Study of highly transparent silica aerogel as a RICH radiator

Transcription

1 Nuclear Instruments and Methods in Physics Research A 553 (2005) Study of highly transparent silica aerogel as a RICH radiator I. Adachi a,, S. Fratina b, T. Fukushima c, A. Gorisˇ ek b, T. Iijima d, H. Kawai c, M. Konishi c, S. Korpar b,e, Y. Kozakai d, P. Krizˇ an b,f, T. Matsumoto g, Y. Mazuka d, S. Nishida a, S. Ogawa h, S. Ohtake h, R. Pestotnik b, S. Saitoh a, T. Seki g, T. Sumiyoshi g, M. Tabata c, Y. Uchida a, Y. Unno a, S. Yamamoto g a IPNS, High Energy Accelerator Research Organization (KEK), Japan b J. Stefan Institute, Ljubljana, Slovenia c Chiba University, Chiba, Japan d Nagoya University, Nagoya, Japan e Faculty of Chemistry and Chemical Engineering, University of Maribor, Slovenia f Faculty of Mathematics and Physics, University of Ljubljana, Slovenia g Tokyo Metropolitan University, Tokyo, Japan h Toho University, Funabashi, Japan Abstract We have developed silica aerogel to be employed as a RICH radiator in the Belle PID upgrade. For this purpose, new aerogel sample with a higher refractive index (n) of 1:05 was successfully produced while keeping the hydrophobic feature. In this novel technique, a new solvent was introduced in the first step of the sol gel process so as to avoid a deterioration of the optical quality, even for n greater than Based on this procedure, the transmission length was obtained to be more than 40 mm at 400 nm wave length, which was twice longer than before, for samples of mm 3 size. In this article, the results of the new aerogel production will be shown together with some design considerations for the RICH radiator. r 2005 Elsevier B.V. All rights reserved. PACS: Ka Keywords: Aerogel; Particle identification; Belle 1. Introduction Corresponding author. Tel.: address: ichiro.adachi@kek.jp (I. Adachi). The successful production of silica aerogel with excellent optical quality has triggered a new application to a Cherenkov radiator in a RICH /$ - see front matter r 2005 Elsevier B.V. All rights reserved. doi: /j.nima

2 I. Adachi et al. / Nuclear Instruments and Methods in Physics Research A 553 (2005) detector. In the course of the Belle spectrometer upgrade programme [1], we have been studying an aerogel based RICH detector as a forward particle identifier [2]. To obtain a reasonable separation of Cherenkov ring images for p and K particles in the momentum region between 1:04:0 GeV=c, the refractive index of the Cherenkov radiator should be around 1.05, which is larger than what we have used in the Belle Aerogel Cherenkov Counter (ACC). In addition, recent studies indicate that multiple radiators consisting of different refractive indices, in which the indices are selected in such a way that the Cherenkov image generated from each layer can be overlapped as a single ring onto a photodetector plane, allows more Cherenkov light to be detected without loosing a single photon angle resolution [3]. In this multipleradiator system, the downstream aerogel must be sufficiently transparent, for most of Cherenkov photons emitted from upstream radiators to pass through it. In a practical accommodation of aerogel radiators for the 4p collider experiment, how to cover a large radiator area as well as how to machine them to be suitable for the radiator shape are significantly important issues. For instance, the flatness of an aerogel sample should be kept at a reasonable level in the case that one can stack two or more aerogel blocks in a real detector. In this paper, a brief sketch of our aerogel production, including our results from the synthesisation of a chemical solvent, is given in the next section. The measurements of the optical properties of aerogels and the crack conditions are also mentioned here. Then, a test of the hydrophobic nature follows in Section 3. In Section 4, a new attempt to produce a monolithic aerogel with multiple layers is explained. After a description of radiator design considerations in Section 5, Section 6 concludes this paper. 2. Aerogel production 2.1. New solvent and transparency In the previous production method utilised for ACC construction [4], aerogel transparencies Average transmission length [mm] H2O 24mol H2O 28mol H2O 32mol DMF ratio [%] Fig. 1. Average transmission length at 400 nm wave length as a function of the DMF mixing ratio for n ¼ 1:05 samples. The quantities of water were also varied. were optimised for lower refractive indices of , and the optical properties were quickly degraded when the refractive index became bigger than To improve this situation, a new solvent of di-methyl-formamide (DMF) was for the first time introduced in the colloidal formation, where alcogel was made. DMF was mixed with methanol, and the sol gel step proceeded in this alcohol mixture. 1 The transparency of the aerogel can be expressed as T ¼ T 0 expð d=lðlþþ, where T 0 and T are the initial and measured intensities of the light with a wave length of l, d is the thickness of the sample, and L is the transmission length, which depends on the wave length. The transmission length of the aerogel block obtained from this method was measured to be about 40 mm at 400 nm wave length, while 23 mm was achieved if only methanol was used as an alcohol solvent. It is considered that this remarkable improvement is based upon the fact that DMF liquid can help to form smaller pores as the alcogel structure from measuring the aerogel internal porosity. We varied some quantities in the liquid components when making alcogel samples, which may have affected the resulting optical characteristics to investigate the transmission dependence on the chemical quantities used in the sol gel process. Fig. 1 indicates the obtained transmission length of the aerogels as a function of the DMF mixing fractions for n ¼ 1:05 samples, where 100% means 1 Matsushita Electric Works Ltd., Japanese patent No

3 148 I. Adachi et al. / Nuclear Instruments and Methods in Physics Research A 553 (2005) only DMF solvent was used. As can be seen in Fig. 1, a transmission length up to 40 mm could be achieved for some samples when more than 70% of DMF liquid was mixed with 28 mol of distilled water. Details of the mixing ratios for the chemical quantities can be found in [5]. Note that at some points the error quoted is large. For a precise verification, we need to increase the number of test samples to be produced. This improvement of higher transmission was validated by a series of test-beam experiments that we carried out. The obtained Cherenkov light yield was increased by about 50% by providing new aerogels. A detailed description can be found in [2]. The aerogel transparency was studied using small samples, approximately mm 3. Larger size aerogel blocks suitable for a realistic design were produced under the condition of 85% DMF and 28 mol distilled water. Fig. 2 shows a picture of the sample that we produced, where the tile size was mm 3. No cracks were found in this block. The transmission length at 400 nm was measured for all samples of n ¼ 1:05. The results are shown in Fig. 3. For 11 out of total 32 tiles, the transmission length exceeded over Fig. 3. Transmission lengths (mm) at 400 nm wave length for samples of n ¼ 1:05. Table 1 Summary of the refractive index measurements Target value Measurement result mm. However, they were widely distributed in the mm range. This could indicate an unstable procedure in the sol gel step, which should be investigated further Refractive index Fig. 2. Photograph of the obtained aerogel ð mm 3 Þ. The refractive index for aerogel obtained after supercritical drying should be controlled as precisely as possible. The measurement of the refractive index was done using the Fraunhofer method with a 405 nm laser, where an aerogel block was set on a rotating table. The resultant values are summarised in Table 1. The obtained values are mostly consistent with our target within 2s errors.

4 I. Adachi et al. / Nuclear Instruments and Methods in Physics Research A 553 (2005) Block thickness and crack To produce crack-free aerogel tiles is also of our concern, since clean aerogels without any cracks can relax the realistic constraints in arranging aerogel RICH radiators located in the small forward space in the Belle spectrometer. We varied the block thickness from 10 to 20 mm with keeping the cross section to be mm 2 ðn ¼ 1:050Þ, and carried out supercritical drying for these samples. We then examined the appearance of any crack by a visual scan, and assigned crack if any small crack was found in the sample. From this test, we found no crack in all of the samples for 10 mm thickness. On the other hand, 87.5% and 70% of the samples were obtained to be crackfree for 15 and 20 mm thickness, respectively. Even for those samples having cracks, they seemed to be tolerable, and did not have a big impact on the practical employment of aerogel radiators. The yield of obtaining the crack-free samples may be improved by adjusting the conditions at the supercritical drying stage. 3. Hydrophobic feature The possession of hydrophobic characteristics is one of the biggest benefits of our sample [4]. This was tested for new aerogel to verify that DMF liquid does not worsen the hydrophobic nature. We simply put an aerogel tile in excess of distilled water, and then compared the transmission lengths before and after this procedure. The transmission length at 400 nm was 25.5 mm before immersing the sample in the fluid; this value was measured to be 28.2 mm later, resulting in no deterioration. Therefore, new aerogel synthesised with new solvent of DMF still maintained its hydrophobic nature as before. where two or more aerogel layers with different refractive indices are arranged. If two aerogel tiles having independent indices can form into a single block, handling multiple radiators can be rather simplified. Moreover, the possible surface effect could be reduced. For these reasons, we produced monolithic tiles with two or more refractive indices stacked up. The present production of two-layer monolithic blocks is made in the following way: we first produce an alcogel for the bottom layer, which normally has a larger index than the other. Then, after the sol gel reaction becomes stable in a couple of minutes chemical liquid for the upper layer is flushed on the bottom alcogel. The subsequent procedures, such as ageing and hydrophobic treatments, are done in the same way as the conventional ones. It is considered that silicadioxide molecules are bonded with each other, during the ageing period, around boundaries between two layers. This makes a single block without adding any adhesive. These operations have not yet been optimised and will be investigated in the future. Fig. 4 is one of our samples, where it consists of two layers of n ¼ 1:045 and with 7 mm thickness each. The transmission for the sample for n ¼ 1:050 and layers is shown in Fig. 5. Note that these values are combined ones with two layers. We obtained a transmission length of 41.1 mm at 400 nm wave 4. Monolithic aerogel with multiple layers A monolithic aerogel block consisting of multiple layers with different refractive indices is quite attractive, especially in the case that one makes a design of the consecutive multiple radiator scheme, Fig. 4. Monolithic aerogel with two layers, where the refractive index for each layer is shown.

5 150 I. Adachi et al. / Nuclear Instruments and Methods in Physics Research A 553 (2005) Transmission(%) n=1.050& wave length(nm) Fig. 5. Transmission (%) as a function of the wave length (nm) for monolithic aerogel consisting of two layers of and length. We have already tried to produce a 3-layer sample. A detailed evaluation will be made. 5. Mechanical considerations In our basic design for the radiator tiling scheme, a hexagonal shape of aerogels has been proposed to reduce light loss at the edge of the tile based on a naı ve calculation. To make our sample hexagonal, we examined a possibility to produce hexagons from square samples with a size of mm 3 using a water-jet machining device; 7 hexagon-shape aerogels were machined and aligned to check how accurately the angles of each corner and the side length were made (Fig.6). The precision was so excellent that no gap between aerogel boundaries could be found, and they could cover a large radiator area without making extra holes at the boundaries. Usually one may find meniscus at the edge of aerogels, which could cause problems in the case that two or more aerogel layers are arranged. By cutting the edges from the square samples, this problem could be simultaneously solved, and the flatness of the radiators could be affirmed. Note that this Fig. 6. Hexagon-shape aerogels ðn ¼ 1:050Þ, where they are laid to form a radiator layer. machining procedure was possible only for hydrophobic aerogels since the water-jet device was used. The way to hold the aerogel radiator in the RICH counter should be considered to realise a detector design. One of the most straightforward methods is to directly adhere aerogels to a supporting wall made of aluminium with an appropriate amount of glue. The issue is to make sure that the optical qualities of aerogels can remain stable after this adhesion procedure, because some kinds of glue can deteriorate aerogel transparencies during long-term operation. We tested various kinds of glues and found that DP did not affect the aerogel quality from observations made for more than 90 days. 6. Conclusion New aerogels with a refractive index larger than 1.03 were synthesised by introducing DMF solvent. The block size produced was mm 3, and the transparency was measured to be 40 mm at 400 nm wave length. Their hydrophobic characteristics were confirmed. According to our index measurements, further adjustments should be made. The crack conditions for each sample were examined and 100% yield was found 2 Sumitomo 3M Limited.

6 I. Adachi et al. / Nuclear Instruments and Methods in Physics Research A 553 (2005) to be crack-free for 1 cm thickness. Two basic studies to envisage a realistic radiator design were performed. One was to machine aerogel blocks to be an applicable shape as a piece of the radiator layer. Hexagonal-shaped aerogel tile was machined from a square sample, and the machining accuracy turned out to be sufficiently precise. The other was selecting the adhesive to fix the aerogels onto a metal wall, where it does not affect optical quality. Based on these studies, aerogel can be used for RICH radiator. References [1] A. Abashian, et al., Belle Collaboration, Nucl. Instr. and Meth. A 479 (2002) 117 K. Abe, et al., SuperKEKB Letter of Intent, KEK Report [2] P. Krizan, et al., Nucl. Instr. and Meth. A., these proceedings. [3] S. Korpar, et al., Nucl. Instr. and Meth. A., these proceedings. [4] I. Adachi, et al., Nucl. Instr. and Meth. A 355 (1995) 390. [5] M. Konishi, et al., in: J.A. Seibert (Ed.), IEEE Nuclear Science Symposium Conference Record, 2004.