Advances in Ceramics for Environmental, Functional, Structural, and Energy Applications
Advances in Ceramics for Environmental, Functional, Structural, and Energy Applications Ceramic Transactions Volume 265
Edited by
Morsi M. Mahmoud Kumar Sridharan Henry Colorado Amar S. Bhalla J. P. Singh Surojit Gupta Jason Langhorn Andrei Jitianu Navin Jose Manjooran
This edition first published 2018 © 2018 The American Ceramic Society
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions. The rights of Morsi M. Mahmoud, Kumar Sridharan, Henry Colorado, Amar S. Bhalla, J. P. Singh, Surojit Gupta, Jason Langhorn, Andrei Jitianu, and Navin Jose Manjooran to be identified as the authors of the editorial material in this work have been asserted in accordance with law. Registered Office John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA Editorial Office 111 River Street, Hoboken, NJ 07030, USA For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com. Wiley also publishes its books in a variety of electronic formats and by print-on-demand. Some content that appears in standard print versions of this book may not be available in other formats. Limit of Liability/Disclaimer of Warranty In view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of experimental reagents, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each chemical, piece of equipment, reagent, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. Library of Congress Cataloging-in-Publication Data is available. ISBN: 9781119543251 ISSN: 1042-1122
Cover design by Wiley Printed in the United States of America. 10 9 8 7 6 5 4 3 2 1
Contents
Preface
Jx
."5&3*"-4'03/6$-&"3&/&3(:"11-*$"5*0/4 Westinghouse Accident Tolerant Fuel Materials
3
Frank Boylan, Peng Xu, Javier Romero, and Ed Lahoda
Investigations of Capacitor Discharge Welding for the Attachment of Endcaps to Molybdenum-Based Nuclear Fuel Rod Cladding
7
Jerry E. Gould, Cem Topbasi, and Bo Cheng
Evaluation of 6 4J Fuel Pellets Sintered in an Argon vs. Vacuum Environment
21
Rita Hoggan, Jason Harp, and Lingfeng He
130$&44*/("/%1&3'03."/$&0'."5&3*"-464*/( .*$308"7&4 &-&$53*$"/%."(/&5*$'*&-%4 6-53" 406/% -"4&34 "/%.&$)"/*$"-803,oo36456.30: 4:.104*6. Microwave-Augmented Crystallization and Decrystallization in Ceramic Processing –– A Phenomenology-Based Commentary
29
Boon Wong
Effects of Elastic Waves at Several Frequencies on Biofilm Formation in Circulating Types of Laboratory Biofilm Reactors
43
Hideyuki Kanematsu, Shogo Maeda, Dana M. Barry, Senshin Umeki, Kazuyuki Tohji, Nobumitsu Hirai, Akiko Ogawa, Takeshi Kogo, Hajime Ikegai, and Yoshimitsu Mizunoe
Resilient Graphitic Carbons from Electro-Thermal Fluidized Bed Reactor
53
Soeren Koester, Eric Salmon, and Carsten Wehling
v
Heavy Clay Body Properties for Hybrid Microwave Firing
59
Garth V A Tayler, Mike Anderson, and Mike Hamlyn
Effective Permittivity and Microwave Heating Characteristics of Electric Conductor and Insulator (Dielectrics) Mixtures
69
Noboru Yoshikawa
Porous Ceramic/Metal Composite Body for DPF (Diesel Particulate Filter) and the Microwave Heating Behavior
79
Noboru Yoshikawa, Chang Chuan Lee, Naoki Inoue, Shoji Taniguchi, and Sergey Komarov
$0/4536$5*0/"/%#6*-%*/(."5&3*"-4'03"#&55&3 &/7*30/.&/5 Study on Preparation of Portland Cement by Using Coal Fly Ash
87
Hui Sun, Miaolian Bian, Dongyang Ma, Shichao Chen, Zhicheng Cao, and Daohong Wu
Corrosion Behaviour of Steel-Reinforcement in C 3 H 7 NO2 S-Admixed Concrete Immersed in Saline/Marine Simulating-Environment
95
Joshua Olusegun Okeniyi and Abimbola Patricia Idowu Popoola
C4H11NO Performance on Steel-Rebar Corrosion in Industrial/ Microbial Simulating Environment
109
Joshua Olusegun Okeniyi and Abimbola Patricia Idowu Popoola
*//07"5*7&130$&44*/("/%4:/5)&4*40'$&3".*$4 (-"44&4"/%$0.104*5&4 Numerical Investigation of Heat Transfer and Reaction Kinetics during 123 the Self-Propagating High-Temperature Synthesis of Silicon Nitride Venkata V. K. Doddapaneni and Sidney Lin
Synthesis of Carbide Ceramics from Activated Carbon Precursors Loaded with Tungstate, Molybdate, and Silicate Anions
137
Grant Wallace, Jerome Downey, Jannette Chorney, Katie Schumacher, Trenin Bayless, Alaina Mallard, Auva Speiser, and Elizabeth Raiha
"%7"/$&4*/%*&-&$53*$."5&3*"-4"/%&-&$530/*$ %&7*$&4 Observation of TI-TI Bonding In TI/CU/PT-Supported Rutile TiO2 (110) 153 Surface: AB Initio Calculations Lei Li, Wenshi Li, Han Qin, Jianfeng Yang, Canyan Zhu, and Ling-Feng Mao
vi
Advances in Ceramics for Environmental, Functional, Structural, and Energy Applications
Effect of Processing Conditions on Electric and Dielectric Properties of Polymer-Derived SiC Ceramics
165
Chengying Xu
5)&5)*/5&3/"5*0/"-4:.104*6.0/(3&&/"/% 4645"*/"#-&5&$)/0-0(*&4'03."5&3*"-4 ."/6'"$563*/("/%130$&44*/( Study on Energy Utilization of High Phosphorus Oolitic Hematite by Different Ironmaking Technologies
177
Hui Sun, Miaolian Bian, Shichao Chen, Dongyang Ma, Zhicheng Cao, and Daohong Wu
/"/05&$)/0-0(:'03&/&3(: &/7*30/.&/5 &-&$530/*$4 )&"-5)$"3&"/%*/%6453:"11-*$"5*0/4 Nanosensors for Detecting Pollutants in Water
187
Shobhan Paul
):#3*%03("/*$*/03("/*$."5&3*"-4'03"-5&3/"5*7& &/&3(: Electrodeposition of Hybrid Sol-Gel Glass Coatings on 304 Stainless Steel for Corrosion Protection
205
Q. Picard, G. Akalonu, J. Mercado, J. Mosa, M. Aparicio, L. C. Klein, and A. Jitianua
463'"$&1301&35*&40'#*0."5&3*"-4 Biofilm Formation on Titanium Alloy Surfaces in a Laboratory Biofilm Reactor
221
Hideyuki Kanematsu, Shun Kanesaki, Hikonaru Kudara, Dana M. Barry, Akiko Ogawa, Takeshi Kougo, Daisuke Kuroda, Nobumitsu Hirai, Hajime Ikegai, and Yoshimitsu Mizunoe
Advances in Ceramics for Environmental, Functional, Structural, and Energy Applications
vii
Preface
This volume contains 20 manuscripts presented during the Materials Science & Technology 2017 Conference (MS&T’17), held October 8-12, 2017 at the David L. Lawrence Convention Center, Pittsburgh, PA. Papers from the following symposia are included in this volume: • 9th International Symposium on Green and Sustainable Technologies for Materials Manufacturing and Processing • Advances in Dielectric Materials and Electronic Devices • Construction and Building Materials for a Better Environment • Innovative Processing and Synthesis of Ceramics, Glasses and Composites • Materials Issues in Nuclear Waste Management in the 21st Century • Materials Development for Nuclear Applications and Extreme Environments • Materials for Nuclear Energy Applications • Nanotechnology for Energy, Healthcare and Industry • Processing and Performance of Materials Using Microwaves, Electric and Magnetic Fields, Ultrasound, Lasers, and Mechanical Work – Rustum Roy Symposium These symposia provided a forum for scientists, engineers, and technologists to discuss and exchange state-of-the-art ideas, information, and technology on a dvanced methods and approaches for processing, synthesis, characterization, and applications of ceramics, glasses, and composites. Each manuscript was peer-reviewed using The American Ceramic Society’s review process. The editors wish to extend their gratitude and appreciation to their symposium co-organizers, to all of the authors for their valuable submissions, to all the participants and session chairs for their time and effort, and to all the reviewers for their comments and suggestions. We hope that this volume will serve as a useful reference for the professionals working in the field of materials science.
Morsi M. Mahmoud Kumar Sridharan Henry Colorado iY
Advances in Ceramics for Environmental, Functional, Structural, and Energy Applications
Amar S. Bhalla J. P. Singh Surojit Gupta Jason Langhorn Andrei Jitianu Navin Jose Manjooran
Advances in Ceramics for Environmental, Functional, Structural, and Energy Applications
x
.BUFSJBMTGPS/VDMFBS &OFSHZ"QQMJDBUJPOT
WESTINGHOUSE ACCIDENT TOLERANT FUEL MATERIALS Frank Boylan1, Peng Xu2, Javier Romero2, Ed Lahoda3 1
Westinghouse Electric Company, Cranberry Township 16066 +1-412-374-4950;
[email protected] 2 Westinghouse Electric Company, Columbia, SC 29061 3 Westinghouse Electric Company, Cranberry Township, PA 16066 ABSTRACT Westinghouse is commercializing two unique accident tolerant fuels (ATFs): silicon carbide (SiC) as produced by General Atomics with uranium silicide (U3Si2) fuel and Cr coated zirconium alloy cladding with U3Si2 fuel. Testing of the cladding alternatives in autoclaves has been performed and samples have begun irradiation at the Massachusetts Institute of Technology Reactor and the Halden Project Reactor. Uranium silicide fuel is undergoing exposure in the Advanced Test Reactor and fuel pins have been removed and are undergoing post irradiation examination (PIE) at the Idaho National Laboratory (INL). This paper provides an update on these activities and a summary of results. INTRODUCTION AND BACKGROUND The Westinghouse Electric Company LLC (Westinghouse) accident tolerant fuel (ATF) program utilizes Cr coated zirconium alloy (CZA) cladding with U3Si2 high density/high thermal conductivity fuel for its lead test rod (LTR) program with irradiation beginning in 2019. The lead test assembly (LTA) program will use both SiC/SiC composites from General Atomics and Cr coated zirconium alloy claddings with the high density/high thermal conductivity U3Si2 pellet which will begin in 2022. Over the past several years, Westinghouse has tested the Cr coated zirconium and SiC claddings in autoclaves and in the Massachusetts Institute of Technology (MIT) reactor and U3Si2 pellets in the Advanced Test Reactor (ATR). High temperature tests at the state-of-the-art facilities in Churchill, PA have been carried out to determine the time and temperature limits for the SiC and Cr coated zirconium claddings. WESTINGHOUSE ATF ACTIVITIES Autoclave Corrosion Testing Westinghouse has performed corrosion testing using the autoclave facility at the Churchill, PA site to screen various coatings and SiC preparation methods for corrosion resistance. As part of a multi-year program, over 12 types of coatings on zirconium alloys and approximately 10 versions of SiC have been tested in autoclaves. As a result of this testing, two coatings (Table I) were identified for testing in the MIT reactor. Testing in the MIT reactor further narrowed the options to the Cr coating. Based on the positive test results, Westinghouse is now exploring methods for production of full length rods for LTRs to be constructed in 2018 for inclusion in a commercial reactor in early 2019.
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
3
Table I – Top Zirconium Alloy Coatings Autoclave Corrosion Performance At 360oC Water Average Zr Corrosion (mg/dm2/day )
Corrosion Rate ( /year)
169
Average Corrosion rate (mg/dm2/day ) 1.07
2.22
7.67
20
0.03
3.27
0.14
Material
Proces s
Vendor
Maximu m Days
TiN/TiAlN
PVD
Cr
Cold spray
Pennsylvani a State University University of Wisconsin, Madison
Initial autoclave and reactor testing indicated relatively high levels of SiC corrosion. Autoclave testing with hydrogen peroxide was used to simulate more aggressive oxidation conditions of the reactor and to explore coolant conditions that would minimize SiC corrosion rates. The full battery of testing has been used to refine the manufacturing parameters of the SiC composites such that along with hydrogen addition to the primary coolant above 40 cc/kg [1], the current corrosion rates for SiC meet or exceed the target 7 microns/year recession rate. For a full core of SiC cladding, this would result in a maximum of 150 kg of SiO2 or about 300 ppm over an 18 month cycle. This is well below the solubility limit of ~700 ppm SiO2 at the coldest steam generator conditions. Note also, that resins are commercially available that could be added to the current resins used to maintain water chemistry to remove SiO2 on a continuous basis. High Temperature Testing The goal of the ATF program is to develop fuels that can withstand post-accident temperatures greater than 1200oC without the cladding igniting in steam or air. Therefore a crucial part of the testing carried out by Westinghouse over the previous year was aimed at quantifying the maximum temperature at which the ATF claddings could operate without excessive corrosion. The test apparatus first used current applied directly to the coated zirconium tubes. However, it was found that as the temperatures increased, issues with the connection of the test piece to the current source caused excessive resistance resulting in excessive heating and then burnout of the samples at the connection point. This direct heating method was then replaced with a graphite rod which was inserted into insulation and then into the test piece. This resulted in very stable heating of the test pieces. CZAs have now been run at up to 1400oC. This is above the Cr-Zr low melting eutectic point of 1333oC. At 1400oC, there was noticeable reaction between the Cr and the Zr. However, there was not the rapid oxidation that uncoated Zr experiences at 1200oC, so that there is likely some reasonable residence time that the cladding could survive at temperatures above 1400oC.
4
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
At temperatures of 1300oC, the Cr coated zirconium alloy was stable for reasonable lengths of time. Combined with the lowering of zirconium oxidation at normal operating temperatures which vastly reduces the formation of zirconium hydrides and therefore embrittlement, the Cr coated zirconium has shown that it will provide significant improvement in the performance during normal operation, transients, design basis accidents and beyond design basis accidents as compared to uncoated zirconium. Similar tests were run with SiC at temperatures from 1600oC up to 1700oC. These tests were run with the graphite heater rod and were terminated only because of excessive corrosion of the heater rod. At 1600oC, the SiC was visually untouched. At 1700oC, there were indications of small beads on the surface (presumably SiO2 from the reaction of SiC with steam) but on the whole, no significant deterioration of the SiC. Changes are being made to the heating rod to increase the flow of He cover gas and to allow accurate weight changes to be made on the SiC rodlets so that kinetic data can be obtained. U3Si2 Testing U3Si2 has never been utilized as pellets inside cladding in Light Water Reactor (LWR) fuel service. There is a lack of data on the behavior of U3Si2 at LWR operating temperatures (estimated to be from 600oC and up to 1200oC during transients). To remedy this lack of data, U3Si2 fuel pellets were manufactured at Idaho National Laboratory (INL) and put into rodlets in the ATR in 2015. The first rodlets came out of ATR at the end of 2016 and are due for destructive post irradiation examination in the summer of 2017 at INL [2]. Preliminary nondestructive testing from neutron radiography of the U3Si2 Pins after exposure of 20 MWd/kgU in the ATR shows very good results with a lack of pellet cracking and distortion. U3Si2 was tested for air and steam oxidation as compared with UO2 using digital scanning calorimeters at both the Westinghouse Columbia facility [3] and at Los Alamos National Laboratory (LANL) [4]. The Westinghouse results indicate that the ignition temperatures for UO2 and U3Si2 are between 400oand 450oC (Table II). The LANL results (Table II) indicates an ignition temperature of about 400oC. The reasons for this difference are being studied. The reactivity of U3Si2 and UO2 are comparable at normal operating conditions (320oC), though the heat generated and mass generated by the oxidation of the U3Si2 is considerably higher than for UO2 at higher temperatures. The effect of this difference in heat release and mass on the stability of the rods is being investigated in rodlet tests in the Churchill autoclaves in the summer of 2017. However, the risk of any reactions between the coolant and the U3Si2 is probably much less than the current 1 to 2 ppm due to current rod failures, because the ATF claddings tend to be much harder than zirconium alloys. Therefore, it is expected that grid to rod fretting leakages will be eliminated. Finally, LANL identified the potential for the formation of a U3Si2-H1.8 compound in the event of a leaker. Further work reported by S. Mašková et al [5] indicated that this would not likely be an issue since the operating temperature of the U3Si2 fuel will be above the decomposition temperature (~550oC) of this compound.
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
5
Table II – Comparison of Westinghouse U3Si2 and UO2 Scanning Calorimeter Results (Heating Rate 2.5°C/min) Material Maximum ͇mass (%) On-set Oxidation T (oC) Reaction enthalpy (-J/g)
UO2 (Air) 4.1% ~410 ~200
U3Si2 (Air) 25% ~450 ~3200
U3Si2 (Steam) 25% ~500 Not available
CONCLUSION The Westinghouse ATF concepts appear to be technically achievable as LTRs and LTAs in the 2019 and 2022 timeframe. Performance issues with SiC, coated cladding, and U3Si2 fuel have been identified and overcome through modifications, engineering, and testing. As with any revolutionary new product, technical challenges may surface, but the robust research and development program that Westinghouse has in place will be used to overcome these challenges. REFERENCES [1] Ed Lahoda, Sumit Ray, Frank Boylan, Peng Xu and Richard Jacko, “SiC Cladding Corrosion and Mitigation,” TopFuel 2016, Boise, Idaho, Paper Number 17450, September 10, 2016. [2] Jason Harp, Idaho National Laboratory, private communication, preliminary examination. . [3] Lu Cai, Peng Xu, Andrew Atwood, Frank Boylan, Edward J. Lahoda, “Thermal Analysis of ATF Fuel Materials at Westinghouse,” ICACC 2017, Daytona Beach, Florida, January 26, 2017. [4] E. Sooby Wood, J.T. White, A.T. Nelson, “Oxidation behaviour of U-Si compounds in air from 25 to 1000 C,” Journal of Nuclear Materials, 484 (2017) pages 245-257. [5] S. Mašková, K. Miliyanchuk, L. Havela, “Hydrogen absorption in U3Si2 and its impact on electronic properties”, Journal of Nuclear Materials, 487 (2017) pages 418-423. ACKNOWLEDGEMENT This material is based upon work supported by the Department of Energy under Award Number DE-NE0008222. This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
6
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
Ste
INVESTIGATIONS OF CAPACITOR DISCHARGE WELDING FOR THE ATTACHMENT OF ENDCAPS TO MOLYBDENUM-BASED NUCLEAR FUEL ROD CLADDING Jerry E. Gould1, Cem Topbasi2, and Bo Cheng2 1 EWI Columbus, OH, U.S.A. 2 EPRI Palo Alto, CA, U.S.A. ABSTRACT In response to the Fukushima nuclear power plant disaster there has been strong interest in fuel rod claddings with enhanced survivability during cooling loss. A strong candidate for such cladding is molybdenum. The high thermal stability of the material is key to improving cladding performance. A challenge to the use of such claddings is providing end sealing once the fuel is loaded. Molybdenum is susceptible to excessive grain growth during welding, resulting in a loss of ductility and toughness. Such grain growth can be minimized (or eliminated) by reducing both bonding temperature, as well as associated exposure times. Capacitor discharge (CD) welding offers significant potential for both. First, it is a solid-state process, functioning below the melting temperature of the substrates and, in addition, has implicit heating times in the range of milliseconds. In this study, CD welding was employed for attaching a range of endcap materials to molybdenum tubes. Endcap materials included Zircaloy-4, Kanthal, and molybdenum. Joint preparation was for an “edge type” weld, employing a taper on the endcaps and the tube wall as the projection. In these trials, relative weldability improved as the end caps were changed from Zircaloy-4 to Kanthal to molybdenum. Poorest combinations (Zircaloy-4 caps) showed excessive deformation of the endcap, with none on the edge of the tube wall. Best combinations (molybdenum endcaps) showed forging of the tube wall end into the opposing substrate. Best-identified practices included welding with molybdenum endcaps, use of a rapid (2.5 ms) current pulse, and a low relative welding force (2.2 kN). Resulting joints showed a solid-state character, no apparent grain growth, and leak tightness. INTRODUCTION On March 11, 2011, the nuclear power plant in Fukushima was disabled by a major tsunami1. During the flooding associated with that tsunami, the facility lost all power resulting in a failure of the cooling systems for the reactor core. The remaining water in the core boiled and the resulting steam reacted with the zirconium-based cladding with additional heat generation, eventually resulting in melting of the core itself1 Following this event, there has been a concerted industry effort to develop fuel cladding with enhanced survivability compared to the existing Zircaloy-4-based technology. Much of the focus has been on developing refractory metal-based cladding2. The material of choice has been molybdenum. Molybdenum is a refractory metal, with a melting point of roughly 2600°C (compared to 1700°C for Zr) that can maintain its mechanical integrity to over 1500°C. Molybdenum, however, has several issues that must be addressed for use in this application. First, molybdenum has a neutron cross section considerably larger than that of Zirconium and comparable to that of steel3. This size difference affects the efficiency of the reactor, as well as can result in swelling of the clad itself. Also, molybdenum has corrosion concerns in the cooling water-based environment and is susceptible to oxidation at higher temperatures2,4. To address these concerns, a new generation, multi-layer cladding approach has been developed4. This cladding includes an interior layer of molybdenum, a thin transition layer of niobium, and an external layer of zirconium. The resulting product offers the opportunity for enhanced thermal
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
7
stability (implying survivability), as well as environmental compatibility in the operating environment. Fabrication of fuel rods, of course, requires attachment of endcaps. For this new generation of claddings, it is implied that the endcap itself is capable of being joined to the hightemperature stability (molybdenum) interior layer. The creation of joints, including a molybdenum element, is generally considered problematic. Refractory metals (including molybdenum) have a body-centered cubic crystal structure that remains unchanged as the material is heated to the melting point. As a result, during welding these materials can experience extensive grain growth. This grain growth has been related to increases in the ductile-to-brittle transition temperature, with associated loss of ductility5,6. Molybdenum is also seen to be incompatible with other candidate materials for fuel rod construction7. For example, with respect to zirconium, a eutectic can be formed resulting in a 300°C melting point suppression. This can result in liquation-related cracking of the resulting joint. In addition, on cooling, this combination can also result in Laves phase formation embrittling the joint. Kanthal is also a candidate material for fuel rods. This is an aluminum containing stainless steel designed for thermal stability. However, in attachment to molybdenum many intermetallic phases can form, potentially embrittling the joint. Previous work has shown that solid-state processes can be an effective approach for the joining of molybdenum alloys8. Solid-state joining methods include technologies such as friction, flash-butt, and upset-butt welding. These processes offer some unique advantages compared to fusion technologies for joining refractory metals. These include lower (below melting) processing temperatures, deformation during joining, and relatively short thermal cycles. CD welding fits into this class of solid-state processes9. This is a resistance-based process, using current flow through the workpieces to generate heat. That current is provided by one or more capacitors charged to relatively high (up to 3000 V) voltages. The current discharge itself is designed to be very rapid. Typical CD welding times can range from a few up to about 100 milliseconds. CD welding is typically done in specially designed load frames. Essentially, the workpieces are placed under an applied force and the current discharged through the stackup. The rapid heating results in thermal softening of the substrates, with subsequent forging to create a joint. It is of note that the process is mechanically, as well as electrically, dynamic (the workpieces are forged together)9. Therefore, the system must be designed to match electrical and mechanical response for optimum joining. In this work, CD welding has been applied for attaching endcaps to specimens representative of candidate molybdenum fuel rods. Work was conducted with three types of endcaps. These included those made from Zircaloy-4, Kanthal, and molybdenum. Work has included design of the joint geometry itself, tooling development, welding trials, and quality assessments. EXPERIMENTAL PROCEDURES As suggested above, endcaps for study were made from Zircaloy-4, Kanthal, and molybdenum. The tubes themselves were from molybdenum. Nominal compositions for each material, based on standard references, are provided in Table I10-11. Tubes for joining included a nominal diameter of 9.5 mm, and a wall thickness of 210 ͮm. Tubes were cut to a nominal length of 22 mm for use as samples. Endcaps were all nominally 9.3 mm in diameter and 3.8-mm thick. All endcaps included a 45-degree taper over the last 1.3 mm of length on one end. This was done to facilitate an edge projection-type joint configuration. A sample endcap with taper is shown in Figure 1. The actual joint geometry prior to welding is provided in Figures 2 and 3. This joint
8
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
geometry is referred to as an “edge projection weld”12. It is commonly used in resistance welding for tube sealing applications. Table I. Representative Compositions for the Endcap and Tube Materials Used in this Study Material Appl. %Zr %Cr %Fe %Sn %Al %Mo %O Source Zircaloy-4 Cap Bal. 0.10 0.20 1.4 0.12% Ref. 10 Kanthal Cap 21 Bal. 5.0 3.0 Ref. 11 Molybdenum Cap Bal. Ref. 12 Molybdenum Tube Bal Ref. 12
Figure 1. A sample endcap with the prepared joint geometry.
Figure 2. Schematic representation of the joint geometry used. Detail “E” of the joint area is provided in Figure 3.
Figure 3. Details of the interaction between the tube edge and endcap. Welding itself was performed on a dedicated pedestal-type CD welder. The power supply included a 1280-ͮF capacitor that could be charged up to 2500 V. This setup yields a peak stored
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
9
energy value of 4 kJ. To accomplish circumferential welding of these thin-wall tubes, special tooling was required that included a collet-type lower tool to restrain the tube (Figure 4). For welding, the tube was placed nominally flush with the end of the collet. An internal aluminum solid support was used to prevent the collet from crushing the tube. As described above, the mechanical response of the system must be matched to that of the current delivery. In this work, a specially designed fast follow-up head was employed. The final configuration, including fast follow-up head and lower tooling, is shown integrated into the welding frame in Figure 5.
Figure 4. The collet-type tooling used to restrain the Mo tubes for welding.
Figure 5. Fast Follow-up head and lower tooling in the welding frame. For these trials, instrumentation was provided for monitoring the secondary current and weld head displacement. Current was monitored with the analog output of a Miyachi Weldchecker MM-326B. Displacement was monitored with a Capacitec 4100-SK position sensor. All process data was collected at a 20-kHz sampling rate on a recording oscilloscope. Weld quality was evaluated using visual and metallographic means, as well as through helium leak testing. Visual assessments were used to assess evidence of bonding, as well as any macroscopic changes to the samples. Metallographic evaluations were conducted on a Nikon metallograph. Samples were prepared using standard techniques and assessed with and without etching. Finally, helium leak testing was performed with a Veeco MS 40 system, with detectable rates < 10-10 Torr. RESULTS Initial welding trials were conducted using the Zircaloy-4 endcaps. Welding was performed using two applied forces (2.2 and 3.3 kN), a range of charge voltages, and three transformer turns ratios (75:1, 139:1, and 213:1). The third (transformer turns ratios) provides
10
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
differing durations of applied power, while the second (charge voltage) controls the amount of current itself. No significant bonding was seen with these caps. Some representative process monitoring data is provided in Figures 6 and 7. These represent trials at the 3.3-kN force, using the two extremes of transformer turns ratios (75:1 and 213:1), respectively. The trial in Figure 6 showed a rise time and peak current of 2.6 ms and 23 kA. The trial for Figure 7 showed a much slower rise time (6.7 ms) and a peak current of 15 kA. All welds showed significant set-down of the cap against the tube. For these trials, the set-downs measured roughly 400-500 ͮm.
Figure 6. Current and displacement Profiles for welding a Zircaloy-4 endcap to a moly tube using rapid discharge. Weld was made with an 800-V charge voltage, 75:1 turns ratio, and 3.3-kN force.
Figure 7. Current and displacement profiles for welding a Zircaloy-4 endcap to a moly tube using slow discharge. Weld was made with a 1000-V charge voltage, 213:1 turns ratio, and 3.3-kN force. Corresponding metallographic data from representative welds is provided in Figures 8 and 9. These are lower (2.2 kN) force trials, again conducted at two different transformer turns ratios (75:1 and 213:1). In both cases, it appears that the measured upset distances were accomplished as the tube end displaced cap material. In neither case is there any apparent local deformation of the tube end itself. In addition, no bonding was evident. Of note, the faster rise time weld showed some flaring of the tube wall that was not apparent at the slower rise times (213:1 windings ratio).
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
11
Figure 8. Overview of the Zircaloy-4 endcap to a moly tube weld using rapid discharge. weld was made with an 800-V charge voltage, 75:1 turns ratio, and 2.2-kN force.
Figure 9. Overview of the Zircaloy-4 endcap to a Moly Tube weld using slow discharge. Weld was made with a 1000-V charge voltage, 213:1 turns ratio, and 2.2-kN force. Trials using the Kanthal endcaps examined a range of forces, transformer windings ratios, and capacitor charge voltages. Generally, these trials differed from those with the Zircaloy-4 in several ways. First, successful welding was possible at both forces (2.2 3.3 kN) and transformer windings ratios (75:1 and 139:1). Second, observed displacements during the trials were significantly less (30-50%) than seen with the Zircaloy-4 endcaps. Third, currents (for a given charge voltage and windings ratio) were slightly higher than those seen with the Zircaloy-4 endcaps. Of note, current rise times for a given transformer windings ratio were similar. Selected monitoring results from the Kanthal endcap trials are provided in Figures 10 and 11. The waveforms shown are taken from welds made using 2.2-kN force and two different transformer windings ratios. Current waveforms are similar, but with higher peak values. Here, peak currents are seen to be roughly 30 kA and 25 kA for the 75:1 and 139:1 windings ratios, respectively. As mentioned, rise times for the two different transformer configurations were similar to that seen for the Zircaloy-4 endcaps, at 2.45 ms and 4.44 ms for the 75:1 and 139:1 windings ratios, respectively. Displacement traces in these plots suggest the reduced total upsets, here at 300-400 ͮm for the two welds shown.
12
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
Figure 10. Current and displacement profiles for welding a Kanthal endcap to a Moly tube using rapid discharge. Weld was made with an 1100-V charge voltage, 75:1 turns ratio, and 2.2-kN force.
Figure 11. Current and displacement profiles for welding a Kanthal endcap to a Moly tube using slow discharge. Weld was made with a 1200-V charge voltage, 139:1 turns ratio, and 2.2-kN force. Associated metallography for a representative Kanthal endcap weld is provided in Figures 12 and 13. The cross-section overview is given in Figure 12. Here it can be seen that there is still significant deformation of the Kanthal endcap relative to the tube wall. However, there is significant forging of the tube end where it contacts the Kanthal. The higher magnification microscopy provided in Figure 13 provides some indication of bond quality at this contact point. Here it can be seen that the joint is fully solid state, with intimate contact of the two substrate materials, and no indication of mixing or intermetallic formation.
Figure 12. Overview of the Kanthal endcap to a Moly tube weld using slow discharge. Weld was made with a 1200-V charge voltage, 139:1 turns ratio, and 3.3-kN force.
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
13
Figure 13. Bond line details of the Kanthal endcap to a Moly tube weld using slow discharge. Weld was made with a 1200-V charge voltage, 139:1 turns ratio, and 3.3-kN force. Trials using the molybdenum endcaps were only conducted using the fast rise time turns ratio on the transformer (75:1). All current rise times were in the range of 2.5 ms. Of note, higher voltages (up to 1600 V) were required for welding the molybdenum endcaps. In addition, higher currents were noted, here up to nearly 60 kA. It is of interest that welding between the endcap and tube occurred over a range of processing voltages/currents, at both the high and low force levels used. Further, measured upset distances were very low, ranging from 50-150 ͮm over the trials that showed successful welding. In these trials, one sample each from the high and low force conditions was also submitted for helium leak testing. A typical process waveform is provided in Figure 14. The rise time on the current waveform is rapid, consistent with the 75:1 windings ratio used. The observed peak current is relatively high at 55 kA. Also, the displacement trace showed little total movement, here only about 50 ͮm. Of interest is the apparent ringing in the displacement trace. This is largely an artifact of the small overall upset distances observed in these trials, with only 50-150-ͮm movement; the inherent stiffness of the displacement transducer mount allowed some ringing. This ringing appears to be on the scale of these small overall displacements.
Figure 14. Current and displacement profiles for welding a Molybdenum endcap to a Moly tube using rapid discharge. Weld was made with a 1550-V charge voltage, 75:1 turns ratio, and 3.3-kN force. As mentioned, two of these specimens were subject to helium leak testing. The leak test configuration is shown in Figure 15. Here, a vacuum is drawn on the inside of the tube and helium dispersed around the circumference of the weld. The welds made at 2.2 kN and 3.3 kN
14
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
showed leak rates of 2.2 × 10-10 Torr and 2.5 × 10-10 Torr, respectively. These leak rates are at the sensitivity limit of the testing device.
Figure 15. A welded assembly including a Molybdenum tube and endcap fixtured for helium leak testing. Duplicate samples at differing weld forces both showed roughly 2.5 × 10-10 Torr leak rates. The associated metallography of a successful weld is provided in Figures 16 and 17. Figure 16 provides an overview of the joint between the molybdenum tube wall and end cap. Contrary to welds made with the other endcap materials, there is significant deformation to the end of the tube wall and almost none of the endcap itself. Further, there is apparent solid-state bonding along the complete length of the tube end/endcap contact length. Finally, there appears to be little or no grain coarsening between the base metal and weld region in either the tube or endcap. A higher magnification view of the bond area itself is provided in Figure 17. In this micrograph, the tube is on the left side and the end cap on the right. Here it can be seen that there is first realignment of the grain texture to parallel along the bond line near the faying surface. In addition, the bond line itself shows no evidence of any defects. Finally, a fine grain size has been retained throughout the weld, typically less than 5 ͮm.
Figure 16. Overview of the Moly endcap to a Moly tube weld using rapid discharge. Weld was made with a 1550-V charge voltage, 75:1 turns ratio, and 3.3-kN force.
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
15
Figure 17. Bond line details of the Moly endcap to a Moly tube weld using rapid discharge. Weld was made with a 1550-V charge voltage, 75:1 turns ratio, and 3.3-kN force. DISCUSSION From the data presented above, it is apparent that the three endcap materials represent widely different responses to CD welding. Molybdenum endcaps demonstrated excellent weldability, forging of both substrates, apparent solid-state bonding, leak tightness, and little or no grain growth. Kanthal endcaps demonstrated some weldability with limited deformation of the molybdenum tube wall and apparent solid-state bonding of the substrates. However, deformation was highly biased toward the Kanthal. Zircaloy-4 endcaps were not found to be weldable to the molybdenum tubes. Results indicated no deformation in the moly tube wall. Rather, the Zicaloy-4 was continuously deformed by the advancing molybdenum tube end, resulting in no evidence of solid-state bonding. This difference in CD weldability (and solid-state weldability) is largely related to the relative forging of the substrates during processing. For effective bonding, strain is required on both substrate contacting surfaces. During CD welding, deformation occurs as the substrates soften with increasing temperature. Figure 18 provides a plot showing the yield strength of molybdenum, Kanthal, and Zircaloy-4 as a function of temperature above 1000 K (727°C). The data here is taken from standard references11,13,14. On this plot, molybdenum maintains >50-MPa yield strength, well above the point where either the Kanthal or Zircaloy-4 have any structural integrity. Of interest is that below roughly 1230 K (957°C), the Zircaloy-4 is stronger than the Kanthal. Above this temperature, however, Kanthal maintains higher yield strengths. The implication here is that this hierarchy of strengths at high temperature progressively increases achievable bonding temperatures. The higher the bonding temperature, the more likely for forging of the molybdenum tube and potential for subsequent solid-state joining. That is certainly borne out in the experimental results.
Figure 18. Variations in yield strength as a function of temperature for the three endcap materials of interest. Data has been taken from standard references10-13.
16
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
As mentioned above and described elsewhere9, the electrical and mechanical response of CD systems must be matched to prevent runaway thermal response and weld metal expulsion. As suggested above, material strength and joint temperature are related. Therefore, for a fixed geometry, the applied force is directly related to the maximum temperature in the weld. As the current (and temperature) rise, the metal will soften and begin to forge. To maintain contact force, the weld head must accelerate at the rate of softening. If not, the effective contact force will decrease and the joint temperature will continue to rise. In the extreme, the melting temperature of one of the substrates can be reached and weld metal expulsion occurs. This expulsion results in localized weld defects, as well as a loss of sealing capability. Previous work9 has established a criterion for ensuring proper system mechanical response. This criterion defines the relationship between the applied force, weight of the moving weld head, current rise time, and observed upset distance for stability. Stability here is defined as maintaining 90% of the welding force through the weld cycle. That criterion is expressed as: (1) where xcollapse is the upset distance, trise is the current rise time, g is the acceleration due to gravity, Fapp is the applied force, and Whead is the weight of the welding head. In this effort, the mechanical response limit is clearly lower (reduced allowable collapse distances) for the 2.2-kN weld force. The relationship described in Equation 1 is then presented in Figure 19 as critical combinations of current rise times and allowable upset distances. Combinations to the left of this line in Figure 19 are mechanically unstable (potentially leading to expulsion), while those to the right are considered stable. The data from all the trials completed at the 2.2-kN weld force are also included on this plot. This includes data from the three endcap material types and the three transformer windings ratios used. In all cases, the actual trials were mechanically stable, even those with the Zircaloy-4 endcaps that showed excessive upset distances.
Figure 19. Mechanical stability limit for the 2.2-kN welding trials completed in this study. The plot includes data for all endcap types and transformer windings ratios. The improved matching between the high-temperature strengths of the substrates described above also was seen to increase effective welding ranges for the various endplug types. This is demonstrated in Figures 20 and 21. These two figures provide plots of upset distance as a function of current for welds made with the three endcaps at conditions of 75:1 turns ratio (nominal 2.5-ms rise time). Figure 20 provides the 2.2-kN data, and Figure 21 that for 3.3-kN welding force. Conditions that resulted in effective welding for each end cap type are circled
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
17
separately. Three aspects of the plot are immediately clear. First, the material combinations that showed effective bonding at lower upset distances had the widest effective welding (current) ranges. This is not surprising based on the discussion above. Better-matched strengths at temperature result in balanced forging across the faying surface, thus minimizing deformation of the end caps (and subsequent upset) necessary for bonding to occur.
Figure 20. Current range for the three endcap materials at 75:1 turns ratio and 2.2-kN weld force.
Figure 21. Current range for the three endcap materials at 75:1 turns ratio and 3.3-kN weld force. CONCLUSIONS This work has investigated the use of CD welding for attaching endcaps to molybdenumbased nuclear fuel rods. Work was performed with thin walled molybdenum tubes attached to end plugs of an array of material types. Endcap materials included Zircaloy-4, Kanthal, and molybdenum. Joint geometry was designed such that the interface between the endcap and tube end created an edge projection. This projection allowed creation of a circumferential seal weld. For each end plug type, trials included variations in capacitor charge voltage, transformer turns ratio, and welding force. Results showed that the molybdenum end plugs readily bonded to the tube end. Successful joints were also made with the Kanthal endcaps. However, these trials resulted in significant deformation of the endcap itself. Finally, it was not possible to bond the Zircaloy-4 to the molybdenum tubes. Variations in weldability were associated with differences in yield strength at temperature. Mismatching between the endcap and tube yield strengths at elevated temperatures led to progressively unequal forging and increasing difficulty in making effective joints. Specific conclusions from this work include: (1) Capacitor discharge welding demonstrated the ability for rapid thermal cycles and the potential for localized forging: Thermal cycles ranging from 5 up to 30 ms were achieved, creating localized deformation at the endcap/tube end interface. (2) The capacitor discharge welding system was configured to prevent thermal-mechanical instabilities during creation of joints: The system design was based on previous work
18
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
defining requirements for weld head weight based on current rise times and planned upset distances. (3) Zircaloy-4 endcaps were not readily weldable to the molybdenum tubes: Joining trials resulted in excessive deformation of the endcap and none on the tube wall end. (4) Kanthal endcaps were successfully joined to the molybdenum tube ends: Best-practice joints showed some deformation of the tube wall end and solid-state bonding to the Kanthal endcap. (5) Excessive endcap deformation was still observed on the Kanthal-to-molybdenum joints: This deformation was less than seen with the Zircaloy-4 endcaps. (6) Molybdenum endcaps were readily joinable to the molybdenum tube ends: Resulting joints showed forging of the tube end against the endplug, resulting in joints with minimal upset. (7) Little or no grain growth was seen in the molybdenum tube material throughout these trials: The rapid thermal cycles associated with CD welding appeared to prevent any significant grain growth in the molybdenum. (8) Weldability between the endcap and tube was closely related to matching the substrate material yield strengths at bonding temperatures: Mismatches in yield strengths resulted in uneven forging across the joint faying surface, lessening the potential for bonding. (9) Effective process operating ranges paralleled the relative mismatch in yield strengths at bonding temperatures: The widest current ranges were seen for the molybdenum end caps; none was experienced for those of the Zircaloy-4 material. (10) Best-identified practices included use of the molybdenum endplugs, rapid current rise times, and lower forces: Trials conducted with the 75:1 turns ratio transformer (2.5-ms rise time) and at the 2.2-kN weld force provided wide current ranges and consistent upset behavior. REFERENCES 1 Fukushima Nuclear Accident Analysis Report, Tokyo Electric Power Company, Inc. (2012). 2 B. Cheng, Y.-J. Kim, and P. Chou, Improving Accident Tolerance of Nuclear Fuel with Coated Mo-Alloy Cladding, Nuclear Engineering and Technology, 48, 16-25 (2016). 3 B.V. Cockeram, R.W. Smith, K.J. Leonard, T.S. Byun, and L.L. Snead, Irradiation Hardening in Unalloyed and ODS Molybdenum during Low Dose Neutron Irradiation at 300°C and 600°C, Journal of Nuclear Materials, 382, 1-23 (2008). 4 S. Bragg-Sitton, Development of Advanced Accident-Tolerant Fuels for Commercial LWRs, Nuclear News, 57(3), 83-91 (2014). 5 J.C. Thornley and A.S. Wronski, The Relation between the Ductile-Brittle Transition Temperature and Grain Size in Polycrystalline Molybdenum, Scripta Metallurgica, 3, 935-38 (1969). 6 A.S. Wronski, A.C. Chilton, and E.M. Capron, The Ductile-Brittle Transition in Polycrystalline Molybdenum, Acta Metallurgica, 17, 751-55 (1969). 7 H. Okamoto, Phase Diagrams for Binary Alloys Desk Handbook, ASM International, Materials Park, OH (2000). 8 A. Ambroziak, Friction Welding of Molybdenum to Molybdenum and to Other Metals, International Journal of Refractory Metals and Hard Materials, 29(4), 462-469 (2011). 9 J.E. Gould, Mechanisms of Solid-State Resistance Welding, ASM Handbook Volume 6A: Welding Fundamentals and Processes, ASM: Metals Park, OH, 171-178 (2011). 10 Metals Handbook, Vol.2 - Properties and Selection: Nonferrous Alloys and SpecialPurpose Materials, ASM International 10th Ed., Metals Park, OH (1990).
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
19
11 12 13
14
20
Kanthal APMT Construction Materials – Datasheet, Sandvik Corporation, (2017). Resistance Welding Manual, RWMA: Philadelphia, PA (2003). K.J. Geelhood, C.E. Beyer, and W.G. Luscher, PNNL Stress/Strain Correlation for Zircaloy (No. PNNL-17700), Pacific Northwest National Laboratory (PNNL), Richland, WA (U.S.) (2008). Refractory Metals and Their Industrial Applications, Robert E. Smallwood, ed. ASTM, Philadelphia, PA (1964).
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
EVALUATION OF U3Si2 FUEL PELLETS SINTERED IN AN ARGON VS. VACUUM ENVIRONMENT Rita Hoggan, Jason Harp, and Lingfeng He Idaho National Laboratory Idaho Falls, Idaho, USA ABSTRACT A comparison study on the effects of sintering U3Si2 pellets under vacuum (~1x10-6 mbar) vs in argon has been carried out as part of a pellet fabrication process for potential light water reactor fuel application. A cross section of pellets sintered in each environment was examined using scanning electron microscopy (SEM) equipped with energy dispersive spectroscopy (EDS). The average density of pellets sintered under vacuum and argon was 11.90 g/cm3 (97.5% theoretical density (TD)) and 11.37 g/cm3 (93.2% TD), respectively. SEM micrographs and EDS revealed the grain structure of pellets sintered under vacuum (vacuum samples) to be angular with a Si rich phase, U5Si4, along the grain boundaries, while grain structure of pellets sintered in argon (argon samples) reveals smaller average grain size and higher porosity with two separate Si rich phase inclusions: U5Si4 and USi. EDS confirmed the main phase in each pellet to be U3Si2 with an O rich phase inclusion. The estimated area percent of U3Si2 and the O rich phase in the samples from SEM image analysis was: argon samples 83.9% and 5.9% and vacuum samples 81.8% and 4.6%, respectively. INTRODUCTION Uranium silicide (U3Si2) fuels with high thermal conductivity are being studied as alternatives to currently used uranium dioxide. U3Si2 pellets were produced by powder metallurgy techniques as part of an accident tolerant fuel concept by Harp et al.1 They reported sintering pellets in both argon and vacuum environments with differences in average sintered pellet density as a result: 11.57 g/cm3 (94.7% theoretical density (TD)) and 11.8 g/cm3 (96.9% TD) respectively. This paper is an evaluation of the microstructure, phase composition, and related density measurements of pellets sintered in each environment. METHODS A batch of pellets fabricated by the process reported by Harp et al.1 were prepared for sintering with three pellets randomly sampled for sintering in each environment: Argon and Vacuum. Three sample pellets were used in each environment as that was the standard sintering batch size according to the size of the tantalum crucible used for all sintering. The argon samples were sintered in a high temperature CM box furnace inside an argon atmosphere glove box with an approximate O2 impurity content of up to 40 ppm. Pellets were sintered for 4 hours at 1500 °C. The vacuum samples were sintered in an RD Webb “Red Devil” graphite furnace operated under vacuum after an argon backfill and evacuation to ~1x10-6 mbar. The same heating schedule for sintering in each environment was followed which included a 2 hour hold at 600 °C, 4 hours at 1500 °C, and allowed to naturally cool. After sintering, the pellet densities were measured via helium gas displacement pcynometry (Quantachrome Micro-Ultrpyc). One pellet sample from each sintering environment was radially sectioned in a Struers precision saw with a silicon carbide blade, mounted in epoxy, and polished to a 1μm finish for examination with a JOEL 7600 Field Emission SEM equipped with EDS. An average of EDS point spectra was used to determine the phases present and validated with XRD results reported previously by Harp et al.1
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
21
Image analysis on a combination of SEM backscattered electron (BSE) and secondary electron images (SEI) micrographs to the intent of ASTM E 1245-03 to estimate the area fraction of the identified phases and porosity. Three fields from each sample were measured at 250×, 500×, and 1000× magnification. A grey level threshold setting was applied for each phase and voids using image analysis software, Image J. RESULTS A visible difference in the argon and vacuum samples were observed after sintering. Fig. 1 shows a representative argon and vacuum as-sintered pellet sample. The argon samples had an appearance of localized surface oxidation compared to the shiny pellet surfaces of the vacuum samples.
Figure 1. Sintered pellet samples.
22
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
Figure 2. SEM images comparing vacuum and argon sintered sample porosity and microstructure. SEM images in Fig. 2 show a comparison of the porosity and microstructure of the samples sintered in argon and vacuum. The average density of the argon and vacuum samples was 11.90 ± 0.07 g/cm3 (97.5% theoretical density (TD)) and 11.37 ± 0.02 g/cm3 (93.2% TD), respectively. The density measurements were in line with those reported by Harp et al. and in the introduction of this paper. Indication of a difference in density is illustrated in the SEI micrographs where the black regions indicate porosity or voids. From image analysis the estimated area percent of porosity/voids was 5.2 ± 1.0 % and 0.27 ± 0.06 %, respectively, for argon and vacuum samples. From the author’s qualitative experience, and also illustrated here, denser samples are less susceptible to polishing induced pullout, indicated by arrows on the SEI argon image. The grain structure in the BSE vacuum sample, Fig. 2, appears more angular and the grains generally larger than the argon sample. The larger observed grain size in the vacuum sample is likely a result of the slower cooling time - allowing a longer time for grain growth relative to that of cooling in argon gas where the gas acts as a mechanism for the samples to cool faster - allowing a shorter time for grain growth.
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
23
Figure 3. SEM images with identified phases for argon and vacuum sintered samples. In higher resolution SEM images, Fig. 3, the different phases are more clearly distinguished in a combination of SEI and BSE images. An average of several EDS point spectra taken over the different shaded regions is provided in Table I. The phases presented in this work have been validated or indicated by previous X-ray diffraction (XRD) data reported by Harp et al. The matrix in each sample is a combination of the two lightest contrast shades in the images. The different shades are a result of crystal orientation and in this case do not indicate differing crystal structure or chemical composition. The matrix in each sample according to table one corresponds to U3Si2. A black oxygen rich phase is seen in the BSE images for both samples and is only distinguished from pores and voids by comparing the SEI and BSE images. Table I indicates this black contrast phase is UO2 A light grey phase, corresponding to U5Si4 according to Table I, appears to have formed preferentially along U3Si2 grain boundaries in the vacuum sample. Two different intermediate grey contrasts shades were observed in the argon sample, likely U5Si4 and USi based on the data in Table I, with the darker contrast corresponding to a higher Si content. Table I. Average EDS point spectra reported in wt. % for identified contrasts/phases. Argon Sample Vacuum Sample descript. black D. grey L. grey matrix black grey Name UO2 USi U5Si4 U3Si2 UO2 U5Si4 Si wt. % 10.91±0.03 9.17±0.03 7.98±0.03 9.00±0.02 U wt. % 87.3 89.10±0.04 90.83±0.03 92.03±0.03 86.7 91.00±0.02 O wt. % 12.7 13.27
24
matrix U3Si2 7.83±0.02 92.17±0.02
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
The crystallographic information for U5Si4 was reported, by Noël et al.2, to exhibit the hexagonal space group P6/mmm. The presence of a U5Si4 phase has not been confirmed by crystallographic data in the specific sintering environment samples evaluated in this study, but previous XRD spectra shown in Fig. 4, on similar samples investigated by Harp et al.1, in retrospect, implies the existence of this U5Si4 phase.
Figure 4. XRD spectra illustrating U5Si4 phase. Fig. 5 shows and enlarged center cut out of the SEI vacuum image in Fig. 3. The oxide phase is more prominently observed in the SEI vacuum image compared to the argon image. That is likely a result of the contrasting during image collection. An iron rich phase was identified with arrows indicating its location in the sample. The iron is thought to be a contaminant in the original uranium feedstock. It is often observed as a small precipitate connected to an oxide precipitate. This iron phase is likely present in both argon and vacuum samples, but was observed only in the vacuum sample.
Figure 5. SEM SEI image of vacuum sample identifying iron rich phase.
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
25
The estimated area fraction for all of the phases identified in each sample is presented in Table II. This data corresponds to the percent of identified phases reported by Harp et al.1 with the major phase: U3Si2, 86 ± 2%; minor phase: USi, 10.5 ± 2.5%; and minor phase: UO2, 3 ± 1%. Table II. Estimated area fraction of phases present via SEM image analysis. Argon Sample voids black D. grey L. grey matrix UO2 USi U5Si4 U3Si2 5.2±1.0 5.9±1.2 4.4±1.0 5.6±1.9 84.0±1.3 Vacuum Sample voids Fe black grey matrix U-Fe-Si UO2 U5Si4 U3Si2 0.3±0.1 0.50±0.11 4.6±0.3 13.9±1.2 81.1±1.1 CONCLUSION A systematic evaluation of the microstructure of U3Si2 pellets sintered in argon compared with U3Si2 pellets sintered under vacuum has been carried out. The major difference in argon samples and vacuum samples is the greater density achieved from a vacuum sinter. This was illustrated in a comparison of the porosity in each sample, and a much higher calculated percentage of voids in the less dense argon sample. The overall U3Si2 phase purity was similar for each sample, ranging between 80 and 85.6 volume percent U3Si2. Other observed differences included the grain structure: shape and size, where the argon sample grains were generally smaller and irregularly shaped with the Si rich phases sporadically included, and the vacuum samples grains were angular shaped with the Si rich phase forming along grain boundaries. A refinement of characterization on the USi phases reported by Harp et al.1 determined in fact, a stoichiometry closer to U5Si4 in both samples and in the argon sample, a combination of USi and U5Si4 - a recently reported phase in the U –Si binary system. An iron rich contaminant phase, not previously reported, was observed in the vacuum sample. Further investigation on this phase is required to determine composition. Other phases identified in the present work and their total percent in the sample correspond to the data reported previously by Harp et al.1 REFERENCES J.M. Harp, P.E. Lessing, and R.E. Hoggan, Uranium Silicide Pellet Fabrication by Powder Metallurgy for Accident Tolerant Fuel Evaluation and Irradiation, J. Nucl. Mater., 466, 728-738 (2015). 2 H. Noël, S. Chatain, T. Alpettaz, C. Guéneau, C. Duguay, and J. Léchelle, Experimental Determination of (U-Si-C) Ternary Phase Diagram at 1000°C and Experimental Points in the Quaternary (U-Pu-Si-C) System, CNRS (2012). 1
26
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
1SPDFTTJOHBOE 1FSGPSNBODFPG .BUFSJBMT6TJOH .JDSPXBWFT &MFDUSJDBOE .BHOFUJD'JFMET 6MUSBTPVOE -BTFST BOE.FDIBOJDBM 8PSLo3VTUVN3PZ 4ZNQPTJVN
MICROWAVE-AUGMENTED CRYSTALLIZATION AND DECRYSTALLIZATION IN CERAMIC PROCESSING – A PHENOMENOLOGY-BASED COMMENTARY Boon Wong, Ph.D. Torrance, California, USA ABSTRACT Most crystallization and decrystallization in ceramic processing are diffusion-controlled, nucleation-and-growth processes. Under coherent, polarized and resonant microwave-irradiation (CPRMWI), the thermodynamics and kinetics of these processes could be significantly augmented. This type of MW-induced-process-augmentation is mainly caused by a remarkable freeenergy contribution from the time-averaged, molar resonant MW-work-input to the initialreactant state of the irradiated system. This free-energy contribution may then increase the driving force for the process and/or reduce the activation free energy for atomic/ionic mobility, thus effectively modifying the free-energy path of the process, and subsequently, promoting the isothermal process feasibility and/or kinetics under CPRMWI. Occurrence of this non-thermal microwave phenomenon (effect) in ceramic crystallization and decrystallization is well exemplified in many documented studies such as the accelerated isothermal crystallization of lithium disilicate (Li2O.2SiO2 or LS2) glass under a resonant, variable-frequency MWI and the rapid solid-state decrystallization (amorphization) of various crystalline 3d-ferrites irradiated under the separated, resonant H-field in a single-mode-MW cavity. The phenomenology-based concept and rationale proposed in this commentary may usher further discussions and research efforts towards the development of a working science-based guide for innovative applications of microwave-augmented crystallization and decrystallization in producing high-performance, engineered ceramics. INTRODUCTION In the history of technology evolution of mankind, engineering applications/practices had quite frequently advanced faster than their scientific rationale/formulation counterparts. This unique pattern of evolution seems to continue on in modern materials technology development. In advanced ceramic processing, for example, during the first decade of this century, microwaveaugmented crystallization and decrystallization processes were successfully applied to rapidly produce an engineered, fully-crystallized glass-ceramic at a relatively low temperature without the aid of hybrid heating,[1] and create innovative, bulk amorphous ceramics below their melting temperatures.[2] Yet, the underlying science required to guide further development of this novel processing technology is still quite lacking.[3] This report presents a phenomenological perspective to comprehend how, and more importantly, why process-augmentations may be realized and accomplished in crystallization and decrystallization under coherent, polarized and resonant microwave-irradiation (CPRMWI). The concept and rationale discussed in this article may initiate a fundamental understanding required for completing a working science-based guide usable for future development on microwaveaugmented crystallization and decrystallization in ceramic processing. MICROWAVE-AUGMENTED NUCLEATION AND GROWTH KINETICS – AN OVERVIEW A majority of the crystallization and decrystallization processes commonly encountered in the ceramic industry are envisioned as a two-step diffusion-controlled, nucleation and growth (N&G) process-sequence – starting with product-nuclei formation (“nucleation”), followed by subsequent product-phase growth (“growth”). Phenomenologically, as per classical kinetic theo-
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
29
ry and irreversible thermodynamics,[4],[5] rates of most diffusion-controlled nucleation or growth processes are governed by Exp. (1) – a generalized expression for the rate of any irreversible, steady-state transformation (transport) process, , resulting from one-dimensional uncoupled matter (mass) flux, driven by its dominant conjugate driving force: = {M}{)D}
where is the process-rate governed by the rate-determining matter flux, {M} is a mobility factor controlled by the rate-determining (matter) transport mechanism, and {)D} is a “drivingforce” function dominated by the driving force for the process: for scenarios of any system under constant temperature and pressure conditions, this driving force is the reduction in (total) molar free energy (chemical potential) of the system via the transformation (transport) process. The phenomenological rate law, as stated in Exp. (1), in the linear regime holds in many fields in classical physics, e.g., in electricity, it is known as the Ohm’s law; in diffusion, known as the Fick’s law; in convection, known as the Poiseuille’s law; and in solid-mechanics, known as the Nabarro-Herring equation for high-temperature, low-stress diffusional creep. Isothermal Microwave-Augmented Nucleation According to the classical nucleation theory and Boltzmann’s molecular energy distribution law, the steady-state nucleation rate per unit material volume at a given temperature under uniform MWI may be expressed as[4],[6] IEM = {M}{)D} ~ {exp[-Ƨ/(RT)]}{exp[-(NA'F*)/(RT)]}
(2)
where ~ is a mathematical symbol denoting “proportional to”. The first term,{exp[-Ƨ/(RT)]}, on the RHS of Exp. (2) expresses the temperaturedependent characteristic of the mobility factor, {M}, which governs the diffusion kinetics of additional parent (reactant) atoms (ions) crossing the parent-to-product interfaces into the newlyformed, critical-sized product-nuclei for their subsequent growth under CPRMWI; where Ƨ is the activation free energy (per mole) for the MW-augmented process, and RT is the product of the gas constant, R, and the thermodynamic temperature, T. Energetically,[4] Ƨ | (Æo - )
(2a)†
where Æo is the activation free energy (per mole) for the mobility governing the diffusion kinetics of reactant atoms (ions) crossing the parent-to-product interfaces under purely-thermal conditions, is the time-averaged, molar resonant MW-work-input to the (electrically and/or magnetically) polarizable reactant during the process under uniform CPRMW-field conditions. On the other hand, the second term, {exp[-(NA'F*)/(RT)]}, on the RHS of Exp. (2) expresses the key feature of the driving-force function, {)D}. Within this function, NA is the Avogadro’s number, 'F*, the nucleation free-energy barrier for every (spherical) critical-sized product-nucleus formation under MWI, is dominantly controlled by the driving force for the process, 'F < 0:[4] 'F* = (16SJ3)/[3('F/:)2] = (16SJ3:2)/[3('F)2] | (16SJ3:2)/[3('Go + - )2] = (16SJ3:2)/[3('Go + )2]
30
(2b)
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
where J is the specific surface free energy of the interfaces separating the product-nuclei and the parent-phase-matrix, : is the molar volume of the material. 'F, the driving force for nucleation (phase transformation) stated in Exp.(2b), may be evaluated based on the reduction in total EMenhanced molar free energy (“EM-chemical” potential) of the system via the process under CPRMWI at constant temperature and pressure:[4] 'F | ('Go + - ) = ('Go + ) < 0
(2bb)†
where 'Go is the change in intrinsic molar free energy (intrinsic chemical potential) of the process occurring under purely-thermal (zero-field, constant temperature and pressure) conditions, and , are the time-averaged, molar resonant MW-work-inputs to the initialreactant and final-product states of the process, respectively, = ( - Ttro for the crystallization scenario under consideration. Note: growth rate, U, is a product of the driving-force function, {)D}, and the mobility factor, {M}, per Exp. (3a). o
conditions and under resonant MW-field, respectively. As accentuated in the figure, the lowering of crystal-growth temperature and the enhancement of isothermal crystal-growth kinetics, both of which were previously observed and reported on the LS2 glass-crystallization conducted under resonant VFMWI,[1] are now quite obviously supported by the proposed phenomenologybased rationale.††††† Single-Mode-MW-Augmented Decrystallization of Crystalline Oxide-Ceramics Under purely-thermal (zero-field) conditions, the (intrinsic) molar free energy (chemical potential) of any crystalline solid is always lower than that of its noncrystalline (amorphous) solid-state counterpart (of the same composition). Yet, under the separated H-field in a 2.45 GHz single-mode-MW cavity, crystalline ferrites with unpaired spins in their 3d-cations such as Fe3O4, CoFe2O4, NiFe2O4, and BaFe12O19 were observed undergoing decrystallization without going through melting in a time duration of seconds.[2a],[2b] Therefore, this type of solid-state
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
35
crystalline-to-amorphous phase transformation, occurring under MWI at ambient pressure below “normal” melting temperature, is considered to be “unexpected” (or even “non-feasible”) within the traditional domain of thermodynamics that deals with systems exclusively coordinated by pressure, volume, temperature, and composition, i.e., PVTN systems.[3],[5a] On the other hand, this “unexpected” but real solid-state phenomenon – MW-induced solid-state decrystallization (amorphization) without passing through actual melting, may be comprehended using a perception of the EM-field-augmented thermodynamics.[4],[7]-[11] For the sake of illustration, we simply consider the isothermal solid-state decrystallization as follows: aD xF Crystalline 3d-ferrite Under CPRMW H-field Amorphous 3d-ferrite with a high magnetic-field @ T below TMP with a low magnetic-field absorptivity absorptivity During decrystallization under the separated, resonant H-field in a single-mode-MW cavity, the remarkable time-averaged, molar resonant magnetization work input to the irradiated magnetic crystalline ferrite (xF), i.e., = | (: PoxFPrmaxHo2), is expected to be far exceeding the magnetization work input to its amorphous counterpart-product (aD) with a low magnetic-field absorptivity, i.e., = | (: PoaDPrHo2); since xFPrmax >> aDPr. As a result, the newly established free-energy path for amorphization of the crystalline ferrite under irradiation may effectively promote both the isothermal feasibility and kinetics of the nucleation and growth sequence during the solid-state decrystallization process occurring within the irradiated ferrite body as illustrated in Figure 4.
Figure 4. Improvement of free-energy path of the isothermal, solid-state decrystallization occurring inside a 3d-ferrite under the influence of the separated resonant H-field of a single-mode MWI. As shown in the figure, the driving force for the isothermal, solid-state decrystallization under purely-thermal conditions is, as expected, not a favorable one, i.e., 'F = 'Go > 0. Yet un-
36
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
der single-mode MWI, the tremendous initial-state, resonant magnetization work input, i.e., = >> ('Go + ) = ('Go + ), could yield a favorable driving force for the process: placing 'F | ('Go + - ) = ('Go + - ) < 0, as per Exp. (2bb), thereby switching the unfavorable purely-thermal process to a favorable one under single-mode MWI. In brief, a tremendous contribution of = to the initial-state free energy of a MW-irradiated ferrite could effectively convert the isothermal, solid-state decrystallization feasibility from non-feasible under purely-thermal conditions, to feasible under appropriate resonant single-mode-MWI. Furthermore, Figure 4 suggests that the tremendous free energy input from = may also significantly enhance both nucleation and growth kinetics of the decrystallization process under irradiation, since this free energy contribution could effectively reduce the energies required for overcoming (1) the nuclei-formation barrier, 'F*, via an augmentation of the driving force for phase change, 'F, as per Exp. (2b), and (2) the activation barrier for atomic/ionic mobility as per Exp. (2a). In brief, free-energy path of the isothermal, solid-state decrystallization of a 3d-ferrite could be effectively improved using the resonant H-field of a single-mode MWI (Figure 4). This improvement may promote not only process feasibility well below the traditional melting point of the system, but also may significantly enhance the isothermal process kinetics. The MW-augmented solid-state decrystallization process may be further envisioned as a solid-state structural change resulting from destruction of bulk “crystallinity” (3D long-ranged periodicity of atomic/ionic arrangement) within the irradiated ferrite, attributed to a tremendous resonant magnetization energy input to the magnetic crystalline solid under the resonant H-field of a single-mode MWI. The overall kinetics of this type of solid-state decrystallization, which is governed by isothermal, simultaneous nucleation and growth of clusters of the MWI-induced atomic/ionic packing defects†††††† generated within the bulk of the irradiated crystalline solid, is best characterized by the MW-augmented JMA equation as already stated in Exp. (4):††††††† X = 1 - exp(-Ktn)
(4)
where X is the volume fraction decrystallized, t is the isothermal decrystallization time elapsed, and n = 4. Also, K | [(S/3)(IEM)(UEM)3] per Exp. (4a), IEM ~ {exp[-Ƨ/(RT)]}{exp[-(NA'F*)/(RT)]} per Exp. (2), and UEM ~ {exp[-Ƨ/(RT)]}{1 - exp['F/(RT)]} per Exp. (3). Under resonant H-field, both the nucleation barrier, 'F*, and the activation free energy, Ƨ, could be significantly reduced; thus the nucleation and growth rates of the MWI-induced packing-defect-clusters inside the crystalline solid, IEM and UEM respectively, may then be highly accelerated. As a result, the rate constant, K, and subsequently the overall 3D decrystallization rate, as predicted by the MW-augmented JMA equation, may also be greatly enhanced. Figure 5 schematically illustrates the sigmoidal solid-state decrystallization kinetics of an irradiated 3d-ferrite under resonant, single-mode MWI at various temperatures as predicted by the MW-augmented JMA equation, i.e., Exp. (4). Within the temperature range, between the crystalline-to-amorphous transformation temperature threshold of the system under resonant MWI, TtrEM,†††††††† and the traditional melting point of the system under purely-thermal (zerofield) conditions, TMP (Ttro), the solid-state decrystallization rate of the irradiated ferrite is revealed to be monotonically increasing with increasing temperature. The rationale for the direct proportionality between the augmented decrystallization rate and the temperature in this heatinginduced phase transformation is rather obvious:
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
37
Figure 5. Temperature-dependent, sigmoidal solid-state decrystallization kinetics of a 3d-ferrite under resonant, single-mode MWI. Non-feasible decrystallization of the same ferrite under purely-thermal conditions at all temperatures below the melting point (TMP/Ttro) is also shown. Increase in temperature increases both the driving force (superheating) and mobility for the MWaugmented decrystallization process, which in turn increase both the MWI-induced defect-cluster nucleation rate, IEM, and the lineal amorphous (defective) phase-growth rate, UEM. Consequently, above the irradiation-induced crystalline-to-amorphous transformation temperature threshold, TtrEM, the overall solid-state decrystallization kinetics of a crystalline 3d-ferrite under CPRMWI will increase monotonically with increasing temperature, as predicted by the MW-augmented JMA equation and exhibited in Figure 5. CONCLUSIONS Based on thermodynamic and kinetic phenomenology, this report explains why scientists and engineers have been quite successfully accomplishing isothermal augmentations on ceramic crystallization and decrystallization using coherent, polarized and resonant microwaveirradiation (CPRMWI). The concept and rationale proposed in this report may encourage further discussions and research efforts towards the development of a working science-based guide for novel applications of microwave-augmented crystallization and decrystallization in producing high-performance, engineered ceramics. Many crystallization and decrystallization transformations in ceramic processing are nucleation-and-growth-controlled processes. Under CPRMWI, the thermodynamics and kinetics of these processes could be significantly augmented. This type of MW-induced processaugmentation is mainly caused by a remarkable increase in free energy of the initial-reactant state of the irradiated system, resulting from a tremendous time-averaged, molar resonant MWwork input. This enormous resonant work/energy contribution may productively enhance the driving force and/or reduce the activation free energy for atomic/ionic mobility, thus effectively improving the free-energy path of the transformation process, subsequently promoting the process feasibility and/or kinetics under CPRMWI.
38
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
In isothermal variable-frequency-MW-augmented crystallization of lithium disilicate (Li2O.2SiO2 or LS2) glass, the molar resonant “dielectric” work input, , resulting from resonant polarization of the irradiated glass, may both increase the driving force for phase-transition and lower the activation free energy for atomic/ionic mobility, thus effectively improving the kinetics (mechanism) of the crystal growth and subsequently, the crystallization process. On the other hand, the molar resonant “magnetic” work contribution, , which originates from resonant magnetization of a (magnetic) crystalline 3d-ferrite irradiated under the separated resonant H-field in a single-mode-MW cavity, may also tremendously augment the thermodynamic driving force and reduce the kinetic barrier-energy, thereby effectively improving the free-energy path and promoting the system-reactivity for decrystallization. Consequently, a crystalline ferrite, irradiated under resonant, single-mode MWI, may undergo a rapid amorphization at temperatures even well below its traditional melting point. FOOTNOTES † Exp. (2a) for activation free energy, Ƨ, and Exp. (2bb) for driving force, 'F, are applicable to both nucleation and growth kinetics discussed in this article. †† Typical electric field amplitude, Eo, used in most microwave processing is 2(gcHr) was documented, where glHr, gcHr are the relative permittivities of the LS2 glass and its glass-ceramic counterpart, respectively. Based on the above dielectric data of LS2 glass and its counterpart-glass-ceramic measured under 2.46 GHz MWI at elevated temperatures, one may infer that the resonant relative permittivity of the highly polarized LS2 glass, glHrmax, induced by the higher-frequency, resonant VFMWI at the crystallization-temperature, could have been enormously (several orders of magnitude) greater than the relative permittivity, gcHr, of the slightly or non-resonant glass-ceramic counterpart-product,[9],[10] i.e., glHrmax >> gcHr during LS2-glass-crystallization conducted under resonant VFMWI. Consequently, = ( - ) = ( - ) | [(: HoEo2)][gcHr - glHrmax] 1,623 K as proposed in the reaction network in Figure 2) in less than 10 seconds after sodium azide starts to decompose (673 K) which is the first chemical reaction in the designed SHS reaction scheme. The temperature then increases rapidly due to the highly exothermic silicon nitridation. The time for pellet temperature to stay above the silicon nitridation temperature (1,623 K) is about 40, 30, and 20 seconds for the points at 1, 2.5, and 4 inches from the ignition surface. The combination of high reaction temperature and sufficient reaction time results in a complete conversion of Si to Si3N4. The slope changes circled in Figure 6 indicate that the endothermic Si3N4 decomposition stops when the temperature is lower than 2,151 K, the proposed Si3N4 decomposition temperature.
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
129
3500 3000
Temperature (K)
2500 2000 1500 1000 500 0 0
20
40
60 Time (s)
80
100
120
Figure 6. The temperature histories along the pellet centerline (Δ: 1.0 inch, : 2.5 inch, : 4.0 inch from the ignition surface) after the ignition when 3Si 4 3 NaN 3 0.2Si 3N 4 is used as the starting mixture. 25000
Concentration (mol/m3)
20000
15000
10000
5000
0 0
20
40
60 Time (s)
80
100
120
Figure 7. The concentration histories of Si (Δ), NaN3 (◊), Na ( ), and Si3N4 ( ) at a centerline point 2.5 inch from the ignition surface when 3Si 4 3 NaN 3 0.2Si 3N 4 is used as the starting mixture. The compositions of a centerline point located 2.5 inches from the ignition surface are calculated and shown in Figure 7. The NaN3 decomposition starts at 58 seconds after the ignition
130
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
and finishes in about 7 seconds; immediately followed by the silicon nitridation (the decrease of Si concentration and the increase of Si3N4) at t= 65 s. The silicon nitridation lasts for about 20 seconds in this case. The sodium (Na) formation from the NaN3 decomposition and vaporization above its boiling point (1,156 K) are also observed from the change of Na concentration. The change of composition distribution are calculated at different times. Figure 8(a) shows the compositions of the initial reactants and Figure 8(b) is the concentration distributions of all solid chemicals in this SHS reaction system. From Figure 8(b), we can estimate at 25 seconds after the ignition the NaN3 decomposition zone is about 0.5 inch long, silicon nitridation reaction zone is about 2 inch long, and the Na vaporization zone is about 0.5 inch long.
Concentration (mol/m3 )
25000 20000 15000 10000 5000 0 0
1
2 3 4 Distance from Ignition Surface (in)
5
6
5
6
(a)
Concentration (mol/m3)
25000 20000 15000 10000 5000 0 0
1
2 3 4 Distance from Ignition Surface (in)
(b) Figure 8. The concentration distribution of Si (Δ), NaN3 (◊), Na ( ), and Si3N4 ( ) along the pellet centerline when 3Si 4 3 NaN 3 0.2Si 3N 4 is used as the starting mixture. (a) initial concentrations at t = 0, (b) the concentration profiles at t = 25 seconds after the ignition. Different amounts of initial Si3N4 was used to study the impact of initial Si3N4 dilution on the SHS reaction when 3 moles of Si and 4/3 moles of NaN3 are used. From Figure 9, it can be concluded that the addition of the Si3N4 diluent effectively reduced the maximum combustion temperature from 3,400 K to 2,800 K when 0.3 mole of Si3N4 is added into the reactant mixture.
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
131
There is no propagation when 0.4 mole of Si3N4 is added into the reactant mixture. Although a lower combustion temperature can reduce the Si3N4 decomposition, it also lower the silicon concentration and significantly reduces the rate of the silicon nitridation according to the Arrhenius equation (Equation (12)). Using this developed model the optimization Si3N4 amount can be determined to maximize the time for the reaction pellet temperature stays between 1,623 K (the nitridation temperature) and 2,151 K (the decomposition temperature). 4000 3500
Temperature (K)
3000 2500 2000 1500 1000 500 0 0
20
40
60 Time (s)
80
100
120
Figure 9. The temperature histories of different amount of Si3N4 diluent in the reactant mixture (Δ: 0, : 0.2, ◊: 0.3, : 0.4), at a centerline point 2.5 inch from the ignition surface when 3Si 4 3 NaN 3 0.2Si 3N 4 is used as the starting mixture. CONCLUSIONS A finite element analysis model of Self-propagating High-temperature Synthesis (SHS) of silicon nitride was successfully developed to investigate the reaction at a normal pressure nitrogen environment (1 atm). Adiabatic temperatures of different starting reactant compositions are calculated. Temperature distribution, temperature history, composition history, composition distribution, and the impact of Si3N4 dilution are studied. This model can be used to optimize the reaction condition for a large scale economical production of silicon nitride. ACKNOWLEDGMENT This work is partially supported by the 2017 Technology Transfer Initiative (TTI) Program of Center for Advanced Research of Energy and Materials (CAREM) of Faculty of Hokkaido University, Japan.
132
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
NOMENCLATURE k Na Vaporization rate constant of sodium, [mol/m3/s] Pre-exponential factor for the decomposition of sodium azide, [sec-1] k NaN 3 ,0 k Si ,0 k Si 3N 4
Pre-exponential factor for silicon nitridation, [sec-1] Silicon nitride decomposition rate constant, [sec-1]
ni p r t u w x y z Cpi C Na C NaN 3
Stoichiometric coefficient of reactants and products Fluid pressure, [Pa] Radius [in] Time [s] Velocity of fluid, [m/s] Moles of nitrogen Moles of silicon Moles of sodium azide Moles of silicon nitride Specific heat of species i [J/mole/ K] Concentration of sodium at any time (t), [mol/m3] Concentration of sodium azide at any time (t), [mol/m3]
CSi C Si 3N 4
Concentration of silicon nitride at any time (t), [mol/m3] Concentration of silicon nitride at any time (t), [mol/m3]
E NaN 3
Activation energy for the decomposition of sodium azide, [J/mole]
E Si F I R T T ad
Activation energy for silicon nitridation, [J/mole] External force applied on the fluid, [N] Identity matrix Gas constant, [8.314 J/mole/K] Temperature [K] Adiabatic temperature [K]
Tf Tp
Temperature of the fluid at any time (t), [K] Temperature of the pellet at any time (t), [K]
T X
Surrounding temperature at any time (t), [K] Distance from the ignition surface, [in] Standard heat of reaction (J) Heat of vaporization of sodium, [J/mole] Density of fluid, [kg/m3] Viscosity of the fluid, [Pa-s] Ratio of green density to theoretical density
H r0 H Na ,v
dC NaN 3 dt dC Si dt dC Si 3N 4 dt
Rate of consumption of sodium azide, [mol/m3/s] Reaction rate of silicon consumption, [mol/m3/s] Reaction rate of silicon nitride formation, [mol/m3/s]
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
133
REFERENCES 1 F.L. Riley, “Silicon Nitride and Related Materials”, J. Am. Ceram. Soc. 83, 245–265 (2000). 2
Berroth, K. Silicon Nitride Ceramics for Product and Process Innovations, .Advances in Science and Technology, 65, 70-77 (2010).
3
S. Hampshire, “Silicon Nitride Ceramics – Review of Structure, Processing and Properties”, Journal of Achievements in Materials and Manufacturing Engineering, 1, 43-50 (2007).
4
H. Vladimir, J.A. Puszynski, “Chemical Engineering Aspects of Advanced Ceramic Materials”, Ind. Eng. Chem. Res., 35, 349-377 (1996).
5
M. Ekelund, B. Forslund, “Carbothermal Preparation of Silicon Nitride: Influence of Starting Material and Synthesis Parameters”, J. Am. Ceram. Soc., 75 (3), 332-339 (1992).
6
A. Atkinson, A.J. Moulson, E.W. Roberts, “Nitridation of High-Purity Silicon”, J. Am. Ceram. Soc., 59, 285 (1976). 7
T. Yamada, “Preparation and Evaluation of Sinterable Silicon Nitride Powder by Imide Decomposition Method”, Am. Ceram. Soc. Bull., 72, 99-106 (1993).
8
W.P. Shen, F. Wang, W.U. Zhuohui, G.E. Changchun, “Effect of Auxiliary Gases on Combustion Synthesis of Si3N4”, J. Mater. Sci. Technol., 21 (5) (2005).
9
G. Xanthopoulou, G. Vekinis, “An Overview of Some Environmental Applications of Selfpropagating High-temperature Synthesis”, Adv. Environ. Res (Oxford, U. K.), 5, 117-128 (2001). 10
S. Lin, J. Selig, H.T. Lin, H. Wang, “Self-propagating High-temperature Synthesis of Calcium Cobaltate Thermoelectric Powders”, J. Alloys Compd., 503, 402-409 (2010). 11
A. Varma, S.R. Alexander, S.X. Alexander, S. Hwang, “Combustion Synthesis of Advanced Materials: Principles and Applications”, Adv. Chem. Eng., 24, 79-226 (1998). 12
S. Lin, R. Wilkins, Z. Henry, “High-pressure Self-propagating High-temperature Synthesis (SHS) of Silicon Carbide-Silicon Nitride Composites”, Ceram. Trans., 103, 51-61 (2000).
13
G.A. Rossetti, R.P. Denkewicz, “Kinetics Interpretation of - and -Si3N4 Formation from Oxide Free High-purity Silicon Powder”, J. Mater. Sci., 24, 3081-3086 (1989). 14
S.W. Yin, L. Wang, L.G. Tong, F.M. Yang, Y.H. L, “Kinetic Study on the Direct Nitridation of Silicon Powders Diluted with -Si3N4 at Normal Pressure”, International Journal of Minerals, Metallurgy and Materials, 20 (5), 493 (2013).
15
H. Ishikawa, K. Oohira, T. Nakajima, T. Akiyama, “Combustion Synthesis of SrTiO3 using Different Raw Materials”, J Alloy Compd., 454 (1-2), 384–388 (2008).
16
C.R. Bowen, B. Derby, “Self-propagating High-temperature Synthesis of Ceramic Materials”, Br. Ceram. Trans., 96 (1), 25 (1997). 17
P.W.M. Jacobs, T.A.R. Kureishy, “Kinetics of Thermal Decomposition of Sodium azide”, J. Chem. Soc., 4718-4723 (1964).
134
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
18
Y. Yu, R.H. Hewins, C.M.O.D. Alexander, J. Wang, “Experimental Study of Evaporation and Isotopic Mass Fractionation of Potassium in Silicate Melts”, Geochim. Cosmochim. Acta., 67 (4), 773-786 (2003).
19
H. Batha, E. Whitney, “Kinetics and Mechanism of the Thermal Decomposition of Si3N4”, J. Am. Ceram. Soc., 56 (7), 365-369 (1973).
20
S. Lin, V. Doddapaneni, J. Lin, T. Akiyama, A. Hiranaka, “Finite Element Analysis of Selfpropagating High-temperature Synthesis (SHS) of Silicon Nitride”, Ceram. Trans., (accepted for publication).
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
135
SYNTHESIS OF CARBIDE CERAMICS FROM ACTIVATED CARBON PRECURSORS LOADED WITH TUNGSTATE, MOLYBDATE, AND SILICATE ANIONS Grant Wallace, Jerome Downey, Jannette Chorney, Katie Schumacher, Trenin Bayless, Alaina Mallard, Auva Speiser, Elizabeth Raiha Montana Tech of The University of Montana Butte, MT, USA ABSTRACT Current commercial carbide ceramic powder production processes require temperatures greater than 1500oC and implement extensive milling operations. A process that could reduce the energy requirements of commercial carbide production could allow for these materials to be more widely applied. In this study, tungstate (WO42-), molybdate (MoO42-), and silicate (SiO32-) anions were adsorbed onto activated carbon. Silicon carbide (SiC) whiskers and mixed crystals of tungsten carbide (WC), tungsten semicarbide (W2C), and tungsten (W) were formed via carbothermal reduction using inert and reducing gas atmospheres at temperatures much lower than what is required commercially (950oC for WC/W2C/W and 1200oC for SiC). Molybdenum carbide (Mo2C) was synthesized at 1100oC under argon gas and MoC was formed at 600oC under H2 at carburization times of 6 h and 8 h. Separation of the WC/W2C/W mixture from the activated carbon matrix was achieved using a heavy liquid separation in lithium metatungstate solution (LMT). The adsorption process for each anion species was statistically-optimized via a central composite response surface analysis. Inductive coupled plasma spectroscopy (ICP-OES) was used to measure adsorption efficiency while the carburization products were characterized using X-ray diffraction and scanning electron microscopy. INTRODUCTION The materials known as ceramic carbides vary greatly in terms of structure and properties. Metal carbides, such as WC, are among the hardest materials in existence. Known as hardmetals, metal carbides like WC possess hardness values approaching that of diamond as well as being resistant to wear and chemical attack at relatively high temperatures. Because of their corrosion and abrasion resistance, carbides are uniquely suited for a number of industrial applications where other traditional metals do not perform as well. Such applications include high-temperature cutting tools, drill bits, surgical implements, and alloying agents1. Tungsten carbide and Mo2C possess catalytic properties that may allow them to act as less-expensive alternatives to noble metal catalysts such as platinum and palladium2. Other ceramic carbides, such as SiC are also characterized by extremely high hardness values and wear resistance. Despite similarities to the hardmetals in terms of mechanical properties, they differ greatly in terms of their physical/chemical structure. While hardmetals consist of an interstitial compound of metal and carbon, SiC is a covalently-bonded structure1. Lighter than the hardmetals, SiC is used in ceramic body armor, brake discs, and high-pressure valves and nozzles3. Due to its thermal resistance and electrical conductivity, SiC is also being considered as an alternative to pure silicon as a semiconductor material for use in hightemperature electronics4. Current ceramic carbide processing methods are energy intensive. Most carbides are produced at relatively high temperatures (>1500oC) by reacting metal or metal oxide powders with carbon. Extensive milling operations are then required to reduce the carbides into a fine powder5. The synthesis of carbide ceramics from aqueous solutions has shown promise as a method for reducing the energy requirements associated with carbide production by reducing carburization temperatures and final particle size. Tungsten and Mo, in the form of tungstate and
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
137
molybdate anions, have been shown to adsorb onto activated carbon in relatively high concentrations, making them suitable candidates for producing commercially viable carbide precursors6,7. Tungsten carbide-cobalt composites (WC/Co) and Mo2C have also been synthesized from solutions of ammonium tungstate, or ammonium molybdate, and sucrose through a drying/carburization process8,9. Silicon carbide has also been synthesized at 1300oC under a H2 atmosphere using wood pulp residue loaded with silicate anions adsorbed from aqueous solution10. The Metallurgical and Materials Engineering Department at Montana Tech is investigating the possibility of using anion-loaded activated carbon precursors to produce carbide ceramics at temperatures lower than most current commercial processes11. Tungstate, molybdate, and silicate anions were adsorbed onto an activated carbon matrix using statistically-optimized conditions to maximize adsorption. Carburization was carried out using the anion-loaded precursors under varying gas compositions and operating temperatures to promote the conversion of the adsorbed anions into carbide ceramics. EXPERIMENTAL PROCEDURE Precursor Preparation and Adsorption Optimization Carbide precursors were produced by adsorbing anions of interest onto activated carbon (Sigma Aldrich, -100/+400 Mesh) from aqueous solutions. One hundred milliliter solutions were prepared with varying concentrations of sodium tungstate, sodium molybdate, and sodium metasilicate. An experimental design matrix was developed for each of the three precursors in order to mathematically model adsorption behavior. To produce W-loaded precursor material for the synthesis of WC, a design matrix, consisting of 30 experiments, was carried out in order to determine the effects of initial W concentration (1000, 10,500, and 20,000ppm), temperature (20, 40, and 60oC), pH (2, 4, and 6,), and reaction time (1, 12.5, 24 h) on the adsorption of tungstate anions onto activated carbon. Tungstate solutions were prepared by dissolving sodium tungstate dihydrate in 100 mL of deionized water. To promote tungstate adsorption, sodium chloride was added to the solution to a concentration of 0.2 M6 and the pH of each solution was altered using hydrochloric acid. A 2.50 g sample of activated carbon was then added to the solution and the solutions were agitated to suspend the activated carbon during adsorption. Solutions carried out at 20oC were agitated at 480 rpm on an orbital shaker table and solutions carried out at elevated temperatures were agitated on a hot/stir plate at 480 rpm. Following adsorption, the activated carbon was separated from the solution by vacuum filtration and dried. The amount of tungstate removed from solution was determined by difference using ICP-OES via an ICP Thermo-Scientific iCAP 6000 instrument. A nearly-identical design matrix was carried out to model the adsorption behavior of molybdate anions. A series of 30 experiments were carried out using the same conditions that were used in the tungstate design matrix. Molybdate solutions were prepared by adding sodium molybdate to 100 mL of deionized water along with sodium chloride so that the NaCl concentration was 0.2 M and the Mo concentration reached the desired amount for the given experiment. Silicate adsorption behavior was modelled using a design matrix consisting of 20 experiments that analyzed the effects of reaction time, temperature, and initial Si concentration on silicate adsorption. For the silicate design matrix, reaction times were set at 1 h, 6 h, and 12 h, temperatures were held at 20oC, 40oC, and 60oC, and Si concentrations were set at 10,000 ppm, 30,000 ppm, and 50,000 ppm. The solution pH was not used as a variable, as the addition of hydrochloric acid induced the precipitation of a silica gel that rendered adsorption impossible in preliminary scoping experiments. Because of this phenomenon, pH became a function of initial
138
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
silicate concentration, which was accounted for in the design of experiments. The ICP data from each design of experiments matrix was fitted to mathematical models of adsorption behavior using the statistical analysis software, DesignExpert 9 (StatEase Inc.). Optimal conditions for maximizing adsorption of each anion were determined and used to produce precursor material for the carburization experiments. Samples of W-loaded precursor were prepared using solutions containing 18,000 ppm W and a pH of 2, where adsorption occurred at 20oC for 2 h. To prepare sufficient quantities of W-loaded precursor, the solution volumes and carbon sample masses were increased to 2 L of solution and 50 g of activated carbon. The same optimized conditions were applied to these scaled-up experiments. Mo-loaded precursor was prepared from solutions containing 18,000 ppm Mo with a pH of 2. Adsorption was carried out at 20oC for 2 h. Si-loaded precursor samples were prepared from solutions containing 50,000 ppm Si, and adsorption occurred at a temperature of 20oC for 2 h. Carburization Experiments Carburization of the W-loaded precursor was carried out in an ATS 3210 split tube furnace using a quartz kiln attached to a rotary electric motor. In initial carburization experiments, 50 g of W-loaded precursor were heated under argon gas to a final temperature of 865oC at a rate of 6oC/min. Once the interior kiln temperature reached 865oC, the argon atmosphere was replaced with a gas mixture consisting of 60% H2, 25% CO, and 15% CH4 by volume with a total volume of 0.5 L/min. Gas flowrates were controlled by a series of three Omega rotameters. The sample was held at 865oC for 8 h before being cooled to room temperature under argon (Experiment name: RGC-2). This procedure was used for multiple carburization runs with changes being made to the final operating temperature and reducing gas composition in order to improve the production of WC on the activated carbon matrix. To improve carburization, temperatures were increased to 950oC and the reducing gas composition was altered to contain 10% H2, 10% CO, and 80% CH4. The mass of sample placed within the rotary kiln was also reduced to 25 g to promote greater solid-to-gas contact. Virgin activated carbon was also added to the W-loaded precursor in a 1:1 mass ratio in order to further promote carburization. This mixture was blended in a ball mill for 5 min prior to being loaded into the furnace and heated at 950oC for 8 h (RGC-17). This blended, carburized material was also blended with an additional amount of activated carbon before undergoing additional carburization at 950oC for 8 h (RGC-18). Carburization of the Mo-loaded precursor was performed in a MTI 1500x GSL tube furnace. A 1.25 g sample of the Mo-loaded precursor was placed inside an alumina sample boat and place in the middle of the furnace tube. Initial experiments were carried out under an inert argon atmosphere for the duration of the experiment. The precursor was heated to 1100oC under argon and held at 1100oC for 4 h, 8 h, and 12 h. Carburization experiments were also carried out under H2 in an attempt to reduce the operating temperatures required for carburization. Samples of Mo-loaded precursor were heated to 600oC under argon at a rate of approximately 5oC/min. Once the furnace temperature reached 600oC, the argon gas flow was replaced with H2 and the precursor was held at 600oC for 4 h, 6 h, and 8 h. The Si-loaded precursor was carburized in the same furnace as the Mo-loaded precursor. Carburization experiments were carried out under both argon and H2 atmospheres. Carburization experiments under argon were carried out at 1400oC, 1300oC, and 1200oC for 20 h. Samples were first heated to 200oC at 5oC/min, and held at temperature for 20 minutes. Then the samples were heated to 600oC at 10oC/min and held at temperature for 20 minutes. At this point, the samples were heated to the final desired temperature at 6-8oC/min and held at temperature for 20 h. Upon completion of the experiment, the furnace was cooled to ambient conditions. This procedure was also used to carburize Si-loaded precursor samples under a H2 atmosphere.
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
139
Samples were heated according to the previous method under argon until the desired operating temperature of 1200oC was reached. Once the furnace reached 1200oC, the argon atmosphere was replaced with a H2 atmosphere and samples were held at temperature for 4, 6, and 8 h. Carburized products were placed in a micronizing mill with an addition of ZnO equal to 10% of the sample mass and milled to -325 Mesh in preparation for analysis via X-ray diffraction (XRD). Samples of carburized Mo-loaded, and W-loaded, precursor were ground to -325 Mesh using a mortar and pestle in preparation for analysis via XRD. Carburized samples were analyzed using X-ray diffraction via a Rigaku Ultima IV diffractometer with a Cu-Kͣ radiation at 40 kV and 40 mA. A semi-quantitative analysis of the crystalline products was conducted using the whole powder pattern fit (WPPF) method. Images of the carburized product were also obtained using the electron backscattering detector on a LEO 1430VP scanning electron microscope (SEM). Carbide Separation Carburized W-loaded precursor was milled for 40 min with ethanol in a stirred ball mill. Following the milling operation, the remaining ethanol was evaporated and the dried material was added to a saturated solution of lithium metatungstate (LMT). A heavy liquid separation was carried out by centrifuging the sample for approximately 60 min. Following centrifugation, the solution was allowed to settle for an additional 60 min before the high-density material was removed from the centrifuge tube via pipette and the remaining LMT was removed by passing the solution through a 0.2 micron syringe filter. RESULTS AND DISCUSSION Adsorption Optimization The adsorption behavior of tungstate, molybdate, and silicate anions onto activated carbon was modelled and optimized using DesignExpert 9. The adsorption of tungstate was measured in grams of W adsorbed per gram of activated carbon (g W/g C). The tungstate adsorption data was modelled using an inverse square root transform combined with a modified quadratic relationship. This model can be expressed according to Equation 1. 1 W
= 0.034320 - 3.09986 × 10-6 A + 9.72219 × 10-6 B + 6.50037×10-8 AB + 8.4568 × 10-11 A2 (1)
where W is defined as the amount of adsorbed W (g W/g C), A is defined as the initial concentration of W in solution (ppm), and B is defined as the pH of the solution. The model’s goodness of fit was verified using a number of statistical diagnostics including analysis of variance (ANOVA) and Cook’s Distance. From the model, it can be observed that the two statistically significant factors affecting W adsorption were the pH of the solution and the initial solution concentration of W. Reaction time and temperature were shown to have no statistical significance on the adsorption behavior of tungstate onto activated carbon11. A model for molybdate adsorption was generated using logarithmic (base 10) transform and a modified quadratic relationship. Molybdate adsorption, expressed as grams of Mo per gram of activated carbon (g Mo/g C), was modelled using Equation 2:
log10 Mo = -1.381 + 1.028 × 10-4 A + 1.404-3 B - 2.794 × 10-9 A2 - 0.018B2
(2)
where Mo represents the adsorption of molybdate anions in units of g Mo/g C, A, represents the initial molybdate concentration of the solution in units of ppm, and B represents the pH of the molybdate solution. Although different mathematical transforms were used to model adsorption behavior, molybdate adsorption behavior was found to be similar to tungstate adsorption. Like tungstate
140
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
adsorption, the adsorption of molybdate anions from solution was determined by initial concentration and pH, while reaction time and temperature were found to be statistically insignificant. Silicate adsorption behavior was modelled using Equation 3: 1 Si
= 5.006 - 5.953 × 10-5 A
(3)
where Si represents the degree of silicate adsorption in g Si/ g C, and A represents the initial solution concentration of Si in ppm. From the model, it can be observed that the only significant factor in determining adsorption of silicate anions from solution is the initial concentration of silicate in solution within the parametric ranges. Response surface curves modelling the adsorption behavior of each anion are presented in Figure 1. For tungstate and molybdate, anion adsorption is plotted as a function of pH and initial anion concentration, while temperature and reaction time are fixed at 20oC and 2 h respectively. Silicate adsorption is plotted as a function of initial Si concentration and temperature, while reaction time is held constant at 2 h.
0.70 0.40
0.50 0.40
Mo Adsorbed (gMo/gC)
W Loading (g W/g C)
0.60
0.30 0.20 0.10 0.00
2.00
pH
5.00 6.00
1000.00
4800.00
8600.00
12400.00
16200.00
0.20 0.10 0.00 2.00 20000.00 16199.36 12398.73
3.00 4.00
0.30
20000.00
3.00 4.00 8598.09
W Conc. (mg/L)
Initial Mo Conc (mg/L)
pH
5.01
4797.45 996.82
6.01
(b)
(a) 0.50
Si Loading (g Si/gC)
0.40 0.30 0.20 0.10 0.00
60.00 50.00 50000.00
40.00
Temperature (deg C)
40000.00 30000.00
30.00
20000.00 20.00
10000.00
Initial Si Conc (mg/L)
(c)
Figure 1: Response Surface Curves for Tungstate (a), Molybdate (b), and Silicate (c) Adsorption Optimal conditions for maximizing anion adsorption were determined from these response surface curves. Maximum tungstate adsorption (Figure 1a) was determined to occur from an initial solution concentration of 18,000 ppm W, a pH of 2, a solution temperature of
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
141
20oC, and a reaction time of 2 h. Experiments carried out using these parameters consistently produced W-loaded precursor with amounts of adsorbed W measuring approximately 0.35 g W/g C [11].. Due to the chemical similarities between tungsten and molybdenum, identical adsorption parameters were used to maximize molybdate adsorption in order develop a more versatile precursor production process. The molybdate adsorption design matrix produced Mo-loaded precursor containing approximately 0.32 g Mo/g C. Optimal conditions for silicate adsorption were fixed at a reaction time of 2 h, a temperature of 20oC, and an initial Si concentration of 50,000 ppm. The lack of an effect on adsorption behavior from changes in temperature is evident on the graph. These operating conditions were found to produce Si-loaded precursor containing 0.27 g Si/g C. Carburization An example of a typical XRD pattern for the carburized W-loaded precursor is shown in Figure 2.
Intensity (cps)
4e+004
3e+004
2e+004
1e+004
0e+000 20
40
60
80
2-theta (deg)
Figure 2: XRD Pattern of Carburized W-Loaded Precursor (950oC, 8 h, 10%H2/10%CO/80%CH4) The large amorphous peak beginning at 0o is due to the presence of the activated carbon matrix while the sharp peaks throughout the pattern indicate a mixture of WC, W2C, and W crystals. Using the WPPF method, it is possible to determine how changes to the experimental procedure affect the carburization of the tungstate anions present on the activated carbon by measuring changes in the relative weight percent values of the WC/W2C/W mixture. The effects of altering gas composition, carburization temperature, and blending the precursor with additional activated carbon can be observed in the data presented in Table I.
142
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
Table I: WPPF Analysis of W-Loaded Precursor Carburization Experiments RGC-2 (8h, 865oC, 60% H2, 25%CO, 15%CH4) Phase RGC2-1 (rel. Wt %) RGC2-2 (rel. Wt %) RGC2-3 (rel. Wt %) WC 56.2 52.1 56.8 W 39.3 42.3 38.7 W2 C 4.5 5.6 4.6 RGC-17 (8h, blended charge, 950oC, 10%H2, 10%CO, 80% CH4) RGC-17-1 RGC-17-2 RGC-17-3 Phase 86.4 81.6 83.4 WC 6.9 6.7 6.9 W 7.5 12.1 10.5 W2 C RGC-18 (8h, blended, charge, 950oC, 10%H2The, 10%CO, 80% CH4) RGC-18-4 RGC-18-5 RGC-18-6 Phase 89.5 85.8 89.6 WC 0 5.0 3.4 W 11.2 10.3 7.6 W2 C The highest degree of conversion to WC was achieved at an operating temperature of 950oC, a carburization time of 8 h, and a gas composition of 10% H2, 10% CO, and 80% CH4. The increase in the CH4 concentration inside the reducing gas atmosphere, coupled with the addition of additional activated carbon also aided in the conversion of the tungstate anions to WC. The XRD patterns of the inert gas carburization experiments involving Mo-loaded precursor are presented in Figure 3. The experimental data are presented from left to right in order of increasing carburization time (4 h, 8 h, and 12 h).
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
143
Figure 3: XRD Patterns of Carburized Mo-Loaded Precursor (1100oC, Ar, 4 h, 8 h, and 12 h) The results of the inert gas carburization experiments show only peaks corresponding to Mo2C present in the XRD patterns for the three experiments carried out at 1100oC. Peak location and intensity are nearly identical across all three experiments. The XRD patterns presented in Figure 1Figure 4 are examples of the results produced by the carburization experiments on Mo-loaded precursor carried out under a reducing gas atmosphere (H2).
144
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
5e+004
Mo C
3e+004
Mo C
Na Cl
Intensity (cps)
4e+004
Mo C
Mo C Na Cl
Mo C
1e+004
Na Cl
Na Cl
Na Cl
2e+004
0e+000 20
40
60
80
2-theta (deg)
(a)
Mo C
3e+004
Mo C
Intensity (cps)
4e+004
Mo C
Mo C Na Cl
1e+004
Na Cl
Na Cl
2e+004
0e+000 20
40
60
80
2-theta (deg)
(b) Figure 4: XRD Pattern of Carburized Mo-Loaded Precursor (600oC, H2, 6 h (a), and 8 h (b)) The XRD results of the reducing gas experiments are very different from the results of the inert gas experiments. Carburization under H2 gas at reaction times greater than 6 h produced MoC rather than Mo2C, and the presence of trace amounts of sodium chloride was detected. The XRD pattern for the carburization of Si-loaded precursor under H2 at 1200oC is shown in Figure 5.
Figure 5: XRD Pattern of Si-Loaded Precursor (1200oC, 4 h, H2)
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
145
The XRD pattern indicates the presence of a single crystalline phase, SiC. This XRD pattern is very similar to the diffraction patters of precursor samples carburized under argon at 1400oC for 20 h. However, the peak intensities are lower in the sample carburized at 1200oC, indicating that the conversion to SiC at 1200oC is less than what occurred at 1400oC. A micrograph of a carburized sample of Si-loaded precursor is presented in Figure 6. This image was taken using a Phenom tabletop SEM with an EBSD detector.
Figure 6: Silicon Carbide "Whiskers" on the Surface of an Activated Carbon Matrix The micrograph shows “whisker-like” crystals of SiC present on the surface of the activated carbon matrix. Although many of the crystals are tens of microns in length, the “whiskers” approach submicron dimensions in diameter. Separation Micrographs of the W-loaded precursor, following the heavy liquid separation in LMT, are presented in Figure 7.
(a) (b) Figure 7: SEM Micrographs of High Density (a) and Low Density (b) Material Following LMT Separation
146
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
A comparison of the two images presented in Figure 7 shows that the LMT separation produced two materials of different density. The settled “High Density” material appears to contain a much greater amount of the WC/W2C/W mixed crystals than the “Low Density” material that did not settle to the bottom of the centrifuge tube. The mixed carbide crystals appear much brighter than the activated carbon matrix using electron back-scattering detection (EBSD) due to their greater molecular weight(s), and these brighter particles appear in higher concentrations in the High Density material (Figure 7a). A second micrograph of the High Density material at higher magnification is presented in Figure 8.
Figure 8: SEM Micrograph of Carbide Crystals Embedded in Syringe Filter Media At this magnification, it can be observed that the High Density material appears to be predominately high-molecular weight crystals and little activated carbon. The dark fiber-like structures are filter media from the syringe filter used to separate the High Density material from the LMT. These micrographs show that a separation of the carbide product from the activated carbon matrix is possible through milling and a heavy liquid separation using LMT.
CONCLUSIONS Statistically-verified mathematical models of the adsorption behavior of tungstate, molybdate, and silicate anions were developed to maximize the adsorption of these anions onto an activated carbon matrix. Maximum adsorption values of 0.35 gW/g C, 0.32 g Mo/g C, and 0.27 g Si/g C were obtained. Carburization of a blended charge of activated carbon and previously-carburized Wloaded precursor at 950oC, using a gas mixture containing 10% H2, 10% CO, and 80% CH4, produced mixed crystals of WC, W2C, and W with the relative weight percent of WC present within this mixture approaching 90%. Separation of the mixed carbide product from the activated carbon matrix is possible through a short milling operation and a heavy liquid separation. Molybdenum carbides were synthesized using argon at 1100oC and H2 at 600oC in the form of Mo2C and MoC respectively. Finally, SiC, in the form of “whisker”-like structures were synthesized at temperatures as low as 1200oC using H2 as a reducing atmosphere. Operating temperatures for carburization under reducing conditions using the anion-loaded precursors are significantly lower than those employed in many commercial carbide processes.
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
147
ACKNOWLEDGEMENTS Research was sponsored by the Army Research Laboratory and was accomplished under Cooperative Agreement Number W911NF-15-2-0020. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the Army Research Laboratory or the U.S. Government. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation herein. REFERENCES A.S. Kurlov, and A.I. Gusev, Tungsten Carbides: Structure, Properties, and Applications in Hardmetals, New York: Springer International Publishing, (2013)
1
2
A. J. Brungs, A. P. York, J. B. Claridge, C. Marquez-Alvarez and M. L. Green, "Dry Reforming of Methane to Synthesis Gas Over Supported Molybdenum Carbide Catalysts," Catalysis Letters, 70, no. 3-4, pp. 17-122, (2000).
3
AZO Materials, "Machining of Silicon Carbide-Process, Applications and Types," [Online]. Available: http://www.azom.com/article.aspx?ArticleID=3271. (Accessed 16 April 2016).
4
J. W. Palmour, J. A. Edmond, H. S. Kong and C. H. Carter Jr., "6H-Silicon Carbide Devices and Applications," Physica B: Condensed Matter, vol. 185, no. 1-4, pp. 461-465, (1993).
5
G. S. Upadhyaya, Cemented Tungsten Carbides: Production, Properties and Testing, Westwood, New Jersey: Noyes Publications, 77-79, (1998).
6
J. J. Cruywagen and A. T. Pienaar, "The Adsorption of Tungsten (VI) on Activated Carbon from 1.0 M Na(H)Cl Solution," Polyhedron, vol. 8, no. 1, pp. 71-76, (1989).
7
K. Ohashi, K. Murakami and K. Yamamoto, "Adsorption Behavior of Tungsten (VI) onto Activated Carbon," The Japan Society for Analytical Chemistry, vol. 32, pp. 313-319, 1983.
8
M. Patel and J. Subrahmanyam, "Synthesis of nanocrystalline molybdenum carbide (Mo2C) by solution route," Materials Research Bulletin, vol. 43, no. 8-9, pp. 2036-2041, 2008.
9
T. Xinglong, Y. Maozhong, P. Shigao, P. Fangcai, Y. Jiangao and D. Yu, "Synthesis of Commercial-Scale Tungsten Carbide-Cobalt (WC/Co) Nanocomposite Using Aqueous Solutions of Tungsten (W), Cobalt (Co), and Carbon," Journal of Materials Science and Chemical Engineering, vol. 2, no. 7, pp. 1-15, 2014. 10
K. Vyshnyakova, G. Yushin, L. Pereselentseva and Y. Gogotsi, "Formation of Porous SiC Ceramics by Pyrolysis of Wood Impregnatied with Silica," International Journal of Applied Ceramic Technology, vol. 3, no. 6, pp. 485-490, 2006. 11
G. C. Wallace, J. P. Downey, J. Chorney, A. Mallard and D. Hutchins, "Synthesis of Carbide Ceramics via Reduction of Adsorbed Anions on an Activated Carbon Matrix," in 8th International Synposium on High-Temperature Metallurgical Processing, San Diego, CA, 2017. 12
Total Materia, "Metal Carbides," November 2006. [Online]. Available: http://www.totalmateria.com/page.aspx?ID=CheckArticle&site=ktn&NM=158. [Accessed 16 April 2016].
13
B. Ozpineci and L. Tolbert, "Silicon Carbide: Smaller, Faster, Tougher," 27 September 2011. [Online]. Available: http://spectrum.ieee.org/semiconductors/materials/silicon-carbide-smallerfaster-tougher. [Accessed 16 April 2016].
148
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
14
Z. Z. Fang, X. Wang, T. Ryu, K. S. Hwang and H. Y. Sohn, "Synthesis, Sintering, and Mechanical Properties of Nanocrystalline Cemented Tungsten Carbide-A Review," International Journal of Refractory Metals and Hard Materials, vol. 27, no. 2, pp. 288-299, 2009.
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
149
"EWBODFTJO%JFMFDUSJD .BUFSJBMTBOE&MFDUSPOJD %FWJDFT
!" #$ % & $ ' ( )$ ( $ *$ +", $ - ( ( .$/ 0$123 2 /' "3 4$5 $ - .$21 $5! ( 1 ( 61! 2* /' 5712$ -3/ 2/ 8 9 $ $-$ !2 2,2 /' 255 ( 2 21,42,1 $ - *!23 //4"/: 61! 2* ,;"/, 5 5 (' (?$5 *, =4/3
" !21,42,1! /' 32$5 !,00/12- !,1'$4 $1 $ $5*;- 6 $ $7 2 / 4$54,5$2 / ! 2/ 6!2 ($2 2" ''42! /' 2" $-!/102 / / 2" '/13$2 / /' 7/ - (= ,7!B, 25* 2" -0 - 2 /' 2" 7/ - ( / 2" $-!/102 / /' 2" 0/0,5$1 32$5 , $ - 2 $2 '/,1 5/: 4//1- $2- ! 2! /' 2" -'42 6 !,1'$4 $1 4/30$1-= $55* 2" $-!/102 / ( 6! $ 5$1(1 4/ 21 7,2 / 2/ 2" /44,11 4 /' 7/ - ( 2"$ , /1 2 $-!/102 / -/!= " ! 7/ - ( 7 ' 2! 2" '/13$2 / /' 2" 5$21$5 4/ -,42 6 0$2" 2" 5$*1= "! '/, -$2 / ! 3$* 7 "50',5 2/ 45$1 '* 2" 1! !2 6 !: 24" ( 0" /3 $= =
! !2 6 1$ -/3 $44!! 33/1* . "$6 $221$42- (1$2 21!2! 14 2 *$1! -, 2/ 2! 5/: 0/:1 4/ !,302 / " (" - ! 2* 2(1$2 / $ - 7221 4/30$2 7 5 2* : 2" 4/3053 2$1* 32$5 /C - !3 4/ -,42/1 . 24" /5/(*= 1$55* 2"* 4/ ! !2 /' 2" 4$0$4 2/1 5 D 32$5 !,5$2/1 32$5 !21,42,1 :"1 $ !,5$2 ( 5$*1 ! !: 24"- $3/ ( $2 5$!2 2:/ - ''1 2 1! !2$ 4 !2$2! , -1 $001/01 $2 6/52$(!EF= 242 / /' 4/ -,42 6 0$2"! /1 ' 5$3 2! ! 5,! 6 74$,! 2"* 7,1 - , -1 32$5 5421/-! 3$D ( 2" 1 , $37 (,/,! -242 / $ - - '' 4,52 - 2 ' 4$2 / = 2 ! :55 D /: 2"$2 !/3 21$ ! 2 / 32$5 /C -! (= + +1 $ - #' $5:$*! $42 2" !,5$2 ( 5$*1 2/ 01'/13 1! !2 6 !: 24" ( 0" /3 $ , -1 2" $-- 2 / $5 32$5 5421/- /' 2 , ( $ - 5= 2" 1!042 2/ 2" 0/0,5$1 -6 4 /' 2 2 455 G:/ 2= $5= 01'/13- 2" " (" 1!/5,2 / 21$ !3 !! / 5421/ 3 41/!4/0* 2/ - 1425* 01/7 2" 4/ -,42 6 0$2" 3/!25* 4/30/!- /' /C*( 6$4$ 4 != #,$ ( 2= $5= '$71 4$2- $ 2 455 2/ !2,-* 1! !2 6 !: 24" ( 01/012 ! $2 1//3 2301$2,1H= ,$ 2= $5= 01/-,4- ,+1 2 $ - +1 2 -6 4! 2/ 6!2 ($2 2" '1 5421/'/13 ( 01/4!!I= / 2= $5= 01/0/!- 2"$2 45/4D: ! $ - 4/, 2145/4D: ! 7 0/5$1 !: 24" ( 7"$6 /1! 6/56- 2" !21,42,1! /' 5 2 $ - 2 5= 2"! C2 ! 6 :/1D! /1 , $2/3! C !2 $2 2" 21'$4 /' ,+1 $ - +1 1$42 : 2" 2" /C - 5$*1 /' /1 +1 2/ 3/-,5$2 2" 4/ -,42 6 0$2" , -1 2" 6/52$(= ($1:$5 2= $5= /7!16- 2" /C*( 6$4$ 4* -1 6 5421/'/13 ( 01/4!! $ /65 5$21$5 2 2 -6 4 :"1 2" 32$5 5421/-! /' 2 5/4$2 $2 7/2" ! -! /' 2" 5$*1 - !2 (, !"- '1/3 2" /13$5 612 4$5 2 2 -6 4!= " ! $1/,!- ,! 2/ 01'/13 2" '/13$2 / /' 4/ -,42 6 0$2" / 2" 5$21$5 !,1'$4 :" 4" $1 : -5* 6!2 ($2- 2" ', -$3 2$5 !,1'$4 !4 4 !2,- != ! -! $!1 2= $5= $ $5*;- 2" 21$ !'/13$2 / 0" /3 $ 72: 2" !,5$2/1 $ - 32$5 !,1'$4 !0,221- : 2" 1J / !@= ' $ !3$55 $3/, 2 /' / ! 2 / ! /1 , / ! $1 !0,221- / 2" !,1'$4 ($1:$5 01/0/!- 5$21$5 2 2 -6 4 :"$2 :/,5- 7
"EWBODFTJO$FSBNJDTGPS&OWJSPONFOUBM 'VODUJPOBM 4USVDUVSBM BOE&OFSHZ"QQMJDBUJPOT
153
"$00 - / 2" !,1'$4 /1 2" !,7!,1'$4K 44,11 4 /' @J /1 J / !K ' 2"* C !2 :"2"1 2" 7/ - ( 4/30/!- /' 2" 4/ -,42 6 0$2" ! /7!16-9K /55/: ( 2"! B,!2 / ! : !,00/! 2"! /1 , $2/3! ' 1!25* $-!/17 / 2" !,1'$4 $ - 2" !2,-* 2" 30$42 /' 2" 1 $-!/102 / 4/ ' (,1$2 / / 2" '/13$2 / /' 7/ - (= = .# 1!25* : 7, 5- LL@ !,01455! 5$22 4 0$1$321! $ M 7 M I=HH N 4 M H=HFF N $! $ 2 $5 !21,42,1 / 2" 01 3 2 6 455 /' 1,2 5 $ M 7 M 9=