Advanced Concept For Chip-Scale Atomic Clocks (A-CSAC)

Advanced Concept For Chip-Scale Atomic Clocks (A-CSAC)

ESA Contract N° 4000130251/20/NL/LW

CSEM Project n° 261-ES.2213

Final Presentation

WebEx Meeting

04.02.2021

Sylvain Karlen

2

• Introduction: The A-CSAC project

• Motivation

• Optical transitions: the future of chip-scale atomic clocks?

• Photonics integrated circuits: a necessary building block

• CSEM two-photon atomic clock

• Chip-scale approach

• CSEM clock test setup

• PIC

• Conclusion

Agenda

3

The A-CSAC project: key numbers

• Start date: April 2020

• End date: December 2021

• Total effort: ~ 2000 engineering hours

• Total cost: 500 kCHF

• Partner: Ligentec

• Key sub-contractor: University of Neuchâtel – LTF

Orolia Switzerland SA

4

• To analyse, define and demonstrate an advanced chip-scale

atomic clock architecture based on hot vapor optical transitions

• Survey, trade-off analysis and architecture definition

• Clock sub-system prototypes design (PIC and atomic reference)

• Clock sub-system prototypes manufacturing and assembling

• Prototypes testing and recommendations for further developments

The A-CSAC project: project goal

Optical transitions

The future of chip-scale atomic

clocks?

6

CSACs: State-of-the-art of RF-based clocks

• Today most commercial CSACs are based on CPT or Double-resonance interrogation

• Typical performances are (order of magnitude)

• Short-term stability: 10-10 to 10-11 at 1s

• Long-term stability (drift): 10-10/month

• Timing error: 1μs/day

• Volume: 10cm3

• Power consumption: 100 mW

• Improvement of the performances are made in two directions:

• SWaP reduction (ASIC, packaging, …) → high TRL

• Stability improvement (pulsed pumping, light-shift reduction, …) → only limited gain

→ The use of an alternative technology was proposed

Microsemi SA.45s CSAC

7

CSACs: Alternative technologies

• Trapped atoms clocks

• 87Rb MOT RF clock (e.g. Szmuk et al. , PRA, 2015)

• 171Yb+ ions trap optical clock (Schwindt et al. RSI, 2016)

→ High stability but complex system

• Molecular and NV center clocks

• Carbonyl Sulfid (OCS) gas (Wang et al., IEEE JSSC, 2019)

• 15N@C60 (Harding, PRL, 2017)

• NV centers (Dmitriev, IEEE SL, 2020)

→ Simple system but stability not yet demonstrated

• Hot vapor optical clock

→ Improved stability demonstrated, relatively simple system

This work

8

CSACs: Hot atomic vapor optical clock architecture

• The most appropriate transition for a hot atomic vapor cell optical clock is the 52S1/2

→ 52D3/2, 5/2 2-photon transition in Rb at 778.1 nm

• Any optical atomic clock requires two parts:

• An atomic reference which consists in a CW laser stabilized on an atomic transition

9

CSACs: Hot atomic vapor optical clock architecture

• The most appropriate transition for a hot atomic vapor cell optical clock is the 52S1/2

→ 52D3/2, 5/2 2-photon transition in Rb at 778.1 nm

• Any optical atomic clock requires two parts:

• An atomic reference which consists in a CW laser stabilized on an atomic transition

• A self-referenced frequency comb to convert the stability from the optical to the RF

domain

10

CSACs: Hot atomic vapor optical clock architecture

• The most appropriate transition for a hot atomic vapor cell optical clock is the 52S1/2

→ 52D3/2, 5/2 2-photon transition in Rb at 778.1 nm

• Any optical atomic clock requires two parts:

• An atomic reference which consists in a CW laser stabilized on an atomic transition

• A self-referenced frequency comb to convert the stability from the optical to the RF

domain

11

2-photon clocks: other realizations

• First investigations into this transition carried out in the 1990s, renewed interest recently with the

development of frequency comb technologies

• Current research lead by two groups in the USA: NIST and the Air Force Research Laboratory

• Vescent awarded 16M$ in Dec. 2021 to develop a clock with < 3·10-14 short-term stability at 1s

Ref. Configuraiton Short-term stability at 1s Flicker floor/long-

term stability

NIS

T

Newman et al.,

Opt. 2019

Clock with DBR, MEMS cell, μ-PMT, 2

Kerr freq. comb

4.4·10-12 ~10-13 floor

Maurice et al.,

Opt. Ex., 2020

Compact optical ref. with DBR, MEMS cell, μ-PMT (<500mW)

2.9·10-12 ~10-13 floor

5·10-12/day drift

Newman et al.,

Opt. 2021

Optical ref. with ECDL, ASG MEMS cell, PMT

1.8·10-13 ~10-14 floor

AFR

L Martin et al.,

PRA, 2018

Clock with doubled 1556nm

narrow-linewidth source and glass-

blown cell

4·10-13 ~10-14 floor

8·10-14/day drift

Photonic integrated

circuits

A necessary building block to

scale-down the size of optical

clocks

PIC: the optical equivalent of integrated circuits

Path towards compact, cost-effective and energy efficient optical systems

Different platform available: Si (SOI), LiNbO3, GaAs, InGaAsP, SiN, InP, SiO2, …

14

PIC in the two-photon clock

Boller et al., Photonics, 2020

Narrow CW source

Wavelength conversion

Obrzud et al., OL 2019

Kerr FC

Wildi et al., OL 2019

Obrzud et al., APL Photonics 2021

CEO frequency detection

Signal mixing / filtering, …This work

CSEM two-photon

atomic clock

16

The A-CSAC approach

17

Atomic reference unit test setup design

DBR laser

μ-PMT

glass-blown

cell

MEMS cell

provided by NIST for testing

18

Atomic reference unit test setup

DBR laser

(not used)

19

Environmental sensitivity testing: light power stability

Measured light power stability Extrapolated clock stability limitation

Out-of-loop light power stability measured

with CSEM custom power regulator (off mode)

Extrapolated effect on the clock stability

with Martin et al. coefficients is sufficient

to reach the expected performances

(10-13 @ 100s, drift < 10-13/day)

20

Environmental sensitivity testing: light power coefficient

Upper bound measured for light-shift

coefficient: 1.9·10-12/%.

Compatible with coefficient from Martin et al.

(1.4·10-13/%)

21

Environmental sensitivity testing: temperature stability

Out-of-loop temperature stability measured with an independent NTC

Extrapolated effect on the clock stability with Martin et al. coefficients is sufficient

to reach the expected performances (10-13 @ 100s, drift < 10-13/day)

Measured temperature stability Extrapolated clock stability limitation

22

Environmental sensitivity testing: temperature coefficient

Upper bound measured for temperature

coefficient: 1.1·10-12/°C.

Compatible with coefficient from Martin et al.

(1.1·10-12/°C)

23

Measurements of clock stability (glass-blown cell)

Clock short-term frequency stability close to the expected value

Limitation is likely due to the laser linewidth (observed transition linewidth of 1.5 MHz)

Long-term stability limited by a large drift already at 100s (3.5·10-12/day)

24

Measurements of clock stability (glass-blown cell)

Linear drift-removed OADEV on selected data demonstrates a potential

for 10-14 stability at 10’000 s

Selected data

26

Preliminary measurements of clock stability (MEMS cell)

Clock short-term frequency stability comparable with glass-blown cells

Limitation is also likely due to the laser linewidth

Long-term stability limited by a large drift already at 100s (9.3·10-12/day)

27

Preliminary measurements of clock stability (MEMS cell)

Linear drift-removed OADEV on the complete dataset does not average down to 10-14

Further measurements are planed outside of the frame of the project

28

2-photon clocks: Comparison of results

Ref. Configuration Short-term

stability

Flicker floor/long-term stability

NIS

T

Newman et al.,

Opt. 2019

Clock with DBR, MEMS cell, μ-PMT, 2

Kerr freq. comb

4.4·10-12 ~10-13 floor

Maurice et al.,

Opt. Ex., 2020

Compact optical ref. with DBR, MEMS cell, μ-PMT (<500mW)

2.9·10-12 ~10-13 floor

5·10-12/day drift

Newman et al.,

Opt. 2021

Optical ref. with ECDL, ASG MEMS cell, PMT

1.8·10-13 ~10-14 floor

AFR

L Martin et al.,

PRA, 2018

Clock with doubled 1556nm

narrow-linewidth source and glass-

blown cell

4·10-13 ~10-14 floor

8·10-14/day drift

CSEM

This work

Clock with DBR, MEMS cell, μ-PMT 1.6·10-12 ~10-13 floor

9.3·10-12/day drift

Clock with DBR, glass-blown cell, μ-PMT

1.2·10-12 ~10-13 floor (10-14 drift removed)

3.5·10-12/day

29

A-CSAC prototype: Compliance to ESA requirements

Parameter Req. Value Expected Value Meas value(glass cell)

Meas value (MEMS cell)

Compliance to req.

Output Frequency 10 MHz to 10 GHz (fixed)

100 MHz (fixed) 100MHz 100MHz C

Frequency stability

ADEV (1 s) 1.0E-10 to 1.0 E-11 1.0 E-12 1.21E-12 1.58E-12 C

ADEV (10 s) 3.2E-11 to 3.2 E-12 3.2 E-13 3.46E-13 4.93E-13 C

ADEV (100 s) 1.0E-11 to 1.0 E-12 1.0 E-13 1.08E-13 1.60E-13 C

Frequency Drift 1.0E-11/day to 1.0E-12/day

1.0 E-13/day 3.5E-12/day 9.3E-12/day C

Timing Error 1μs/day to 1μs/10 day

<1μs/10 day 150ns/day

15μs/10 day

400ns/day

40μs/10day

NC

Power Consumption 10 mW to 100 mW n/a (lab breadboard)

Volume 5 cm3 to 10 cm3 n/a (lab breadboard)

30

The A-CSAC approach

31

• Independant testing of 4 functionalities:

PIC design

32

• Three iterations

• Geometries designed at

CSEM

• Layout designed by

Ligentec

• Fabricaton by Ligentec

• Testing by CSEM

PIC manufacturing and testing

33

• Different concepts evaluated

• Ligentec‘s ExSpot coupling at 1550nm

• Lensed fibers into narrow, tapered WGs at 1550nm

• Lensed fibers and broad WGs at 780nm

• Results:

• At 780nm transmission of around 30% (~5dB loss) per facet

• Lower than expected for simulations but within the specs

• At 1550nm for ExSpot transmission of around 55% per facet

• At 1550nm for lensed fibers similar transmission of around 50 to

60% per facet

Input and output coupling – at 1550 nm and at 780 nm

34

• Characterization with broadband supercontinuum source

• Problem with too small spacings in design lead likely to not full

separation of the waveguides. Only a few couplers showed the

desired behavior

Dichroic characterization

35

• Characterization with supercontinuum source

Dichroic characterization on PIC2 - Reproducibility

Blue and orange curve are two different polarizations

36

PIC CEO detection

• Comparing f-2f and f-3f

• Using f-3f

• Input power: 60mW

• Input pulse duration: 120fs

• Laser rep rate: 100MHz

37

• Principle successfully demonstrated

• Integration possible with external

photodiode

PIC combination of supercontinuum and 780nm beatnote

Beat notes with the

1st nearest teeth

Rep. rate

photodiode

780nm in

1560nm in

alignment microscope

5mm

PIC

photodiode

780nm in

1560nm in

alignment microscope

5mm

PIC

Beat note with the

2nd nearest tooth

38

Conclusion

• What was shown:

• 2-photon clock with CSAC compatible elements in the atomic reference unit

reaching 10-14 floor stability, drift removed

• Required optical fuctionalities for 2-photon clock operation independently

included in a PIC

• What is left to further studies:

• A 2-photon clock reaching a 10-14 floor, drift included

• A PIC-based frequency comb with CSAC compatible dimensions

• A PIC-based narrow CW laser allowing a few 10-13 at 1s short-term clock stabiltiy

39

• The target is a compact clock with two intermediate already useable

clock but with a less challenging SWaP

Developpment plan

40

Thank you for your attention!

[email protected]