A novel fluidized bed process to produce fine-grade artificial lightweight aggregates

A novel fluidized bed process to produce fine-grade artificial lightweight aggregates
of 10
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  A novel fluidized bed process to produce fine-gradeartificial lightweight aggregates Satoshi Kimura a, *, Kaoru Kimura  b , Hidehiro Kamiya a  , Masayuki Horio a  a  The Graduate School of BASE, Tokyo University of Agriculture and Technology, Naka-cho 2-24-16, Koganei-shi, Tokyo 184, Japan  b  Niijimabussan Co. Ltd., Touyou 3-26-26 Koutou-ku, Tokyo 135, Japan Received 20 October 2003; received in revised form 10 May 2004; accepted 11 August 2004Available online 21 September 2004 Abstract Fine-grade artificial lightweight aggregates (ALAs) with particle diameters within 300–600  A m were manufactured by using a fluidized bed of spherical mullite particles. The product aggregates contained uniformly dispersed closed bubbles and showed excellent characteristics of light weight, high strength, and low water absorption, all of which are of ideal properties for lightweight concrete production. By applying mercury intrusion test, tensile strength was estimated from isostatic compressive strength with excellent reliability. High strength during pumping and casting concrete mix is expected for the fine ALAs from fluidized bed process. D  2004 Elsevier B.V. All rights reserved.  Keywords:  Artificial lightweight aggregate; Volcanic silicate; Closed pore structure; Fluidized bed 1. Introduction Fine-grade artificial lightweight aggregates (ALAs),once they were successfully made from ceramic materials,would be effective to mix into portland cement to make pumpable and castable lightweight concrete. Character-istics required for ALAs are lightweightness, high strength,and low water absorption capacity. Low water absorptioncharacteristics make us possible to avoid difficultiesassociated with water absorption, such as choking during pumping and poor workability in concrete casting. ALAshave a potential in reducing constructing technical limitsfor ultra-high-rise buildings/long span structures andimproving their being earthquake-proof. Highly porousALAs would also provide precast siding board withefficient thermal insulation property. Especially, fine-gradeALAs of diameters below 600  A m should be effective toimprove both reliability and strength of concrete structuresor to produce thinner siding boards because of theworkability and castability of the concrete mix prepared by using them.To develop low-water-absorption ALAs, we have tomake the pores in them not open ones but small closed bubbles with a uniform size distribution. Kimura et al. [1]and Tachibana et al. [2] proposed a method to produce porous ceramics particles, which have a bubbly structure by high-temperature oxidation of SiC as a foaming agent in the molten phase of volcanic silicate pellets. So far, thehigh-temperature foaming has been conducted in a rotarykiln, where nonuniform foaming, overmelting, andunwanted agglomeration have been inevitable due to thetemperature nonuniformity and high particle–particleinteraction force in a rotary kiln. Rotary kilns are,accordingly, not appropriate to produce fine particleswhose diameter is less than 1 mm, because severeagglomeration cannot be suppressed in them. Fluidized bed processes are expected to prevent agglomeration because of its much less particle-to-particle interactionthan in a rotary kiln. This is because in a fluidized bed, particles are suspended by fluid drag force. From the 0032-5910/$ - see front matter   D  2004 Elsevier B.V. All rights reserved.doi:10.1016/j.powtec.2004.08.002* Corresponding author. Tel.: +81 42 388 7067; fax: +81 42 386 3303.  E-mail address: (S. Kimura).Powder Technology 146 (2004) 111–  same reason, fluidized beds have superb characteristics of good solid mixing and rapid heat transfer to prevent inhomogeneous heat treatment. b Shirasu balloons  Q  , i.e., glassy balloons made fromvolcanic silicates produced from Shirasu soil of southernKyushu plateau of Japan, have been produced withfluidized beds containing inert bed material [3–5] where balloons are foamed by rapid evaporation of moisture inthe molten glassy phase within several seconds. A balloonusually contains one or several bubbles. Shirasu balloonshave good lightweightness but particularly those greater than 50  A m in diameter lack hardness. b Sludge lite  Q  , ALAs having diameters within 0.6–3.5 mm produced by Tokyo metropolitan bureau of sewage, has been produced f rom sewage sludge with a multistagespouted bed [6,7]. The sludge aggregates are foamed during nonuniform melting. A product particle is composed of awell-melted inner part which shows a bubbly structure and anonmelted skin layer, where pores are open and areconnected to inner bubbles. Accordingly, sludge aggregatesare highly water-absorbing (i.e., 8.4–15.4%).In the above two techniques, agglomeration has beenavoided by short contact time or by incomplete melting.However, to produce fine-grade high-performance ALAs,it is necessary to achieve a situation close to the completemelting so that bubbles are uniformly distributed, poresare closed, and the grain surface is sufficiently smoothwith no pore opening. To make such foaming structuresfrom volcanic silicates, a sufficiently long holding timeand gradual temperature rise are necessary during shrink-ing period due to softening and foaming of raw materials[7]. Thus, it should be the core issue of processdevelopment for fine-grade high-performance ALAs to prevent agglomeration of well-melted particles during along holding time.In this study, fine-grade ALAs of diameters within 300– 600 A mforlightweightconcretesareexperimentallyobtained byusingalaboratoryscalefluidizedbedofinertbedmaterialsand areexamined byfocusingon howthe foaming tookplacein the bed without defluidization. The properties of the product granules were evaluated in terms of apparent density,tensile strength, and isostatic compressive strength. 2. Raw material and pretreatment 2.1. Raw material and a foaming method  The raw material for the ALA is porous biotite rhyolite(Kouka-seki, in Japanese), which belongs to volcanicsilicates. The material has been produced in the NiijimaIsland of Tokyo. As shown in Table 1, it is rich in SiO 2 and Al 2 O 3 , containing a little amount of Fe 2 O 3 , havingdurability and fireproof  [2]. The biotite rhyolite has been known as the raw material for porous ceramics throughhigh-temperature heat treatment with coexist ing a siliconcarbide (SiC) as a foaming agent (cf., Ref. [1]). The SiC foaming method consists of pulverization of the raw material, addition of the foaming agent , gran-ulation, and heat treatment as shown in Fig. 1. When volcanic silicates in raw pellets are heat-treated (1423– 1523 K) to the molten glassy condition, SiC particles aredispersed into the molten glassy matrix and then react with SiO 2  under oxidizing atmosphere to evolve CO or CO 2  gas in the glass matrix. For homogeneous foaming inthe SiC foaming method, a high amount of glassy SiO 2 , alow amount impurity (especially Fe 2 O 3  and carbon), and avery fine grain size of raw material (below 10  A m) aredesirable [1]. 2.2. Raw pellets and pretreatments Raw pellets used in the presented work contained mainly pulverized biotite rhyolite, 0.2 wt.% of SiC as a foamingagent, and 3.0 wt.% of bentonite as a binder, as shown inTable 2. With a drum granulator, the mixed powder was pelletized and then classified to 0.25–0.5 mm by sieving toobtain raw green pellets. Properties of raw pellets are shownin Table 3. Table 1Analysis of biotite rhyolite (wt.%)SiO 2  Al 2 O 3  Fe 2 O 3  CaO MgO K  2 O Na 2 O Ig.loss78.7 12.3 0.87 0.85 0.09 2.72 4.01 0.39Fig. 1. Prescription to obtain ALA.Table 2Composition of raw pelletsMaterials Rules Content (wt.%)Biotite rhyolite Raw material 96.8Silicon carbide Foaming agent 0.2Bentonite Binder 3.0 S. Kimura et al. / Powder Technology 146 (2004) 111–120 112  Raw pellets were heat-treated for 30 min at 1173–1273K for calcination. A fine alumina powder ( d   psv =6.46  A m)was used to coat the raw pellets to prevent agglomeration. 3. Experiments 3.1. Fluidized bed sintering  A steel tube (  D =38 mm) placed in an electric furnace wasused as a fluidized bed column. The schematic of theexperimental setup is shown Fig. 2. A fixed bed of alumina  balls ( d   p =2 mm, fill height=0.28 m) was used as the air distributor and preheater. Fluidizing air was supplied via amass flow controller to maintain the excess gas velocity( u 0  u mf  ) constant by a computer-assisted system regardlessof the temperature in the bed. The bed temperature wasmeasured by using a K-type thermocouple immersed in the bed. The bed pressure drop was measured by a differential pressure sensor. Minimum fluidization velocity of the bedmaterial at elevated temperature was estimated by the Wen– Yu correlation [8].Three heat treatment conditions were tested: case 1—raw pelletsfluidizedwithoutinertbedmaterial(i.e.,100wt.%raw pellets); case 2—small amount of raw pellets in inert beds,case2-1: 96.0 wt.%ofnonspherical aluminaparticles and4.0wt.% raw pellets, case 2-2: 95.3 wt.% of spherical mullite beads and 4.7 wt.% raw pellets; case 3—high amount of raw pellets (i.e., 35.5 wt.% raw pellets) in a bed of inert bedmaterials.In case 1, raw pellets of 64.8 g (to a static bed height of 60mm) were charged into the column to a height of 60 mm andheated up to 1473 K at a rate of 5 K/min. In case 2, inert bedmaterials (130.6 g alumina particles in case 2-1 and 117.6 gmullite beads in case 2-2) were charged into the column to aheight of 60 mm and heated to 1423 K under a fluidizingcondition.Thenrawpellets of5.4g(equivalenttoastaticbedheight of 5 mm) were fed from the top of the column into thefluidized bed within 120 s. Fluidization was continued for another 800 s. In case 3, inert bed materials of 58.8 g mullite beads was charged in the column to a height of 30 mm andheated under a fluidizing condition. Then, raw pellets of 32.4g (equivalent to a static bed height of 30 mm), preheated toabout 1273 K in a box furnace, were fed from the column topinto the fluidized bed of predetermined temperature (1423– 1473 K) within 120 s, and the fluidization was continued for another 500–2000 s. Par ticle properties of bed materials arealso included in Table 3. 3.2. Determination of product properties3.2.1. Apparent density and water absorption ratio Concerning the characterization of fine aggregates (par-ticlesize5–0.15mm)forstructuralconcretes,ALAsapparent density, water absorption after 24 h in water, and saturatedsurface-dry condition (SSD condition) have been defined byJIS A1134, where the flow cone test has been adopted toexamine the SSD condition.In the flow cone test, ALAs are dipped in water for 24 hand then dried slowly by an oven. During the drying process,theyarepickedupandfilledupintoaflowcone,amoldintheform of frustum of a cone, and then the mold is liftedvertically. In case surface moisture is still present, the fineaggregates will retain the molded shape. If they have reachedthe SSD condition, the slight slump of the shaped aggregatescan be detected. Some angular particles or lightweight  particles with a high proportion of fines may not slump intheconetestuponreachingtheSSDcondition.Insuchacase,removing fines of diameter below 150  A m is allowed.However, because the reactor size in the present stage of this research was still too small to prepare sufficient sampleamount for JIS A1104, which requires a sample of roughly1.6kg,determinationofSSDcondition,apparentdensity,andwater absorption were done in our srcinal way. First of all,the SSD condition of ALA particles was achieved by dipping Table 3Properties of raw pellet and inert bed materialsParticles Particlesize d   p  ( A m)Surface-to-volume mean particle size d   psv  ( A m)Apparent density q  p (kg/m 3 )Minimumfluidizationvelocity u mf   (m/s)Raw pellet 500–250 309 1990 0.809Alumina powder (nonspherical)420–180 258 3980 0.0758Mullite beads(spherical)425–200 280 2820 0.0688Fig. 2. Experimental apparatus. S. Kimura et al. / Powder Technology 146 (2004) 111–120  113   particlesinwater(293K)for24handdryingbyhotairof328K.TodetecttheSSDcondition,particleswereputintoaglasstube (ID 24 mm), and then the tube was rolled over 90 8  to seeifparticles sticktothetubewall.Ifnot,theyweresupposedto be in the SSD condition. Such a sample was then divided intotwo parts to determine the apparent density in the SSDcondition and the water content. About 10 g of ALAs in theSSD condition was taken out, and their weight,  m SSD , wasmeasured. Then, they were brought to a sufficiently driedcondition in an oven at 373 K for 24 h, and then their weight  m d wasmeasured.Thewaterabsorption Q  wasdeterminedbyEq. (1). Q  ¼  m SSD    m d m d 100 % ½  ð 1 Þ The remaining part of the ALA sample in the SSDcondition was used to measure their density  q SSD  by puttingit into a picnometer (water capacity 50 ml). Apparent density of dry ALAs  q  p  was calculated by Eq. (2). q  p  ¼  q SSD 100100  þ  Q  ð 2 Þ The porosity of ALAs  e  was determined from Eq. (3)from  q  p  and true density  q t   determined after fine grinding. e  ¼  1   q  p q t  ð 3 Þ 3.2.2. Tensile strength The tensile strength of ALAs  F  t   was calculated by thefollowing Hiramatsu correlation [9] from load  w  under monoaxial compression:  F  t   ¼  2 : 8  w p d  2 p ð 4 Þ where, particle diameter   d   p  was determined by a slide gage,and experimental values of load  w  was measured by amonoaxial compression tester. A monoaxial compressiontester consisted of a universal testing machine (ShimadzuAGS-G) with a flat zig and an electronic balance. Theelectronic balance was set under the crosshead. A particle to be measure was put on the center of the pan of the electronic balance. Then, the crosshead was sent down at 0.5 mm/min,and the crush load was measured. The tensile strength of ALAs was determined from an average of 30 measurementsfor each lot. 3.2.3. Mercury intrusion test  To produce precast siding boards by high-pressureextrusion casting, the highly isostatic compressive strengthof ALAs is an important property to prevent breaking andwater absorbing during extrusion casting. In this work,isostatic compressive strength of 0.5 MPa was taken as thetarget value according to Kamio [10]. The isostatic com-  pressive strength of ALAs was determined by a mercury porosimeter (Shimadzu micromeritics pore sizer 9310).About 0.1 g of ALAs was placed in the cell of 0.384 cm 3 ,and after evacuation, mercury was introduced and pressurewas raised up to 200 MPa. The collapse of the solid bubblewalls under high pressure can be detected from an increase inthe amount of intruded mercury. 4. Results and discussions 4.1. Thermomechanical analysis of raw pellets and defluid-ization behavior  Fig. 3 shows results of thermomechanical analyisis(TMA) of raw pellet by Rigaku TAS-100 under 5 kPa load Fig. 3. Relative shrinkage of raw pellets in TMA.Fig. 4. Defluidization behavior of raw pellets with no inert bed material.Fig. 5. The bed pressure drops with inert bed materials. S. Kimura et al. / Powder Technology 146 (2004) 111–120 114  and 5 K/min of heating rate. Here,  L  is static packed bedheight of 0.5 g of sample, and  D  L  is change of packed bedheight.This result indicates that biotite rhyolite pellets shrank due to sintering at above 1173 K and swelled due tofoaming above 1390 K. Fig. 4 shows a typicaldefluidization behavior of the fluidized bed detected witha differential pressure sensor via a pressure probeimmersed in the bed. When the raw pellets were heatedwithout any inert bed materials, the bed pressure drop Fig. 6. Comparison of surface and bubble structure of aggregates (a) appearance, (b) cross-section, (c) surface, and (d) bubbles structure. Left: fluidized bed product 1423 K 500 s  q  p =1520 kg/m 3 ; right: rotary kiln test product 1473 K   q  p =1024 kg/m 3 . S. Kimura et al. / Powder Technology 146 (2004) 111–120  115
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks