Vibration-Free Pulse Tube Cryocooler System for Gravitational Wave Detectors II - Cooling Performance and Vibration -

Transcription

1 1 Vibration-Free Pulse Tube Cryocooler System for Gravitational Wave Detectors II - Cooling Performance and Vibration - R. Li A, Y. Ikushima A, T. Koyama A, T. Tomaru B, T. Suzuki B, T. Haruyama B, T. Shintomi B and A. Yamamoto B A Cryogenics Division, Sumitomo Heavy Industries, Ltd Yato, Nishitokyo, Tokyo, , Japan B High Energy Accelerator Research Organization (KEK) 1-1 Oho, Tsukuba, Ibaraki, , Japan ABSTRACT A vibration-free pulse tube cryocooler system has been developed for gravitational wave detectors. A commercially available 4K pulse tube cryocooler (SRP-052A, Sumitomo Heavy Industries, Ltd.) with a cooling capacity of 0.5W at 4.2K was applied in the system. In order to reduce the vibration of the 4K pulse tube cryocooler down to an ultra low level, (1) two vibration reduction stages (VR stages), (2) a cold head supporting frame and (3) a valve unit mounting table were introduced as major components of the system. The cooling capacities of 15 W at 45 K and 0.4 W at 4.2 K were available at the first and the second VR stages simultaneously. Concerning the vibration, on the other hand, the displacement due to the elastic deformation of the pulse tubes was effectively reduced to be less than 1 m. In the direction parallel to the pulse tubes especially, the displacement has been lowered down to 5 m, which was two to three orders of magnitude smaller than that of the original 4 K pulse tube cryocooler, SRP-052A. INTRODUCTION Pulse tube cryocooler has no moving in its low temperature part, and thus has less vibration than Gifford-McMahon (G-M) cycle cryocooler or Stirling cycle cryocooler in general. In the last ten years, many kinds of pulse tube cryocooler have been commercialized by a number of manufactures, and have been already applied as a substitution of liquid nitrogen or liquid helium in various sensitive applications 1-4. Compare to the G-M or Stirling cycle cryocooler, pulse tube cryocooler has a significantly smaller acceleration of the cold head actually. A previous investigation indicated that the overall acceleration of the cold head for 4 K pulse tube cryocooler is smaller than that for 4 K G-M cryocooler about two orders of magnitude 5. Concerning the displacement of cold stage, however, it is comparable between the 4 K pulse tube cryocooler and the 4 KG-M cryocooler. The vibration of pulse tube cryocooler, therefore, is still not low enough for many sensitive applications. In the Cryogenic Laser Interferometer Observatory (CLIO) 6, which is a prototype of the Large-scale Cryogenic Gravitational wave Telescope (LCGT) 7, a number of mirrors and radiation shields are designed to be cooled by pulse tube cryocoolers. Although there are several

2 2 vibration isolation components in the CLIO as a matter of course, a pulse tube cryocooler systems with a vibration level as low as possible is still essential for reaching a desired sensitivity by these state-of-the-art detectors. By using a commercially available 4 K pulse tube cryocooler, SRP-052A, manufactured by Sumitomo heavy industries, Ltd., a vibration-free pulse tube cryocooler system has been developed in this work. A number of key components, such as (1) two vibration reduction stages (VR stages), (2) a cold head supporting frame and (3) a valve unit mounting table, were introduced for reducing the vibration from the 4 K pulse tube cryocooler. In the present paper, the design concept of the system and the major components were described. The cooling performance and the vibration of the vibration-free system were also reported and discussed. DESIGN OF CRYOCOOLER SYSTEM Design Concept As described in the paper presented by Dr. Tomaru, T. in this 13 th International Cryocooler Conference, the CLIO includes a number of mirrors which were designed to be cooled down to 20 K by about 10 units of pulse tube cryocooler 8. Liquid helium was a competitive option for cooling these mirrors without mechanical vibration. The site of the CLIO, however, is located in a remote area, Kamioka mine. For a long interval observation, a regular refill of liquid helium will be very inconvenient and uneconomic, and a cooling system with 4 K pulse tube cryocooler was chosen. There were two major design concepts for developing a pulse tube cryocooler system with an ultra low level vibration in this work. The first concept was to make use of a commercially available 4 K pulse tube cryocooler. To develop a special pulse tube cryocooler that has very low level vibration with a reasonable cooling capacity was a considerable approach. After a feasibility study from the viewpoint of development difficulty level, time and expense, etc., to use a commercially available 4 K pulse tube cryocooler was decided. A previous investigation has pointed out that a commercially available 4 K pulse tube cryocooler has a certain level vibration, especially for the cold stage displacement 5. The investigation has also made it clear that there are two kinds of vibration in a pulse tube cryocooler. One is an overall cold head vibration that is caused by the movement of compressor, rotary valve unit and working gas flow. Another is cold stage vibration which is generated by elastic deformation of the pulse tubes and the regenerator tubes due to the periodic pressure oscillation inside of cryocooler. Based on the investigation, the second design concept was to introduce several vibration reduction components for separating these two kinds vibration and for reducing these vibration effectively. Figure 1 is the schematic drawing of the pulse tube cryocooler system developed for the CLIO, according to the design concepts mentioned above. Major Components Besides a 4 K pulse tube cryocooler, the major components of the vibration-free pulse tube cryocooler system were two vibration reduction stages, a cold head supporting frame and a valve unit mounting table. The followings are the details of these components. 4 K Pulse Tube Cryocooler. The 4 K pulse tube cryocooler employed in this development is manufactured by Sumitomo Heavy Industries Limited. The cryocooler is named as SRP-052A, and its specification of cooling capacity was 0.5 W at 4.2 K and 20 W at 45 K simultaneously with 7 kw power consumption. The cryocooler consists of a cold head of two-stage type, a rotary valve unit driven by a synchronous motor, two flexible hoses of 20 meter long and a helium compressor as same as that for 4 K G-M cryocooler. The vibration of the pulse tube cryocooler has been previously investigated compared with a 4 K G-M cryocooler. In brief, the overall acceleration of the cold head was smaller than that of G-M cryocooler about two orders

3 3 Connecting Tube Upper Flange Bellows FRP pipes Heatlinks Supporting Frame Cryostat Valve Unit Lower Flange 1st Cold Stage 1st VR Stage 2nd Cold Stage 2nd VR Stage Valve Unit Table Figure 1. Schematic drawing of vibration-free pulse tube cryocooler system. The compressor and the flexible hoses are not shown. of magnitude, and the displacement of the cold stage is almost the same as that of G-M cryocooler. An attractive advantage of SRP-052A is that the rotary valve unit is separated from the cold head with a self-sealing connection. The standard setup recommended by the manufacture is to connect the rotary valve unit to the cold head directly without any extension connecting tube for getting maximum cooling performance. From the viewpoint of vibration reduction, this configuration is enable to separate the rotary valve unit away from the cold head for reducing the vibration from the compressor and the valve unit as low as possible. Vibration Reduction Stages. In order to reduce the cold stage vibration effectively, two vibration reduction stages were introduced for the first and the second stages of the cold head respectively (Fig.1). The VR stages are the interface for delivering cooling capacities to the mirrors and radiation shields of the CLIO. The VR stages were made of copper, thermally linked to the original cold stages of the cryocooler with braided wires, and tightly supported by eight pipes connected to the top flange (the lower flange in Fig. 1) of the cryostat. For the first VR stage, the braided wires were made of oxygen free high conductivity copper, and for the second VR stage, they are braided wires of high purity aluminum. The copper wire has a higher thermal conductivity at the temperatures above ~45 K. The high purity aluminum wire has the advantages of light weight, excellent thermal conductivity below 20 K, and very smaller Young s modulus 9. The Young s modulus of the aluminum wire is only about 1/3 of that of copper wire. The pipes for supporting the VR stages, on the other hand, is required that have poor thermal conductivity and larger Young s modulus. In this work, the FRP pipes made of alumina fiber (A-FRP) were applied for both VR stages. The A-FRP is well known that has lower thermal conductivity below 40 K and larger Young s modulus than that of glass fiber reinforced plastic (G-FRP). Cold Head Supporting Frame. The cold head supporting frame is an important component for separating the overall cold head vibration and the clod stage vibration. The cold head was fixed on the upper flange and supported by the supporting frame (Fig.1). The frame was made of H-section steel and has four strong poles. There was a bellows between the upper flange and the lower flange for shaping a vacuum space for the cryocooler. With this configuration, the weight of the cold head was perfectly supported by the upper flange and the supporting frame, and the whole weight of the VR stages was supported by the lower flange and the cryostat. It is expected that the effect of the VR stages will be maximized by this configuration

4 4 Valve Unit Mounting Table. Because the compressor and the driving motor of the rotary valve unit generate a considerable mechanical vibration, an effective vibration isolation between the rotary valve unit and the cold head is important for the present development. A steel table weighted more than 140 kg was used for mounting the rotary valve unit (Fig.1). In addition, the table was also applied as an anchor for the helium gas lines from the compressor. Between the rotary valve unit and the cold head, a connecting tube of 40 cm long has been employed. This configuration is not recommended for getting the maximum cooling performance, but is important and convenient for reducing the vibration from the compressor and the driving motor. COOLING PERFORMANCE As described above, the VR stages and the valve unit mounting table were introduced into the pulse tube cryocooler system for vibration reduction. In the view of cooling performance, these two improvements will certainly impact the cooling capacities available from the system. The extended connecting tube between the valve unit and the cold head is expected to reduce the cooling capacity compared with the original cryocooler. The braided wires of the VR stages will produce a finite temperature difference between the VR stages and the original cold stages. Effect of Valve Unit Separation The effect of the valve unit separation was evaluated by comparing the load maps of the cryocooler with and without the connecting tube of 40 cm long. In figure 2, two load maps show the comparison, but all data was measured without the VR stages. For the original configuration of SRP-052A (i.e. without the connecting tube and joint the rotary valve unit directly to the cold head), the cooling performance was 20 W at 40.8 K for the first stage and 0.5 W at 4.08 K for the second stage simultaneously. After the introduction of the 40 cm long connecting tube, the cooling capacities becomes 20 W at 42.2 K on the first stage and 0.5 W at 4.17 K on the second stage simultaneously. The decrease of the cooling capacities is clear, but is limited. The cooling performance meets the specification of SRP-052A still, although the connecting tube was employed. Load Map on the VR stage In the CLIO, there are several vibration isolations between the 4 K cryocooler system and Second Stage Temperature (K) Standard Setup Valve Unit 40 cm Separated (0W, 0.5W) (20W, 0.5W) (0W, 0W) (20W, 0W) 60 Hz First Stage Temperature (K) Figure 2. Comparison of load maps with and without the connecting tube of 40 cm.

5 5 2nd Stage Temperature [K] Low Vibration Condition : (1st) Cu-wire / (2nd) Al-wire 0.7W 0.5W 0.3W 0W 15W st Stage Temperature [K] 20W Cold Stage VR Stage 0.7W 0.5W 0.3W 0W 15W 20W Figure 3. Load map of the VR stages compared with that of load map of the cold stages. the 20 K mirrors. Because some of the isolation have a poor capability for heat conduction, the cooling capacity available from the VR stages was required as large as possible. Figure 3 shows the load map of the VR stages with solid lines. As a reference, the load map of the cold stages measured at the same time is represented by broken lines in the figure. The typical point with head loads of 15 W at the first stage and 0.5W at the second stage, the temperature of the first VR stage was 43.7 K and that of the second VR stage was 4.43 K. At the same moment, the cold stage temperatures were 41.2 K and 4.15 K respectively. Because the heat leak from room temperature via the A-FRP pipes was estimated only about 0.3 W to the first VR stage and about 10 mw to the second VR stage, the loss of cooling capacity is almost caused by the thermal resistance of the heat link of braided wires. The wires of the heat link were made of high purity metals and were well heat-treated. A high thermal conductivity of the braided wires is expected. The dominant thermal resistance is considered come from the contact surfaces of the heat link, faced to the cold stages or the VR stages. Although the VR stages somewhat affected the cooling performance negatively, the cryocooler system is able to simultaneously deliver net cooling capacities of 15 W at 45 K and 0.4 W at 4.2 K at least. VIBRATION OF VR STAGE The overall vibration of the 4 K pulse tube cryocooler system has been reported by Dr. Tomaru, T. in the paper entitled: Vibration-free pulse tube Cryocooler system for Gravitational Wave Detectors I Reduction of vibrations 8. In present paper, the displacement of the cold stages and the VR stages were focused. Setup of Vibration Measurement Figure 4 illustrates the setup of vibration measurement for the pulse tube cryocooler system. The rotary valve unit was mounted on the valve unit table. The weights of the cold head and the VR stages were separately supported by an upper flange supporting frame and a lower flange supporting frame. The displacement was measured by a laser displacement sensor (LC-2420, Keyence, Co.), which was rigidly held on the lower flange supporting frame. An air damper was employed for the supporting frame of lower flange for insolating the vibration from the floor. The vibration was measured at room temperature, and the pressure oscillation in the cold head

6 6 Y Upper Flange Support X Pressure Sensor Displacement Meter Z Air Dumper Lower Flange Support Personal Computer Spectrum Analyzer Figure 4. Setup of vibration measurement for pulse tube cryocooler system. was monitored acquired at the same time. The data of the displacement and the pressure were acquired by a spectrum analyzer (OR24, Oros, SA) and a personal computer. As shown in the Fig.4, the vertical direction parallel to the pulse tubes was defined as Z-axis, and the direction of 15rotated from the connecting tube of 40 cm was represented as X-axis. Displacement of VR Stage The measurement result of the second VR stage displacement is shown in figure 5, 6 and table 1. For making a comparison, the displacement of the second cold stage measured with the same setup is also illustrated and listed. The displacements of the cold stage were 5.3m, 5.7m and 8.5m for the X, Y and Z directions respectively, but for the VR stage, the displacements for each axis were reduced down to less than 1m. The details of the VR stage displacement were shown in figure 6 by enlarging the vertical axis of Fig. 5. Figure 6 exhibits the second VR stage displacement for the cases of the pulse tube cryocooler turned off Figure 5. Displacement of VR stages and clod stags.

7 7 DISPLACEMENT, m X-axis ON OFF PRESSURE TIME, s PRESSURE, MPa DISPLACEMENT, m Y-axis ON OFF PRESSURE TIME, s PRESSURE, MPa DISPLACEMENT, m Z-axis ON OFF PRESSURE TIME, s Figure 6. Displacement of VR stages for both cases of cryocooler turned on and turned off. PRESSURE, MPa and turned on, and represents the pressure wave inside of the cold head too. From Fig. 6, it is found that there was a higher level noise mixed in the displacement data for both cases of the cryocooler turned on and turned off. Including these noises, the displacements of the second VR stage were 0.42m, 0.65m and 5m for the X, Y and Z directions respectively. Especially for the Z-axis, there was almost no difference between the cases of the cryocooler turned on and turned off in the displacement. It is means that the vibration of the VR stage in this direction was successfully reduced down to background vibration level. In this sense, it is can say that the cryocooler system has achieved a vibration-free level at least in the Z direction. Figure 7 shows the spectrum of the displacement for the cold stage and the VR stage. For the second cold stage, the peaks corresponding to the operating frequency (1.2Hz) and its higher X-axis Y-axis Z-axis Displacement [m] 1.E-03 1.E-04 1.E-05 1.E-06 1.E-07 1.E-08 1.E-09 1.E-10 1.E-11 1.E Frequency [Hz] Displacement [m] 1.E-03 1.E-04 1.E-05 1.E-06 1.E-07 1.E-08 1.E-09 1.E-10 1.E-11 1.E Frequency [Hz] Displacement [m] Figure 7. Spectrum of second VR satge displacement. 1.E-03 1.E-04 1.E-05 1.E-06 1.E-07 1.E-08 1.E-09 1.E-10 1.E-11 1.E Frequency [Hz]

8 8 harmonics were sharp and high. For the second VR stage, the peak for at the operating frequency was still confirmed, but it was remarkable lower than that of the cold stage. Indeed, the peak value was 6.4m for cold stage at 1.2Hz, but became only 10 nm for the VR stage. For the X and Y directions, the peak value at 1.2 Hz was reduced from 2.6 m to 0.12m and 4.6m to 0.31m respectively, Both the time series data and the spectrum data indicate that the vibration was reduced by one to two orders of magnitude in the X and Y directions and by two to three orders of magnitude in the Z direction. The effect of the VR stage was somewhat small for the X and Y directions. The reason is considered that was caused by the coupling of the pulse tube and regenerator tube, but the detail is necessary to investigate in the future. Concerning the vibration of the first cold stage and the first VR stage, it was also evaluated in this work. Since the results for the first stages were similar to those for the second stages, they were skipped in this paper. The displacement of the first stages, however, was much smaller because the length of the first pulse tube and the first regenerator were shorter. CONCLUSION A vibration-free pulse tube cryocooler system has been developed for gravitational wave detectors. In the cryocooler system, a commercially available 4 K pulse tube cryocooler has been employed. For an effective vibration reduction, (1) vibration reduction stages, (2) a cold head supporting frame and (3) a valve unit mounting table ware introduced as major components in the cryocooler system. Although the VR stages and the valve unit mounting table affect the cooling performance negatively, the cryocooler system was able to deliver net cooling capacities of 15 W at 45 K and 0.4 W at 4.2 K simultaneously, and gives a large cooling margin for the mirrors and vibration isolation components of the CLIO. The major components introduced by this work remarkably reduced the vibration of the original 4 K pulse tube cryocooler. For X-axis and Y-axis, the displacement was reduced by one to two orders of magnitude, and was down to less than 1m. For the Z-axis, especially, the displacement was lowered to be less than 5 m and was two to three orders of magnitude smaller than that of the original cold head. The vibration was as low as expected for the gravitational wave detectors in the CLIO. REFERENCES 1. Li, R., Ishikawa, A., Koyama, T., Ogura, T., Aoki, A. and Koizumi, T., A Compact Liquid Nitrogen Recondenser with a Pulse Tube Cryocooler, Advances in Cryogenic Engineering, vol. 45, Kluwer Academic/Plenum Publishers, New York (2000), pp.?-?. 2. Xu, M.Y., Yan, P.D., Koyama, T., Ogura, T. and Li, R., Development of a 4 K Two-stage Pulse Tube Cryocooler, Cryocooler 12, Kluwer Academic/Plenum Publishers, New York (2003), pp Wang, C. and P.E. Gifford, Performance Characteristics of a 4 K Pulse Tube in Current Applications, Cryocooler 11, Kluwer Academic/Plenum Publishers, New York (2001), pp C. Lienerth, G. Thummers and C. Heiden, Low-Noise Cooling of HT-SQUIDs by Means of a Pulse Tube Cooler with Additional Vibration Compensation, Proceedings of the 18th International Cryogenic Engineering Conference, Narosa Publishing House, New Delhi (2000), pp Tomaru, T., Suzuki, T., Haruyama, T., Shintomi, T., Yamamoto, A., Koyama, T., Li, R. and Matsubara, Y., Vibration Analysis of Cryocoolers, Jaurnal of the Cryogenic Society of Japan, Vol. 38, No. 12 (2003), pp Ohashi, M. et al., Design and Construction Status of CLIO, Class. Quantum Grav., Vol. 20, No. 17 (2003) pp. S599-S Kuroda, K. et al., Large-Scale Cryogenic Gravitational Wave Telescope, Int. J. Mod. Phys. D, Vol. 8 (1999) p.557.

9 9 8. Tomaru, T., Suzuki, T., Haruyama, T., Shintomi, T., Sato, N., Yamamoto, A., Ikushima, Y., Li, R., Akutsu, T., Uchiyama, T. and Miyoki, S., Vibration-free pulse tube Cryocooler system for Gravitational Wave Detectors I Reduction of vibrations, to be published in Cryocooler 13, Kluwer Academic/Plenum Publishers, New York (2005). 9. Kasahara, K., Tomaru, T., Uchiyama, T., Suzuki, T., Yamamoto, K., Miyoki, S., Ohashi, M., Kuroda, K. and Shintomi, T., Study of the Heat Links for a Cryogenic Laser Interferometric Gravitational Wave Detector, Jaurnal of the Cryogenic Society of Japan, Vol. 39, No. 1 (2004), pp