish the nature of the January 25 emission and the frequency of such events.

4.4 Potential members requiring further study

Below we present notes on six systems which we have assigned a membership of ‘?’ in Table 3 (provided in full as Supporting Information with the online version of the paper). Improved spectroscopic, astrometric or photometric data is required to unequivocally assign membership of these stars to the 32 Ori group.

4.4.1 2MASS J05525572$-$0044266

2M0552$-$0044 was classified as a detached, Algol-type eclipsing binary with a period of 0.86 d by the Catalina Sky Survey (CSS; Drake et al., 2014), and as such makes it a rare example of an eclipsing M dwarf (unresolved spectral type M3). Its phased $V_{\rm CSS}$ light curve is plotted in the top panel of Fig. 12. We confirm binarity from two radial velocities of +65.5 and $-$39.3 km s*-1* ($\Delta t=357$ d, see Table 5) and a variable, double-peaked H$\alpha$ emission line. We obtained five more WiFeS observations of the star during 2017 January 7–9 for the purposes of establishing a preliminary orbital solution. Although the secondary component was not visible in the CCF, we estimated its velocity at each epoch by measuring the velocity offset between H$\alpha$ peaks and adding this to the primary velocity derived from cross-correlation. One observation did not show a double H$\alpha$ line, implying its radial velocity (18.4 km s*-1*) is close to systemic. Using these velocities we fit Keplerian orbits of period 0.8589884 d (Drake et al., 2014) using standard least squares methods and derive the solution listed in Table 7 and the middle panel of Fig. 12. The system has a total mass of $(M_{1}+M_{2})\sin^{3}i=0.6$ M*⊙* and separation $a\sin i=3.2~{}R_{\odot}$. Assuming $i=90^{\circ}$, this corresponds to component masses of approximately 0.44 and 0.16 M*⊙. Given the agreement between the fitted systemic velocity of $20.9\pm 2.3$ km s-1* and the 20.6 km s*-1* expected of a 32 Ori group member at that position, we consider 2M0552$-$0044 a highly likely member pending further velocity measurements and improved photometry.

The CSS light curve contains 192 measurements over 7.3 years and is not well sampled around the eclipses. Rotationally modulated photometric variation due to star-spots appearing and disappearing during that time also likely contributes to scatter in the regions outside the eclipses. This may affect the period determination and makes deriving temperature ratios and radii problematic. Moreover, the APASS $V$-band magnitude ($15.15\pm 0.06\,\rm{mag}$) differs significantly from the CSS photometry. The accuracy of these data is limited by the transformation to $V_{\rm CSS}$ from the unfiltered survey photometry, and is only appropriate for the G-dwarf calibrators used by the survey. From 445 Landolt standard stars, Drake et al. (2013) derived $V=V_{\rm CSS}+0.31(B-V)^{2}+0.04$, which for $(B-V)_{\rm APASS}=1.51$ yields $V=14.2$ mag. After re-deriving this relation we find it is well-defined for standards as red as 2M0552$-$0044, so the remaining discrepancy is likely due to binarity affecting the transformation. We therefore adopt the APASS photometry, which provides a CMD position appropriate for a 25 Myr-old M3 star at $\sim$90 pc.

We are obtaining high-cadence photometry and further radial velocity observations of the system and will present a full reanalysis and characterisation, including final 32 Ori group membership, in an upcoming work. Until then we note that 2M0552$-$0044 is Li-poor, satisfies both selection criteria ($\Delta_{\rm{PM}}=2.5$ mas yr*-1*, $d_{\rm{kin}}=92$ pc), and is likely associated with the X-ray source 1RXS J055257.7$-$004424.

4.4.2 BD+08 900AB (=HD 34081AB)

A 4 arcsec A7+F2 near-equal brightness binary, BD+08 900AB is resolved in UCAC4 and we obtained spectra for both components during good seeing (1.5 arcsec FWHM) WiFeS observations in 2015 September and 2017 January. BD+08 900B may be an SB2 while the primary is a possible fast rotator (see Section 3.2.1). Both stars have radial velocities which are consistent with 32 Ori group membership at $\lesssim$ 2 km s*-1* and BD+08 900B is an excellent proper motion match ($\Delta_{\rm PM}=1.5$ mas yr*-1*) at a kinematic distance of 82 pc. BD+08 900A’s UCAC4 proper motion is significantly smaller and a poor match at any distance. The combined system was observed with Hipparcos ($d=120_{-16}^{+22}$ pc) and Gaia ($97^{+6}_{-5}$ pc). Adopting the astrometry of the latter, BD+08 900AB is only 3 km s*-1* from the 32 Ori group mean space motion and the $\textrm{EW[Li]}=130$ mÅ we measure for BD+08 900B is similar to young F-type stars in Fig. 7. Until the binary nature of BD+08 900B is confirmed we refrain from assigning the system to the 32 Ori group, but note it is a strong kinematic candidate.

4.4.3 HD 37825

HD 37825 is a new SB3 system (see Fig. 6) which we observed at four epochs (see Table 4), one of which (2016 February 21) exhibited narrow, single lines and a radial velocity of 28 km s*-1*, which must be near-systemic. This is in reasonable agreement with the $\sim$20 km s*-1* expected of a 32 Ori group member at that position and there is also a moderately good proper motion match to the group ($\Delta_{\rm PM}=6.8$ mas yr*-1*) at a kinematic distance of 84 pc. The 2016 October and 2017 January spectra showed clearly resolved double lines with a weaker tertiary component. However, only two components are visible in the CCF, separated by $\sim$ 140 km s*-1*. The system does not have a Hipparcos or Gaia DR1 parallax or proper motion. The $\textrm{EW[Li]}=130$ mÅ we measured in the single-lined spectrum is typical of a young F5 star (see Fig. 7). Higher resolution spectroscopy of the system and a velocity curve would be useful in confirming its membership.

4.4.4 2MASS J05053333+0044034

2M0505+0044 satisfies both UCAC4 selection criteria ($\Delta_{\rm{PM}}=5.0$ mas yr*-1*, $d_{\rm{kin}}=90$ pc) and also shows depleted lithium, consistent with its M2.5 spectral type. The star’s radial velocity, however, is rather discrepant ($\Delta_{\rm{RV}}=13.4$ km s*-1*). Coupled with a broad H$\alpha$ line ($\Delta v\approx 300$ km s*-1* at 10 per cent of peak flux; \textrm{EW[H\alpha]}=-12.3 Å) and strong He i 5876 and $6678\,\rm{\AA}$, Na d, and X-ray emission (1RXS J050533.4+004421), this could be indicative of spectroscopic binarity. The star’s CCF is not broadened, suggesting it is not a fast rotator or SB2. Choosing to ignore the $|\Delta_{\rm{RV}}|<5$ km s*-1* velocity threshold would have implications for other candidates and so we retain 2M0505+0044 as a possible member requiring further velocity measurements.

4.4.5 2MASS J05561307+0803034

2M0556+0803 easily satisfies both selection criteria ($\Delta_{\rm{PM}}=0.5\,\rm{mas\,yr^{-1}}$ and $d_{\rm{kin}}=97\,\rm{pc}$), whereas its mean radial velocity is somewhat discrepant ($\Delta_{\rm{RV}}=7.6\,\rm{km\,s^{-1}}$). On its own, this would suggest non-member status, however, the star’s CCF is rather broad ($\textrm{FWHM}=3.8$ px) and a double-peaked He i $5876\,\rm{\AA}$ emission line suggests SB2 binarity. Furthermore, 2M0556+0803 lies in a region of the CMD surrounded by other Li-depleted group members (to within the uncertainties on its photographic $V$-band magnitude), consistent with an M4 spectral type. We therefore retain 2M0556+0803 as a possible member to be re-examined when further spectroscopic observations become available.

4.4.6 2MASS J05572121+0502158

2M0557+0502 satisfies the proper motion selection criterion ($\Delta_{\rm{PM}}=1.3$ mas yr*-1*) and at a kinematic distance of 115 pc was one of a small number of distance outliers in our sample observed with WiFeS. The star’s radial velocity, however, is strongly inconsistent with membership ($\Delta_{\rm{RV}}=30.4$ km s*-1*). The presence of a strong Li line ($\rm{EW[Li]}=290$ mÅ) suggests 2M0557+0502 is young, but we also find its rising spectrum is significantly reddened [$E(B-V)\approx 0.6$ mag], with an underlying spectral type of K2. This level of reddening is more than an order of magnitude greater than the group mean of $E(B-V)=0.03$ mag (see Section 5.1). Assuming $(V-K_{\rm{s}})_{\circ}\approx 2.3\,\rm{mag}$ appropriate for an early K-dwarf and a de-reddened $M_{V}\approx 7.2$ mag at 115 pc, 2M0557+0502 lies 1–2 mag below the 32 Ori group CMD sequence and is unlikely to be a member, especially if it is an SB1. Given the strong reddening and large radial velocity it could be a Li-rich background giant but we seek a better spectroscopic characterisation before finalising its evolutionary and membership status.

4.5 Comparison to previous non-spectroscopic surveys

Since its initial discovery by Mamajek (2007), there have been no surveys dedicated specifically to discovering new members of the 32 Ori group, with new members being individually added on the basis of common attributes like proper motion and radial velocity (Mace et al., 2009; Franciosini & Sacco, 2011). Given that in this work we aim to characterise the stellar population of the 32 Ori group, it is important that we place our new members in context by re-examining previous non-spectroscopic memberships from large surveys of Galactic open clusters and associations.

As part of their global catalogue of Milky Way clusters, Kharchenko et al. (2013) recovered the 32 Ori group, arguing that it comprises 40 members within a radius of 2.2°, lies at a distance of 95 pc and has an age of $32\,\rm{Myr}$. They found a mean proper motion and radial velocity of ($\mu_{\alpha}\cos\delta,\mu_{\delta})\simeq(10.0,-32.2)\,\rm{mas\,yr^{-1}}$ ($\sigma_{\mu}\simeq\pm 0.8\,\rm{mas\,yr^{-1}}$) and $13.1\,\rm{km\,s^{-1}}$, respectively. Kharchenko et al. assigned membership probabilities using a combination of three metrics; one kinematic (based on the proper motion of a given object with respect to the cluster mean) and two photometric (based on positions in the $K_{\rm{s}}$, $J-H$ and $J-K_{\rm{s}}$ CMDs). Stars which satisfied all three metrics with a probability greater than 61 per cent were classified as ‘most probable’ members. Cross-matching their 40 most probable members with the 45 stellar group members of the current study, we find only three objects in common: HR 1807, HD 35714 and HD 36338. In other words, the Kharchenko et al. selection criteria omit 32 Ori as a member of the 32 Ori group! Furthermore, cross-matching their members against potential kinematic members from UCAC4 and URAT1 (see Section 2.1), we find 13 and 40 objects in common, respectively. Aside from HR 1807, HD 35714 and HD 36338, none of the other candidates have CMD positions or $\Delta_{\rm PM}$ values consistent with membership in the 32 Ori group.

The main reason the Kharchenko et al. (2013) membership for late-type objects is particularly unreliable is due to their reliance on near-infrared 2MASS CMDs. At low effective temperatures ($T_{\rm{eff}}\lesssim$ 4000 K), isochrones in both the $K_{\rm{s}}$, $J-H$ and $J-K_{\rm{s}}$ CMDs become vertical and essentially degenerate with age. This removes any meaningful photometric distance information and means that the study was reliant solely on deeper but less accurate PPMXL proper motions for membership. We note that within their cluster radius of $2.2\degr$ there are 15 new members from Table 3 which could have provided matches. Of these, 13 failed the kinematic test and two failed the photometric tests. Clearly these tests are too restrictive and the PPMXL proper motions not accurate enough to rely on for membership determinations.

The 32 Ori group is also listed (as Mamajek 3) in the catalogue of optically-visible open clusters by Dias et al. (2002, v.2.5; 2005). Recently, Dias et al. (2014) used UCAC4 astrometry to provide membership probabilities for individual stars by fitting the observed proper motion distribution in a region surrounding each cluster with two elliptical bivariate populations. Based on an apparent diameter of 250 arcmin, Dias et al. (2014) identified over $2.8\times 10^{4}$ UCAC4 counterparts, of which $2.3\times 10^{4}$ were assigned to the 32 Ori group, making it the twelfth most populous cluster in the entire catalogue. From these members they estimated a mean group proper motion of ($\mu_{\alpha}\cos\delta,\mu_{\delta})=(0.55,-2.75)\,\rm{mas\,yr^{-1}}$, which is not only inconsistent with our proper motion (see Table 2), but also the proper motion listed in the original Dias et al. (2002) catalogue. Cross-matching our 45 stellar members with the sources identified by Dias et al., we find only 15 objects in common (11 of which are previously known members, see Table 1; 28 lie outside the adopted diameter of the group). None has a membership probability greater than 69 per cent. As with the Kharchenko et al. (2013) study, 32 Ori itself appears to be a non-member of its own group, making it hard to place much credibility in the Dias et al. results. Their membership probabilities are based solely on the proper motion of an object relative to the mean, which alone is not sufficient to unambiguously demonstrate membership.

5 Properties of the 32 Ori group

Below we use the previously known and new members of the 32 Ori group to investigate its global properties; namely interstellar reddening, age, circumstellar disc frequency and spatial structure.

5.1 Reddening and extinction

Before attempting to determine an age for the 32 Ori group from its CMD, the effects of interstellar reddening must first be accounted for. The 32 Ori group lies at a distance of $\sim$ 90 pc, and whilst reddenings for stars at this distance are typically low [$E(B-V)\lesssim 0.05\,\rm{mag}$], the complex shape of the Local Bubble and surrounding clouds is such that the reddening is not necessarily negligible (Reis et al., 2011).

In Table 8 we list $E(B-V)$ estimates for the bright B-, A- and F-type members of the 32 Ori group using the best available spectral types and photometry. For the two B-type members (unresolved 32 Ori and HR 1807) we adopt $UBV$ photometry from Mermilliod (2006) and combine this with the Q-method (Pecaut & Mamajek, 2013) to derive negligible reddenings of $E(B-V)=0.01$ and 0.00 mag, respectively. Using resolved $BV$ photometry from Tycho-2, and adopting spectral types of B5 and B7 for 32 Ori A and B from Edwards (1976), and A7 and A7.5 for HD 36002 and HD 36823, we estimate $E(B-V)=0.01$, 0.06, 0.03 and 0.02 mag, respectively. For the five other members, we fit the $BVJHK_{\rm{s}}$ photometry (and for HR 1807, $U$-band from Johnson, 1966) to the intrinsic dwarf colour locus of Pecaut & Mamajek (2013), calculating $\chi^{2}$ fits for spectral energy distributions (SEDs) spaced in $T_{\rm{eff}}$ by 10 K and in steps of $E(B-V)=0.01\,\rm{mag}$. The quoted $E(B-V)$ values reflect the distribution of those fits for which the $\chi^{2}$ probability $q>5$ per cent over the range of $T_{\rm{eff}}$ reflecting a $\pm$1 subtype uncertainty in spectral type (where B9.5V to A0V was counted as a full step in subtype). The uncertainties in the $E(B-V)$ estimates from the SED fitting are typically $\pm$0.02 mag.

Based on the values in Table 8 we adopt a median reddening toward the group of $E(B-V)=0.03\pm 0.01$ mag, which translates to a $V$-band extinction of $A_{V}=0.10\pm 0.03$ mag and a $K_{\rm{s}}$-band extinction of $A_{K_{\rm{s}}}=0.011\pm 0.004$ mag (following Bilir et al., 2008). The range of values [$E(B-V)\lesssim$ 0.06 mag] are suggestive of small amounts of patchy extinction across the core of the group. Note that the $K_{s}$-band extinction is smaller than the typical 2MASS photometric errors.

5.2 Isochronal age

Fig. 13 shows the $M_{V}$, $V-K_{\rm{s}}$ CMD of the 45 stellar members of the 32 Ori group. We preferentially use trigonometric parallaxes to transform the apparent $V$-band magnitudes listed in Table 3 to $M_{V}$, however in the majority of cases these are not available and so we adopt the kinematic distances from our selection process (see Section 2.1). Note that for parallaxes from the recent Gaia DR1 we include the additional systematic uncertainty of $\pm 0.3$ mas as discussed in Gaia Collaboration et al. (2016). Our adoption of the best-fit kinematic distances is in contrast to Bell et al. (2015) in which we simply adopted the mean group distance for the 11 stars listed in Table 1 without parallaxes (barring HD 36823 which was not included in that study), and has resulted in a less tight colour-magnitude sequence in Fig. 13. Uncertainties on $M_{V}$ incorporate the uncertainties in the proper motions used to calculate the kinematic distances (typically several mas$\,\rm{yr^{-1}}$) and result in distance uncertainties in the range of 8–30 per cent.

We derive an isochronal age for the 32 Ori group using the same method as that described in previous papers (see e.g. Bell et al., 2014, 2015), namely by fitting two-dimensional probability distributions to the CMD using the $\tau^{2}$ fitting statistic of Naylor & Jeffries (2006) and Naylor (2009). In brief, the probability distributions are created using stellar evolutionary models and not only include binarity but also incorporate an empirical colour-$T_{\rm{eff}}$ relation and bolometric corrections from observations of low-mass Pleiades members, as well as theoretical corrections for the dependence on surface gravity. For our isochronal age analysis, we have adopted the following interior models: Dartmouth (Dotter et al., 2008), PARSEC (Bressan et al., 2012), BHAC15 (Baraffe et al., 2015), and Pisa (Tognelli, Prada Moroni & Degl’Innocenti, 2015). Note that the Pisa models are based on the same calculations as described in Tognelli et al. (2015), however they cover a mass range of $0.08<M/\rm{M_{\odot}}<9$. As per Section 5.1, we have included the effects of interstellar reddening, adopting a group mean $E(B-V)=0.03\,\rm{mag}$. Table 9 lists the individual isochronal ages for the 32 Ori group in addition to our mean isochronal age of $25\pm 5\,\rm{Myr}$ ($\pm 4\,\rm{Myr}$ statistical, $\pm 3\,\rm{Myr}$ systematic).

5.3 Lithium depletion boundary age

The lithium depletion boundary (LDB) is defined as the sharp transition between stars which have contracted sufficiently that their core temperatures have reached the critical value to burn Li and thus exhibit depleted levels of Li in their photospheres, and those stars at slightly lower masses which have not yet reached this critical temperature and show undepleted levels. Over the past few years the LDB has been extolled as the least model-dependent, absolute age-dating technique for coeval populations with ages of between $\sim$20 and 200 Myr (see e.g. Soderblom et al., 2014).

As shown in Fig. 13, the LDB in the 32 Ori group is defined by the Li-poor M4.5 member 2MASS J05313290+0556597 (THOR 30) and the Li-rich ($\textrm{EW[Li]}=600$ mÅ) M4.5 member 2MASS J05264073+0712255 (THOR 32). We note that neither of these stars appear to be an unresolved binary which could significantly affect the determination of the LDB luminosity. To derive an LDB age for the group we adopt the method described in Binks & Jeffries (2014, 2016b), which involves defining a region in the CMD at which the LDB is located and then fitting curves of constant luminosity corresponding to 99 per cent Li depletion. These curves are created using stellar evolutionary models and after transforming $L_{\rm{bol}}$ and $T_{\rm{eff}}$ to absolute $V$-band magnitude and $V-K_{\rm{s}}$ colour using the pre-MS bolometric correction and colour relations of Pecaut & Mamajek (2013). Our threshold for Li-poor stars (EW[Li] $<100\,\rm{m\AA}$) ensures that, in comparison to the Li-rich stars which exhibit undepleted levels of Li consistent with measurements of young stars at birth (see e.g. Palla et al., 2007), the difference in depletion between the Li-poor and Li-rich stars is greater than a factor of 100, and so calculating the LDB age in such a manner is entirely justified (see also Jeffries & Oliveira, 2005; Tognelli et al., 2015).

To define the LDB region we simply adopt the central position between the two stars and the separation in both colour and magnitude, yielding $V-K_{\rm{s}}=5.462\pm 0.094$ and $M_{V}=11.308\pm 0.055$ mag. These values are consistent with those reported by Binks & Jeffries (2014) for the BPMG. The LDB age is then calculated by fitting the Li depletion curves to this point and the uncertainty on the age calculated from the size of the region in both colour and magnitude. Note that we have also included an additional uncertainty of 0.1 mag in colour and 0.3 mag in $M_{V}$ to reflect likely uncertainties in both the photometric calibration (especially the inhomogeneous $V$-band photometry) and kinematic distances. Table 9 lists the individual LDB ages derived from the following sets of evolutionary models: Dartmouth (including a new prescription for magnetic fields, $\langle Bf\rangle=2.5\,\rm{kG}$; Feiden & Chaboyer, 2013, 2014), BHAC15 and Pisa, which together yield a mean LDB age of $23\pm 4\,\rm{Myr}$ ($\pm 4\,\rm{Myr}$ statistical, $\pm 1\,\rm{Myr}$ systematic). The individual LDB ages from the non-magnetic Dartmouth, BHAC15 and Pisa models all agree to within 1 Myr, whereas the magnetic Dartmouth models imply an older, yet consistent, age for the group. These findings are similar to those reported by Malo et al. (2014) and Binks & Jeffries (2016b) for the BPMG.

5.4 Final adopted age

The age analyses above clearly demonstrate that the isochronal and LDB ages for the 32 Ori group are in agreement. Combining these two age determinations, we calculate a final adopted age of $24\pm 4\,\rm{Myr}$ ($\pm 4\,\rm{Myr}$ statistical, $\pm 2\,\rm{Myr}$ systematic). We note that this age is essentially identical to the $23\pm 3\,\rm{Myr}$ age for the BPMG derived by Mamajek & Bell (2014). The uncertainty in our final adopted age is driven by the statistical uncertainty which stems from i) the reasonably large uncertainties on the kinematic distances and ii) the inhomogeneous $V$-band photometry collated from the available literature. Both of these points will be directly addressed by Gaia data releases in the coming years, by providing parallaxes and well-calibrated, homogeneous $G$-band photometry for all members of the group. This will naturally lead to a tighter sequence in the CMD compared to that shown in Fig. 13 and will allow more statistically-robust age determinations.

5.5 Circumstellar disc frequency

Given that the 32 Ori group appears to be essentially coeval with the BPMG, which is known to harbour several optically-thin debris discs (e.g. $\beta$ Pic itself, 51 Eri and AU Mic), it is conceivable that such discs may also be present in the 32 Ori group. These could present ideal targets for direct imaging in an attempt to discover gas giant planets and further constrain potential planetary formation mechanisms. As discussed in the Introduction, the Spitzer IRAC and MIPS survey described by Shvonski et al. (2016) demonstrates that 4/14 members (not including HD 36823) exhibit $24\,\rm{\mu m}$ excess emission, and hence these stars are likely debris disc candidates. In this Section we search for further evidence of circumstellar material around the combined 32 Ori group census.

To identify potential circumstellar discs we cross-matched our membership against the AllWISE catalogue (Cutri & et al., 2014). All 45 stars have counterparts in the $W1$ $(3.4\,\rm{\mu m})$, $W2$ $(4.5\,\rm{\mu m})$ and $W3$ $(12\,\rm{\mu m})$ bands, however only 19 were detected in the $W4$ $(22\,\rm{\mu m})$ band. Fig. 14 plots the $K_{\rm{s}}-W1$, $K_{\rm{s}}-W2$, $K_{\rm{s}}-W3$ and $K_{\rm{s}}-W4$ infrared colours as a function of spectral type for our 32 Ori group members. To determine whether a given star exhibits an infrared excess we use the pre-MS colour sequence of Pecaut & Mamajek (2013), which essentially defines photospheric colours for stars with ages of $\simeq$ 5–40 Myr. Due to a combination of saturation effects and a paucity of early-type stars within the Local Bubble, the Pecaut & Mamajek photospheric colours only cover spectral types of F0 and later in the WISE bandpasses. For earlier spectral types we adopted the atlas9 odfnew synthetic colour indices as calculated by Pecaut & Mamajek, which are based on the atmospheric models of Castelli & Kurucz (2004) and for which we assumed log$\,g=4.5$ dex, and interpolated within these for the $T_{\rm{eff}}$ corresponding to the specific spectral type as prescribed by the dwarf spectral type-$T_{\rm{eff}}$ relation of Pecaut & Mamajek. We define an excess in a given WISE band as any object which lies greater than 3$\sigma$ above the Pecaut & Mamajek relation, where $\sigma$ is the photometric uncertainty on the observed colour. We stipulate, however, that for a source to be labelled an excess object it must also exhibit excesses in each of the longer wavelength WISE bandpasses.

The terminology of circumstellar discs can vary depending on which criteria one adopts, thus making like-for-like comparisons between different regions problematic. We therefore adopt the observational criteria described in Luhman & Mamajek (2012) which has been used in several recent studies of nearby young moving groups and associations (see e.g. Kraus et al., 2014; Pecaut & Mamajek, 2016). Following Luhman & Mamajek, we can eliminate any full, transitional or evolved discs which all require an excess of $E(K_{\rm{s}}-W4)>3.2\,\rm{mag}$, yielding a disc fraction of $<$ 3.9 per cent (68 per cent confidence level) for such discs. Debris discs are classified as objects with excesses of $E(K_{\rm{s}}-W4)<3.2\,\rm{mag}$, which comfortably covers all of the excess objects identified in the bottom-right panel of Fig. 14. Of the 19 objects with $W4$ detections, we identify six – HR 1807 (THOR 2), 2MASS J05284050+0113333 (THOR 4B), HD 36338 (THOR 6), 2MASS J05274313+1446121 (THOR 17A), 2MASS J05274855+0645459 (THOR 27) and 2MASS J05243009+0640349 (THOR 38) – as exhibiting excess emission and derive a debris disc fraction of $32^{+12}_{-8}$ per cent based on binomial statistics (see Cameron, 2011). Of these six stars, THOR 2, 17A and 38 also display varying degrees of excess $W3$ emission. The mid-infrared AllWISE photometry and excesses for these debris disc candidates are listed in Table 10. In Fig. 15 we show the SEDs for these objects, created from a combination of optical, near- and mid-IR photometric data including Tycho-2 $(BV_{\rm{T}})$, APASS DR9 $BVgri$, 2MASS $JHK_{\rm{s}}$ and AllWISE $W1$–$W4$. Using the Virtual Observatory SED Analyser (VOSA; Bayo et al., 2008) we fitted the observed SEDs with the solar-metallicity BT-Settl CIFIST atmospheric models of Allard, Homeier & Freytag (2011), computed adopting the revised solar abundances of Caffau et al. (2011). Note that due to the upper limit of 7000 K on the CIFIST models, in the case of HR 1807 we instead used the atlas9 odfnew models of Castelli & Kurucz (2004).

The obvious dataset against which to compare our census of circumstellar discs is the recently published Spitzer survey by Shvonski et al. (2016). Our WISE analysis corroborates that of Shvonski et al. with neither study finding evidence for the presence of full or warm dusty discs based on a lack of excess detections in all four IRAC bands as well as the WISE $W1$ and $W2$ bands. Of the four debris disc candidates identified by Shvonski et al., we find that both HR 1807 (THOR 2) and HD 36338 (THOR 6) are $22\,\rm{\mu m}$ excess sources, however we do not detect any excess emission from either HD 35499 (THOR 5) or TYC-112-1486-1 (THOR 10), despite Shvonski et al. determining both exhibit $24\,\rm{\mu m}$ emission at $>4\sigma$ above typical photospheric values. The cause of this discrepancy is unclear as neither object is flagged as variable in the AllWISE catalogue (but see Melis et al., 2012) and visual inspection of the co-added images suggest nothing unusual (e.g. nearby sources, non-uniform background). In the case of TYC 112-1486-1 this could be a sensitivity issue as the signal-to-noise ratio of the $W4$-band measurement is only $\sim$3. The remaining four debris disc candidates identified in this study are all new members and thus were not included in the Shvonski et al. study.

Of the six objects identified as exhibiting $W4$ excesses, the four M dwarf debris disc candidates are of particular note. Studies have demonstrated that as stellar mass decreases, so does the fraction of associated debris discs (see e.g. Lestrade et al., 2009). To date, only three M dwarf debris discs have been confirmed via scattered light observations, all with ages similar to or younger than the 32 Ori group (AU Mic, TWA 7 and TWA 25; see Choquet et al., 2016). If the new debris disc candidates presented in this study (as well as those in the BPMG; see Binks & Jeffries, 2016a555Recently, Silverberg et al. (2016) claimed to have identified an M dwarf debris disc candidate (WISE J080822.18-644357.3) in the 45 Myr-old Carina association. Their membership is based solely on the SPM4 proper motion, which when combined with the BANYAN II Bayesian membership tool (Gagné et al., 2014), provides a membership probability of 93.9 per cent. Given this reliance on a single observable, it is therefore perturbing to note that the PPMXL $\mu_{\delta}$ value differs from that provided by SPM4 by $\simeq$ 19 $\rm{mas\,yr^{-1}}$. Prior to confirming membership, we await an improved proper motion and spectroscopic observations of the star, most importantly a radial velocity in agreement with the $\sim$ 21 km s*-1* expected of a Carina member at that position. ) can be confirmed by ancillary mid- and far-IR observations, then these would not only represent some of the oldest and lowest-mass stars with such discs, but they would also provide ideal targets for follow-up direct imaging which may help to enhance our understanding of the processes governing the production and dynamics of dust and planetesimals in such systems during an epoch which is believed to be important for terrestrial planet formation.

Fig. 16 shows the debris disc fraction of the 32 Ori group compared to other nearby young groups/associations and clusters (based on excess emission at either 22 or $24\,\rm{\mu m}$). The disc fractions for all other regions have been taken from the compilation of Zuckerman et al. (2012) and the uncertainties calculated as above using binomial statistics. The ages and uncertainties have been compiled from Barrado y Navascués, Stauffer & Jayawardhana (2004) for IC 2391, $\alpha$ Per and the Pleiades, Jeffries & Oliveira (2005) for NGC 2547, Bell et al. (2015) for TWA, $\eta$ Cha, the BPMG and Tuc-Hor/Columba, and Pecaut & Mamajek (2016) for UCL/LCC.

5.6 Spatial structure of the 32 Ori group

Fig. 17 shows the $XYZ$ spatial and $UVW$ velocity distributions for the 45 stellar members of the 32 Ori group. Note that due to unreliable literature radial velocities for both HR 1807 and HD 35714, and missing radial velocities for HD 36823 and HD 36002 (SB2), these stars are not shown in the $UVW$ panels of Fig. 17. From these distributions we estimate a mean group position of $(X,Y,Z)=(-89.1\pm 2.1,-26.0\pm 1.3,-24.0\pm 0.7$) pc and velocity of $(U,V,W)=(-13.1\pm 0.4,-18.9\pm 0.2,-9.0\pm 0.2)\,\rm{km\,s^{-1}}$, where the uncertainties on each represent the standard error of the mean. Due to the use of kinematic distances for the majority of the stars shown in Fig. 17, we would advise that, until additional Gaia parallaxes become available, the mean group $UVW$ velocity listed in Table 2 be adopted for future searches for additional members.

Examining the $XYZ$ distribution we see that the geometry of the 32 Ori group is broadly ellipsoidal and elongated toward the Galactic centre, with ($\Delta_{X},\Delta_{Y},\Delta_{Z}$) $\sim$ $(60,25,20)$ pc. This is quite different from more filamentary/sheet-like geometries like those observed in TWA and Tuc-Hor (see Weinberger, Anglada-Escudé & Boss, 2013; Kraus et al., 2014). Based on our derived age for the group, and assuming that all members formed within a region $\ll$1 pc, the current dispersion in the $XYZ$ plane suggests a one-dimensional internal velocity dispersion on the order of 1–$2\,\rm{km\,s^{-1}}$, which is consistent with those of other young moving groups and associations in the Local Bubble (Mamajek, 2016). Furthermore, as argued by Kraus et al. (2014) for Tuc-Hor, the age of the 32 Ori group is much less than one Galactic orbital period and so we would not expect the tidal field of the Milky Way to have significantly influenced the current geometry of the group. Hence, the ellipsoidal shape more likely reflects the initial star formation conditions in which the group formed and could be more indicative of an originally compact cluster which has since become gravitationally unbound and is slowly dispersing into the Galactic field, as proposed for the $\epsilon$ Cha group (Murphy, Lawson & Bessell, 2013).

6 Conclusions

We have undertaken the first large-scale systematic stellar census of the nearby but poorly-studied 32 Orionis group. The main results from this study are as follows:

Based on spectroscopic follow-up of candidate members selected from UCAC4 and URAT1, we have identified a total of 30 new members of the 32 Ori group; 29 M dwarf members and one A-type star which forms a co-moving common proper motion pair with one of the new members. Members were confirmed by combining Li absorption, H$\alpha$ emission and radial velocity information with kinematic distances and proper motions. This study has increased the number of known group members by a factor of three, bringing the total number of 32 Ori group members to 46. 2. 2.

We have unambiguously identified the Li depletion boundary (LDB) of the 32 Ori group. Using stellar evolutionary models we derive ages from both isochronal fitting ($25\pm 5\,\rm{Myr}$) and LDB analyses ($23\pm 4\,\rm{Myr}$), which we combine to calculate a final adopted age for the 32 Ori group of $24\pm 4\,\rm{Myr}$ ($\pm 4\,\rm{Myr}$ statistical, $\pm 2\,\rm{Myr}$ systematic). This age implies that the 32 Ori group is coeval with the somewhat closer $\beta$ Pictoris moving group. 3. 3.

We have searched for the presence of circumstellar discs around the 45 stellar members of the 32 Ori group using the AllWISE catalogue. As with other groups and associations of similar age, we find no evidence for prevailing warm, dusty discs, however we have identified several possible debris discs based on excess emission in the WISE $W4$-band at 22 $\mu$m. From our limited sample of stars with $W4$ detections, we estimate a debris disc fraction of $32^{+12}_{-8}$ per cent for the group.

Acknowledgements

CPMB acknowledges support from the Swiss National Science Foundation (SNSF). SJM acknowledges support from a University of New South Wales Vice Chancellor’s Fellowship. EEM acknowledges support from the University of Rochester School of Arts and Sciences, NSF grant AST-1313029, and the NASA NExSS program. Part of this research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration. The authors would like to thank the ANU TAC for the generous allocation of telescope time which made this study possible as well as Alex Binks for discussions regarding LDB ages, and John Stauffer and Sarah Schmidt for discussions regarding the emission feature seen in 2M0536+1300. The authors would also like to thank the referee Ben Zuckerman for comments which have improved the paper. This research has made extensive use of the VizieR and SIMBAD services provided by CDS as well as the Tool for OPerations on Catalogues And Tables (TOPCAT) software package (Taylor, 2005). This publication makes use of VOSA, developed under the Spanish Virtual Observatory project supported from the Spanish MICINN through grant AyA2011-24052. This work has made use of data from the European Space Agency (ESA) mission Gaia (http://www.cosmos.esa.int/gaia), processed by the Gaia Data Processing and Analysis Consortium (DPAC, http://www.cosmos.esa.int/web/gaia/dpac/consortium). Funding for the DPAC has been provided by national institutions, in particular the institutions participating in the Gaia Multilateral Agreement.