Exploiting higher-order resonant modes for axion haloscopes
Jinsu Kim, SungWoo Youn, Junu Jeong, Woohyun Chung, Ohjoon Kwon, Yannis K. Semertzidis
EExploiting higher-order resonant modes for axionhaloscopes
Jinsu Kim , , SungWoo Youn , ∗ , Junu Jeong , , WoohyunChung , Ohjoon Kwon , Yannis K. Semertzidis , Center for Axion and Precision Physics Research, Institute for Basic Science,Daejeon 34047, Republic of Korea Department of Physics, Korea Advanced Institute of Science and Technology,Daejeon 34141, Republic of KoreaE-mail: *[email protected]
27 November 2019
Abstract.
The haloscope is one of the most sensitive approaches to the QCD axionphysics within the region where the axion is considered to be a dark matter candidate.Current experimental sensitivities, which rely on the lowest fundamental TM modeof a cylindrical cavity, are limited to relatively low mass regions. Exploiting higher-order resonant modes would be beneficial because it will enable us to extend thesearch range with no volume loss and higher quality factors. This approach has beendiscarded mainly because of the significant degradation of form factor, and difficultywith frequency tuning. Here we introduce a new tuning mechanism concept whichboth enhances the form factor and yields reasonable frequency tunability. A proof ofconcept demonstration proved that this design is feasible for high mass axion searchexperiments.
Keywords : axion, haloscope, higher-order mode, dielectric, tuning mechanism
1. Introduction
The axion is a consequence of the PQ mechanism proposed to solve the strong-CP problem in particle physics [1]. If the mass falls within a certain range, it hascosmological implications as cold dark matter [2]. A methodological approach to detectthe axion signal employs cavity haloscopes, where, under a strong magnetic field, axionsare converted to microwave photons resonating with a cavity mode [3]. The axion-to-photon conversion power is given by P a → γγ = g aγγ ρ a m a B V C min( Q L , Q a ) , where g aγγ is the axion-to-photon coupling, ρ a is the local halo density, m a is the axionmass, B is the external magnetic field, V is the cavity volume, C is the form factor, a r X i v : . [ phy s i c s . i n s - d e t ] N ov xploiting higher-order resonant modes for axion haloscopes Q L and Q a are the cavity (loaded) and axion quality factors. As the axion mass isa priori unknown, all possible mass ranges need to be scanned. From the experimentalpoint of view, the figure of merit is the scan rate, which is written as dfdt = (cid:18) (cid:19) (cid:18) P a → γγ k B T sys (cid:19) Q a Q L , where SNR is the signal-to-noise ratio, k B is the Boltzmann constant, and T sys is thenoise temperature of the system.The form factor has dependence on the cavity geometry and resonant mode: C ≡ (cid:12)(cid:12)(cid:12)(cid:82) V (cid:126)E c · (cid:126)B d x (cid:12)(cid:12)(cid:12) (cid:82) V (cid:15) ( x ) | (cid:126)E c | d x (cid:82) V | (cid:126)B | d x , (1)where (cid:126)E c is the electric field of the cavity mode under consideration, and (cid:15) ( x ) is thedielectric constant inside the cavity volume. For cylindrical cavities, the TM mode isconventionally considered since it yields the largest form factor in a homogeneous staticmagnetic field [4].To date, the axion haloscope is the most promising approach with sensitivity to theQCD axion models [5, 6]in the mass range between 10 and 10 µ eV, where the axionis considered to be a candidate for cold dark matter. However, current experimentalsensitivities are limited to relatively low mass regions since cavity-based experimentstypically employ a single resonant cavity for a large detection volume and adopt thelowest resonant mode for the maximum form factor [7, 8, 9, 10]. Some cavity designshave been proposed to efficiently explore higher mass regions with minimal volume losswhile relying on the same resonant mode [11, 12].As an alternative method to extend the search range towards higher mass regions,it could be beneficial to exploit the higher-order resonant modes in a cylindrical cavity.This would enable us to access higher frequency regions without volume loss and evenwith higher quality factors, as summarized in Table 1. However, as shown in Fig. 1,high-degree variations in the cavity EM field give rise to out-of-phase field components,which, under an external magnetic field, results in negative contributions to the formfactor in Eq. 1. The negative effect becomes larger with the increasing order of theTable 1: Parameters of a cylindrical cavity for different resonant modes: resonantfrequency ( f ), volume ( V ), quality factor ( Q ), and form factor ( C ). f , V , and Q are values relative to those of the TM mode, while C is the absolute value. Mode f V Q C TM f V Q . × f V . × Q . × f V . × Q xploiting higher-order resonant modes for axion haloscopes rz plane) of the TM modes, listed in Table 1, ofa cylindrical cavity. The maximum field strength and cavity radius are normalized tounity.resonant mode, as can be seen in Table 1. This significantly reduces the experimentalsensitivity, and consequently, the higher modes have not been considered for axion searchexperiments. However, a (periodic) structure of dielectric material can suppress the out-of-phase electric field component(s), which would enhance the form factors substantially,making reasonable sensitivities achievable. For cylindrical cavities, a (periodic) layer(s)of dielectric hollow(s) can be considered for this purpose. Figure 2 provides an exampleof such a dielectric effect and intuitive design for the TM mode of a cylindrical cavity.In this article, we examine several tuning mechanisms utilizing higher-order resonantmodes, in particular the TM mode, to find a suitable approach for high mass axionsearches. (a) (b) Figure 2: (a) Effect of a dielectric medium (in cyan) on the electric field of the TM mode. The black solid and green dashed lines are the electric field profiles with andwithout the medium. (b) A dielectric hollow structure for a cylindrical cavity toimplement the effect. xploiting higher-order resonant modes for axion haloscopes mode is considered for further study. In Sec. 3, weexamine several frequency tuning schemes by breaking the field symmetry along eachdirection of the cylindrical coordinate system, in the longitudinal, radial and azimuthaldirections. Finally, a simple and realistic tuning mechanism is chosen to demonstrateits feasible application to axion haloscopes in Sec. 4.
2. Analytic solutions for a cylindrical cavity with a dielectric hollow
The general electromagnetic field solutions for the TM modes of a cylindrical cavity,where the field symmetry is conserved along the longitudinal and azimuthal directions,are given in the cylindrical coordinate system ( r, φ, z ) by (cid:126)E = ( AJ ( √ (cid:15)kr ) + BY ( √ (cid:15)kr ))ˆ z,(cid:126)B = √ (cid:15)c [ AJ ( √ (cid:15)kr ) + BY ( √ (cid:15)kr )] ˆ φ, (2)where J α and Y α are the Bessel’s functions of the first and second kinds for the non-negative integer α , (cid:15) is the dielectric constant, and k is the wave number in vacuum. A and B are coefficients to be determined by boundary conditions. With a cylindricaldielectric hollow in the cavity as seen in Fig. 2(b), there are three distinct regions -inner, medium, and outer regions, which are denoted by the subscripts i , m , and o ,respectively, with the inner and outer regions assumed to be in vacuum, i.e., (cid:15) = 1.For the inner region, B i = 0 is required as Y diverges at the origin. The electricfield amplitude is normalized to unity, i.e., A i = 1, such that the coefficients in the otherregions are scaled accordingly. In the medium region filled with a dielectric materialwith (cid:15) >
1, the continuity of the fields at the inner and outer surfaces is imposed by theboundary conditions J ( kr i ) = A m J ( √ (cid:15)kr i ) + B m Y ( √ (cid:15)kr i ) ,J ( kr i ) = √ (cid:15)A m J ( √ (cid:15)kr i ) + √ (cid:15)B m Y ( √ (cid:15)kr i ) , and A m J ( √ (cid:15)kr o ) + B m Y ( √ (cid:15)kr o ) = A o J ( kr o ) + B o Y ( kr o ) , √ (cid:15) [ A m J ( √ (cid:15)kr o ) + B m Y ( √ (cid:15)kr o )] = A o J ( kr o ) + B o Y ( kr o ) , (3)where r i and r o are the inner and outer radii of the dielectric hollow. It is straightforward xploiting higher-order resonant modes for axion haloscopes A m = J ( kr i ) Y ( √ (cid:15)kr i ) − J ( kr i ) Y ( √ (cid:15)kr i ) / √ (cid:15)J ( √ (cid:15)kr i ) Y ( √ (cid:15)kr i ) − J ( √ (cid:15)kr i ) Y ( √ (cid:15)kr i )= − π kr i (cid:2) √ (cid:15)J ( kr i ) Y ( √ (cid:15)kr i ) − J ( kr i ) Y ( √ (cid:15)kr i ) (cid:3) ,B m = − J ( kr i ) J ( √ (cid:15)kr i ) − J ( √ (cid:15)kr i ) J ( kr i ) / √ (cid:15)J ( √ (cid:15)kr i ) Y ( √ (cid:15)kr i ) − J ( √ (cid:15)kr i ) Y ( √ (cid:15)kr i )= π kr i (cid:2) √ (cid:15)J ( kr i ) J ( √ (cid:15)kr i ) − J ( kr i ) J ( √ (cid:15)kr i ) (cid:3) , and A o = J ( √ (cid:15)kr o ) Y ( kr o ) − J ( √ (cid:15)kr o ) Y ( kr o ) √ (cid:15)J ( kr o ) Y ( kr o ) − J ( kr o ) Y ( kr o ) A m + Y ( √ (cid:15)kr o ) Y ( kr o ) − Y ( √ (cid:15)kr o ) Y ( kr o ) √ (cid:15)J ( kr o ) Y ( kr o ) − J ( kr o ) Y ( kr o ) B m = − π kr o (cid:8)(cid:2) J ( √ (cid:15)kr o ) Y ( kr o ) − J ( √ (cid:15)kr o ) Y ( kr o ) √ (cid:15) (cid:3) A m + (cid:2) Y ( √ (cid:15)kr o ) Y ( kr o ) − Y ( √ (cid:15)kr o ) Y ( kr o ) √ (cid:15) (cid:3) B m (cid:9) ,B o = − J ( √ (cid:15)kr o ) J ( kr o ) − J ( √ (cid:15)kr o ) J ( kr o ) √ (cid:15)J ( kr o ) Y ( kr o ) − J ( kr o ) Y ( kr o ) A m − J ( kr o ) Y ( √ (cid:15)kr o ) − J ( kr o ) Y ( √ (cid:15)kr o ) √ (cid:15)J ( kr o ) Y ( kr o ) − J ( kr o ) Y ( kr o ) B m = π kr o (cid:8)(cid:2) J ( √ (cid:15)kr o ) J ( kr o ) − J ( √ (cid:15)kr o ) J ( kr o ) √ (cid:15) (cid:3) A m + (cid:2) J ( kr o ) Y ( √ (cid:15)kr o ) − J ( kr o ) Y ( √ (cid:15)kr o ) √ (cid:15) (cid:3) B m (cid:9) . The boundary condition for a perfect electric conductor requires a vanishingtangential component of the E -field, i.e. E z ( r = R ) = 0, for which only TM n modesare allowed. This yields from Eq. 2 A o J ( k n R ) + B o Y ( k n R ) = 0 , (4)from which the resonant frequency is extracted as f n = ω n / π = k n c/ π . In thisarticle, we particularly consider n = 3, i.e., TM resonant mode, for further study. The optimal thickness of the dielectric hollow, r o − r i , is determined by maximizing aphysical quantity, V C Q . This quantity is relevant to the experimental sensitivity inEq. 1 and called the scan rate factor in this manuscript. The individual parameters canbe expressed in terms of the electric field E in the following manner.For a cylindrical cavity with radius R and length L , the volume is V = πR L . Theform factor in Eq. 1 under a uniform external magnetic field in the z direction becomes C = | (cid:82) V E z d x | V (cid:82) V (cid:15) ( (cid:126)x ) | (cid:126)E | d x . xploiting higher-order resonant modes for axion haloscopes φ and z directions, the integrals are simplified as (cid:90) V E z d x = 2 πL (cid:90) R rE z dr, (5)and (cid:90) V (cid:15) ( (cid:126)x ) | (cid:126)E | d x = 2 πL (cid:88) j (cid:15) j (cid:90) j r | (cid:126)E | dr, (6)with j denoting the three distinct regions. The cavity quality factor is obtained from arelation Q = GR s , where G is the geometry factor and R s is the surface resistance. The mode dependentgeometry factor is given by G = µ ω (cid:82) V | (cid:126)H | d x (cid:72) S | (cid:126)H | d x , (7)where µ is the vacuum permeability and (cid:126)H is the magnetic field strength of the resonantmode under consideration. For normal conductors, the surface resistance in the radiofrequency regime is expressed as R s = 1 δσ , where δ is the skin depth, δ ≡ (cid:112) /µ σω , and σ is the metal conductivity.Using geometrical symmetries and the fact that stored electric and magnetic fieldsshare an equal amount of energy, the volume and surface integrals in Eq. 7 become,respectively, (cid:90) | (cid:126)H | d x = 1 µ (cid:90) (cid:15) ( (cid:126)x ) | (cid:126)E | d x, (8)and (cid:73) | (cid:126)H | d x = 2 πRL | (cid:126)H ( r = R ) | + 2 (cid:90) end | (cid:126)H | d x. (9)The last term in Eq. 9 is for the end surfaces and is further developed to (cid:90) end | (cid:126)H | d x = 1 L (cid:90) | (cid:126)H | d x = 1 Lµ (cid:90) (cid:15) ( (cid:126)x ) | (cid:126)E | d x. By plugging Eq. 6 into Eqs. 8 and 9, we obtain (cid:90) | (cid:126)H | d x = 2 πLµ (cid:88) j (cid:15) j (cid:90) j r | (cid:126)E | dr (cid:73) | (cid:126)H | d x = 2 πRL | (cid:126)H ( r = R ) | + 4 πµ (cid:88) j (cid:15) j (cid:90) j r | (cid:126)E | dr. It is noted that all the integrals required to calculate the scan rate factor, V C Q ,are expressed in two forms: (cid:90) R rE z dr and (cid:90) j r | (cid:126)E | dr. xploiting higher-order resonant modes for axion haloscopes r i and r o , are foundby requiring d ( V C Q ) /dr = 0, which unfortunately cannot be resolved analytically.Instead, we obtain the optimal values from a numerical approach using WolframMathematica computer program [13], as shown in Fig. 3(a). It is found that the optimalthickness of a dielectric hollow with (cid:15) corresponds to approximately λ/ √ (cid:15) , where λ isthe wavelength of the EM wave of the TM resonant mode with the dielectric hollowplaced, which is approximately equal to the cavity radius R , as can be inferred fromFig. 2(a) ‡ . We found that this optimized dimension is consistent with that determinedby a simulation study, using COMSOL Multiphysics R (cid:13) software [14], where we modeleda cylindrical cavity with a 90 mm diameter and 100 mm height, and a dielectric hollowwith (cid:15) = 10. Figure 3(b) shows the dependence of the scan rate factor and resonantfrequency on the thickness of the hollow. The presence of the optimized dielectricstructure makes an enhancement to the form factor by almost an order of magnitude(from 0.05 to 0.45), resulting in a substantial improvement in scan rate (about 60 timeshigher than that in its absence). (a) (b) Figure 3: (a) Two-dimensional plot of the scan rate factor as a function of the inner andouter radii of a dielectric cylindrical hollow. Mathematica was used to compute V C Q along with the EM solutions driven in the text. The red cross, which reads r i = 0 . r o = 0 .
52, represents the optimal point that maximizes the scan rate factor. (b)Scan rate factor and resonant frequency as a function of the thickness (relative to thecavity radius R ) of the dielectric hollow with (cid:15) = 10. The dashed line represent theoptimal thickness ( λ/ √ (cid:15) ), which yields the highest sensitivity. ‡ For reference, the wavelength of the TM mode for an empty cylindrical cavity with a radius R isgiven by λ = 2 πR/χ = 0 . R , where χ = 8 .
654 is the 3rd zero of the Bessel’s function J ( kr ). xploiting higher-order resonant modes for axion haloscopes
3. Frequency tuning mechanisms
Since the axion mass (equivalently the frequency of the converted photon) is a prioriunknown, a cavity must be tunable to cover a wide frequency range for a given cavitydimension. In this section, we examine plausible frequency tuning mechanisms for theTM mode utilizing the dielectric structure depicted in Sec. 2. We consider symmetrybreaking of the electric field along the longitudinal, radial and azimuthal directions ofthe cylindrical coordinate system.
Recently, there was a study that showed a potential application of Bragg reflectorswith a dielectric hollow using the TM mode of cylindrical cavities [15]. The studypresented a frequency tuning scheme which breaks the field symmetry along the cavityaxis by splitting a dielectric cylindrical hollow in the longitudinal direction. Since theperformance of the mechanism was already evaluated in that study, no attempt is madein this article for any further studies.
A conventional way to tune the resonant frequency of a cylindrical cavity is to breakthe transverse symmetry by translating a (pair of) dielectric or conducting rod(s) inthe radial direction [7, 8]. For a cylindrical dielectric cavity, such symmetry breakingcan be achieved by splitting the dielectric hollow vertically into pieces and moving themapart along the radial direction. For simplicity, a two-piece scheme, which is illustratedin Fig. 4, is considered in this study.Figure 4: Illustration of the tuning scheme along the radial direction using two half-hollow cylinders. The color represents the strength of the electric field for the TM mode.The initial size of the hollow cylinder is chosen from the value derived in Sec. 2. Wefound however that even though a thickness of λ/ √ (cid:15) yielded the highest sensitivity, itresulted in a lack of tunability. A dedicated simulation study, some results of which areshown in Fig. 5, indicates that the sensitivity and tuning range need to be compromised. xploiting higher-order resonant modes for axion haloscopes § of V C Q and ∆ f , the frequencytuning range, we find that λ/ √ (cid:15) is the optimal thickness with reasonable tunability( ∼
9% with respect to the central frequency) and high sensitivity ( ∼
80% of that for λ/ √ (cid:15) on average). The maximum tuning range is achieved when the two half cylindersare separated by approximately λ/ √ (cid:15) . However, the mechanical design for displacingthe long hollow pieces in the radial direction could be challenging to implement in suchsensitive experiments.Figure 5: Simulation results of scan rate factor and tunable frequency range fordifferent dielectric thicknesses (DTs) relying on the radial tuning mechanism. Thesame cavity dimension and dielectric property as in Sec. 2.2 were used. The areas incolor, corresponding to the figure of merit newly defined in the text, are 0.29 (blue),0.47 (orange) and 0.07 (green), in unit of m · MHz, respectively.
The study performed in Sec. 2.2 showed that the resonant frequency has a strongdependence on the thickness of the dielectric hollow, as seen in Fig. 3(b). Based onthis feature, we designed a new tuning mechanism, which consists of a double layer ofconcentrically segmented dielectric hollow pieces, with one layer of the segments rotatingwith respect to the other. In this scheme, the outer layer of dielectric segments is fixedin position, while the inner layer is turned along the azimuthal direction simultaneouslyby a single rotator. Figure 6 visualizes how this scheme alters the effective thicknessof the dielectric hollow over the course. The corresponding electric field distributionsare also shown in Fig. 7. This tuning mechanism concept essentially relies on thetransformation of continuous rotational symmetry into a discrete rotation symmetry, tocreate a tuning capability. It is noted that since the resonant frequency is tunable usinga single rotator, this design provides a more reliable tuning mechanism, one which isalso easy to implement. § Strictly speaking, the new figure of merit is an integration of the scan rate factor over the tuningrange, i.e., (cid:82) ∆ f ( V C Q ) df . xploiting higher-order resonant modes for axion haloscopes -like mode with the 6-segmenttuning mechanism applied to suppress the negative field component. Each distributioncorresponds to the rotational angle of the inner layer in Fig. 6.The design of the tuning system was optimized based on simulation studies byconsidering three parameters: 1) number of segments; 2) thickness of the double layer;and 3) inner radius of the inner layer. Assuming that there were no (or weak) correlationsbetween one another, the parameters were scanned independently to find the optimalvalues which maximized the aforementioned figure of merit, (cid:82) ∆ f ( V C Q ) df . Using amodel of a copper cylindrical cavity with dielectric segments of (cid:15) = 10, it was foundthat a six segment design gives the best performance. The optimal thickness of thelayer pair and the radius of the inner layer were obtained to be approximately λ/ √ (cid:15) and 0.37 R , which are consistent with the analytically calculated values. The relativethickness between the inner and outer layers was also fine tuned, such that a thickerinner layer relative to the outer layer yielded a slightly better tunability, with the overallsensitivity remaining intact. We observed that the frequency tuning range of the TM resonant mode was about 6% with respect to the central frequency and the form factorwas enhanced to greater than 0.33 (compared to 0.05 for an empty cavity) over theentire tuning range. Sections 3.1 − mode, based on symmetrybreaking along the longitudinal, radial, and azimuthal directions. Using the COMSOLsimulation tool, the performance of the individual mechanisms was evaluated in termsof scan rate factor and frequency tunability. We modeled a cylindrical cavity with the xploiting higher-order resonant modes for axion haloscopes mode. The area below each line corresponds to the figure of merit defined inthe literature. The equivalent performance for an octuple-cell cavity [12], which relieson the TM mode, is also compared.
4. Experimental demonstration
A cavity and a tuning system were fabricated to demonstrate the experimental feasibilityof the chosen tuning mechanism. We employed a split-type cavity design, which wasinitially introduced in Ref. [16]. Made of oxygen-free high conductivity copper, thecavity consists of three identical pieces. The assembly builds a cylindrical cavity with a90 mm inner diameter and 100 mm inner height and introduces a narrow hole throughthe center of both the top and bottom ends. Inside each cavity piece, two outerdielectric segments were placed at fixed positions, as shown in Fig. 9 (a). Highly purealuminum oxide (99.7% Al O with (cid:15) = 9 .
8) was chosen as the dielectric material. Sixsmaller dielectric pieces, supported by a pair of wheel-shaped structures at the topand bottom, composed the inner layer of the tuning system, as shown in Fig. 9 (b).The support structure was made of polytetrafluoroethylene (PTFE), which has a lowdielectric constant ( (cid:15) = 2 .
1) and a low dissipation factor (on the order of 10 − at 4 K).The bottom structure was fabricated to have an extended rod piece in the middle, whichwas attached through the cavity hole to a single rotational piezo actuator outside thecavity to simultaneously turn the inner layer. The overall structure of the tuning system xploiting higher-order resonant modes for axion haloscopes wheel mechanism. (a) (b) (c) Figure 9: Cavity design. (a) One of three pieces of the split cavity with two outerdielectric segments fixed in place. (b) Assembled inner layer composed of six smallersegments and a pair of wheel-shaped structures. (c) Photo of the partially assembledcavity with the tuning system mounted.The assembled cavity was installed in a cryogenic system and brought to a lowtemperature of around 4 K. The resonant frequencies and quality factors were measuredusing a network analyzer, through transmission signal between a pair of monopole RFantennae weakly coupled to the cavity. A frequency map of the resonant modes ofthe cavity was drawn as a function of the tuning step, while rotating the inner layerof the tuning system. We observed the periodic behavior of the resonant modes overa full tuning process. Figure 10 shows the frequency map over a single cycle, whichcorresponds to the rotational angle of 60 ◦ . The bell-shaped curve with its frequencyspanning from 7.02 to 7.32 GHz corresponds to the TM mode. This frequency regionis about three times higher than the TM resonant frequency of the same cavity, f = 2 .
55 GHz. Symmetric and smooth frequency curves over a tuning range of ∼
300 MHz indicates the stability of the tuning mechanism with reasonable tunability.The measured quality factors varied between 110,000 in the low frequency regions and90,000 for the high frequency regions, which were consistent with the simulation results.This verifies that the design is a plausible approach for utilizing the higher-order resonantmodes for axion haloscopes.
5. Conclusions
We exploited higher-order resonant modes for axion haloscope experiments to extendthe search range towards high mass regions. In particular, we examined various tuningmechanisms for the TM mode by introducing a structure of dielectric material, whichsubstantially enhanced the form factor. The general EM solutions for a cylindrical cavitywere obtained analytically and found to be consistent with the numerical calculations. xploiting higher-order resonant modes for axion haloscopes mode. A few mode crossings by TE modes are also observed.Depending on the way that the field symmetry is broken, three schemes can beconsidered: frequency tuning along the longitudinal, radial, and azimuthal directions.For each scheme, a tuning system was optimally designed based on simulation studiesby maximizing a new figure of merit, (cid:82) ∆ f ( V C Q ) df . Among those, the tuning schemethat relies on the symmetry breaking along the azimuthal direction, showed the highestsensitivity over a reasonable tuning range with a realistic design. The experimentalfeasibility of this mechanism was demonstrated using a three-piece copper cavity and adouble-layer of alumina hollow segments. We conclude that higher-order resonant modesutilizing suitable frequency tuning mechanisms are certainly applicable to haloscopesearches for high mass dark matter axions. Acknowledgments
This work was supported by IBS-R017-D1-2019-a00 / IBS-R017-Y1-2019-a00.
References [1] R. D. Peccei and H. R. Quinn,
CP Conservation in the Presence of Pseudoparticles , Phys. Rev.Lett. , 1440 (1977); R. D. Peccei and H. R. Quinn, Constraints imposed by CP conservationin the presence of pseudoparticles , Phys. Rev. D , 1791 (1977).[2] L. F. Abbott and P. Sikivie, A cosmological bound on the invisible axion , Phys. Lett. B , 123(1983); J. Preskill, M.B. Wise and F. Wilczek,
Cosmology of the invisible axion , Phys. Lett. B , 127 (1983); M. Dine and W. Fischler,
The not-so-harmless axion , Phys. Lett. B , 137(1983).[3] P. Sikivie,
Experimental Tests of the “Invisible” Axion , Phys. Rev. Lett. , 1415 (1983).[4] R. Bradley et al. , Microwave cavity searches for dark-matter axions , Rev. Mod. Phys. , 777(2003).[5] J. E. Kim, Weak-interaction singlet and strong CP invariance , Phys. Rev. Lett. , 103 (1979); M. xploiting higher-order resonant modes for axion haloscopes A. Shifman, A. I. Vainshtein, and V. I. Zakharov,
Can confinement ensure natural CP invarianceof strong interactions? , Nucl. Phys. B , 4933 (1980).[6] A. P. Zhitnitsky,
On the possible suppression of axion hadron interactions , Yad. Fiz. , 497 (1980)[Sov. J. Nucl. Phys. , 260 (1980)]; M. Dine, W. Fischler and M. Srednicki, A simple solutionto the strong CP problem with a harmless axion , Phys. Lett. B , 199 (1981).[7] N. Du et al. (ADMX Collaboration),
Search for Invisible Axion Dark Matter with the Axion DarkMatter Experiment , Phys. Rev. Lett. , 151301 (2018).[8] L. Zhong et al. (HAYSTAC Collaboration),
Results from phase 1 of the HAYSTAC microwavecavity axion experiment , Phys. Rev. D , 092001 (2018).[9] B. T. McAllister et al. , The ORGAN Experiment: An axion haloscope above 15 GHz , Phys. DarkUniv. , 67 (2017).[10] S. Youn, Axion Research at CAPP/IBS , Int. J. Mod. Phys. Conf. Ser. , 1660193 (2016); J. Choi et al. , First axion dark matter search with toroidal geometry , Phys.Rev. D , 061102 (2017).[11] J. Jeong et al. , Phase-matching of multiple-cavity detectors for dark matter axion search , Astropart.Phys. , 33 (2018).[12] J. Jeong et al. , Concept of multiple-cell cavity for axion dark matter search , Phys. Lett. B ,412 (2018).[13] Wolfram Research, Inc., Mathematica, Version 8.0, Champaign, IL (2010).[14] COMSOL Multiphysics R (cid:13) et al. , Tunable Supermode Dielectric Resonators for Axion Dark-MatterHaloscopes , Phys. Rev. Applied , 014028 (2018).[16] S. Ahn et al. , Magnetoresistance in copper at high frequency and high magnetic fields , J. Instrum.,12