Possible applications of a novel type of photon counting instrument for

Transcription

1 Possible applications of a novel type of photon counting instrument for Intensity Interferometry observations University of Padova Workshop on Stellar Intensity Interferometry p y y Salt Lake City January 2009

2 Introduction During the last years we realized in Padova two similar instruments, AquEYE and IquEYE, for astronomical applications. They are essentially extremely fast photon counters, with the capability of time tagging the collected photons with a 50 ps time accuracy and storing all the timing data in a mass memory. This type of instrument is really versatile because it allows to operate independently d with distant t tl telescopes if a suitable clock synchronization can be obtained. We are planning to further develop this type of instruments for possible applications that can range from quantum observations with future ELTs, as measurement of second and higher order correlation functions from remote light sources, to intensity interferometry with existing telescopes as VLT and Keck. 2

3 The team Many people are participating to the realization of this project: Univ. Padova: C. Barbieri, I. Capraro, G. Naletto, T. Occhipinti, E. Verroi, P. Zoccarato, V. Da Deppo, C. Facchinetti, C. Germanà, E. Giro, M. Parrozzani, F. Tamburini, M. Zaccariotto, L. Zampieri INAF Rome: A. Di Paola, INAF Cagliari: P. Bolli, C. Pernechele INAF Catania: S. Billotta, G. Bonanno, Collaborations: D. Dravins (Lund), A. Cd Cadez (Ljubljana) 3

4 Outline Some history: QuantEYE Description of AquEYE and of some of the obtained results Description of IquEYE and of someof of the obtained results (very preliminary) Results of the JointAsiago Ljubljana Crab pulsar observation (preliminary) Instrument present limitationsandand possible ways to overcome them Future applications 4

5 QuantEYE proposal In Sept. 2005, we completed a study (QuantEYE, the ESO Quantum Eye) in the frame of the studies for the 100 m OWL telescope. The main goal of the study was to demonstrate the possibility to reach the ps timeresolution needed to bring quantum optics concepts into the astronomical domain, with two main scientific aims in mind: Measure the entropy of the light through the statistics of the photon time of arrival (TOA) Demonstrate the feasibility of HBTII 5

6 Why studying the photon time statistics? 6

7 Why Extremely Large Telescopes? The above mentioned quantum correlations are fully developed ontimescales of the order of the inverse opticalbandwidth bandwidth. For instance, with the very narrow band pass Δλ =0.1 nm in the visible, through a definite polarization state, typical time scales are 10 ps. However, the photon flux is very weak even from bright stars, so that only Extremely Large Telescopes (ELTs) can bring Quantum Optical effects in the astronomical reaches. 7

8 QuantEYE QuantEYE was conceived for measuring second and higher order correlation functions in the collected photon stream (up to 1 GHz) from OWL with the highest timeresolution (better than0 0.1ns) ns). 8

9 Key limitation: the detector The most critical point, and driver for the possible optical designs of QuantEYE, was the availabilityof very fast and accurate photon counting detectors. Imaging g PC detectors (ICCD, ICMOS, MCP) either do not allow fast time tagging of the detected events, or have a rather low maximum total count rate Non imaging PC detectors (PMT, SPADs) either have a relatively low QE, or have a small sensitive area SPADs are preferable: a 50 ps time resolution with count rates as high as 10 MHz can be obtained, with standard voltages and QE. However, even if the time resolution could be acceptable for this application, the total count rate was still two orders of magnitude smaller than what was necessary! 9

10 Solution: splitting the problems To suitably design the system and to overcome both the SPAD limitations and the difficulties of a reasonable optical design (coupling the 100 m pupil / 600 m focal length of OWL with a single 50 μm detector!), we decided to split the problems. In practice, we designed QuantEYE subdividing the system pupil into N N sub pupils, each of them focused on a single SPAD (so giving a total of N 2 distributed SPAD's). In such a way, a sparse SPAD array (SSPADA) coping with the required very high h count rate could be obtained. The SSPADA is sampling the telescope pupil, so a system of N 2 parallelsmallertelescopes isrealized realized, each one actingasa as a fast photometer. 10

11 QuantEYE optical design Schematic view of the telescope pupil subdivision 11

12 Advantages of this optical design The global lcount rate is statistically ttiti increased dby a factor N 2 with respect to the maximum count rate of a single SPAD. In the assumption of having N = 10 (100SPAD's), the globalcount rate becomes 1 GHz (one photon every 100 ns on each SPAD) Simpler optical design Detector redundancy By suitable cross correlations correlations of the detected signal, a digital HBT intensity interferometer is realized among a large number of different sub apertures across the full OWL pupilp 12

13 Overall QuantEYE block diagram The overall system: two heads, controls, storage, time unit. 13

14 While expecting the realization ofthe future E ELT ELT, we decided to apply the described concept to realize a much smaller version of the instrument, compatibly also with the few available funds. We named this instrument AquEYE, the Asiagoquantum eye: it has been applied to the AFOSC camera of the Asiago Cima Ekar (Italy) 182 cm Telescope. AquEYE 14

15 AquEYE optomechanical design A simple way of realizing this small prototype was to consider an opticalconfiguration configuration inwhich the telescope pupil isdivided in four parts only by means of a pyramidal mirror. 15

16 AquEYE subsystems AFOSC focus Pyramid Focusing lenses Filters SPAD 16

17 Selected detectors As best compromise, the selected detectors are SPADs produced by Italian company MPD. Their main drawbacks are the small sensitive area (50 µm diameter) and a 70 ns dead time. 17

18 Advantages of multiple detectors Differences between the photon times of arrival for 1 or 4 SPADs. (some MHz total rate) 18

19 AquEYE electronics schematics 19

20 Time referencing and tagging The CAEN TDC board samples the collected events at 40 GHz (25 ps time resolution), multiplying a reference frequency at 40 MHz. To maintain the desired 100 ps time accuracy over hours of observation avoiding too expensive solutions like Hydrogen maser or Cesium clock, a rubidium oscillator coupled to a Trimble Mini T GPS disciplined OCXO (Oven Controlled X tal Oscillator) has been used as external reference frequency to the CAEN TDC board. This clock is extremely accurate on short term, but has a drift for long periods. To remove this drift, the PPS signal from GPSDO (GPS Disciplined oscillator, which is synchronized within 25 ns rms to UTC) is given in input to the CAEN board and time tagged together with ihthe events. Then a post process linear fit analysis of the collected PPS allows to estimate the rubidium drift, and to remove it. 20

21 Data handling Obviously, all the data have to be stored and preliminary analyzed at their production rate. To store and analyze all the collected data a central storage unit with a capacity of 1 TB has been used. The arrival time of each photon is given as input to an asynchronous post processor which guarantees data integrity for the following scientific investigation. 21

22 The light curve of the Crab pulsar Average Crab pulse profile from Asiagodata (blue) and from 4 m Kitt Peak telescope data (red; Fordham et all, ApJ. 581, 2002). The measured period in Asiago was P = s, to be compared with the P = s extrapolated from Jodrell Bank ephemerides. 22

23 IquEYE Thanks to the positive experience of AquEYE, it has been decided to realize IquEYE, a more complex instrument for applications to larger telescopes, as NTT and TNG. 23

24 IquEYE optical layout 24

25 IquEYE opto mechanics 25

26 IquEYE electronics Monitor camera and motor controls CAEN board GPS CAEN board (redundant) Rubidium clock Control & data analysis & storage server Acquisition server 26

27 IquEYE block diagram ATFU AquEYE Time and Frequency Unit EQuA Electronics for Quantum Astronomy Optics Telescope, Optical AquEYE and Detectors Mass Storage QuAS Quantum Astronomy Software Scientific DATA 27

28 Equa schematics 28

29 The Crab pulsar at NTT 29

30 The PSR B NTT (2009) HSP on HST (1993) CTIO 4 m (1985) 30

31 At the other extreme: Eta Carinae 31

32 The Asiago Ljubljana experiment The Ljubljana telescope (80 cm diameter) is 230 km far from Asiago. 32

33 Joint observations of the Crab pulsar On October 2008 we performed joint observations of the Crab pulsar. Both the observatories were equipped with a breadboard ACTS (Accurate and Certified Time System) clock unit provided by Thales Alenia Space. This is an experimental setup to simulate the characteristics of timing of the future Galileo system. ACTS assures a time accuracy of 25 ns on UTC, and certifies the time. These units were used to have a common clock, with which we tried to synchronize the two observations. 33

34 Obtained results The obtained data have not been completely analyzed yet. The preliminary results (determination of the initial phase of the Crab pulsar period) show that the two measurements were about 100 µs out of phase. This value is much larger than expected and suspected; ; investigations are going gon to understand the reason of this discrepancy. However the pulsar period obtained by this measurement was in agreement within 1 ns with the value given by the value obtained by means of Jodrell Bank ephemerides, demonstrating a perfect internal clocking. 34

35 Total count rate Instrument present limits The used SPADs have two outputs: an extreme timingaccurate (25 ps) NIM, which limits the linearity range of the detector to about 2 MHz; an about 10 times less accurate TTL, which gives up to 12 MHz count rate. To have the best timing, we used the NIM output, accepting a low count rate. The used CAEN board limits the total output count rate to 8 MHz Detector dead time The used SPADs have an about 75 ns dead time, limiting the single channel maximum rate (f (if TTL output is used) but mainly inhibiting the capability of detecting very time close photons 35

36 Possible instrument improvements (I) Total count rate CAEN people assured that they willincrease increase the board output band. Anyway, we could simply use more boards in parallel It is rather difficult to improve the MPD SPAD time accuracy performance. However, SPAD technology is fast improving: several companies are now producing them, and SPAD arrays are becoming available. It is reasonable to suppose that in a few years it will be possible to have more performing SPADs Detector dead time The use of multiple detectors statistically allows to greatly reduce this problem. The higher h the detector number, the higher the probability of detecting very time close photons, substantially reducing to zero the dead time. 36

37 Possible instrument improvements (II) The optical design can be improved. In fact, the present design is a consequence of the limited availabilityof suitable detectors. Presently, the detector limitations imposed a multi channel optical design, with all the related complexity. If SPAD arrays will be available in the future, a much simpler optical design will be possible. The timing accuracy can be improved In future it will be possible to use the better GNSS Galileo receiver with the aim to achieve a better synchronization to UTC. 37

38 Future developments We are planning to bring IquEYE also to TNG, which is very similar to NTT. A hypothesis under investigation is to leave IquEYE (upgraded) as a resident instrument for NTT. The next step will be to realize another version of this instrument to be brought to one of the existing 8 10 m telescopes (for example the Very Large Telescope at Cerro Paranal, Chile, or the Large Binocular Telescope in Tucson, or Keck on Mauna Kea). We have already applied to be funded for this experiment, and contacts have already been taken with VLT. We are also considering the possibility of mounting a quantum detector in the central pixel of the Cherenkov light collector MAGIC (Major Atmospheric Gamma Imaging Cherenkov) (Roque de los Muchachos, Canarias, Spain). 38

39 HBTII possible application This type of apparatus could be used with a network of telescopes, allowingfor example multi dimensional HBTII performed by means of post process data analysis. 39

40 Possible performance of HBTII applications Simulations have been performed to verify the possibility of realizing HBTII withthis this type of instrument. Test conditions: λ = 500 nm Δλ = 3 nm QE = 0.7 Losses = 0.3 Detector dead time = 70 ns Number of detectors = 4 Two cases have been considered: 8 m telescopes, 1 ns time accuracy, 2 hours integration time 1.8 m telescopes, 20 ns time accuracy (Tempo2), 4 hours i.t. 40

41 Possible performance of HBTII applications 41

42 Comments on the plot Small telescopes are rather inefficient for HBTII applications, but could be used with long exposure times To synchronize the observations, the photon TOA s have to be homogenized at the solar system baricentre by suitable s/w, as Tempo2. The time error associated with Tempo2 is 20 ns: this is the error in time that has been considered for the present instrumentation applied to 1.8 m telescope. However it is not clear how it should be considered in these applications. The SNR ratio is proportional to the square root of the integration time: very long observations can be done Flattening of the lines is due to saturation of the SPAD because of high rate. If more SPADs can be used, the SNR can linearly increase 42

43 Another simulation 43

44 Other possible applications The realized instrument allows to perform measurements of other very fast phenomena: Variabilities close to black holes Free electron lasers in magnetars Flare stars Lunar occultations CV Exoplanetary transits 44

45 Conclusions The characteristics of QuantEYE, AquEYE and IquEYE, the instruments studied, realized and tested have been reviewed. The proposed designs are very modular, and can be easily adapted to any optical telescope. The performed tests showed that this type of instrument performs very well as extremely fast photon counters / photometers. The instrument characteristics make it very suitable for HBTII applications also with the present design. It is reasonable to expect that in a few years much better performance can be obtained, mainly improving the time tagging accuracy. The adopted philosophy ofstoring allthe collected data allowsthe possibility of using network of telescopes, also located in different sites. 45