Orpheus: Extending the ADMX QCD Dark-Matter Axion Search to Higher Masses 3rd Workshop on Microwave Cavities and Detectors for Axion Research Gianpalo Carosi, Raphael Cervantes , Seth Kimes, Parashar Mohapatra, Rich Ottens, Gray Rybka University of Washington 08/23/2018 1 / 27
Outline 1. Detecting Axions with Haloscopes. 2. Motivation for Dielectric Haloscopes and Open Resonators. 3. Orpheus Concept. 4. Orpheus Progress. 5. Outlook. 2 / 27
Searching for Axions with the ADMX Haloscope 2 V eff = | dV � B ext · � E a | � P a ∝ B 2 ext QV eff , 2 dV ε r | � E a | B 2 � ext 3 / 27
Becomes increasingly difficult at higher frequency 2 dV � B ext · � V eff = | E a | � P a ∝ B 2 ext QV eff , 2 dV ε r | � B 2 E a | � ext � � Resonator wall E a B ext � � dV � B ext · � � E a � > 0, V is small � � � � � dV � B ext · � � V is large, � = 0 E a � � � 4 / 27
Solution: Dielectric Haloscopes Higher frequency with more volume and better axion coupling. 2 V eff = | dV � B ext · � E a | � P a ∝ B 2 ext QV eff , 2 B 2 dV ε r | � E a | � ext � � Resonator wall Dielectric E a B ext Low-loss dielectric ∼ λ/ 2 thick placed every other half-wavelength. � � dV � B ext · � � V is large, E a � > 0 � � � 5 / 27
ADMX Orpheus Concept Goal : Dielectrically Loaded Fabry-Perot Open Resonator threaded by a dipole magnet. Tunes with cavity length. Search for axion-like particles at 15-18 GHz. Open → Less ohmic losses → higher Q. Open → Sparser spectrum → less mode crossings. Challenges : ◮ Designing optics. Maintain good mode with high Q from 15-18 GHz. ◮ Moving mirror and dielectrics in LHe. ◮ Obtaining dipole magnet. 6 / 27
Orpheus Science Reach Within a few years. + 10.1103/PhysRevD.91.011701 *arXiv:1403.4594 7 / 27
Progress Towards Orpheus: Preliminary Cryogenic Design Submerged in LHe. Think of an accordian! 8 / 27
Room-Temperature Prototyping Cheap and fast prototyping. Test mechanics and electronics. Strategy 1. Study empty cavity. Get analytical solution, simulations, and measurement to agree with each other. Gain confidence. 2. Load cavity with dielectrics. Simulate and measure. See if they agree. 9 / 27
Standing waves in FP Cavities Sustain TEM mnq modes. Think of Gaussian beams. m, n: nodes in transverse plane. q: nodes along cavity. Analytical formula: f mnq = ( q + 1) f o + ( f o /π )(1 + m + n ) cos − 1 (1 − 2 L / r o ), f o = c 4 L Mirror focusing to reduce diffraction. Mode tunes with cavity length. 10 / 27
Simulation Transmission of TEM 00 − 18 mode Simulation 130 Lorentzian Fit Q = 14688.2 18 = 16.005 GHz f 00 140 S21 (dB) 150 160 170 15.900 15.925 15.950 15.975 16.000 16.025 16.050 16.075 16.100 Frequency (GHz) Simulated for different cavity lengths. 11 / 27
Measure Transmission of TEM 00 − 18 mode Cavity Length = 17cm (0, 0, 17) (0, 0, 18) (0, 0, 19) 10 (1, 1, 18) 20 (1, 1, 17) S21 (dB) 30 40 50 15.0 15.5 16.0 16.5 17.0 17.5 18.0 Frequency (GHz) Q L between 1000 and 5000. Measured Qs don’t match simulation, perhaps because of different coupling. 12 / 27
Empty Orpheus Mode Map 18.0 Analytical (0,0,18) 17.5 17.0 Frequency (GHz) 16.5 16.0 15.5 15.0 16.0 16.5 17.0 17.5 18.0 18.5 19.0 Cavity Length (cm) 13 / 27
Empty Orpheus Mode Map 18.0 Analytical (0,0,18) Simulation (0,0,18) 17.5 17.0 Frequency (GHz) 16.5 16.0 15.5 15.0 16.0 16.5 17.0 17.5 18.0 18.5 19.0 Cavity Length (cm) Simulations agree with analytical formula. 14 / 27
Empty Orpheus Mode Map 18.0 0 Analytical (0,0,18) Simulation (0,0,18) 0,0,18 17.5 10 Measured S21 (dB) 17.0 20 Frequency (GHz) 16.5 30 16.0 40 15.5 50 15.0 60 16.0 16.5 17.0 17.5 18.0 18.5 19.0 Cavity Length (cm) Resonant frequencies for analytical formula, simulation, and experiment agree! 15 / 27
Empty Orpheus Mode Map 18.0 0 Analytical (0,0,18) 0,0,19 Simulation (0,0,18) 0,0,18 17.5 10 17.0 Measured S21 (dB) 20 Frequency (GHz) 0,0,17 16.5 30 16.0 40 15.5 1,1,18 50 1,1,17 15.0 60 16.0 16.5 17.0 17.5 18.0 18.5 19.0 Cavity Length (cm) Resonant frequencies for analytical formula, simulation, and experiment agree! Other modes where predicted. No mode crossings! 16 / 27
Room-Temperature Prototyping with Delrin TEM 00 − 18 mode is the good mode for axion coupling. Can we track this mode while we tune it? 17 / 27
Delrin: Predict resonant frequency through simulation Simulation 120 Lorentzian Fit Q L = 15238.5 18 = 17.107 GHz f 00 130 140 S21 (dB) 150 160 170 17.000 17.025 17.050 17.075 17.100 17.125 17.150 17.175 17.200 Frequency (GHz) ε r = 3 . 7 at 100 Hz. Don’t know at 15-18 GHz. Simulation done with lossless Delrin. Q L comparable to empty resonator for this frequency at this cavity length. Q depends highly depends on loss tangent and impedance changing parameters (e.g. mirror thickness, hole aperture size). 18 / 27
Delrin setup: Measure Transmission of TE 00 − 18 mode 32.5 Measurement Lorentzian fit 35.0 37.5 40.0 S21 (dB) 42.5 Q = 135.2 45.0 F = 16.059GHz 47.5 50.0 15.6 15.8 16.0 16.2 16.4 Frequency (GHz) Q is much lower, as expected. Delrin is very lossy. Will need better dielectrics to understand substructure. 19 / 27
Delrin setup mode map 18.0 17.5 30 17.0 Measured S21 (dB) Frequency (GHz) 35 16.5 40 16.0 45 15.5 15.0 50 12.5 13.0 13.5 14.0 14.5 15.0 Cavity Length (cm) Mode structure is apparent but messier. Expected with lower Q. 20 / 27
Delrin setup mode map 18.0 TEM 00 18 resonances Simulation 17.5 30 17.0 Measured S21 (dB) Frequency (GHz) 35 16.5 40 16.0 45 15.5 15.0 50 12.5 13.0 13.5 14.0 14.5 15.0 Cavity Length (cm) Mode structure is apparent but messier. Expected with lower Q. Simulation resonances agree with experiment. Should be better with improved mechanical structure. Also, don’t know ε r . 21 / 27
Delrin setup mode map 18.0 TEM 00 18 resonances Simulation experiment 17.5 30 17.0 Measured S21 (dB) Frequency (GHz) 35 16.5 40 16.0 45 15.5 15.0 50 12.5 13.0 13.5 14.0 14.5 15.0 Cavity Length (cm) Mode structure is apparent but messier. Expected with lower Q. Simulation resonances agree with experiment. Should be better with improved mechanical structure. Better Q would allow us to understand substructure. 22 / 27
Progress Towards Orpheus: Magnet Making 3,250 windings. Niobium titanium wire 0.3 mm in diameter. 1 T. Prototyping different manufacturing methods. 23 / 27
Next Steps ◮ Improve mechanical design. ◮ Improve and understand optics. ◮ Develop DAQ, electronics, and motor systems. ◮ Cryogenic tests. First results in 2-3 years. 24 / 27
How to get to DFSZ sensitivity Scan rate equation from ADMX hf Assume Quantum Limited Amplifiers. Then T sys = 2 k B = 0 . 43K. Let Q L = 10 5 , SNR = 3 . 5 , V eff = VC lmn , f = 18GHz. If dt = 1GHz / year, then B 2 V eff = 200LT 2 df 25 / 27
Summarize 1. Dielectric haloscopes could look for ∼ 100 µ eV axions. 2. ADMX Orpheus experiment is a haloscope consisting of a Fabry-Perot Cavity loaded with evenly-spaced dielectrics. It will explore 15 to 18 GHz. 3. Room-temperature table top resonators have been characterized. Improvements underway. 4. Cryogenic setup in development. 5. First results in 2-3 years. 26 / 27
Acknowledgements Support was provided by the Heising Simons Foundation and the U.S. Department of Energy through Grants No. DE-SC0009723,No. DE-SC0010296, No. DE-FG02-96ER40956, No. DEAC52-07NA27344, and No. DE-C03-76SF00098. 27 / 27
Recommend
More recommend
