A kilometer-scale asteroid inside Venus's orbit
W.-H. Ip, B. T. Bolin, F. J. Masci, Q. Ye, E. A. Kramer, G. Helou, T. Ahumada, M. W. Coughlin, M. J. Graham, R. Walters, K. P. Deshmukh, C. Fremling, Z.-Y. Lin, J. W. Milburn, J. N. Purdum, R. Quimby, D. Bodewits, C.-K. Chang, C.-C. Ngeow, H. Tan, C. Zhai, P. van Dokkum, M. Granvik, Y. Harikane, L. A. Mowla, K. B. Burdge, E. C. Bellm, K. De, S. B. Cenko, C. M. Copperwheat, R. Dekany, D. A. Duev, D. Hale, M. M. Kasliwal, S. R. Kulkarni, T. Kupfer, A. Mahabal, P. J. Mróz, J. D. Neill, R. Riddle, H. Rodriguez, E. Serabyn, R. M. Smith, J. Sollerman, M. T. Soumagnac, J. Southworth, L. Yan
AA kilometer-scale asteroid inside Venus’s orbit
W.-H. Ip ∗ , B. T. Bolin , ∗ , ∗∗ , F. J. Masci , Q. Ye , E. A. Kramer , G. Helou ,T. Ahumada , M. W. Coughlin , , M. J. Graham , R. Walters , K. P. Deshmukh ,C. Fremling , Z.-Y. Lin , J. W. Milburn , J. N. Purdum , R. Quimby , ,D. Bodewits , C.-K. Chang , C.-C. Ngeow , H. Tan , C. Zhai , P. van Dokkum ,M. Granvik , , , Y. Harikane , , L. A. Mowla , K. B. Burdge , E. C. Bellm ,K. De , S. B. Cenko , , C. M. Copperwheat , R. Dekany , D. A. Duev ,D. Hale , M. M. Kasliwal , S. R. Kulkarni , T. Kupfer , A. Mahabal ,P. J. Mr ´oz , J. D. Neill , R. Riddle , H. Rodriguez , E. Serabyn , R. M. Smith ,J. Sollerman , M. T. Soumagnac , , J. Southworth , L. Yan Inst. of Astr., NCU, Taiwan, Div. of Phys., Math. and Astr., CIT, Pasadena, CA, IPAC, CIT, Pasadena, CA, Dept. of Astr., University of Maryland, College Park, MD, JPL, CIT, Pasadena, CA, School of Phys. and Astr., UMN, Minneapolis, MN, Dept. of Eng., IIT Bombay, Powai, India, COO, Pasadena, CA, Dept. of Astr., SDSU, San Diego, CA, Kavli Inst., Univ. of Tokyo, Tokyo, Japan, Phys. Dept., Leach Science Center, Auburn University, Auburn, AL, Astr. Dept., Yale Univ., New Haven, CT, Dept. of Phys., Univ. of Helsinki, Finland Div. of Space Tech., Univ. of Tech. Kiruna, Sweden Dept. of Phys. and Astr., UCL, London, UK, NOAJ, Tokyo, Japan, Dept. of Astr., UW, Seattle, WA, Astrophys. Sci. Div., NASA GSFC, Greenbelt, MD, Joint Space-Science Inst., UMD, College Park, MD, Astrophys. Research Inst., LJMU, Liverpool, UK, Kavli Inst., Univ. of California, Santa Barbara, CA, Dept. of Astr., Stockholm Univ., Stockholm, Sweden, LBNL, Berkeley, CA, Dept. of Particle Phys. and Astrophys., WIS, Rehovot, Israel ∗ These authors contributed equally to this work. ∗∗ To whom correspondence should be addressed; [email protected] a r X i v : . [ a s t r o - ph . E P ] S e p ne-sentence summary The discovery by the Zwicky Transient Facility of the first asteroid interior to the orbit Venus,2020 AV , may imply an additional source of asteroids in the inner Solar System. Abstract
Near-Earth asteroid population models predict the existence of asteroids lo-cated inside the orbit of Venus. However, despite searches up to the end of2019, none have been found. Here we report the discovery by the ZwickyTransient Facility of the first known asteroid located inside of Venus’ orbit,2020 AV , possessing an aphelion distance of 0.65 au and ∼ is the largest of its kind, we find that its discoveryis surprising in the context of population models where the expected count isclose to zero. If this discovery is not a statistical fluke, then 2020 AV may comefrom a yet undiscovered source population of asteroids interior to Venus, andcurrently favored asteroid population models may need to be adjusted. Main Text
Almost all 1 million known asteroids are exterior to Earth’s orbit with only a fraction of apercent located entirely inside its orbit (
1, 2 ). No asteroids have yet been directly observed thatexist entirely within the orbit of Venus despite dynamical models extrapolated from the knownpopulation of asteroids that predict their existence in small numbers ( ). Here we report thefirst discovery of an asteroid interior to the orbit of Venus, 2020 AV ( ), that was first detectedby the Zwicky Transient Facility (ZTF) on the Samuel Oschin Telescope (
7, 8 ) on 2020 January4 in four separate 30 s r -band exposures and moving ∼ was made in the evening twilight sky while it was ∼ )on 2020 January 8 and the Kitt Peak Electron Multiplying CCD Demonstrator (KPED) mountedon the Kitt Peak 84-inch telescope ( ) on 2020 January 9 confirmed this asteroid as having anaphelion distance of ∼ ± ∼ σ significance, confirming the discovery of the firstinner-Venus asteroid (Fig. 2).Spectroscopic observations of 2020 AV made using the Keck telescope on 2020 January 23indicate a reddish surface corresponding to colors of g - r = 0.65 ± r - i = 0.23 ± ∼
900 nm corresponding to i - z = 0.11 ±
11, 12 ) consistent with an origin fromthe inner Main Belt where S-type asteroids are the most plentiful ( ) and in agreement withthe expectations of near-Earth asteroid (NEA) models that predict asteroids with the orbitalelements of 2020 AV (Tab. S1, Fig. S2) should originate from the inner Main Belt (
5, 14 ).Such models also predict that 2020 AV will have a surface reflectivity of ∼ % ( ). Ifso, 2020 AV has a size of ∼ ± < is one of the largest inner-Venus asteroids in the Solar System( ). In addition, dynamical N-body simulations of 2020 AV indicate that its orbit is stable on ∼
10 Myr timescales, entering into temporary resonances with the terrestrial planets and Jupiterbefore its orbit evolves onto close-encounter paths with the gas giant leading to its eventualejection from the Solar System (Fig. S3, see also refs. 18).We estimated the number of inner-Venus asteroids expected to have been discovered by3TF by using synthetic inner-Venus asteroids generated by the NEA population model (Fig. 4,see also ref. 5) combined with ZTF’s completeness at recovering known asteroids during theZTF survey (Figs. S1 and S4). The number of inner-Venus asteroids generated in our Syntheticpopulation within a 68.2 % confidence interval encompassing 2020 AV ’s size is 0.29 ± . . withthe main source of uncertainty being from the estimate on 2020 AV ’s size. The completenessof detecting inner-Venus asteroids is . ± . (Fig. 4), therefore, we expect 0.05 ± . . inner-Venus asteroids to have been discovered during our observations. Despite its low probability, apossible explanation for our detection of 2020 AV is a random chance discovery from the near-Earth asteroid population. However, history has shown that the first detection of a new classof objects is usually indicative of another source population c.f., such as the Kuiper Belt withthe discovery of the first Kuiper Belt Objects 1992 QB and 1993 FW ( ). Therefore, 2020AV could have originated from a source of asteroids located closer to the Sun, such as near thestability regions located inside the orbit of Mercury at ∼
20, 21 ). Acknowledgements
Based on observations obtained with the Samuel Oschin Telescope 48-inch and the 60-inchTelescope at the Palomar Observatory as part of the Zwicky Transient Facility project. ZTFis supported by the National Science Foundation under Grant No. AST-1440341 and a collab-oration including Caltech, IPAC, the Weizmann Institute for Science, the Oskar Klein Centerat Stockholm University, the University of Maryland, the University of Washington, DeutschesElektronen-Synchrotron and Humboldt University, Los Alamos National Laboratories, the TANGOConsortium of Taiwan, the University of Wisconsin at Milwaukee, and Lawrence Berkeley Na-tional Laboratories. Operations are conducted by COO, IPAC, and UW.SED Machine is based upon work supported by the National Science Foundation under4rant No. 1106171The KPED team thanks the National Science Foundation and the National Optical Astro-nomical Observatory for making the Kitt Peak 2.1-m telescope available. We thank the observa-tory staff at Kitt Peak for their efforts to assist Robo-AO KP operations. The KPED team thanksthe National Science Foundation, the National Optical Astronomical Observatory, the CaltechSpace Innovation Council and the Murty family for support in the building and operation ofKPED. In addition, they thank the CHIMERA project for use of the Electron Multiplying CCD(EMCCD).The authors would like to thank Alessandro Morbidelli for useful discussion in the inter-pretation of the first inner-Venus asteroid discovery as well as providing the synthetic asteroidpopulation used to model our survey efficiencies.M. W. Coughlin acknowledges support from the National Science Foundation with grantnumber PHY-2010970.C.F. gratefully acknowledges the support of his research by the Heising-Simons Foundation(
Gaia ( ), processed by the Gaia
Data Processing andAnalysis Consortium (DPAC, onsortium ). Funding for the DPAC has been provided by national institutions, in particularthe institutions participating in the Gaia
Multilateral Agreement.
Authors Contributions
W.-H.I. helped initiate, design and secure P48 time for the ZTF Twilight. B.T.B. discovered2020 AV in the ZTF data and realized it had an orbit interior to Venus’, led the study, wrote themanuscript, reviewed and scanned all candidate asteroid detections for the duration of this study,helped design and secure time for the Twilight survey and for follow-up, prepared and executedfollow-up observations, reduced the photometric and spectroscopic data, did the astrometry,orbit determination, helped generate the synthetic NEA population, calculated the survey com-pleteness and estimates of the inner-Venus asteroid population. F.J.M. served as the ScienceData System Lead of the ZTF collaboration, wrote and maintained software for extraction ofmoving objects in ZTF data, calculated the survey completeness, maintained the ZTF data sys-tem, helped secure P48 time for the ZTF Twilight survey, helped design and initiate the Twilightsurvey. Q.Y. contributed code used to schedule the ZTF survey and contributed improvementsto the moving object identification pipeline, and helped initiate and design the Twilight sur-vey. E.K. monitored the NEA discovery performance of the ZTF survey. G.H. helped withthe interpretation of the completeness calculation, helped prepare the manuscript, served as thePrimary Investigator of the team observing NEAs with ZTF and as the Data Archive Directorof the ZTF collaboration, helped design the cadence, secure P48 time for the ZTF Twilightsurvey and is one of its initiators. E.K. monitored the NEA discovery performance of the ZTFsurvey. G.H. served as the Primary Investigator of the team observing NEAs with ZTF andas the Data Archive Director of the ZTF collaboration, helped design the cadence, secure P48time for the ZTF Twilight survey and is one of its initiators. E.K. monitored the NEA discoveryperformance of the ZTF survey. G.H. served as the Primary Investigator of the team observing6EAs with ZTF and as the Data Archive Director of the ZTF collaboration, helped design thecadence, secure P48 time for the ZTF Twilight survey and is one of its initiators. T.A. and M.C.performed follow-up observations of 2020 AV with KPED and contributed to the reduction ofphotometry. M.G. served as the Project Scientist of ZTF, secured time with Keck I and tookspectroscopic observations of 2020 AV with Keck I/LRIS and reduced spectral data. R.W.performed follow-up observations of 2020 AV with SEDM. K.P.D. contributed to planningfollow-up observations of 2020 AV . C.F. contributed to the planning of spectroscopy obser-vations of 2020 AV with Keck and spectroscopic data reduction. Z.-Y.L. contributed to theinterpretation of the 2020 AV spectroscopy data. J.W.M. contributed to follow-up observationsof 2020 AV . J.P. and R.Q. contributed to the reduction of photometry and follow-up obser-vations of 2020 AV . C.-K.C., C.-C.N. and H.T. contributed to securing time with the P48 forZTF Solar System observations. C.Z., C.M.C and J.S. were members of the team that conductsfollow-up observations of ZTF Solar System discoveries. P.D. and L.A.M. helped secure timefor follow-up observations of 2020 AV . M.G. helped generate the synthetic NEA population.K.B.B. and K.D. contributed to the reduction of Keck/LRIS spectroscopy data of 2020 AV .E.C.B. served as the Survey Scientist of the ZTF collaboration and scheduled the observationstaken with the P48 during the ZTF Twilight survey. S.B.C. contributed to the planning of P48time used in the Twilight survey. D.A.D, A.M. and P.J.M. wrote and maintained software andmachine learning algorithms for the ZTF data pipeline and identification of sources in ZTFdata. A.M. served as the Machine Learning Lead for the ZTF collaboration. R.D. served asthe Project Manager of the ZTF collaboration and contributed to the design and constructionof the ZTF camera. D.H. contributed to the design and construction of the ZTF camera. M.K.served as the Primary Investigator of the Global Relay of Observatories Watching TransientsHappen (GROWTH) team for follow-up of ZTF discoveries. S.R.K. served as the PrincipalInvestigator of the ZTF collaboration and survey. T.K. and M.T.S. contributed to the ZTF cali-7ration pipeline. J.D.N. contributed to SEDM operations and observations of 2020 AV . R.R.contributed to maintaining ZTF operations. H.R. contributed to SEDM operations. E.S. con-tributed to KPED operations and observations of 2020 AV . R.M.S. served as the Lead CameraEngineer for the ZTF survey and lead the design and construction of the ZTF camera. J.S. con-tributed to SEDM operations and observations of 2020 AV . L.Y. contributed to planning andsecuring of telescope time for follow-up observations of 2020 AV .8igure 1: (A) Discovery 30 s r-band image of 2020 AV taken on 2020 January 4 UTC where2020 AV is the detection located in the circle. (B) Composite image containing the four discov-ery 30 s r-band exposures covering 2020 AV made by stack on the rest frame of the backgroundstarsover a 22 minute time interval. The first detection has been labeled. The asteroid was mov-ing ∼ ∼
15 arcseconds spacing between the detections of 2020 AV . The cardinal directions andspatial scale are indicated for reference. 9igure 2: Orbital configuration of 2020 AV , Earth, Venus and Mercury at the time of 2020AV ’s discovery looking from above the orbital plane of the inner Solar System. The orbitplotted for 2020 AV in this figure is taken from the orbital elements in Tab. S1. The periheliondirections of 2020 AV and the planets are plotted with dotted lines. The aphelion directions of2020 AV and the planets are plotted with solid straight lines.10igure 3: Visible wavelength reflectance spectrum taken of 2020 AV with the LRIS instru-ment on Keck I on 2020 January 23 plotted as blue dots. The error bars on the spectrum datapoints correspond to 1- σ uncertainty. The spectrum has been normalized to unity at 550 nmindicated by the black cross. The spectrum presented was obtained by combining two spectrafrom the blue camera using the 600/4000 grating and the red camera using the 400/8500 gratingwith a 560 nm dichroic (
22, 23 ). The data have been rebinned and smoothed by a factor of10 using an error-weighted mean. The spectral range of S, V and C-type asteroids from theBus-DeMeo asteroid taxonomic catalogue ( ) are over-plotted with the S-type spectrum mostclosely resembling the spectra of 2020 AV . 11igure 4: Comparison between the (A) semi-major axis, a , (B) eccentricity, e , (C) inclination i and (D) absolute magnitude H distribution of the number of synthetic inner-Venus asteroidsgenerated from the NEA model ( ) (grey histograms) and the completeness of synthetic inner-Venus asteroids detected in the survey simulation (red histogram). The 1- σ error bars on thecompleteness are determined assuming Poissonian statistics. The vertical dashed black lineindicates the value of the element for 2020 AV from Tab. S1. The absolute magnitude rangeof 15 to 18 corresponds to asteroids in the size range of ∼ % surfacereflectivity. The number of objects from the NEA model has been oversampled by a factor of1,000. 12 eferences and Notes
1. R. P. Binzel, V. Reddy, T. L. Dunn,
The Near-Earth Object Population: Connections toComets, Main-Belt Asteroids, and Meteorites (2015), pp. 243–256.2. R. Jedicke, et al. , Surveys, Astrometric Follow-Up, and Population Statistics (2015), pp.795–813.3. G. W. Wetherill,
Icarus , 96 (1979).4. P. Michel, V. Zappal`a, A. Cellino, P. Tanga, Icarus , 421 (2000).5. M. Granvik, et al. , Icarus , 181 (2018).6. B. T. Bolin, et al. , Minor Planet Electronic Circulars (2020).7. E. C. Bellm, et al. , Publications of the Astronomical Society of the Pacific , 018002(2019).8. M. J. Graham, et al. , Publications of the Astronomical Society of the Pacific , 078001(2019).9. N. Blagorodnova, et al. , Publications of the Astronomical Society of the Pacific , 035003(2018).10. M. W. Coughlin, et al. , Monthly Notices of the Royal Astronomical Society , 1412(2019).11. S. J. Bus, R. P. Binzel,
Icarus , 146 (2002).12. F. E. DeMeo, R. P. Binzel, S. M. Slivan, S. J. Bus,
Icarus , 160 (2009).13. F. E. DeMeo, B. Carry,
Nature , 629 (2014).134. M. Granvik, et al. , Nature , 303 (2016).15. F. E. DeMeo, B. Carry,
Icarus , 723 (2013).16. R. P. Binzel, et al. , Icarus , 41 (2019).17. A. Morbidelli, et al. , Icarus , 113631 (2020).18. S. Greenstreet,
Monthly Notices of the Royal Astronomical Society (2020).19. D. Jewitt, J. Luu,
Nature , 730 (1993).20. N. W. Evans, S. Tabachnik,
Nature , 41 (1999).21. A. Shannon, A. P. Jackson, D. Veras, M. Wyatt,
Monthly Notices of the Royal AstronomicalSociety , 2059 (2015).22. J. B. Oke, et al. , Publications of the Astronomical Society of the Pacific , 375 (1995).23. J. K. McCarthy, et al. , Blue channel of the Keck low-resolution imaging spectrometer (1998), vol. 3355 of
Society of Photo-Optical Instrumentation Engineers (SPIE) Confer-ence Series , pp. 81–92.24. S. Keys, et al. , Publications of the Astronomical Society of the Pacific , 064501 (2019).25. E. C. Bellm, et al. , Publications of the Astronomical Society of the Pacific , 068003(2019).26. F. J. Masci, et al. , Publications of the Astronomical Society of the Pacific , 018003(2019).27. D. A. Duev, et al. , Monthly Notices of the Royal Astronomical Society , 4158 (2019).28. Q. Ye, et al. , Astronomical Journal , 70 (2020).149. R. Dekany, et al. , Proceedings of SPIE (2016), vol. 9908 of
Society of Photo-Optical In-strumentation Engineers (SPIE) Conference Series , p. 99085M.30. Q. Ye, et al. , Publications of the Astronomical Society of the Pacific (2019).31. M. Rigault, et al. , Astronomy & Astrophysics , A115 (2019).32. D. A. Perley,
Publications of the Astronomical Society of the Pacific , 084503 (2019).33. Gaia Collaboration, et al. , Astronomy & Astrophysics , A1 (2016).34. Gaia Collaboration, et al. , Astronomy & Astrophysics , A1 (2018).35. H. Raab, Astrometrica: Astrometric data reduction of CCD images (2012).36. K. C. Chambers, et al. , ArXiv e-prints (2016).37. H. A. Flewelling, et al. , ArXiv e-prints (2016).38. H. Rein, S.-F. Liu,
Astronomy & Astrophysics , A128 (2012).39. H. Rein, D. S. Spiegel,
Monthly Notices of the Royal Astronomical Society , 1424(2015).40. D. Nesvorn´y, F. Roig, S. Ferraz-Mello,
Astronomical Journal , 953 (2000).41. C. de la Fuente Marcos, R. de la Fuente Marcos,
Monthly Notices of the Royal AstronomicalSociety , L6 (2020).42. M. Granvik, et al. , Astronomy & Astrophysics , A52 (2017).43. B. T. Bolin, M. Delbo, A. Morbidelli, K. J. Walsh,
Icarus , 290 (2017).154. A. Morbidelli, R. Gonczi, C. Froeschle, P. Farinella,
Astronomy & Astrophysics , 955(1994).45. A. Milani, G. F. Gronchi,
Theory of Orbital Determination (Cambridge University Press,2010).46. J. Wisdom,
Icarus , 51 (1983).47. E. Bowell, et al. , Asteroids II pp. 399–433 (1988).48. P. Vereˇs, et al. , Icarus , 34 (2015).49. R. Jedicke, B. Bolin, M. Granvik, E. Beshore,
Icarus , 173 (2016).50. S. P. Naidu, S. R. Chesley, D. Farnocchia,
AAS/Division for Planetary Sciences MeetingAbstracts , AAS/Division for Planetary Sciences Meeting Abstracts (2017), p. 112.04.16 ist of Supplementary Materials
MethodsDiscovery, follow-up and orbital determinationTwilight Survey strategyDiscovery, follow-up and characterization observationsVisual imaging/spectroscopy reduction and astrometryDynamical evolutionComparison with the NEA populationEstimating the ZTF inner-Venus asteroid population completenessFigs. S1-S4Tab. 1References (24-50)
Supplementary Material
Methods
Discovery, follow-up and orbital determination:
The initial discovery observations of 2020AV on 2020 January 4 were made in four sidereally-tracked 30 s r-band exposures by the48-inch Samuel Oschin Telescope. The observations were made during evening astronomicaltwilight while the telescope was pointing at 25.5 degrees elevation and the center of the tele-scope’s field of view was pointing through 2.3 airmasses. The brightness of 2020 AV during itsdiscovery observations were r ∼ having FWHM ∼ during the observations was . ”/min resulting in no significant trailing losses ofthe individual detections of 2020 AV in the discovery images (Fig. 1, A-B).The preliminary orbit of 2020 AV determined from the discovery observations was firstclassified as an Apollo-like orbit based on the initial discovery observations with a ∼ e ∼ ). Af-ter including the SEDM and KPED observations made on 2020 January 8 and 9, the semi-major axis became a ∼ ± e ∼ ± Q ∼ ± Q = 0.654 ± > σ confidence interval smaller than the 0.72 au periheliondistance of Venus (Supplementary Material Tab. S1 and Fig. S2). No pre-discovery observa-tions of 2020 AV were located in the ZTF archive. Twilight Survey strategy:
Astronomical twilight time is used by ZTF to search for Solar Sys-tem objects at small Solar elongations <
60 degrees from the Sun in a program called the “Twi-light Survey” during which 2020 AV was discovered. Searching for Solar System objects soclose to the Sun during twilight is accomplished by using the 47 sq. degree field of view andminimum elevation of the 48-inch Samuel Oschin Telescope to point the telescope as close tothe horizon as possible during evening and morning Astronomical twilight when the Sun is ∼ ). This results in being able to search the night sky to as close as ∼
35 degrees from the Sun, corresponding to potentially detecting objects as close as ∼ ∼
470 square degrees of sky coverage per Twilight Survey session and covering an area of thesky with ∼
20 degrees elevation and airmass < ). Each Twilight Survey session lasts 20-25 minutes whereeach field is image four times with 30 s exposures in r -band, resulting in a ∼ ∼ >
10 degrees/day by the ZTF processingsystem (
26, 27 ). The limiting magnitude in each of the 30 s r -band exposures is degraded dueto the extinction at its higher airmass observations and higher sky background resulting in alimiting magnitude closer to V ∼ V ∼
7, 8 ).The Twilight Survey that our results are based on began on 2019 September 20 and wasbased on an earlier version that ran during Winter 2018 and Summer 2019 ( ). The currentTwilight Survey is ongoing, but poor weather in February and March limits this analysis tosurvey dates between 2019 September 19 and 2020 January 30. The current Twilight Surveyexecutes its 10 field coverage each night, weather-permitting, alternating between evening andmorning twilight. The preliminary version of the Twilight Survey ran on a more sporadic ca-dence where it only operated on a 3-day cadence alternating between evening and morningtwilight. We sought to improve our survey coverage and rate of self-recovery of candidate dis-coveries by moving to the current every night cadence. In total, the Twilight Survey was carriedout 90 times between 2019 September 20 and 2020 January 30 split into 47 mornings and 43evenings. The complete Sun-centered sky coverage of these 100 observing sessions is presentedin Fig. S1. In total, ∼ Discovery, follow-up and characterization observations:
Zwicky Transient Facility, ZTF:
The ZTF camera consists of 16 separate 6144 x 6160-pixelarrays on a single CCD camera mounted on the 48-inch Samuel Oschin Telescope at PalomarObservatory and is robotically operated. The plate scale of the camera is 1.01 arcseconds/pixeland has a 7.4-degree x 7.4-degree field of view (
7, 29 ). The data processing pipeline producedimages differenced from reference frames and removes or masks most detector artifacts. Tran-19ient are extracted from the images and several algorithms are used to identify slower movingobjects that appear as round PSF detections in the images ( ) and to extract fast-moving objectsthat appear as streaked detections (
27, 30 ). Moving objects can be identified in images taken ineach of its g , r and i band filters. For the purpose of the Twilight Survey, 30 s exposures wereused with the r filter. Seeing was measured to be ∼ in the discovery images. Spectral Energy Distribution Machine, SEDM:
Observations taken by the SEDM used the Rain-bow Camera consisting of two identical Princeton Instruments Pixis 2048B eXelon model2048 x 2048 pixel CCDs mounted on the Palomar 60-inch telescope and is robotically oper-ated (
9, 31 ). The Rainbow Camera has a 13 arcminute x 13 arcminute field of view dividedinto four ∼ u , g , r and i anda 0.125 arcseconds/pixel spatial scale. Only the r filter quadrant was used for our follow-upobservations with SEDM with non-sidereally tracked 30 s exposures. Seeing conditions were ∼ Kitt Peak Electron Multiplying CCD Demonstrator, KPED:
The KPED instrument is mountedon the Kitt Peak 84-inch telescope and consists of a 1024 x 1024 pixel Electron MultiplyingCCD camera and is robotically operated ( ). The camera has a spatial scale of 0.26 arcsec-onds/pixel and a 4.4 arcminute x 4.4 arcminute field of view. The camera is capable of readingout at a rate of 1 Hz and of individual exposures times up to 10 s. Our observations used 10 sexposures in r -band and were sidereally tracked due to the short exposure time. Seeing condi-tions were ∼ Keck I Telescope:
The Low Resolution Imaging Spectrometer (LRIS) ( ) on the Keck I tele-scope was used to observe 2020 AV on 2020 January 23 in spectroscopy mode (Program IDC272, PI M. Graham). Both the blue camera consisting of a 2 x 2K x 4K Marconi CCD ar-ray and the red camera consisting of a science grade Lawrence Berkeley National Laboratory20K x 4K CCD array were used. Both cameras have a spatial resolution of 0.135 arcsec/pixel.The 1.0-arcsecond wide slit was used with the 560 nm dichroic with ∼ % transmission effi-ciency in combination with the 600/4000 grating for the blue camera and the 400/8500 gratingfor the red camera providing a spectral resolution of 0.4 nm and 0.7 nm, respectively (
22, 23 ).A total exposure time of 600 s over two integrations were taken in seeing conditions of ∼ ∼ ∼ Visual imaging/spectroscopy reduction and astrometry:
All visual imaging data was re-duced using custom code for bias and flat field detrending. The LRIS spectroscopic data werereduced using flat field, dark current and arc lamp exposures with the LPipe spectroscopy re-duction software ( ). The Gaia data release 2 catalog (
33, 34 ) was used with the ZTF datareduction pipeline ( ) to produce an astrometric solution on ZTF data and with the Astromet-rica software ( ). Photometric calibration was performed using the Pan-STARRS1 catalogdatabase (
36, 37 ). Dynamical evolution:
We used the rebound
N-body orbit integration package ( ) using the IAS15 adaptive time step integrator ( ) to determine the orbital history of 2020 AV . Usingthe multi-variate distribution of the orbital parameters presented in Tab. S1 and Fig. S1, weintegrated several 100 clone orbits of 2020 AV forwards and backward 30 Myrs with a nom-inal timestep of 14 h. Because it is an adaptive time-step integrator, IAS15 will decrease thetime of this time step during close encounters to avoid discrepancies in the orbits of test par-ticles resulting from too coarse time steps. We find that 2020 AV has experienced numerous,21 is its occasionalcapture into mean motion resonances with Venus such as the 3:2 mean motion resonance at ∼ ). At the precision of the current orbit, 2020 AV will remain in resonance withVenus for ∼ during its next apparition inthe fall of 2020 may improve the orbital accuracy to allow for longer-term investigation of thisresonant behavior (
18, 41 ). In any case, the orbit of 2020 AV is firmly confined within the orbitof Venus with having an aphelion distance of < prevents predicting its orbital behav-ior on timescales exceeding a few 10 Myrs, it is apparent from its orbital evolution that it isa transitory inhabitant of the inner Venus region of the Solar System. The majority of orbitalclones have several lunar distances encounters with Venus and the Earth within 10-20 Myr thathave the effect of exciting their orbits onto trajectories crossing with the orbit of Jupiter. Thecrossing of the orbit of Jupiter has the effect of exciting test particles’ orbits onto hyperbolictrajectories after a close encounter with the giant planet sending them out of the Solar Systemas seen for one of the orbital clones in Fig. S3 (C). The majority of the clones are ejected fromthe Solar System during close encounters with Jupiter after ∼
20 Myr consistent with the mean ∼
10 Myr lifetimes of near-Earth objects originating from the Main Belt (
5, 17 ) and independent22ntegrations by Greenstreet 2020.
Comparison with the NEA population:
One of the dynamical pathways for inner-Venus as-teroids is to originate from the Main Asteroid belt through source regions located near variousmajor planetary resonances ( ). If we assume that 2020 AV originated from the Main Beltas an asteroid family fragment ( ) before crossing inside of the orbit of Venus, asteroids withorbits similar to 2020 AV according to the Granvik et al. (2018) NEA model most likely origi-nate with a ∼ % probability from the ν resonance that forms the boundary of the inner MainBelt at 2.2 au ( ). The second most likely source of 2020 AV with a ∼ % probability are theHungaria asteroid population located just exterior to the Main Belt at 2.0 au ( ) and the thirdmost likely at ∼ % being the 3:1 mean motion resonance with Jupiter located in the Main Beltat 2.5 au ( ).Given its dynamical preference for originating from the inner Main belt, it is no surprise that2020 AV has a S-type spectrum as seen in Fig. 3 since the majority of asteroids located in theinner Main Belt have spectra similar to S-types ( ). We can take the spectral and dynamicalclassification one step further by comparing its source region probability with the NEA albedomodel of Morbidelli et al. (2020) to estimate its surface reflectivity as an independent constrainton its spectral type. Comparing our source region probabilities with the NEA albedo model, itis likely that 2020 AV has an albedo of ∼ ). Combining our albedo estimate from the NEA albedo model with our estimateof the absolute magnitude of H = 16.4 ± in the H - G system ( ) from ourdetermination of its orbit, we measure a diameter of ∼ ± ). Our uncertainties on the diameter of 2020 AV are conservatively estimated by assuming a 1- σ uncertainty of ∼ ∼ H of asteroids from the Minor Planet Center catalog ( ), roughlytaking into account the uncertainty caused by the unknown phase function of 2020 AV and the23arge ∼ ◦ phase angle the asteroid was observed at in 2020 January ( ). In addition, theNEA model predicts that there are ∼ 18 of which ∼ % or ∼ ). This number of inner-Venusasteroids predicted by the NEA model shrinks to 0.29 ± . . when we only consider asteroidsbrighter than in the 1- σ range of the H of 2020 AV , < H = 16.4 ± Estimating the ZTF inner-Venus asteroid population completeness: We generate a syntheticpopulation of inner-Venus asteroids to compare with our discovery of 2020 AV with the pre-dicted inner-Venus asteroid population from the NEA model ( ). We oversample a mediumresolution version of the NEA model with d a = 0.05 au, d e = 0.02, d i = 2.0 ◦ , d H = 0.25 by afactor of 1,000 generating 1,168,279 asteroids with 15 18 of which 2,521 are inner-Venusasteroids by definition with an aphelion distance < a , e , i and H distributions ofour synthetically generated inner-Venus asteroids from the NEA population models are plottedin Fig. 4 (A-D).To simulate the discovery and observations of our synthetic inner-Venus asteroids by ZTF,we use the complete list of ZTF telescope pointings used during the Twilight Survey between2019 September 20 UTC and 2020 January 30 UTC plotted in Fig. S2 with a survey simula-tor ( ). We took into account trailing-losses on objects’ apparent brightnesses in the surveysimulation. The output of the survey simulator produces a list of synthetic inner-Venus asteroidsdetections from the survey simulation that can be used to roughly estimate the completeness ofthe ZTF survey in detecting inner-Venus asteroids. We refine our synthetic inner-Venus aster-oid observation estimates by calculating completeness for each synthetic inner-Venus asteroiddetection in each field. The per object per field completeness is estimated by comparing actualdetections of known moving objects serendipitously observed in the Twilight Survey fields withthe predicted number of known objects detected in the fields. The Twilight Survey per object24ompleteness as a function of V magnitude is presented in Fig. S4 (C). A function of the form (cid:15) ( V ) = (cid:15) (cid:20) V − V lim V width (cid:21) − (1)is used to interpolate the per-field completeness with (cid:15) = 0.87 representing the maximum pos-sible completeness for detecting moving objects, V lim = 20.60 mag, representing the limitingmagnitude of the survey where the completeness drops to half for detecting moving objectsand V width = 0.74 representing the width of the transition in the drop of the completeness indetection faint moving objects ( ) and is plotted as a red line in Fig. S4 (C). We note that thevalue of (cid:15) = 0.87 is remarkably close to the fill factor of ZTF ( ) suggesting that the limitingfactor in detecting bright moving objects by ZTF is the detector layout rather than the ZTFprocessing pipeline ( ). In addition to the limiting magnitude, trailing losses due to the skyplane motion of the inner-Venus asteroids could further decrease the completeness calculatedwith Eq. (1) ( ). However, as seen in Fig. S4 (B), the vast majority of synthetic inner-Venusasteroid detections have a sky plane rate of motion of 1.6 degrees/day or slower which doesn’tresult in significant trailing of the detections given the typical 2 arcseconds seeing at the Palo-mar observing site. The lack of preference for slower-moving objects seen in the rate of motiondistribution of the detected number of objects plotted in red in Fig. S4 (B) compared to therate of motion distribution of synthetic inner-Venus asteroids suggests that losses due to trailingwithin the inner-Venus asteroid population are negligible.The per inner-Venus asteroid, j , per session, n , completeness, (cid:15) j,n ( V j,n ) is given by itsper session field visible magnitude, V j,n using equation Eq. (1). If a synthetic inner-Venusasteroid is not seen > n , (cid:15) j,n ( V j,n ) = 0 . Thevast majority of synthetic inner-Venus asteroids have V magnitude < 20 as seen in Fig. S4 (A)resulting in the majority of inner-Venus asteroid detections located in Twilight Survey fieldshaving (cid:15) j,n ( V j,n ) > % . 25he probability of detecting a single Synthetic inner-Venus asteroid, j , p j ,over the total n max Twilight Survey sessions is given by the following equation p j = 1 − n max (cid:89) n =0 [1 − (cid:15) j,n ( V j,n )] (2)where n max = 89 corresponding to 90 individual Twilight Survey sessions. The number ofdetected synthetic inner-Venus asteroids is weighted per synthetic inner-Venus asteroid a , e , i and H bin by each detected object’s p j . The completeness of synthetic inner-Venus aster-oids detected by the synthetic ZTF survey per synthetic inner-Venus asteroid a , e , i and H binis calculated by dividing the weighted number of synthetic inner-Venus asteroids detected persynthetic inner-Venus asteroid a , e , i and H bin by the total number of synthetic inner-Venus as-teroids generated from the NEA model per synthetic inner-Venus asteroid a , e , i and H bin. Thecompleteness of inner-Venus asteroids detected per a , e , i and H bin is plotted with a red line inFig. 4 (A-D). Marginalizing over the complete a , e , i and H distribution of the synthetic inner-Venus asteroid population results in an integrated completeness of ∼ ∼ % confidence interval encompass-ing 2020 AV ’s H magnitude. 26igure S1: Skyplane distribution of Twilight Survey coverage occurring between 2019 Septem-ber 19 UTC and 2020 January 30 UTC. The star plots the discovery locations of inner-Venusasteroid 2020 AV . The color scale is the number of ZTF visits per square degree as a functionof ecliptic longitude and latitude. 27able S1: Orbital elements of 2020 AV based on observations collected between 2020 January4-23 UTC. The orbital elements are shown for the Julian date (JD) shown using the software Find Orb by Bill Gray. The 1- σ uncertainties are given in parentheses. The value and 1- σ uncertainties for H is provided by the JPL Small-Body Database Browser entry for 2020 AV .ElementEpoch (JD) 2,458,871.5Time of perihelion, T p (JD) 2,458,907.045 ± (0.019)Semi-major axis, a (au) 0.55536 ± (8.51x10 − )Eccentricity, e ± (0.000222)Perihelion, q (au) 0.456893 ± (0.000192)Aphelion, Q (au) 0.653817 ± (0.000825)Inclination, i ( ◦ ) 15.880 ± (0.006)Ascending node, Ω ( ◦ ) 6.7065 ± (0.0032)Argument of perihelion, ω ( ◦ ) 187.305 ± (0.017)Mean Anomaly, M ( ◦ ) 275.350 ± (0.019)Absolute Magnitude, H ± (0.8)28igure S2: Corner plot of 30,000 samples from the multivariate distribution of orbital elementsof 2020 AV from the covariance obtained with the orbit fit from observations between 2020January 4-23 UTC. The central value and the 1- σ uncertainty for each parameter value is givenat the top of each column in the corner plot. 29igure S3: (A) Evolution of the semi-major axis of 2020 AV . The plateaus in the evolution ofthe semi-major axis separated by jumps are due to 2020 AV crossing resonances with Venusand Jupiter. At around 0.05 Myrs, 2020 AV entered a 3:2 mean motion resonance with Venusthat lasts for a few 0.1 Myrs. (B) The evolution of the aphelion (orange) and perihelion (blue)distances of 2020 AV integrated to ± 10 Myrs. The current aphelion (dashed line) and peri-helion distances (dash-dot line) are plotted as horizontal lines for Venus (cyan) and Mercury(red) and Earth (purple). The aphelion distance of 2020 AV spends the majority of the simu-lation within the perihelion distance of Venus (0.718 au) and the perihelion distance less thanthe aphelion distance of Mercury (0.467 au) for most of the simulation and crosses Mercury’sperihelion distance (0.307 au) at Myrs during the simulation having a perihelion of ∼ ∼ ∼ 22 Myrs andsubsequent perturbations from the other planets results in 2020 AV eventually increasing in itsaphelion distance until it encounters Jupiter and is ejected from the Solar System at ∼ 28 Myrs.The majority of the clones of 2020 AV are lost in a similar manner after 10-20 Myrs.30igure S4: (A) Comparison of the synthetic inner-Venus asteroid apparent V magnitude dis-tribution with the weighted V magnitude distribution of detected inner-Venus asteroids in thesurvey simulation. Detections are weighted using Eqs. (1) and (2). (B) Same as (A), but forthe synthetic inner-Venus asteroid’s rate of motion. The vertical dashed lines in (A) and (B)are the values of the apparent V magnitude and rate of motion of 2020 AV on 2020 January 4UTC. (C) Detection efficiency as a function of V magnitude. The 1- σ error bars are determinedassuming Poissonian statistics. Eq. (1) using (cid:15) = 0.87, V lim = 20.60, V widthwidth