an experiment for the measurement of the dynamical Casimir effect

Surface mechanical motionStrong limitation for a moving layer: INERTIA€δxE = 1 2 ρVω 2 δx 2ω / 2 π = 10 GHzδx > 1 nmMechanical powerP ~ kW - MWTo be transferred @ 10 GHzVery inefficient technique: to move the electrons giving the reflectivityone has to move also the nuclei with large waste of energyAlso, problems with the mechanical designInstead of moving the complete layer one can produce oscillations on thesurfaces. Maximum deformations before damage δ = ω a / v s , v s sound velocityGiuseppe Ruoso - Les Houches - June 9, 2005Maximum boundary velocity ~ 50 m/s

How to produce high frequency motioni) Acoustic waves in solidsIn the 60’s Bömmel-Dransfeldproduced GHz acoustic waves inquartz using piezo excitation, butexciting all modes. Motion of asingle mode δx

Surface effective motionVan der Walls forcesmeasured at µm distances onsilicon sample change underillumination of the sampleParametric excitation of e.m.waves using a dense plasma layerin a cavity; the layer is created byirradiating a semiconductor filmwith femtosecond laser pulsesUse of laser pulse onsemiconductor to change veryrapidly its index of refractionGiuseppe Ruoso - Les Houches - June 9, 2005

Surface effective motion IIGenerate periodic motion by placing the reflecting surface intwo distinct positions alternativelyPosition 1 - metallic platePosition 2 - microwave mirror with driven reflectivityMetalplateVariablemirrorMicrowaveUSEP1 P2Semiconductors under illumination can change their dielectric properties and becomefrom completely transparent to completely reflective for microwaves.Light with photon energyhν > E band gap of semiconductorEnhances electron density in the conduction bandLaser ON - OFFOn semiconductorTime variable mirrorGiuseppe Ruoso - Les Houches - June 9, 2005

MIR - the experimental schemeA semiconductor layer (thickness ~ 1 mm)is placed on one end of a niobium superconductingcavity. Cavity resonance ν r.Using an amplitude modulated (at frequencyf) laser light the semiconductorswitches from transparency to reflection,thus producing an effective motion.whenf = 2 ν rPhotons will be produced in the vacuumand will be picked-up by the antennaWith this set-up the amplitude ofthe motion in very large, thusproviding a large effective velocityΔT = 100 psT = 1 µs..n = 1 , 2 , 3 , 4, ... . , 10 3P las = 10 nJ/100 psGiuseppe Ruoso - Les Houches - June 9, 2005

Experimental issuesEffective mirror• the semiconductor when illuminated behavesas a metal (in the microwave band)• timing of the generation and recombinationprocesses• quality factor of the cavity with insertedsemiconductor• possible noise coming fromgeneration/recombination of carriersLaser system• possibility of high frequencyswitching• pulse energy > threshold forcomplete reflectivity• number of consecutive pulsesDetection system• minimum detectable signal• noise from blackbodyradiationGiuseppe Ruoso - Les Houches - June 9, 2005

Semiconductor as a reflectorReflection curves for Si and CuLight pulseExperimental set-upResults:• Perfect reflectivity for microwave Si, GaAs: R=1;• Light energy to make a good mirror ≈ 1 µJ/cm 2Time (µs)Giuseppe Ruoso - Les Houches - June 9, 2005

Timing of the reflection processIs it feasible to make the mirror appear and disappear at GHz frequencies?mirror appearancerise time of the laser pulse(e¯ transition time ≈ fs)mirror disappearancerecombination time of e¯property of the semiconductorτ needed is 20-40 psLow Temperature Grown GaAsGiuseppe Ruoso - Les Houches - June 9, 2005

Cavity with semiconductor wallComputer model ofa cavity with asemiconductorwafer on a wallFundamental mode TE 101: the electric field Ea = 7.2 cmb = 2.2 cml = 11.2 cm=c2⎛ 1 ⎞⎜ ⎟⎝ a ⎠2⎛1⎞+ ⎜ ⎟⎝ l ⎠2f rQ L = πντ measured ≈ 3 · 10 6≅ 2.4899 GHzGiuseppe Ruoso - Les Houches - June 9, 2005600 µm thick slab of GaAs

Detection electronicsMinimum detectable signal inside the cavity

Detection electronics 2• Performed a sensitivity measurement feeding a test signalP in=10 -12 dBm = 10 -15 W Δt = 1.6 µsE RF pulse=10 -15 J/s × 1.6 10 -6 s = 1.6 × 10 -21 J = 10 -2 eV ⇒10 3 RF photonsOscilloscope averagingafter 100 sweeps:• Measured the cavity output when flashing the ligth on thesemiconductor --> NO signal detected --> NO noise inducedby the plasmaGiuseppe Ruoso - Les Houches - June 9, 2005

Laser systemPulsed laser with rep rate 5 GHz, pulse energy ~10 µJ, trainof 10 3 - 10 4 pulses, slightly frequency tunable ~ 800 nmLaser master oscillator5 GHz, low powerPulse pickerOptical amplifierTotal number of pulses limited by theenergy available in the opticalamplifier: good estimate 1000-5000(Total energy 10 - 50 mJ)Each train repeated every few secondsGiuseppe Ruoso - Les Houches - June 9, 2005

Detailed set-upWorking temperature4 - 9 KComplete freeze out ofcarriers insemiconductorNo background noisefrom thermal photonsCavity resonance in therange 2 - 3 GHzSemiconductorthickness ~ 1 mmGiuseppe Ruoso - Les Houches - June 9, 2005

Giuseppe Ruoso - Les Houches - June 9, 2005The apparatus

Giuseppe Ruoso - Les Houches - June 9, 2005Inside the cryostat

Detection schemeN pulsesSteps5. Look for signal with τ ~ Q / 2πν r1. Find cavity frequency ν r2. Wait for empty cavity3. Set laser system to 2 ν r4. Send burst with > 1000 pulsest p= 1/ 2 ν rCharged cavity.Will decay with itstime constantExpected number of photons:Niobium cavity with TE 101 ν r = 2.5 GHz (22 x 71 x 110 mm 3 )Semiconductor GaAs with thickness δx = 0.6 mmSingle run with ~ 5000 pulsesN ≥ 10 4 photonsGiuseppe Ruoso - Les Houches - June 9, 2005

Check listSeveral things can be employed to disentangle a real signal from aspurious oneSignal (a.u.)Change temperature of cavityEffect on black body photonsChange laser pulse rep. frequency1.61.41.210.80.60.85 0.9 0.95 1 1.05 1.1Laser pulse frequency (a.u)Loading of cavity with real photons (isour system a microwave amplifier?)Power inside cavity at end of laser pulses (a.u.)14121086420Determine vacuum effect from severalmeasurements with pre-loaded cavity0 1 2 3 4 5 6Power inside cavity at t = t0 (a.u.)-change recombination time of semiconductor-change width of semiconductor layerGiuseppe Ruoso - Les Houches - June 9, 2005

ConclusionThe experiment MIR seems feasibleFinish R & D within 1 year• complete laser system with high power part• develop a system to characterize semiconductors (autocorrelation)Giuseppe Ruoso - Les Houches - June 9, 2005