Results From Testing of a Monolithic Interferometer with Keck ET at APO and Plans for the Future: Towards ET DFDI Long-Term Stability and High Precision Fleming1, S.W., Ge1, J., Wan1, X., Lee1, B., Mahadevan1, S., Kane1, S.R., Groot1, J., van Eyken1, J. 1. ABSTRACT We present engineering results using a monolithic interferometer at the Keck ET instrument at Apache Point Observatory. Testing demonstrated ~15 m/s differential drift for our best spectra pair over ~6 days and ~0.1 degree temperature stability. This interferometer was not temperature compensated and not located in a vacuum. These results indicate use of a final interferometer design using CaF2 spacers, temperature compensation and a sealed enclosure will enable an ~30% flux increase and also provide long-term stability to the ET instrument at KPNO by removing the actively-controlled PZT movable mirror. The design goal is to achieve 0.01 degree temperature stability around the entire instrument enclosure. This should result in an interferometer drift (thermal component only) of ~1 m/s. Our goal is an instrument precision of ~0.5-1 m/s and a stability of ~1 m/s. Combined with a planned upgrade of the dispersing element to R ~ 20,000 and expanded wavelength coverage from 0.38-0.65 μm, the KPNO ET instrument (that will then be known as EXPERT) will serve as an excellent follow-up instrument to APO long-period candidates as well as a research vessel for long-period binaries, brown dwarfs, intermediate-period planets, multiple-planet systems or planets with masses comparable to ice giants and super-Earths. A second, similar instrument may be built at the Lijiang 2.4 m telescope at Yunnan Observatory, China providing multi-longitudinal, high-precision RV measurement capability via collaboration. 3. NEW RESULTS FROM APO During Aug. 2007 we tested a new prototype monolithic interferometer using our Keck ET multi-object instrument at Apache Point Observatory (APO), New Mexico, USA. The interferometer was not thermally compensated but was housed in a specially constructed enclosure (Fig. 4). Thermal probes measured the temperature drift above and below the enclosure seen in Fig. 4 (but not inside the enclosure where the actual interferometer was housed). Temperature stability over 6 days was ~0.1ºC (Fig. 5.). Using calibration lamp images (Tungsten light through our Iodine cell) we traced the fringe drift over these 6 days to estimate stability of this new monolithic interferometer. Results from one fiber-pair are shown in Fig. Fig. 4: The passive interferometer used at APO for testing, 6. The fibers track one another with an RMS of ~15 m/s. Based on these results, the next-gen version of the monolithic seen here in its metallic housing. One window is viewable in this photo (left side). This interferometer was an interferometer that will include temperature compensation advanced prototype and was not thermally compensated or and high-quality environmental control should achieve our vacuum-sealed. The final, next-gen version will be design goal of 1 m/s thermal RV drift. Based on the optical thermally compensated and housed under superior stability design calculations, temperature compensation will improve conditions. stability by at least a factor of 25 (even one order of magnitude converts 15 m/s to ~1.5 m/s). In addition, the radial velocity stability is approximately linear with temperature stability, so improving our thermal stability can further improve the instrumental thermal drift. The optical design of our next-gen monolithic interferometer predicts ~200 m/s per ºC thermal drift. When housed in a thermally-controlled environment (e.g., vacuum, sealed enclosure, etc.) with 0.01 ºC stability our design goal of 1 m/s instrumental stability can be achieved. Fig. 8 shows the measured absolute velocity from one particular spectrum and the theoretical value if dominated soley by thermal drift (red line). The agreement is excellent, suggesting that the intrumental drift is dominated (at least to a level of ~100 m/s or less) by the thermal component. 2. PREVIOUS RESULTS – KPNO 2005 As a first test using a passive interferometer, a prototype interferometer was built using a Newport broad-band non-polarizing beamsplitter, BK7 glass window from Edmund Optics and mirrors from Thorlabs. A copper ring was used as a spacer to create a delay in one arm of the interferometer. The interferometer was not housed in a specially designed, environmentally-controlled environment such as a vacuum or high-quality sealed room, although the current Kitt Peak instrument room does have temperature stability of ~0.2° C. Observations were conducted from May 27 to June 6, 2005 using the 2.1 meter telescope at Kitt Peak National Observatory (KPNO), Arizona, USA. In addition to science targets, reference stars typically used during observations with the actively-controlled interferometer were included as a feasibility study to determine if a passive interferometer design is a viable option. Results are shown below (Fig. 1-3). Fig. 5: Temperature stability over 6 days. Stability is ~0.1°C over the 6 days outside the interferometer enclosure. The oscillations are caused by a known day-night cycle effect and will be improved upon in the final MARVELS instrument enclosure. Current KPNO longterm temperature stability is 0.2-0.3 °C. Fig. 1: Observations of the RV-stable star 36 Uma. Estimated RMS over ~8 days is 10.0 m/s. Fig. 2: Observations of the planet-bearing star 55 Cnc. RMS over ~8 days is 12.5 m/s. Planet orbit taken from Exoplanets Encyclopedia1 and incorporates a 4-planet non-interacting model. 4. The EXPERT Instrument The MARVELS survey will uncover many intermediate and long-period brown dwarf and extrasolar planet candidates that will require follow-up observations and long-term monitoring. Fig. 7 (left) shows a few intermediate-period companions from the pilot program using a MARVELS prototype instrument from 2006-2007 that require additional monitoring to confirm the period and m sini(i). EXPERT is a planned upgrade to the current Exoplanet Tracker (ET) instrument at KPNO. EXPERT will feature higher resolution (R ~ 20,000 vs. 6,000), improved throughput and long-term stability by eliminating all moving components currently used in the ET instrument. Based on the results from the prototype monolithic interferometer presented in Section 3, with a thermally-compensated Fig. 7 (above): Intermediate-period companions next-gen interferometer and sufficient from the MARVELS pilot program (2006-2007). temperature stability our design goal of 1 m/s The MARVELS survey, once it begins in Fall 2008, will have many intermediate and long-period thermal drift should be obtainable. EXPERT is brown dwarf and extrasolar planet candidates for planned for construction and testing during the Spring of 2008. which an instrument like EXPERT at KPNO can be used as a follow-up instrument. Fig. 6: Radial velocity drift over 6 days of measurements. RMS error is measured to be 14.4 m/s. Based on these measurements, our next-gen version that will be temperature compensated and with high-quality thermal stability should meet our design goal of 1 m/s thermal drift. Fig. 3: Observations of the planet-bearing star rho Crb. An estimated RMS of 11.1 m/s over ~8 days is obtained. Planet orbit taken from the Exoplanets Encyclopedia1. Acknowledgments The authors would like to thank Rich Kron, Bill Boroski and Jim Gunn for allocating engineering and science time at the SDSS telescope for testing the Keck ET instrument and science observations, and the staff of KPNO and APO for their support and expertise. We acknowledge the Keck foundation for funding the Keck ET instrument that performed the MARVELS pilot program survey. We thank NASA, the NSF and the University of Florida for supporting the ET instrument. S.W.F. acknowledges support from the Florida Space Grant Fellowship. Background image credit: http://homepage.mac.com/merussell/serenity_cover_minimalist.jpg 1: http://exoplanet.eu/ Fig. 8: Measured absolute velocity vs. the theoretical prediction if solely determined by temperature drifting (red line). The excellent agreement suggests that at a level of ~100 m/s or less, our instrumental drift is temperaturedominated. A lag in the x-direction is expected as temperature shifts outside the interferometer enclosure (used to calculate the red line) slowly change the temperature inside the interferometer enclosure (the measured velocity drifts represented as black X's).