@misc{oai:ir.soken.ac.jp:00000606,
 author = {田島, 健 and タジマ, ツヨシ and TAJIMA, Tsuyoshi},
 month = {2016-02-17, 2016-02-17},
 note = {KEKB, B-Factory at KEK, is a double ring collider of 3.5GeV positrons and 8GeV electrons.  The two rings were named LER (Low Energy Ring) and HER (High Energy Ring) after their energies.  They were constructed in the tunnel that was used for TRISTAN.  One of the main physics objectives of this facility is the detection of so-called CP-violation, which may give an answer to the question why the amount of matter and anti-matter is not the same.<br /><br />With regard to the accelerator, KEKB is a very challenging accelerator in the sense that it has to deal with ampere-class current (2.6A for LER and 1.1A for HER). which is about two orders of magnitude higher than that of TRISTAN and has never been used with any machine before.  To achieve a high current without suffering from single bunch instabilities, the beam will consist of some 5000 bunches, i.e., every RF bucket filled with a bunch of several 1010 positrons/electrons.  In this case, the HOM (Higher Order Mode) fields that are excited by a bunch at accelerator components, such as RF cavities, are likely to affect the following bunches because the decay times of the fields are sometimes longer than the bunch interval, e.g., 2ns for KEKB.<br /><br />This situation makes the beams susceptible to coupled bunch instabilities, i.e. intolerable growth of the oscillatory coupled-bunch motions.  For the success of KEKB, it is essential to lower the HOM impedances so that the threshold currents are higher than the designed currents or they are manageable with a bunch-by-bunch feedback system.  The most effective way to achieve this is to lower Q's (quality factors) and R/Q's (effective shunt impedances) of most HOM's in RF cavities since RF cavities are normally the main source of HOM's.*  This can be embodied by so-called HOM-damped structures.  At KEK, we use two types of such structures, one is called ARES (Accelerator REsonantly coupled Structure) and the other is SCC (SuperConducting Cavity).<br /><br />ARES is the inveniton of creative people at KEK.  Using a three-cell structure with a large energy-storing cell, it can lower the R/Q of fundamental mode to 15.  Whereas, SCC is inherently suited for high intensity machines, such as KEKB, since the aperture of the beam pipe can be larger with enough accelerating gradients, i.e., reduced interaction with a beam, and by virtue of the high gradients, the number of total cavities can be reduced and thus the total impedance can be lowered as well as saving capital costs.  Furthermore, at KEK, we can utilize the refrigerator that was used for TRISTAN, a substantial cost saving.<br /><br />The damped structure for SCC has large beam pipes with absorbing material.  Most of the HOM's go through these pipes and are absorbed.  This scheme, however, that puts microwave absorbing material, such as ferrite, in the same SCC structure and vacuum was a new approach and its feasibility had to be proved.  Therefore, one of the key components for the success of this scheme was the HOM absorber, which is the subject of this thesis.<br /><br />The HOM absorber for SCC requires sufficient HOM damping, power handling capability (>5W/cm2), UHV (Ultra High Vacuum) compatibility (<〜1x10-9 Torr), free of particulates that enter and degrade the cavity, and some surface electric conductivity to prevent charge-up and sparks.<br /><br />Due to the favorable results of preliminary tests on HOM damping, outgassing rate, high availability and short delivery time, we chose to use a microwave-absorbing ferrite IB-004 from TDK, Inc.  However, the most difficult part of the development was to bond this material with high integrity to meet the above-mentioned requirements.  Neither brazing nor soldering of commercial tiles (60x60x4mm3) were successful due to cracking during the cooling down from brazing/soldering temperatures.  Also, we thought of making a cylindrical-shaped ferrite, but experts at TDK commented that making a thin (〜4mm) cylinder of large diameter (220/300mm) ferrite would be very difficult and probably crack.  Forming a thin layer with, for example, plasma spraying, would probably result in a very porous surface and be incompatible with UHV and free-of-particulate conditions.<br /><br />The idea, that we demonstrated its usefulness and high potential for more applications, is sinter-bonding of pre-sintered ferrite powder directly onto the inner surface of a copper beam pipe with HIP (Hot Isostatic Press).  Having developed the proper tooling and optimizing the conditions, we succeeded in establishing a technique to make full-size absorbers that met all the requirements.  SEM (Secondary Electron Microscopy) and EDX (Energy Dispersive X-ray) analyses showed high integrity of the bonding, namely, a 90μm-wide transition region was formed, which consists of a mixture of ferrite and copper with a narrower band of silicon.  This firm bonding appears to have been accomplished by a combination of interlocking and diffusing the copper and ferrite.  In addition, acoustic tomography showed high uniformity of the bonding through the entire area.<br /><br />A 1/3-size (109 mm in ferrite outer diameter) model was first manufactured for various evaluation tests using a TM01 mode of 2.45GHz 5kW power source.  A maximum absorbed power of 3.95kW (average power density of 8.3W/cm2 with the maximum density of 29W/cm2) was reached without any problem.  In addition, at this experiment, the effectiveness of having tapers at the edges to avoid excess heating was confirmed, thereafter 25-mm-long tapers have been machined on both ends.<br /><br />Then, two full-size absorbers, 220/300mm in outer diameter of ferrite, 120/150mm in length and 4mm in thickness, were manufactured.  A milestone experiment, the first beam test, was performed at TRISTAN MR (Main Ring) in 1995.  We installed a full-size (300-mm in diameter) absorber at the Nikko section of the ring with tapers on both sides and tested its performance.  There was no spark, discharge, damage or degradation up to an available maximum single bunch current of 4.4mA or 2.8x1011 electrons per bunch, which is 20 times that of KEKB-HER.<br /><br />The second important test was carried out at TRISTAN AR (Accumulation Ring) in 1996.  A fully equipped superconducting cavity of KEKB shape was assembled with HOM absorbers attached on the upstream and downstream sides, installed in the east tunnel and tested for three periods of time.  SCC could store 0.57A, about half of the design current at KEKB-HER, without any instability due to cavity HOM's.  No degradation of cavity performance was observed and up to a total power of 4.2kW was recorded at the absorbers without any problem, which is about 80% of the power expected at KEKB-HER.  After these tests, micro-cracks (10μm>, invisible with the naked eye) were found in the ferrite, but the following intense investigation showed that these cracks were present before these tests and were caused by baking.  In other words, these tests demonstrated that cracks do not affect the necessary performance of superconducting cavities.<br /><br />Finally, a total of 8 absorbers were assembled with four KEKB-HER superconducting cavity modules and installed in the Nikko D11 tunnel in the summer of 1998.  These modules have shown gap voltages of 2.5-3MV without beams, which is much higher than the design value of 1.5MV, assuring a significant margin for stable operation.  Since the commissioning started in December of 1998, the system has been operating very smoothly.  As of January 31, 2000, the maximum current achieved so far is 0.51A, about half of the designed value.  All HOM absorbers have been showing similar behavior and the maximum power absorved so far is about 2.5kW per module, which is consistent with calculations., 総研大乙第72号},
 title = {Development of Higher-Order-Mode (HOM) Absorbers forKEKB Superconducting Cavities},
 year = {}
}