Fig. 1: Photocurrent pulse timing jitter measurement concepts and results.
From: Attosecond electronic timing with rising edges of photocurrent pulses
a Concept and experimental setup for characterizing relative timing jitter between femtosecond optical pulse train and the generated photocurrent pulse train. MLL mode-locked laser, EDFA erbium-doped fibre amplifier, VDL variable optical delay line, ϕM unidirectional electro-optic phase modulator, ϕb π/2 nonreciprocal phase bias, EOS-TD electro-optic sampling-based timing detector27,28, VOA variable optical attenuator. Note that RIN control is optional. b Photocurrent pulse profiles (voltage across a 25-ohm impedance) measured by a 33-GHz real-time oscilloscope. A 12-GHz p–i–n photodiode (20-V bias) and a 12-GHz MUTC photodiode (8-V bias) for different average photocurrent levels are used. c Timing jitter power spectral densities measured by the EOS-TD at the middle of rising edges of (i) 12-GHz p–i–n (black solid), (ii) 12-GHz MUTC (blue solid), (iii) 22-GHz MUTC photodiodes (violet solid) and at the middle of falling edges of (iv) 12-GHz p–i–n (black dashed), (v) 12-GHz MUTC (blue dashed) and (vi) 22-GHz MUTC photodiodes (violet dashed). (vii) EOS-TD measurement background noise for 22-GHz MUTC photodiode case (grey solid). (viii) Estimated fibre path length jitter (dark yellow solid, see Methods). The bottom curves indicate the integrated root-mean-square (r.m.s.) timing jitters, which are (i) 180 as, (ii) 140 as, (iii) 80 as, (iv) 12 fs, (v) 11 fs, (vi) 6.5 fs, when integrated from 1 Hz to 1 MHz offset frequency.
