1 Introduction

Frequency stable laser systems are a key technology for future space missions aiming at Earth observation, tests of fundamental physics, or gravitational wave detection. The Laser Interferometer Space Antenna (LISA) for gravitational wave astronomy [1, 2] and Earth observation missions such as the Next Generation Gravity Mission (NGGM) and the Gravity Recovery and Climate Explorer follow-on mission (GRACE-FO) [3, 4] rely on laser interferometric measurements with coherent optical links between laser systems with Nd:YAG lasers at 1064 nm situated on remote satellites. The missions have high demands on the laser frequency noise in their measurement bandwidth especially at low Fourier frequencies between 0.1 mHz and 1 Hz, asking for active frequency stabilization to optical frequency references. Potential frequency references need to be robust systems to withstand launch vibrations and thermal cycling and to fulfill the requirements on the frequency noise of the relevant mission. The requirements on the frequency noise of a laser system for LISA are [5]

$$\begin{aligned} S_\mathrm{f}(f) \le 280\,\text {Hz}/\sqrt{\text {Hz}} \times \sqrt{1+\left( \frac{2.8\,\text {mHz}}{f}\right) ^4} \end{aligned}$$
(1)

and for NGGM is [3, 6]

$$\begin{aligned} S_\mathrm{f}(f) \le 30 \text { to } 40\,\text {Hz}/\sqrt{\text {Hz}} \times \sqrt{1+\left( \frac{10\,\text {mHz}}{f}\right) ^2} \end{aligned}$$
(2)

both for Fourier frequencies between 0.1 mHz and 1 Hz. Several macroscopic frequency references have been developed recently with respect to deployment in space missions based on fiber-based or free-beam unequal arm length interferometer [7, 8] and on rugged optical cavities with robust mounting concepts [9,10,11,12,13,14] that partly fulfill the requirements on the frequency noise.

In addition to the demands on the frequency stability, space missions with a laser transponder scheme like LISA and concepts within NGGM further need to synchronize the frequency of the laser systems on remote spacecraft for heterodyne detection within the bandwidth of photo detector systems of the order of 20 MHz. This frequency synchronization is not inherent to macroscopic frequency references, but can be achieved by absolute frequency references provided by optical transitions in atoms or molecules. For laser systems at 1064 nm, the well-known hyperfine transitions in molecular iodine provide optical absolute frequency reference [15,16,17] accessible with Doppler-free spectroscopy techniques at the second harmonic of 1064 nm, i.e., at 532 nm. The absolute frequency of the transition was precisely measured [17] and different iodine references were compared in national laboratories and many international comparisons showing frequency offsets of only 10 kHz or less [16, 18, 19] and the frequency repeatability of individual setups was shown to be better than 1 kHz [20]. In addition to accuracy, setups with fractional frequency instability below \(4\times 10^{-15}\) have been demonstrated [21, 22], which also meets the requirements on the frequency noise of planned space missions. Several laboratory setups of iodine references have been recently demonstrated for space applications, either as simple absolute frequency reference with accuracy of 10 MHz [23] or with performance close to or in agreement with the requirement on the frequency noise of the LISA or NGGM mission [22, 24,25,26].

Here, we present an absolute optical frequency reference based on modulation transfer spectroscopy [27] of molecular iodine that was developed with respect to qualification for potential deployment in space missions. This engineering model of a flight-like iodine reference is designed to be robust enough to survive launch vibrations and thermal cycling and to provide frequency stability and accuracy compliant with requirements of planned space missions. To ease the implementation of a space-born iodine frequency reference, we also test an iodine cell filled at unsaturated iodine vapor pressure, which avoids the need for cooling of the cell that is usually necessary to set the iodine vapor pressure.

Section 2 introduces the spectroscopic setup, called engineering model (EM), and the technology used for its realization. Section 3 presents the laser system and the stabilization scheme. Section 4 describes the environmental tests and reports on the frequency noise and frequency reproducibility of the EM setup before and after the tests. Furthermore, the performance of the setup with the unsaturated iodine cell is presented. Section 5 gives a conclusion and an outlook.

2 Engineering model setup

The engineering model (EM) of the spectroscopy setup was built using a special assembly integration technique to realize an ultra-stable modulation transfer spectroscopy (MTS) setup with balanced detection. The fiber-coupled setup features a specially designed, compact multi-pass absorption cell that allows for a long effective absorption length on a small footprint and detectors for control of the optical power and cancellation of residual amplitude modulation (RAM).

2.1 Optical bench

The MTS setup, shown in Fig. 1 and described in the following, was assembled on a baseplate made of fused silica with dimensions of 38 cm \(\times\) 18 cm \(\times\) 4 cm using an assembly integration technique described in a later section.

Fig. 1
figure 1

Photograph of the engineering model for modulation transfer spectroscopy of molecular iodine. The baseplate has dimensions of 38 cm \(\times\) 18 cm \(\times\) 4 cm and is made of fused silica

Full size image
Fig. 2
figure 2

Schematic drawing of the EM spectroscopy setup. TFP thin-film polarizing beam splitter, P polarizer, BS beam splitter, RM rotation mount with wedged glass plates, PD photo detector, NC noise canceling photo detector, C collimator

Full size image

A to scale schematic of the EM spectroscopy setup is shown in Fig. 2. The pump beam is coupled to the setup through a polarization maintaining (pm) fiber and launched with an optical power of about 8 mW from a collimator (\(C_\mathrm{Pump}\)) with a beam diameter of 2.8 mm at a beam height of 15 mm. A pair of AR-coated wedged glass plates in glass-rotation mounts (RM) are foreseen for fine adjustment of the overlap of pump and probe beam. A fraction of 300 mW of the pump beam is separated at a beam sampler (BS 1) and focused onto a photo detector (PD 1) for the detection and control of residual amplitude modulation (RAM). The pump beam is transmitted by a thin-film polarizing beam splitter (TFP 1), passes the multi-pass iodine cell nine times, and is reflected at TFP 2 and guided to a photo detector (PD 2) for the detection and control of the optical pump power.

The probe beam is sent to the setup through a separate pm fiber and is launched from an identical collimator (\(C_\mathrm{Probe}\)) with an optical power of 3 mW. Glass sheet polarizer (P) with polarization extinction ratio of 100,000:1 are integrated in both collimators to ensure clean, linear polarization. For balanced detection, the beam is split at TFP 3 into a reference beam and a probe beam with optical powers of 2 and 1 mW, respectively. The probe beam is guided to the cell, where a fraction of 200 mW is split off before the cell for stabilization of the probe beam power. The probe beam counter-propagates the pump beam through the iodine cell with orthogonal polarization and is reflected by TFP 1 and guided to a noise canceling detector (NC), adapted from [28], that is used to suppress common mode intensity noise in the detection bandwidth of the MTS signal by up to 20 dB. To adjust the optical power of the reference beam, it passes a half-wave plate (\(\lambda\)/2) and TFP 4 and is guided to the noise canceling detector (NC).

2.2 Multi-pass absorption cell

We have designed a multi-pass cell that provides an absorption length of 90 cm on a small footprint of 10 cm \(\times\) 10 cm using internal reflection from the windows of the cell (cf. Figs. 1, 3).

Fig. 3
figure 3

CAD-drawing exploded view of the multi-pass absorption cell. The hollowed cell body has dimensions of 10 cm \(\times\) 10 cm \(\times\) 3 cm. Green fused silica, blue AR coating, orange HR coating

Full size image

The absorption cell features a robust rectangular design, that is optimized for the assembly integration technique used to integrate the EM. The cell body has a well-defined geometry with angle tolerances of \(\pm 10^{\prime \prime }\), so that the beam stays parallel to the surface of the optical bench. The cell consists of a hollowed spacer with dimensions of 10 cm \(\times\) 10 cm \(\times\) 3 cm with two holes for a cold finger pin and a filling pin both glass-welded to 1” glass blanks and two windows (cf. Fig. 3). All parts are made of fused silica and held together by optical contacts. The windows of the cell are wedged and AR-coated for 532 nm on the outer surface to avoid interference of ghosting beams and stray light. The inner surface of the windows is mainly HR-coated, except for a small AR-coated part, where the laser beams enter and exit the cell. A relatively short cold finger design is used taking into account the expected vibrational loads. The iodine vapor pressure inside the cell is controlled by cooling of some piece of solid iodine, condensed in the cold finger, at \(-10\,^{\circ }\)C corresponding to a pressure of 1.33 Pa using a small Peltier element, a Pt100 temperature sensor, and a commercial temperature controller. The laser frequency shift with the temperature of the cold finger is −300 Hz/K, corresponding to −2.2 kHz/Pa.

Two multi-pass cells were manufactured at Meopta-optika, s.r.o. and filled with pure \(^{127}\)I\(_2\) at the Institute of Scientific Instruments of the Czech Academy of Sciences (ISI) in Brno [29].

2.3 Assembly integration

With respect to the vibrational and thermal loads expected for a space mission and for maintenance-free long-term operation of the iodine reference, the EM spectroscopy module was built using a specific assembly integration (AI) technology. The AI technology must offer the possibility for benign alignment of the optical setup during its integration and high thermal and mechanical stability and long-term stability afterward. Here, the optics are directly bonded to a glass baseplate using a space-qualified two-component epoxy.

The adhesive bonding integration procedure was initially developed for the implementation of a high sensitivity laser interferometer developed in the LISA project [30, 31] and was also used in our previous iodine setup on elegant breadboard (EBB) level [22]. For the integration of the EM setup, two types of space-qualified two-component adhesives were used. A highly fluid adhesive (Hysol EA 9313, Henkel (The citation of a company’s name is for information only and does not constitute an endorsement)) was used for glass-to-glass joints, where both materials have the same coefficient of thermal expansion (CTE). This results in very thin adhesive layers in the few micrometer range [30] and, therefore, high alignment maintenance during the integration of the optical components. For compensating the CTE mismatch between glass and metal parts, a viscous adhesive (Hysol EA 9361, Henkel (The citation of a company’s name is for information only and does not constitute an endorsement)) was used, where a defined layer thickness of 0.2 mm was preset using appropriate nylon threads as spacer within the adhesive layer.

The EM was integrated on a baseplate made of fused silica with a CTE of \(0.55\times 10^{-6}~\)K\(^{-1}\). The surface of the fused silica baseplate was polished to \(\le \lambda\)/10 and its surface as well as the contact surface of the optics were matted to increase the surface for adhesive bonding and to avoid accidental optical contacting during the integration. The mirrors and TFPs have dimensions of 20 mm \(\times\) 35 mm \(\times\) 8 mm. The optics and the absorption cell are also made of fused silica to avoid mechanical stress between the components during thermal cycling caused by differential CTE. The rotation mounts for the wave plates are made of Titanium with a CTE of \(8.6\times 10^{-6}\) K\(^{-1}\) to match the CTE of the zero-order wave plates and to minimize thermal distortion. The mounts for the collimators and adapter plates, which are glued to the side of the baseplate for mounting of the photo detectors, are made of Invar (CTE \(1.6\times 10^{-6}\) K\(^{-1}\)), minimizing CTE mismatch with the fused silica baseplate. Three half-inch glass blanks are glued to the baseplate to define the position of the absorption cell. The mirrors and beam splitter feature a slight wedge of \(1^{\circ }\) and an AR coating on their backside to avoid etalon effects and stray light. Back reflections from the photo diodes were carefully dumped.

3 Laser system and frequency stabilization

A mobile laser system prepares the spectroscopy beams at 532 nm. Electro-optic modulators at 1064 nm are used for frequency shifting, modulation as well as power and RAM control and highly efficient fiber pigtailed second harmonic generation (SHG) modules are used in a final step to provide light at 532 nm via polarization maintaining fibers to the spectroscopy module.

Fig. 4
figure 4

Schematic of the laser system for modulation transfer spectroscopy (MTS) with beam preparation at 1064 nm and feedback control of pump and probe power as well as cancellation of RAM

Full size image

3.1 Laser system

The laser system, schematically shown in Fig. 4, is based on a Nd:YAG laser, that provides 500 mW at 1064 nm. About 5 mW of the output power is split off by a beam sampler and fiber coupled for beat-note frequency measurements. The main part of the optical power is split into a pump and a probe beam with a half-wave plate and a polarization beam splitter. The pump beam is frequency shifted by 80 MHz and frequency modulated at 280.3 kHz and up to 500 kHz deviation by AOM 1 and the 1st diffraction order is coupled into the pm input fiber of the first SHG module (SHG 1). The probe beam passes AOM 2 which is used for intensity stabilization, where the zeroth diffraction order is launched into the pm input fiber of the second SHG module (SHG 2). The infrared pump and probe beam are then frequency doubled by the fiber pigtailed periodically poled lithium niobate (PPLN) waveguide SHG modules (NTT Electronics (The citation of a company’s name is for information only and does not constitute an endorsement)) that incorporate a thermistor and a Peltier element for the stabilization of the phase-matching temperature, here at 41.5 and 56.5 \(^{\circ }\)C. Up to 40 mW of optical output power at 532 nm are available in both, the pump and probe beam, that are fiber coupled to the EM setup. Unconverted infrared light is filtered out by bending losses in the 532 nm pm output fibers of the SHG modules. Because of the frequency doubling, the frequency offset between probe and pump beam and the phase modulation index of the pump beam are doubled as well.

3.2 Electronics and laser control

The electronics features feedback control of the optical power of the pump and probe beam as well as control of residual amplitude modulation (RAM) at the EM with electro-optic actuators in the laser system, cf. Fig. 4, and is similar to the system described in [22].

The probe beam power is detected by PD 3 and controlled by a servo filter with a bandwidth of 1 kHz using AOM 2. The pump power is detected by PD 2 and controlled by an identical servo filter using AOM 1. The RAM at 280.3 kHz is detected at PD 1 to control its influence on the frequency stability of the system [32]. We apply analog IQ demodulation, similar to the digital system described in [33, 34], using power splitters and double-balanced mixers to detect and suppress the in-phase and quadrature components of the RAM. The amplitude modulation generated by the RAM servo is added to the control signal for the pump beam power applied to AOM 1 to cancel the RAM. LC-bandpass filter at 280.3 kHz with a Q-factor of 2 are integrated in the RF paths for RAM and frequency stabilization to filter out harmonics of the modulation frequency.

Laser frequency stabilization is realized with fast feedback to the laser PZT with a control bandwidth of \(\approx\)10 kHz using servo electronics with two integrators at corner frequencies around 3 kHz and 900 Hz, and slow feedback with an additional integrator with time constant of 1 s to the temperature of the Nd:YAG crystal.

4 Experimental results

The frequency stability of the EM was characterized in beat-note measurements at 1064 nm with another iodine frequency reference using a frequency counter. The second iodine reference was realized on elegant breadboard level (EBB) and is detailed in [22]. Environmental tests were carried out to test the robustness of the EM spectroscopy setup with respect to thermal and vibrational loads. The beat-note measurements were carried out before and after each environmental test to check for potential degradation of the system. All offset frequencies and frequency noise spectral densities presented in this paper refer to optical frequencies at 1064 nm.

4.1 Environmental tests

The EM was subject to vibrational loads up to \(25.1~\mathrm{g_{rms}}\) in all three axes as well as thermal cycling from −20 to \(+60\,^{\circ }\)C. The parameters of the vibrational and thermal tests are listed in Tables 1 and 2, respectively. To avoid the risk of loosing the whole EM in case of a possible damage of the optical bench or of the multi-pass cell, the cell was not bonded to the optical bench and both were tested separately.

Table 1 Specifications of vibration loads applied to EM baseplate and multi-pass cell in all three directions
Full size table
Table 2 Specifications for thermal cycling of EM ground-plate and multi-pass cell
Full size table

Two vibrational tests were carried out. In the first run, the first multi-pass cell was tested and the optical bench was tested only partially because the shaker-table got broken. In the second run, the test of the optical bench was completed without iodine cell. Unfortunately, the first iodine cell failed during the vibrational tests. The failure occurred at the optical contact of the filling pin plate, so that the cell was vented with the air. This might have been caused by residual thermal stress induced during the glass welding of the filling pin. Since all other optical contacts showed no degradation, this problem might be mitigated by a more careful design of the filling pin. The optical bench and the detectors endured the vibrational loads without any damage. After the first vibrational test, the frequency stability measurements of the EM setup were carried out with the second iodine cell.

Thermal cycling was first carried out using the specifications according to the LISA technology package (LTP) engineering model developed for the LISA Pathfinder mission and, after visual inspection, with specifications according to the European cooperation for space standardization (ECSS), both summarized in Table 2. The optical bench and the first cell also endured the thermal cycling without damage, and the spectroscopy setup could be rebuild and activated without readjustments or significant changes in performance as reported in the next section.

4.2 Frequency stability and reproducibility

Fig. 5
figure 5

Frequency noise spectral density of the beat-note at 1064 nm between the EM and the EBB iodine reference before and after each environmental test. The performance of the EBB setup and the requirements of planned space mission, cf. Eqs. (1) and (2), are shown for reference. The EM fulfills the requirements on frequency noise for planned space mission and shows no degradation of the performance after the environmental tests

Full size image
Fig. 6
figure 6

Frequency offset at 1064 nm between the EM and the EBB before the environmental tests and after each environmental test. The reproducibility of the optical frequency is determined to be within 1 kHz

Full size image

To determine the frequency stability and frequency reproducibility of the EM setup before and after the environmental tests, it was locked to the transition R(56)32-0:\(a_{10}\) and the frequency stability was determined from the beat-note with the EBB iodine reference locked to the same transition. The amplitude spectral density of frequency noise was calculated from the beat-note time-record using the LTPDA toolbox [35] and is shown in Fig. 5. The frequency noise of the EBB reference, shown in black, was reported previously [22] and does not limit the measurements. The spectra of all four measurement runs exhibit white frequency noise below 10 Hz/Hz\(^{1/2}\) for Fourier frequencies above 50 mHz, corresponding to fractional frequency instability below \(2\times 10^{-14}/\tau ^{1/2}\). For Fourier frequencies below 5 mHz, the spectra show random walk noise on a level of \(S_\mathrm{f}(f) = 1 \times 10^{-1}\,\mathrm{Hz}/\mathrm{Hz}^{1/2}/f\), where the origin of the random walk noise is yet unknown. The frequency noise achieved with the EM fulfills the requirements on planned space missions, such as LISA or NGGM, with large margins and no significant difference in the performance before and after the tests is visible.

The frequency offset from the nominal beat frequency between the iodine references before and after the environmental tests is shown in Fig. 6. The nominal beat-note frequency is 60 MHz and originates from different frequency offsets between the pump and probe beams in the laser systems of the EM and the EBB. The mean frequency offset of the beat-note frequency between the EM and the EBB for all measurement runs, shown in Fig. 6, is 632 Hz with a standard deviation of all four measurements of 107 Hz. These results are comparable to frequency offsets and frequency repeatability found in international comparisons of iodine frequency references [16, 18,19,20]. The frequency reproducibility is, therefore, sufficient to reliably synchronize laser systems on remote satellites as required, e.g., for LISA type missions, to within the bandwidth of the photo detectors (\(\approx\)20 MHz).

4.3 Frequency noise with unsaturated cell

For space missions, cooling of the cold finger is a technical difficulty as it requires additional hardware and thermal design considerations. In a proof of principle measurement, we investigated the performance of the EM with an unsaturated cell, which does not require cooling [36]. To this end, the cell was re-filled at ISI Brno [37] with an unsaturated pressure of 1.3 Pa corresponding to a cold finger temperature of \(-10\,^{\circ }\)C and integrated again on the optical bench of the EM.

Fig. 7
figure 7

Frequency noise spectral density of the beat-note at 1064 nm between the EM with unsaturated cell and the ULE cavity reference. The frequency noise of the ULE cavity and the requirements on planned space missions, cf. Eqs. (1) and (2) are shown for comparison

Full size image

The frequency noise of this setup locked again to the hyperfine-transition R(56)32-0:\(a_{10}\) was determined in a beat-note measurement with a cavity stabilized laser system and is shown in Fig. 7. The measurements were done with the cavity stabilized laser system instead of the EBB reference, since the latter was not available for this measurement. The frequency noise of the cavity stabilized laser system, shown in black, does not limit the measurement. The frequency noise spectrum of the EM setup with unsaturated cell shows white frequency noise on a level of 40 Hz/Hz\(^{1/2}\) for Fourier frequencies between 2 mHz and 10 Hz. The frequency stability of the EM with the unsaturated cell, therefore, also fulfills the requirement on the frequency stability for missions like LISA or NGGM in a setup, where the cooling of the iodine cell can be omitted. This is a great technological advantage and simplifies the application of iodine references on space missions.

5 Conclusion and outlook

We presented a compact and ruggedized, quasi-monolithic spectroscopy setup that was realized using a special designed multi-pass absorption cell and an assembly integration technique developed for the implementation of mechanical and thermal stable optical systems.

The main parts of the spectroscopy setup, the optical bench and the multi-pass cell, were subject to environmental test including vibrational loads up to \(25.1~\mathrm{{g_{rms}}}\) and thermal cycling from to +60 \(^{\circ }\)C. With our laser system locked to the spectroscopy module, we find the frequency noise below 10 Hz/Hz\(^{1/2}\) at Fourier frequencies above 100 mHz which fulfills the requirement on the frequency noise of planned space missions like LISA or NGGM. The repeatability of the absolute frequency of the iodine reference was found to be better than 1 kHz.

We further investigated the performance of the same setup with an unsaturated iodine cell and found similar results in terms of frequency stability with white frequency noise on a level of 20 Hz/Hz\(^{1/2}\). The unsaturated cell significantly simplifies the implementation of a space-qualified iodine reference, since it avoids the need for cooling and temperature control of the cold finger of the absorption cell.

We believe that an iodine reference, such as the one developed in this work, is an interesting candidate for space missions that require laser systems with high long-term frequency stability or an absolute frequency reference at 1064 nm. To demonstrate the technical maturity of the iodine reference, a sounding rocket mission with a laser system based on a micro-integrated diode laser at 1064 nm locked to the iodine reference is planned for end of 2017 [38].