RF-Based Detector for Measuring Fiber Length Changes with Sub-5 Femtosecond Long-Term Stability. J. Zemella 1, V. Arsov 1, M. K. Bock 1, M. Felber 1, P. Gessler 1, K. Gürel 3, K. Hacker 1, F. Löhl 1, F. Ludwig 1, H. Schlarb 1, S. Schulz 2, A. Winter 1, L. Wissmann 2 1 Deutsches Elektronen-Synchrotron (DESY), Hamburg, Germany 2 Institute for Experimental Physics, Hamburg University, Germany 3 Department of Physics, Bilkent University, Ankara, Turkey FEL 2009, Liverpool, United Kingdom J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 1 / 13
Introduction Laser-Based Synchronization System at FLASH. Current State Goal: Synchronization system with a long-term stability of sub-10 fs Master Oscillator low BW PLL Direct use of the laser pulses (BAM) Master Laser Oscillator Distribution Laser Synchronization: Pump-Probe Seed-Laser EO-Laser active stabilized optical fiber links ~5 fs RF-Generation Modelocked Erbium-doped Master Laser Oscillator with 216 MHz repetition rate Distribution of the laser pulses to 14 endstations using optical fiber links Link stabilization with optical cross correlator (OCC) Endstations like beam arrival-time monitor (BAM), two-color OCC or local RF generation (Sagnac loop) J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 2 / 13
Introduction Motivation for RF-Based Detector. Optical Cross Correlator and Conventional RF-Phase Detector Optical Cross Correlator: Necessary: Exact pulse overlap, dispersion compensation, feedback Rather complex, cost intensive but allows fs or < fs resolution. Femtosecond timing not required for most endstations Conventional RF-phase detector: Limitations: AM to PM, offset drifts of the mixer, thermal phase drifts of the photo detection process and the filter Long-term drift 50 100 fs Alternative solution: Amplitude measurement of high harmonics of the interference pattern of two superimposed pulse trains. Less complex, less expensive system J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 3 / 13
Detection Principle Frequency Spectrum of the Photodiode Output. Basics Laser pulse train leads to a frequency comb Frequency lines are spaced by f 0 = 1/T 0 The superposition of two laser pulse trains (I 1 = I 2 ) leads to: Modulated frequency comb Modulation of the n th -harmonic: I(nf 0) cos 2 (πnf 0 t) Intensities of the harmonics depend on the temporal offset t Intensity a.u. Intensität a.u. 0 2 4 6 8 0 5 10 15 0 0 J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 4 / 13
Detection Principle Frequency Spectrum of the Photodiode Output. Basics Laser pulse train leads to a frequency comb Frequency lines are spaced by f 0 = 1/T 0 The superposition of two laser pulse trains (I 1 = I 2 ) leads to: Modulated frequency comb Modulation of the n th -harmonic: I(nf 0) cos 2 (πnf 0 t) Intensities of the harmonics depend on the temporal offset t Intensity a.u. Intensität a.u. 0 2 4 6 8 0 0 5 10 15 0 J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 4 / 13
Detection Principle Frequency Spectrum of the Photodiode Output. Basics Laser pulse train leads to a frequency comb Frequency lines are spaced by f 0 = 1/T 0 The superposition of two laser pulse trains (I 1 = I 2 ) leads to: Modulated frequency comb Modulation of the n th -harmonic: I(nf 0) cos 2 (πnf 0 t) Intensities of the harmonics depend on the temporal offset t Intensity a.u. Intensität a.u. 0 2 4 6 8 0 0 5 10 15 0 J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 4 / 13
Detection Principle Frequency Spectrum of the Photodiode Output. Basics Laser pulse train leads to a frequency comb Frequency lines are spaced by f 0 = 1/T 0 The superposition of two laser pulse trains (I 1 = I 2 ) leads to: Modulated frequency comb Modulation of the n th -harmonic: I(nf 0) cos 2 (πnf 0 t) Intensities of the harmonics depend on the temporal offset t Intensity a.u. Intensität a.u. 0 2 4 6 8 0 5 10 15 0 0 J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 4 / 13
Detection Principle Frequency Spectrum of the Photodiode Output. Basics With the observation of one harmonic a change of the temporal offset is possible Change of the observed harmonic n-times larger for the n th -harmonic Observing two harmonics seperated by a minimum resp. maximum of the modulation eliminates amplitude dependence Intensität a.u. 0 5 10 15 0 Difference signal a.u. 0.5 0.5 I cos 2 (πnf 0 t) cos 2 (π(n + 1)f 0 t) 0 1 J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 5 / 13
Detection Principle Frequency Spectrum of the Photodiode Output. Basics With the observation of one harmonic a change of the temporal offset is possible Change of the observed harmonic n-times larger for the n th -harmonic Observing two harmonics seperated by a minimum resp. maximum of the modulation eliminates amplitude dependence Intensität a.u. 0 5 10 15 0 Difference signal a.u. 0.5 0.5 I cos 2 (πnf 0 t) cos 2 (π(n + 1)f 0 t) 0 1 J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 5 / 13
Detection Principle Frequency Spectrum of the Photodiode Output. Basics With the observation of one harmonic a change of the temporal offset is possible Change of the observed harmonic n-times larger for the n th -harmonic Observing two harmonics seperated by a minimum resp. maximum of the modulation eliminates amplitude dependence Intensität a.u. 0 5 10 15 0 Difference signal a.u. 0.5 0.5 I cos 2 (πnf 0 t) cos 2 (π(n + 1)f 0 t) 0 1 J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 5 / 13
Detection Principle Frequency Spectrum of the Photodiode Output. Basics With the observation of one harmonic a change of the temporal offset is possible Change of the observed harmonic n-times larger for the n th -harmonic Observing two harmonics seperated by a minimum resp. maximum of the modulation eliminates amplitude dependence Intensität a.u. 0 5 10 15 0 Difference signal a.u. 0.5 0.5 I cos 2 (πnf 0 t) cos 2 (π(n + 1)f 0 t) 0 1 J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 5 / 13
Detection Principle Frequency Spectrum of the Photodiode Output. Basics With the observation of one harmonic a change of the temporal offset is possible Change of the observed harmonic n-times larger for the n th -harmonic Observing two harmonics seperated by a minimum resp. maximum of the modulation eliminates amplitude dependence Intensität a.u. 0 5 10 15 0 Difference signal a.u. 0.5 0.5 I cos 2 (πnf 0 t) cos 2 (π(n + 1)f 0 t) 0 1 J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 5 / 13
Setup Optical Part. Schematics of the Superpostion of the both Pulse Trains to end-station 50/50 FRM ~20m collimator 1 2 3 b Laser half wp quarter wp movable mirror collimator 2 detector inloop J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 6 / 13
Setup Optical Part. Schematics of the Superpostion of the both Pulse Trains 50/50 FRM 50/50 coupler detector outloop ~20m collimator 3 collimator 1 1 2 a b 3 Laser half wp half wp quarter wp movable mirror collimator 2 detector inloop J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 6 / 13
Setup RF-Part. Balanced Detection Scheme. Power detection ADC Filter 44 f₀ = 9.53 GHz opt. Signal Filter n f₀, (n+1) f₀ Power detection ADC Filter 45 f₀ = 9.75 GHz Photodiode with 10 GHz bandwidth Power-detector: Zero Bias Schottky Detector ADC with 1 MHz sampling rate and a bandwidth of 40 MHz J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 7 / 13
Results Calibration. The Voltage Change of the Detector Channels. 2 nd -order polynomial is fitted to the data to calculate the voltage into time 5 dv dt mv 10 15 ps voltage (V) 0 5 0.1 0.15 10 5 0 5 10 time (ps) Blue: Red: Green: Black: Inloop detector 44 f 0 = 9.53 GHz Inloop detector 45 f 0 = 9.75 GHz Outloop detector 44 f 0 = 9.53 GHz Outloop detector 45 f 0 = 9.75 GHz J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 8 / 13
Results 50 h Long-term Measurement. Balanced Time Change of the Inloop and Outloop Detector t 1,2 = 1 2 (t 9.53 GHz + t 9.75 GHz ) balanced time (fs) 400 200 0 200 400 600 800 0 10 20 30 40 50 time (h) Red: Black: Inloop detector t pp = 1.24 ps Outloop detector t pp = 1 ps Inloop detector measures fiber length changes twice Measurement bandwidth: 500 Hz J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 9 / 13
Results 50 h Long-term Measurement. Time Difference of the Inloop and Outloop Detector time difference of the detectors (fs) 30 20 10 0 10 20 t = 1 2 t 1 t 2 30 0 10 20 30 40 50 time (h) Peak-to-peak of the time difference: t pp = 20 fs Standard deviation of the time difference over 50 h: t = 4.6 fs Resolution of one detector: Blue: t Res = 3.2 fs Measurement bandwidth: Blue 500 Hz Red 10 mhz J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 10 / 13
Results Application of the Detector Length Change Measurement of PSOF Link Temperture change for the fiber T = ±3 C time difference of the inloop and outloop detector (fs) 15 10 5 0 5 10 15 0 1 2 3 4 5 6 7 time (h) t pp = 55 fs @ ±3 C, link length 20 m T k = fs/m K t = 3 fs (RMS) t Res = 2.1 fs (RMS) J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 11 / 13
Conclusion and Outlook Conclusion and Outlook. New detection principle based on interference pattern of two superimposed pulse trains. Drift-free because of the use of only one photodiode and an amplitude measurement instead of a phase measurement. Long-term resolution over 50 h of 3.2 fs could be achieved. Try to use the scheme for longer fiber links. Comparison with the optical cross correlator. Install a stabilized link to connect the photo injector laser at FLASH to the synchronization system. J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 12 / 13
Conclusion and Outlook Acknowledgements. On behalf of the FLASH-LbSyn-Team and involved DESY-Groups Thank you for your attention! J. Zemella (DESY Hamburg) RF-Detector Design FEL 2009-08-27 13 / 13