3.1

Side band phase noise (SBPN)

The measurement principle is presented in Figure 4. The LH, with sideband modulation on, is attenuated and combined with a Local Oscillator (LO) laser. The beat note between these two lasers is recorded with a fast photodetector. After filtering, this measurement is composed of two frequencies: the carrier-LO peak at 2.4 GHz+Δf and the sideband-LO peak at the frequency Δf. These two frequencies are isolated with an RF splitter and filters. The 2.4 GHz+Δf frequency is used to offset lock the LO laser to the carrier.

Figure 4:

Setup for sideband phase noise measurement

As the LO laser is locked on the carrier, the sideband-LO beat note contains the phase noise generated by the LH Electro Optical Modulator (EOM) and all other components within the signal path. This signal is sent to a phase detector for comparison to the reference signal.

The reference signal is generated by mixing the outputs of the synthesizer generating the 2.4 GHz and the 2.4 GHz+Δf carriers-LO beat note. This signal is considered as the reference since it does not travel through the EOM and the subsequent LH system. The phase detector signal is recorded and processed to calculate the associated Power Spectral Density (PSD). To meet the requirement, the phase noise induced by the test bench shall be minimized.

In order to validate the test bench is verified on a MO followed by an EOM. The performed measurement is presented in Figure 5. For this measurement, the EOM is seeded with a 2.4GHz at 13.8dBm signal (cable losses not taken into account).

Figure 5 :

Sideband phase noise test bench noise floor

The measurement depicted in Figure 5 reaches the specification at low and high frequency. Between 0.1Hz and 1mHz, the measurement setup is out of specification mainly due to thermal issue.

During the setup validation, it was noted that temperature drifts are the most frequent origin of long-term drifts. The increase of phase noise density at low frequencies is very likely due to temperature drift of the Radio Frequency (RF) components of the setup (amplifiers, mixers, splitters, filters, cables). An improved thermal stabilization should improve the test bench noise floor.

3.2

Frequency stability

The requirement on frequency stability spectral density limit S_υ for the pre-stabilized laser mode, is depicted in Figure 6.

Figure 6:

Frequency noise requirement

To be able to assess the achieved laser performances, two dedicated test benches had to be developed, one covering the low frequency range [30uHz; 3kHz] and one the high frequency range [3kHz; 10MHz]. Both are presented in the following.

Low frequency range [30 uHz; 3 kHz] test bench

The setup developed for this frequency range is depicted in Figure 7. The LH is characterized by measuring the beatnote frequency noise of the NASA laser against a more stable reference laser. To obtain this reference light, we lock a self-referenced optical frequency comb to the optical frequency of a 1560 nm Continuous Wave (CW) laser (from OEwaves) locked to an ultrastable cavity and we correct the cavity drift by a reference maser leading to the combination of the best frequency stability performances of the two references (cavity and maser).

Figure 7:

Architecture of frequency noise measurement setup for the [30u 3 kHz] frequency range.

In more details, this stabilization is achieved via an offset lock of a 1560nm laser on the cavity. Then, one optical mode of a homemade 600MHz repetition rate self-referenced frequency comb (FC) is phase locked to this cavity-stabilized 1560nm laser. The phase of the frequency comb repetition frequency and the phase of a synthesizer referenced to the maser are compared. The obtained phase error signal is fed to the 1560 nm Continuous Wave (CW) laser control electronics to change its optical frequency via its offset lock to the reference cavity. In this way the long-term stability of the maser is transferred to the 1560 nm CW laser, correcting the drift of the cavity. As such the system delivers:

Finally, to measure the frequency noise of the NASA laser, stable comb light at 1064nm is used to phase lock a 1064 nm CW laser (from Redfern Integrated Optics (RIO)) considered as local oscillator (LO). The beatnote of the LO laser and the Laser Head under test is then recorded with a photodetector and analyzed with a frequency counter and a phase noise test set.

The developed cavity and the frequency comb are depicted in Figure 8.

Figure 8:

CSEM reference cavity and frequency comb

In Figure 8, on the left, the optical reference cavity is shown. The cavity is a 1560nm Fabry Perot cavity with a finesse of 200,000 and maintained at its zero coefficient of thermal expansion temperature. To reach the required stability, the cavity is maintained under vacuum, thermally stabilized via active temperature stabilization and passive thermal shields. The CSEM design is based on [3] and [4]. The cavity stabilization digital electronics including the offset lock and the referencing on the maser through the frequency comb repetition rate was developed by CSEM in the framework of the ESA OSRC project [5].

In Figure 8 right, the frequency comb setup is represented. The 600 MHz diode-pumped solid-state frequency comb shown on the top left of the picture was designed and manufactured by CSEM. The output light of this comb is split to seed two independent fiber amplifiers. The first amplifier is used to self-reference the frequency comb with a standard f-2f interferometer. The stabilization is performed via the frequency comb pump diode current with home-made ultralow noise electronics. The second amplifier arm is used to generate the 1064nm light to phase lock the 1064nm CW LO.

High frequency range [3 kHz; 10 MHz]

At high frequencies, the measurement approach relies on a Michelson interferometer that converts laser frequency fluctuations into intensity fluctuations. The optical schematic of the Michelson interferometer is depicted in Figure 9. The light from the LH is attenuated and sent to a non-polarizing beam splitter. In both interferometer arms, retro reflectors based on right angle prisms are used. Reflections from both arms interfere in the beam splitter and the two outputs are sent to a balanced photodetector having the appropriate bandwidth. To cover the needed range of the measurement, an interferometer with an arm length difference of 2m and a related Free Spectral Range (FSR) of 75 MHz is used. Balanced photodetection is used to remove the noise due to the Relative Intensity Noise (RIN)

Figure 9:

Michelson interferometer for frequency noise measurement in the [3 kHz; 10 MHz] range.

To validate the setup, a measurement performed using an OEwaves MO is presented in Figure 10. This MO does not reach the specification but was the lowest high frequency noise laser available at CSEM premises before the NASA uNPRO arrived. It is thus not possible to fully validate the test bench limit with this laser.

Figure 10:

Blue curve represent the high frequency, frequency noise requirement. The red line is the measurement realized on a MO laser (OE-waves). The black curve represents the dark noise of the photodetector.

Figure 10 shows that, between 100kHz and 1MHz, the test bench can measure the LISA specification. Between 1kHz and 100kHz, the measurement is limited by the OEwaves laser that does not reach the specification. It is thus impossible to fully conclude if the test bench can measure at the specification level over the full bandwidth. This being said, later a measurement of the NASA uNPRO will show that the bench is well capable to meet the specification. Below 1kHz, the setup is governed by mechanical noise and the measure is limited by the test bench itself. An improvement on the mechanical isolation should improve the measurement in the future.

3.3

Output Power Stability

Output power stability is one of the most important performance requirements of the LH. The specified output power stability spectral densities are depicted in Figure 11.

Figure 11 :

Output power stability RIN from the development requirement specification in orange

As this requirement covers more than 10 orders of magnitude of frequency range, three different setups corresponding to the low frequency range [100 uHz ;1 Hz], the intermediate frequency range [100 mHz ;10 MHz] and the high frequency range [100 kHz ;5 GHz] of RIN were required. They are presented separately in the following.

Low frequency [30 uHz;1 Hz]

To reach the low frequency range requirement, an active stabilization of the amplifier is required. NASA developed a stabilization test bench and duplicate the measurement photodiode (PD) and light picking stage to perform an out of loop measurement as well. NASA made available its low frequency power test / stabilization bench for the test campaign at CSEM. A picture of the bench is shown in Figure 12.

Figure 12:

Picture of NASA power test / stabilization bench

The power stability depicted in Figure 13 was measured with the out of loop detector of the NASA measurement setup. This measurement shows that the test bench is able to measure the requirement.

Figure 13:

Measurement of power stability with out of loop detector

Intermediate frequency [100 mHz ;10 MHz]

The intermediate frequency measurement setup is presented in Figure 14. The LH output is first attenuated with an output coupler and then focalized on an off-the-shelf photodetector and the signal is recorded with a vector signal analyzer.

Figure 14 :

Intermediate frequency power stability test bench.

For the test benches validation, a measurement shall be performed on a laser within in specification. Such a laser is available from the previous ESA activity Monalisa where the CSEM developed a LH BB and test two different seed lasers [6][7].

Figure 15 shows the intermediate frequency RIN test bench validation performed with the measurement of the CSEM LH BB seeded with a RIO laser.

Figure 15 :

Power stability intermediate frequency measured wit CSEM LH BB with RIO seed laser

As one can see in Figure 15, the noise floor is sufficient for RIN detection below the noise density specifications from 100 mHz to 3MHz. At higher frequency the measurement is limited by the laser noise. A measurement performed on CSEM LH BB with OEwaves seed laser (Figure 17) shows that the intermediate frequency measurement test bench can reach the specification at frequency higher than 3MHz. The light blue curve represents the PD dark noise and thus the measurement limitation.

High frequency [100 kHz ;5 GHz]

The bench is depicted in Figure 16. The LH is first attenuated with an output coupler and coupled in a fiber. The signal is detected with a 13 GHz bandwidth photodetector. Its output signal is analyzed by an electrical signal analyzer.

Figure 16:

High frequency test bench

The proposed test bench was validated using CSEM LH BB with OEwaves seed laser. The measurement is shown in Figure 17.

Figure 17:

Power stability high frequency measured wit CSEM LH BB with OEwaves seed laser

As one can see in Figure 17, starting from 100 kHz the measurement overlaps with the one from the intermediate frequency range. At high frequency the power stability measurements reach the 10^–8/VHZ level are obtained. This is in line with the power stability requirement from the ESA specification.