I winded a few meters of optical fibers on a PZT tube which is driven by a signal generator, then made a fiber optical resonator with direction coupler. The source which I used is a 1550nm laser and the detecor is made of InGaAs, the signal is demonstrated on a oscilloscope.
What I observed is that the resonant curve on the oscilloscope is swinging, which means the resonant dip is drifting during the modulation process.
I wonder whether there's something wrong with my PZT tube or the phenomenon is normal while modulating with PZT. If it's normal, how should I remove or suppress the drift of resonant dip?
If I understand your configuration correctly, your laser is connected to one port of the directional coupler. A few metres of optical fibre are wound around a piezo-electric cylinder, with ends connected to input and output ports of the directional coupler. The remaining coupler port is connected to a detector whose electrical output is monitored by an oscilloscope.
How stable is the temperature? The temperature coefficient of refractive index of silica is around 8.5 10-6 K-1 (Leviton, 2006) The linear thermal expansion coefficient of silica is 0.5 10-6 K-1. In 5 m fibre, the optical path length will change by 45 10-6 m K-1 , so roughly 0.3 wavelengths for 0.01 K temperature change. Temperature stability better than 0.001 K is desirable.
The thermal expansion coefficient of the PZT could further increase the temperature sensitivity.
How stable is the laser temperature and drive current? A typical 1550 nm DFB might have temperature coefficient of wavelength around 0.1 nm / K (12 GHz / K optical frequency coefficient). Free spectral range of a 5 m fibre resonator cavity is 0.33 pm, or 0.04 GHz.
The change of DFB optical frequency with drive current might be 0.6 GHz / mA, so current stability of a few parts per million, and temperature stability better than 0.0001 K is desirable, depending on the phase stability you require.
Note also that stretching the fibre in the PZT will change the fibre birefringence. If the input polarisation is not aligned with the birefringence axes, the polarisation state of the output will vary with PZT drive. This will change the magnitude of the demodulated signal. If normal (not polarisation maintaining) fibre is used, the output will be sensitive to disturbance and temperature change in the fibre between laser and directional coupler.
https://arxiv.org/pdf/0805.0091
http://www.kayelaby.npl.co.uk/general_physics/2_3/2_3_5.html
@Alan Robinson
Your detailed answer is appreciated. I preferred the modulation of PZT on fiber length is not linear because the process of stretching can’t be exactly periodic, but I didn’t expect the temperature fluctuation could affect the resonant curve so much. So it seems that I should operate the laser with a high precision power supply and keep the entire platform in a thermostatic chamber. Hope it works. Thank you again for your kindly help!
It is necessary to use piezoelectric ceramics and a fiber with equal coefficients of temperature expansion, consistent with the geometry of the device.
@Nikolay Krobka
Thanks for your answer. I checked some reference and it seems that the coefficient of barium titanate ceramics is nearly an order of magnitude larger than the one of fiber. Does it mean that I should use piezoelectric ceramics made of other materials?
If you want to avoid the thermal expansion of fiber, I am suggesting two solutions.
1) An electro-optical system with a temperature-drift compensation of a fiber link where the delay drift is measured by a radio frequency phase comparison. A (coarse) wavelength division multiplexing technology is used for combining compensation and user signals. Temperature and other delay variations (e. g. vibrations) are compensated by an internal heated/cooled fiber spool and/or piezo-controlled fiber stretcher. This one is in the range of ps stability.
https://www.researchgate.net/publication/301542448_Multi-Purpose_Constant-Delay_Optical_Link
2) If you want to have few fs stability, you need the system where the temperature and the vibration-induced fiber group-delay variations are compensated by adjusting the wavelength of the laser and making use of the fiber chromatic dispersion.
Distribution of RF-clock signal over single optical fiber with temperature-drift and vibration compensation. Available from: https://www.researchgate.net/publication/269329508_Distribution_of_RF-clock_signal_over_single_optical_fiber_with_temperature-drift_and_vibration_compensation
Conference Paper Multi-Purpose Constant-Delay Optical Link
Conference Paper Distribution of RF-clock signal over single optical fiber wi...
@Bostjan Batagelj
Your suggestion provides me a brand new solution to my experiment. I will try to construct feedback loop behind laser source to stable wavelength and check the drift of resonant curve. The compensation of using fiber stretcher is also considered. Thanks for your help and paper!
Hi Chenge, In the case that you will need some help with fabrication of devices, go to www.inlambda.com and check Products or Solutions os simply contact them. You can go directly to jurij[/at\]inlambda.com.
@Bostjan Batagelj
I'll contact inlambda if necessary. Thanks for your kind information and well noted.
- Badges
- Science topic