Hydrophobic silica aerogel production at KEK
Makoto Tabata, Ichiro Adachi, Hideyuki Kawai, Takayuki Sumiyoshi, Hiroshi Yokogawa
aa r X i v : . [ phy s i c s . i n s - d e t ] D ec Hydrophobic silica aerogel production at KEK
Makoto Tabata a,b, ∗ , Ichiro Adachi c , Hideyuki Kawai b , Takayuki Sumiyoshi d , Hiroshi Yokogawa e a Institute of Space and Astronautical Science (ISAS), Japan Aerospace Exploration Agency (JAXA), Sagamihara, Japan b Department of Physics, Chiba University, Chiba, Japan c Institute of Particle and Nuclear Studies (IPNS), High Energy Accelerator Research Organization (KEK), Tsukuba, Japan d Department of Physics, Tokyo Metropolitan University, Hachioji, Japan e Advanced Materials Development Department, Panasonic Electric Works Co.,Ltd., Kadoma, Japan
Abstract
We present herein a characterization of a standard method used at the High Energy Accelerator Research Organization(KEK) to produce hydrophobic silica aerogels and expand this method to obtain a wide range of refractive index ( n = 1.006 − Keywords:
Silica aerogel, Refractive index, Cherenkov radiator, Dust collector, SAXS
1. Introduction
Silica aerogel is a useful solid material used in vari-ous scientific instruments because of its properties such astransparency, low bulk density, and unique refractive in-dex. The mechanical and optical characteristics originatefrom an amorphous three-dimensional structure of silicaparticles and pores. The bulk density of silica aerogel isdetermined by the ratio of silica and pore air and can befreely adjusted in the production process. The porosityreaches up to 99.7% or more. The refractive index n ofsilica aerogel is proportional to its density ρ : n − kρ ,where k is a constant.As reported by a large number of articles in this jour-nal, one of the most successful applications of silica aero-gel is as a Cherenkov radiator. The first use of aerogelsas a Cherenkov medium was reported in 1974 [1]. Atthat time, it was mentioned that the lowest refractive in-dex available for aerogels was n = 1 .
01 and the highestwas n = 1 .
20, which was attained by heating. In addi-tion, ultralow-density aerogels were fabricated at the JetPropulsion Laboratory, California Institute of Technologyfor collecting comet dust, and the Stardust spacecraft byNASA successfully retrieved dust samples from the cometWild 2 in 2006 [2]. Recently, aerogels as dust capturershave also been applied in fusion plasmas [3]. Furthermore,silica aerogel is considered to be a muonium-emitting ma-terial and is used as a source of ultracold muon beams [4].The need for silica aerogel is thus increasing.The history of aerogel development in Japan dates backto the 1980s [5]. At present, the study of aerogel pro- ∗ Corresponding author.
Email address: [email protected] (MakotoTabata) duction in Japan is carried forward by a collaboration in-volving Japan’s High Energy Accelerator Research Orga-nization (KEK), Chiba University, and Panasonic (Mat-sushita) Electric Works. As described in this article, wecan manufacture silica aerogel with n = 1.006 − n = 1 .
03 were usedfor an electron-positron veto-threshold Cherenkov counterin the Laser Electron Photon Experiment at SPring-8(LEPS) [6] and for the detection of cosmic-ray antiprotonsin the Balloon-borne Experiment with SuperconductingSpectrometer (BESS) [7]. Also, Matsushita n = 1 .
03 aero-gel was employed for a ring imaging Cherenkov (RICH)counter to separate pions and kaons in the HERMES ex-periment at DESY [8]. Moreover, our low-density aerogelswere used as the space debris capture medium in a series ofMicro-Particles Capturer (MPAC) experiments by JAXA(NASDA) on the International Space Station (ISS) [9].As already reported in our previous articles [10, 11, 12],we are using new production methods to strongly pushforward the development of aerogels with ultrahigh andultralow densities. The novel aerogels will be used in BelleII, which is a super B -factory experiment at KEK [13],and for the Tanpopo, which is an astrobiological missionplanned for the Japanese Experiment Module (JEM) inthe ISS [14], in addition to the next generation of nuclearexperiments at J-PARC. The new production methods willfurther widen the opportunities to employ aerogels in sci-entific instruments. However, our conventional productionmethod remains a key technique to produce high-quality Preprint submitted to Nuclear Instruments and Methods A August 21, 2018 erogels for various applications, and provides the basisfor new production methods. Here we will review the con-ventional production method and its capabilities.In this article, we summarize the conventional and stan-dard technique for producing hydrophobic silica aerogel atKEK and clarify the key features of this technique. Themethods we use to make basic optical measurements arealso described. The optical properties (i.e., refractive in-dex and transmittance) are given as important parame-ters characterizing aerogels. Finally, we discuss the rela-tionship between the optical properties and the fine struc-ture of aerogels based on bulk density measurements anda small-angle X-ray scattering (SAXS) technique.All aerogel samples described in this article were manu-factured by one of the authors (M.T.) at Chiba Universityand examined at KEK between 2004 and 2011. Typicaldimensions of aerogel tiles were 10 × × andthe smallest sample from trial production had dimensionsof 4 × × . Our production method allows us toproduce large monolithic tiles. For example, 26 × × aerogel tiles with n ∼ n < . . < n < .
2. Silica-aerogel production method
Our present method (the third method, following thesingle- and two-step methods) of producing silica aero-gel, developed in the 1990s, is based on the KEK method[17]. To construct threshold aerogel Cherenkov coun-ters (ACCs) [18] for the Belle detector [19], hydrophobicaerogel with n = 1.01 − of aerogel was installedin the ACC in 1998 and, until the end of the Belle ex-periment in 2010, the ACC system played an importantrole in separating pions from kaons. In 2004, we mod-ernized the KEK method by introducing the solvent N , N -dimethylformamide [DMF, HCON ( CH ) ] in the wet-gel(solvogel) synthesis process [15]. With the modernizedKEK method, we can produce highly transparent aero-gels with n ∼ .
05 [20], and we have since been workingto optimize chemical-preparation recipes. The modernizedKEK method has four important features: • Simple solvogel synthesis by methyl silicate 51; • Selection of solvents for solvogel synthesis accordingto the desired refractive index; • Hydrophobic treatment by hexamethyldisilazane;
Figure 1: Water drop on hydrophobic (water-resistant) aerogel (ID= LGG3-4b, n = 1 . • Supercritical drying by carbon dioxide.The fundamental process of solvogel synthesis is basedon the following two successive reactions: Si ( OCH ) + 4 H O → Si ( OH ) + 4 CH OH, m Si ( OH ) → ( SiO ) m + 2 m H O. Tetramethoxysilane [ Si ( OCH ) ] can be simultaneouslyhydrolyzed, condensed, and polymerized in appropriatesolvents to create silica solvogel with the help of a basiccatalyst (ammonia aqueous solution). Instead of tetram-ethoxysilane, we use the commercially available ”methylsilicate 51” (MS51) to simplify the solvogel synthesis pro-cess. MS51 is prepared by polymerizing tetramethoxysi-lane into an oligomer (average degree of polymerization is4), which has a high silica content (51% by weight).The selection of solvents, ethanol, methanol, and DMFfor the solvogel synthesis affects the optical performance ofaerogels. The relative quantity of solvents to total solvo-gel volume is the principle factor determining the aerogelrefractive index. For high transparency aerogels, appro-priate solvents should be selected according to the desiredrefractive index. As expected from the chemical equationabove, the most basic solvent is methanol, which is used ina wide range of n ≥ .
02. However, at n ∼ .
02, shrinkageof aerogels synthesized with methanol increases and theirtransparency decreases. This problem was solved by us-ing ethanol (99.5) for n < .
02 [17]. On the other hand,transparency in the high-refractive-index range was im-proved by introducing DMF, which was used mixed withmethanol for n < .
06 and alone for n ≥ .
06 [15]. So far,DMF has been introduced in the range of n ≥ . − OH ) on the surface of SiO par-ticles are likely to be charged and can easily react withother ions. Absorption of moisture into hydrophilic aero-gels is a particularly serious problem. Hydroxyl groups are2 able 1: Starting materials used in aerogel production process. Materials Manufacturer (Distributor)Methyl silicate 51 Fuso Chemical Co., Ltd.Distilled water Wako Pure Chemical Industries, Ltd.Ethanol (99.5)* Wako Pure Chemical Industries, Ltd.Methanol Wako Pure Chemical Industries, Ltd. N , N -dimethylformamide Wako Pure Chemical Industries, Ltd.28% Ammonia solution Wako Pure Chemical Industries, Ltd.Silazane, Z-6079 Dow Corning Toray Co., Ltd.Ethanol (99)** Japan Alcohol Trading Co., Ltd.Liquefied carbon dioxide Showa Tansan Co., Ltd.* Only for solvogel synthesis. ** For other processes (e.g., rinse).thus replaced with trimethylsiloxy groups [ − OSi ( CH ) ]by adding the hydrophobic reagent hexamethyldisilazane[(( CH Si ) N H ] [21]:2( − OH ) + (( CH ) Si ) N H → − OSi ( CH ) ) + N H . Water-resistant aerogels are obtained after supercriticaldrying (SCD) (see Fig. 1).Because the fine structure of silica networks is easily de-stroyed if solvogels are dried in air, they should be dried bythe SCD method. When the solvent ethanol in solvogels isextracted from silica networks by natural evaporation, thepath in the pressure-temperature phase diagram of ethanolintersects the boiling line. Because it is accompanied bysignificant ethanol-volume changes, the fine silica networksare lost. To avoid this, we change the solvent phase fromliquid to gas by going around the critical point in the SCDmethod. The simplest way to extract ethanol is to goaround the critical point of ethanol. The critical pressureand temperature of ethanol are 6.4 MPa and 243.1 ◦ C, re-spectively. Although this approach is certainly availablefor at least n < .
02, higher-refractive-index solvogels syn-thesized with DMF are broken in the ethanol SCD method.For this reason, solvogels in any refractive-index range areusually dried by the carbon dioxide SCD method. In thismethod, by using an autoclave, solvogels go around thecritical point of carbon dioxide after the ethanol in thesolvogel is replaced by liquefied carbon dioxide under highpressure. The critical pressure and temperature of carbondioxide are 7.4 MPa and 31.1 ◦ C, respectively. Using car-bon dioxide is safe because it is nonflammable and has alow critical temperature.
We begin by preparing two solutions A and B. SolutionA is made by adding MS51 to the solvent and solution B ismade by adding 28% ammonia solution to distilled water.The starting materials for the aerogel production processare listed in Table 1. Solutions A and B are quickly mixedin a polyethylene beaker at room temperature and stirredwell for 30 s. The mixed solution is carefully poured into apolystyrene mold and immediately covered with a lid. Toobtain more transparent aerogels, the amount of ammonia solution should be adjusted to complete the gel forma-tion reaction in several minutes. Furthermore, dependingon the solvent used, the solvogel is covered with a smallamount of ethanol or methanol, and the mold is coveredagain with the lid to avoid evaporation of the solvent inthe solvogel.To strengthen the three-dimensional
SiO network, thesynthesized solvogel in the mold is placed in an air-tightvessel and aged at room temperature for 1 week. In thisaging period, the solvogel slightly shrinks and separatesfrom the mold; thus, forming a solvogel tile. After 1 week, the lid of the solvogel mold is removed andthe air-tight vessel is filled with ethanol (99) so that thesolvogel is immersed in ethanol. Next, the vessel is sealedagain and the solvogel is left for 3 days to be impregnatedwith ethanol.For the hydrophobic treatment, the solvogel is de-tached from the mold and transferred into a stainless steelpunched tray in the same ethanol, and then the hydropho-bic reagent is poured into the ethanol. The volume ratioof hydrophobic reagent to ethanol is 1:9. The solvogel iskept in the solution at room temperature for 3 days.Ammonia generated by the hydrophobic reaction is ex-tracted by significant quantities of ethanol, and the solvo-gel is then transferred into fresh ethanol. The ethanol isreplaced by new ethanol every 2–3 days for a total of twotimes.
At this stage of the aerogel production, ethanol stillremains in the voids between the silica networks of thesolvogel. Ethanol can be extracted without cracking thesolvogel by using the carbon dioxide SCD method. Super-critical extraction equipment (autoclave) was constructedfor this purpose (see Fig. 2).After the solvogel is positioned in the autoclave, the au-toclave is filled with fresh ethanol and sealed. Fig. 3 showsa history of the internal pressure and temperature in theautoclave. SCD begins with the free injection of lique-fied carbon dioxide from a cylinder at room temperature.3 igure 2: Homebuilt supercritical extraction equipment ( CO auto-clave) installed at Chiba University. The autoclave with a heater ismounted on an electronic balance. A pump is used to send lique-fied carbon dioxide from a cylinder to the autoclave. Based on thetemperature measured with a type-K thermocouple inserted into theautoclave, the output of the heater is adjusted by a controller. Theautoclave has a capacity of 7.6 liters. Up to 10 solvogel tiles with di-mensions of 11 × × can be dried simultaneously. This sizeis sufficient for test production; for mass production of large aerogeltiles, we can use the large autoclave (140 liters) at the Mohri Oil MillCo., Ltd. The injection stops when the autoclave pressure is equalto that of the cylinder, after which the autoclave is keptat the cylinder pressure ( ∼ ∼
12 h).During this time, the liquefied carbon dioxide mixes wellwith the ethanol and penetrates the solvogel.To increase the pressure as high as 8.0 MPa above thecritical pressure of carbon dioxide (7.4 MPa), liquefied car-bon dioxide is pumped into the autoclave. When the pumpis stopped, the autoclave is heated at 10 ◦ C/h to 40 ◦ Cabove the critical temperature of carbon dioxide (31.1 ◦ C).Note that the solvogel is below the subcritical conditionbecause the autoclave is still rich in ethanol, whose criticaltemperature is 243.1 ◦ C. The internal pressure in the au-toclave increases with increasing temperature. When thepressure reaches 10.0 MPa, the extraction of the ethanoland carbon dioxide fluid mixture begins by opening an P r e ss u r e [ M P a ] T e m p e r a t u r e [ ℃ ] PressureTemperature
Figure 3: History of internal pressure and temperature in the auto-clave (run 143). The open and solid squares represent the changesin pressure (left axis) and temperature (right axis), respectively. output valve to keep the pressure. Initially, almost all ofthe extracted fluid is ethanol.When the temperature reaches 40 ◦ C, injection by pump-ing liquefied carbon dioxide starts again. Extraction of themixed fluid continues and the pressure is maintained in therange of 8.0 to 11.0 MPa. It takes 30 h from the beginningof the extraction before the extracted fluid is sufficientlyrich in carbon dioxide. To extract all the ethanol fromthe solvogel, the temperature is raised to 80 ◦ C at 10 ◦ C/h.The conditions of 8.0 − ◦ C are maintainedtypically for at least 10 h. The concentration of ethanolvapor in the extracted fluid determines when the injectionof liquefied carbon dioxide is finished. When an ethanol-inspection tube shows a concentration of 200 ppm or less,the pump is stopped.The autoclave operation for the pressure- andtemperature-reduction process is crucial for obtainingcrack-free aerogels. Light, crack-free aerogels with n ≤ .
03 are relatively easy to make. However, a slow pressure-reduction rate is important for thick crack-free aerogelswith n ≥ .
04. The pressure is first reduced below thecritical pressure (7.4 MPa) at high temperature (80 ◦ C).The rate of pressure reduction is 0.5 − aser AerogelRotating table Screen L d m δ m α Figure 4: Setup of refractive index measurement. The notation isexplained in the text. The table can be manually rotated, and a gridsheet is fixed to the screen.
3. Optical measurements
For all aerogel tiles manufactured, the important opticalparameters (i.e., refractive index and transmittance) weremeasured by hand, tile by tile, in the visible range. Ow-ing to the recent progress in the RICH detectors by usingan aerogel radiator, the need to focus on optical measure-ments is growing so that the Cherenkov angle from un-scattered Cherenkov photons emitted by aerogels can bereconstructed.
Manufactured aerogels can be tagged by their refractiveindex at a wavelength ( λ ) of 405 nm, which can be mea-sured by the laser Fraunhofer method with a blue-violetsemiconductor laser. The measurement setup is shown inFig. 4. The Fraunhofer method gives the refractive in-dex of aerogels relative to air and is based on the prismformula: nn air = sin (cid:18) α + δ m (cid:19) (cid:20) α/ (cid:21) ,δ m = tan − ( d m /L ) , where n air ∼ . α = π/ δ m is theminimum angle of deviation, d m is the minimum displace-ment of the laser spot, and L is the distance between theaerogel and the screen. The distance L was approximately1.8 m in our case. The averaged refractive index was de-termined by d m measured at four corners of each aerogeltile, where the laser penetration depth was 5 mm from thevertex of the aerogel. The laser-spot spread on the screendepends on the optical properties, including the aerogel re-fractive index. After passing through the aerogels, typicallaser-spot diameters were 1, 6, and 16 mm (at equivalentresolution, ∆ n = 0.0001, 0.0008, and 0.002) for n ∼ Distance [cm]0 5 10 15 20 T r a n s m itt a n ce [ % ] Figure 5: Transmittance at 400 nm as a function of distance betweenthe downstream surface of the aerogel (ID = PD156b, n = 1 . t = 11 . The transmittance through 10 to 20 mm of aerogel wasmeasured using the generalized spectrophotometer HitachiU-4100 (or U-3210 prior to 2008). This system allowed usto measure transmittance from 200 to 800 nm. The mea-surement chamber consisted of a continuous light source,sample stage, and light-integrating sphere. The transmit-ted light was collected by the integrating sphere and intro-duced into a photomultiplier tube (PMT). The diameterof the entrance to the integrating sphere was 20 mm and,without aerogels, the spot of the light source spread to ap-proximately 10 × at the entrance of the integratingsphere. Because the entrance of the integrating sphere waslarger than the spot of the light source, a portion of thelight scattered by the aerogel could enter the sphere. As aresult, the transmittance for a given wavelength dependson the distance between the aerogel tile and the integrat-ing sphere, as shown in Fig. 5. To gather as effectivelyas possible only the unscattered light, we decided to placethe downstream surface of the aerogel tile 10 cm in frontof the entrance to the integrating sphere. This distancewas measured for several samples and found to be repro-ducible.Fig. 6 shows a resulting transmittance curve as a func-tion of wavelength. The sample aerogel tile manufacturedwith DMF has a refractive index of n = 1 .
044 and a thick-ness of 20.8 mm. It is known that light transmission inaerogels is dominated by Rayleigh scattering: T ( λ, t ) = A exp( − Ct/λ ) , where T ( λ, t ) is the transmittance, A and C are parame-ters, and t is the aerogel thickness. The parameter C iscalled the ”clarity coefficient” and is usually measured inunits of µ m /cm. The parameters obtained from the fit-ting are A = 1 and C = 0 . ± . µ m /cm. The5 avelength [nm]200 300 400 500 600 700 800 T r a n s m itt a n ce [ % ] / ndf χ ± ± χ ± ± χ ± ± Figure 6: Aerogel transmittance curve for aerogel ID = PDR21-2a, n = 1 . t = 20 . T = A exp( − Ct/λ ) are A = 1 and C = 0 . ± . µ m /cm. Theupper limit of the parameter A was set to 1 in the fitting procedure.The corresponding transmission length was calculated to be 50 mmat λ = 400 nm. upper limit of the parameter A was set to 1 in the fittingprocedure. The transmission length Λ T ( λ ), which is defined asΛ T ( λ ) = − t/ ln T ( λ ), is a useful parameter for comparingtransparencies of aerogel tiles with differing refractive in-dices and thicknesses because, provided the surface effectcan be neglected, Λ T is independent of thickness. We usu-ally evaluate the transmission length at λ = 400 nm, whichis where typical photon detectors have peak quantum ef-ficiency. The improvement of the transmission length forany given refractive index is our biggest goal.Fig. 7 shows the distribution of transmission length inthe range n = 1.006 − n > . n = 1 .
025 we obtained the most transparent aerogel tile(Λ T = 40 mm) for aerogels synthesized with methanol.For both n < .
025 and n > . n = 1.040 − n = 1 . n = 1 .
050 and 1.060, respectively. Aerogels with n < . Refractive index1.00 1.05 1.10 1.15 T r a n s m i ss i on l e ng t h [ mm ] EthanolMethanolDMF
Figure 7: Distribution of transmission length at λ = 400 nm over awide refractive-index range. The refractive index was measured at λ = 405 nm. A total of 142 aerogel tiles are represented by separatedots: triangles, circles, and squares denote aerogels synthesized withethanol, methanol, and DMF, respectively. The transmission lengthwas calculated from transmittance measured with the spectropho-tometer and from the measured thickness of the aerogels.
4. Fine structure of silica aerogel
The transparency of aerogels depends on their nanos-tructure (i.e., their three-dimensional configuration) andthe size of the primary
SiO particles and pores. Thecharacteristic length scale of the nanostructure formed inthe solvogel synthesis process was determined by scanningelectron microscopy to be on the order of 10 nm. Experi-ence shows that reducing the gelation time leads to moretransparent aerogels. Aerogels synthesized using DMF aregenerally more transparent than those synthesized usingmethanol; a result that we attribute to the fact that DMFshould form a finer SiO structure. To verify this hypoth-esis, we performed SAXS experiments, which we describenext. SAXS is the diffusive scattering produced by the con-trast of the nanoscale electron density of the sample (i.e.,Thomson scattering). A typical scattering angle is 2 θ =0 ◦ to 5 ◦ with respect to the incident X-rays. For porousmaterials such as silica aerogel, information regarding thesize and shape of the particles and pores may be obtainedby SAXS because it is sensitive to structures 1 to 100 nmin size. Bragg’s law applied to crystalline materials relatesthe scattering (or diffraction) angle to the lattice spacing d of crystals:2 d sin θ = λ, λ is the X-ray wavelength. In SAXS, the intensity I of scattered X-rays is measured as a function of the scat-tering parameter q = 4 π sin θ/λ . From Bragg’s law andthe definition of the scattering parameter, the structuralspacing of crystals is given by d = 2 π/q . We prepared approximately 1-mm-thin aerogel platesusing methanol or DMF as solvent for SAXS measure-ments and also manufactured 1-cm-thick aerogels with atile shape as a reference for optical measurements. Theirrefractive indices and transmission lengths were n = 1 . n = 1 .
061 and 23 mm, respectively, when DMFwas used as solvent.The SAXS measurements were conducted at theNishikawa Laboratory, Department of Chemistry, ChibaUniversity on a NANO-Viewer (Rigaku) SAXS system,which consists of a monochromatized X-ray generator ( Cu K α , λ = 1.54 ˚A), a focusing multilayer optic with threeslits, and a detector. Except for the space where the sam-ple holder was installed, the X-ray path was in vacuum( <
100 Pa) to avoid scattering of X-rays by air. X-raysscattered by the aerogel plates were detected with a two-dimensional imaging plate. The imaging plates were an-alyzed with an R-AXIS DS3C scanner (Rigaku). Moredetails on the SAXS method are given in Ref. [22].
Fig. 8 shows the X-ray intensity as a function ofthe scattering parameter for aerogels synthesized withmethanol or DMF. We conclude that the particle or poreshapes of the two aerogels were the same because theSAXS profiles are similar to each other. However, the twocurves intersect. The scattering parameters taken at thepeak intensities are q = 0 . ± . . ± . − for methanol and DMF, respectively, which meansthat the structural spacings were clearly different. Finerstructural information appears at the higher range of scat-tering parameter. Although we should analyze the entirerange of scattering parameter for more details [23], this re-sult supports the conclusion that aerogels synthesized withDMF form a finer particle structure, making them moretransparent than those synthesized with methanol.
5. Density measurement
When we use aerogels as a Cherenkov radiator, eachaerogel tile is optically characterized by its refractive in-dex, as described in Sec. 3. However, for aerogels used ascosmic dust collectors in planetary science, each aerogeltile is usually characterized by its bulk density. Acqui-sition of cosmic material is very important in planetaryscience. Micron-size dust particles traveling at hyperve-locity (typically ∼
10 km/s) in a low earth orbit (LEO)or in deep space can create impact craters on the aerogel ] −1 Scattering parameter, q [nm 0.0 0.5 1.0 1.5 2.0 2.5 3.0 S AX S i n t e n s it y , I [ a r b . un it s ] Constant 12.38 Methanol
Constant 6.825 DMF
Figure 8: Scattered X-ray intensity I ( q ) as a function of scatteringparameter q for aerogels synthesized with methanol (dotted line) orDMF (solid line) as solvent. The SAXS profiles take into account theeffects of the intensity fluctuation of the incident X-ray beam and theX-ray absorption of the aerogels. The intensities were fitted aroundthe peaks by Gaussian functions, and the parameters obtained areshown on the graph. surface and dig into the aerogel with impact tracks in theintact capture process. Thus, for this application, the den-sity of aerogels is a barometer of the hypervelocity captureperformance.Although only aerogels with low refractive indices (be-low 1.016, which corresponds to a density ρ ≤ .
06 g/cm )are used as dust samplers, the density of all manufacturedaerogels, including those with higher refractive indices, arealways measured. From the mass, dimensions, and thick-ness, the density of aerogels can be gravimetrically deter-mined. The mass is measured with an electronic balance toan accuracy of 10 − g. Our hydrophobic aerogels have nochange in mass for at least ten years after production. Thedimensions, including the thickness, are measured with ascale in increments of 0.25 mm along the four sides of theaerogel tile. The thickness measurement is also necessaryto calculate transmission length. The form of an aero-gel tile is fixed in the solvogel synthesis process, wherethe prepared solution coagulates with a meniscus on thesurface of solvogels. Because of the meniscus geometry,aerogel tiles are approximately 0.5 mm thinner near thecenter. Transparent aerogels allow us to estimate averagethicknesses through the sides. Note that the Panasonicproducts (e.g., SP-50) have no meniscus (thickness varia-tion ∆ t < . n ( λ ) = 1 + k ( λ ) ρ , where k is a constant that7 Density [g/cm0.0 0.1 0.2 0.3 0.4 0.5 0.6 R e fr ac ti v e i nd e x / ndf χ ± ± χ ± ± χ ± ± Methanol / ndf χ ± ± χ ± ± χ ± ± DMF / ndf χ ± ± χ ± ± χ ± ± Ethanol
Figure 9: Refractive index measured at λ = 405 nm as a functionof aerogel density. The aerogels, which are identical to those inFig. 7, were synthesized with ethanol, methanol, and DMF and areindicated by triangles, circles, and squares, respectively. The dottedlines represent best fits by linear functions to the three data series.The k values are labeled as ”Slope” in the legends. depends on the wavelength of light and is given as 0.21cm /g in the particle data booklet [24], which is consis-tent with the values given in Ref. [25]. Our hydrophobicaerogels with added trimethylsiloxy groups show larger k to some extent. Fig. 9 shows the relationship betweenrefractive index and density for aerogels synthesized withethanol, methanol, and DMF. We found different valuesfor k depending on the solvent used: at λ = 405 nm, k =0.251, 0.283, and 0.300 cm /g for ethanol, methanol, andDMF, respectively. This result reflects the nanostructuraldifferences between aerogels synthesized with different sol-vents. Although the difference in k between methanol andDMF is small, the refractive index of aerogels synthesizedwith DMF is always larger than that of aerogels synthe-sized with methanol (at the same density).The measurement of dimensions is advantageous forcomputing the shrinkage ratio of aerogel tiles. From itsevaluation, we can understand that fluctuations in the re-fractive index between several aerogel tiles or between pro-duction batches are attributed in large part to fluctuationsin the shrinkage ratio. The longitudinal shrinkage ratio asa function of refractive index is shown in Fig. 10. Thisratio is defined as l/l , where l is the length of the sidesof a manufactured final aerogel and l is the size of themolds for the solvogel synthesis. In Fig. 10, we see twogroups of data near n ∼ .
045 for aerogels synthesizedwith DMF. This result is a reflection of the dependenceof the shrinking ratio on aerogel size; that is, in this case,
Refractive index1.00 1.05 1.10 1.15 L ong it ud i n a l s h r i nk a g e r a ti o EthanolMethanolDMF
Figure 10: Aerogel longitudinal shrinkage ratio as a function of re-fractive index (at 400 nm). Triangles, circles, and squares representthe aerogels synthesized with ethanol, methanol, and DMF, respec-tively. small aerogels (52 cm ) shrank more than large aerogels(183 cm ). When we adjust the refractive index in aero-gel preparation, it is important to take into account theshrinkage.
6. Conclusion
We characterized our original (and conventional)method of producing hydrophobic silica aerogel. Sincethe early 2000s, several thousand aerogel tiles have beenmanufactured at Chiba University and Panasonic ElectricWorks under the Belle upgrade program and for other pur-poses. We examined all the aerogels produced, and 142samples were selected and described in detail in terms oftheir basic properties. The introduction of DMF as solventin the solvogel synthesis process resulted in improvementsin aerogel transparency in the high-refractive-index range.SAXS measurements revealed that finer aerogel structurecan be formed by using DMF as solvent.
Acknowledgments
We are grateful to the members of Particle Physics Lab-oratory of Chiba University for their assistance in aerogelproduction. We are also grateful to Dr. Y. Hatakeyamaof Chiba University for his full support with the SAXS ex-periments. This work was partially supported by a Grant-in-Aid for JSPS Fellows (No. 07J02691 for M.T.) and aGrant-in-Aid for Scientific Research (C) (No. 17540284and 19540317 for I.A.) from the Japan Society for the Pro-motion of Science (JSPS) and a Grant-in-Aid for ScientificResearch on Innovative Areas (No. 21105005) from theMinistry of Education, Culture, Sports, Science and Tech-nology (MEXT). This publication was in part supported8y the National Institute for Fusion Science (NIFS) in theNational Institute of Natural Sciences (NINS) of Japan.
References [1] M. Cantin, et al., Nucl. Instr. and Meth. 118 (1974) 177.[2] D. Brownlee, et al., Science 314 (2006) 1711.[3] H. Bergs˚aker, et al., J. Nucl. Mater. 415 (2011) S1089; M. Tang,et al., J. Nucl. Mater. 415 (2011) S1094.[4] Y. Fujiwara, et al., RIKEN Accel. Prog. Rep. 44 (2011) in press.[5] H. Kawai, et al., Nucl. Instr. and Meth. 228 (1985) 314.[6] Y. Saito, et al., in: IEEE Nuclear Science Symposium, Confer-ence Record, 2005, p. 2456.[7] Y. Asaoka, et al., Nucl. Instr. and Meth. A 416 (1998) 236.[8] P. Carter, HERMES Collaboration, Nucl. Instr. and Meth. A433 (1999) 392.[9] Y. Kitazawa, et al., JAXA Special Publication, Proceedings ofInternational Symposium on ”SM/MPAC&SEED Experiment,”JAXA-SP-08-015E, 2009, p. 49.[10] M. Tabata, et al., in: IEEE Nuclear Science Symposium, Con-ference Record, 2005, p. 816.[11] M. Tabata, et al., Nucl. Instr. and Meth. A 623 (2010) 339.[12] M. Tabata, et al., Biol. Sci. Space 25 (2011) 7.[13] T. Abe, et al., Belle II Technical Design Report, KEK Report2010-1.[14] A. Yamagishi, et al., Trans. JSASS Space Tech. Japan 7 (2009)Tk 49.[15] I. Adachi, et al., in: IEEE Nuclear Science Symposium, Con-ference Record, 2004, p. 687; I. Adachi, et al., Nucl. Instr. andMeth. A 553 (2005) 146.[16] T. Iijima, et al., Nucl. Instr. and Meth. A 548 (2005) 383.[17] I. Adachi, et al., Nucl. Instr. and Meth. A 355 (1995) 390.[18] T. Sumiyoshi, et al., Nucl. Instr. and Meth. A 433 (1999) 385.[19] A. Abashian, et al., Nucl. Instr. and Meth. A 479 (2002) 117.[20] I. Adachi, et al., Nucl. Instr. and Meth. A 595 (2008) 180.[21] H. Yokogawa, M. Yokoyama, J. Non-Cryst. Solids 186 (1995)23.[22] Y. Hatakeyama, et al., J. Phys. Chem. C 113 (2009) 3917.[23] A. Rigacci, et al., J. Non-Cryst. Solids 285 (2001) 187.[24] W.-M. Yao, et al., Particle Data Group, J. Phys. G 33, (2006)1.[25] G. Poelz, Nucl. Instr. and Meth. A 248 (1986) 118; A.R.Buzykaev, et al., Nucl. Instr. and Meth. A 433 (1999) 396; A.F.Danilyuk, et al., Nucl. Instr. and Meth. A 494 (2002) 491.[1] M. Cantin, et al., Nucl. Instr. and Meth. 118 (1974) 177.[2] D. Brownlee, et al., Science 314 (2006) 1711.[3] H. Bergs˚aker, et al., J. Nucl. Mater. 415 (2011) S1089; M. Tang,et al., J. Nucl. Mater. 415 (2011) S1094.[4] Y. Fujiwara, et al., RIKEN Accel. Prog. Rep. 44 (2011) in press.[5] H. Kawai, et al., Nucl. Instr. and Meth. 228 (1985) 314.[6] Y. Saito, et al., in: IEEE Nuclear Science Symposium, Confer-ence Record, 2005, p. 2456.[7] Y. Asaoka, et al., Nucl. Instr. and Meth. A 416 (1998) 236.[8] P. Carter, HERMES Collaboration, Nucl. Instr. and Meth. A433 (1999) 392.[9] Y. Kitazawa, et al., JAXA Special Publication, Proceedings ofInternational Symposium on ”SM/MPAC&SEED Experiment,”JAXA-SP-08-015E, 2009, p. 49.[10] M. Tabata, et al., in: IEEE Nuclear Science Symposium, Con-ference Record, 2005, p. 816.[11] M. Tabata, et al., Nucl. Instr. and Meth. A 623 (2010) 339.[12] M. Tabata, et al., Biol. Sci. Space 25 (2011) 7.[13] T. Abe, et al., Belle II Technical Design Report, KEK Report2010-1.[14] A. Yamagishi, et al., Trans. JSASS Space Tech. Japan 7 (2009)Tk 49.[15] I. Adachi, et al., in: IEEE Nuclear Science Symposium, Con-ference Record, 2004, p. 687; I. Adachi, et al., Nucl. Instr. andMeth. A 553 (2005) 146.[16] T. Iijima, et al., Nucl. Instr. and Meth. A 548 (2005) 383.[17] I. Adachi, et al., Nucl. Instr. and Meth. A 355 (1995) 390.[18] T. Sumiyoshi, et al., Nucl. Instr. and Meth. A 433 (1999) 385.[19] A. Abashian, et al., Nucl. Instr. and Meth. A 479 (2002) 117.[20] I. Adachi, et al., Nucl. Instr. and Meth. A 595 (2008) 180.[21] H. Yokogawa, M. Yokoyama, J. Non-Cryst. Solids 186 (1995)23.[22] Y. Hatakeyama, et al., J. Phys. Chem. C 113 (2009) 3917.[23] A. Rigacci, et al., J. Non-Cryst. Solids 285 (2001) 187.[24] W.-M. Yao, et al., Particle Data Group, J. Phys. G 33, (2006)1.[25] G. Poelz, Nucl. Instr. and Meth. A 248 (1986) 118; A.R.Buzykaev, et al., Nucl. Instr. and Meth. A 433 (1999) 396; A.F.Danilyuk, et al., Nucl. Instr. and Meth. A 494 (2002) 491.