1. The XMM-Newton Slew Survey Richard Saxton, M. Pilar Esquej Michael Freyberg, Bruno Altieri Andrew Read

2. Overview - 390 slew datasets in archive - PN, MOS-1/2 exposures in Medium filter with the observing mode set to that of previous pointed observation - Average slew length is 70 degrees - Data available for slews > 30 mins - Closed slew, open slew, closed slew. Open slew speed = 90 degrees / hour, i.e. on-source time <~14 secs - Area covered to date ~8000 deg 2 (~20% of sky)

3. Pilot study 1: SAS worked fine after small OAL change. Tangential projection not valid over whole slew. Long slews need to be subdivided to maintain astrometry. Divide slew into 1 deg 2 images and recalculate sky positions Source search using near-standard pipeline eboxdetect/emldetect combination tuned for ~zero background. Process 9 slews, source searching single 0.2-15 keV image (flag=0, pattern<=4), flat exposure map To check if slew data scientifically useful

4. Initial impressions Source extended into a 4 arcmin streak due to 2.6 second frame time Full frame (73ms) mode streak = 6 arcsecs – not a problem Extra pn sensitivity + additional MOS background means little to be gained from analysing MOS slews Epic-pn attitude reconstruction very good MOSPN PN (eFF) Extended full frame (200ms) streak of 18 arcsecs Slew direction

5. Optical Loading Assessment ~10% sources coincident with bright stars - Optical loading or genuine X-ray sources? - Could be an issue but not dominant DSS images of 139 pilot#1 slew positionsDSS on slew positions + 1hour RA

6. Spurious Sources… … due to crab off-axis !Sources 1 & 3 of Pilot#1

7. Pilot Study 1: Summary Source searched images from 9 pn slews and found 139 sources down to det_ml=10, ~0.5 source/deg 2 Identifications Total sources = 139 ROSAT survey IDs = 63 Bright stars (optical loading?), detector flashes, Crab off- axis & spurions + 1 source observed twice in separate slews leave… ~50 unidentified

8. Extended source analysis Source 2: 2x2 arcmin FF mode point sources are fitted well by a PSF, e.g. source 4 Pilot#1 sources - 5x5 arcmin images Enough counts to detect extension in brightest sources

9. Abell 3581 Easily detected as extended in <~14 second exposure (EFF mode) Slew direction

10. Conclusions from pilot 1 - MOS slew data effectively useless - ~10% of sources coincident with bright stars - optical loading? - Expect to find ~ 4,000 sources from current data - Sensitive to extension in brighter sources New operations strategy (now in place):  MOS slews used for calibration  All PN slews larger than 15 minutes down-linked and processed  Medium filter if FF,eFF,LW and Closed for other modes

11. Pilot study 2: Three long slews ~ 110 degrees, 1 each of FF, EFF and LW modes Three long slews ~ 110 degrees, 1 each of FF, EFF and LW modes 3 source-searching strategies investigated 3 source-searching strategies investigated A) 1XMM scheme - separate 0.2-0.5, 0.5-1.0, 1.0-2.0 and 2.0-12.0 keV images – 23 detections (mean of 14.4 counts) A) 1XMM scheme - separate 0.2-0.5, 0.5-1.0, 1.0-2.0 and 2.0-12.0 keV images – 23 detections (mean of 14.4 counts) B) Search single 0.2-12 keV, pattern<=4 image – 110 detections, many spurious B) Search single 0.2-12 keV, pattern<=4 image – 110 detections, many spurious C) Search 0.2-0.5 (pattern 0) + 0.5-12.0 (patt <=4), single image - 64 detections (mean of 8.6 counts) C) Search 0.2-0.5 (pattern 0) + 0.5-12.0 (patt <=4), single image - 64 detections (mean of 8.6 counts) Best results from scheme C: we lose band-specific count rates and hardness ratios but can calculate these a posteriori with eregionanalyse on individual images. Best results from scheme C: we lose band-specific count rates and hardness ratios but can calculate these a posteriori with eregionanalyse on individual images. Also almost all spurious detector features removed (using only singles in very softest band) Also almost all spurious detector features removed (using only singles in very softest band) To investigate optimum processing and source search strategy

12. Usage of Exposure Maps Slew searched using variable exposure maps and detector mask – no unusual effects. Next to last degree of slew Final degree of slew Uneven exposure at end of slew due to closed-loop.

13. Conclusions from pilot 2 - Source density again ~0.5 / deg² - Use full-band image: 0.2-0.5 (patt=0) + 0.5-12.0 keV (patt<=4) & soft-band (~ROSAT) image: 0.2-0.5 (patt=0) + 0.5-2.0 keV (patt<=4) & hard-band image: 2.0-12.0 keV (patt<=4) i.e. 3 Surveys - Use exposure maps and detector masks - Find individual sub-band count rates after source search - Only use FF, EFF and LW mode data - For uniform survey can apply an exposure threshold

14. The XMM Slew Survey : current processing scheme - Processing of slew data (creation of event files) - Creation of images, exposure maps (full-, soft- and hard-band) - Source search full-, soft- and hard-band images using exposure maps - Make catalogue of detections - Make DSS images for each detection - Check for and flag spurious sources - Check for and flag optical loading due to bright stars - Find hardness ratios from individual band images - Make clean catalogue of sources - Cross-correlate with Rosat - Cross-correlate with other catalogues (In preparation…) (In preparation…) - Analyse extensions - Co-add slews and re-search - Optical follow-up…

15. Current Status - Initial processing and event file creation for all available Slews - 374 Slews: 297 of which are FF (206), eFF (61) or LW (30) - 54 Slews: Images & exposure maps created & source-searched (31 FF, 19 eFF, 4 LW) - Slew catalogue frozen (01/04/05) - 780 sources in total (0.2-12.0 keV) band - 645 sources in soft (0.2-2.0 keV) band - 96 sources in hard (2.0-12.0 keV) band - At faint end, - 68 soft-band sources, not detected in the total band - 20 hard-band sources, not detected in the total band - 20 hard-band sources, not detected in the total band - Total of 868 sources in ~1800 deg ² ~0.5 sources per square degree Currently searched slews

16. Current Problems - Processing problems affect ~40% of slews – Concentrating at present on slews which process correctly first time – Finding solutions to particular problems as we progress - Some slews contain sections with high exposure, related to the closed-loop slew phase. Sub-Images with large exposures are excluded at present - High background slews (~25%) are excluded at present - from 297 FF, eFF, LW slews, we are initially processing 219 and expect to find ~3600 sources - from 297 FF, eFF, LW slews, we are initially processing 219 and expect to find ~3600 sources - but, we can probably recover a lot of the high background - but, we can probably recover a lot of the high background slews by subsetting the data using GTIs slews by subsetting the data using GTIs

17. UY Vir DSS Image 2’×2’ Total band (0.2-12 keV) Hard band (2-12 keV)Soft band (0.2-2 keV) (RASS src) Blue circle in X-ray image indicates a detection

18. Four Galaxies ESO443-IG005Mrk 352 MCG 04-53-021NGC5899 Non-RASS

19. Two Merging Galaxy Pairs NGC 4748 II Zw 742

20. Soft source not detected in full energy band Hard source not detected in full energy band Extra background in full-energy band (sometimes just 1 photon!) can flip statistics below the threshold (DET_ML=10) Faint sources detected in soft/hard- band but not in full-band

21. Sources detected in separate slews - Source detected in 3 separate slews: Rev 0570, Rev 0573, Rev 0574 - Known ROSAT source (+faint galaxy) - Evidence for variability (factor ~2)

22. Well Known Sources N132D (4.1″ pixels)

23. 1RXS J134143.6-165736 LMXRB RASS~145 ct/s Pseudo-MIPs (e.g. 2 ~9keV Photons/pixel in a frame) LW mode Extremely Bright Sources

24. … and other oddities… Southern source - RASS Northern source – non-RASS

25. Optical Loading – A Problem? 0.2-2.0 keV counts V Magnitude CD 36-1289 : X-ray emission from this V=10.5 star, but none from neighbours at V=6.5 and V=7.6 Several examples like this - No correlation between source counts and Vmag for bright stars - Consistent with optical loading predictions: 5 counts expected above 200eV only for stars brighter than V=3.75 No optical loading problems – Genuine X-ray sources Medium Filter Stars in Slew

26. Correlations with 2MASS Blue squares: Number of 2MASS sources vs. distance to slew source Green diamonds: Equivalent distribution for random sky positions (same number of sources) Red triangles: Difference in distributions i.e. Blue minus Green Pointing accuracy of Slew survey ~6″ … but long tail…?

27. - 407 Slew source matches with ROSAT BSC & FSC out to 120″ - Distribution of distance offsets shown - Small error (includes RASS error), but long tail…? Correlations with ROSAT

28. Attitude Problem Two types of positional error (1) Real error of ~6″ (2) Error of 0-60″ (mean 30″) but only in slew direction Likely due to the current ATS attitude file having only 1sec time resolution Slew direction AB Dor Also may be cause of patterns in exposure maps ‘Error ellipse’ around source (slew-oriented) Optimistic we can remove error (2) by re-processing the attitude data (RAF?) Error (1) remains, but easily small enough to allow good optical follow-up

29. Correlations with ROSAT - RASS count rate versus XMM Slew count rate for 407 matches - Line shows 10:1 factor - ~45% of Slew sources with <5cts are RASS sources - comparable with whole sample

30. Correlations with ROSAT All Slew sources: Hardness ratio vs Galactic latitude Green circles: No RASS counterpart Red crosses: RASS counterpart Hard sources not seen by ROSAT RASS sources on average softer than non-RASS sources Hardness

31. Correlations with ROSAT All Slew sources in Galactic coordinates Green circles: RASS counterpart Red crosses: No RASS counterpart No obvious trend

32. Survey Characteristics Source Count DistributionExposure Time Distribution Minimum @ ~ 4 countsMaximum @ ~ 10 seconds - Can calculate flux limits for the XMM-Slew survey

33. RASS 5.0x10 -13 (92% of sky) EMSS 3.0x10 -12 (2% 0f sky) HEAO-1 3.6x10 -11 (all-sky) Exosat 5.0x10 -11 (?% of sky) RXTE 1.0x10 -11 (all-sky), but with only 1 ° positional accuracy Flux limits

34. Schedule - Initial processing of slew data (finished) - Creation of images, source searching, rejection of spurious sources (Feb to May) - Cross correlation with ROSAT, other catalogues, DSS, SDSS etc. (Apr to July) - Creation of final catalogue for ingestion into XSA (Aug to Sep) - Ingestion into XSA (end of September) - Investigate extended sources (begin May) - Search co-added slews (TBD) - 2 new (ESAC) trainees, Diego Bermejo, Veronica Lazaro, recently started on project Now

35. Conclusions from current Slew analysis - All available (374) Slews processed (~20% of sky) (Revs 314-785, → 100% of sky by mid-2014) - 54 Slews source-searched to date (using best strategies): 868 sources found in 1800 deg² (~4.5% of sky) ~0.5 sources per square degree - Soft X-ray band detection limit close to ROSAT All-Sky Survey - Hard X-ray band detection limit deepest ever: > 10× deeper than EXOSAT, HEAO-1 ~ 2.5× deeper than RXTE (only has 1° positional accuracy) - XMM-Slew positional accuracy ~6″ Additional ‘attitude-error’ (30″ mean), but only along slew Confident that this attitude-error can be drastically reduced - Aim to recover science from high-background slews - Encountering & solving problems en route – Still on schedule…