Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Axion Detection with Precision Frequency Metrology

Axion Detection with Precision Frequency Metrology 1 1 1, a) Maxim Goryachev, Ben T. McAllister, and Michael E. Tobar ARC Centre of Excellence for Engineered Quantum Systems, School of Physics, University of Western Australia, 35 Stirling Highway, Crawley WA 6009, Australia (Dated: 12 September 2019) We investigate a new class of galactic halo axion detection techniques based on precision frequency and phase metrology. Employing equations of axion electrodynamics, it is demonstrated how a dual mode cavity exhibits linear mode-mode coupling mediated by the axion upconversion and axion downconversion processes. The approach demonstrates phase sensitivity with an ability to detect axion phase with respect to externally pumped signals. Axion signal to phase spectral density conversion is calculated for open and closed loop detection schemes. The fundamental limits of the proposed approach come from the precision of frequency and environment control electronics, rather than fundamental thermal uctuations allowing for table-top experiments approaching state-of-the-art cryogenic axion searches in sensitivity. Practical realisations are considered, including a TE-TM mode pair in a cylindrical cavity resonator and two orthogonally polarised modes in a Fabry-P erot cavity. INTRODUCTION axions creating additional photons with a di erent fre- quency (such that energy is conserved) at a rate related to the number of pre-existing photons and the number of Axions are theoretical weakly-interacting sub-eV par- 1 axions. Prior work has considered a single pumped RF ticles that can be formulated as a primary component of cavity mode interacting with axions, followed by detect- dark matter. With mounting evidence pointing towards 13,14 2,3 ing a small signal in an orthogonally polarised mode . lower mass particles , particles such as axions are be- In this work we consider more general cases of two mode- coming increasingly promising dark matter candidates. axion interactions. Axions arise as a result of an elegant solution to the 4 Despite the apparent similarity between the DC and strong CP problem in QCD , and are expected to have 5 AC detection schemes, they belong to di erent classes properties consistent with dark matter . Confounding of detectors. Since virtual photons or static elds carry experimental e orts to detect axions is the fact that the no phase, the traditional Sikivie haloscope detectors (us- axion mass is largely unknown, with only weak bounds ing DC magnetic or electric elds) belong to the class from cosmology and theory. Despite this, a number of 6{11 of phase insensitive systems. On the other hand, the axion detection experiments are already underway . AC scheme considered in this work relies on pump- In perhaps the most common axion detection tech- ing signal(s) carrying relative phase as well as sepa- nique, it is generally agreed that dark matter axions rate phases relative to the axion signal. Thus, the de- can be detected via the (inverse) Primako e ect, a tected signal as well as the overall result would have two photon-axion interaction. Nowadays, in a typical a footprint of these phases. This fact draws analogies (Sikivie) detector , these hypothetical particles interact with existing ampli ers that can be grouped into DC with a DC magnetic eld, or virtual photons, to produce (phase insensitive) ampli ers, where energy is drawn real photons whose frequency corresponds to the mass from static power supply, and parametric (phase sen- of axions. This scheme employs one or several tuneable sitive) ampli ers, where energy comes from oscillating microwave cavities, serving as resonant antennas, with elds. The second type gives more freedom allowing im- the output coupled to the lowest noise ampli ers so gen- proved ampli cation/detection schemes based on quadra- erated photons may be detected with the greatest sen- ture squeezing . Thus, in this work we expand investi- sitivity possible. In principle, a similar detector using gations into a novel class of axion detectors employing static electric eld is also possible although due to consid- the phase sensitive approach. erable mismatch between electric and magnetic compo- nents and relative diculty in creating extremely strong electric elds in large volumes, such detectors are never I. AXION ELECTRODYNAMICS DESCRIPTION realised in practice. The third Sikivie-like axion detection technique is represented by the type of detectors utilising The Hamiltonian density of the photon-axion system RF or microwave elds instead of static ones: indeed, the consist of the conventional electromagnetic, axion and Primako process works equally well with real photons 13,14 interaction parts: instead of virtual ones . In these schemes the exist- ing photons of a given frequency interact with cosmic H = H +H +H : (1) EM a int The free electromagnetic Hamiltonian density is usually represented using the vectors of electric and magnetic a) Electronic mail: michael.tobar@uwa.edu.au elds, E and B respectively, or a vector potential and its arXiv:1806.07141v3 [physics.ins-det] 11 Sep 2019 2 conjugate momentum, A and : as (moving from Hamiltonian density to Hamiltonian): Z Z h i h i " 1 1 3 3 2 2 2 2 2 2 H = " cg  d rE  B + d rE  B ; (6) int 0 a 1 2 2 1 H = E + c B =  + " c (r A) ; EM 0 2 2 " V V (2) where it is assumed that material properties dictate that for each mode E  B = 0. Quantising the two modes, i i where c is the speed of light, and " is the dielectric one arrives at the following interaction Hamiltonian in permittivity of free space. terms of creation-annihilation operators: The axion part of the system for a laboratory size ex- periment may be represented by the uniform eld  and ~g y y H = i ! !  (c + c )(c c )+ int 1 2 1 2 1 its canonical conjugate : 2 1 (7) y y 2  (c + c )(c c ) 2 1 2 1 2 (3) H = + V (); 2m where two dimensionless coecients  and  represent 1 2 overlap between the two modes: where m is axion mass. Choosing the normal harmonic m ! a 2 potential V () =  , this Hamiltonian reduces to 1 = p d r(e  b ); 1 1 2 a simple harmonic oscillator. On the other hand, for an V V 1 2 V (8) experiment of reasonable size and duration, this represen- = p d r(e  b ): tation is excessive; axion dynamics cannot be observed as 2 2 1 V V 1 2 V the associated time constant should be extremely large compared to the experimental duration. The apparent These coecients can span from one, when two modes nite `quality factor' of axion-induced photon signal is are of the same shape and fully orthogonal, to zero when due to the velocity distribution of axion dark matter in they exhibit no overlap. Finally, it is possible to separate the galactic halo, rather than due to response time of swapping and parametric parts: an `axion mode'. In this case, as it is usually done, the h i y y y y axion part is simply represented as an external signal H = i~g   (c c c c ) +  (c c c c ) ; (9) int e 1 2 + 1 2 2 1 1 2 with amplitude  and angular frequency ! that can be varied in time due to axion velocity. where  =   , and g = ! ! is the e ective 1 2 e 1 2 It is widely accepted that the axion-photon interaction trilinear coupling. part in a laboratory size experiment can be represented A particular type of dynamics described by the interac- in the following form: tion Hamiltonian (7) depends on the resonant frequencies of the photon modes ! and axion signal ! . In partic- i a " cg  ular, the following regimes can be identi ed (assuming 0 a (4) H = " cg  E B =  (r A); int 0 a ! ! > 0): " 2 1 where  is a scalar axion-like eld, and g is the com- ! = ! + ! ; axion downconversion a a 2 1 (10) monly presented axion-photon coupling constant. ! = ! ! ; axion upconversion a 2 1 where neither of the angular frequencies is zero. A case II. TWO MODES AXION ELECTRODYNAMICS where one of cavity frequencies is zero would correspond to the Sikivie detector with virtual photons represent- ing one of the modes. This case has been widely studied We consider an electromagnetic cavity with two modes theoretically with a few experimental realisations in dif- of two angular frequencies ! and ! . Given a resonant 1 2 6,16 ferent frequency ranges . structure, each mode n is characterised by a certain dis- To demonstrate the di erence between the upconver- tribution of electric (E (r)) and magnetic eld (B (r)) n n sion and down conversion cases, we transform the system in a certain nite volume: into the rotating frame associated with the frequencies of both modes and apply the rotating wave approximation E (r) =  u (r) = iE (c c )e (r); n n n V;n n n (RWA). For this purpose, the axion signal  may be de- (5) composed in terms of complex amplitudes a exp(i! t) B (r) = A r u (r) = E (c + c )b (r); n i n V;n n n and a exp(i! t). Conditions (10) give the two follow- n a ing Hamiltonians after corresponding RWA when fast ro- tating terms are removed (all couplings and signals are where c (c ) are creation (annihilation) operators for small as dictated by the weak signal detection problem): ~! the mode n, E = , and e (r) and b (r) are V;n n n 2" V 0 n y y unit vectors representing the mode polarization. H = i~g  (ac c a c c ); D e + 1 2 1 2 (11) Each of these two photonic modes is coupled to an y y H = i~g  (a c c ac c ); U e 1 2 2 1 axion signal via the term in Eq. (4) that can be rewritten 3 (A) to-noise ratio of the corresponding axion detector. The second class of strategies is related to measuring photon mode frequency shifts introduced by the axion mediated coupling terms (11) and has previously been discussed in the context of other hidden sector particle searches . In this approach, discussed in detail in Section IV, instead of detecting power coming from the modes, one is look- (B) ing for mode frequency deviations associated with the new physics. III. POWER DETECTION APPROACH The excess power detection method is the most widespread approach to weak signal detection including the axion and paraphoton searches. According to this FIG. 1: Graphical representation of the two modes paradigm, one is looking for extra power coming from an interacting through axion coupling in (A) upconversion experimental setup over the expected noise oor. Typ- and (B) downconversion cases. ically the noise oor is associated with thermal uctua- tions of a measured device, i.e. cavity, and noise temper- ature of the amplifying electronics. Thus, such a scheme in the corresponding interaction pictures. Graphical rep- is usually realised in a cryogenic environment minimising resentations of these processes is given in Fig. 1. In all thermal uctuations. both cases, the problem is reduced to a system of two The same approach can be realised using the analysed piecewise coupled modes, as a may be understood as a double mode system. Although, the potential for exter- complex coecient. The downconversion case Hamilto- nal signal pumping of oscillating elds brings new possi- nian represents the parametric interaction found, for ex- bilities to the problem. Indeed, the axion coupling may ample, in the blue side band regime of optomechanical be regarded as a mixing term in the presence of strong ex- systems .The upconversion case is a swapping (or beam ternal pumping resulting in certain analogies that can be splitter) interaction that corresponds to the red side- 22,23 drawn with parametric ampli ers , which have found band regime in optomechanics . For the upconversion numerous applications on the forefront of physics. Para- and downconversion cases, the corresponding Heisenberg metric ampli ers' phase sensitivity and related phenom- equations of motion are respectively: ena allow them to beat the quantum or thermal limit or any other physical constraints to achieve extraordinary in c = i! c c g  ac i 2 b ; i i i i i e j i levels of sensitivity. Such devices could be built on di er- dt (12) ent physical principles, for example, nonlinearities from in c = i! c c + g  ac i 2 b ; i i i i i e + i j i superconducting junctions, nonlinear crystals, mechani- dt 24{28 cal resonators or magnon systems . To analyse such in where j 6= i, and b are a loss rate and input signal systems, it is customary to split the problem into two for the nth mode respectively. steps by splitting variables into nonlinear large pumped Given the systems described by the Hamiltonians (11), amplitudes and small uctuations giving linearised equa- one can imagine two classes of detection strategies. The tions of motions . In Appendix A, we follow this line rst class relies on excess power detection introduced by of thought to analyse the sensitivity of the externally the axion-photon coupling. This approach, further dis- pumped double mode axion detector in the downconver- cussed in Section III, is an extension to standard axion sion and upconversion cases. The result (Eq. (A4) for 6,12,16 detectors using strong DC magnetic elds where the upconversion and Eq. (A6) for the downconversion) the role of the DC eld is played by the modes' oscillat- suggests that the power in one mode due to an axion ing elds. Besides the numerical values for the physically signal is proportional to the power stored in the other achievable eld strengths, the major di erence between mode. This result is in accordance with the standard the current proposal and the standard virtual eld tech- DC magnetic eld detector where the axion power at nique is the appearance of phases, making the experiment the cavity output is proportional to B , squared mag- DC phase sensitive. It is worth noting that the Hamiltonian netic eld strength, i.e. magnetic eld energy. Regarding for the downconversion case may be interpreted as a tri- the amplitudes, rather than powers, the main di erence linear Hamiltonian coupling pumped cavities to an ex- is additional phase  (upconversion) or + (down- m m ternal signal. This Hamiltonian has been considered pre- conversion) that is not measurable with a standard power 18 14 viously in the literature in di erent physical contexts detector . It is worth noting that the resulting spectral 19,20 including its application to parametric ampli cation . density from Eq. (A4) is identical to the previously pro- This property might be exploited to increase the signal- posed method in the case when one mode is pumped on 4 resonance and the signal is observed through the other A. Axion Induced DC Frequency Shifts mode. In practice, direct power detection using the dou- ble mode approach would be inferior to the scheme in- The system equations of motion in the interaction pic- volving virtual photons (DC magnetic eld) due to prac- ture for the downconversion (10) can be written in the tical impossibility to create oscillating elds of matching following form: strength. Instead of measuring the power of uctuations from C = ( i )c + g  AC ; 1 1 1 1 e + dt a single mode, a better strategy is to measure cross- (14) correlation between the two modes of the system. As- C = ( i )c + g  AC ; 2 2 2 2 e + suming statistical independence of the eld uctuations dt in coming into the modes b , the cross correlation spectrum where C and C are slowly varying amplitudes, is the 1 2 n of both channels is reduced to: loss rate for the mode n,  is the detuning frequency of a cavity with angular frequency ! such that  + i 1 2 2 = ! +! ! , and A is the axion eld. Introducing g  C C 2 1 2 a 1 2 D/U e i(  ) 2 1 (13) S [ ] = e S [ ]; =  i , the eigenvalues of this system are n n n j(i )(i )j 1 2 + 1 D 2  2 2 2 2 e =  ( ) + 4g  jAj where upconversion and downconversion cases are dif- 2 e 2 2 (15) ferent only through the phase factors giving the identical 2 2 2 2 =  i  ( i ) + g  jAj + + cross-correlation spectrum (13) with a di erent multiplier e + !  due to di erent mode overlapping integrals (8). as well as their complex conjugates, where = ( It means that by observing the power, the two cases are )=2 and  = (   )=2. 2  1 2 indistinguishable. One consequence of this is that by a For the upconversion case, the equations of motion single measurement, one searches for axions in both the could be written as follows: upconversion and downconversion regimes. The sensi- tivities of these cases can vary because of the di erent C = ( i )c g  AC ; overlap integrals. In case of a candidate detection, one 1 1 1 1 e 2 dt (16) would need either to modify geometry and thus  or ver- ify the two cases separately with a di erent combination C = ( i )c + g  A C ; 2 2 2 2 e 1 dt of frequencies ! and ! . 1 2 with the eigenvalues: + 1 1 2 U 2 2 2 2 e =  ( ) + 4g  jAj 1 2 2 2 (17) IV. FREQUENCY MEASUREMENT APPROACH q 2 2 2 2 =  i  ( i ) + g  jAj + + Frequency measurement techniques have found consid- and their complex conjugate. erable attention in precision sensing technology. They In the simplest and most sensitive form, one tunes both have been successfully applied in such elds such as par- detuning frequencies to zero ( = 0) and match the ticle detection, bio-sensing, magnetic eld and mass sen- losses ( = ). The resulting relative shift of eigenfre- 1 2 29{31 21,32{34 sors, etc , as well fundamental physics tests . quencies due to coupling to axion signal of amplitude jAj Unlike the power detection scheme, the fundamental is limit in frequency measurements comes from the best achievable frequency stability of the measurement parts D D D 2 2 2 2 @e = e e = g ! !  jAj ; 1 2 + a + + and thus it is not directly related to the Nyquist noise (18) U U U and the ambient temperature, although cryogenic cool- @e = e e = g ! !  jAj: a 1 2 ing may signi cantly improve the detection limit due to Whereas for the upconversion case, splitting is directly higher quality factors. It is generally accepted that one proportional to the axion amplitude, in the downconver- can detect frequency variations on the order of 10 rela- sion case only imaginary part is sensitive to the pres- tive to the system linewidth. With such systems as super- ence of axions. To compare the two cases of detection conducting cavities and sapphire, it is possible to achieve in more detail, we calculate deviation of real and imagi- linewidths of the order of 1 Hz at microwave frequencies nary components of the eigenvalues from its values in the giving unprecedented sensitivity of the frequency mea- axion-free case (jAj = 0) as a function of the normalised surement approach. g  jAj eff axion coupling strength  = shown in Fig. 2. In this section we consider a possible application of fre- quency metrology to axion searches enabled by the pro- The upconversion case (dashed line) demonstrates the ex- posed dual mode approach. As before, we split the dis- pected linear dependence of the real component of eigen- cussion into the downconversion and upconversion parts. frequency deviation on the coupling for small and the energy conservation, but rather indicate the regime in which the electromagnetic modes are slightly detuned from the axion frequency. Such a regime is analogous to the situation in which the cavity mode frequency of the standard DC haloscope is slightly shifted from the axion frequency. The detuning frequency f plays the role of the Fourier spectrum in the generated noise and could span over a few decades. In this situation, the axion amplitude appears as a slowly varying parameter A in the EOMs (14) and (16). The detailed analysis of this approach is given in Appendix B where we reformulate the problem 36{38 in terms of slowly varying real magnitude and phase instead of the complex amplitude representation (14) and (16). The analysis in Appendix B, reveals a transfer function FIG. 2: DC sensitivity of the dual mode detector from one of the axion quadratures to the phase uctu- eigenvalues to the normalised axion coupling strength . ations of the output signal (Eq.(B10) for upconversion and Eq. (C1) for downconversion). Using this transfer function, the phase noise spectrum of the output signal as stated by Eq. (18). This regime demonstrates eigenfre- for the ith mode may be represented as follows: quencies shifts equal to axion signal strength normalised 2 2 to frequency units. This case requires precise matching g  x D/U  j e i S (f ) = S (f ) + S (f ); (19) ';i of the modes in terms of losses and detuning frequen- 2 2 2 2 f + x f + i i cies. With nonzero di erence between mode losses where the rst component is the spectrum of phase in- the sensitivity in the lower coupling limit decreases and duced by the axion signal, and S (f ) is the power spec- is limited only to the imaginary part of the eigenvalue. A trum of the axion eld, in units of jAj , where A repre- In the downconversion case, although the sensitivity sents the axion eld, and the second term is due to tech- is reduced, the axion coupling appears as a modi cation nical phase uctuations S (f ) of the pump signal. The to the imaginary part of eigenfrequencies: the linewidth result explicitly depends on the ratio of steady state am- gets narrower when coupling increases. The change in the plitudes x in both modes. Thus, the overall sensitivity imaginary part is shown with the green curve in Fig. 2. i may be boosted using this ratio. It is also important to At  = 1, the axion term balances the sum of the losses note that despite the fat that the axion signal is ltered leading to the steady state oscillation regime. In this by the resonator (appears as the rst order transfer func- regime, photons created by axion downconversion, bal- tion in the phase space), the overall signal-to-noise ratio ance the cavity losses and the system starts oscillat- is constant as the technical uctuations are also ltered ing. Since the axion signal is extremely weak, this regime by the same ltering function. is not achievable even for the most narrow-linewidth cav- The phase noise spectrum (19) could be measured us- ities. In order to boost the sensitivity of the frequency ing the phase measurement setup shown in Fig. 3. Here, measurement technique in the downconversion case, one the output of each cavity is mixed with a local oscillator may look for compensation of the cavity losses.Indeed, of the same frequency as the pump. By varying the phase by introducing external or internal gain into one or two between the pump signal incident on the cavity and the modes, the e ective cumulative losses decrease result- mixer , one can access to both quadratures of the eld ing in lower oscillation threshold. uctuations in the cavity. Such an approach is capable in principle of detecting phase uctuations in rms amplitude of 2 10 rad= Hz at Fourier frequencies above a few B. Open Loop Axion Induced Spectral Density of Phase kiloHertz . Such kind of measurement is not possible Measurements with the traditional DC- eld axion detector. The sensitivity plots of the pumped phase noise mea- In practice, measurement of DC frequency shifts is surement experiment is shown in Fig. 4. These sensitivi- technically challenging. For fundamental tests the sit- ties are based on comparing the size of the axion-induced uation is worsened by the requirement to verify or veto phase-shifts, which are taken to be of magnitude the candidate signals. The problems may be solved by modulating one of the system parameters as, for exam- 1 g ! ! x a 1 2 D/U j S =  jAj ; ple, proposed in the case of a paraphoton search . An- ';i;a 2 f + 2 x other approach is to search for axion signals in the Fourier spectrum. For such an approach, we consider a regime S (f ) with the minimum detectable phase-shift where of small axion detuning: ! = !  ! + 2f where a 1 2 2f  ! . Note that these relations do not violate in this case the background noise, S (f ) is the phase 1  6 Phase Shift Filter Cavity CAST Phase Shift FIG. 3: An open loop realisation of the dual mode axion detection scheme. Two external local oscillators (LO) are used to excite two axion coupled modes. FIG. 4: Sensitivity of the cavity phase noise noise spectral density of the output we are measuring (as measurement experiment comparing to the axion opposed to the phase noise spectrum, S ), and t is the models and existing limits due to the CAST experiment. averaging time . The green and blue lines in g. 4 represent the poten- Phase Shift tial sensitivity to the up and downconversion regimes of this kind of experiment. The pump signal is taken to orig- Filter Cavity Gain inate from a state of the art, frequency stabilized cryo- genic sapphire oscillator (CSO) operating at the noise oor of a partially cryogenic frequency discrimination Gain system and white noise background. We further as- sume that the ratio of powers in the two modes is 1000, such that the readout mode has 1000 times less power dissipated in it. The mode structures, quality factors, resonance details and overlap integrals are discussed in Phase Shift section VI. Generally speaking, such a search would op- erate by detuning the two modes by some distance in fre- FIG. 5: Loop oscillator approach to dual mode axion quency space, and then searching the Fourier phase-noise detection. Two positive feedback loops utilise axion spectrum of the one of the resonances (say the higher fre- coupled modes as frequency selective elements. quency resonance for instance) for peaks corresponding to axion-induced phase-shifts. C. Loop Oscillator Axion Induced Spectral Density of Of course, for a given resonator geometry the upcon- Phase Measurements version technique will never be able to reach the same high-mass range as the downconversion technique, and Instead of relaying on an external frequency source, so the extension of the green line into the same frequency range as the blue line in g. 4 represents the sensitivity if one may construct an oscillator using the double mode cavity as a frequency selective element. Such a measure- we were to construct some resonator with the same sen- sitivity, but with mode frequencies such that this range ment setup is shown in Fig. 5. Discussion of such a sys- tem is given in Appendix C where the spectrum of phase was achievable with the upconversion technique. This is uctuations is given as follows: presented for the purpose of the direct comparison of the two techniques. By searching a few MHz in Fourier space h 2i 2 2 for such peaks, we are sensitive to slightly detuned ax- g  x D/U e  j S (f ) = 1 + S (f ) + S (f ) ; ';i 2 2 2 ions such that ! = !  ! + 2f , where f is our range a 1 2 f f + x of Fourier frequencies. We may then further detune the (20) resonances and repeat the process, gradually excluding a large section of the axion mass range. Of course, aver- where S are technical uctuations inside the oscillator aging for longer can yield improvements, and the above loop, and S (f ) is de ned as in (19). The multiplying plots are based on an averaging time of 30 days per 100 factor in this result is due to the Leeson e ect . MHz. The amount of frequency space covered in that The sensitivity plots of the loop oscillator experiment 30 day period will then depend on the range of Fourier for the room temperature and cryogenic implementations frequencies measured. are shown in Fig. 6. These sensitivities are calculated in KSVZ/DSFZ Models 7 consider di erent hybrid implementations. For example, one mode is used with a feedback loop as an oscillator, and the other is employed to measure the axion coupling induced phase shift with a help of external (highly stable) CAST frequency source. V. DEGENERATE MODE BROADBAND EXPERIMENT An interesting class of detection techniques may be de- veloped based on the degenerate frequency case scenario, where we require the two modes have equal frequencies (! = ! = !), and axion frequency (mass) is small 1 2 (!  !). Under these requirements, only the upconver- sion case is feasible. In the degenerate frequency case, considering only pure tones would put an exclusive re- quirement on the axion mass being near zero. The ac- FIG. 6: Sensitivity of the room temperature and tual signal would be searched for in a wide range of o set cryogenic versions of the loop oscillator experiment (Fourier) frequencies as discussed in Section IV and de- comparing to the axion models and existing limits due veloped in Appendices B and C. The obvious advantage to the CAST experiment. The dashed line is the of this approach is its broadband nature. Indeed, by mea- extension of the upconversion non-degenerate case down suring phase and amplitude noise in the degenerate case, to low frequencies. one has direct access to a few decades of Fourier frequen- cies. Such broadband measurements could be directly realised with the modern recon gurable digitizers. a similar fashion to the externally pumped loop experi- To analyse the Equations of Motion in the degenerate ment above. Indeed, the experiment would operate in the case, one may apply the same logic as in Appendix A, same way, by detuning the resonances some amount and B, C.Thus, the results obtained in Section III and IV are searching Fourier space, before detuning the resonance valid for the degenerate case as well. further and repeating the search. We again compare the The possibility to design a cavity with two orthogonal magnitude of axion-induced phase shifts - in this case modes of the same frequencies have been demonstrated given by 43,44 before . For the microwave frequency range, one may s argue that the mode degeneracy is not achievable due 1 p x to unavoidable imperfections of a cavity. Such imper- D/U j S = 1 + g ! !  jAj ; a 1 2 ';i;a 2 2 fections introduce coupling between the two modes that f f + x results in avoided level crossings. And it is due to this phenomenon the modes never coincide in the frequency S (f ) with the minimum detectable phase shift where space. Though, this is true, it is feasible to match the in this case the background noise, S (f ) is the phase frequencies within the mode bandwidth: by minimising noise spectral density of the loop oscillator that we are the all sorts of "imperfections", e.g. strongly coupled ex- measuring the output of (as opposed to the phase noise ternal probes, one may reduce mode-mode coupling to a spectrum, S ), and t is the averaging time. value that is much smaller than the mode bandwidth. In this case, no avoided level crossing could be observed. An For the cryogenic case, we are again assuming that the oscillator is a state of the art CSO operating at the example of a microwave cavity that can be used for the degenerate two mode axion sensing is discussed further noise oor of a partially cryogenic frequency discrimi- nation system and white noise background, and for the in Section VI. room temperature measurements we assume frequency Another practical problem immediately encountered stabilized loop oscillators based on copper cavities with with the degenerate frequency measurement scheme is in- the same properties as the open loop experiment, again ability to apply lters and separate the two signals. For discussed in more detail in section VI, and ampli ers with example, in the frequency detection scheme with two os- e ective added noise temperatures of 50 K, dissipating 1 cillators, the oscillators will synchronize due to unavoid- W in the resonator, operating at the noise oor of the able coupling between the two resonances of the same frequency discriminator and white noise background. It frequency. A possible solution to this problem is to treat is important to note that for the upconversion experi- the signals as one and use an external clock as a reference ment the sensitivities linearly decrease down to very low as depicted in Fig. 7. values of the frequency. Projected sensitivity of the degenerate broadband ex- In addition to purely frequency and phase measure- periment operated at both room and cryogenic temper- ment schemes presented in Fig. 3 and Fig. 5, one may ature ranges is shown in Fig. 8. This plot is calculated KSVZ/DSFZ Models 8 Phase Shift Filter Cavity Gain CAST SN1987A FIG. 7: A possible realisation of the degenerate mode detection scheme where two axion coupled modes of the same frequency are used as a frequency selective element for a feedback loop. The generated signal is compared against an external Local Oscillator (LO). FIG. 8: Sensitivity of the room temperature and based on the same parameters and phase expressions as cryogenic versions of the degenerate mode experiment the detuned loop oscillator experiment, simply setting comparing to the axion models and existing limits due the two resonance to the same frequency and sampling to CAST experiment and SN1987A. the phase noise spectrum up to 100 MHz. We again as- sume 30 days averaging time for this 100 MHz span - this will be limited to some extent by the achievable fre- A. TE-TM mode Cylindrical Cavity Resonator quency resolution of the digitiser. We note that these limits curve upwards as the loop oscillator phase noise As discussed, the sensitivity limits above are calcu- hits the white phase noise oor some distance into the lated based on the example of a microwave cavity with Fourier spectrum. As a result the straight, linear curves two orthogonally polarized resonances. Many other ge- for the exclusion limits in g. 6 (and indeed, the dashed ometries are possible, but for the purposes of demon- red line in g. 8) represent extending the atter part of strating the sensitivity of these techniques we have mod- the \broadband" exclusion limits, as it is in principle pos- elled the following. We take a 29.2 mm radius cylindrical sible to achieve this level of sensitivity at any point in the copper cavity with a TM mode frequency of 9 GHz, accessible frequency space, by simply setting the detun- and a TE mode frequency tunable from 6.5 to 9 GHz ing frequency accordingly and sampling only a narrow as the cavity height tunes from 18.5 to 83.6 mm. The region of Fourier space. Such an experiment would natu- loaded quality factors of both resonances are taken to rally scan far slower than 30 days per 100 MHz, but the be 10,000, which is readily achievable in copper at these curves in gs. 8 and 6 represent the limits of sensitivity frequencies. In this con guration, taking the stationary achievable with room temperature and cryogenic loop os- frequency TM mode to be mode \2" in the equations cillators, with feasible parameters, and within reasonable for the overlap integrals, we nd that  varies from -0.39 laboratory time scales. It is important to note that these to -0.50 as over the tuning range, whilst  varies from schemes, and indeed the open loop scheme presented in 0.46 to 0.57. Generally speaking, the analytical expres- g. 4, do carry technical diculties, and will require care- sions for these overlap integrals are messy and complex, ful design. For instance, these limits rely on the imple- and it is preferable to calculate them numerically for a mentation of an improved frequency stabilisation system given set of modes and cavity geometry. We found that for a CSO. However, this is not beyond the realm of what the magnitude of these expressions generally increased is achievable, and the presented limits should beviewed with the cavity aspect ratio, where h is the height of as the potential reach of this technique. the cavity, and a the radius, up to some maximum value achieved for aspect ratios greater than  5. B. Orthogonally Polarised Modes in a Fabry-P erot Cavity VI. SOME PRACTICAL REALISATIONS Another interesting orthogonally polarized mode All dual mode techniques considered in the present scheme to consider relies on Gaussian beam modes in work rely on a possibility to design a cavity that ex- Fabry-P erot cavities. Gaussian beams have the advan- hibits signi cant orthogonality of two modes represented tage that the electric and magnetic elds have the same by Eq. (8). These formulae give the cavity form factor pro le, simply rotated 90 degrees with respect to one an- that has to be as large as possible. other. If we take the electric eld of a given Gaussian KSVZ/DSFZ Models 9 beam to be polarized on the x-direction and propagating (f >10GHz) axion searches. Moreover, optical in the z-direction such that cavities might be used to probe otherwise unacces- sible regions of THz and infrared spectrum as well E = E(r; z) x; (21) as millimiter-wave and microwave frequencies; Liquid-Helium temperature operation (> 4K) then the magnetic eld is where only a limited number of components such B / E(r; z) y (22) as cavity and ampli ers have to be at low tem- perature. This factor removes the need of dilu- This means, that if we were to take two orthogonally po- tion refrigeration that is a key component in tra- 6{9,11 larized Gaussian beam modes in the same Fabry-P erot ditional haloscopes making the whole exper- cavity, with the same frequency, the variables  and iment available to a broader audience. Although would be 2 and 0 respectively. This has promising dilution refrigeration might give some incremental implications for high mass axion searches, via the down- improvement in the axion search, all ultra-stable conversion technique. microwave and optical clocks and oscillators do not However, we may also boost the sensitivity of the up- require temperatures below 4K. conversion technique, by employing two modes orthogo- access to higher and lower frequency ranges. The nally polarized in the same cavity, but of di erent fre- fact that actual axion mass is either the sum or dif- quency, ie a di erent number of wavelengths between the ference of working frequencies opens a possibility to two mirrors. This would give a non-zero value of the search for axions in less accessible frequency rages. parameter, which would open up the possibility of lower For instance, working around 20GHz, one is able mass axion searches with such structures. to probes axion masses in the vicinity of 40GHz, where experiments are signi cantly more dicult; CONCLUSIONS limited power levels (P < 100W cryogenically and 1 W at room temperature). Although, the sen- We demonstrated that precision frequency and phase sitivity does explicitly depend on power levels, on metrology could be used as a highly sensitive tool in dark the current calculation only limited power levels are 14,49,50 matter detection. Because measurements are made in the used unlike in some other proposals ; space of frequencies and phases rather than in the space axion phase sensitive. Comparing to DC magnet of amplitudes, the fundamental limit is set by control haloscopes, the dual frequency method is able to electronics rather than due to bare thermal uctuations. provide additional information about axions, par- This fact allows us to introduce a number of detection ticularly its phase relative to pump signals, al- schemes matching or exceeding sensitivity of state-of-the- though that might lead to more complicated de- art cryogenic axion detectors. This includes a broad band tection schemes; low mass axion detection based on the degenerate case 51{54 and Fourier spectra. In addition to bare axion amplitude KSVZ/DSFZ achievable. It is estimated that detection, the method allows to deduce relative phase of the cryogenic dual mode experiment is able to the axion signal. The proposed new class of detectors achieve the limit of the widely accepted axion dark may be understood as a phase sensitive detection scheme matter models. On the other hand, even a table- similar to phase sensitivity of parametric ampli ers. In top search may lead to competitive limits on dark summary, the main advantages of the proposed frequency matter; control method are: broadband search for low mass axions is 6{11 55,56 magnet-free. Unlike traditional haloscopes , the possible . It is demonstrated that a wide- proposed method does not require a strong DC band search that does not require tuning is magnetic elds; possible. SQUID-free. All sensitivities calculated in this work are based on usage of traditional low noise ACKNOWLEDGEMENTS semiconductor ampli ers. Though superconduct- ing technology might be used in the future, its pres- This work was supported by the Australian Research ence is not crucial contrary to traditional metods ; Council grant numbers DP160100253 and CE170100009. cavity volume independence. Although cavity vol- ume in uences many parameters of the experiment REFERENCES such as resonance frequencies and quality factors, the sensitivity is not directly proportional to this 11,45{48 parameter unlike in traditional haloscopes . J. Jaeckel and A. Ringwald, Annual Review of Nuclear and Par- This removes the major obstacle for higher mass ticle Science 60, 405 (2010). 10 2 33 R. Barkana, Nature 555, 71 (2018), arXiv:1803.06698 [astro- A. Lo, P. Haslinger, E. Mizrachi, L. Anderegg, H. Muller,  M. Ho- ph.CO]. hensee, M. Goryachev, and M. E. Tobar, Physical Review X 6, J. D. Bowman, A. E. E. Rogers, R. A. Monsalve, T. J. Mozdzen, 011018 (2016). and N. Mahesh, Nature 555, 67 (2018). M. Goryachev, Z. Kuang, E. N. Ivanov, P. Haslinger, H. Muller,  R. D. Peccei and H. R. Quinn, Phys. Rev. Lett. 38, 1440 (1977). and M. E. Tobar, IEEE Transactions on Ultrasonics, Ferro- J. Ipser and P. Sikivie, Phys. Rev. Lett. 50, 925 (1983). electrics, and Frequency Control, IEEE Transactions on Ultra- B. T. McAllister, G. Flower, E. N. Ivanov, M. Goryachev, sonics, Ferroelectrics, and Frequency Control 65, 991 (2018). J. Bourhill, and M. E. Tobar, Physics of the Dark Universe C. R. Locke, S. Munro, M. E. Tobar, E. N. Ivanov, and 18, 67 (2017). G. Santarelli, in IEEE International Frequency Control Sympo- A. Wagner, G. Rybka, M. Hotz, L. J. Rosenberg, S. J. Asztalos, sium and PDA Exhibition Jointly with the 17th European Fre- G. Carosi, C. Hagmann, D. Kinion, K. van Bibber, J. Hoskins, quency and Time Forum, 2003. Proceedings of the 2003 (2003) C. Martin, P. Sikivie, D. B. Tanner, R. Bradley, and J. Clarke, pp. 350{354. Phys. Rev. Lett. 105, 171801 (2010). E. Rubiola, The Cambridge RF and microwave engineering se- W. Chung, Proceedings, 15th Hellenic School and Workshops ries (Cambridge University Press, 2008) p. 220. on Elementary Particle Physics and Gravity (CORFU2015): M. Goryachev and S. Galliou, Ultrasonics 51, 966 (2011). Corfu, Greece, September 1-25, 2015, PoS CORFU2015, 047 M. Goryachev, Cryogenic BAW Resonators and Oscillators (2016). (Lambert Academic Publishing, 2012). 9 39 J. Choi, H. Themann, M. J. Lee, B. R. Ko, and Y. K. Se- E. N. Ivanov and M. E. Tobar, Review of Scienti c Instruments, mertzidis, Phys. Rev. D 96, 061102 (2017). Review of Scienti c Instruments 80, 044701 (2009). 10 40 A. Caldwell, G. Dvali, B. Majorovits, A. Millar, G. Ra elt, J. Re- M. E. Tobar, P. Wolf, A. Fowler, and J. G. Hartnett, Phys. Rev. dondo, O. Reimann, F. Simon, and F. Ste en (MADMAX Work- D71, 025004 (2005), [Erratum: Phys. Rev.D75,049902(2007)], ing Group), Phys. Rev. Lett. 118, 091801 (2017). arXiv:hep-ph/0408006 [hep-ph]. 11 41 B. M. Brubaker, L. Zhong, Y. V. Gurevich, S. B. Cahn, S. K. E. N. Ivanov, M. E. Tobar, and R. A. Woode, IEEE Transactions Lamoreaux, M. Simanovskaia, J. R. Root, S. M. Lewis, S. Al Ke- on Ultrasonics, Ferroelectrics, and Frequency Control 45, 1526 nany, K. M. Backes, I. Urdinaran, N. M. Rapidis, T. M. Shokair, (1998). K. A. van Bibber, D. A. Palken, M. Malnou, W. F. Kindel, M. A. D. Leeson, Proc. IEEE 54, 329 (1966). Anil, K. W. Lehnert, and G. Carosi, Phys. Rev. Lett. 118, K. Shigeru and K. Yoshio, Electronics and Communications in 061302 (2017). Japan (Part II: Electronics) 78, 31 (1995). 12 44 P. Sikivie, Physical Review Letters 51, 1415 (1983). M. D. Harvey and A. G. White, Optics Communications 221, A. C. Melissinos, Phys. Rev. Lett. 102, 202001 (2009). 163 (2003). 14 45 P. Sikivie, arXiv:1009.0762 [hep-ph] (2010). J. Jeong, S. Youn, S. Ahn, J. E. Kim, and Y. K. Semertzidis, C. M. Caves, Physical Review D 26, 1817 (1982). Physics Letters B 777, 412 (2018). 16 46 S. J. Asztalos, G. Carosi, C. Hagmann, D. Kinion, K. van Bib- J. Jeong, S. Youn, S. Ahn, C. Kang, and Y. K. Semertzidis, ber, M. Hotz, L. J. Rosenberg, G. Rybka, J. Hoskins, J. Hwang, Astroparticle Physics 97, 33 (2018). P. Sikivie, D. B. Tanner, R. Bradley, and J. Clarke, Physical B. T. McAllister, G. Flower, L. E. Tobar, and M. E. Tobar, Review Letters 104, 041301 (2010). Phys. Rev. Applied 9, 014028 (2018). 17 48 M. Aspelmeyer, T. J. Kippenberg, and F. Marquardt, Rev. Mod. M. Goryachev, B. T. McAllister, and M. E. Tobar, Physics Let- Phys. 86, 1391 (2014). ters A 382, 2199 (2018), special Issue in memory of Professor P. D. Nation and M. P. Blencowe, New Journal of Physics 12, V.B. Braginsky. 095013 (2010). P. W. Graham, I. G. Irastorza, S. K. Lamoreaux, A. Lindner, G. P. Agrawal and C. L. Mehta, Journal of Physics A: Mathe- and K. A. van Bibber, Ann. Rev. Nucl. Part. Sci. 65, 485 (2015), matical, Nuclear and General 7, 607 (1974). arXiv:1602.00039 [hep-ex]. 20 50 R. Gambini, Physical Review A 15, 1157 (1977). K. Ehret, M. Frede, S. Ghazaryan, M. Hildebrandt, E.-A. Kn- S. R. Parker, G. Rybka, and M. E. Tobar, Physical Review D abbe, D. Kracht, A. Lindner, J. List, T. Meier, N. Meyer, 87, 115008 (2013). D. Notz, J. Redondo, A. Ringwald, G. Wiedemann, and B. Yurke and E. Buks, Journal of Lightwave Technology, Journal B. Willke, Physics Letters B 689, 149 (2010). of Lightwave Technology 24, 5054 (2006). J. E. Kim, Phys. Rev. Lett. 43, 103 (1979). 23 52 C. Eichler and A. Wallra , EPJ Quantum Technology 1, 2 (2014). M. Shifman, A. Vainshtein, and V. Zakharov, Nuclear Physics M. A. Castellanos-Beltran and K. W. Lehnert, Applied Physics B 166, 493 (1980). Letters, Applied Physics Letters 91, 083509 (2007). M. Dine, W. Fischler, and M. Srednicki, Physics Letters B 104, B. A. Kochetov and A. Fedorov, in 2016 8th International Con- 199 (1981). ference on Ultrawideband and Ultrashort Impulse Signals (UW- A. R. Zhitnitsky, Sov. J. Nucl. Phys. 31, 260 (1980), [Yad. BUSIS) (2016) pp. 112{116. Fiz.31,497(1980)]. 26 55 G. Cerullo and S. De Silvestri, Review of Scienti c Instruments, Y. Kahn, B. R. Safdi, and J. Thaler, Phys. Rev. Lett. 117, Review of Scienti c Instruments 74, 1 (2003). 141801 (2016). 27 56 A. Olkhovets, D. W. Carr, J. M. Parpia, and H. G. Craig- B. T. McAllister, M. Goryachev, J. Bourhill, E. N. Ivanov, and head, in Technical Digest. MEMS 2001. 14th IEEE Interna- M. E. Tobar, (2018), arXiv:1803.07755 [physics.ins-det]. tional Conference on Micro Electro Mechanical Systems (Cat. No.01CH37090) (2001) pp. 298{300. T. Br acher, P. Pirro, and B. Hillebrands, Parallel pumping for Appendix A: Derivation of Two Mode Power Sensitivity magnon spintronics: Ampli cation and manipulation of magnon spin currents on the micron-scale, Physics Reports 699, 1 (2017). V. Frank and Y. Lan, Nanophotonics 1, 267 (2012). As it is mentioned in Section III, the problem of axion H. Du, G. Zhou, Y. Zhao, G. Chen, and F. S. Chau, Applied detection with two pumped modes has much in common Physics Letters, Applied Physics Letters 110, 061110 (2017). 22,23 with the analysis of parametric ampli ers , where one M. R. Foreman, J. D. Swaim, and F. Vollmer, Advances in optics and photonics 7, 168 (2015). is looking at manipulation of very weak signals on top of M. Nagel, S. R. Parker, E. V. Kovalchuk, P. L. Stanwix, J. G. strong pumps. Following the standard analysis in the Hartnett, E. N. Ivanov, A. Peters, and M. E. Tobar, Nature eld of parametric ampli ers, rstly, the steady state Communications 6, 8174 EP (2015). problem is solved for the strong pump signals. In the 11 case of Josephson parametric ampli ers, these equations axion coupling terms, the solution Eq. (A2) is valid for of motions are nonlinear. Secondly, the system dynamics this case as well. And, the equation of motion for the is analysed for small uctuations using linearised equa- small uctuations are then given as: tions. For this reason, each eld component is split into in strong pumped parts (C and B ) and small uctuations i ti n U m ec = ec g  eaC e n n n e m in dt (A5) (ec and b ): in i 2 b ; i i! t n pn c = C e + ec (t) e ; n n n where  = ! + ! ! is axion detuning in the U a pm pn in in i' in i! t n e pn (A1) b = B e + b (t) e ; upconversion case. The solution in the frequency domain n n n i! t can be given as follows: a = ea(t)e ; g  C in e m U i where ! and B represent pumping frequency and pn n ec [ ] = ea[ ]e (i magnitude for the nth mode and the axion signal has p (A6) only the small component ea. One may extend the model 2 in i b [ ]: to include an additional output \probe" and explicitly (i distinguish coupling and intrinsic losses. In the present In both upconversion and downconversion cases, Equa- analysis, b may be associated with technical uctuations tions (A6) and (A4), the mode uctuations consist of a and noise, while ea is the component that has to be de- noise component associated with the input signal b[ tected. The latter quantity is assumed to be slowly time- and signal component ea[ ] which is ampli ed by the varying to account for the "Quality factor" or linewidth strong signal deposited in the other mode C . of the axion signal that appears in many cosmological m models. Also, the phases of the pump signals ' are in- troduced with respect to the axion phase that is set to Appendix B: Axion Generated Phase Spectrum in a Dual zero. Mode Cavity On the rst step, due to extreme weakness of axions, the axion mediated coupling terms could be neglected In this Appendix we consider a case when an axion resulting into two linear independent equations. The re- signal is searched in the Fourier spectrum of phase noise sulting amplitudes are of the double mode cavity. In this case, we relax the re- quirement of the axion frequency to be exactly the sum or in i('  ) n m C = B e ; (A2) n di erence of resonant frequencies of two modes. Instead, n n this property is only approximately held true where the where = ! ! and phases are chosen to set the detuning constitute the Fourier frequency. In this case, n n pn coecents C real. ! = !  ! + where ! . Under this condition, n a 1 2 1 On the second step, the equations of motion are writ- the axion amplitude appear as a slowly varying param- eter in the EOMs. The objective of this derivation is to ten for small varying uctuations ec , ea and b where only n n nd a transfer function from this small and slow time- terms of the rst order in ec and ea are retained. Denoting varying quantity to amplitudes and phase uctuations = i + , the equations of motion in the Markovian n n n of the output signals of the cavity. The transfer func- and "slowly varying envelope" approximations: tion is derived by transferring the equations of motion d for the complex amplitudes to the equations for real sig- i t+i D m ec = ec + g  eaC e n n n e + m nal phases and magnitude and linearizing them around a dt (A3) in steady state point. i 2 b ; The starting point of this derivation is the equations of motion for the complex amplitudes C and C of the 1 2 where  = ! ! ! is axion detuning in the D a p1 p2 two modes in the upconversion case. Additionally, we downconversion case that can be set to zero. This EOM introduce two pump signals B and B , so the equations 1 2 can be easily solved by transforming the problem into the of motion can be written as follows: frequency domain: g  C e + m C = ( i ( ))C g  A( )C + B ( ); D i 1 1 1 1 e 2 1 ec [ ] = ea[ ]e (i p (A4) C = ( i ( ))C + g  A ( )C + B ( ); 2 2 2 2 2 e 1 2 in i b [ ]: (i ) (B1) For the upconversion case, the analysis follows the where  is the "slow time". Time variation of the detun- same steps. Moreover, since on the zeroth order step ing coecients  accounts for uctuations of the reso- considering the strong pump case does not include the nance frequencies of the modes as well as DC detuning 12 of the pump signals. Instead of complex amplitudes, we The linearised EOMs for the small uctuations are need to rewrite the equations of motion in terms of real phases and amplitudes: C ( ) = x ( ) exp(j' ( )) and i i i xe = xe x Q ( )+ 1 1 1 2 c B ( ) = y ( ) exp(j ( )). Here y and  ( ) represents d i i i i i amplitude and phase uctuations of the pump signals. e ye cos( ' ) y sin( ' )( 'e ); 1 1 1 1 1 1 1 1 The EOMs in the real amplitude-phase representations are written as follows: x 'e = x  + xe  x Q 1 1 1 1 1 1 2 s +ye sin( ' ) + y cos( ' )( 'e ); 1 1 1 1 1 1 1 1 (B6) x = x x Q ( ) + y cos( ' ); 1 1 1 2 c 1 1 1 xe = xe + x Q ( )+ 2 2 2 1 c d x y 2 1 ' =  ( ) Q ( ) + sin( ' ); ye cos( ' ) y sin( ' )( 'e ); 1 1 s 1 1 2 2 2 2 2 2 2 2 d x x 1 1 (B2) x 'e = x  + xe  x Q ( ) 2 2 2 2 2 2 1 s x = x + x Q ( ) + y cos( ' ); 2 2 2 1 c 2 2 2 d +ye sin( ' ) + y cos( ' )( 'e ): d x y 2 2 2 2 2 2 2 2 1 2 ' =  ( ) Q ( ) + sin( ' ); 2 2 s 2 2 d x x 2 2 that can be simply written in the matrix form as where Q and Q are two quadrature of the axion signal c s (B7) X = HX + A + N; de ned as follows: where X = [xe ; 'e ; xe ; 'e ] is a vector of state variables, 1 1 2 2 A = [x Q ( );x Q ( )=x ; x Q ( );x Q ( )=x ] Q ( ) = g  jAj cos( ( ) + ' ' ); 2 c 2 s 1 1 c 1 s 1 c e a 2 1 (B3) is a vector of axion signal, N is a vector of technical uc- Q ( ) = g  jAj sin( ( ) + ' ' ); s e a 2 1 tuations coming from the pump signal and internal mode instabilities. The system matrix is and  is the axion phase. 0 1 x 0 0 1 1 1 These are nonlinear equations with varying coecients B C 0 0 B 1 C that cannot be solved exactly. Instead, we linearise them x H = B C : (B8) @ 0 0  x A 2 2 2 for small signal uctuations around large steady state amplitudes. Here we assume that both axion signal and 0 0 input phase and amplitude uctuations are very small. Thus, we can split each variable and time varying pa- It is worth noting that in this linearised approxima- rameter into two components strong steady and small tion, all state variable uctuations as well as the axion uctuating: x = x + xe , y = y + ye , ' = ' + 'e and signal are fully uncoupled due to their extreme small- i i i i i i i i i e ness and ltered by the de ned system transfer matrix. =  + . As for the axion signal, it has only the small i i i Moreover, the two modes are completely uncoupled giv- uctuating part as described in the introduction to this ing two independent signals. The corresponding solutions Appendix. The steady state equations for the constant for technical uctuations have been previously analysed part are: 36{38 and can be found elsewhere . For this reason, we are interested only in axion-phase relationship that is given as follows: x = y cos( ' ); 1 1 1 1 1 0 =  x + y sin( ' ); 1 1 1 1 (B4) 'e [s] =  Q [s] i j;i c x = y cos( ' ); 2 2 2 2 2 2 s + 2 s +  + i i (B9) 0 =  x + y sin( ' ); s + 2 2 2 2 i Q [s] j;i s s + 2 s +  + i i that are two independent sets of two equations. Solv- where s is the Laplace variable, = is a ratio of j;i ing these equations give the steady state point (y ; ' ) in i i i stored amplitudes, + sign attributes to the second mode, terms of input signal parameters (y ;  ) and detunings and the is for the rst one. Q [s] and Q [s] are Laplace c s transforms of the two quadratures of the axion signals. It is important to empathise that these quadratures are de ned with respect to the phase di erence of the cavity x = p ; ' =  arctan : (B5) modes ' ' . Thus, generally the result is phase de- i i i 1 2 2 2 +  i i i pendent and could depend on the instance of time when 13 it starts. In other words, the axion signal provides an follows absolute time scale for this kind of experiment. In the case when both cavities are pumped on reso- 'e [s] = R [s]+ i j;i c nance ( = 0), transfer function (B9) is reduced to the i 2 2 2 s + 2 s +  + i i rst order low pass lter: (B14) s + R [s]; j;i s s + 2 s +  + 'e [s] i j;i i U i i H [s] = = : (B10) Q [s] s + s i which is di erent from the upconversion solution (B9) It is apparent from this result that the output phase com- only in the signs of each term as well as de nition of the ponent due to the axion signal is scaled by the ratio of axion quadratures. In these equations R [s] and R [s] c s magnitudes in both modes. If one is going to measure are Laplace transform of the axion signal de ned as in ' , it is advantageous to increase the x magnitude and (B13). Here the axion signal quadratures are de ned 1 2 keep x as small as possible. On the other hand when with respect to the sum of the phases of signals in both x ! 0, the carrier signal become undetectable. modes ' + ' . Thus like in the upconversion case, ex- 1 2 perimental results could be phase sensitive. In the case To derive the sensitivity of the setup over a range of of two cavities pumped on resonance, the result is also Fourier frequencies, we compare result (B10) to the trans- reduced to the rst order transfer low pass lter function: fer function for technical phase uctuations. Without lack of generality, we assume that technical uctuations are dominated by external uctuations  [s] whose trans- i 'e [s] i j;i i 37,38 H [s] = = : (B15) fer function in the phase space is . It can be s+ i R [s] s + s i shown that the internal cavity uctuations lead to a sim- ilar relation. A ratio of the axion signal and phase noise This result is identical to that of the upconversion case uctuations at the output of the cavity gives a constant (B10) up to the de nition of the axion quadratures and signal-to-noise ratio: the geometry factor. The corresponding signal-to-noise ratio is jAj SNR = g  : (B11) e j;i jAj i i D SNR = g  : (B16) e + j;i i i If the cavity is pumped on resonances, the same result is obtained for the case of limits due to external signal uctuations. Appendix C: Axion Generated Phase Spectrum in a Dual For the downconversion case, the equations of motion Loop Oscillator in terms of real phases and amplitudes are A loop oscillator is a frequency selective positive feed- x = x + x R ( ) + y cos( ' ); 1 1 1 2 c 1 1 1 back system in which two conditions of existence of sus- tained oscillations are ful lled: small signal gain across d x y 2 1 ' =  + R ( ) + sin( ' ); 1 1 s 1 1 the loop is greater than 1 and open loop phase shift is d x x 1 1 (B12) integer multiple of 2. These conditions make the small x = x + x R ( ) + y cos( ' ); 2 2 2 1 c 2 2 2 signal solution for the oscillator divergent, although due to amplitude limiting nonlinearity the self sustained sig- d x y 1 2 nals gets saturated as some value of circulating power. ' =  + R ( ) + sin( ' ): 2 2 s 2 2 d x x 2 2 To analyse phase-amplitude uctuations in such a sys- tem, in addition to cavity parameters, one needs to make where the axion signal quadratures are assumptions on the type of nonlinearity, ampli er gain and phase shift as well as ampli er-cavity couplings. In R ( ) = g  jAj cos( ( ) ' ' ); c e + a 2 1 this case, it is possible to apply the same type of analy- (B13) R ( ) = g  jAj sin( ( ) ' ' ): sis as described in Appendix B. On the other hand, one s e + a 2 1 may assume a certain level of circulating power, e.g in the form of magnitudes in the cavity modes x and x , The solution of the corresponding steady state equa- 1 2 and analyse the system in the small signal regime. In this tion is the same as in the upconversion case described regime, an ideal ampli er has a unity transfer function in above. Dynamics of the small magnitude-phase uctua- the phase space. The resulting system for the ith mode tions can be described by the same matrix equation (B7) is demonstrated in Fig. 9. The phase uctuations at the and system matrix (B8) with a di erent axion input vec- output of the oscillator are found as a sum of contribu- tor: A = [x R ( ); x R ( )=x ; x R ( ); x R ( )=x ] . 2 c 2 s 1 1 c 1 s 1 The corresponding axion-phase relationship appears as tions from the technical noise  and axion signal R (or i s 14 Q ): h i i j;i 'e = 1 +  + R ; i s s s + h i  (C1) i j;i 'e = 1 +  + Q ; i s s s + where the rst term constitutes the Leeson e ect for the FIG. 9: Schematic representation of the axion to phase technical phase uctuations in the loop and the second conversion in a feedback oscillator in the phase space. term represents the axion induced phase uctuations. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png High Energy Physics - Phenomenology arXiv (Cornell University)

Axion Detection with Precision Frequency Metrology

Download PDF
14 pages

Loading next page...
 
/lp/arxiv-cornell-university/axion-detection-with-precision-frequency-metrology-tBgZe0gRYK

References (49)

ISSN
2212-6864
eISSN
ARCH-3335
DOI
10.1016/j.dark.2019.100345
Publisher site
See Article on Publisher Site

Abstract

1 1 1, a) Maxim Goryachev, Ben T. McAllister, and Michael E. Tobar ARC Centre of Excellence for Engineered Quantum Systems, School of Physics, University of Western Australia, 35 Stirling Highway, Crawley WA 6009, Australia (Dated: 12 September 2019) We investigate a new class of galactic halo axion detection techniques based on precision frequency and phase metrology. Employing equations of axion electrodynamics, it is demonstrated how a dual mode cavity exhibits linear mode-mode coupling mediated by the axion upconversion and axion downconversion processes. The approach demonstrates phase sensitivity with an ability to detect axion phase with respect to externally pumped signals. Axion signal to phase spectral density conversion is calculated for open and closed loop detection schemes. The fundamental limits of the proposed approach come from the precision of frequency and environment control electronics, rather than fundamental thermal uctuations allowing for table-top experiments approaching state-of-the-art cryogenic axion searches in sensitivity. Practical realisations are considered, including a TE-TM mode pair in a cylindrical cavity resonator and two orthogonally polarised modes in a Fabry-P erot cavity. INTRODUCTION axions creating additional photons with a di erent fre- quency (such that energy is conserved) at a rate related to the number of pre-existing photons and the number of Axions are theoretical weakly-interacting sub-eV par- 1 axions. Prior work has considered a single pumped RF ticles that can be formulated as a primary component of cavity mode interacting with axions, followed by detect- dark matter. With mounting evidence pointing towards 13,14 2,3 ing a small signal in an orthogonally polarised mode . lower mass particles , particles such as axions are be- In this work we consider more general cases of two mode- coming increasingly promising dark matter candidates. axion interactions. Axions arise as a result of an elegant solution to the 4 Despite the apparent similarity between the DC and strong CP problem in QCD , and are expected to have 5 AC detection schemes, they belong to di erent classes properties consistent with dark matter . Confounding of detectors. Since virtual photons or static elds carry experimental e orts to detect axions is the fact that the no phase, the traditional Sikivie haloscope detectors (us- axion mass is largely unknown, with only weak bounds ing DC magnetic or electric elds) belong to the class from cosmology and theory. Despite this, a number of 6{11 of phase insensitive systems. On the other hand, the axion detection experiments are already underway . AC scheme considered in this work relies on pump- In perhaps the most common axion detection tech- ing signal(s) carrying relative phase as well as sepa- nique, it is generally agreed that dark matter axions rate phases relative to the axion signal. Thus, the de- can be detected via the (inverse) Primako e ect, a tected signal as well as the overall result would have two photon-axion interaction. Nowadays, in a typical a footprint of these phases. This fact draws analogies (Sikivie) detector , these hypothetical particles interact with existing ampli ers that can be grouped into DC with a DC magnetic eld, or virtual photons, to produce (phase insensitive) ampli ers, where energy is drawn real photons whose frequency corresponds to the mass from static power supply, and parametric (phase sen- of axions. This scheme employs one or several tuneable sitive) ampli ers, where energy comes from oscillating microwave cavities, serving as resonant antennas, with elds. The second type gives more freedom allowing im- the output coupled to the lowest noise ampli ers so gen- proved ampli cation/detection schemes based on quadra- erated photons may be detected with the greatest sen- ture squeezing . Thus, in this work we expand investi- sitivity possible. In principle, a similar detector using gations into a novel class of axion detectors employing static electric eld is also possible although due to consid- the phase sensitive approach. erable mismatch between electric and magnetic compo- nents and relative diculty in creating extremely strong electric elds in large volumes, such detectors are never I. AXION ELECTRODYNAMICS DESCRIPTION realised in practice. The third Sikivie-like axion detection technique is represented by the type of detectors utilising The Hamiltonian density of the photon-axion system RF or microwave elds instead of static ones: indeed, the consist of the conventional electromagnetic, axion and Primako process works equally well with real photons 13,14 interaction parts: instead of virtual ones . In these schemes the exist- ing photons of a given frequency interact with cosmic H = H +H +H : (1) EM a int The free electromagnetic Hamiltonian density is usually represented using the vectors of electric and magnetic a) Electronic mail: michael.tobar@uwa.edu.au elds, E and B respectively, or a vector potential and its arXiv:1806.07141v3 [physics.ins-det] 11 Sep 2019 2 conjugate momentum, A and : as (moving from Hamiltonian density to Hamiltonian): Z Z h i h i " 1 1 3 3 2 2 2 2 2 2 H = " cg  d rE  B + d rE  B ; (6) int 0 a 1 2 2 1 H = E + c B =  + " c (r A) ; EM 0 2 2 " V V (2) where it is assumed that material properties dictate that for each mode E  B = 0. Quantising the two modes, i i where c is the speed of light, and " is the dielectric one arrives at the following interaction Hamiltonian in permittivity of free space. terms of creation-annihilation operators: The axion part of the system for a laboratory size ex- periment may be represented by the uniform eld  and ~g y y H = i ! !  (c + c )(c c )+ int 1 2 1 2 1 its canonical conjugate : 2 1 (7) y y 2  (c + c )(c c ) 2 1 2 1 2 (3) H = + V (); 2m where two dimensionless coecients  and  represent 1 2 overlap between the two modes: where m is axion mass. Choosing the normal harmonic m ! a 2 potential V () =  , this Hamiltonian reduces to 1 = p d r(e  b ); 1 1 2 a simple harmonic oscillator. On the other hand, for an V V 1 2 V (8) experiment of reasonable size and duration, this represen- = p d r(e  b ): tation is excessive; axion dynamics cannot be observed as 2 2 1 V V 1 2 V the associated time constant should be extremely large compared to the experimental duration. The apparent These coecients can span from one, when two modes nite `quality factor' of axion-induced photon signal is are of the same shape and fully orthogonal, to zero when due to the velocity distribution of axion dark matter in they exhibit no overlap. Finally, it is possible to separate the galactic halo, rather than due to response time of swapping and parametric parts: an `axion mode'. In this case, as it is usually done, the h i y y y y axion part is simply represented as an external signal H = i~g   (c c c c ) +  (c c c c ) ; (9) int e 1 2 + 1 2 2 1 1 2 with amplitude  and angular frequency ! that can be varied in time due to axion velocity. where  =   , and g = ! ! is the e ective 1 2 e 1 2 It is widely accepted that the axion-photon interaction trilinear coupling. part in a laboratory size experiment can be represented A particular type of dynamics described by the interac- in the following form: tion Hamiltonian (7) depends on the resonant frequencies of the photon modes ! and axion signal ! . In partic- i a " cg  ular, the following regimes can be identi ed (assuming 0 a (4) H = " cg  E B =  (r A); int 0 a ! ! > 0): " 2 1 where  is a scalar axion-like eld, and g is the com- ! = ! + ! ; axion downconversion a a 2 1 (10) monly presented axion-photon coupling constant. ! = ! ! ; axion upconversion a 2 1 where neither of the angular frequencies is zero. A case II. TWO MODES AXION ELECTRODYNAMICS where one of cavity frequencies is zero would correspond to the Sikivie detector with virtual photons represent- ing one of the modes. This case has been widely studied We consider an electromagnetic cavity with two modes theoretically with a few experimental realisations in dif- of two angular frequencies ! and ! . Given a resonant 1 2 6,16 ferent frequency ranges . structure, each mode n is characterised by a certain dis- To demonstrate the di erence between the upconver- tribution of electric (E (r)) and magnetic eld (B (r)) n n sion and down conversion cases, we transform the system in a certain nite volume: into the rotating frame associated with the frequencies of both modes and apply the rotating wave approximation E (r) =  u (r) = iE (c c )e (r); n n n V;n n n (RWA). For this purpose, the axion signal  may be de- (5) composed in terms of complex amplitudes a exp(i! t) B (r) = A r u (r) = E (c + c )b (r); n i n V;n n n and a exp(i! t). Conditions (10) give the two follow- n a ing Hamiltonians after corresponding RWA when fast ro- tating terms are removed (all couplings and signals are where c (c ) are creation (annihilation) operators for small as dictated by the weak signal detection problem): ~! the mode n, E = , and e (r) and b (r) are V;n n n 2" V 0 n y y unit vectors representing the mode polarization. H = i~g  (ac c a c c ); D e + 1 2 1 2 (11) Each of these two photonic modes is coupled to an y y H = i~g  (a c c ac c ); U e 1 2 2 1 axion signal via the term in Eq. (4) that can be rewritten 3 (A) to-noise ratio of the corresponding axion detector. The second class of strategies is related to measuring photon mode frequency shifts introduced by the axion mediated coupling terms (11) and has previously been discussed in the context of other hidden sector particle searches . In this approach, discussed in detail in Section IV, instead of detecting power coming from the modes, one is look- (B) ing for mode frequency deviations associated with the new physics. III. POWER DETECTION APPROACH The excess power detection method is the most widespread approach to weak signal detection including the axion and paraphoton searches. According to this FIG. 1: Graphical representation of the two modes paradigm, one is looking for extra power coming from an interacting through axion coupling in (A) upconversion experimental setup over the expected noise oor. Typ- and (B) downconversion cases. ically the noise oor is associated with thermal uctua- tions of a measured device, i.e. cavity, and noise temper- ature of the amplifying electronics. Thus, such a scheme in the corresponding interaction pictures. Graphical rep- is usually realised in a cryogenic environment minimising resentations of these processes is given in Fig. 1. In all thermal uctuations. both cases, the problem is reduced to a system of two The same approach can be realised using the analysed piecewise coupled modes, as a may be understood as a double mode system. Although, the potential for exter- complex coecient. The downconversion case Hamilto- nal signal pumping of oscillating elds brings new possi- nian represents the parametric interaction found, for ex- bilities to the problem. Indeed, the axion coupling may ample, in the blue side band regime of optomechanical be regarded as a mixing term in the presence of strong ex- systems .The upconversion case is a swapping (or beam ternal pumping resulting in certain analogies that can be splitter) interaction that corresponds to the red side- 22,23 drawn with parametric ampli ers , which have found band regime in optomechanics . For the upconversion numerous applications on the forefront of physics. Para- and downconversion cases, the corresponding Heisenberg metric ampli ers' phase sensitivity and related phenom- equations of motion are respectively: ena allow them to beat the quantum or thermal limit or any other physical constraints to achieve extraordinary in c = i! c c g  ac i 2 b ; i i i i i e j i levels of sensitivity. Such devices could be built on di er- dt (12) ent physical principles, for example, nonlinearities from in c = i! c c + g  ac i 2 b ; i i i i i e + i j i superconducting junctions, nonlinear crystals, mechani- dt 24{28 cal resonators or magnon systems . To analyse such in where j 6= i, and b are a loss rate and input signal systems, it is customary to split the problem into two for the nth mode respectively. steps by splitting variables into nonlinear large pumped Given the systems described by the Hamiltonians (11), amplitudes and small uctuations giving linearised equa- one can imagine two classes of detection strategies. The tions of motions . In Appendix A, we follow this line rst class relies on excess power detection introduced by of thought to analyse the sensitivity of the externally the axion-photon coupling. This approach, further dis- pumped double mode axion detector in the downconver- cussed in Section III, is an extension to standard axion sion and upconversion cases. The result (Eq. (A4) for 6,12,16 detectors using strong DC magnetic elds where the upconversion and Eq. (A6) for the downconversion) the role of the DC eld is played by the modes' oscillat- suggests that the power in one mode due to an axion ing elds. Besides the numerical values for the physically signal is proportional to the power stored in the other achievable eld strengths, the major di erence between mode. This result is in accordance with the standard the current proposal and the standard virtual eld tech- DC magnetic eld detector where the axion power at nique is the appearance of phases, making the experiment the cavity output is proportional to B , squared mag- DC phase sensitive. It is worth noting that the Hamiltonian netic eld strength, i.e. magnetic eld energy. Regarding for the downconversion case may be interpreted as a tri- the amplitudes, rather than powers, the main di erence linear Hamiltonian coupling pumped cavities to an ex- is additional phase  (upconversion) or + (down- m m ternal signal. This Hamiltonian has been considered pre- conversion) that is not measurable with a standard power 18 14 viously in the literature in di erent physical contexts detector . It is worth noting that the resulting spectral 19,20 including its application to parametric ampli cation . density from Eq. (A4) is identical to the previously pro- This property might be exploited to increase the signal- posed method in the case when one mode is pumped on 4 resonance and the signal is observed through the other A. Axion Induced DC Frequency Shifts mode. In practice, direct power detection using the dou- ble mode approach would be inferior to the scheme in- The system equations of motion in the interaction pic- volving virtual photons (DC magnetic eld) due to prac- ture for the downconversion (10) can be written in the tical impossibility to create oscillating elds of matching following form: strength. Instead of measuring the power of uctuations from C = ( i )c + g  AC ; 1 1 1 1 e + dt a single mode, a better strategy is to measure cross- (14) correlation between the two modes of the system. As- C = ( i )c + g  AC ; 2 2 2 2 e + suming statistical independence of the eld uctuations dt in coming into the modes b , the cross correlation spectrum where C and C are slowly varying amplitudes, is the 1 2 n of both channels is reduced to: loss rate for the mode n,  is the detuning frequency of a cavity with angular frequency ! such that  + i 1 2 2 = ! +! ! , and A is the axion eld. Introducing g  C C 2 1 2 a 1 2 D/U e i(  ) 2 1 (13) S [ ] = e S [ ]; =  i , the eigenvalues of this system are n n n j(i )(i )j 1 2 + 1 D 2  2 2 2 2 e =  ( ) + 4g  jAj where upconversion and downconversion cases are dif- 2 e 2 2 (15) ferent only through the phase factors giving the identical 2 2 2 2 =  i  ( i ) + g  jAj + + cross-correlation spectrum (13) with a di erent multiplier e + !  due to di erent mode overlapping integrals (8). as well as their complex conjugates, where = ( It means that by observing the power, the two cases are )=2 and  = (   )=2. 2  1 2 indistinguishable. One consequence of this is that by a For the upconversion case, the equations of motion single measurement, one searches for axions in both the could be written as follows: upconversion and downconversion regimes. The sensi- tivities of these cases can vary because of the di erent C = ( i )c g  AC ; overlap integrals. In case of a candidate detection, one 1 1 1 1 e 2 dt (16) would need either to modify geometry and thus  or ver- ify the two cases separately with a di erent combination C = ( i )c + g  A C ; 2 2 2 2 e 1 dt of frequencies ! and ! . 1 2 with the eigenvalues: + 1 1 2 U 2 2 2 2 e =  ( ) + 4g  jAj 1 2 2 2 (17) IV. FREQUENCY MEASUREMENT APPROACH q 2 2 2 2 =  i  ( i ) + g  jAj + + Frequency measurement techniques have found consid- and their complex conjugate. erable attention in precision sensing technology. They In the simplest and most sensitive form, one tunes both have been successfully applied in such elds such as par- detuning frequencies to zero ( = 0) and match the ticle detection, bio-sensing, magnetic eld and mass sen- losses ( = ). The resulting relative shift of eigenfre- 1 2 29{31 21,32{34 sors, etc , as well fundamental physics tests . quencies due to coupling to axion signal of amplitude jAj Unlike the power detection scheme, the fundamental is limit in frequency measurements comes from the best achievable frequency stability of the measurement parts D D D 2 2 2 2 @e = e e = g ! !  jAj ; 1 2 + a + + and thus it is not directly related to the Nyquist noise (18) U U U and the ambient temperature, although cryogenic cool- @e = e e = g ! !  jAj: a 1 2 ing may signi cantly improve the detection limit due to Whereas for the upconversion case, splitting is directly higher quality factors. It is generally accepted that one proportional to the axion amplitude, in the downconver- can detect frequency variations on the order of 10 rela- sion case only imaginary part is sensitive to the pres- tive to the system linewidth. With such systems as super- ence of axions. To compare the two cases of detection conducting cavities and sapphire, it is possible to achieve in more detail, we calculate deviation of real and imagi- linewidths of the order of 1 Hz at microwave frequencies nary components of the eigenvalues from its values in the giving unprecedented sensitivity of the frequency mea- axion-free case (jAj = 0) as a function of the normalised surement approach. g  jAj eff axion coupling strength  = shown in Fig. 2. In this section we consider a possible application of fre- quency metrology to axion searches enabled by the pro- The upconversion case (dashed line) demonstrates the ex- posed dual mode approach. As before, we split the dis- pected linear dependence of the real component of eigen- cussion into the downconversion and upconversion parts. frequency deviation on the coupling for small and the energy conservation, but rather indicate the regime in which the electromagnetic modes are slightly detuned from the axion frequency. Such a regime is analogous to the situation in which the cavity mode frequency of the standard DC haloscope is slightly shifted from the axion frequency. The detuning frequency f plays the role of the Fourier spectrum in the generated noise and could span over a few decades. In this situation, the axion amplitude appears as a slowly varying parameter A in the EOMs (14) and (16). The detailed analysis of this approach is given in Appendix B where we reformulate the problem 36{38 in terms of slowly varying real magnitude and phase instead of the complex amplitude representation (14) and (16). The analysis in Appendix B, reveals a transfer function FIG. 2: DC sensitivity of the dual mode detector from one of the axion quadratures to the phase uctu- eigenvalues to the normalised axion coupling strength . ations of the output signal (Eq.(B10) for upconversion and Eq. (C1) for downconversion). Using this transfer function, the phase noise spectrum of the output signal as stated by Eq. (18). This regime demonstrates eigenfre- for the ith mode may be represented as follows: quencies shifts equal to axion signal strength normalised 2 2 to frequency units. This case requires precise matching g  x D/U  j e i S (f ) = S (f ) + S (f ); (19) ';i of the modes in terms of losses and detuning frequen- 2 2 2 2 f + x f + i i cies. With nonzero di erence between mode losses where the rst component is the spectrum of phase in- the sensitivity in the lower coupling limit decreases and duced by the axion signal, and S (f ) is the power spec- is limited only to the imaginary part of the eigenvalue. A trum of the axion eld, in units of jAj , where A repre- In the downconversion case, although the sensitivity sents the axion eld, and the second term is due to tech- is reduced, the axion coupling appears as a modi cation nical phase uctuations S (f ) of the pump signal. The to the imaginary part of eigenfrequencies: the linewidth result explicitly depends on the ratio of steady state am- gets narrower when coupling increases. The change in the plitudes x in both modes. Thus, the overall sensitivity imaginary part is shown with the green curve in Fig. 2. i may be boosted using this ratio. It is also important to At  = 1, the axion term balances the sum of the losses note that despite the fat that the axion signal is ltered leading to the steady state oscillation regime. In this by the resonator (appears as the rst order transfer func- regime, photons created by axion downconversion, bal- tion in the phase space), the overall signal-to-noise ratio ance the cavity losses and the system starts oscillat- is constant as the technical uctuations are also ltered ing. Since the axion signal is extremely weak, this regime by the same ltering function. is not achievable even for the most narrow-linewidth cav- The phase noise spectrum (19) could be measured us- ities. In order to boost the sensitivity of the frequency ing the phase measurement setup shown in Fig. 3. Here, measurement technique in the downconversion case, one the output of each cavity is mixed with a local oscillator may look for compensation of the cavity losses.Indeed, of the same frequency as the pump. By varying the phase by introducing external or internal gain into one or two between the pump signal incident on the cavity and the modes, the e ective cumulative losses decrease result- mixer , one can access to both quadratures of the eld ing in lower oscillation threshold. uctuations in the cavity. Such an approach is capable in principle of detecting phase uctuations in rms amplitude of 2 10 rad= Hz at Fourier frequencies above a few B. Open Loop Axion Induced Spectral Density of Phase kiloHertz . Such kind of measurement is not possible Measurements with the traditional DC- eld axion detector. The sensitivity plots of the pumped phase noise mea- In practice, measurement of DC frequency shifts is surement experiment is shown in Fig. 4. These sensitivi- technically challenging. For fundamental tests the sit- ties are based on comparing the size of the axion-induced uation is worsened by the requirement to verify or veto phase-shifts, which are taken to be of magnitude the candidate signals. The problems may be solved by modulating one of the system parameters as, for exam- 1 g ! ! x a 1 2 D/U j S =  jAj ; ple, proposed in the case of a paraphoton search . An- ';i;a 2 f + 2 x other approach is to search for axion signals in the Fourier spectrum. For such an approach, we consider a regime S (f ) with the minimum detectable phase-shift where of small axion detuning: ! = !  ! + 2f where a 1 2 2f  ! . Note that these relations do not violate in this case the background noise, S (f ) is the phase 1  6 Phase Shift Filter Cavity CAST Phase Shift FIG. 3: An open loop realisation of the dual mode axion detection scheme. Two external local oscillators (LO) are used to excite two axion coupled modes. FIG. 4: Sensitivity of the cavity phase noise noise spectral density of the output we are measuring (as measurement experiment comparing to the axion opposed to the phase noise spectrum, S ), and t is the models and existing limits due to the CAST experiment. averaging time . The green and blue lines in g. 4 represent the poten- Phase Shift tial sensitivity to the up and downconversion regimes of this kind of experiment. The pump signal is taken to orig- Filter Cavity Gain inate from a state of the art, frequency stabilized cryo- genic sapphire oscillator (CSO) operating at the noise oor of a partially cryogenic frequency discrimination Gain system and white noise background. We further as- sume that the ratio of powers in the two modes is 1000, such that the readout mode has 1000 times less power dissipated in it. The mode structures, quality factors, resonance details and overlap integrals are discussed in Phase Shift section VI. Generally speaking, such a search would op- erate by detuning the two modes by some distance in fre- FIG. 5: Loop oscillator approach to dual mode axion quency space, and then searching the Fourier phase-noise detection. Two positive feedback loops utilise axion spectrum of the one of the resonances (say the higher fre- coupled modes as frequency selective elements. quency resonance for instance) for peaks corresponding to axion-induced phase-shifts. C. Loop Oscillator Axion Induced Spectral Density of Of course, for a given resonator geometry the upcon- Phase Measurements version technique will never be able to reach the same high-mass range as the downconversion technique, and Instead of relaying on an external frequency source, so the extension of the green line into the same frequency range as the blue line in g. 4 represents the sensitivity if one may construct an oscillator using the double mode cavity as a frequency selective element. Such a measure- we were to construct some resonator with the same sen- sitivity, but with mode frequencies such that this range ment setup is shown in Fig. 5. Discussion of such a sys- tem is given in Appendix C where the spectrum of phase was achievable with the upconversion technique. This is uctuations is given as follows: presented for the purpose of the direct comparison of the two techniques. By searching a few MHz in Fourier space h 2i 2 2 for such peaks, we are sensitive to slightly detuned ax- g  x D/U e  j S (f ) = 1 + S (f ) + S (f ) ; ';i 2 2 2 ions such that ! = !  ! + 2f , where f is our range a 1 2 f f + x of Fourier frequencies. We may then further detune the (20) resonances and repeat the process, gradually excluding a large section of the axion mass range. Of course, aver- where S are technical uctuations inside the oscillator aging for longer can yield improvements, and the above loop, and S (f ) is de ned as in (19). The multiplying plots are based on an averaging time of 30 days per 100 factor in this result is due to the Leeson e ect . MHz. The amount of frequency space covered in that The sensitivity plots of the loop oscillator experiment 30 day period will then depend on the range of Fourier for the room temperature and cryogenic implementations frequencies measured. are shown in Fig. 6. These sensitivities are calculated in KSVZ/DSFZ Models 7 consider di erent hybrid implementations. For example, one mode is used with a feedback loop as an oscillator, and the other is employed to measure the axion coupling induced phase shift with a help of external (highly stable) CAST frequency source. V. DEGENERATE MODE BROADBAND EXPERIMENT An interesting class of detection techniques may be de- veloped based on the degenerate frequency case scenario, where we require the two modes have equal frequencies (! = ! = !), and axion frequency (mass) is small 1 2 (!  !). Under these requirements, only the upconver- sion case is feasible. In the degenerate frequency case, considering only pure tones would put an exclusive re- quirement on the axion mass being near zero. The ac- FIG. 6: Sensitivity of the room temperature and tual signal would be searched for in a wide range of o set cryogenic versions of the loop oscillator experiment (Fourier) frequencies as discussed in Section IV and de- comparing to the axion models and existing limits due veloped in Appendices B and C. The obvious advantage to the CAST experiment. The dashed line is the of this approach is its broadband nature. Indeed, by mea- extension of the upconversion non-degenerate case down suring phase and amplitude noise in the degenerate case, to low frequencies. one has direct access to a few decades of Fourier frequen- cies. Such broadband measurements could be directly realised with the modern recon gurable digitizers. a similar fashion to the externally pumped loop experi- To analyse the Equations of Motion in the degenerate ment above. Indeed, the experiment would operate in the case, one may apply the same logic as in Appendix A, same way, by detuning the resonances some amount and B, C.Thus, the results obtained in Section III and IV are searching Fourier space, before detuning the resonance valid for the degenerate case as well. further and repeating the search. We again compare the The possibility to design a cavity with two orthogonal magnitude of axion-induced phase shifts - in this case modes of the same frequencies have been demonstrated given by 43,44 before . For the microwave frequency range, one may s argue that the mode degeneracy is not achievable due 1 p x to unavoidable imperfections of a cavity. Such imper- D/U j S = 1 + g ! !  jAj ; a 1 2 ';i;a 2 2 fections introduce coupling between the two modes that f f + x results in avoided level crossings. And it is due to this phenomenon the modes never coincide in the frequency S (f ) with the minimum detectable phase shift where space. Though, this is true, it is feasible to match the in this case the background noise, S (f ) is the phase frequencies within the mode bandwidth: by minimising noise spectral density of the loop oscillator that we are the all sorts of "imperfections", e.g. strongly coupled ex- measuring the output of (as opposed to the phase noise ternal probes, one may reduce mode-mode coupling to a spectrum, S ), and t is the averaging time. value that is much smaller than the mode bandwidth. In this case, no avoided level crossing could be observed. An For the cryogenic case, we are again assuming that the oscillator is a state of the art CSO operating at the example of a microwave cavity that can be used for the degenerate two mode axion sensing is discussed further noise oor of a partially cryogenic frequency discrimi- nation system and white noise background, and for the in Section VI. room temperature measurements we assume frequency Another practical problem immediately encountered stabilized loop oscillators based on copper cavities with with the degenerate frequency measurement scheme is in- the same properties as the open loop experiment, again ability to apply lters and separate the two signals. For discussed in more detail in section VI, and ampli ers with example, in the frequency detection scheme with two os- e ective added noise temperatures of 50 K, dissipating 1 cillators, the oscillators will synchronize due to unavoid- W in the resonator, operating at the noise oor of the able coupling between the two resonances of the same frequency discriminator and white noise background. It frequency. A possible solution to this problem is to treat is important to note that for the upconversion experi- the signals as one and use an external clock as a reference ment the sensitivities linearly decrease down to very low as depicted in Fig. 7. values of the frequency. Projected sensitivity of the degenerate broadband ex- In addition to purely frequency and phase measure- periment operated at both room and cryogenic temper- ment schemes presented in Fig. 3 and Fig. 5, one may ature ranges is shown in Fig. 8. This plot is calculated KSVZ/DSFZ Models 8 Phase Shift Filter Cavity Gain CAST SN1987A FIG. 7: A possible realisation of the degenerate mode detection scheme where two axion coupled modes of the same frequency are used as a frequency selective element for a feedback loop. The generated signal is compared against an external Local Oscillator (LO). FIG. 8: Sensitivity of the room temperature and based on the same parameters and phase expressions as cryogenic versions of the degenerate mode experiment the detuned loop oscillator experiment, simply setting comparing to the axion models and existing limits due the two resonance to the same frequency and sampling to CAST experiment and SN1987A. the phase noise spectrum up to 100 MHz. We again as- sume 30 days averaging time for this 100 MHz span - this will be limited to some extent by the achievable fre- A. TE-TM mode Cylindrical Cavity Resonator quency resolution of the digitiser. We note that these limits curve upwards as the loop oscillator phase noise As discussed, the sensitivity limits above are calcu- hits the white phase noise oor some distance into the lated based on the example of a microwave cavity with Fourier spectrum. As a result the straight, linear curves two orthogonally polarized resonances. Many other ge- for the exclusion limits in g. 6 (and indeed, the dashed ometries are possible, but for the purposes of demon- red line in g. 8) represent extending the atter part of strating the sensitivity of these techniques we have mod- the \broadband" exclusion limits, as it is in principle pos- elled the following. We take a 29.2 mm radius cylindrical sible to achieve this level of sensitivity at any point in the copper cavity with a TM mode frequency of 9 GHz, accessible frequency space, by simply setting the detun- and a TE mode frequency tunable from 6.5 to 9 GHz ing frequency accordingly and sampling only a narrow as the cavity height tunes from 18.5 to 83.6 mm. The region of Fourier space. Such an experiment would natu- loaded quality factors of both resonances are taken to rally scan far slower than 30 days per 100 MHz, but the be 10,000, which is readily achievable in copper at these curves in gs. 8 and 6 represent the limits of sensitivity frequencies. In this con guration, taking the stationary achievable with room temperature and cryogenic loop os- frequency TM mode to be mode \2" in the equations cillators, with feasible parameters, and within reasonable for the overlap integrals, we nd that  varies from -0.39 laboratory time scales. It is important to note that these to -0.50 as over the tuning range, whilst  varies from schemes, and indeed the open loop scheme presented in 0.46 to 0.57. Generally speaking, the analytical expres- g. 4, do carry technical diculties, and will require care- sions for these overlap integrals are messy and complex, ful design. For instance, these limits rely on the imple- and it is preferable to calculate them numerically for a mentation of an improved frequency stabilisation system given set of modes and cavity geometry. We found that for a CSO. However, this is not beyond the realm of what the magnitude of these expressions generally increased is achievable, and the presented limits should beviewed with the cavity aspect ratio, where h is the height of as the potential reach of this technique. the cavity, and a the radius, up to some maximum value achieved for aspect ratios greater than  5. B. Orthogonally Polarised Modes in a Fabry-P erot Cavity VI. SOME PRACTICAL REALISATIONS Another interesting orthogonally polarized mode All dual mode techniques considered in the present scheme to consider relies on Gaussian beam modes in work rely on a possibility to design a cavity that ex- Fabry-P erot cavities. Gaussian beams have the advan- hibits signi cant orthogonality of two modes represented tage that the electric and magnetic elds have the same by Eq. (8). These formulae give the cavity form factor pro le, simply rotated 90 degrees with respect to one an- that has to be as large as possible. other. If we take the electric eld of a given Gaussian KSVZ/DSFZ Models 9 beam to be polarized on the x-direction and propagating (f >10GHz) axion searches. Moreover, optical in the z-direction such that cavities might be used to probe otherwise unacces- sible regions of THz and infrared spectrum as well E = E(r; z) x; (21) as millimiter-wave and microwave frequencies; Liquid-Helium temperature operation (> 4K) then the magnetic eld is where only a limited number of components such B / E(r; z) y (22) as cavity and ampli ers have to be at low tem- perature. This factor removes the need of dilu- This means, that if we were to take two orthogonally po- tion refrigeration that is a key component in tra- 6{9,11 larized Gaussian beam modes in the same Fabry-P erot ditional haloscopes making the whole exper- cavity, with the same frequency, the variables  and iment available to a broader audience. Although would be 2 and 0 respectively. This has promising dilution refrigeration might give some incremental implications for high mass axion searches, via the down- improvement in the axion search, all ultra-stable conversion technique. microwave and optical clocks and oscillators do not However, we may also boost the sensitivity of the up- require temperatures below 4K. conversion technique, by employing two modes orthogo- access to higher and lower frequency ranges. The nally polarized in the same cavity, but of di erent fre- fact that actual axion mass is either the sum or dif- quency, ie a di erent number of wavelengths between the ference of working frequencies opens a possibility to two mirrors. This would give a non-zero value of the search for axions in less accessible frequency rages. parameter, which would open up the possibility of lower For instance, working around 20GHz, one is able mass axion searches with such structures. to probes axion masses in the vicinity of 40GHz, where experiments are signi cantly more dicult; CONCLUSIONS limited power levels (P < 100W cryogenically and 1 W at room temperature). Although, the sen- We demonstrated that precision frequency and phase sitivity does explicitly depend on power levels, on metrology could be used as a highly sensitive tool in dark the current calculation only limited power levels are 14,49,50 matter detection. Because measurements are made in the used unlike in some other proposals ; space of frequencies and phases rather than in the space axion phase sensitive. Comparing to DC magnet of amplitudes, the fundamental limit is set by control haloscopes, the dual frequency method is able to electronics rather than due to bare thermal uctuations. provide additional information about axions, par- This fact allows us to introduce a number of detection ticularly its phase relative to pump signals, al- schemes matching or exceeding sensitivity of state-of-the- though that might lead to more complicated de- art cryogenic axion detectors. This includes a broad band tection schemes; low mass axion detection based on the degenerate case 51{54 and Fourier spectra. In addition to bare axion amplitude KSVZ/DSFZ achievable. It is estimated that detection, the method allows to deduce relative phase of the cryogenic dual mode experiment is able to the axion signal. The proposed new class of detectors achieve the limit of the widely accepted axion dark may be understood as a phase sensitive detection scheme matter models. On the other hand, even a table- similar to phase sensitivity of parametric ampli ers. In top search may lead to competitive limits on dark summary, the main advantages of the proposed frequency matter; control method are: broadband search for low mass axions is 6{11 55,56 magnet-free. Unlike traditional haloscopes , the possible . It is demonstrated that a wide- proposed method does not require a strong DC band search that does not require tuning is magnetic elds; possible. SQUID-free. All sensitivities calculated in this work are based on usage of traditional low noise ACKNOWLEDGEMENTS semiconductor ampli ers. Though superconduct- ing technology might be used in the future, its pres- This work was supported by the Australian Research ence is not crucial contrary to traditional metods ; Council grant numbers DP160100253 and CE170100009. cavity volume independence. Although cavity vol- ume in uences many parameters of the experiment REFERENCES such as resonance frequencies and quality factors, the sensitivity is not directly proportional to this 11,45{48 parameter unlike in traditional haloscopes . J. Jaeckel and A. Ringwald, Annual Review of Nuclear and Par- This removes the major obstacle for higher mass ticle Science 60, 405 (2010). 10 2 33 R. Barkana, Nature 555, 71 (2018), arXiv:1803.06698 [astro- A. Lo, P. Haslinger, E. Mizrachi, L. Anderegg, H. Muller,  M. Ho- ph.CO]. hensee, M. Goryachev, and M. E. Tobar, Physical Review X 6, J. D. Bowman, A. E. E. Rogers, R. A. Monsalve, T. J. Mozdzen, 011018 (2016). and N. Mahesh, Nature 555, 67 (2018). M. Goryachev, Z. Kuang, E. N. Ivanov, P. Haslinger, H. Muller,  R. D. Peccei and H. R. Quinn, Phys. Rev. Lett. 38, 1440 (1977). and M. E. Tobar, IEEE Transactions on Ultrasonics, Ferro- J. Ipser and P. Sikivie, Phys. Rev. Lett. 50, 925 (1983). electrics, and Frequency Control, IEEE Transactions on Ultra- B. T. McAllister, G. Flower, E. N. Ivanov, M. Goryachev, sonics, Ferroelectrics, and Frequency Control 65, 991 (2018). J. Bourhill, and M. E. Tobar, Physics of the Dark Universe C. R. Locke, S. Munro, M. E. Tobar, E. N. Ivanov, and 18, 67 (2017). G. Santarelli, in IEEE International Frequency Control Sympo- A. Wagner, G. Rybka, M. Hotz, L. J. Rosenberg, S. J. Asztalos, sium and PDA Exhibition Jointly with the 17th European Fre- G. Carosi, C. Hagmann, D. Kinion, K. van Bibber, J. Hoskins, quency and Time Forum, 2003. Proceedings of the 2003 (2003) C. Martin, P. Sikivie, D. B. Tanner, R. Bradley, and J. Clarke, pp. 350{354. Phys. Rev. Lett. 105, 171801 (2010). E. Rubiola, The Cambridge RF and microwave engineering se- W. Chung, Proceedings, 15th Hellenic School and Workshops ries (Cambridge University Press, 2008) p. 220. on Elementary Particle Physics and Gravity (CORFU2015): M. Goryachev and S. Galliou, Ultrasonics 51, 966 (2011). Corfu, Greece, September 1-25, 2015, PoS CORFU2015, 047 M. Goryachev, Cryogenic BAW Resonators and Oscillators (2016). (Lambert Academic Publishing, 2012). 9 39 J. Choi, H. Themann, M. J. Lee, B. R. Ko, and Y. K. Se- E. N. Ivanov and M. E. Tobar, Review of Scienti c Instruments, mertzidis, Phys. Rev. D 96, 061102 (2017). Review of Scienti c Instruments 80, 044701 (2009). 10 40 A. Caldwell, G. Dvali, B. Majorovits, A. Millar, G. Ra elt, J. Re- M. E. Tobar, P. Wolf, A. Fowler, and J. G. Hartnett, Phys. Rev. dondo, O. Reimann, F. Simon, and F. Ste en (MADMAX Work- D71, 025004 (2005), [Erratum: Phys. Rev.D75,049902(2007)], ing Group), Phys. Rev. Lett. 118, 091801 (2017). arXiv:hep-ph/0408006 [hep-ph]. 11 41 B. M. Brubaker, L. Zhong, Y. V. Gurevich, S. B. Cahn, S. K. E. N. Ivanov, M. E. Tobar, and R. A. Woode, IEEE Transactions Lamoreaux, M. Simanovskaia, J. R. Root, S. M. Lewis, S. Al Ke- on Ultrasonics, Ferroelectrics, and Frequency Control 45, 1526 nany, K. M. Backes, I. Urdinaran, N. M. Rapidis, T. M. Shokair, (1998). K. A. van Bibber, D. A. Palken, M. Malnou, W. F. Kindel, M. A. D. Leeson, Proc. IEEE 54, 329 (1966). Anil, K. W. Lehnert, and G. Carosi, Phys. Rev. Lett. 118, K. Shigeru and K. Yoshio, Electronics and Communications in 061302 (2017). Japan (Part II: Electronics) 78, 31 (1995). 12 44 P. Sikivie, Physical Review Letters 51, 1415 (1983). M. D. Harvey and A. G. White, Optics Communications 221, A. C. Melissinos, Phys. Rev. Lett. 102, 202001 (2009). 163 (2003). 14 45 P. Sikivie, arXiv:1009.0762 [hep-ph] (2010). J. Jeong, S. Youn, S. Ahn, J. E. Kim, and Y. K. Semertzidis, C. M. Caves, Physical Review D 26, 1817 (1982). Physics Letters B 777, 412 (2018). 16 46 S. J. Asztalos, G. Carosi, C. Hagmann, D. Kinion, K. van Bib- J. Jeong, S. Youn, S. Ahn, C. Kang, and Y. K. Semertzidis, ber, M. Hotz, L. J. Rosenberg, G. Rybka, J. Hoskins, J. Hwang, Astroparticle Physics 97, 33 (2018). P. Sikivie, D. B. Tanner, R. Bradley, and J. Clarke, Physical B. T. McAllister, G. Flower, L. E. Tobar, and M. E. Tobar, Review Letters 104, 041301 (2010). Phys. Rev. Applied 9, 014028 (2018). 17 48 M. Aspelmeyer, T. J. Kippenberg, and F. Marquardt, Rev. Mod. M. Goryachev, B. T. McAllister, and M. E. Tobar, Physics Let- Phys. 86, 1391 (2014). ters A 382, 2199 (2018), special Issue in memory of Professor P. D. Nation and M. P. Blencowe, New Journal of Physics 12, V.B. Braginsky. 095013 (2010). P. W. Graham, I. G. Irastorza, S. K. Lamoreaux, A. Lindner, G. P. Agrawal and C. L. Mehta, Journal of Physics A: Mathe- and K. A. van Bibber, Ann. Rev. Nucl. Part. Sci. 65, 485 (2015), matical, Nuclear and General 7, 607 (1974). arXiv:1602.00039 [hep-ex]. 20 50 R. Gambini, Physical Review A 15, 1157 (1977). K. Ehret, M. Frede, S. Ghazaryan, M. Hildebrandt, E.-A. Kn- S. R. Parker, G. Rybka, and M. E. Tobar, Physical Review D abbe, D. Kracht, A. Lindner, J. List, T. Meier, N. Meyer, 87, 115008 (2013). D. Notz, J. Redondo, A. Ringwald, G. Wiedemann, and B. Yurke and E. Buks, Journal of Lightwave Technology, Journal B. Willke, Physics Letters B 689, 149 (2010). of Lightwave Technology 24, 5054 (2006). J. E. Kim, Phys. Rev. Lett. 43, 103 (1979). 23 52 C. Eichler and A. Wallra , EPJ Quantum Technology 1, 2 (2014). M. Shifman, A. Vainshtein, and V. Zakharov, Nuclear Physics M. A. Castellanos-Beltran and K. W. Lehnert, Applied Physics B 166, 493 (1980). Letters, Applied Physics Letters 91, 083509 (2007). M. Dine, W. Fischler, and M. Srednicki, Physics Letters B 104, B. A. Kochetov and A. Fedorov, in 2016 8th International Con- 199 (1981). ference on Ultrawideband and Ultrashort Impulse Signals (UW- A. R. Zhitnitsky, Sov. J. Nucl. Phys. 31, 260 (1980), [Yad. BUSIS) (2016) pp. 112{116. Fiz.31,497(1980)]. 26 55 G. Cerullo and S. De Silvestri, Review of Scienti c Instruments, Y. Kahn, B. R. Safdi, and J. Thaler, Phys. Rev. Lett. 117, Review of Scienti c Instruments 74, 1 (2003). 141801 (2016). 27 56 A. Olkhovets, D. W. Carr, J. M. Parpia, and H. G. Craig- B. T. McAllister, M. Goryachev, J. Bourhill, E. N. Ivanov, and head, in Technical Digest. MEMS 2001. 14th IEEE Interna- M. E. Tobar, (2018), arXiv:1803.07755 [physics.ins-det]. tional Conference on Micro Electro Mechanical Systems (Cat. No.01CH37090) (2001) pp. 298{300. T. Br acher, P. Pirro, and B. Hillebrands, Parallel pumping for Appendix A: Derivation of Two Mode Power Sensitivity magnon spintronics: Ampli cation and manipulation of magnon spin currents on the micron-scale, Physics Reports 699, 1 (2017). V. Frank and Y. Lan, Nanophotonics 1, 267 (2012). As it is mentioned in Section III, the problem of axion H. Du, G. Zhou, Y. Zhao, G. Chen, and F. S. Chau, Applied detection with two pumped modes has much in common Physics Letters, Applied Physics Letters 110, 061110 (2017). 22,23 with the analysis of parametric ampli ers , where one M. R. Foreman, J. D. Swaim, and F. Vollmer, Advances in optics and photonics 7, 168 (2015). is looking at manipulation of very weak signals on top of M. Nagel, S. R. Parker, E. V. Kovalchuk, P. L. Stanwix, J. G. strong pumps. Following the standard analysis in the Hartnett, E. N. Ivanov, A. Peters, and M. E. Tobar, Nature eld of parametric ampli ers, rstly, the steady state Communications 6, 8174 EP (2015). problem is solved for the strong pump signals. In the 11 case of Josephson parametric ampli ers, these equations axion coupling terms, the solution Eq. (A2) is valid for of motions are nonlinear. Secondly, the system dynamics this case as well. And, the equation of motion for the is analysed for small uctuations using linearised equa- small uctuations are then given as: tions. For this reason, each eld component is split into in strong pumped parts (C and B ) and small uctuations i ti n U m ec = ec g  eaC e n n n e m in dt (A5) (ec and b ): in i 2 b ; i i! t n pn c = C e + ec (t) e ; n n n where  = ! + ! ! is axion detuning in the U a pm pn in in i' in i! t n e pn (A1) b = B e + b (t) e ; upconversion case. The solution in the frequency domain n n n i! t can be given as follows: a = ea(t)e ; g  C in e m U i where ! and B represent pumping frequency and pn n ec [ ] = ea[ ]e (i magnitude for the nth mode and the axion signal has p (A6) only the small component ea. One may extend the model 2 in i b [ ]: to include an additional output \probe" and explicitly (i distinguish coupling and intrinsic losses. In the present In both upconversion and downconversion cases, Equa- analysis, b may be associated with technical uctuations tions (A6) and (A4), the mode uctuations consist of a and noise, while ea is the component that has to be de- noise component associated with the input signal b[ tected. The latter quantity is assumed to be slowly time- and signal component ea[ ] which is ampli ed by the varying to account for the "Quality factor" or linewidth strong signal deposited in the other mode C . of the axion signal that appears in many cosmological m models. Also, the phases of the pump signals ' are in- troduced with respect to the axion phase that is set to Appendix B: Axion Generated Phase Spectrum in a Dual zero. Mode Cavity On the rst step, due to extreme weakness of axions, the axion mediated coupling terms could be neglected In this Appendix we consider a case when an axion resulting into two linear independent equations. The re- signal is searched in the Fourier spectrum of phase noise sulting amplitudes are of the double mode cavity. In this case, we relax the re- quirement of the axion frequency to be exactly the sum or in i('  ) n m C = B e ; (A2) n di erence of resonant frequencies of two modes. Instead, n n this property is only approximately held true where the where = ! ! and phases are chosen to set the detuning constitute the Fourier frequency. In this case, n n pn coecents C real. ! = !  ! + where ! . Under this condition, n a 1 2 1 On the second step, the equations of motion are writ- the axion amplitude appear as a slowly varying param- eter in the EOMs. The objective of this derivation is to ten for small varying uctuations ec , ea and b where only n n nd a transfer function from this small and slow time- terms of the rst order in ec and ea are retained. Denoting varying quantity to amplitudes and phase uctuations = i + , the equations of motion in the Markovian n n n of the output signals of the cavity. The transfer func- and "slowly varying envelope" approximations: tion is derived by transferring the equations of motion d for the complex amplitudes to the equations for real sig- i t+i D m ec = ec + g  eaC e n n n e + m nal phases and magnitude and linearizing them around a dt (A3) in steady state point. i 2 b ; The starting point of this derivation is the equations of motion for the complex amplitudes C and C of the 1 2 where  = ! ! ! is axion detuning in the D a p1 p2 two modes in the upconversion case. Additionally, we downconversion case that can be set to zero. This EOM introduce two pump signals B and B , so the equations 1 2 can be easily solved by transforming the problem into the of motion can be written as follows: frequency domain: g  C e + m C = ( i ( ))C g  A( )C + B ( ); D i 1 1 1 1 e 2 1 ec [ ] = ea[ ]e (i p (A4) C = ( i ( ))C + g  A ( )C + B ( ); 2 2 2 2 2 e 1 2 in i b [ ]: (i ) (B1) For the upconversion case, the analysis follows the where  is the "slow time". Time variation of the detun- same steps. Moreover, since on the zeroth order step ing coecients  accounts for uctuations of the reso- considering the strong pump case does not include the nance frequencies of the modes as well as DC detuning 12 of the pump signals. Instead of complex amplitudes, we The linearised EOMs for the small uctuations are need to rewrite the equations of motion in terms of real phases and amplitudes: C ( ) = x ( ) exp(j' ( )) and i i i xe = xe x Q ( )+ 1 1 1 2 c B ( ) = y ( ) exp(j ( )). Here y and  ( ) represents d i i i i i amplitude and phase uctuations of the pump signals. e ye cos( ' ) y sin( ' )( 'e ); 1 1 1 1 1 1 1 1 The EOMs in the real amplitude-phase representations are written as follows: x 'e = x  + xe  x Q 1 1 1 1 1 1 2 s +ye sin( ' ) + y cos( ' )( 'e ); 1 1 1 1 1 1 1 1 (B6) x = x x Q ( ) + y cos( ' ); 1 1 1 2 c 1 1 1 xe = xe + x Q ( )+ 2 2 2 1 c d x y 2 1 ' =  ( ) Q ( ) + sin( ' ); ye cos( ' ) y sin( ' )( 'e ); 1 1 s 1 1 2 2 2 2 2 2 2 2 d x x 1 1 (B2) x 'e = x  + xe  x Q ( ) 2 2 2 2 2 2 1 s x = x + x Q ( ) + y cos( ' ); 2 2 2 1 c 2 2 2 d +ye sin( ' ) + y cos( ' )( 'e ): d x y 2 2 2 2 2 2 2 2 1 2 ' =  ( ) Q ( ) + sin( ' ); 2 2 s 2 2 d x x 2 2 that can be simply written in the matrix form as where Q and Q are two quadrature of the axion signal c s (B7) X = HX + A + N; de ned as follows: where X = [xe ; 'e ; xe ; 'e ] is a vector of state variables, 1 1 2 2 A = [x Q ( );x Q ( )=x ; x Q ( );x Q ( )=x ] Q ( ) = g  jAj cos( ( ) + ' ' ); 2 c 2 s 1 1 c 1 s 1 c e a 2 1 (B3) is a vector of axion signal, N is a vector of technical uc- Q ( ) = g  jAj sin( ( ) + ' ' ); s e a 2 1 tuations coming from the pump signal and internal mode instabilities. The system matrix is and  is the axion phase. 0 1 x 0 0 1 1 1 These are nonlinear equations with varying coecients B C 0 0 B 1 C that cannot be solved exactly. Instead, we linearise them x H = B C : (B8) @ 0 0  x A 2 2 2 for small signal uctuations around large steady state amplitudes. Here we assume that both axion signal and 0 0 input phase and amplitude uctuations are very small. Thus, we can split each variable and time varying pa- It is worth noting that in this linearised approxima- rameter into two components strong steady and small tion, all state variable uctuations as well as the axion uctuating: x = x + xe , y = y + ye , ' = ' + 'e and signal are fully uncoupled due to their extreme small- i i i i i i i i i e ness and ltered by the de ned system transfer matrix. =  + . As for the axion signal, it has only the small i i i Moreover, the two modes are completely uncoupled giv- uctuating part as described in the introduction to this ing two independent signals. The corresponding solutions Appendix. The steady state equations for the constant for technical uctuations have been previously analysed part are: 36{38 and can be found elsewhere . For this reason, we are interested only in axion-phase relationship that is given as follows: x = y cos( ' ); 1 1 1 1 1 0 =  x + y sin( ' ); 1 1 1 1 (B4) 'e [s] =  Q [s] i j;i c x = y cos( ' ); 2 2 2 2 2 2 s + 2 s +  + i i (B9) 0 =  x + y sin( ' ); s + 2 2 2 2 i Q [s] j;i s s + 2 s +  + i i that are two independent sets of two equations. Solv- where s is the Laplace variable, = is a ratio of j;i ing these equations give the steady state point (y ; ' ) in i i i stored amplitudes, + sign attributes to the second mode, terms of input signal parameters (y ;  ) and detunings and the is for the rst one. Q [s] and Q [s] are Laplace c s transforms of the two quadratures of the axion signals. It is important to empathise that these quadratures are de ned with respect to the phase di erence of the cavity x = p ; ' =  arctan : (B5) modes ' ' . Thus, generally the result is phase de- i i i 1 2 2 2 +  i i i pendent and could depend on the instance of time when 13 it starts. In other words, the axion signal provides an follows absolute time scale for this kind of experiment. In the case when both cavities are pumped on reso- 'e [s] = R [s]+ i j;i c nance ( = 0), transfer function (B9) is reduced to the i 2 2 2 s + 2 s +  + i i rst order low pass lter: (B14) s + R [s]; j;i s s + 2 s +  + 'e [s] i j;i i U i i H [s] = = : (B10) Q [s] s + s i which is di erent from the upconversion solution (B9) It is apparent from this result that the output phase com- only in the signs of each term as well as de nition of the ponent due to the axion signal is scaled by the ratio of axion quadratures. In these equations R [s] and R [s] c s magnitudes in both modes. If one is going to measure are Laplace transform of the axion signal de ned as in ' , it is advantageous to increase the x magnitude and (B13). Here the axion signal quadratures are de ned 1 2 keep x as small as possible. On the other hand when with respect to the sum of the phases of signals in both x ! 0, the carrier signal become undetectable. modes ' + ' . Thus like in the upconversion case, ex- 1 2 perimental results could be phase sensitive. In the case To derive the sensitivity of the setup over a range of of two cavities pumped on resonance, the result is also Fourier frequencies, we compare result (B10) to the trans- reduced to the rst order transfer low pass lter function: fer function for technical phase uctuations. Without lack of generality, we assume that technical uctuations are dominated by external uctuations  [s] whose trans- i 'e [s] i j;i i 37,38 H [s] = = : (B15) fer function in the phase space is . It can be s+ i R [s] s + s i shown that the internal cavity uctuations lead to a sim- ilar relation. A ratio of the axion signal and phase noise This result is identical to that of the upconversion case uctuations at the output of the cavity gives a constant (B10) up to the de nition of the axion quadratures and signal-to-noise ratio: the geometry factor. The corresponding signal-to-noise ratio is jAj SNR = g  : (B11) e j;i jAj i i D SNR = g  : (B16) e + j;i i i If the cavity is pumped on resonances, the same result is obtained for the case of limits due to external signal uctuations. Appendix C: Axion Generated Phase Spectrum in a Dual For the downconversion case, the equations of motion Loop Oscillator in terms of real phases and amplitudes are A loop oscillator is a frequency selective positive feed- x = x + x R ( ) + y cos( ' ); 1 1 1 2 c 1 1 1 back system in which two conditions of existence of sus- tained oscillations are ful lled: small signal gain across d x y 2 1 ' =  + R ( ) + sin( ' ); 1 1 s 1 1 the loop is greater than 1 and open loop phase shift is d x x 1 1 (B12) integer multiple of 2. These conditions make the small x = x + x R ( ) + y cos( ' ); 2 2 2 1 c 2 2 2 signal solution for the oscillator divergent, although due to amplitude limiting nonlinearity the self sustained sig- d x y 1 2 nals gets saturated as some value of circulating power. ' =  + R ( ) + sin( ' ): 2 2 s 2 2 d x x 2 2 To analyse phase-amplitude uctuations in such a sys- tem, in addition to cavity parameters, one needs to make where the axion signal quadratures are assumptions on the type of nonlinearity, ampli er gain and phase shift as well as ampli er-cavity couplings. In R ( ) = g  jAj cos( ( ) ' ' ); c e + a 2 1 this case, it is possible to apply the same type of analy- (B13) R ( ) = g  jAj sin( ( ) ' ' ): sis as described in Appendix B. On the other hand, one s e + a 2 1 may assume a certain level of circulating power, e.g in the form of magnitudes in the cavity modes x and x , The solution of the corresponding steady state equa- 1 2 and analyse the system in the small signal regime. In this tion is the same as in the upconversion case described regime, an ideal ampli er has a unity transfer function in above. Dynamics of the small magnitude-phase uctua- the phase space. The resulting system for the ith mode tions can be described by the same matrix equation (B7) is demonstrated in Fig. 9. The phase uctuations at the and system matrix (B8) with a di erent axion input vec- output of the oscillator are found as a sum of contribu- tor: A = [x R ( ); x R ( )=x ; x R ( ); x R ( )=x ] . 2 c 2 s 1 1 c 1 s 1 The corresponding axion-phase relationship appears as tions from the technical noise  and axion signal R (or i s 14 Q ): h i i j;i 'e = 1 +  + R ; i s s s + h i  (C1) i j;i 'e = 1 +  + Q ; i s s s + where the rst term constitutes the Leeson e ect for the FIG. 9: Schematic representation of the axion to phase technical phase uctuations in the loop and the second conversion in a feedback oscillator in the phase space. term represents the axion induced phase uctuations.

Journal

High Energy Physics - Phenomenology – arXiv (Cornell University)

Published: Jun 19, 2018

There are no references for this article.