THE CHANDRA ACIS SURVEY OF M33: X-RAY, OPTICAL, AND RADIO PROPERTIES OF THE SUPERNOVA REMNANTS

The vast majority of supernova remnants (SNRs) in nearby galaxies have been identified using optical interference-filter imagery where the ratio of [S ii] to Hα emission discriminates between SNRs and H ii regions. Pioneered by Mathewson & Clarke (1973), this method is based on the fact that radiative shocks contain cooling zones with temperatures near 10,000 K, where S⁺ is the dominant ionization stage of S, whereas in (bright) H ii regions S⁺⁺ is the dominant ionization state. As a result, [S ii]:Hα ratios of 0.4 to 1 are common in SNRs, whereas in bright H ii regions the ratio is typically 0.1 (D'Odorico & Sabbadin 1976; Levenson et al. 1995).

M33, a member of the Local Group, is the nearest late-type spiral galaxy. At a distance of 817 ± 58 kpc (Freedman et al. 2001), an SNR with a diameter of 20 pc would have an angular size of 50. The galaxy is relatively face-on (i = 56° ± 1°; Zaritsky et al. 1989), lies along a line of sight with low Galactic absorption ($N_H \sim \mbox{\rm $6 \times 10^{20} \rm \:cm^{-2}$}$; Dickey & Lockman 1990; Stark et al. 1992) and, as a result, has been an obvious choice for characterizing the SNR population in nearby galaxies. D'Odorico et al. (1978) identified the first three SNRs in M33 using interference-filter imagery; other subsequent studies, principally by D'Odorico et al. (1980), Long et al. (1990), and Gordon et al. (1998), have increased the number of remnant candidates to nearly 100.

In contrast to M33, the vast majority of SNRs in the Galaxy were first recognized from radio observations, based on their extended non-thermal emission. Since most SNRs are located near the Galactic plane implying high line-of-sight absorption to most objects, even today a majority of the Galactic SNRs remain undetected in [S ii] and Hα. Radio searches for SNRs have also been carried out in nearby galaxies; these provide an independent search technique that can detect several classes of SNRs, including Crab-like remnants, very young objects like Cas A, and the so-called Balmer-dominated SNRs, which emit little or no [S ii] and thus are unlikely to be discovered in optical Hα-[S ii] searches. In the case of M33, Goss et al. (1980) were the first to detect radio emission from three of the bright optical SNRs, and the number of radio detections has increased with time. Gordon et al. (1999) used images constructed from data taken with the Very Large Array (VLA) and the Westerbork Synthesis Radio Telescope (WSRT) to detect 56 of the optical candidates as radio sources. However, no one has successfully carried out a truly independent survey to find extended, non-thermal radio sources as has been done in the Galaxy; furthermore, no Crab-like SNRs have been found in M33 (Reynolds & Fix 1987).

The Einstein Observatory was the first X-ray experiment with sufficient sensitivity and angular resolution to detect extragalactic SNRs at X-ray wavelengths, although most of these were, not surprisingly, in the Magellanic Clouds (Long & Helfand 1979; Seward & Mitchell 1981). Long et al. (1981), in the initial Einstein study of M33, suggested that the bright X-ray source M33 X-3, located near NGC 592, was an SNR because of its very soft X-ray spectrum. Markert & Rallis (1983) reported that this source was time-variable, and its nature was, for a time, uncertain. Today we know this object as G98-21, the brightest M33 X-ray SNR (Gordon et al. 1993; Gaetz et al. 2007). Later, using ROSAT, Long et al. (1996) found X-ray counterparts to 10 optically identified SNRs. More recently, Misanovic et al. (2006) reanalyzed the XMM-Newton data originally discussed by Pietsch et al. (2004), identifying 15 non-variable X-ray sources with optical SNRs and suggesting 11 other X-ray sources might be SNRs based on their X-ray spectra. Ghavamian et al. (2005) inspected interference-filter images of the candidates and found that two, XMM068 and XMM270, were located along lines of sight to nebulae with elevated [S ii]:Hα ratios that had been missed by Gordon et al. (1998) as well as earlier optical observers.¹²

Although XMM-Newton has sufficient sensitivity to detect the brighter M33 remnants, Chandra is the first X-ray observatory with sufficient sensitivity and spatial resolution that one might expect to discover new SNRs in galaxies as close as M33 using X-ray observations alone. With a resolution of 05 on-axis, and <5'' up to 8' off-axis (Weisskopf et al. 2002), Chandra should resolve most remnants with diameters ≳15 pc. Ghavamian et al. (2005) showed the effectiveness of Chandra for detecting SNRs, finding a total of 22 X-ray sources that aligned with optically identified SNRs in the data that existed at that time, which had typical exposures of 50 ks. For this reason among others, we have undertaken a series of deep observations of M33 with Chandra, the ChASeM33 survey. Plucinsky et al. (2008) provided a "First Look" at the results from ChASeM33 based on approximately half of the data; there we identified 26 SNRs from the list of Gordon et al. (1998).¹³ In addition, Gaetz et al. (2007) provided a detailed analysis of G98-21 (M33 X-3), which appears as a 5'' shell on the edge of a bright H ii region.

The purpose of the present study is to provide a full analysis of the ChASeM33 survey relevant to SNRs, in combination with supplemental optical and radio data. A detailed analysis of the survey as a whole, and the point sources that have been detected, will be reported by R. Tüllmann et al. (2010, in preparation). The remainder of this paper is organized as follows: we describe the ChASeM33 survey and our approach to extracting information about SNRs in Section 2. We then discuss our construction of a set of candidate SNRs (Section 3) and report the detection of more than half this sample as X-ray sources in Section 4. We next describe our attempts to better characterize the sample using a combination of new and archival optical and radio observations of M33 (Section 5) and compare the X-ray, optical, and radio properties of the SNRs in Section 6. We explore the question of why we have detected some objects and not others in Section 7 and then describe the images and spectra of the brightest X-ray SNRs in M33 in Section 8. We discuss which Galactic SNRs we would have detected in M33 and compare the samples of SNRs in M33 to those from other galaxies in Section 9. We summarize our conclusions in Section 10. Finally, in the Appendix, we provide an atlas of X-ray and optical images of M33 SNRs and SNR candidates, including details on why some might have been misclassified.

The strategy for the ChASeM33 survey has been described in detail by Plucinsky et al. (2008). The survey comprises ACIS-I observations of seven overlapping fields which, as shown in Figure 1, cover the inner region of M33 to a radius of about 18', or 4.3 kpc. Of the 98 SNRs cataloged by Gordon et al. (1998), 93 (or about 95% of the SNRs identified in the last systematic optical-only search of M33) lie within the area covered by ChASeM33. All of the data for the survey have now been obtained. Each field was observed for a total of ∼200 ks, but because of overlaps the total exposure over a substantial portion of the inner galaxy exceeds 400 ks. As a result, the limiting 0.35–2 keV sensitivity is about 2 × 10³⁴ erg s^-1 (2σ), assuming emission from a thermal plasma with kT of 0.6 keV, a metal abundance of 0.5 times solar, and a line-of-sight absorption of 1 × 10²¹ cm⁻².¹⁴

Zoom In Zoom Out Reset image size

The starting point for our analysis is the same screened data that are being used for the creation of the point-source catalog; details can be found in R. Tüllmann et al. (2010, in preparation). These data include not only those obtained by us as part of ChASeM33 but also ACIS-I data from observations which are available from the Chandra archive (ObsIDs 1730 and 2023). Limited data from chip S3 of the ACIS spectroscopic array exist, but we have not included these in our analysis because this chip has higher background and a very different energy response from that of the I-array CCDs. The ChASeM33 data were obtained between 2005 September and 2006 November. Each observation of a field, except that of Field 6, was split into two epochs to enable a more comprehensive search for variable sources. Due to scheduling constraints ObsID 7208, which is part of Field 6 epoch 1, was observed at a different roll angle and has much lower exposure; we analyzed this observation separately. The data used, including the exposure resulting in good data at each position, are summarized in Table 1.

Table 1. ChASeM33 Observation Log

Field	R.A.	Decl.	Roll Ang.	Exposure
	(J2000)	(J2000)	($\deg$)	(ks)
F1e1	01:33:51.15	+30:39:20.5	308.5	93.1
F1e2	01:33:50.18	+30:39:51.4	142.0	92.1
F2e1	01:34:13.20	+30:48:03.0	140.2	94.3
F2e2	01:34:13.46	+30:48:04.3	127.9	98.2
F3e1	01:33:33.30	+30:48:55.6	140.2	89.4
F3e2	01:33:33.47	+30:48:56.5	132.2	98.4
F4e1	01:33:08.20	+30:40:10.5	262.2	96.7
F4e2	01:33:09.20	+30:40:41.1	97.8	90.7
F5e1	01:33:27.20	+30:31:39.4	145.7	100.5
F5e2	01:33:27.40	+30:31:40.7	135.7	89.2
F6e1	01:34:06.48	+30:30:26.6	224.2	86.4
F6e2	01:34:07.92	+30:30:49.3	103.2	98.7
F7e1	01:34:33.54	+30:39:00.4	94.9	88.5
F7e2	01:34:32.55	+30:38:29.6	265.7	95.6
1730	01:33:50.91	+30:39:54.3	108.6	46.7
2023	01:34:34.63	+30:47:58.2	106.6	88.3
7208	01:34:06.99	+30:30:18.9	259.4	11.5

Download table as:  ASCIITypeset image

Our goal was to extract information from the X-ray data about all known and candidate SNRs in M33. For our analysis we used ACIS Extract (AE) version 2009-08-12 (Broos et al. 2002), which is designed specifically to deal with source crowding, to determine source properties like net counts, fluxes, and significances, and to extract spectra simultaneously from multiple observations with overlapping fields of view. The general approach we used to account for the spatial extent of the SNRs, designed for the present case where source extent is of the same order of magnitude as the point-spread function (PSF), is described here.

We began with separate lists of SNR/SNR candidates and point-source positions. The construction of our SNR/SNR candidate list is discussed in more detail in Section 3, but the list basically consists of all of the objects we and others have suggested are SNRs as well as objects that we felt were plausible candidates based on inspection of the X-ray data and other archival data at our disposal. Although our interest is in the SNRs, the point-source list is also required so that the SNR fluxes can be calculated using background regions that exclude point sources. Since our analysis of the SNRs took place in parallel with the analysis of the entire ChASeM33 survey, our point-source list consisted of a near-final version of the point-source catalog, with those sources we have identified with SNR candidates removed.

We first processed all of the sources together as if they were point sources. The primary purpose of this step was to obtain two sets of region files, one describing the PSF at the position of all of the SNRs in each of the observations and the other describing the areas of each X-ray image affected by point sources.¹⁵ The PSFs and background regions for a given source vary, depending on the roll angle and off-axis angle of each observation.

Since SNRs in M33 are not point sources but have typical sizes of 3'' to >30'', the count rates extracted by AE in the first processing step (which used region files appropriate for point sources) were not expected to yield accurate fluxes (or upper limits) for the SNRs; the region files must be modified to reflect both the physical size of the SNR and the local PSF in each X-ray image. We used the optical interference-filter images discussed in Section 3.1 as well as (on-axis) X-ray data to estimate the intrinsic size of each SNR or candidate object. We then used the AE point-source extraction regions at the position of each SNR convolved with these intrinsic sizes to create appropriate region files for each SNR for each of the various X-ray fields in which that SNR was observed. There were two objects for which a point source was either clearly inside a large-diameter object or was close enough to the SNR that the expanded region file included the point source in the pointings where the SNR was observed far from the telescope axis (so the PSF was large). In these cases, we constructed special region files that excluded the point sources from the extraction regions. We then reran AE using these tailored region files to extract count rates for the objects in our SNR list.¹⁶

AE is designed to handle the crowding of point sources, and it adjusts extraction regions to reduce the effects of source crowding during the source extraction process. Inserting our own region files for the extended SNRs interferes with this process. To assure ourselves that the regions we were using for source extraction and background subtraction were consistent and correct, we created a number of tools to overlay background and region files for the SNRs and nearby point sources on the X-ray images. Since the M33 field is not exceptionally crowded, there were no problems with AE attempting to resize the extraction regions in most cases. We did, however, eliminate a few observations from our analysis if there was too much overlap between the SNR and a contaminating point source at large off-axis angles (>7'), or where the SNR was very close to the edge of the field. As a result of this effort, we re-iterated our analysis until we were satisfied that we had obtained accurate count rates for all SNRs and SNR candidates.

When we began this study there were roughly 100 objects in M33 that had been suggested as SNRs based almost entirely upon optical interference-filter imagery. Follow-up optical spectroscopy has been carried out for 72 of these objects and has confirmed that the ratio of [S ii]:Hα is high in essentially all objects observed. Most (98) of these objects are summarized in the study reported by Gordon et al. (1998) in which a number of us participated. We used the Gordon list as our initial set of candidates. Some of these objects are classic spherical shells of high surface brightness, but the majority are not (see Section 7).

The Gordon et al. (1998) list is not complete, in part because the images available at the time did not include all of the outer portions of the galaxy, and in part because the identification of an SNR using the [S ii]:Hα ratio at a particular flux or surface-brightness limit depends upon the amount of confusion by other sources of emission—H ii regions or other diffuse gas in the vicinity. A few objects, including the two XMM-Newton-discovered sources noted earlier, were simply overlooked. These facts, together with the existence of better optical and radio data than those used by Gordon et al. (1998; some first reported here), have led us to expand the list of candidates based on new interference-filter imagery and on the X-ray data, as we describe in this section.

We were fairly liberal in identifying candidates, since X-ray detection provides strong confirmation that an object is a remnant, and our goal is to construct a complete remnant catalog. This means that some objects in the candidate list may not be SNRs but, rather, related entities such as wind-blown bubbles or superbubbles. The vast majority of objects are, however, bona fide SNRs. We return to the question of both the completeness of our SNR list and the possibility that some objects are not SNRs in Section 7.3, when we have the X-ray count rates and upper limits in hand.

3.1. New Optically Selected SNR Candidates

3.2. X-ray-selected SNR Candidates

3.3. The Complete SNR List

Our complete list of 137 SNRs and SNR candidates is presented in Table 3. The objects have diameters ranging from 8 to 179 pc; the median diameter is 44 pc, corresponding to a middle-aged SNR.¹⁹ The list includes all 98 of the objects described by Gordon et al. (1998), three additional objects based on our assertion that G98-43, G98-57, and G98-97 are multiple SNRs, 23 objects that were selected because of their optical properties, and 13 that were included based on their X-ray emission characteristics. The distribution of candidate objects in X-ray and optical images of the southern spiral arm of the galaxy is presented in Figure 2; close inspection shows that many of the candidates are indeed X-ray sources.

Zoom In Zoom Out Reset image size

Table 3. M33 SNRs and SNR Candidates—Basic Data

Name	Other*	R.A.	Decl.	Dia.	Morph.	Env.	Radio**	Sur. Bright†Hα	LHα	[N ii]:Hα	[S ii]:Hα	[S ii]-rat.	Spec. Ref.
		(J2000)	(J2000)	(pc)					(erg s−1)
L10-001	G98-01	01:32:30.37	+30:27:46.9	123	A	2	y	1.6e−16	9.6e+36	0.18	0.77	1.30	G98
L10-002	G98-02	01:32:31.41	+30:35:32.9	29	C	3	y	4.6e−16	1.6e+36	0.26	0.44	1.34	G98
L10-003	G98-03	01:32:42.54	+30:20:58.9	100	A	2	y	2.0e−16	8.0e+36	0.22	0.55	1.47	G98
L10-004	G98-04	01:32:44.83	+30:22:14.6	39	A	1	n	1.5e−16	9.3e+35	...	...	...	...
L10-005	XMM068	01:32:46.73	+30:34:37.8	45	A	1	n	1.5e−16	1.2e+36	0.62	0.78	1.45	MMT-BCSa,c
L10-006	G98-05	01:32:52.76	+30:38:12.6	56	A	1	?	1.9e−16	2.4e+36	0.29	0.55	1.48	S93
L10-007	(X-ray)	01:32:53.36	+30:48:23.1	73	A	1	n	3.6e−17	7.6e+35	...	...	...	...
L10-008	G98-06	01:32:53.40	+30:37:56.9	51	A	2	n	4.2e−16	4.5e+36	0.21	0.61	1.44	S93
L10-009	G98-07	01:32:54.10	+30:25:31.8	39	C	1	n	2.5e−16	1.5e+36	...	...	...	...
L10-010	G98-08	01:32:56.15	+30:40:36.4	93	A	1	y	1.9e−16	6.7e+36	0.21	0.81	1.47	S93
L10-011	G98-09	01:32:57.07	+30:39:27.1	20	A'	2	c	1.8e−15	2.9e+36	0.28	0.84	1.11	hecto
L10-012	G98-10	01:33:00.15	+30:30:46.2	52	B	3	n	2.5e−15	2.7e+37	...	...	...	...
L10-013	G98-11	01:33:00.42	+30:44:08.1	33	C	1	c	1.2e−16	5.3e+35	<0.27	0.47	1.11	G98
L10-014	G98-12	01:33:00.67	+30:30:59.3	46	B	3	n	1.9e−15	1.6e+37	...	...	...	...
L10-015	G98-13	01:33:01.51	+30:30:49.6	28	B	3	n	6.7e−16	2.1e+36	...	...	...	...
L10-016	(Opt)	01:33:02.93	+30:32:29.6	51	A	2	n	2.4e−16	2.5e+36	0.37	0.97	1.39	MMT-BCSa
L10-017	G98-14	01:33:03.57	+30:31:20.9	33	A	1	y	3.3e−16	1.4e+36	0.34	1.07	1.43	G98
L10-018	G98-15	01:33:04.03	+30:39:53.7	30	A	1	y	1.2e−15	4.4e+36	0.30	0.49	1.73	S93
L10-019	G98-16	01:33:07.55	+30:42:52.5	71	A	2	n	3.3e−16	6.6e+36	...	...	...	...
L10-020	(Opt)	01:33:08.98	+30:26:58.9	51	A	2	n	1.7e−16	1.8e+36	...	...	...	...
L10-021	G98-17	01:33:09.87	+30:39:34.9	67	A	3	y	2.1e−16	3.8e+36	0.38	0.66	1.38	S93
L10-022	G98-18	01:33:10.18	+30:42:22.0	27	B	1	n	2.0e−16	6.0e+35	0.39	0.86	1.36	hecto
L10-023	G98-20	01:33:11.10	+30:39:43.7	25	A'	2	c	3.3e−16	8.5e+35	0.36	0.78	1.11	hecto
L10-024	G98-19	01:33:11.28	+30:34:23.5	99	A	2	y	3.0e−16	1.2e+37	0.32	0.49	1.53	S93
L10-025	G98-21	01:33:11.76	+30:38:41.5	25	A	3	c	8.2e−16	2.0e+36	0.21	0.55	1.39	MMT-BCS
L10-026	(X-ray)	01:33:16.73	+30:46:10.3	73	B	2	n	2.5e−16	5.2e+36	...	...	...	...
L10-027	G98-22	01:33:17.44	+30:31:28.5	44	C	2	y	1.3e−16	1.0e+36	0.50	0.91	1.46	G98
L10-028	(Opt)	01:33:18.80	+30:27:04.4	179	A	1	n	1.2e−16	1.5e+37	...	...	...	...
L10-029	(Opt)	01:33:18.94	+30:46:51.9	66	A	1	n	9.6e−17	1.6e+36	...	...	...	...
L10-030	G98-23	01:33:21.64	+30:31:31.1	76	A	1	n	2.0e−16	4.7e+36	...	...	...	...
L10-031	G98-24	01:33:22.67	+30:27:04.0	20	A	1	c	4.4e−16	6.9e+35	0.41	0.95	1.39	hecto
L10-032	G98-25	01:33:23.85	+30:26:13.5	21	A	1	c	9.6e−16	1.8e+36	0.51	1.06	1.31	BK85
L10-033	G98-26	01:33:27.07	+30:47:48.6	67	C	2	y	1.3e−16	2.4e+36	0.27	0.57	1.46	MMT-BCS
L10-034	G98-27	01:33:28.08	+30:31:35.0	32	B	2	c	5.7e−16	2.4e+36	0.41	0.55	1.41	G98
L10-035	XMM156	01:33:28.96	+30:47:43.5	19	C	2	n	4.1e−16	5.8e+35	0.20	0.36	1.52	MMT-BCS
L10-036	G98-28	01:33:29.05	+30:42:17.0	18	A	1	c	5.0e−15	6.5e+36	0.59	1.13	1.11	S93
L10-037	G98-29	01:33:29.45	+30:49:10.8	32	B	1	y	1.8e−16	7.3e+35	0.39	0.75	1.27	G98
L10-038	G98-30	01:33:30.21	+30:47:43.8	51	C	3	?	2.6e−16	2.7e+36	0.25	0.31	1.58	MMT-BCS
L10-039	G98-31	01:33:31.25	+30:33:33.4	13	A	1	c	1.5e−14	9.6e+36	0.86	0.95	0.78	MMT-BCS
L10-040	G98-32	01:33:31.34	+30:42:18.3	55	A	2	n	2.4e−16	2.8e+36	0.26	0.65	1.61	S93
L10-041	(Opt)	01:33:31.80	+30:31:01.1	97	A	1	n	4.5e−17	1.7e+36	...	...	...	...
L10-042	G98-33	01:33:35.14	+30:23:07.5	43	A	1	n	1.3e−16	9.6e+35	...	...	...	...
L10-043	(X-ray)	01:33:35.39	+30:42:32.1	85	C	3	n	1.6e−16	4.6e+36	...	...	...	...
L10-044	G98-34	01:33:35.61	+30:49:23.0	29	A	1	n	3.3e−16	1.1e+36	0.37	1.01	1.43	hectoa
L10-045	G98-35	01:33:35.90	+30:36:27.4	30	A'	2	c	5.8e−15	2.0e+37	0.57	0.82	1.10	MMT-BCS
L10-046	G98-36	01:33:37.09	+30:32:53.5	39	A	1	n	2.2e−16	1.3e+36	0.48	1.01	1.39	MMT-BCS
L10-047	G98-37	01:33:37.75	+30:40:09.2	50	A	1	n	1.1e−16	1.1e+36	0.62	1.11	1.43	S93
L10-048	G98-38	01:33:38.01	+30:42:18.2	21	A	2	n	5.3e−16	9.7e+35	0.40	0.66	1.40	S93
L10-049	G98-39	01:33:40.66	+30:39:40.8	43	A	2	n	1.9e−16	1.4e+36	...	...	...	...
L10-050	G98-41	01:33:40.73	+30:42:35.7	71	C	2	n	2.8e−16	5.6e+36	0.44	0.61	1.44	S93
L10-051	G98-40	01:33:40.87	+30:52:13.7	60	A	2	y	1.0e−16	1.5e+36	0.31	0.69	1.37	G98
L10-052	G98-42	01:33:41.30	+30:32:28.4	33	C	2	y	2.0e−16	8.9e+35	...	...	...	...
L10-053	G98-43A	01:33:41.71	+30:21:04.1	34	A	2	y	1.1e−15	5.1e+36	0.23	0.44	1.47	G98
L10-054	G98-43B	01:33:42.24	+30:20:57.8	45	A	2	n	4.1e−16	3.3e+36	...	...	...	...
L10-055	G98-44	01:33:42.91	+30:41:49.5	44	C	2	y	1.6e−16	1.2e+36	...	...	...	...
L10-056	G98-45	01:33:43.49	+30:41:03.8	23	B	2	y	3.8e−16	8.3e+35	0.54	0.89	1.45	G98
L10-057	G98-46	01:33:43.70	+30:36:11.5	36	C	3	n	2.4e−16	1.2e+36	0.37	0.54	1.40	S93
L10-058	(Opt)	01:33:45.26	+30:32:20.1	67	A	2	n	4.0e−16	7.2e+36	...	...	...	...
L10-059	(Opt)	01:33:47.46	+30:39:44.7	42	A	1	n	1.7e−16	1.2e+36	...	...	...	...
L10-060	G98-48	01:33:48.35	+30:39:28.4	14	C	2	?	1.4e−15	1.0e+36	0.51	0.74	1.49	G98
L10-061	G98-47	01:33:48.50	+30:33:07.9	60	B	3	y	1.1e−15	1.6e+37	0.40	0.74	1.43	S93
L10-062	(Opt)	01:33:49.75	+30:30:49.7	73	A	1	n	8.9e−17	1.9e+36	...	...	...	...
L10-063	(Opt)	01:33:49.90	+30:30:16.7	54	A	1	n	4.8e−17	5.5e+35	...	...	...	...
L10-064	G98-49	01:33:50.12	+30:35:28.6	48	B	1	y	2.2e−16	2.0e+36	0.53	0.83	1.35	S93
L10-065	G98-50	01:33:51.06	+30:43:56.2	50	B	2	y	5.4e−16	5.4e+36	0.46	0.63	1.44	S93
L10-066	G98-52	01:33:51.67	+30:30:59.6	59	A	1	y	9.0e−17	1.3e+36	0.43	1.65	1.21	G98
L10-067	G98-51	01:33:51.71	+30:30:43.4	45	C	1	n	1.2e−16	9.5e+35	...	...	...	...
L10-068	(Opt)	01:33:52.15	+30:56:33.4	109	A	2	n	5.6e−17	2.7e+36	...	...	...	...
L10-069	G98-53	01:33:54.28	+30:33:47.9	47	A	1	y	3.6e−16	3.2e+36	0.36	0.82	1.40	S93
L10-070	G98-54	01:33:54.51	+30:45:18.7	21	B	3	c	2.2e−15	3.7e+36	0.47	0.83	1.41	G98
L10-071	G98-55	01:33:54.91	+30:33:11.0	20	A	2	c	5.0e−15	8.1e+36	0.48	0.83	1.12	S93
L10-072	(Opt)	01:33:55.01	+30:39:57.3	32	B	1	n	1.1e−16	4.5e+35	...	...	...	...
L10-073	(Opt)	01:33:56.49	+30:21:27.0	57	A	1	n	1.7e−16	2.2e+36	...	...	...	...
L10-074	G98-57A	01:33:56.97	+30:34:58.7	31	B	2	y	3.2e−16	1.2e+36	0.35	0.67	1.39	MMT-BCS
L10-075	G98-56	01:33:57.13	+30:40:48.5	44	A	1	?	1.3e−16	1.0e+36	0.70	1.16	1.46	G98
L10-076	G98-57B	01:33:57.13	+30:35:06.1	21	B	2	n	7.5e−16	1.3e+36	...	...	...	...
L10-077	G98-59	01:33:58.06	+30:32:09.6	24	C	3	?	8.6e−16	1.9e+36	0.30	0.40	1.45	S93
L10-078	G98-58	01:33:58.07	+30:37:54.6	16	A	1	n	1.0e−15	1.1e+36	0.69	1.19	1.35	hecto
L10-079	(X-ray)	01:33:58.15	+30:48:36.4	58	A	2	n	3.2e−16	4.3e+36	...	...	...	...
L10-080	G98-60	01:33:58.42	+30:36:24.3	8	A	1	n	1.2e−15	2.9e+35	0.64	0.96	1.22	hecto
L10-081	G98-62	01:33:58.51	+30:33:32.3	34	C	3	c	3.6e−16	1.7e+36	0.44	0.73	1.44	S93
L10-082	G98-61	01:33:58.52	+30:51:54.3	56	A	1	y	1.2e−16	1.5e+36	0.52	1.07	1.42	G98
L10-083	G98-65	01:33:59.93	+30:34:21.2	31	B	3	n	1.4e−15	5.2e+36	0.32	0.63	1.64	G98
L10-084	G98-64	01:34:00.31	+30:42:19.4	32	A	1	c	4.4e−16	1.9e+36	0.81	1.23	1.39	hecto
L10-085	G98-63	01:34:00.32	+30:47:24.1	43	B	1	n	1.7e−16	1.2e+36	...	...	...	...
L10-086	FL236	01:34:00.60	+30:49:04.2	13	C	1	n	1.4e−16	9.5e+34	0.51	0.80	1.03	hectoa
L10-087	G98-66	01:34:01.34	+30:35:20.2	48	A	2	n	1.2e−16	1.1e+36	0.41	0.90	1.38	MMT-BCSa,b
L10-088	G98-67	01:34:02.24	+30:31:06.8	60	A	1	y	1.2e−16	1.7e+36	0.43	0.96	1.49	S93
L10-089	(X-ray)	01:34:03.31	+30:36:22.9	92	A	2	n	3.7e−16	1.3e+37	...	...	...	...
L10-090	G98-68	01:34:03.48	+30:44:43.8	42	A	1	y	1.6e−16	1.1e+36	0.61	1.03	1.37	S93
L10-091	G98-69	01:34:04.26	+30:32:57.1	36	B	1	y	4.3e−17	2.3e+35	...	...	...	...
L10-092	G98-70	01:34:07.23	+30:36:22.0	101	A	1	n	1.9e−16	7.7e+36	0.38	0.71	1.35	S93
L10-093	FL261	01:34:07.50	+30:37:08.0	20	B	1	n	1.4e−16	2.1e+35	0.57	0.95	1.32	hectoa,b
L10-094	XMM244	01:34:08.37	+30:46:33.2	20	C	2	n	3.3e−16	5.1e+35	0.49	0.75	1.21	MMT-BCSa,d
L10-095	G98-71	01:34:10.02	+30:47:14.9	23	A	1	n	1.9e−16	4.0e+35	0.55	0.83	1.26	MMT-BCSb
L10-096	G98-73	01:34:10.70	+30:42:24.0	18	A	1	c	3.7e−15	5.0e+36	0.59	1.25	1.15	S93
L10-097	G98-72	01:34:11.04	+30:38:59.9	14	B	2	n	3.1e−16	2.5e+35	0.65	1.15	1.27	G98
L10-098	G98-74	01:34:12.69	+30:35:12.0	67	A	2	n	2.2e−16	3.8e+36	0.36	0.46	1.95	S93
L10-099	G98-75	01:34:13.02	+30:48:36.1	51	C	3	n	3.0e−16	3.1e+36	...	...	...	...
L10-100	G98-76	01:34:13.65	+30:43:27.0	26	A	1	n	1.6e−16	4.2e+35	0.51	0.87	1.25	S93
L10-101	G98-77	01:34:13.71	+30:48:17.5	58	B	2	n	2.4e−16	3.2e+36	...	...	...	...
L10-102	G98-80	01:34:14.10	+30:34:30.9	39	B	3	n	8.5e−16	5.1e+36	0.26	0.46	1.42	S93
L10-103	G98-79	01:34:14.35	+30:41:53.6	48	A	1	n	1.1e−16	1.0e+36	0.61	1.08	1.34	S93
L10-104	G98-81	01:34:14.38	+30:39:41.6	39	A	1	n	1.7e−16	1.0e+36	0.52	1.05	1.50	S93
L10-105	G98-78	01:34:14.41	+30:53:51.9	50	A	1	n	2.8e−16	2.8e+36	0.39	1.02	1.40	G98
L10-106	(Opt)	01:34:14.67	+30:31:50.9	66	A	1	n	4.9e−17	8.6e+35	...	...	...	...
L10-107	G98-82	01:34:15.57	+30:32:59.9	33	A	2	y	2.1e−16	9.3e+35	0.44	0.74	1.50	S93
L10-108	G98-84	01:34:16.31	+30:52:32.7	77	A	2	y	1.5e−16	3.6e+36	0.23	0.80	1.59	MMT-BCSa,b
L10-109	G98-83	01:34:17.00	+30:51:47.1	25	B	3	?	2.0e−15	5.2e+36	0.25	0.28	1.36	MMT-BCSc
L10-110	(Opt)	01:34:17.03	+30:33:57.7	54	B	3	n	4.4e−16	5.1e+36	...	...	...	...
L10-111	G98-85	01:34:17.61	+30:41:23.3	51	A	1	y	2.3e−16	2.4e+36	0.58	1.04	1.47	S93
L10-112	(Opt)	01:34:18.32	+30:54:05.8	84	A	1	n	4.9e−17	1.4e+36	...	...	...	...
L10-113	G98-86	01:34:19.28	+30:33:45.9	38	A	3	n	1.1e−15	6.4e+36	0.29	0.62	1.42	S93
L10-114	G98-87	01:34:19.87	+30:33:56.0	22	A	2	n	1.6e−15	3.1e+36	0.33	0.74	1.43	S93
L10-115	G98-88	01:34:23.23	+30:25:24.9	47	A	1	n	9.5e−17	8.4e+35	...	...	...	...
L10-116	XMM270	01:34:23.27	+30:54:23.9	42	A	1	n	1.4e−17	9.8e+34	0.52	0.86	1.48	MMT-BCSa
L10-117	G98-89	01:34:25.09	+30:54:58.1	66	A	2	n	3.5e−16	6.0e+36	0.31	0.78	1.45	MMT-BCSa,b
L10-118	G98-90	01:34:25.41	+30:48:30.9	53	A	1	n	8.0e−17	8.9e+35	0.44	1.06	1.70	G98
L10-119	FL312	01:34:25.87	+30:33:16.8	33	A	1	n	5.4e−17	2.4e+35	...	...	...	...
L10-120	G98-91	01:34:29.61	+30:41:33.4	43	C	2	?	1.6e−16	1.2e+36	0.60	0.82	1.37	G98
L10-121	G98-92	01:34:30.29	+30:35:44.8	54	A	1	y	1.3e−16	1.5e+36	0.36	1.04	1.39	G98
L10-122	(Opt)	01:34:31.85	+30:56:41.5	111	A	1	n	1.4e−16	6.9e+36	...	...	...	...
L10-123	G98-93	01:34:32.41	+30:35:32.6	41	A	1	n	1.6e−16	1.0e+36	...	...	...	...
L10-124	G98-94	01:34:33.02	+30:46:39.2	11	A'	3	n	9.4e−15	4.5e+36	0.29	0.66	0.97	BK85
L10-125	(X-ray)	01:34:35.41	+30:52:12.7	39	B	1	n	8.6e−17	5.1e+35	...	...	...	...
L10-126	G98-95	01:34:36.22	+30:36:23.6	46	A	1	n	1.2e−16	1.1e+36	...	...	...	...
L10-127	G98-96	01:34:39.00	+30:37:59.8	86	A	1	n	1.5e−16	4.3e+36	...	...	...	...
L10-128	G98-97B	01:34:40.73	+30:43:36.4	22	A	2	n	1.3e−15	2.6e+36	0.33	0.77	1.43	hectoa
L10-129	G98-97A	01:34:41.10	+30:43:28.3	39	A	2	y	1.2e−15	7.2e+36	0.41	1.07	1.36	BK85
L10-130	(Opt)	01:34:41.23	+30:43:55.4	35	A	1	n	7.6e−17	3.7e+35	...	...	...	...
L10-131	(Opt)	01:34:41.89	+30:37:35.2	156	A	1	n	8.7e−17	8.4e+36	...	...	...	...
L10-132	G98-98	01:34:44.62	+30:42:38.8	55	C	1	n	2.4e−16	2.9e+36	...	...	...	...
L10-133	(Opt)	01:34:54.88	+30:41:17.0	75	A	1	n	3.0e−16	6.9e+36	...	...	...	...
L10-134	(Opt)	01:34:56.44	+30:36:23.2	58	A	1	n	5.0e−17	6.8e+35	...	...	...	...
L10-135	(Opt)	01:35:00.36	+30:40:05.0	65	A	2	n	1.3e−16	2.1e+36	...	...	...	...
L10-136	(Opt)	01:35:01.22	+30:38:17.1	128	A	1	n	8.1e−17	5.3e+36	...	...	...	...
L10-137	(Opt)	01:35:03.18	+30:37:09.6	127	A	1	n	7.4e−17	4.7e+36	...	...	...	...

Notes. ^*Other names or for new objects whether the object was initially suspected to be SNR based on optical or X-ray data. ^**The letter "y" implies reported by Gordon et al. (1999) as a radio-detected SNR; "c" detected by Gordon et al. and confirmed by our new radio observations; "?" claimed as a radio-detected SNR by Gordon et al. but seen in our higher-resolution data as radio emission arising from a nearby H ii region or point source, not the SNR; "'n" not yet detected at radio wavelengths. As discussed in the text, because our radio observations were designed for studies of small-diameter SNRs, we did not expect to detect all of the SNRs detected by Gordon et al. (1999). ^†Hα surface brightness in units of erg cm⁻² s⁻¹ arcsec⁻²; both surface brightness and L_Hα include some contribution from [N ii] λλ6548, 6583. ^aFirst spectroscopic confirmation that [S ii]:Hα ratio supports a classification as an SNR. ^b[S ii]:Hα varies for two spectra of separate sections; average shown. ^cG98-83 has a low S ii:Hα ratio for an SNR and this makes its identification as an SNR suspect. ^d[S ii]:Hα varies for two spectra of separate sections; average shown.

Download table as:  ASCIITypeset images: 1 2 3

The results of our search for X-ray-emitting SNRs in M33 are presented in Table 4. Of the 137 objects in our list of SNR candidates, 131 were in regions observed as part of the ChASeM33 survey of M33, and 82 (58) were detected at significance levels of 2σ (3σ) or greater.²⁰ A detailed discussion of the brightest objects appears in Section 8. X-ray and optical images and some background information about the remainder of the detected sources are presented in the Appendix. Of the 95 SNRs in our revised accounting of the Gordon et al. (1998) list which fall within the region observed by Chandra, 61 (45) were detected at 2σ (3σ) or greater. These numbers include G98-57A, GKL98-57B, GK98-97A and GKL98-97B, all of which were detected at more than 3σ.²¹ Of the 23 new objects in our list of SNR candidates selected from optical interference-filter imagery, only 8 (2) were detected in X-rays. That a smaller fraction of these new objects were detected in X-rays compared to the Gordon et al. (1998) sample is unsurprising given that they are typically larger and lower (optical) surface-brightness objects; indeed, some may not be SNRs (see the Appendix). As expected, all of the 13 objects that initially attracted our attention because of their X-ray properties were detected in the final AE analysis.

Luminosities (and upper limits) for the objects in our list of SNRs and SNR candidates are presented in Figure 3 as a function of the diameters determined from our inspection of the optical and X-ray images. To convert from count rates to luminosities, we assumed a spectrum for a thermal plasma with 0.5 solar abundances, a temperature kT of 0.6 keV and an absorbing column of 1 × 10²¹ cm⁻². The observed luminosities for detected SNRs vary by a factor of 500, from 2.4 × 10³⁴ erg s^-1 to 1.2 × 10³⁷ erg s^-1 in the 0.35–2.0 keV band. The brightest SNRs have diameters of 20–30 pc. Smaller-diameter SNRs tend to be fainter, as do larger ones. However, even at relatively small diameters of 20–40 pc, 10 optically identified SNRs were not detected in X-rays. Some SNRs were detected in X-rays throughout the size range from the smallest up to ∼100 pc, though many of the larger ones are quite faint.

Zoom In Zoom Out Reset image size

Four objects with diameters of 80 pc or greater were detected. Two of these—L10-043 (85 pc) and L10-089 (92 pc)—were identified initially because of their X-ray properties, but two others—G98-70 (101 pc) and L10-122 (111 pc)—were selected based solely on their optical properties. Their detection as extended sources supports their identification as SNRs although, if they are isolated SNRs, they would be among the largest such objects known. An alternative possibility that these are superbubbles in which SNe have exploded is discussed in Section 7.3.

We first consider whether any of our X-ray detections represent chance coincidences. The new point-source catalog will contain about 670 entries, of which 7% (46) are SNRs. The Chandra survey area covers about 1000 arcmin². The summed areas of the region files for SNRs and candidates with diameters less than 60 pc is 2 arcmin²; for candidates with diameters greater than 60 pc the total area is about 5 arcmin². Assuming the sources are randomly distributed on the sky, we thus expect about one chance coincidence for SNRs smaller than 60 pc and about three for larger diameter objects. We do find one point source located within the boundary of one of the SNRs, L10-137; the effect of this point source was removed in calculating the values reported in Table 4. We conclude that possible point-source contamination to the detections we report is very small.

Figure 4 displays the distribution of objects as a function of diameter and galactocentric radius. The fraction of X-ray detections is roughly constant, apart from the effect that can be attributed to a larger number of large-diameter objects at larger galactocentric radii.

Zoom In Zoom Out Reset image size

The distribution of SNR diameter is shown in Figure 5 which compares the distribution for all observed SNR candidates (in black), those detected at >2σ in blue and those detected at >3σ in red. While there are some small (<30 pc) diameter objects that are not detected, the fraction of detections is high below 40 pc and gradually falls at larger diameters. The drop-off in detection frequency is more rapid for the 3σ limit than at 2σ as one would expect if large-diameter objects were typically fainter. A substantial fraction of the objects detected between 2σ and 3σ have diameters in the 40–80 pc range. The median diameter is 38 (33) pc for objects detected at 2σ (3σ), compared to the median diameter of 44 pc for the entire sample. The median diameter for objects detected with L_X > 10³⁵ erg s⁻¹ is 29 pc; for 10³⁶ erg s⁻¹ it is 19 pc.

Zoom In Zoom Out Reset image size

Although there are a few SNRs bright enough for detailed spectral analysis (see Section 8), most are faint. Nonetheless, it is clear that the spectra of the SNRs are soft compared to other M33 (and background) source populations. This is illustrated in Figure 6, which shows histograms of the numbers of SNRs and point sources detected in M33 with Chandra as a function of hardness ratio. Here, we have defined the hardness ratio, M − S/Total, using the numbers of counts in the 0.35–1.2 keV (S), 1.2–2.6 keV (M), and 0.35–8 keV (Total) bands. With this choice of bands, the SNRs occupy a region centered around −0.5, whereas the point sources are centered near +0.5. The width of the SNR distribution is broad compared to that of the point-source distribution, largely a consequence of the high fraction of the SNRs that are faint, rather than an indication of real spectral variations among the SNRs. There is a tail of soft sources in the point-source distribution; as discussed in Section 3, we inspected the positions of all of these sources in optical images and find no evidence that they are SNRs. A number of them can be identified with foreground stars. The clear separation between the majority of point sources and the SNRs supports our assertion that we are including few if any chance associations in our list of detected SNRs.

Zoom In Zoom Out Reset image size

We performed tests to see if there were trends in hardness ratios with the X-ray count rates and/or SNR diameters. We tested the hardness ratios defined above, as well as hardness ratios based on the bands 0.35–1 keV and 1–2 keV. One might expect such trends if larger diameter and/or fainter SNRs had lower shock velocities, but none was seen, presumably because of the limited signal-to-noise ratio of the spectra and limited spectral resolution of ACIS-I. There was a general narrowing of the distribution as the number of counts increases, as one would expect from counting statistics.

We have detected many more SNRs than the 26 reported by Plucinsky et al. (2008) in the First Look survey. There are several reasons for this: (1) we have tailored the extraction regions for the actual sizes of the SNRs; (2) we have selected an energy bandpass of 0.35–2 keV optimized for the relatively soft spectra of SNRs; and (3) we are using the full ChASeM33 data set, whereas the First Look analysis was carried out only on data obtained during the first year of observations. Of the 26 sources reported in the First Look survey and associated with SNRs, there are two—G98-42 and G98-83, identified with FL174 and FL286—that do not appear in our list of detections. In the First Look survey, sources were associated with SNRs by positional coincidence only using a match radius of 10''. In these two cases, our refined analysis shows that there were point sources near but outside the region defined by the optical SNR. As noted earlier, Ghavamian et al. (2005) analyzed a different set of Chandra data including some data from the S-array but covering only part of M33; they detected 23 SNRs including 22 from the list of Gordon et al. (1998), five of which were not detected in the First Look analysis. The Ghavamian et al. (2005) study was specifically focused on recovering SNRs from the data and used energy bands and extraction regions optimized for SNRs. Our study confirms that all but one of the SNRs reported by Ghavamian et al. (2005) are X-ray emitters. The one SNR that was not detected by us is G98-83; with 5.1 ± 5.3 net counts, it falls below our 2σ detection threshold.

In order to better characterize the sample, we have gathered all of the available information on the objects in our list of SNRs and SNR candidates, using data that were already available and, when necessary, acquiring new optical and radio observations. We describe here the results of these efforts.

5.1. Optical Flux Measurements

As discussed in Section 3.1, we used interference-filter imagery from Massey et al.'s (2006) LGGS data and our own imagery from the Burrell Schmidt telescope to search for new SNRs in M33. We used these same data to determine the apparent sizes of SNRs and to measure Hα and [S ii] fluxes from the objects in our sample. We examined each of the SNRs and candidates and created elliptical region files, varying the major and minor axes and the position angle to match the extent of the Hα and [S ii] emission. For full or partial shells where the size could be estimated with little ambiguity, we chose regions to match. For the minority of objects with amorphous shapes or where only an individual filament is visible, we simply chose an ellipse that most nearly encloses the visible emission. For all but a handful of objects, we used identical regions for both optical and X-ray flux extractions, the exceptions being objects that were clearly more extended in one band or the other.

To measure the optical flux, we used the imcnts task (part of the IRAF external package xray) to extract the flux within the elliptical regions defined above from the flux-calibrated,²² continuum-subtracted images. In a few instances where a nearby bright star had been poorly subtracted, we excluded a small region around the star from the extraction region. We measured the flux using data from both the LGGS and Schmidt surveys, and in the vast majority of objects obtained results that agreed within ±30%. (Instances of disagreement were mostly very small objects where the lower resolution of the Schmidt survey caused much of the flux to spill outside the extraction region.)

The results are given in Table 3, where the "diameter" represents the geometric mean between major and minor axes, converted to pc assuming a distance to M33 of 817 kpc. The surface brightness is the actual mean value within the elliptical region (as reported by imcnts), and the luminosity L_Hα is determined simply by scaling to 817 kpc, with no absorption; both of the latter quantities were measured from the LGGS. While the actual luminosities will be slightly higher once absorption is taken into account, the absorption is approximately uniform across the face of M33, so the relative values should change little. Also, it is important to note that the LGGS bandpass of 81 Å passes [N ii] λλ6548, 6583, so the "Hα" surface brightness and luminosity values include most of the [N ii] emission as well (the relative intensity of which will vary systematically relative to Hα with galactocentric radius).

The Hα luminosities for the SNRs range from 1 × 10³⁵ to 3 × 10³⁷ erg s^-1, with a median value of 1.9 × 10³⁶ erg s^-1, while the Hα surface brightness ranges from 1.4 × 10⁻¹⁷ to 1.5 × 10⁻¹⁴ erg cm⁻² s⁻¹arcsec⁻² with a median value of 2.2 × 10⁻¹⁶ erg cm⁻² s⁻¹arcsec⁻². As noted earlier the optical searches for SNRs in M33 are largely surface-brightness limited, as is clear from Figure 7, which shows the Hα luminosities of objects in our list as a function of diameter. At smaller diameters, SNRs and candidates have a large range of luminosities, while at large diameters there are no low-luminosity objects since these objects would have such faint average surface brightnesses that they would fall below our detection threshold.

Zoom In Zoom Out Reset image size

5.2. Optical Spectroscopy

5.3. High-resolution Radio Imaging

The most detailed survey of SNRs in M33 at radio wavelengths was carried out by Gordon et al. (1999). Based on a series of 6 and 20 cm observations obtained at the VLA and the WSRT and originally discussed by Duric et al. (1993), they constructed a catalog of 186 sources along the line of sight to M33. They identified 53 of these with objects previously suggested as SNRs based on optical imagery and follow-up spectroscopy. Their 7'' beam was sufficient to resolve the larger remnants (>30 pc), but not the smaller ones; in addition, it left their study vulnerable to source confusion, especially in crowded regions near H ii complexes. They did not propose any new SNRs based solely on their radio data.

In order to complement our high-resolution X-ray images, and to search for small-diameter, young remnants and PWNe, we undertook new radio observations of M33 using the VLA in its A configuration. Data were taken at both 6 cm and 20 cm in bandwidth-synthesis mode to minimize bandwidth smearing effects at large off-axis angles. At 6 cm we used a mosaic of eight fields to cover 621 arcmin², roughly co-extensive with the radio coverage of Gordon et al. (1999) and somewhat less than observed with Chandra; a single pointing at 20 cm covering 721 arcmin² was also obtained. These images have resolutions of ∼05 and ∼2'' at 6 and 20 cm, respectively; the typical map noise levels were 15–25 μJy.

Of the 53 objects identified as radio-detected SNRs by Gordon et al. (1999), 50 lie in the region covered by our new radio images. Of these, 23 are detected, with 21 having optical diameters smaller than 10''. This bias toward small-diameter objects results from the fact that our high-resolution data are not sensitive to angular scales that are more than a few times the diameter of the synthesized beam (∼2'' at 20 cm). For another 19 of their detections, some of which were rather weak, we record no emission, although all of these remnants are larger than 8''; the one additional small-diameter source from Gordon et al. (1999) is in a noisy part of our image, and no meaningful limit can be obtained from our data. For the seven remaining sources, we detect radio emission within <10'' of the optical SNR position, but not from the SNR itself; in four cases the emission is coincident with a nearby H ii region, and in the other three it is consistent with one or more point sources (most likely background objects). These objects are flagged in Table 3.

We also cross-correlated our radio catalog with the complete X-ray source list in order to search for new SNR candidates, particularly small-diameter young PWNe analogous to the Crab Nebula which might have been missed on the grounds of X-ray extent (too small), X-ray spectrum (too hard), or optical data (confused in H ii regions, lacking high [S ii]:Hα ratios, etc.). Using a matching radius of 3'' to account for extended sources in both wavelength regimes as well as the larger positional uncertainties of off-axis X-ray detections, we find 34 coincidences. Eleven of these are known SNRs. The majority of the remainder are point-like X-ray sources coincident with point-like radio sources (or, in one case, with a classical double radio source) and are likely to be background active galactic nuclei. Nine objects, however, show some evidence of radio extent, although most are weak (∼200 μJy) radio sources and are only marginally larger than the beam. We examined the optical emission at the location of each of these sources and found nothing to suggest any of them are normal shell-type remnants. Better radio observations are required before any of these can be declared bona fide radio SNRs.

One source, however, is worthy of further consideration. FL281 is distinguished by its hard X-ray spectrum approximated by a power law with a photon spectral index of 2.1.²³ It is coincident with a faint Hα emission region. The radio source is just resolved at 6 cm, with an FWHM of 05 ± 01 or ∼2 pc; it also has a flat radio spectrum, with α_radio = 0.2. The X-ray and radio characteristics match those of a young PWN. The field of the candidate object is shown in Figure 8. We discuss this object further in Section 9.

Zoom In Zoom Out Reset image size

An obvious question is whether the X-ray properties of SNRs correlate with their observed properties at other wavelengths. Such comparisons are largely not possible with Galactic samples because of the effects of absorption in the Galactic plane and because there are significant uncertainties in the distances to most Galactic SNRs. However, all the SNRs in M33 are at essentially the same distance, and the line-of-sight absorption to M33 is quite low. There are various effects that limit the uniformity and utility of our sample, including spatial differences in X-ray exposure, confusion in crowded regions in the optical, somewhat limited coverage beyond the D₂₅ optical elliptical isophot (diameter ∼ 32'), and poor sampling of low spatial frequencies in our radio observations. Nonetheless, the data we have on M33 are likely the most homogenous data set that exists for SNRs today in the sense that effectively the entire galaxy has been covered to uniform depth in each wavelength band and significant numbers of SNRs have been detected in each band. If correlations do exist then one would expect them to appear in the M33 sample. We have carried out a number of such comparisons and tests to see whether the optical or radio properties of SNRs are good predictors of X-ray emission and vice versa. In all cases, we have used both the Gordon et al. (1998) list of SNRs and our expanded list. The Gordon et al. list has the advantage that it was constructed without knowledge of the X-ray properties, while our new list is somewhat larger and presumably more complete, although it contains some objects added as a consequence of their X-ray selection.

A comparison of L_X and Hα surface brightness is presented in Figure 9. Most of the SNRs in our sample were identified optically, and optical searches in M33 (and other galaxies of comparable distance) are largely surface-brightness-limited, except for smallest diameter objects. On the other hand, except for the largest (>100 pc diameter) objects, X-ray detections of SNRs in the ChASeM33 survey are primarily limited by flux, because the background count rates are so low. In some sense, then the comparison of L_X to Hα surface brightness is the cleanest from an observational perspective. Figure 9 shows that the most luminous X-ray objects tend to be the SNRs with the highest Hα surface brightness. These are the objects that are interacting with dense media and hence are the ones evolving on the shortest timescales. The median (average) Hα surface brightness is 2.8(15) × 10⁻¹⁶ erg cm⁻² arcsec⁻² for SNRs with X-ray luminosities brighter than 10³⁵ erg s⁻¹ and 1.6(2.8) × 10⁻¹⁶ erg cm⁻² arcsec⁻² for SNRs fainter than 10³⁵ erg s⁻¹. Brighter X-ray SNRs tend to have higher Hα surface brightness, but the scatter is quite large.

Zoom In Zoom Out Reset image size

The corresponding comparison of L_X to Hα luminosity is presented in Figure 10. Again, the most luminous X-ray SNRs tend to be the SNRs with the highest Hα luminosities. The X-ray luminosities of the SNRs in M33 are nearly always less than the optical luminosities as measured by Hα, and in many cases much less. It is worth remembering in this regard that most of the radiant energy from the SNRs is likely to be in lines in the UV or IR. The large amount of variation in the X-ray-to-optical luminosity ratio is not surprising. The optical lines originate in relatively dense recombining gas with temperatures of 10,000–20,000 K and short cooling timescales, while the X-ray emission originates in tenuous material with temperatures of a million degrees or more and relatively long cooling timescales. In Galactic SNRs, where the structure of an SNR is clearly resolvable at both X-ray and optical wavelengths, there is seldom a good detailed correlation between the X-ray and optical emission.

Zoom In Zoom Out Reset image size

The [S ii] λλ6717, 6731 ratio is density sensitive. Since one anticipates rough pressure equilibrium between the X-ray gas and the gas in the recombination zone where [S ii] is generated, one might anticipate a relatively strong correlation between the [S ii] ratio and X-ray emission, either with L_x, or possibly with $\sqrt{L_x/r^3}$. The available [S ii] line ratios are plotted against remnant diameter in the left panel of Figure 11. Smaller diameter objects are more likely to show departures from the low-density limit than are large-diameter ones. High density is a good predictor of X-ray detectability; of the 17 objects with [S ii] line ratios measured to be less than 1.3 (corresponding to a density higher than about 140 cm³; Cai & Pradhan 1993), 16 were detected in our survey. Furthermore, as shown in the right panel, in most cases the higher the density in the [S ii] zone the higher the X-ray luminosity. The median (average) line ratio for objects with L_x > 10³⁵ erg s⁻¹ is 1.31 (1.25) whereas for objects with lower values (including the objects that were not detected), the ratio is 1.42 (1.39). Once again, however, there are exceptions: G98-21, the highest luminosity object in our sample, has an [S ii] λλ6717, 6731 ratio of 1.39, not far from the low-density limit.

Zoom In Zoom Out Reset image size

While the radio survey of Gordon et al. (1999) appears to have some deficiencies associated with lack of sufficient spatial resolution as described in Section 5.3, it remains the most complete survey of SNRs in M33 at radio wavelengths. The faintest sources they detected have flux densities of 0.2 mJy, low enough to have detected some, but not the majority of, Galactic SNRs at the distance of M33. For example, RCW86, Puppis A, PKS1209-52, and all of the historical SNe except for SN1006 would have been detected, but the Cygnus Loop would have been missed. Of the 53 objects they identified as radio SNRs based on positional coincidence with an optically identified SNRs, all but three—G98-01, G98-03, and G98-43—are in the regions covered by our X-ray survey; another seven are misidentifications arising from source confusion as noted in Section 5.3, none of which are X-ray detected. Of the (43) real radio-optical matches in our survey region, 29 (26) or 67% (60%) were detected at 2σ (3σ) in our X-ray survey. This compares to a 64% (47%) detection rate of X-ray counterparts by Chandra of the sources in the sample of optically identified SNRs which was available to Gordon et al. (1999). One might expect that these radio SNRs would have been preferentially detected in our Chandra X-ray survey, but the effect is fairly small. There are 52 objects in our sample derived from the Gordon et al. (1998) list which are not known (now) to be radio sources²⁴ and consequently whose identification as an SNR excluding the ChASeM33 observations is based solely on their optical properties. Of these, 32 (19) or 62% (37%) were detected in X-rays.

Moreover, the radio flux as measured by Gordon et al. (1999) is not a good predictor of X-ray detectability; the median radio flux for the detected X-ray SNRs (0.6 mJy) is actually lower than the median for the sources that were not detected in X-rays (0.8 mJy). This somewhat surprising finding is due in part to problems associated with identification of the radio sources with SNRs, as was discussed in Section 5.3; G98-83, for example, is bright in Gordon et al. (1999), but really is a nearby H ii region based on our new radio data. On the other hand, X-ray brightness is a good predictor of radio detection; the brightest seven X-ray SNRs are all detected by Gordon et al. (1999). A deeper, radio survey with complete uv coverage and scaled-array spectral indices is needed to understand the relationship between the radio and X-ray properties of M33 SNRs.

Our overall conclusion from all of these comparisons of observables in different wavelength bands is that correlations are weak. It is generally the case that extreme objects are extreme at all wavelengths; if one considers average or median values for the brightest X-ray objects, these tend to have the highest optical surface brightness, the most significant departures from the low-density limit as measured by the [S ii] line ratio, and are the radio brightest. However, the range of variation is large. The inventory of X-ray gas and gas at 10,000 K where the optical emission arises is simply very different in different SNRs.

Of the 131 objects in our sample observed by Chandra, we have detected more than half (62%). At least for the 69 X-ray detected objects that were selected initially based on a large [S ii]:Hα ratio and the absence of obvious young star clusters, their detections provide strong support for their identification as SNRs. For the 13 X-ray-selected objects, X-ray detection is not supporting evidence, of course; that comes from the fact that these objects have optical characteristics expected from SNRs. On the other hand, 49 objects were not detected, including 34 Gordon et al. (1998) SNR candidates and 15 new candidates we identified from our search of the LGGS and Schmidt plates. The fraction of undetected objects is higher for objects with larger diameters, but some small-diameter objects were missed as well. What factor(s) determine X-ray detectability?

On an observational level, the answer to this question is simple: many of the objects were detected fairly near our sensitivity limit, and it is clear from Figure 3 that SNRs with similar diameters have widely differing luminosities; had our sensitivity limit been two or three times lower, it is quite likely that more of the objects would have been detected. In addition, while the survey was fairly uniform, objects were observed at varying off-axis angles, as shown in Table 4, and with differing exposure times; a few objects were likely missed because they fell below our sensitivity limit at their positions in the survey fields. A secondary concern, a possibility we return to in Section 7.3, is that a small fraction of the objects may have been misclassified as SNRs.

7.1. The Importance of Local Gas Density

7.2. Do Optical Morphology and Environment Influence X-ray Detectability?

7.3. Supernova Remnants to Superbubbles and Diffuse Ionized Gas

7.4. A Final Evaluation of our Sample

Based on the considerations discussed above, we have systematically reviewed the SNR candidate list in Table 3, using all available data to identify the subset of objects that had been admitted as candidate SNRs based on our fairly inclusive criteria, but that seem in some ways peculiar or unlikely to represent the remnants of single SN explosions. Our review has identified 26 of these objects in both categories outlined above. A moderately large shell or diffuse emission region with one or more stars located centrally within the shell at the very least raises suspicion that stellar winds (possibly in conjunction with one or more SNe) may be partially responsible for creating the observed nebula. A variation on this theme are the large arcs of enhanced [S ii]:Hα emission protruding from bright H ii regions which could be "blow-outs" involving stellar winds and/or potentially SNR shocks. Likewise, isolated filaments of enhanced [S ii]:Hα in diffuse regions of emission (without clearly associated shell-like morphology) could be examples of the "bright DIG" phenomenon. Table 6 provides a summary of this assessment along with an indication of the reason for questioning the veracity of the SNR identification. These 26 objects constitute 18% of the sample. Further details about all the objects, including those in Table 6, are presented in the Appendix.

Table 6. Possible Superbubbles and Peculiar SNR Candidates^a

Object	Alternate Name	Comments
L10-001	G98-01	Very large SNR or fossil superbubble with few stars remaining. Not in ChASeM33 fields
L10-003	G98-03	Likely superbubble. Not in ChASeM33 fields
L10-010	G98-08	Very large SNR or fossil superbubble with few stars remaining
L10-012	G98-19	Large shell with bright central star(s); likely superbubble
L10-026	(X-ray)	Likely stars associated and only slightly enhanced [S ii]; likely superbubble
L10-028	(Opt)	[S ii] enhanced, but stars seen in projection; likely superbubble
L10-030	G98-23	Associated stars likely; old SNR or superbubble both possible
L10-035	XMM156	Bright, extended X-ray emission but compact [O iii] optical source; peculiar SNR
L10-043	(X-ray)	Associated stars likely; complex region; old SNR or superbubble both possible
L10-050	G98-41	Long filament of enhanced [S ii] associated with larger region of emission
L10-055	G98-44	Long linear filament of enhanced [S ii] in much larger region of emission
L10-068	(Opt)	Possible blow-out on east side of H ii region, but projected stars not clearly associated
L10-080	G98-60	Single optical filament on edge of H ii region but X-rays associated with one side; peculiar SNR
L10-089	(X-ray)	Possible blow-out from bright H ii region to west; budding superbubble?
L10-092	G98-70	Cluster of centrally located stars; likely superbubble
L10-098	G98-74	Associated stars likely; old SNR or superbubble both possible
L10-108	G98-84	Enhanced [S ii] filaments in much larger region of emission; old SNR or fossil superbubble both possible
L10-117	G98-89	Possible associated stars; old SNR or superbubble both possible
L10-119	FL312	Possible but uncertain association of X-ray source with faint emission nebula seen in Hα
L10-120	G98-91	Modest sized linear filament in more diffuse gas. [S ii] enhanced, but no shell or associated structure
L10-122	(Opt)	Exceedingly large size (165 × 85 pc) makes a single SNR origin unlikely; no clearly associated stars
L10-131	(Opt)	Large size and star cluster associated; likely superbubble
L10-132	G98-98	Clumpy, faint Hα with marginally too low [S ii]:Hα; stars seen in projection; likely superbubble
L10-133	(Opt)	Large shell with central stars; also central bright, compact emission region; likely superbubble
L10-136	(Opt)	Large shell with marginally enhanced [S ii]:Hα; bright star(s) centrally located; likely superbubble
L10-137	(Opt)	Large shell with stars associated; likely superbubble

Note. ^aSee individual object descriptions in the Appendix for more details.

Download table as:  ASCIITypeset image

We emphasize that we found no reason to question the identification of the majority of objects in our SNR list. The bulk of the 137 objects in Table 3 are certainly SNRs; the fact that 82 of them were detected in X-rays strongly supports the SNR identification for these objects. It is also quite possible that some of the objects we have listed as questionable will ultimately turn out to be SNRs; several were detected in X-rays. Given current capabilities across the electromagnetic spectrum there simply is no way observationally to separate all of the borderline cases cleanly.

A Venn diagram providing a snapshot of the current status of detections of the M33 SNRs sample is shown in Figure 12. There are 29 objects that have been detected as optical emission nebulae with elevated [S ii] compared to Hα, as X-ray sources, and also as radio sources. The 8 X-ray only objects are associated with some optical emission, either Hα or [O iii], but do not appear to show elevated [S ii]. Similar diagrams are provided by Pannuti et al. (2007) for the galaxies discussed briefly in Section 9.2. All of these diagrams are strongly affected by the sensitivities of searches at various wavelengths. In particular, for M33, while Gordon et al. (1999) did detect a substantial number of the SNRs that had been identified optically at radio wavelengths, their survey sensitivity and angular resolution were not sufficient for them to identify a radio-only sample of SNRs. Indeed our higher angular resolution radio observations of M33 showed that seven of the 53 objects that they thought were coincident with optically identified SNRs were nearby H ii regions or background objects. Except for the Galaxy and the Magellanic Clouds, optical searches remain the most effective way to identify SNRs. The [S ii]:Hα ratio provides a fairly clean diagnostic and photons are relatively plentiful.

Zoom In Zoom Out Reset image size

The Chandra observations reported here represent about the best one can hope to do with the current generation of X-ray instrumentation. Of our total sample of 137 objects, 131 were within our survey region and 82 were detected at X-ray wavelengths. For eight of these, the primary reason for believing they are SNRs is that they are (1) extended and (2) soft X-ray sources.

The next major advance in the M33 SNR sample is likely to occur with the advent of the EVLA. This will be important not only for finding more PWNe, but also for finding and detecting additional normal shock-dominated remnants. It should provide both the angular resolution to construct radio images for all of the known SNRs and the sensitivity and polarimetric fidelity to find the remnants that have been missed in optical surveys; i.e., the PWNe and young remnants without the radiative shocks needed to produce [S ii]. At optical wavelengths, the concern as one goes fainter is confusion with DIG masquerading as a shock-heated material; at radio wavelengths, it is background galaxies that happen to lie along the line of sight toward Hα emission. Higher quality (or for about one-third of the objects, initial) optical spectroscopy would also be very helpful in clarifying which of the objects are bona fide SNRs and what their physical properties are. Finally, a new interference-filter optical imaging study of M33 would also be desirable, especially if the observations could be obtained with subarcsec seeing conditions; this would allow a more sensitive search for small-diameter objects and more accurate measurements of the optical sizes. The currently available data set at X-ray wavelengths represents close to the limit of what can be expected with the current generation of X-ray observatories, and it is thus unlikely that the X-ray sample in M33 will be increased significantly for some time to come.

There are seven SNRs in M33 that stand out in terms of their X-ray brightness—G98-21 (29 pc diameter), G98-28 (22 pc), G98-29 (40 pc), G98-31 (17 pc), G98-35 (34 pc), G98-55 (24 pc), and G98-73 (22 pc). These remnants are sufficiently bright to allow extraction and fitting of their spectra, and also comparison of their X-ray, optical, and radio morphologies. Interestingly, all of these are intermediate-sized objects (∼15–40 pc across). In addition, most show evidence of significant interaction with surrounding material (e.g., strongly enhanced X-ray emission at localized spots on their boundaries). There are sufficient counts in the X-ray spectra of these seven SNRs to allow a more detailed characterization of each remnant with more complex spectral models. In the following, we first discuss the comparison of the X-ray and optical images of each remnant and then discuss results of our X-ray spectral model fits.

8.1. Multiwavelength Spatial Analyses

In order to take advantage of Chandra's imaging quality, we performed 10 iterations of a Lucy–Richardson deconvolution on these objects using the IDL astron package max_likelihood. For each object, we selected a subset of pointings which were closest to on-axis, avoiding cases where the source fell on a detector chip gap. The data for each object were merged and filtered to include only the 0.35–4 keV band. For the deconvolution kernel, we constructed a merged PSF by merging ray traces performed at the appropriate off-axis angles, as determined using the CIAO tool dmcoords. The ray traces used SAOtrace²⁶ to trace the optics and MARX²⁷ to project onto the detector plane.

Figures 13 and 14 show four-panel X-ray/optical/radio comparisons for each of these SNRs. The leftmost panel in each row is a color composite made from continuum-subtracted Hα (red), [S ii] (green) image, and [O iii] (blue) LGGS images, with superposed radio contours from our VLA observations of M33. The remaining panels in each row are the deconvolved X-ray image, the raw X-ray image, and the PSF for the observations of the fields used. In addition, the corresponding atlas images of these objects in the Appendix show gray-scale renderings (left to right) of the X-ray data, Hα, [S ii], and V-band images.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

G98-21 (or L10-025) is the brightest X-ray SNR in M33, and a thorough study of this object has been published by Gaetz et al. (2007). This remnant is located on the outskirts of the giant H ii region NGC 592. As noted earlier, this object was suggested as an SNR by Long et al. (1981) using Einstein, and its identification as an SNR was firmly established when Gordon et al. (1993) showed it to be a fairly bright, non-thermal radio source. The Chandra images, analyzed in detail by Gaetz et al. (2007), showed the X-rays to be a nearly circular shell with a diameter of 20 pc (5'') that is brighter on the eastern side (i.e., toward the center of NGC 592).

The optical/X-ray comparison is shown in the top row of Figure 13 (see also Figure 23, third row). While a bright star or tight cluster is seen in projection on the eastern limb of the optical shell, the [S ii] image clearly shows the portion of the object dominated by radiative shocks. In this object, the morphology of the optical shell differs from that of the X-ray SNR, with an arc of optical emission on the south side extending beyond the western edge of the X-ray shell. The brightest portion of the X-ray ellipse actually lies somewhat east of the most conspicuous optical emission. This morphology of G98-21 is unusual; while it is common to see differences in the relative surface brightness of features at optical and X-ray wavelengths, rarely is the overall morphology so qualitatively different. If the bright eastern X-ray emission is due to enhanced density in the direction of NGC 592, then we would expect enhanced optical emission there as well, contrary to what Figure 13 shows. The [O iii] emission of the SNR (image not shown) follows a similar pattern to that seen in [S ii]. Gaetz et al. (2007) suggest that photoionization by NGC 592 prevents recombination in the eastern limb of the SNR.

G98-28 (or L10-036) is shown in the second row of Figure 13 (see also Figure 26, second row). This SNR is located away from any obvious local optical nebulosity, although it is about 1' away from the northwestern edge of the giant H ii region NGC 595. The optical object appears to be a bright partial shell, open to the east–southeast with an enhancement at its southern end. The X-ray image shows little evidence of a shell, but rather is centered on the very high surface brightness optical shell. The morphology of the object is nearly identical in Hα, [S ii], and [O iii]. We inspected images from the Spitzer archive of this region and these show that the SNR is also detected at 24 μm, indicating the likely presence of warm dust in the radiative shocks. Optical echelle spectra (Blair et al. 1988) show bulk motions of 345 km s^-1, confirming high velocities. The appearance of G98-28 in the V-band continuum image is likely due to the transmission of bright [O iii] emission lines through the broadband filter. The reason for the dominance of emission from the western side of the shell is not clear from these data.

G98-29 (or L10-037), shown in the third row of Figure 13 (see also Figure 26, third row), is a large limb-brightened X-ray SNR about 32 pc across. The original and deconvolved images are very similar in this case. The SNR was detected by Gordon et al. (1999) at radio wavelengths, but is not contained in our higher resolution radio maps, which is most likely due to the fact that our observations were less sensitive to objects of this size. Optically, the SNR is relatively isolated from other nebulosity. The optical emission has a knotty appearance, with only a hint of emission in the NW.

G98-31 (or L10-039), shown in the bottom row of Figure 13 (see also Figure 27, top row), is quite small at X-ray wavelengths and the emission appears centrally peaked, with only a hint of shell-like structure extending to the north. The X-ray emission corresponds to a bright core of optical emission seen in the figure and the high-surface-brightness component of this core is very similar. The [S ii] line ratio is 0.8 for this object, indicating high density. Optical echelle spectra (Blair et al. 1988) show velocities of 330 km s^-1 for this core. The core also appears strong in [O iii], and it may be this emission being passed by the V-band filter that causes the extended emission in this image. In a wider view, one can see that the object sits centrally in a larger Hα shell with brighter knots on the east and north sides.

At a stretch that reveals faint structures, the optical emission from G98-31 displays an unusual morphology, with what appear to be two thin arcs or shells extending from the core to the north and east. While somewhat reminiscent of the rings around SN1987A, the physical scale of the thin arcs in G98-31 is nearly 60 times larger, extending some 30 pc. Long-slit spectra confirm that these arcs show elevated [S ii]:Hα ratios, but they do not show the high velocities seen in the core. The arcs are present but less well defined in the [O iii] image (not shown). It is conceivable that the arcs may represent an older SNR located along the same line of sight.

G98-35 (or L10-045), shown in Figure 14, top (see also Figure 28, third row), is another complicated object with a bizarre morphology. A very bright central core of optical emission is surrounded by asymmetrical loops or arcs, one to the east–southeast and one to the northwest. The eastern lobe is high in [S ii] while the northwestern lobe has lower, but still elevated [S ii] emission. The [O iii] emission looks quite similar. Hubble Space Telescope (HST) observations with the original WF/PC camera by Blair & Davidsen (1993) showed the bright core to be a broken shell of very high surface brightness emission. This can be seen in the [S ii] panel of the figure, which is scaled to show the highest surface-brightness regions. Echelle data by Blair et al. (1988) show bulk motions of 313 km s⁻¹ for the core, but the velocities are significantly lower in the portions of the loops measured by their east–west oriented slit.

Most of the X-ray emission stems from the very bright optical core. However, there is clearly emission extending beyond the core—especially toward the eastern lobe where faint X-ray emission peaks just inside the optical rim. Without more information, we define the SNR to include the core and both lobes, which appear to form a coherent structure as seen in the Hα panel of the figure, but clearly the core itself is kinematically distinct.

G98-55 (or L10-071), shown in the middle panel of Figure 14 (see also Figure 34, bottom row), is another very high surface brightness optical SNR on the eastern edge of an H ii complex with bright condensations and extensive diffuse emission. In [S ii] and [O iii], the SNR is distinct from the H ii region, and is dramatically brighter on its western side, toward the bulk of the H ii region. The southern edge of a shell is also visible, and the complete shell can be seen with altered contrast settings. The X-ray emission shows a well-developed shell, brightest in the southwest, where the optical and radio emission are also brightest. Echelle spectra (Blair et al. 1988) show bulk motions of 272 km s^-1. The smudge in the V-band continuum frame is likely due to [O iii] emission transmitted through the bandpass.

G98-73 (or L10-096), shown in the bottom panel of Figure 14 (see also Figure 40, bottom row), is a small, diffuse shell-like SNR showing bright knots on the southeast and northwest limbs. A very bright, compact H ii region ∼6'' to the southeast is apparently not related. The [O iii] emission (not shown here) reveals a complete shell, with the brighter knots less evident than in the other emission-line images. Echelle spectra (Blair et al. 1988) show bulk motions of 185 km s^-1. The X-ray emission is barely resolved and appears to fill the shell. The compact H ii region is a 24 μm source; the SNR, however, is not.

8.2. X-ray Spectral Analyses

Abundance analysis can in principle yield clues to the SN type, while the derived temperature and ionization timescale can yield information on the evolutionary state of the remnant. The extracted spectrum for each remnant is shown in Figure 15. Using XSPEC, we first fit all of the spectra to simple plane-parallel shock models (pshock (Borkowski et al. 2001) in XSPEC with NEIVERS set to the default of 1.1). We fixed the Galactic absorption along the line of sight to be 6 × 10²⁰ cm⁻², using the tbabs model (Wilms et al. 2000) in XSPEC with the abund parameter set to wilm and the cross sections to vern (Verner et al. 1996). We fixed the metallicity for absorption in M33 to be 0.5 times solar, but allowed the column in M33 to vary using the tbvarabs model in XSPEC. We fit the shock temperatures, the overall metallicities, and the ionization ages for the shocks. The results are given in Table 7, and the model fits are compared to the data (with residuals plotted below) in Figure 15. The simple plane-parallel shock models reproduce the spectra of G98-21, G98-28, G98-29, G98-55, and G98-73 fairly well. The metal abundances are sub-solar as expected for M33, varying from 0.18 to 0.66. The best-fit shock temperatures are of order 0.6 keV (except for G98-29 and G98-73 which have best-fit temperatures of 0.84 keV and 1.05 keV, respectively), and χ²_ν values, while exceeding unity per degree of freedom, are reasonable. Thus, the fits for these objects are consistent with shocks propagating into M33 ISM material.

Zoom In Zoom Out Reset image size

Table 7. Plane-parallel Shock Model^a Fits

Parameter	G98-21	G98-28	G98-29	G98-31	G98-35	G98-55	G98-73
NH (1021 cm−2)	0.0[0.0,0.3]b	0.0[0.0,0.3]	0.0[0.0,2.0]	5.8[4.3,7.3]	1.0[0.0,3.2]	0.0[0.0,1.1]	0.0[0.0,1.0]
kT (keV)	0.57[0.55,0.59]	0.61[0.60,0.66]	0.84[0.63,1.40]	0.43[0.38,0.48]	0.78[0.65,0.99]	0.60[0.53,0.66]	1.05[0.66,2.47]
Abundance	0.36[0.32,0.42]	0.46[0.35,0.69]	0.46[0.24,1.17]	0.19[0.14,0.28]	0.18[0.10,0.35]	0.33[0.23,0.45]	0.66[0.33,1.81]
τ ($\mbox{$\rm 10^{11}\rm \:\rm cm$}^{-3} {{\rm s}}$)	5.23[4.44,6.14]	4.61[3.01,5.87]	0.88[0.32,1.68]	0.56[0.34,0.98]	3.40[1.55,7.98]	2.29[1.58,3.06]	0.43[0.25,0.62]
Norm (10−4)	1.95[1.71,2.33]	0.27[0.19,0.35]	0.10[0.03,0.21]	2.09[1.30,3.30]	0.24[0.13,0.43]	0.31[0.23,0.46]	0.05[0.02,0.14]
χ2ν	1.39	1.33	0.83	3.13	1.39	0.95	1.10

Notes. ^aXSPEC model: tbabs(tbvarabs*pshock). See the text for explanation. ^b90% (χ²_min + 2.706) confidence intervals are indicated in brackets.

Download table as:  ASCIITypeset image

The pshock fits to the spectra of G98-31, and to a lesser degree G98-35, are not as good. It is evident from Figure 15 that the region around 1.4 keV, which contains the Mg xi triplet, is fitted particularly poorly. To improve the fit we considered the vpshock model, a model similar to pshock but in which individual abundances can vary. In our case, we allowed the abundances of the O, Ne, and Mg to be free, since these elements are expected to emerge from core-collapse SNe and may have a significant signal in our bandpass. We also allowed the abundance of Fe to vary, so that the abundances of O, Ne, and Mg could be compared to Fe. The results are given in Table 8, and the model fits are compared to the data (again with residuals plotted below) in Figure 16. Clearly, the fit around the Mg xi triplet region is significantly improved in these models, as evidenced by the improvement in the reduced χ² from 3.13 to 1.57 for G98-31 and from 1.39 to 0.64 for G98-35.

Zoom In Zoom Out Reset image size

Table 8. Variable Abundance Plane-parallel Shock Model^a Fits

Parameter	G98-31	G98-35
NH (1021 cm−2)	3.5[1.2,5.6]b	0.0[0.0,1.9]
kT (keV)	0.48[0.40,0.66]	0.60[0.49,0.67]
O	0.21[0.15,0.31]	1.06[0.32,4.00]
Ne	0.23[0.17,0.34]	0.99[0.51,2.64]
Mg	0.61[0.48,0.85]	0.93[0.37,2.26]
Fe	0.10[0.09,0.19]	0.15[0.07,0.41]
τ ($\mbox{$\rm 10^{11}\rm \:\;$}{\rm cm}^{-3}\;{\rm s}$)	1.55[0.84,2.76]	> 24.8
Norm (10−4)	1.42[0.53,2.82]	0.24[0.10,0.59]
χ2ν	1.57	0.64

Notes. ^aXSPEC model: tbabs(tbvarabs*vpshock). See the text for explanation. ^b90% (χ²_min + 2.706) confidence intervals are indicated in brackets.

Download table as:  ASCIITypeset image

For G98-31, the variable abundance fit resulted in [O], [Ne], [Mg] abundances of 0.21 (0.15–0.31 at 90% confidence), 0.23 (0.17–0.34), and 0.61 (0.48–0.85), respectively, and an [Fe] abundance of 0.10 (0.09–0.19), where [Z] indicates the abundance of element Z relative to the solar abundance. The other metals were held to the value of 0.19 solar based on the (pshock) model fits. While the absolute abundances of O, Ne, and Mg are not especially high, the [O]/[Fe] and [Ne]/[Fe] ratios are ∼2 and the [Mg]/[Fe] ratio is ∼6; we can say with some confidence that O, Ne, and Mg are all enhanced relative to Fe. Optically, G98-31 is dominated by swept-up material, so the enhanced abundances in the X-ray analysis indicates enrichment by a significant mass of O, Ne, and Mg from SN ejecta. This argues for a core-collapse explosion of a high-mass progenitor (see Rauscher et al. 2002, and references therein).

For G98-35, the variable abundance fit resulted in [O], [Ne], and [Mg] abundances of 1.06 (0.32–4.00), 0.99 (0.51–2.64), and 0.93 (0.37–2.26), respectively, and an [Fe] abundance of 0.15 (0.07–0.41), while the other metals were held to the value of 0.18. The O, Ne, and Mg appear to be enhanced with respect to Fe ([O]/[Fe], [Ne]/[Fe], and [Mg]/[Fe] are all ∼6), and enhanced compared to the mean metallicity derived from the simpler pshock model fits of 0.18, but the uncertainties are large. We consider it likely that G98-35 also resulted from the core-collapse of a high-mass star, and that there is some evidence for this in the abundance pattern derived from the fits, but the case is not as strong as for G98-31. Interestingly, as with G98-31, there is no clear evidence for peculiar abundances in the optical data for this object.

For both G98-31 and G98-35, we assert that the unusual spectra of these objects are significantly contaminated by ejecta and core-collapse SNe are required since the lower Z elements appear to be overabundant relative to Fe. Beyond that, it is very difficult to be more precise about the nature of the progenitors. While the spectra for G98-31 and G98-35 are clearly different from those of the remaining objects in our bright sample, the signal-to-noise ratio in our spectra is fairly low, and it is likely that there is a substantial forward-shock contribution to the emission we observe. Neither of these objects show evidence of anomalous abundances in optical spectra, although the velocity widths of emission lines are significant (indicating relative youth). Furthermore, in Chandra-based studies of Galactic and Magellanic Cloud SNRs, individual ejecta features are typically isolated and analyzed whereas our spectra of G98-31 and G98-35 represent global averages with ejecta contaminated by overlying blast wave emission in each object. This explains why the abundance ratios in G98-31 and G98-35 are not as extreme as expected from pure-ejecta material (see, e.g., Nomoto et al. 1997), but it does not help in trying to determine the mass of the progenitor. The fact that we do not detect the K-shell emission from the higher Z elements, especially Fe, also produces considerable uncertainty. It is possible that detailed model fits to specific SNR models could improve this picture but that is beyond the scope of this paper.

We can estimate the age, the initial ambient density, the explosion energy, and the swept-up mass of these seven brightest M33 remnants if we make the somewhat oversimplified assumption that they are in the Sedov phase of evolution. Although there is evidence from the X-ray and optical morphology (and in the case of G98-31 and G98-35, X-ray ejecta emission) that these remnants are not yet fully evolved to the Sedov phase, this is still a useful exercise for a qualitative look at the properties of these seven objects. For this analysis, we adopted the sizes of the remnants as derived from the X-ray images, the unabsorbed flux in the 0.35–5.0 keV band, and the temperature as obtained from the X-ray spectral fits. For an assumed distance of 817 kpc to M33 and an assumed emissivity of the thermal plasma models pshock and vpshock appropriate for the best-fitted temperature, we can follow the formulation in Cox (1972) to derive the parameters of a remnant in the Sedov phase. For G98-21, G98-28, G98-29, G98-55, and G98-73, we adopted the fitted parameters from the pshock models listed in Table 7; and for G98-31 and G98-35, the fitted parameters from the vpshock models listed in Table 8. The inferred shock velocity is directly proportional to the square root of the fitted X-ray temperatures,²⁸ and since all of the remnants except G98-29 and G98-73 have an inferred temperature of around 0.6 keV, the inferred shock velocities are all in a narrow range from 600 to 700 km s^-1. For G98-29 and G98-73, the X-ray temperatures of 0.84 keV and 1.05 keV give shock velocities of 800 km s^-1 and 900 km s^-1, respectively.

The inferred ages are about 5000 yr, with G98-31 the youngest at 3300 yr and G98-29 the oldest at 9700 yr. These are most likely overestimates of the true age, since assuming a Sedov model for a remnant which has not yet reached the Sedov phase leads to an overestimate in most cases. The inferred initial ambient densities are around a few per cm^-3, ranging from n₀ ∼ 0.20 cm^-3 for G98-29 to n₀ ∼ 5 cm^-3 for G98-31. The inferred explosion energies are all in the range 0.3 to 0.55 × 10⁵¹ erg, except for G98-21 and G98-29 with energies of 1.5 × 10⁵¹ erg and 1.4 × 10⁵¹ erg, respectively. Finally, only two of the remnants appear to have swept-up a mass ≳100 M_☉: G98-21 and G98-29 have swept up an estimated 250 M_☉ and 162 M_☉. As a group, these seven brightest remnants are rather homogeneous in terms of their inferred properties, with G98-21 and G98-29 exhibiting the greatest difference from the others. The picture that emerges is one of fairly young SNRs (ages around 5000 yr) which went off in fairly dense regions of the ISM (n₀ ∼ 0.2–5cm⁻³). These characteristics are consistent with the fact that these remnants are the most X-ray luminous in our sample. G98-21 and G98-29 may have been overenergetic explosions in a dense medium which has resulted in their high X-ray luminosities.

9.1. Comparison with Remnants of Historical Galactic Supernovae

9.2. Comparison with the Local Group Remnant Population

The samples of SNRs most directly comparable to those in M33 are those in other Local Group galaxies, namely the Large and Small Magellanic Clouds (LMC and SMC) and M31. In their ROSAT-based survey of SNRs in the LMC, Williams et al. (1999) reported detections of 31 SNRs in the LMC using the PSPC and the HRI: this sample includes the extremely young object SN 1987A. Williams et al. (1999) also identified six other known X-ray-detected SNRs that were not detected or observed by ROSAT: 0450-70.9, 0500-702 (N186D), 0513-692, 0524-664 (DEM L175a), 0527-658 (DEM L204), and 0527-658 (DEM L241). Of these, 0450-70.9 and 0527-658 have been the subjects of recent pointed observations with XMM-Newton (see Williams et al. 2004 and Bamba et al. 2006, respectively). Therefore, excluding SN 1987A, a total of 36 LMC SNRs have been detected in X-rays. The diameter distribution of the X-ray-detected SNRs peaks near 40 pc, but also includes 0450-70.9 which, with a size of 98 × 70 pc, is the largest single X-ray-emitting SNR known (Williams et al. 2004). These X-ray-detected SNRs constitute nearly all of the well-studied SNRs in the LMC; Payne et al. (2008) recently asserted that there are 56 known SNRs and another 20 candidate SNRs in the LMC based on recent radio studies, but very few of the objects not yet detected in X-rays have been confirmed through follow-up optical or X-ray investigation. In the SMC, 18 SNRs are known to exist (Filipović et al. 2008). All have been detected in X-rays; in addition, those authors identify a new candidate X-ray SNR which was also detected in the radio but not in the optical, along with a candidate radio SNR detected (weakly) in the optical but not in the X-ray. The diameter distribution of the SMC SNRs peaks near 45 pc. Finally, M31 has a significant number of optically identified SNRs and candidates, at least 221 as summarized and discussed by Matonick & Fesen (1997). Most of the sample is due to Magnier et al. (1995), who reported 178 objects, of which less than half of these would have passed the [S ii]:Hα criterion we have used here for optically selected SNRs. Of the M31 SNRs and candidates, fewer than 30 were detected in X-rays by Pietsch et al. (2005) using XMM-Newton. It is possible that because of the high inclination angle at which M31 is observed, significant internal absorption within the galaxy's disk prevents many SNRs from being detected at X-ray energies. Individual SNRs in M31 have been detected with Chandra and discussed (see, e.g., Williams et al. 2005), but a comprehensive study of SNRs in M31 using all of the available ACIS data has not yet been published.

The luminosity function for SNRs in M33 is shown in Figure 17. There are 29 SNRs in M33 with 0.35–2 keV luminosities in excess of 10³⁵ erg s⁻¹, and seven in excess of $\rm 10^{36}\rm \;{\rm erg\:s^{-1}}$. The figure also shows the 0.35–2 keV luminosity function for SNRs in the LMC and SMC. Despite their much smaller distances, neither the LMC nor the SMC has been searched to the same limit over the entire face of the galaxy that we have achieved for M33. However, it is at the bright end of the luminosity functions that discrepancies arise. As was pointed out initially by Ghavamian et al. (2005), the number of high-luminosity SNRs in the LMC exceeds that for M33. We find only two SNRs with L_x > 10^36.5 erg s^-1 in M33, whereas the LMC has eight such objects. It is not clear what accounts for this difference, but it may simply arise from small-number statistics. The bright end of the LMC SNR luminosity function comprises a very heterogeneous collection of objects. The brightest of these, N132D, still shows evidence of SN ejecta from a core-collapse explosion. Some of the other bright remnants have pure Balmer-dominated spectra and are thought on a variety of grounds to arise from Type Ia SNe. Others like N49 and N63A are interacting with dense material. The brightest SNR in M33, G98-21, is also bright because it is expanding into a dense medium. G98-31, the second brightest X-ray SNR, is nearly point-like in X-rays and appears to be emitting primarily from shocks in the ejecta.

Zoom In Zoom Out Reset image size

The star formation rate (SFR) in M33 is about 0.2  M_☉ yr⁻¹ (see, e.g., Verley et al. 2007), similar to that in the LMC 0.19–0.26 M_☉ yr⁻¹(Kennicutt et al. 1995; Whitney et al. 2008), and so the rate at which core-collapse SNe occur should be similar. The agreement of the curves at 10³⁵ erg s⁻¹ is comforting in this regard. By contrast, the SFR for the SMC is about 0.05 M_☉ yr⁻¹ (Kennicutt et al. 1995; Wilke et al. 2004), and, as expected, the number of remnants is about one-third that of the other two galaxies. At 2 × 10³⁴ erg s^-1, the number of SNRs in M33 exceeds that in the LMC. Assuming this does not represent incompleteness in the LMC X-ray sample at these luminosities, the easiest way to explain this is to argue that the SNRs in the LMC are evolving faster than they are in M33, and hence fade earlier. This would seem to require the typical density of the ISM in the LMC to be higher than in M33, but it is not obvious why that would be. Considerable data have been obtained for the LMC in recent years at X-ray, radio, and optical wavelengths, much of which is not yet published. Given the fact that the Magellanic Clouds are much closer than M33, it should be possible ultimately to create a more complete, better characterized sample of SNRs there than in M33. It will be interesting to compare the luminosity functions once this is done.

Recently, Pannuti et al. (2007) examined X-ray SNR populations in five spiral galaxies—M81, M101, NGC 2403, NGC 4736 (M94), and NGC 6946—that are outside the Local Group but still relatively nearby (within 7 Mpc). These five galaxies host a total of 138 objects identified as SNRs based on [S ii]:Hα ratios and 50 objects suggested to be SNRs based on their radio properties (namely, non-thermal spectral indices with Hα counterparts). Nine of the optically identified objects and 12 of the radio sources have X-ray counterparts within 2''–25, and so most of the associations are likely real. The limiting luminosities for the X-ray observations of the galaxies examined by Pannuti et al. (2007) were higher (∼2 × 10³⁶ erg s^-1) than for M33, but many of the trends were similar. As is the case with M33, the mean diameter of the objects with X-ray counterparts is smaller than that of SNRs without counterparts (24 ± 4 pc and 62 ± 6 pc, respectively).

9.3. PWNe in M33

The ChASeM33 survey and our supporting multiwavelength analysis has enabled us to identify the largest well-characterized sample of SNRs that exists for any galaxy, except perhaps our own. Of the 137 SNR candidates in this sample, we observed 131 with Chandra and determined the X-ray luminosities for 82, with upper limits on the remaining 49. The six other objects were outside or at the very edges of the ChASeM33 survey fields. The detected SNRs have L_X ranging from a minimum of 2 × 10³⁴ erg s^-1 to a maximum of 9 × 10³⁶ erg s^-1. From the Chandra data and our improved measurements of the optical and radio properties of the SNRs—including determination of the Hα luminosities of the sample, compilation of the spectroscopic data, and new VLA observations—we find the following.

• 1.  
  There are no missing SNRs dominated by thermal emission brighter than L_x≃4 × 10³⁵ erg s^-1 within our survey area. Our spectroscopic search of the brighter Chandra sources in ChASeM33 should have revealed them all, regardless of whether they had been identified previously. All of them should have been recognized by their soft, plasma-dominated spectra or, if they were larger than about 10 pc, extended X-ray emission, even if they had been missed in the optical and radio surveys.
• 2.  
  There appear to be no direct analogs of Cas A, Kepler, or Tycho in M33, or more precisely in the regions of M33 that were well observed with Chandra. Tycho's SNR, the faintest of these, should have produced ∼500 counts in the energy range 0.35–2 keV in 400 ks, unless the absorption along the line of sight in M33 to such an SNR is higher than expected. The only well-known historical SNR that could have been missed is SN 1006, which would be at or below our detection limit. In FL281, however, we may have discovered the first PWN in M33.
• 3.  
  Two of the brightest SNRs in M33—G98-31 and G98-35—show evidence of abundance anomalies expected if their X-ray emission is contaminated by material produced in the SN explosion. Their spectra are suggestive of a core-collapse SN, which is unsurprising since most of the SNe in M33 should have arisen from core-collapse SNe. However, the abundances are not as extreme as the youngest core-collapse SNRs observed in the LMC and SMC; the resulting weaker lines and limited signal-to-noise ratio of the spectra do not enable us to place meaningful constraints on progenitor masses. The optical spectra of these SNRs do not reveal fast ejecta-dominated filaments, and as a result it is unlikely these SNRs are extremely young objects.
• 4.  
  There is no characteristic luminosity for an SNR of a certain size, because an SNR expanding into an ISM with a density of 0.1 cm⁻³ (for instance) evolves more slowly and to lower luminosity than one expanding into an ISM with a density of 0.5 cm⁻³. Unfortunately, without estimates of current expansion velocities, there is no easy way with the existing data to remove or calibrate the effects of local environment on the other properties of SNRs one might want to measure.
• 5.  
  There are no strong correlations between the X-ray luminosities of M33 SNRs and their properties at other wavelengths, although extreme objects at one wavelength tend to be extreme at other wavelengths. X-ray emission is not the dominant source of radiative energy loss for most SNRs, at least as measured in the 0.35–2 keV band. For most SNRs, the Hα luminosity exceeds the 0.35–2 keV X-ray luminosity.

Support for this work was provided by the National Aeronautics and Space Administration through Chandra Award Number G06-7073A issued by the Chandra X-ray Observatory Center, which is operated by the Smithsonian Astrophysical Observatory for and on behalf of the National Aeronautics Space Administration under contract NAS8-03060. P.P.P. and T.J.G. acknowledge support under NASA contract NAS8-03060. P.F.W. and E.K.M. acknowledge additional support from the National Science Foundation through grants AST-0307613 and AST-0908566. The work by R.H.B. was partly performed under the auspices of the US Department of Energy by Lawrence Livermore National Laboratory under contract DE-AC52-07NA27344. We acknowledge our extensive use of the optical data on M33 from the Local Group Galaxies Survey, and are grateful to P. Massey and his colleagues for obtaining these data and for making them freely available. This work has made extensive use of SAOImage DS9 (Joye & Mandel 2003), developed by the Smithsonian Astrophysical Observatory. We also acknowledge the heroic efforts of the referee, M. Filipović, for an extremely careful reading of the original manuscript and numerous comments that have improved this paper.

X-ray and optical images of the SNRs detected in M33 are provided in Figures 18–50, with four SNRs shown in each figure and four panels per figure. The four panels include (from left to right) ChASeM33 data, continuum-subtracted Hα and [S ii] images, and the V-band continuum image of the region, all from the LGGS data. Where possible, the X-ray images were constructed from the subset of the ChASeM33 observations where the SNR was within 5' of the center of the Chandra field. A scale bar is shown in the rightmost panel of each four-panel figure. When appropriate, comments about the appearance in [O iii] λ5007 are also included, although these data are not shown. A white contour on the X-ray and V-band panels shows the adopted extraction region we have set to define each object's extent. The following notes summarize what is known about each object.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

• 1.  
  L10-001=G98-01 (Figure 18)This is a large, irregular nebula somewhat resembling the letter "S" that stands out from the background confusion in the [S ii] image. No [O iii] emission is indicated and only a couple of stars are seen in projection; they do not appear to be associated directly with the nebula. This object was outside the region of the X-ray survey, being on the far western side of the survey region. Existing optical spectra confirm the elevated [S ii]:Hα ratio and, surprisingly, indicate a density significantly above the low-density limit. The object is either a very old, large SNR or a fossil superbubble whose exciting stars have largely evolved and disappeared.
• 2.  
  L10-002=G98-02 (Figure 18)This is a faint region of elevated [S ii]:Hα ratio that overlaps a bright H ii region. We can only verify the faint northern portion due to confusion and set the extraction region based on this. However, the southern extension seen in the [S ii] image is suggestive of a larger shell. Only an upper limit of 7 × 10³⁴ erg s^-1 on the X-ray emission is available. Optical spectra imply a density somewhat elevated above the low-density limit.
• 3.  
  L10-003=G98-03 (Figure 18)G98-03 is located to the south of the region surveyed for ChASeM33 and thus no X-ray data are available. The object is a large, faint shell in [S ii]:Hα with several bright stars interior to and adjacent at least in projection. One compact H ii region adjacent to a star is seen centrally located in the shell, something that is sometimes seen in superbubbles such as DEM301/N70 in the LMC. This combination of features makes a superbubble interpretation most likely for this object.
• 4.  
  L10-004=G98-04 (Figure 18)This is a faint, limb-brightened thick shell filled with diffuse emission in Hα and [S ii]. No [O iii] is visible. No optical spectra are available, and it is outside the X-ray region surveyed by ChASeM33.
• 5.  
  L10-005=XMM068 (Figure 19)This is the object also known as XMM068, reported previously by Ghavamian et al. (2005). A nearly complete but somewhat irregular broken shell of emission is seen in Hα and [S ii], and is also present (but somewhat fainter) in [O iii]. Although the identification is solid from the imagery available, an MMT-BCS long-slit spectrum reported in this paper provides a solid confirmation as an SNR, and shows that densities are in the low-density limit. The modest X-ray detection is consistent with a shell with some interior emission as well.
• 6.  
  L10-006=G98-05 (Figure 19)This is a faint, almost complete and circular limb-brightened shell in Hα and [S ii]. [O iii] emission looks similar but is even fainter. The X-ray remnant appears to have been detected separate from a bright X-ray point source present just east of the SNR. Densities are in the low-density limit of the [S ii] ratio. Another SNR, G98-06, is centered 20'' to the southeast.
• 7.  
  L10-007 (Figure 19)This faint X-ray source was identified as part of the "extended X-ray source" analysis. Comparing the position against the optical data revealed a large, very faint, patchy circle of emission that was present in Hα and [S ii]. The exceedingly faint optical nebulosity has a somewhat enhanced [S ii]:Hα ratio of 0.32–0.36, but given the difficulties of measuring this ratio for such a faint object and the likely association of X-ray emission, it was included in our candidate list. A few stars are seen in projection against the nebula, but their association with it is not clear. No obvious [O iii] emission is present. No spectral data are yet available. The X-ray morphology appears to be center-filled, although the detection significance is low (2.7σ).
• 8.  
  L10-008=G98-06 (Figure 19)This SNR is a nearly circular limb-brightened shell very similar to and only slightly smaller than G98-05 (20'' to the northwest). The SNR is seen most clearly in [S ii] since it overlaps with more extended, structured Hα emission on the south and east sides. Only the brightest part of the Hα shell is present in [O iii]. Optical spectra show low densities, and only an upper limit is available in X-rays.
• 9.  
  L10-009=G98-07 (Figure 20)G98-07 corresponds to an ill-defined patch of emission in [S ii] and Hα, with diffuse, faint emission toward the north and a brighter patch in the south. A region is simply defined that encompasses the observed emission. No optical spectrum is yet available. The source was outside the region explored by ChASeM33.
• 10.  
  L10-010=G98-08 (Figure 20)This is a large, loopy, faint emission structure with elevated [S ii]:Hα ratio. [O iii] is faintly present from the brightest regions in Hα. Optical spectra show low densities and only an upper limit is available in X-rays. Although a couple of stars are seen in projection against this structure, no fossil association is obviously present. This is either an exceeding large SNR or a fossil superbubble whose primary stars have all evolved.
• 11.  
  L10-011=G98-09 (Figure 21)This very high surface brightness optical SNR is present in Hα, [S ii], and [O iii], and has been confirmed spectroscopically to be shock-heated by a number of investigators over the years, including new MMT Hectospec data reported here. HST images of the object with the original WF/PC instrument published by Blair & Davidsen (1993) resolve the bright patch into a partial shell open to the northwest. The X-ray emission arises primarily at the location of the bright optical emission. Presumably the shock wave from an SN explosion centered northwest of the bright patch must be encountering a significant density enhancement at this location on the shell. The somewhat oversized extraction region is an attempt to account for the object being larger than the obvious bright emission regions. Echelle spectra by Blair et al. (1988) show bulk velocities of order 260 km s^-1. The [S ii] ratio confirms an elevated electron density for this object.
• 12.  
  L10-012=G98-10 (Figure 20)This is a faint, broken ring of optical emission with enhanced [S ii]:Hα ratio extending from the western side of a complex H ii region. [O iii] emission is seen primarily from the western limb of the shell. The X-ray detection implies an X-ray luminosity of about 1.2 × 10³⁵ erg s^-1. Stars present within the H ii region, and even within the nominal shell of the SNR, make it likely that one of the massive stars responsible for the H ii emission exploded as a core-collapse SN. No spectral data exist for this candidate. Two other Gordon et al. (1998) SNRs, G98-12 and G98-13, are also associated with this H ii complex, but neither are detected in X-rays.
• 13.  
  L10-013=G98-11 (Figure 21)This moderately strong X-ray source corresponds with an unremarkable optical SNR candidate consisting of two brighter knots of optical emission surrounded by more diffuse emission that may or may not be related to the SNR. The [O iii] emission looks similar to the other optical images. The X-rays give more of a filled shell appearance although some indication of a shell may be present. In this case, the X-ray emission shows the overall extent of the object better than the optical. The relatively low [S ii]:Hα ratio of 0.47 for this object (compared to many other objects on our list) is still above the nominal threshold for shock heating, and may be artificially low due to potential background subtraction complications. Interestingly, a moderately high density for the bright knots is indicated by the [S ii] ratio despite the overall low optical surface brightness.
• 14.  
  L10-014=G98-12 (Figure 20)G98-12 is a half-shell of bright Hα-[S ii] emission on the north side of the H ii region that has also spawned G98-10 and 13. [O iii] emission is present but very faint from the Hα partial shell. No optical spectrum is yet available, but X-ray emission appears to have been detected.
• 15.  
  L10-015=G98-13 (Figure 20)G98-13 is a faint SNR buried in a complex region. Careful inspection of the Hα and [S ii] images shows a faint oval shell with a brighter spot on the southern edge and a fainter enhancement on the northern side. [O iii] is faint but present and brightest on the southeastern limb of the shell. No optical spectrum is yet available, but the source appears to have been detected at X-ray wavelengths.
• 16.  
  L10-016 (Figure 21)This is a large, oval, diffuse, limb-brightened shell of low surface brightness identified in our Schmidt survey. In Hα, the object is surrounded by faint, patchy emission in the wider field, but the SNR stands out in the [S ii] image. The shell is barely discernible in [O iii] and brighter on the north and east sides, while Hα is brighter on the west. X-ray emission is confidently detected and seems to fill the shell. Our new MMT-BCS data on this object confirm elevated [S ii] emission and show densities at the low-density limit.
• 17.  
  L10-017=G98-14 (Figure 21)G98-14 is a well-formed, limb-brightened shell about 1' northeast of the H ii region that hosts G98-10, G98-12, and G98-13. [O iii] is present but very faint. No optical spectrum is yet available, and only an upper limit on X-ray emission is available in our data.
• 18.  
  L10-018=G98-15 (Figure 21)This X-ray source corresponds with a moderately bright, partially filled shell of emission in Hα and [S ii]. Extensions to the north and west of the main shell in Hα appear to be associated but do not have elevated [S ii]. Much fainter, diffuse [O iii] emission with a brighter knot on the west side are coincident with the shell. Oddly, the low-density limit is indicated by the observed [S ii] ratio for this fairly bright optical SNR. The X-ray emission seems to fill the shell rather than show the outer limb. Blair et al. (1988) found velocities in excess of 180 km s⁻¹ for this object. Based on our inspection of images of M33 in the Spitzer archive, a 24 μm source appears to coincide with this SNR.
• 19.  
  L10-019=G98-16 (Figure 22)G98-16 is a large, somewhat oddly shaped limb-brightened shell most clearly delineated in the [S ii] image. A bright H ii region is visible just to the south, with a stellar association clearly embedded. A faint ridge of emission is present 30'' to the east. The shell is present in [O iii] but more confused with surrounding emission. No optical spectrum is yet available, and only an upper limit on X-ray emission is available in our data.
• 20.  
  L10-020 (Figure 22)This object was found in our Schmidt survey. It is on the western edge of a large "ring" H ii region, and its full morphology is somewhat masked in the Hα image, but it appears as a limb-brightened shell filled with faint diffuse emission. [S ii] reveals the SNR somewhat better. A very faint but complete ring is seen in the [O iii] image. A couple of stellar sources are seen in projection, but are not centrally located in the shell. The X-ray detection is below 3σ, but the X-rays appear to be interior to the shell.
• 21.  
  L10-021=G98-17 (Figure 22)G98-17 is a faint, very oblong ring of [S ii] emission with a brighter spot on its western side. A corresponding structure is seen in Hα, but is confused by other overlying emission. An optical spectrum confirms enhanced [S ii]:Hα ratio, and low densities. Some of this extended structure is visible at very low surface brightness in [O iii]. Our X-ray data only provide an upper limit. A brighter, more compact SNR (G98-20) is present about 25'' to the east, and both objects are in the northern outskirts of the H ii region that hosts the brightest M33 SNR, G98-21.
• 22.  
  L10-022=G98-18 (Figure 22)This is another example where the X-ray emission better defines the extent of the object than the optical emission does. This moderately bright X-ray source coincides with a fairly faint, partial shell of optical emission brightest on the northeast side. X-ray emission seems to fill the shell rather than follow the optical shell. An [O iii] patch corresponds with the brightest spot of Hα and [S ii] emission. Our MMT Hectospec data confirm a very high [S ii]:Hα ratio and a low density for the optical nebula.
• 23.  
  L10-023=G98-20 (Figure 23)This is another example where the X-ray emission better defines the extent of the object than the optical emission does. This fairly bright X-ray source appears as a limb-brightened shell brightest on the east side. A moderately bright optical knot of Hα, [S ii], and [O iii] emission corresponds with the southeastern limb of the SNR, which is located in the northern outskirts if the H ii region that also hosts G98-21 (Gaetz et al. 2007). Under the best contrast setting, the knot almost resolves into an elongated, partial shell open to the west–northwest. Another larger diameter optical SNR, G98-17, lies ∼25'' to the southwest but is not apparently detected in X-rays. A fairly high density is indicated for G98-20 in our MMT Hectospec data, consistent with the high optical surface brightness.
• 24.  
  L10-024=G98-19 (Figure 23)This is a large, loopy shell with a centrally located bright star or tight association and marginally enhanced [S ii]:Hα = 0.41–0.48 and no evidence for X-ray emission. Both stellar winds and/or possible SNR shocks may be involved in raising the observed [S ii]:Hα ratio, and this appears more consistent with the object being primarily classified as a superbubble.
• 25.  
  L10-025=G98-21 (Figure 23, see also Figure 13, top row)This is the brightest M33 X-ray SNR, and has been discussed at length by Gaetz et al. (2007) and above in Section 8. Because of the discrepancy between the physical extent of X-ray and optical regions and the complexity of the surrounding optical emission, the optical extraction region was adjusted from the X-ray region (shown here).
• 26.  
  L10-026 (Figure 23)This object was first noticed from the diffuse, extended X-ray search. A faint, well-formed shell in Hα and [S ii] is seen on the east side, and a more complicated structure extends toward the west. A very bright, compact H ii region appears just to the south. Stars visible within the shell make a combination of stellar winds and SNRs possible as the source of the shocks, and a superbubble identification likely. No optical spectrum is available for this newly identified object.
• 27.  
  L10-027=G98-22 (Figure 24)This is a faint, oddly shaped limb-brightened nebula that stands out best in [S ii]. A star or stars are seen projected within the shell, but a direct association is unclear. Previous optical spectra confirm elevated [S ii]:Hα ratio and densities in the low-density limit. Only an upper limit is available for the X-ray emission. Another SNR, G98-23, lies ∼1' due east.
• 28.  
  L10-028 (Figure 24)This is a very large, faint emission ring brightest is on the southern side and has an image-derived [S ii]:Hα = 0.5–0.55. However, a cluster of stars is seen in projection along the major axis of the ellipse, including apparently one stellar object with excess Hα emission directly associated. The physical size of this nebula, ∼200 pc, would also seem to make a superbubble origin more likely for this object. Stellar winds and/or multiple previous SNe may contribute shocks that have raised the observed [S ii]:Hα ratio. X-rays were not detected from the object.
• 29.  
  L10-029 (Figure 24)This is a faint, limb-brightened shell, very regular on the south side but somewhat irregular in the north, found as part of the Schmidt survey. A brighter spot appears on the southwest side of the shell and close to two stars, arguing for a possible association. An adjacent nebula with very similar characteristics lies just to the west and south. No optical spectrum is yet available, although the object appears to have been detected at X-ray wavelengths.
• 30.  
  L10-030=G98-23 (Figure 24)This is a large, faint, limb-brightened shell broken in the east and with a bright spot in the southwest. The nebula looks identical but fainter in [S ii] compared with Hα, and [O iii] is only marginally detected from the brightest region of Hα. Several bright stars, including some that form a horseshoe pattern, inhabit the interior, raising the possibility of a superbubble interpretation for this object. However, no optical spectrum is yet available. The object was not detected in X-rays.
• 31.  
  L10-031=G98-24 (Figure 25)This is a classic optical "filled smoke ring" SNR in Hα and [S ii], with comparable [O iii] emission as well. It is about 24 pc in diameter, making it similar to the Cygnus Loop in our Galaxy. Our Hectospec data confirm the elevated [S ii]:Hα ratio and densities near the low-density limit. Only an upper limit on X-ray emission is available in our data.
• 32.  
  L10-032=G98-25 (Figure 25)This is an elongated optical SNR with a bright partial shell on the northeast side and fainter, more diffuse emission trailing off to the southwest. Only faint [O iii] emission is present and it is more uniform. Optical spectra show a very high [S ii]:Hα ratio and a density somewhat above the low-density limit. The X-ray emission appears to fill the optical shell. The southwestern X-ray source is harder and is apparently a separate source from the SNR.
• 33.  
  L10-033=G98-26 (Figure 25)The X-ray luminosity is less than 7 × 10³⁴ erg s^-1. This optical remnant is unremarkable, primarily standing out as a diffuse, filled circle on the [S ii] image. Spectra (Table 3) confirm elevated [S ii]:Hα but indicate very low densities.
• 34.  
  L10-034=G98-27 (Figure 25)This SNR lies on the southern side of a somewhat larger "ring" H ii region, with other emission surrounding to the south. The SNR shows a fairly diffuse morphology, with a hint of shell-like structure in Hα on its northern and southern sides. Very faint and fairly diffuse emission is also present in [O iii], but is complicated by other photoionized emission in the region. X-ray emission is well detected and seems to fill the shell rather than showing the limb. The [S ii] ratio is very nearly at the low-density limit in existing spectra. A number of stars are seen in projection in and near the SNR, but their direct association is not assured.
• 35.  
  L10-035=XMM156 (Figure 26)This object is a special case, and is potentially an example of a new oxygen-dominated optical SNR coincident with a fairly bright soft extended X-ray source. The X-ray source is clearly extended into an oblong shell brightest in the northwest. The surrounding region is complex at optical wavelengths, including two adjacent Gordon et al. (1998) SNRs (G98-26 to the west and G98-30 to the east), knotty and diffuse emission regions, and several bright stellar sources, including one that is essentially coincident with the SNR position. Careful inspection of the continuum-subtracted emission-line images, and in particular the [O iii] image (not shown here), implied the presence of emission at a level above that expected from an incomplete stellar subtraction. An MMT-BCS long-slit spectrum confirms the presence of strong [O iii] emission with little or no directly corresponding emission in Hα and [S ii], from a region directly adjacent to the star, but from a region that is much smaller than the X-ray SNR. This object could in principle be similar to the object N132D in the LMC, where the optical O-rich knots are concentrated in an interior region much smaller than the extended X-ray shell. No high-velocity emission was detected in our optical spectrum, however. This object is deserving of further study.
• 36.  
  L10-036=G98-28 (Figure 26)This bright, compact optical SNR is discussed in Section 8. Under appropriate contrast settings, this oblong optical source shows indications of a partial shell open to the east (similar to G98-09). Blair et al. (1988) found velocities in excess of 340 km s⁻¹ for this object. Moderately high densities are indicated by the [S ii] ratio, and a faint 24 μm source aligns with the optical SNR.
• 37.  
  L10-037=G98-29 (Figure 26)This is an example where the X-ray emission better defines the extent of the object than the optical emission does, showing a filled circular patch of emission that is brightest on the southeastern side. The optical emission shows bright knots in the southeast that appear to be the brightest portions of a much fainter shell, also visible in high contrast and on the [O iii] image. Densities moderately above the low-density limit are indicated by existing spectra. A small faint Hα patch just west of the shell is not obviously related to the SNR.
• 38.  
  L10-038=G98-30 (Figure 26)G98-30 is an exceedingly ill-defined patch of enhanced [S ii] emission within extended Hα emission. It is associated with the same regions that include G98-26 and XMM156. Our new MMT-BCS long-slit data show only a marginally enhanced [S ii]:Hα ratio of 0.31, although severe background subtraction problems make this uncertain. No X-ray detection is claimed, so only an upper limit is reported. The identification of this source as an SNR is somewhat questionable.
• 39.  
  L10-039=G98-31 (Figure 27)This is a very bright X-ray source that aligns with a very high surface brightness optical SNR with a bizarre morphology. It was discussed in Section 8. The appearance of apparent diffuse continuum emission at the SNR position could be an artifact of the high surface brightness [O iii] emission lines being passed by the broadband filter image. Blair et al. (1988) found velocities in excess of 330 km s⁻¹ for this object. A strong 24 μm source is coincident.
• 40.  
  L10-040=G98-32 (Figure 27)This is a moderately large, nearly complete shell SNR located about 20'' east of G98-28, in the outskirts of the H ii region NGC 595. The X-ray detection is below 3σ, but suggests a luminosity of 3.7 × 10³⁴ erg s^-1. Some emission interior to the shell may be present, and densities at the low-density limit are indicated by previous spectra. Any [O iii] emission is very faint and confused in the outer emission from NGC 595.
• 41.  
  L10-041 (Figure 27)This is a large diameter, very low surface brightness diffuse optical nebula found in our Schmidt survey with comparable Hα and [S ii]. No [O iii] emission is seen. While several stellar objects are seen in projection against this object, they are not an organized association and their connection to the nebula is uncertain. Such an object in a more complicated region of the ISM would almost surely have gone unnoticed. The X-ray image appears to show emission and appears to fill the large shell, but excess is less than 1σ and so the object is reported as an upper limit in Table 4. The association of even faint X-ray emission with such a large, faint nebula is somewhat unusual.
• 42.  
  L10-042=G98-33 (Figure 27)This is a faint but classic smoke ring SNR about 40 pc in diameter and slightly brighter on its southwest limb. It is visible just outside the glow from a very bright star some 40'' to the northeast. No [O iii] emission is seen. No optical spectrum is yet available, and only an upper limit on X-ray emission is available in our data.
• 43.  
  L10-043 (Figure 28)A faint ∼90 pc shell with somewhat enhanced [S ii]:Hα = 0.44–0.47 is buried in a complex region. A large, bright H ii region is visible to the southwest and a bright, more compact H ii region is projected onto or within the eastern limb of the shell. The projection of stars within the shell favor a superbubble interpretation. No optical spectrum is available. However, X-rays are detected with a luminosity of 1.2 × 10³⁵ erg s^-1, which is somewhat unusual for a superbubble.
• 44.  
  L10-044=G98-34 (Figure 28)This oblong SNR is shell-like, but appears flattened on the northern side. Very weak [O iii] emission is indicated, and the X-ray emission is weak but seems to be primarily interior to the shell. Our new Hectospec data for this object confirm the elevated [S ii]:Hα ratio for this object and show that densities are in the low-density limit.
• 45.  
  L10-045=G98-35 (Figure 28)This is another bright X-ray source coincident with a very high surface brightness optical SNR with bizarre morphology. It is discussed in Section 8. Very high electron densities are indicated by optical spectroscopy, including the new MMT spectra presented in this paper. Blair et al. (1988) found velocities in excess of 310 km s⁻¹ for this object.
• 46.  
  L10-046=G98-36 (Figure 28)This is a fairly diffuse, filled SNR with a shell visible, especially in the north. Faint [O iii] emission especially along the eastern side is indicated. The extended X-ray source seems to fill the shell rather than align with the limb. A high [S ii]:Hα ratio and a density near the low-density limit is indicated by our new MMT-BCS optical spectra.
• 47.  
  L10-047=G98-37 (Figure 29)This is a faint optical SNR, diffuse and filled but with indications of a partial shell on the west side in Hα and [S ii]. Interestingly, it is the east side of the shell that is primarily visible in [O iii] although it is still quite faint. As with the previous source, the X-ray emission seems to fill the shell rather than showing the outer rim. A high [S ii]:Hα ratio and a density at the low-density limit is indicated by previous spectra.
• 48.  
  L10-048=G98-38 (Figure 29)This is a small enhanced patch of emission at the southern end of a long, faint finger of emission. Its appearance is complicated by the need to subtract a star that is centrally located (at least in projection). No appreciable [O iii] emission is seen. Previously published spectra apparently confirm the [S ii]:Hα ratio, and a density near the low-density limit is indicated. Only an upper limit is available on the X-ray emission from this object.
• 49.  
  L10-049=G98-39 (Figure 29)This is a faint, clumpy, irregular optical SNR most visible from the surrounding faint contaminating emission in the [S ii] image. Very faint [O iii] emission is present as well. Again, the faint X-ray emission seems to fill the region rather than showing a rim. No optical spectrum is available for this object.
• 50.  
  L10-050=G98-41 (Figure 29)The long, broken filament identified by Gordon et al. (1998) with enhanced [S ii] may be the brightest southeastern portion of a fainter, much more extensive emission shell barely visible in the Figure. A bright H ii region appears coincident with the northern limb of this shell. Some moderately bright stars are seen in projection within the shell, but their association with the shell is unclear. While stellar winds or even possible SNR shocks may be involved in raising the observed [S ii]:Hα ratio, this appears more consistent with being primarily classified as a superbubble or even a budding supershell. Only an upper limit is available on X-ray emission.
• 51.  
  L10-051=G98-40 (Figure 30)This is a faint, moderately large limb-brightened shell SNR. It is unremarkable in Hα and appears as part of a line of faint emission stretching from northeast to southwest, but the shell stands out more conspicuously in [S ii]. Previous spectra confirm the enhanced [S ii]:Hα and show a density very near the low-density limit. The X-ray detection within the defined contour of the shell is officially an upper limit, but there is a curious clustering of X-rays near the center of the shell that may be significant.
• 52.  
  L10-052=G98-42 (Figure 30)G98-42 is a marginally enhanced smudge of emission extended in a roughly north–south direction within a complex region in the southern spiral arm. It is not detected in X-rays, but two X-ray sources corresponding with nearby stellar objects are present just to the south of the candidate. No optical spectrum is yet available.
• 53.  
  L10-053=G98-43A (Figure 30)This object is located in the LGGS south field and is outside the region surveyed for ChASeM33. The shell identified by Gordon et al. (1998) is the brighter shell in the figure, with a fainter, diffuse shell to the southeast that also appears to have enhanced [S ii]. The nebula Gordon et al. (1998) identified as G98-43 is what we designate as G98-43A. (The fainter oval region we designate as G98-43B and is discussed next.) This is either a single SNR with bizarre morphology or two adjacent SNRs that are either physically associated or projected along the line of sight. A spectrum of G98-43A confirms elevated [S ii]:Hα ratio and an [S ii] density in the low-density limit.
• 54.  
  L10-054=G98-43B (Figure 30)G98-43B is a faint emission region visible in the LGGS data due to its greater sensitivity compared with previous surveys. Interestingly, it is visible (faintly) in Gordon et al.'s figure for this object but was not called out as a separate object or part of the G98-43 SNR candidate. This object is located in the LGGS south field and is outside the region surveyed for ChASeM33. This is either a single SNR with bizarre morphology or two adjacent SNRs that are either physically associated or projected along the line of sight. The brighter shell to the northeast we designate as G98-43A (see above). No spectrum is yet available for G98-43B.
• 55.  
  L10-055=G98-44 (Figure 30)This object may be a more extreme version of the phenomenon seen in G98-41 (L10-050). The long, filament with enhanced [S ii] is clearly part of a much larger structure, but unlike G98-41, no larger potential shell-like structure is evident. No [O iii] is seen from the filament, and it was not detected as an X-ray source. Bright H ii regions are seen to the west. Stellar winds and/or ancient SNe shocks cannot be eliminated as contributors to the enhanced [S ii], but a clear definition as a superbubble is also not obvious. Without further information, the origin of this filament cannot be stated. No spectrum yet exists for G98-44 to confirm the enhanced [S ii]:Hα indicated by imagery.
• 56.  
  L10-056=G98-45 (Figure 31)This is a faint partial shell optical SNR just NW of a bright, compact H ii region. The shell is brightest on the east side, and the X-ray emission seems to be centrally located in the shell. No obvious [O iii] emission is present for this SNR. Low densities are indicated by the [S ii] ratio in existing spectra.
• 57.  
  L10-057=G98-46 (Figure 31)This is a faint, diffuse, oblong filament on the outskirts of a much brighter "ring" H ii region. No [O iii] emission is indicated, and the X-ray detection is just above 3σ, with no particular structure seen to the X-ray emission. Existing optical spectra show an [S ii]:Hα ratio of 0.54 and low densities.
• 58.  
  L10-058 (Figure 31)This is one of the newly identified optical SNR candidates. The object appears as a moderately large kidney-bean shaped limb-brightened shell. An indentation with brightening on the east side gives the impression that the shock is encountering denser material there. However, the object is not detected in X-rays and no optical spectrum is yet available.
• 59.  
  L10-059 (Figure 31)This is one of the newly identified optical SNR candidates, located in a busy field just west of the galaxy nucleus. Careful subtraction of the background revealed a faint, perfectly oval diffuse emission region with just a hint of limb brightening. Several bright stars are seen projected within the shell, but their association with the nebula is uncertain. The X-ray image and the excess in counts measured for the object give the impression that the source is detected, but the location of the source within the scattering region from the bright nuclear source makes the measurement somewhat uncertain. No optical spectrum is available.
• 60.  
  L10-060=G98-48 (Figure 32)G98-48 is a rather ill-defined patch of enhanced [S ii]:Hα about 20'' southeast of L10-059 and hence even closer to the nuclear source. Nonetheless, the X-ray detection of this source appears to be convincing despite the official statistic showing it below 3σ. A bright, compact H ii region is directly adjacent to the north. The object is barely visible in [O iii]. Existing optical spectra confirm the enhanced [S ii]:Hα and show the densities to be in the low-density limit.
• 61.  
  L10-061=G98-47 (Figure 32)This is a fairly large, diffuse emission nebula directly adjacent to a bright, compact H ii region, with significant other surrounding emission in the southern spiral arm. The region of interest is most visible in the [S ii] image where most of the photoionized gas is fainter. There are two diffuse but somewhat patchy lobes, one protruding to the east and the other to the northeast from the compact H ii region. Both lobes show significantly enhanced [S ii] emission and are likely shock heated. (The object is even visible as an extended green nebula in the central panel of Figure 2, below and right of center.) The X-ray emission seems to fill the region of both lobes as opposed to being in a shell-like structure, and does not seem to extend into the compact H ii region to the west. Blair et al. (1988) used an E–W slit at echelle resolution and saw velocities in excess of 100 km s⁻¹ for the eastern portion (they called this object M33-20). The low-density limit is indicated by the [S ii] ratio. The extent of the source is only approximated by the oval extraction region.
• 62.  
  L10-062 (Figure 32)This is a large, very faint circular emission region with a hint of limb-brightening especially in the Hα image. It is just 20'' west of two other SNRs, G98-51 and 52, which are visible at the edge of the frame. L10-062 was discovered in the Schmidt survey and is probably just enough fainter than these other SNRs that it was not confidently detected in earlier surveys. Several stars are seen in projection in and near the shell, but their association is uncertain. X-rays are not detected, and no optical spectrum is yet available.
• 63.  
  L10-063 (Figure 32)This SNR candidate is an exceedingly faint circular region of emission identified in the Schmidt survey. A single star is seen projected within the region but its association is unknown. X-rays are not detected, and no optical spectrum is yet available.
• 64.  
  L10-064=G98-49 (Figure 33)The optical SNR looks similar in all three optical bands (including [O iii]), showing roughly half of an optical shell bright on the west side and open to the east. An extension to the south in Hα does not appear to be part of the SNR. Although below a 3σ detection, the X-rays appear to coincide more with the interior of the SNR as opposed to the shell. The extent is measured assuming the full shell follows the curvature of the visible portion. A fairly low density, but above the low-density limit, is indicated by the [S ii] ratio in previous spectra.
• 65.  
  L10-065=G98-50 (Figure 33)This is a moderately bright optical SNR with a well-defined partial shell on the south and east sides. Fainter patchy and diffuse emission fills the interior of the shell and extends somewhat to the west and north. A bright, compact H ii region appears just to the north and another lies 20'' to the east. Both of these H ii regions are bright Spitzer 24 μm sources. Only a few stars appear in projection and their association is now known. X-rays are not confidently detected. Previous optical spectra confirm the enhanced [S ii]:Hα ratio and show densities in the low-density limit.
• 66.  
  L10-066=G98-52 (Figure 33)G98-52 is the northern object of a pair of adjacent SNRs (the other is G98-51). G98-52 is a faint, diffuse, nearly circular SNR with a hint of limb-brightening in Hα and possibly faint X-ray emission (below 2σ). Very faint [O iii] emission is seen from the northern rim. Despite the optical faintness, an intermediate density is indicated by the [S ii] ratio in existing spectra. To the extent one can tell, the X-ray emission seems to fill the shell.
• 67.  
  L10-067=G98-51 (Figure 33)G98-51 is comparable in size and surface brightness to G98-52 (just to the north), but is less well defined, appearing as a faint, diffuse nebula somewhat extended in the northwest–southeast direction. No X-ray detection is claimed and no optical spectra are available.
• 68.  
  L10-068 (Figure 34)This is a large, very faint shell of enhanced [S ii] emission (0.42–0.55) located on the eastern edge of a bright H ii complex. It was located in our Schmidt survey. The physical size is ∼100 × 120 pc. Several bright stars and/or tight clusters are projected within the shell. Although SN shocks are likely to be involved, stellar winds could also be important in producing the observed feature, making a superbubble interpretation likely. A blow-out from the H ii region is another possibility, although a full shell seems to be present in [S ii], so this could be a projection effect. Only an X-ray upper limit is available.
• 69.  
  L10-069=G98-53 (Figure 34)This is a somewhat irregular, patchy shell SNR about 10'' west of an extended north–south finger of H ii emission. No [O iii] is visible. The X-ray emission appears centrally located in the shell. Existing spectra indicate densities near the low-density limit.
• 70.  
  L10-070=G98-54 (Figure 34)This is a relatively faint SNR on the western side of a complex region of H ii and related emission. A partial shell structure is visible in Hα with a brighter knot on the north side that also shows in [S ii] and [O iii]. The X-ray emission seems to be concentrated on the eastern side of the SNR, where the shock may be interacting with denser surrounding gas. The neighboring H ii region to the east also seems to have been detected as an extended source. Existing spectra indicate densities near the low-density limit.
• 71.  
  L10-071=G98-55 (Figure 34; also Figure 14, middle row)This is a very bright X-ray SNR and is also high surface brightness in all three optical bands. It was discussed in Section 8. An echelle spectrum by Blair et al. (1988) shows motions at 270 km s⁻¹ associated with this object. Fairly high densities are indicated by the optical [S ii] ratio.
• 72.  
  L10-072 (Figure 35)Unlike many of the new SNR candidates identified in our Schmidt survey, this one is relatively modest in size and fairly bright. A well-defined half-shell is present, open to the north. The remainder of the object is invisible in the optical data, but the curvature of the visible portion has been used to define the size. Despite the appearance of X-ray emission in the region, it is not significant given the fairly high background in the region. No optical spectrum is yet available to confirm the source.
• 73.  
  L10-073 (Figure 35)This is a faint, isolated, limb-brightened shell identified in our Schmidt survey, but little is known. Only an X-ray upper limit is available, and no optical spectra have yet been obtained.
• 74.  
  L10-074=G98-57A; L10-076=G98-57B (Figure 35)The previous candidate G98-57 has been separated into two candidate objects as follows. G98-57A is a nearly circular, diffuse, enhanced [S ii] patch in a region of fairly high contamination by other emission. Interestingly, [O iii] emission forms a partial shell visible only on the southwest side of the object. Two small arcs or fingers of enhanced [S ii] emission are visible just north of (but separate from) the SNR proper, and this region also appears to have X-ray emission, but the relation of this to the SNR is unclear. We designate this region as G98-57B. Our new MMT spectra of G98-57A confirm the elevated [S ii]:Hα ratio and show a [S ii] ratio near the low-density limit.
• 75.  
  L10-075=G98-56 (Figure 35)This is a faint, diffuse SNR with only a hint of limb brightening. It is located in a fairly busy portion of the southern spiral arm, with several nearby compact H ii regions and significant extended diffuse emission. A bright star on the eastern limb of the SNR does not appear to be associated. Existing optical spectra confirm the elevated [S ii]:Hα ratio and show densities to be low. Only an X-ray upper limit is available.
• 76.  
  L10-077=G98-59 (Figure 36)G98-59 is an ill-defined and somewhat marginal candidate in many regards. Gordon et al. (1998) apparently identified the brighter optical knot at the center of the figure as the SNR, but this knot is just the brightest portion of a more extended region of emission with comparable [S ii]:Hα ratio. Nonetheless, a previous optical spectra seems to confirm [S ii]:Hα = 0.40, just qualifying it as an SNR candidate under our definition. No [O iii] emission is seen, but interestingly, our inspection of data from the Spitzer archive indicates the whole region shown, including the SNR region as defined here, shows significant 24 μm emission. No X-ray detection is claimed for this object.
• 77.  
  L10-078=G98-58 (Figure 36)This is a relatively small, bright partial shell SNR at optical wavelengths, brightest in the south. The SNR appears flattened on the north instead of round. The X-ray emission appears to correspond more with the northern limb, implying that we are seeing primarily the SE side of the SNR shell at optical wavelengths and the northern rim in X-rays. [O iii] emission is prominent only from the southern partial shell portion. Our new Hectospec data show a low density, but above the low-density limit.
• 78.  
  L10-079 (Figure 36)This object was identified in our diffuse, extended X-ray source search, and inspection of the optical showed a faint nebular structure extended in the east–west direction, in close proximity to two compact H ii regions, one just to the north and a smaller one to the south. There are no clearly associated stars, and an SNR in a very inhomogeneous medium is indicated. No optical spectra are yet available.
• 79.  
  L10-080=G98-60 (Figure 36)This is a small diameter, moderately bright optical SNR located in the outer reaches of a bright H ii region complex. [O iii] is present but faint, and gives an impression of a partial shell open to the north. A stellar source is located directly adjacent to the south, but does not appear to be related directly. X-rays are detected but no shape is visible. Our new Hectospec data indicate an intermediate density for this object.
• 80.  
  L10-081=G98-62 (Figure 37)The optical morphology of this SNR is poorly defined, consisting primarily of a filament or arc of emission at the southwest edge of an H ii complex. It is unclear whether the fainter arc of emission extending to the south and east are directly related or not. The X-ray emission appears centered southwest of this filament, leaving the impression that the optical emission represents just the northeast edge of the object as the shock encounters the outskirts of the H ii region. The extent of this poorly defined object is estimated based on both the optical and X-ray emission. Earlier optical spectroscopy confirms the elevated [S ii]:Hα ratio and shows a low density for this object.
• 81.  
  L10-082=G98-61 (Figure 37)This is a somewhat irregular partial shell SNR at Hα and [S ii] wavelengths. Extended emission visible in Hα does not appear to be associated with the SNR directly. Interestingly, in [O iii] the SNR appears diffuse and filled (but faint), showing the full extent. The X-ray emission seems to fill the object rather than be associated with the bright optical regions. Low densities are indicated by the [S ii] ratio in existing spectra.
• 82.  
  L10-083=G98-65 (Figure 37)This is a faint, ill-defined optical SNR embedded in a complex region of extended optical emission. In [S ii], there is some indication of a partial shell on the west side, connecting to a possible section on the east side that is confused by other emission. No [O iii] emission is clearly associated with the shell. Large groupings of stars are present just to the east and west of the defined region, although no direct association with the SNR is assumed. Existing spectra confirm elevated [S ii]:Hα ratio and show the low-density limit for this object. X-rays appear to be detected, although the source of the X-ray emission seen to the west–northwest of the SNR is not obvious at optical wavelengths.
• 83.  
  L10-084=G98-64 (Figure 37)This is a moderately bright X-ray source coincident with a classic, shell-like Hα and [S ii] SNR reminiscent of the Cygnus Loop in our Galaxy. The [O iii] emission is somewhat fainter and more evenly distributed, with some limb brightening, but is entirely consistent in extent. New Hectospec data give a density just above the low-density limit, and confirm a very elevated [S ii]:Hα ratio. The X-ray emission fills the shell, and appear brightest in the south exactly where the Hα shell is faintest. A star located on the southeastern rim does not appear to be associated.
• 84.  
  L10-085=G98-63 (Figure 38)The optical counterpart of this SNR is peculiar. It appears almost as a partial shell open to the north, but this partial shell appears as a series of four clumps or small arcs of emission. The X-ray emission is centered to the north, implying that we are seeing one side (the southern side) of the object at optical wavelengths. [O iii] emission is visible from the same partial shell, although its distribution is somewhat different in detail from that seen in Hα and [S ii]. No spectrum is available for this object, although the image ratio indicates [S ii] is plenty strong to indicate shock heating. The few faint stars seen in projection do not appear to be related directly to the SNR.
• 85.  
  L10-086=FL236 (Figure 38)This optical counterpart was found via the search for optical emission at the positions of soft X-ray sources. Two faint knots of comparable Hα and [S ii] emission are seen, which may represent bright spots on a small, partial shell. There is no obvious emission in [O iii]. It is unlikely this optical counterpart would have been identified without the X-ray source drawing attention to it. Our Hectospec data confirm the elevated [S ii]:Hα ratio and show a fairly high density for the optical knots. The X-ray emission arises on and just north of the optical knots, indicating they may be on the southern edge of an SNR centered slightly north of the optical knots. Our defined region takes this interpretation into account.
• 86.  
  L10-087=G98-66 (Figure 38)This is a moderately large, faint diffuse SNR with a hint of limb-brightening or shell-like emission in Hα and [S ii], located in a fairly busy nebular field, but still clearly defined. It stands out best on the [S ii] frame. A faint western extension seen in Hα is not seen in [S ii], and is likely not part of the SNR. No distinct [O iii] is visible. X-ray emission seems to fill the object, although emission from a point source on the southeast rim of the SNR affects the X-ray source appearance. Our new MMT-BCS spectra confirm the elevated [S ii]:Hα ratio and indicate a density slightly above the low-density limit.
• 87.  
  L10-088=G98-67 (Figure 38)G98-67 is a moderately bright partial shell SNR centrally located between two H ii region (to the north and south). The shell is brightest on the eastern side. A moderately bright stellar source is seen in projection inside the shell but is not obviously associated. [O iii] appears faintly but seems to be somewhat anti-correlated with the Hα (thus showing the northwest side of the shell). The X-ray detection is weak enough so as to make the X-ray morphology uncertain, but no indication of the shell is seen. Optical spectra from previous work indicate the low-density limit for this object.
• 88.  
  L10-089 (Figure 39)This large arc of enhanced [S ii]:Hα (0.62–0.79) is located on the eastern side of a bright H ii region and appears to be part of a large scale blow-out from the H ii region. While a number of stars appear in projection, it is not clear whether these stars are associated with the nebular arc or with the general population in the field. An excess of X-ray emission over the background indicates the source is detected. The large physical size (∼70 × 120 pc) makes a multiple SN/stellar wind superbubble origin possible for this intriguing object.
• 89.  
  L10-090=G98-68 (Figure 39)This is a fairly faint but classic limb-brightened shell-type SNR at all three optical wavelengths. The distribution in [O iii] appears slightly different, but the extent is the same. X-ray emission is detected at just over 3σ and seems to be centrally located although it is hard to say with confidence. Existing optical spectra indicate densities slightly above the low-density limit.
• 90.  
  L10-091=G98-69 (Figure 39)G98-69 is an exceedingly faint, irregular, and patchy optical SNR. In a more complicated region it probably could not have been identified, but it is located in an unconfused region. Its size is ill-determined, especially in Hα and [S ii]. However, faint [O iii] emission (not shown here) appears to be somewhat more extensive than the other lines. The X-ray emission is faint, but seems to anti-correlate with the Hα, perhaps indicating a faint shell. No optical spectra yet exist for this object.
• 91.  
  L10-092=G98-70 (Figure 39)G98-70 is a large (∼100 pc), patchy emission structure with some indications of shell edges or filaments in conjunction with more diffuse emission. [O iii] emission is too faint to show structure. A handful of stellar sources are seen in projection against the nebula, which, along with its size, imply a possible superbubble interpretation is appropriate. The X-ray emission is faint and not well defined, registering at only 2.3σ. Existing optical spectra indicate densities just slightly above the low-density limit.
• 92.  
  L10-093=FL261 (Figure 40)This optically faint, small partial shell was found via the search for optical emission at the positions of soft X-ray sources. Emission is comparable in Hα and [S ii], and optical [O iii] emission (not shown here) looks identical to Hα. Our new Hectospec data confirm the elevated [S ii]:Hα ratio and show densities significantly above the low-density limit. As with FL236, the X-ray emission is centered just north of the optical partial shell, lending credence to the idea that the optical is showing the southern side of a larger SNR centered to the north.
• 93.  
  L10-094=XMM244 (Figure 40)This moderately strong X-ray source aligns with a diffuse, poorly defined optical SNR candidate in a complex stellar and nebular field. However, a region with comparable Hα and [S ii] emission does correspond with the X-ray source, and similar diffuse emission is present in [O iii] as well. A grouping of fairly bright stars surrounds the position. A long-slit MMT-BCS spectrum (this paper) confirms the identification and indicates moderate densities in the optical nebula.
• 94.  
  L10-095=G98-71 (Figure 40)This is a relatively small, classic shell-like SNR at all three optical wavelengths. A small break in the southeast is the only part of the shell missing, and almost no emission is seen interior to the shell. Using the optical data to define the size, the associated X-ray emission seems to correlate with the southern half of the SNR, although the detection is quite faint (5.7 × 10³⁴ erg s^-1). A long-slit MMT-BCS spectrum (this paper) confirms the identification and indicates moderate densities in the optical nebula.
• 95.  
  L10-096=G98-73 (Figure 40)This is a bright X-ray source associated with a high surface brightness, small-diameter SNR, as discussed in Section 8. Blair et al. (1988) found velocities in excess of 180 km s⁻¹ for this object.
• 96.  
  L10-097=G98-72 (Figure 41)All three optical emission line images are consistent, showing an arc or partial shell of emission open to the southwest. The arc is in the outskirts of a large, bright H ii region to the northwest of the SNR. The X-ray detection is only at the 2.2σ level, but is consistent with the partial shell. The extent of the object is based on an extension of the visible part of the shell. An intermediate density is indicated by previous optical spectroscopy.
• 97.  
  L10-098=G98-74 (Figure 41)This is a faint, diffuse emission nebula with marginally enhanced [S ii]:Hα = 0.41-0.49 from image analysis. Faint [O iii] emission appears to coincide with the somewhat brighter northeastern quadrant of the nebula as seen in Hα. A number of bright stars are seen in projection against the nebula (although their association with the nebula is not ironclad) and a compact emission region is also seen in projection on the Hα frame, which is a feature seen in some of the other purported superbubbles. A large, old SNR identification is not outside the realm of possibility. However, a faint superbubble appears to be the preferred classification. Very low densities are indicated by existing optical spectra. The upper limit to the X-ray luminosity is 2.4 × 10³⁴ erg s^-1.
• 98.  
  L10-099=G98-75 (Figure 41)G98-75 is an ill-defined SNR candidate consisting of a few scattered optical filaments with enhanced [S ii]:Hα emission. The bright compact H ii region to the southeast is a strong Spitzer 24 μm source, and another SNR, G98-77, is located just to the southeast of that. Any [O iii] emission from the SNR is very faint although the H ii regions are detected. No optical spectrum is available, and the object was not detected in X-rays.
• 99.  
  L10-0100=G98-76 (Figure 41)This is a faint but fairly complete shell-like SNR at all three optical wavelengths, somewhat brighter in the northeast. The X-ray detection is only at the 3σ level and little can be said about the distribution of the X-ray emission. Intermediate densities are indicated by existing spectroscopy.
• 100.  
  L10-101=G98-77 (Figure 42)This ill-defined SNR candidate consists of two groupings of faint irregular filaments that are plausibly portions of the same shell, with the brighter filaments aligning with the eastern side and some fainter emission in the north. An ellipse is fit under these assumptions. No appreciable [O iii] emission is seen. Another SNR candidate, G98-75, is located on the top edge of the frame shown. No optical spectrum is available. The object does appear to have been detected in X-rays.
• 101.  
  L10-102=G98-80 (Figure 42)G98-80 is an ill-defined collection of knotty filaments within a fairly confused region of stars and extended emission. Two separate groupings of filaments more or less define the upper half of a shell, which we use to define the region of interest. Existing optical spectra show [S ii]:Hα = 0.46, which may be a lower limit given the overlying emission, and low densities. No X-ray detection is claimed for this object.
• 102.  
  L10-103=G98-79 (Figure 42)G98-79 is a faint, shell-like SNR located just to the southeast of a small H ii region. It is visible faintly in all three optical bands, looking somewhat like a θ in Hα and [S ii]; the [O iii] emission is very faint and somewhat enhanced on the southern side, but the shell is otherwise similar in extent. The X-ray emission seems to be center-filled, based on inspection of the X-ray image, but the detection is less than 2σ level. Existing optical spectra indicate densities somewhat above the low-density limit.
• 103.  
  L10-104=G98-81 (Figure 42)G98-81 is a fairly diffuse oval SNR with hints of limb-brightening in Hα and [S ii], especially on the south side of the SNR. The [O iii] emission is very faint, but appears brighter in the northwest side. X-ray detection at just under 3σ does not allow X-ray morphology to be determined. The low-density limit is indicated by existing optical spectra.
• 104.  
  L10-105=G98-78 (Figure 43)This is a fairly well-defined shell-like SNR, with a faint protrusion to the east in Hα and [S ii] that gives the overall shape a somewhat triangular appearance. The [O iii] emission comes primarily from the north and west limbs of the shell. A stellar source is seen projected just inside the northwest limb of the shell, but is not clearly related. We have defined the extent using the main shell, ignoring the faint eastern extension. X-ray emission seems to fill the upper two-thirds of the shell, but no enhancement corresponding to the shell itself is seen. Low densities are indicated by the [S ii] ratio in existing spectra.
• 105.  
  L10-106 (Figure 43)This is an exceedingly faint shell at Hα and [S ii] detected as part of the Schmidt survey. A very compact Hα emission source coincident with a faint star in the continuum frame is seen in projection against the shell, but is not clearly associated. With a diameter of ∼70 pc and no clearly associated stars, this could be a large, old SNR that has nearly faded from visibility. The object appears to have been detected in X-rays. Unfortunately, no optical spectra are available.
• 106.  
  L10-107=G98-82 (Figure 43)G98-82 is a faint, somewhat patchy, diffuse, circular SNR with a probably unrelated stellar source in projection against the northeast limb. Its extent is well defined despite other nearby diffuse and structured emission. The [O iii] emission is relatively strong compared with many other SNRs in M33, and the distribution of [O iii] light is fairly uniform compared with Hα. X-rays are detected (3.1σ) and the morphology to some extent seem to be center-filled. The low-density limit is indicated by existing optical spectra. To the south of the SNR lies a faint extended "banana" of emission with only slightly lower [S ii]:Hα ratio but whose origin is unknown.
• 107.  
  L10-108=G98-84 (Figure 43)G98-84 is a peculiar, large collection of filaments which we interpret as being a partial shell of emission. The filaments are part of a much larger network of filamentary and diffuse emission that is only slightly lower in [S ii]:Hα ratio. Some of the network of filaments shows up faintly in [O iii] as well. The region defined for the object is 60 × 100 pc and a handful of stars are seen projected within, raising the possibility of a superbubble interpretation. Two compact, high surface-brightness nebulae are also seen in the Hα frame but are not related. A long-slit MMT-BCS spectrum (this paper) confirms the enhanced [S ii]:Hα ratio and indicates densities in the low-density limit. No X-ray detection is claimed for this object.
• 108.  
  L10-109=G98-83 (Figure 44)This candidate corresponds to a partial shell of raised [S ii]:Hα emission within a confused region of other emission and H ii regions to the north. From careful inspection of the [S ii] frame, we identify a partial oval shell and define the region based on this. There is [O iii] emission in the region, but it is not clearly associated with the Hα-[S ii] shell. A star is projected onto the northern limb and one interior, but their association is unclear. A new MMT-BCS spectrum shows an [S ii]:Hα ratio of 0.28, but confusion from overlying emission makes this uncertain. The density indicated is slightly above the low-density limit. No X-ray detection is claimed for this object.
• 109.  
  L10-110 (Figure 44)This object, which stands out as a ring on the [S ii] frame despite being very confused in the Hα frame, was found as part of the Schmidt survey. Faint X-ray emission appears to have been observed, but no optical spectra are available.
• 110.  
  L10-111=G98-85 (Figure 44)This SNR is very similar in size and morphology to G98-79 (which is ∼1' to the northwest) except that it is somewhat brighter. While a shell is prevalent in Hα and [S ii], some emission is seen in projection toward the center of the shell as well. The [O iii] emission arises primarily from the outer shell only. As with many other X-ray detected objects, the X-ray emission seems to be more center-filled. Low densities are indicated by existing spectroscopy of [S ii].
• 111.  
  L10-112 (Figure 44)This exceedingly faint, nearly circular shell was identified in the Schmidt survey. Such an object would be very hard to detect in a more confused region. It is not clear that any of the stars in the region are associated directly with the nebula. With a diameter near 90 pc, this could be an old SNR that has nearly faded beyond visibility. No X-ray detection is claimed for this object, and no optical spectra are available.
• 112.  
  L10-113=G98-86 (Figure 45)This SNR is the limb-brightened ring that stands out best on the [S ii] image, despite the significant confusion caused by the photoionized regions directly adjacent to the southeast and southwest. There is an enhancement in [S ii] on the northeast limb of the shell that, in principle, could be a separate partial shell SNR open on the east side, but without independent confirmation we list this as one object. Another SNR, G98-87, is visible just 25'' to the northeast, and much of the general emission visible is only slightly less enhanced in [S ii]:Hα than the SNR candidate(s). A number of stars are seen projected in and near the SNRs. Existing spectra confirm the elevated [S ii]:Hα ratio and show densities near the low-density limit. While some X-ray emission appears to align with the northeast limb, it is statistically significant only at the 2.5σ level.
• 113.  
  L10-114=G98-87 (Figure 45)This object is a moderately bright but patchy and ill-defined SNR embedded in a region of extensive diffuse but structured emission. Two arcs looking somewhat like crab pincers extend eastward from the brightest section and appear to be part of the SNR. [O iii] emission is present and fairly prominent. The source is detected in X-rays but the morphology is ill-defined. Another SNR, G98-86, is adjacent 25'' to the southwest and also has a poorly defined morphology (but is apparently not detected at X-ray wavelengths). Despite the moderately high surface brightness, previous optical spectra indicate low densities for this object.
• 114.  
  L10-115=G98-88 (Figure 45)G98-88 is a faint, fairly diffuse shell SNR, seen mainly in Hα and [S ii]. A faint extension from the shell to the southeast gives the impression of a possible break-out in this quadrant. The X-ray emission is detected at the 2.5σ level and seems to fill the shell. No optical spectrum is available for this SNR.
• 115.  
  L10-116=XMM270 (Figure 45)This extended but ill-defined X-ray source corresponds with a faint, unresolved knot of Hα and [S ii] emission that would not likely have been identified as interesting without the X-ray data pointing the way. Indeed, the optical emission is directly adjacent to a faint star, making it difficult to confirm from imagery alone. The optical knot may just be a bright spot on the southeastern edge of the X-ray source centered to the northwest. The extent chosen is based on the X-ray data. A new MMT-BCS long-slit spectrum reported in this paper confirms the shock heated nature of the knot but the densities are in the low-density limit.
• 116.  
  L10-117=G98-89 (Figure 46)This is a large, diffuse but limb-brightened SNR with an H ii region directly to the southwest. [O iii] emission is nearly non-existent, and a few stars are seen in projection toward the shell, but their relationship to the object is unclear. X-rays are detected and seem to imply a center-filled morphology. Densities at the low-density limit are indicated by [S ii] in the MMT-BCS spectrum reported here. With a diameter near 70 pc, this is either a very old SNR or a superbubble.
• 117.  
  L10-118=G98-90 (Figure 46)This is a very faint, oblong, shell-like SNR brightest in the south. [O iii] is barely visible. The X-ray detection is above 3σ and the emission seems to arise in the interior of the shell. The low-density limit is indicated by existing optical spectra.
• 118.  
  L10-119=FL312 (Figure 46)This exceedingly faint emission nebula was identified as possibly being associated with the soft, extended X-ray source FL312 as part of our soft-source search. It is too faint optically to determine a reliable [S ii]:Hα ratio from imaging and no spectrum exists. The X-ray emission is roughly coincident with the optical nebula, but there is no sense that it fills the shell or is necessarily a direct counterpart. It is not outside the realm of possibility that this source represents a Balmer-dominated SNR in M33, but its true character (and the association of optical and X-ray emission) remains uncertain.
• 119.  
  L10-120=G98-91 (Figure 46)This patch of enhanced [S ii]:Hα emission is a less extreme version of the phenomenon seen in [G98-41 and G98-44] and described above. G98-91 is even fainter and more diffuse than these other objects, although the enhanced [S ii] region appears to be connected with more extensive and diffuse emission on the Hα image. No stars appear to be directly associated with the emission, and no fainter shell-like structure appears to be associated. Existing spectra appear to confirm the candidate as an SNR and show a density very near the low-density limit of the [S ii] ratio. The nature of this nebula remains uncertain. The ChASeM33 observations yield an upper limit to the X-ray luminosity of 2.9 × 10³⁴ erg s^-1.
• 120.  
  L10-121=G98-92 (Figure 47)This moderately large, broken shell SNR is brightest on the west side in Hα and [S ii]. It is visible in [O iii] as well, although the detailed [O iii] morphology is somewhat different and the shell is more uniform and complete. The X-ray detection is of low significance (2σ) but the emission arises within the shell. Another SNR, G98-93, lies 30'' to the southeast, and is not detected in X-rays. Densities close to the low-density limit are indicated by [S ii] in existing spectra.
• 121.  
  L10-122 (Figure 47)This object is only peculiar due to its huge size and oblong shape (165 × 85 pc). The observed [S ii]:Hα = 0.65–0.72 easily qualifies the object as being dominated by shocks. This remarkable object was found as part of our Schmidt survey, and is located in the far-northern region of the galaxy. Unlike many of the Schmidt survey objects, this one is fairly bright, appearing as a huge, elongated but complete limb-brightened shell in Hα. [O iii] emission is also prominent and looks very similar to Hα except for being brighter along the southeastern and southern rim regions. [S ii] is enhanced in selected regions around the shell, especially in the east, north, and upper west sides. Filaments to the west of the main shell may also have elevated [S ii], but the association of these filaments with the main shell is not entirely clear. While some stars are seen in projection against this large shell, they are not centrally concentrated like they are in several nearby H ii regions, and their association with the nebula is unclear. The X-ray detection is marginally significant at 2.3σ. Because of the huge size, a fossil superbubble would seem to be indicated.
• 122.  
  L10-123=G98-93 (Figure 47)This SNR is a diffuse, oval emission nebula just north of a small, compact H ii region. A hint of limb brightening on the west side is evident, especially in [S ii]. Some [O iii] emission is present from the northern third of the object. No X-ray detection is claimed for this object, and no optical spectra are available.
• 123.  
  L10-124=G98-94 (Figure 47)G98-94 appears as two bright optical Hα and [S ii] knots in the southern extended emission related to the giant H ii region NGC 604. Only the southeastern of the two knots has associated [O iii] emission. It is not entirely clear whether this is a single SNR or a double object similar to DEM L316A+B in the outskirts of 30 Doradus in the LMC ((Williams & Chu 2005). We have defined a single extraction region, which appears to be filled with fairly bright X-ray emission, and treat this as a single SNR. Earlier optical spectra show intermediate densities are indicated. Blair et al. (1988) found velocities in excess of 220 km s⁻¹ for this object.
• 124.  
  L10-125 (Figure 48)This X-ray source (3.7σ detection) was identified as part of the "extended X-ray source" analysis. Comparing the position against the optical data revealed a large, very faint, patchy partial shell of emission (brighter on the east side) that was present in Hα and [S ii]. Faint [O iii] emission (near the level of detectability in the current optical data) is also present. A small H ii region just south and slightly west of the shell is not obviously related to the SNR candidate. X-ray morphology appears to be center-filled. No optical spectrum is yet available.
• 125.  
  L10-126=G98-95 (Figure 48)G98-95 is a faint, diffuse SNR with a modest brightening on the west side. In [O iii], the bright spot seems to show a small, half shell that would be indicative of a much smaller object than indicated by the fainter diffuse emission. (The same structure is hinted at in the Hα image.) No X-ray detection is claimed for this object, and no optical spectra are available.
• 126.  
  L10-127=G98-96 (Figure 48)G98-96 is a large, faint, limb-brightened shell that is brightest on its southwest side. No [O iii] emission is visible. While a few stars are projected within the shell, their association with the nebula is not obvious. No X-ray detection is claimed for this object, and no optical spectra are available. Another new candidate found in the Schmidt survey, L10-131, is located just to the southeast of the field shown.
• 127.  
  L10-129=G98-97A³¹ (Figure 48)The object listed as G98-97 in Gordon et al. (1998) has been broken into two separate SNR listings, based on a careful assessment of the optical imagery. G98-97A is a large north–south oval of bright Hα and [S ii] emission. The east and west sides of the oval are even more prominent in [O iii]. X-ray emission is well detected and seems to fill the shell. Blair et al. (1988) found velocities of nearly 210 km s⁻¹ for this object. Existing optical data show densities somewhat above the low-density limit.
• 128.  
  L10-128=G98-97B (Figure 48)Directly adjacent to the northwest of G98-97A we have identified G98-97B as a separate complete shell, either overlapping in projection or possibly physically associated. G98-97B is about half the size of G98-97A, but its [O iii] emission is considerably fainter, and the X-ray detection is clear but of lower significance than for G98-97A. Our new Hectospec data on this object confirm elevated [S ii] emission and show densities in the low-density limit.
• 129.  
  L10-130 (Figure 49)This newly detected object appears as a partial shell in Hα and [S ii]  open to the north. In [O iii], the full shell is seen, filling the region shown in the figure. The object is located about 25'' north of the G98-97AB pair. No X-ray detection is claimed for this object, and no optical spectra are available.
• 130.  
  L10-131 (Figure 49)This is a huge (120 × 200 pc), oblong nebula found as part of the Schmidt survey. Its size and central clustering of stars makes a superbubble origin likely. No X-ray detection is claimed for this object, and no optical spectra are available. Source L10-127 is in the northwest corner of the frame shown.
• 131.  
  L10-132=G98-98 (Figure 49)This is a clumpy but faint and diffuse emission nebula with a clustering of moderately bright stars seen in projection. While Gordon et al.'s (1998) assessment was that the [S ii]:Hα ratio was above 0.4, our image assessment shows a somewhat lower value of 0.30–0.39. If stellar wind shocks or SNe are involved, this activity has not been successful in organizing the emission into a shell-like structure. It seems likely to us that this emission is associated with the stars in the region and is not clearly an SNR. No optical spectra are available and only an X-ray upper limit is available from our survey.
• 132.  
  L10-133 (Figure 49)This is a moderately bright ∼80 pc shell found in our Schmidt survey. There is a larger faint Hα halo just visible on the Hα image. There is a bright, compact emission region near center of shell, and several bright stellar condensations within the shell (and likely associated with it). Upon careful assessment, the [S ii]:Hα ratio is marginally too low (0.35–0.38). A superbubble is the preferred classification for this nebula. No optical spectra are available and only an X-ray upper limit is available from our survey.
• 133.  
  L10-134 (Figure 50)This faint, oval complete shell with interior diffuse optical emission was found in our Schmidt survey. Very faint [O iii] emission is present as well. Unlike some of the other Schmidt survey objects, this one does not appear to have associated stellar sources. Although it is fairly large in physical size (∼60 pc), an old, large SNR identification is likely. The X-ray detection is only marginally significant at 2.5σ. No optical spectra are available.
• 134.  
  L10-135 (Figure 50)This is a faint, somewhat irregular shell found in our Schmidt survey, in a region contaminated with other extended emission. Our assessment of the optical data indicate a shell-like structure with enhanced [S ii]:Hα that we identify as the candidate. A faint arc on the east side in Hα may demarcate a break-out on that side of the shell, but the [S ii] image is too faint to discern clearly. No optical spectra are available and only an X-ray upper limit is available from our survey.
• 135.  
  L10-136 (Figure 50)This large shell (∼130 pc) was identified in our Schmidt survey as possibly having enhanced [S ii]:Hα emission. A careful assessment indicates a somewhat low ratio of 0.30–0.37. A single bright star or tight cluster is projected within the large shell, and a smaller interior shell may be indicated in the Hα image. A superbubble interpretation seems likely for this object. No optical spectra are available and only an X-ray upper limit is available from our survey.
• 136.  
  L10-137 (Figure 50)This large, faint optical shell was identified in our Schmidt survey, and is located on the far eastern edge of the survey region. This object is very similar to L10-136. In this case, the shell is even more uniform and has a sharper limb, and a couple of bright stars or tight associations seem clearly associated. The [S ii]:Hα ratio seems genuinely enhanced at 0.58–0.71, but the physical size is large (∼130 pc). Spectroscopy with the MMT-BCS reported in this paper confirms the significantly enhanced ratio for the nebula, but the presence of several stars near the center of the shell and the large physical size make a superbubble interpretation likely, similar to the object N70 in the LMC (see Danforth & Blair 2006, and references therein). The enhanced [S ii]:Hα ratio makes it likely that SNR shocks in addition to stellar winds are responsible for the excitation of the nebula. While some faint, distributed X-ray emission may be present, the bulk of the X-rays arise in a small, central region adjacent to the brightest stars. A hint of [O iii] emission is present, but only on the southeast side of the shell.