|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001795408
007 cr mnu|||uuuuu
008 070709s2006 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001566
Kennedy, Jermaine L.
Investigations of fiber optic temperature sensors based on Yb:Y3Al5O12
h [electronic resource] /
by Jermaine L. Kennedy.
[Tampa, Fla] :
b University of South Florida,
ABSTRACT: This dissertation presents the development of temperature sensors which employ a fiber-optic probe consisting of single crystal YB3BAlB5BOB12B (YAG) fiber with a phosphor of short length grown directly onto one end using the laser heated pedestal growth method. The response of all the crystalline temperature sensors derives from the temperature-dependent decay time of fluorescence. Yb3+P ions served as the fluorescer, while the addition of various rare-earth codopants (i.e., NdP3+ and ErP3+) with YbP3+ provided an additional path in the form of phonon assisted energy transfer. With the additional nonradiative decay path, the temperature sensors exhibited a more desirable response. A thermally compensated fluorescence decay rate fiber optic temperature sensor was demonstrated for the first time experimentally to the best of our knowledge to make accurate surface temperature measurements. Overall, this novel technique is envisioned to aid in the perpetual challenge of precise surface temperature measurements in comparison to current methods, with the emphasis in the area of rapid thermal processing of semiconductors.
Dissertation (Ph.D.)--University of South Florida, 2006.
Includes bibliographical references.
Text (Electronic dissertation) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 164 pages.
Adviser: Nicholas Djeu, Ph.D.
Phonon assisted energy transfer.
Surface temperature measurements.
x Applied Physics
t USF Electronic Theses and Dissertations.
Investigations of Fiber Optic Temperature Sensors Based on Yb:Y 3 Al 5 O 12 by Jermaine L. Kennedy A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Physics College of Arts and Sciences University of South Florida Major Professor: Nicholas Djeu, Ph.D. Wei Chen, Ph.D. Dennis Killinger, Ph.D. Myung Kim, Ph.D. Pritish Mukherjee, Ph.D. Date of Approval: April 6, 2006 Keywords: energy transfer, multiphonon rela xation, phonon assisted energy transfer, thermal compensation, surface temperature measurements Copyright 2006, Jermaine L. Kennedy
ACKNOWLEDGMENTS During the past four years of my gra duate dissertation work, I have benefited from the guidance, mentorship, and advice from a pioneer in the field of laser physics, Dr. Nicholas Djeu. My major professor Dr. Djeu has continuous ly reinforced the importance of completing all ta sks correctly through the years. I would like to thank Dr. Wei Chen, Dr. Dennis Killinger, Dr. Myung Kim, Dr. Pritish Mukherjee, and Dr. Boris Shekhtman for serving on my dissert ation committee. I would also like to thank Dr. Anthony Buonaquisti for always having time to have many enlightening discussions with me. I would also like to thank Mr. Sam Valenti for his expertise in the field of machining, plus Mr. Phil Bergeron, Mrs. Evelyne Keeton-Williams, and Ms. Sue Wolfe for their assistance over the years. Sp ecial thanks MicroMater ials Inc. (MMI) of Tampa, FL for their generosity with the donation of laser diodes and digital signal processor to facilitate in the completion of this work.
iv LIST OF TABLES Table 4.1. Selected properties of YAG. 48 Table 4.2 Ionic radii of rare earth ions. 49
i TABLE OF CONTENTS LIST OF TABLES iv LIST OF FIGURES v LIST OF ABBREVIATIONS xi ABSTRACT xiii CHAPTER 1 INTRODUCTION 1 1.1 Background Information 1 1.2 Scientific and Industrial Need s for Novel Temperature Sensors 3 1.3 Requirements of Temperature Sensors in Coal Gasifiers 7 1.4 Requirements of Temperature Sensors for Surface Measurements 9 1.5 Review of Non-Optical Temperature Measurement Devices 9 1.5.1 Electrical Thermometry 10 1.5.2 Acoustic Thermometry Methods 14 1.6 Review of Optical Based Temperature Measurement 14 1.6.1 Distributed Temperature Sensors 15 1.6.2 Remote Temperature Pyrometers/Radiation Thermometry 16 1.6.3 Interferometric Temperature Sensors 18 1.6.4 Fiber-Optic Fluorescence Decay Rate (FDR) Temperature Sensors 19 220.127.116.11 Historical Foundations on Fluorescence Decay Rate Temperature Sensors 20 CHAPTER 2 ELEMENTARY PRINCIP LES OF FLUORESCENCE DECAY RATE (FDR) THERMOMETRY 24 2.1 The Fluorescence Process 24 2.2 Elements Altering the Fluorescence Process of Rare Earth Impurity Ions 27 2.2.1 Rare Earth Impurity Ion Concentration 28 2.2.2 Effects of Saturation 29 2.2.3 Host Lattice Associated Impurities 30 2.2.4 Interactions of Differe nt Rare Earth Impurity Ions 30 2.2.5 Rare Earth Impurity Ion Particle Size Effects 32 2.3 Temperature Effects on the Fluorescence Lineshift, Intensity, and Absorption 32
ii CHAPTER 3 THEORIES OF MU LTIPHONON RELAXATION AND PHONON-ASSISTED ENERGY TRANSFER IN RARE EARTH DOPED CRYSTALS 34 3.1 Theory of Multiphonon Relaxation In Rare Earth Doped Crystals 35 3.1.1 Oribt-Lattice Interactions 36 3.1.2 Energy Gap Dependence 38 3.1.3 Temperature Dependence 39 3.2 Theory of Phonon Assisted Ener gy Transfer In Rare Earth Doped Crystals 40 3.2.1 Energy Gap Dependence of Transfer Probability 41 3.2.2 Temperature Dependence of Transfer Probability 42 CHAPTER 4 FABRICATION OF FIBE R-OPTIC FLUORESCENCE DECAY RATE TEMPERATURE SENSORS 43 4.1 The Laser Heated Pedestal Growth (LHPG) Process 44 4.2 Advantages of the Laser Heated Pedestal Growth (LHPG) Technique 45 4.3 The LHPG Station at the University of South Florida 46 4.4 Properties of Yttrium Aluminum Garnet (YAG) 47 4.5 Fabrication of the Single Cr ystal YAG Leads for Temperature Sensing 48 4.6 Preparation and Fabrication of the Rare Earth Doped YAG Phosphor Tips 49 4.7 Distribution of Rare Earth Ions In YAG 54 CHAPTER 5 YTTERBIUM BASED FLUORESCENCE DECAY RATE FIBER OPTIC TEMPERATURE SYSTEMS 56 5.1 Spectroscopic Properties of Yb:YAG 57 5.1.1 Experimental Details for Yb Doped YAG Sensors 59 5.1.2 Expermintal Results for Yb Doped YAG Sensors 61 5.2 Spectroscopic Properties of Nd,Yb:YAG 88 5.2.1 Experimental Results for the Nd,Yb:YAG Sensors 88 5.3 Spectroscopic Properties of Er,Yb:YAG and Er:YbAG 98 5.3.1 Experimental Results for the Er,Yb:YAG and Er:YbAG Sensors 98 CHAPTER 6 THERMALLY COMPENSATED FLUORESCENCE DECAY RATE TEMERATURE SENSOR 107 6.1 Thermally Compensated Fluorescence Decay Rate Temperature Sensor Model 109 6.2 Experimental Details and Results for the Thermally Compensated FDR Temperature Sensor 118 CHAPTER 7 CONCLUSIONS 130 REFERENCES 132
iii APPENDICES 139 Appendix A: Fluorescence Stability Appendix B: Thermal Compensation Model 160 ABOUT THE AUTHOR End Page
v LIST OF FIGURES Figure 1.1 Nickel-Chromium vs. Nickel -Aluminum Thermocouple Diagram. 11 Figure 2.1 Absorption and fluorescence transitions according to the configurational coordinate model. 26 Figure 2.2 Three level system with one fluorescence level and quick decay process. 26 Figure 5.1 Relevant energy level diagra m of ytterbium pertaining to the operation of fluorescence decay rate temperature sensing. 58 Figure 5.2 Schematic of the optical layout for fluorescence decay rate measurements. 60 Figure 5.3 Temporal evolution of 1040 nm fluorescence of Yb 3+ ions ( 2 F 5/2 2 F 7/2 ) at 300 K after pulsed excitation of the Yb 3+ ions at 940 nm. 63 Figure 5.4 Temporal evolution of 1040 nm fluorescence of Yb 3+ ions ( 2 F 5/2 2 F 7/2 ) at 1,714 K after pulsed excitation of the Yb 3+ ions at 940 nm. 64 Figure 5.5 Temporal evolution (in semi-logarithmic scale) of 1040 nm fluorescence of Yb 3+ ions ( 2 F 5/2 2 F 7/2 ) at 300 K after pulsed excitation of the Yb 3+ ions at 940 nm. 65 Figure 5.6 Temporal evolution (in semi-logarithmic scale) of 1040 nm fluorescence of Yb 3+ ions ( 2 F 5/2 2 F 7/2 ) at 1,714 K after pulsed excitation of the Yb 3+ ions at 940 nm. 66 Figure 5.7 Low temperature expanded view of the experimental data on fluorescence decay rate versus temperature for the 5%Yb:YAG sensor. 67 Figure 5.8 Experimental data on fluorescence decay rate versus temperature for the 5%Yb:YAG sensors. 68 Figure 5.9 12 hour fluorescence decay ra ta data for 5%Yb:YAG phosphor at 1,149 K +/1.3 K. 69
vi Figure 5.10 12 hour fluorescence decay ra ta data for 5%Yb:YAG phosphor at 1,248 K +/0.68 K. 70 Figure 5.11 12 hour fluorescence decay ra ta data for 5%Yb:YAG phosphor at 1,344 K +/0.70 K. 71 Figure 5.12 12 hour fluorescence decay ra ta data for 5%Yb:YAG phosphor at 1,439 K +/0.84 K. 72 Figure 5.13 12 hour fluorescence decay ra ta data for 5%Yb:YAG phosphor at 1,530K +/0.46 K. 73 Figure 5.14 12 hour fluorescence decay ra ta data for 5%Yb:YAG phosphor at 1,622 K +/1.04 K.. 74 Figure 5.15 12 hour fluorescence decay ra ta data for 5%Yb:YAG phosphor at 1,714 K +/0.60 K.. 75 Figure 5.16 Relative sensitivity versus temperature (in semi-logarithmic scale) for 5%Yb:YAG sensor. 78 Figure 5.17 Low temperature expanded vi ew of the experimental data on fluorescence decay rate versus temperature for the 10%Yb:YAG sensor. 79 Figure 5.18 Low temperature expanded vi ew of the experimental data on fluorescence decay rate versus temperature for the 20%Yb:YAG sensor. 80 Figure 5.19 Low temperature expanded vi ew of the experimental data on fluorescence decay rate versus temperature for the 50%Yb:YAG sensor. 81 Figure 5.20 Experimental data on fluorescence decay rate versus temperature for the 10%Yb:YAG, 20%Yb:YAG, and 50%Yb:YAG sensors. 82 Figure 5.21 Relative sensitivities versus temperature for the 10%Yb:YAG, 20%Yb:YAG, and 50%Yb:YAG sensors. 83 Figure 5.22 Experimental data on fluorescence decay rate versus temperature for the YbAG sensor. 84 Figure 5.23 The multiphonon transition rate from the 2 F 5/2 manifold of Yb 3+ in YbAG is shown as a function of temperature. 87
vii Figure 5.24 Energy levels in ytterbiu m and neodymium relevant to the operation of Nd,Yb:YAG sensors. 89 Figure 5.25 Experimental data on fluorescence decay rate versus temperature for the 2%Nd,20%Yb:YAG sensor. 91 Figure 5.26 Low temperature expanded vi ew of the experimental data on fluorescence decay rate versus temperature for the 2%Nd,20%Yb:YAG sensor. 92 Figure 5.27 Relative sensitivity vers us temperature for 2%Nd,20%Yb:YAG sensor. 93 Figure 5.28 Experimental setup for side fluorescence measurements for the 2%Nd,20%Yb:YAG sensor. 96 Figure 5.29 Experimental side fluorescence measurements for the 2%Nd,20%Yb:YAG sensor. 97 Figure 5.30 Energy levels in erbium and ytterbium relevant to the operation of Er,Yb:YAG sensors. 99 Figure 5.31 Experimental data on fluorescence decay rate versus temperature for the 2%Er,20%Yb:YAG sensor. 100 Figure 5.32 Temporal evolution of 1040 nm fluorescence of Yb 3+ ions ( 2 F 5/2 2 F 7/2 ) at 300 K and 1,145 Kafter pulsed excitation of the Yb 3+ ions at 940 nm for th e 2%Er,20%Yb:YAG sensor. 102 Figure 5.33 Experimental data on fluorescence decay rate versus temperature for the 5%Er:YbAG and 10%Er:YbAG sensors. 105 Figure 5.34 Low temperature expanded vi ew of the experimental data on fluorescence decay rate versus temperature for the 5%Er:YbAG and 10%Er:YbAG sensors. 106 Figure 6.1 Simplified model for the thermal transport mechanisms of the thermally compensated FDR YbAG sensor. 110 Figure 6.2 Phosphor temperature versus gap length using the theoretical equation 6.1 with various la ser dissipation values. 115
viii Figure 6.3 Phosphor temperature differe nce between the gap distances of 0 m and 100 m for each initial heating condition versus the initial heating temperatures at the specific gap distances of 0 m (hollow circles) and 100 m (squares). 116 Figure 6.4 Phosphor temperature versus ga p length at various laser dissipation values using the theoretical e quation 6.1 with extractions of specific gap length distances of 10 m and 50 m from Fig. 6.2. 117 Figure 6.5 Schematic of the experiment al layout for fiber-optic thermally compensated YbAG temperature sensor. 119 Figure 6.6 Low temperature fluorescence decay rate versus temperature for the YbAG phosphor. 122 Figure 6.7 Relative sensitivity of the YbAG sensor between room temperature and 140 C. 123 Figure 6.8 Experimental data of the phos phor temperature versus gap distance for various laser launching dissipation values for the YbAG phosphor. 125 Figure 6.9 Phosphor temperature versus gap length at various launching laser dissipation values with extractions of specific gap distances of 0 m and 100 m from Fig. 6.8. 127 Figure 6.10 Phosphor temperature differen ces between the ga p distances of 0 m and 100 m for each internal heati ng data set in Fig. 6.8 versus the initial heat temperatures for each specific gap distance of 0 m (hollow circles) and 100 m (squares). 128 Figure A.1 12 hour fluorescence decay rate data for 10%Yb:YAG phosphor at 1,118 K +/1.8 K. 139 Figure A.2 12 hour fluorescence decay rate data for 10%Yb:YAG phosphor at 1,215 K +/2.0 K. 140 Figure A.3 12 hour fluorescence decay rate data for 10%Yb:YAG phosphor at 1,311 K +/2.2 K. 141 Figure A.4 12 hour fluorescence decay rate data for 10%Yb:YAG phosphor at 1,399 K +/1.6 K. 142 Figure A.5 12 hour fluorescence decay rate data for 10%Yb:YAG phosphor at 1,483 K +/2.3 K. 143
ix Figure A.6 12 hour fluorescence decay rate data for 10%Yb:YAG phosphor at 1,575 K +/5.4 K. 144 Figure A.7 12 hour fluorescence decay rate data for 10%Yb:YAG phosphor at 1,665 K +/2.0 k. 145 Figure A.8 12 hour fluorescence decay rate data for 10%Yb:YAG phosphor at 1,686 K +/2.6 K. 146 Figure A.9 12 hour fluorescence decay rate data for 20%Yb:YAG phosphor at 1,118 K +/1.9 K. 147 Figure A.10 12 hour fluorescence decay rate data for 20%Yb:YAG phosphor at 1,221 K +/1.5 K. 148 Figure A.11 12 hour fluorescence decay rate data for 20%Yb:YAG phosphor at 1,317 K +/2.6 K. 149 Figure A.12 12 hour fluorescence decay rate data for 20%Yb:YAG phosphor at 1,403 K +/2.9 K. 150 Figure A.13 12 hour fluorescence decay rate data for 20%Yb:YAG phosphor at 1,490 K +/2.0 K. 151 Figure A.14 12 hour fluorescence decay rate data for 20%Yb:YAG phosphor at 1,571 K +/2.8 K. 152 Figure A.15 12 hour fluorescence decay rate data for 20%Yb:YAG phosphor at 1,661 K +/1.9 K. 153 Figure A.16 12 hour fluorescen ce decay rate data for 50%Yb:YAG phosphor at 1,124 K +/1.2 K. 154 Figure A.17 12 hour fluorescen ce decay rate data for 50%Yb:YAG phosphor at 1,221 K +/1.8 K. 155 Figure A.18 12 hour fluorescen ce decay rate data for 50%Yb:YAG phosphor at 1,316 K +/2.2 K. 156 Figure A.19 12 hour fluorescen ce decay rate data for 50%Yb:YAG phosphor at 1,406 K +/1.9 K. 157 Figure A.20 12 hour fluorescen ce decay rate data for 50%Yb:YAG phosphor at 1,495 K +/2.7 K. 158
x Figure A.21 12 hour fluorescence decay ra te data for 50%Yb:YAG phosphor at 1,581 K +/2.8 K. 159 Figure B.1 Schematic model for thermal transport between the phosphor tip and sample. 160 Figure B.2 Schematic model for ther mal transport be tween the phosphor sidewalls and sample. 162
xi LIST OF ABBREVIATIONS A Amps CJC Cold Junction Condensation CW Continuous Wave Er Erbium FDR Fluorescence Decay Rate FWHM Full Width at Half Maximum IGCC Integrated Gasification Combined Cycle IR Infrared LHPG Laser Heated Pedestal Growth mm Millimeter mW Milliwatt Nd Neodymium NIR Near Infrared nm Nanometer RF Radio Frequency RTD Resistance Temperature Detectors UV Ultraviolet m Micrometer YAG Yttrium Aluminum Garnet YbAG Ytterbium Aluminum Garnet
xii Yb Ytterbium Y Yttrium
xiii Investigations of Fiber Optic Temperature Sensors Based on Yb:Y 3 Al 5 O 12 Jermaine L. Kennedy ABSTRACT This dissertation presents the developm ent of temperature sensors which employ a fiber-optic probe consis ting of single crystal Y 3 Al 5 O 12 (YAG) fiber with a phosphor of short length grown directly ont o one end using the laser heat ed pedestal growth method. The response of all the crysta lline temperature sensors de rives from the temperaturedependent decay time of fluorescence. Yb 3+ ions served as the fluorescer, while the addition of various rare-e arth codopants (i.e., Nd 3+ and Er 3+ ) with Yb 3+ provided an additional path in the form of phonon assist ed energy transfer. With the additional nonradiative decay path, the temperature sensor s exhibited a more desirable response. A thermally compensated fluorescence decay rate fiber optic temperature sensor was demonstrated for the first time experiment ally to the best of our knowledge to make accurate surface temperature measurements. Overall, this novel technique is envisioned to aid in the perpetual challenge of preci se surface temperature measurements in comparison to current methods, with the empha sis in the area of rapid thermal processing of semiconductors.
1 CHAPTER 1 INTRODUCTION 1.1 Background Information on the Proposed Research As one of the seven basic quantities in the SI (International System of Units) system, temperature is probabl y the most measured physical parameter since virtually every process in nature and industry is temperature dependent. The accelerating proliferation of technology in the applied fiel ds of science, engin eering and medicine has led to an ever-increasing vari ety of situations th at requires the knowle dge of the accurate temperature of some system, process, component or specimen. As a result, new techniques and instrumentation are need ed depending on temperature measurement requirements in different processes and working environments. One novel instrumentation system that is ad aptable to the needs of a wide variety of situations is based on fluorescing materi als. The thermal dependence of fluorescent thermometry may be exploited to provide for emissivity-independent, contact or noncontact, optical alternative to other more conventional t echniques, e.g., those employing pyrometry, thermocouples or thermistors. There are certain situ ations in which the advantages of fluorescence based thermo metry over other methods make it the only useful approach. The motivation of this research is twofold. The first is to meet the recent increasing needs for temperature sensors capab le of operating accurate ly and reliably in harsh environments, such as coal-based pow er generation and dist ribution industries,
2 nuclear power industries, glass and metal ma nufacturing/processing i ndustries and other high temperature chemically corrosive envi ronments. The second motivation of this research is to make accurate measurements of surface temperatures. In general, optical sensors offer many advantageous features over conventional elec trical sensors for applications in harsh environments. These include: immunity to electromagnetic interference (EMI), resistan ce to chemical corrosion, high sensitivity, large bandwidth, capability of remote operation, small size, light weight, absence of ground loops and capability of operation at high temperatures [1,2]. These advantages have promoted worldwide research activities in the area of optical fiber se nsor technologies for harsh environments. The sensor prototype developed in this research is intended for non-intrusive, direct temperature measurement, in the primary and secondary stages of slagging gasifiers and the precise measurement of surf ace temperatures. These gasifiers are used in the coal-based power gene ration industries, where a sens or is required to withstand extremely harsh environments imposed by high temperature, high pressure and corrosive chemical materials. In this researc h, innovative techniques are presented for high temperature sensors capable of operating at temperatures up to 1600C. These sensors fulfill the need for real-tim e monitoring and long-term di rect measurement of high temperatures in extremely harsh environments The efforts devoted to innovative sensing techniques focus mainly on the following issu es: 1.) Fabrication ma terial selection and structure design of probe. 2.) Optimizati on of sensor prototype and performance evaluation. 3.) Sensor prototype implementation.
3 1.2 Scientific and Industr ial Needs for Novel Temperature Sensors The majority of commercially available temperature measurement instruments are made using only a few basic types of devices : thermocouples, resistance thermometers, liquid-in-glass thermometers and radiation thermometers. These traditional measurement instruments have been in use for many decades The major sources of their instability or drift, as well as possible systematic errors are well understood. As temperature sensing technology matures, many traditional techni ques are still unsuitable for specific measurement requirements. Conventional temperature measurement se nsors are generally inexpensive while being widely available for scientific and i ndustrial applications. Innovation and research and development activities in temperature measurements need to be pursued to enhance the performances attainable by traditional techniques concerning measurement sensitivity, accuracy, response time and range limitations. Temperature measurement requirements necessary in todays modern industrial environments include monitoring surface/internal temperature profiles, measurements in hostile environments and gas temperature measurements. All of these environments en tail non-intrusive temperature acquisitions in which the sensor does not affect the medium of interest. One representation of a harsh environment is an entrained flow slagging gasifier, which is one of the main units among coal ga sification facilities. In the new emerging coal-fired power plants for advanced power generation, the coal gasification technique  has been developed to generate electricity and other high-value energy products without extremely harmful byproducts. Rather than burn ing coal directly, coal gasification reacts coal with steam using carefully cont rolled amounts of air or oxygen under high
4 temperatures and pressures. Hot exhaust from the gas turbine is then fed into a steam turbine, producing a second source of power This unique inte grated gasification combined cycle (IGCC) configuration of tu rbines offers major improvements in power plant efficiencies compared with conventional coal combustion. To optimize performance for these IGCC plants, certain relevant physical parameters should be monitored and controlled precisely for coal gasification processes, such as real-time accurate and reliable monitoring of temperatures at various locations in a coal gasifier, pressure distribution monitoring, burning material flow patterns inside the gasifier, air or oxygen monitoring, etc. The accurate continuous monitoring of the in terior temperature at several strategic locations enables the timely adjustment s of the input streams for an optimum performance. Prevailing harsh conditions of high temperature, high pressure and reactive constituents within the gasifier make extr aordinary demands on the temperature sensor. The gasifier must be operated at a temperature high enough for th e ash in the fuel, such as coal, to melt and become sufficiently fluid to flow out of the gasifier through the bottom tap-hole. Load changes will also affect the temperature in the gasifier and downstream, which would require adjusting the operati ng conditions. Operat ing under the optimum temperature would cause the molten slag to become viscous or solidify, plugging up the tamp-hole and preventing additional slag from draining out of the gasifier. Eventually the gasifier has to be shut down, cooled off, and the slag, now in the form of a hard vitereous rock, has to be manually chipped out to be removed from inside the gasifier. The shut down to clean up the slag may ta ke weeks, resulting in a lengthy loss of production. Operation above the optimal te mperature would significantly shorten the
5 lifetime of the refractory lining. In addition, more of the alkali species in the ash would be volatilized, reacting with the ash particles entrained in the gas to form low temperature eutectics which deposit in the cooler sections of the gasifi er or downstream equipment, such as the boiler, causing plugging problems. Exceeding the operating temperature will also reduce the conversion effici ency of the gasification proc ess in the production of the synthesis gas. In order to realize the full economic potential of the gasification systems, there is an increasing need to utilize a wide variety of feedstock in additi on to coals, such as biomass, refuse, wood wastes, etc. in the ga sification plants. The ash properties of these various feedstocks vary signi ficantly, and would thus requir e operating the gasifier at different temperatures to facilitate slag tapp ing. The flow condition inside the gasifier is highly turbulent, with many pockets of gas r ecirculation zones. Te mperature in various regions of the gasifier could be widely di fferent. The temperature at the exit of the burners could be above 1927 C (3500 F), where as close to the wall temperatures could drop to less than 1316 C (2400 F). Real-time accurate an d reliable monitoring of temperatures at various locations in a gasifier is thus highly desirable. Various methods for measuring temperatur e in harsh environments have been investigated in the past [4-8]. Among th ese are optical and acoustical pyrometers, and high temperature thermocouples. However, due to the stringent surroundings involving entrained molten slag, typical temperature sensors and measurement devices are very difficult to apply [9-13]. At present, retr actable thermocouples protected by thermowells are being used for gasification systems. Du e to decalibrations of the thermocouple and corrosion of the thermowell, the total exposure time of the sensor probe is limited to only
6 a few days to a few weeks. Decalibra tions which are caused by the diffusion of impurities into the thermocouple junction at high temperatures, is an inherent problem. This makes thermocouples unsuitable for ex tended deployment even if the thermowell corrosion problem can be overcome. Non-contact optical pyrometer applications involve an infrared transparent high temperature window on the gasifier wall to main tain a large pressure differential, while allowing transmission of the infrared ra diation emitted by the product gases to the detector placed outside the gasifier. Obstruction of the sight-path opening for the pyrometers in the refractory wall by molten sl ag is a major problem. Shifting of the refractory lining, which could have a thickness of 2 feet or more, during heat-up and operation processes could also cause blocka ges of the sight-path. In addition, the measurement is also subject to interfer ence from the radiation emitted by entrained particles or the relatively cold refractory walls. Direct contact temperature measurement is preferred since it will give the measurement for the specific location of intere st. Several of these de vices installed at the critical locations inside the gasifier coul d thus provide the temperature profile, and performance of the gasifier could be monitored and improved by making adjustments in a timely manner. However, no direct contact m easuring devices are available to date due to material issues. The highl y corrosive molten slag attack s both metals and ceramics. Ceramic materials are also susceptible to a ttack from alkali vapors in the gas [14-18]. This situation suggests for innovative techniques that can operate in gasifier harsh environments for real-time, reliable monito ring of temperature to be developed.
7 Another prevailing challenge in thermome try involves the precise measurement of surface temperatures. Radiometry is curre ntly the most conventional means in the noncontact technique while being highly unreliable unless the em issive properties of the surface are highly characterized. When this t echnique is used emissivity and reflection errors result, then the apparent and not the ac tual temperature is indicated. The second objective of the proposal is to demonstrate the ability of a fluores cence-decay-rate (FDR) fiber-optic temperature sensor to measure surface temperature with high accuracy when operated in a thermally compensated mode. 1.3 Requirements for Temperature Sensors in Coal Gasifiers For optical temperature sensors to be usef ul in coal gasifier environments, there are several requirements that must be met. The operating temperature in the coal gasifier is in the range of 1200C 1600C, depending on the physical locations in the chamber. The high temperature is the main reason most electronic sensors are inapplicable. Some optical sensors cannot be deployed because of the limitations of the thermal properties pertaining to the fabrication materials used. For example, silica optical fibers can only withstand temperatures up to 800C before th e dopants start to thermally diffuse. To measure high temperatures accurately in a wi de measurement range with high resolution, proper fabrication materials ar e essential, as well as a simple and stable mechanical structure of the sensing probe. Pressures as high as 500 psi can be encountered in the gasifier chamber. In order to be able to su rvive in such high pre ssure environments, the optical temperature sensing probe must be constructed and fabricated with ample
8 mechanical strength while having its optical pa ths entirely sealed to provide the necessary protection. Optical temperature sensors for volatile e nvironment sensing applications must be thermally stable. Temperature related degradation mechanisms, such as thermal shock, thermal cycling, thermal stress, and thermal fatigue from high heat fluxes, must be considered in the elevated temperature sensor design for long term stable measurements. Rigorous mechanical structures and special fabrication materials are crucial to promote the thermal stability and deployability. Th ese sensors must be capable of remote operation and be flexible enough for easy depl oyment. Features such as mechanical vibration-proof, high mechanic al strength, and remote mon itoring and control capability are thus inevitable. With temperatures exceeding 1200C, pressures exceeding 500 psi, and chemically corrosive agents such as alkalis, sulfur, transition metals and steam, it is hard to find a material that is impervious to such an extensive corrosive attack. Conventionally, commercially available temper ature sensors exhibit greatly abbreviated lifetimes due to the hostile environment. Proper fabrication materials are needed to implement the sensing probe to be chemica lly corrosion resistant. In addition, the sensors are required to have self-calibration capability so that the guiding fiber loss variations and the source power fluc tuations can be fully compensated. As the market for optical temperature sensors for harsh environment grows rapidly, the cost of the sensors and instrume ntation is becoming a concern of increasing importance. In order to achieve successful commercialization, optical temperature sensor systems must be robust as well as low cost. This requires that the complexity of the
9 sensor system is kept to the minimum and th e technique and process of fabricating sensor probes have the potential of allowing mass production. 1.4 Requirements of Temperature Sens ors for Surface Measurements Any apparatus brought near to, or in contact with, the source-body surface changes the heat-transfer rates between the surf ace area and its environm ent. The rates of surface chemical reactions, evaporation, and co ndensation may be similarly affected. To obtain high accuracy in the contact mode, the te mperature sensor has to be fused onto the surface by cement or inserted into a well dr illed in the sample. The lack of such measures results in probes with accur acies no better than 5-10% . The first error in surface measurements stems from the actual area of contact that is generally a small fraction of the apparent contact area because of the surface roughness of the source-body to be measured, generati ng a layer of high thermal resistance between the sample and the probe. The probe itself ma y cause thermal perturbations at the point of contact on the source-body to be measured. The result of this blanketing or thermal damming is that the temperature to be measured is altered by the means of measurement causing large inaccuracies in surface temperature measurements. The proposed FDR temperature sensor will avoid these problema tic issues that have lingered for decades. 1.5 Review of Non-Optical Based Te mperature Measurement Devices In science and technology, temperature is defined in terms of the amount of heat transferred in a Carnot cycle . This is not the most practical way of measuring temperature, and in practice many differ ent techniques are used depending on the
10 temperature measurement requirements. In practice, every temperature measurement involves the use of certain calib rated transducers to convert a measurable quantity into a temperature value. These tran sducers to convert changes in the temperature into other measurable physical quantities, such as el ectromotive force (thermocouple), volumetric expansion (liquid thermomete r), resistance (restistance te mperature detector-RTD), radiated energy (radiation thermometer), dimensional change (bimetallic thermometer), or some other characteristics of a material that varies reproduci bly with temperature [21,3]. For high temperature measurements in excess of 1000 C, the existing nonoptical measurement techniques are very limited. Possible choices include high temperature thermocouples and acoustic methods. 1.5.1 Electrical Thermometry The most universal electrical temperatur e sensors found in society today detect deviations in resistance or voltage with temperature. Resistance thermometry is an electrical method that is based on the variation of resistance for temperature measurements. A resistance thermometer make s use of the change of resistivity in a metal wire with temperature. As elect rons migrate through a metal, the thermal vibrations of the atoms in the crystal latti ce impede them. The impedance and resistivity are both proportional to temperat ure and thus temperature depende nt. This effect is very marked in pure metals. Resistance thermometers are usually more sensitive than other electrical temperature sensi ng devices but operate over a smaller range (248 K 1125 K). Potential limitations for these devices are cau sed by alterations in lead temperature and excess power to th e circuits .
Thermocouples are the most common electrical thermometry device used to make temperature measurements. These devices depe nd on the principle of the Seebeck effect: when a conductor is placed in a temperature gr adient, electrons diffuse along the gradient and an electromagnetic field (emf), or therm ovoltage, is generated. The magnitude of the emf depends on the material and also on its physical condition. To measure the generated thermal emf, the circuit must be completed using a second different conductor to form a common junction. The thermocouple emf is then the difference between the emfs generated in the two conducto rs. The voltage between the wires is proportional to the ambient temperature. These devices typica lly operate in the range of 225 K 1975 K, varying depending on the type of thermocouple used. Figure 1.1: Nickel-chrom ium vs. nickel-aluminum thermocouple diagram. Standard tables show the voltage pr oduced by thermocouples at any given temperature, so for example in the above diagram Fig. 1, the K type thermocouple at 300C will produce 12.2mV. Unfortunately it is not possible to simply connect up a voltmeter to the thermocouple to measure this voltage, because the connection of the voltmeter leads will make a second, undesire d thermocouple junction. To make accurate 11
12 measurements, this must be compensated for by using a technique known as cold junction compensation (CJC). The output from ther mocouples is small which makes them not immune to error. Typical outputs are of the or der of millivolts to tens of millivolts. The major limitation in the use of thermocouples fo r precise thermometry ar ises from the fact that unwanted emfs are generated in i nhomogeneous thermoelements passing through temperature gradients. The useful temperature range for a partic ular thermocouple is somewhat arbitrary. One can argue that the upper limit is determined by the melting points of the thermocouple materials, but in practice the useful upper limit is usually well below the melting point and depends upon such factors as operating time at el evated temperatures, environment, thermoelement diameter, and th e exact conditions of use. The lower limit is also somewhat arbitrary, but it is usually determined by impractically small thermopower. There are numerous types of thermoc ouples available, which are formed by different metals. Eight types of them are standa rdized , including T, E, K, J, N, B, R, S type thermocouples. Different types of th ermocouples are characterized to measure up to certain levels with different resolutions. For high temperatures over 1000 C, Type B, R, S thermocouples are commercially available. In chemically corrosive environments, high temperature thermocouples th at utilize precious metals are used and have a limited life of only a few days due to their suspectibil ity to attack from corrosive chemicals. They drift significantly unde r high temeperature environm ents for long-term operation and are susceptible to decalib rations due to diffusion of impurities into the thermocouple
13 junction at high temperatures making them unsuitable for extended deployment in harsh environments. Resistance temperature detectors (RTDs) ma ke use of the fact that resistance to flow of electricity in a wire changes with temperature. Pl atinum is the most commonly used wire material. Wirewound and thin f ilm RTDs are two types of RTDs normally found commercially today. Wirewound RTDs consist of wire wound on a bobbin, which is enclosed in glass. For thin-film RTDs, a film is etched onto a ceramic substrate, and sealed. RTDs are more accurate and stable than thermocouples, but cannot be used to monitor extremely high temperatures. One limitation to impact RTDs on thermocouples has to do with application. RTDs do not f unction properly at temper atures greater than 650 C . Therefore, in cer tain industries and extremely high temperature applications such as heat treating, oven cont rol, jet engine testing, stee l making and metal fabrication, RTDs cannot be used. The exotic metals platinum and tungsten thermocouples will continue to dominate in thes e high temperature applications for the foreeable future of electrical thermometry. Like RTDs, thermistors also change resi stance with changing temperatures, but they are more sensitive than either RTDs or thermocouples. Ther mistors change their resistance much more significantly than RTDs with changing temperature. However, this change is highly nonlinear. Due to their ex treme sensitivity and non linearity, thermistors are limited to measuring temperatures of a few hundred degrees Celcius. They are less rugged than RTDs, further limiting their application.
14 1.5.2 Acoustic Methods It is well known that the speed of sound in a material depends on the temperature [23,24]. Temperature can thus be measur ed by detecting the speed of sound that propagates inside a material. This technol ogy is especially useful for measuring gas temperature in a combustion chamber, where it is difficult to measure the temperature using inserted probes due to low thermal mass and low conductivity of gases, and the strong radiation coupling of the walls of the encl osure to the sensor at high temperatures. Using the gas itself as a temperatur e sensor overcomes these problems. The prevailing difficulty w ith acoustic methods derive from the speed of sound being strongly dependent on the composition of the gas along the path, which is generally not homogeneous in the combustion chamber. Soot particles slow the acoustic wave significantly, and will result in large error. Furthermore, a dilemma arises from the refraction of the sound wave front by the de nsity and temperature gradients in the chamber. These are often tu rbulent which distort the wave fronts, making the accurate determination of time of flight problematic. Since the sound wave travels with the gas, the apparent speed of sound will also be str ongly affected by the flow velocity through the Doppler effect. Practical systems usually attempt to compensate for this effect by performing measurements in both directions. High temperature measurement uncertainties up to 30 C at 1000 C have been claimed for this method . 1.6 Review of Optical Based Temperature Measurement Optical sensors are devices in which optical signals are transformed in a reproducible way by an external physical stimul us such as temperature, strain, pressure,
15 strain, etc. An optical beam is characterized by several variables, such as variation with time, intensity, spectrum, phase and state of polarization. Many different physical phenomena related to these characteristics ar e used to perform sensing functions. Temperature sensors probably constitute the largest class of commercially available optical sensors. Besides bulk optics based sensors, a wide variety of temperature sensors using fiber optics have b een developed . They offer numerous significant advantages over electric sensors, su ch as small size, light weight, immunity to electromagnetic interference, etc. The main existing techniques for optical thermometry are remote pyrometers, thermal expansion th ermometers, fluorescence thermometers, and thermometers based on optical scattering including Raman and Raleigh scattering. 1.6.1 Distributed Temperature Sensors The interaction of light and the lattice vibration when light is incident on a substance causes a change in the energy level of the lattice vibrati on and a shift in the energy of the light that is scattered. This leads to scattering of light with a different wavelength from that of the incident light. This phenomenon is called Raman scattering, and the two components of the differing wave lengths are termed Stokes light and antiStokes light. The method used in fiber optic distributed temperature sensors is based on the fact that the intensity of the li ght components generated by Raman scattering when a light pulse is propagated through a body compos ed of optical fibers depends on the temperature of that body  This makes it possible to determine temperature distribution via Raman scattering intensities generated at various points along the optical
16 fiber path. The temperature measurement range of the fiber optic distributed temperature sensor was narrow at 223 K to 423 K using c onventional optical fibers, thus limiting the objects that could be measured. Later versions of this sensor were able to detect up to 873 K with an accuracy of 0.1 K . The utmost limitation of these sens ors are derived from optical losses. When optical loss increase is equivalent in the two components of the Raman scattering light, the accuracy of the measurement is not aff ected, but the distance at which measurement is possible decreases. However, if the optical lo ss increase of the two components differs, both measurement distance and accuracy suffer. Spontaneous Brillouin scattering based sy stems are also temperature dependent and provide a signal that is an order of magnitude greater than Raman scattering. Brillouin scattering in optical fibers result fr om the interaction between the incident light beam and thermally generated acoustic waves in the fiber. The advances in narrow bandwidth pulsed laser technology and low loss optical fibers have allowed the Brillouin signal to be separated from the Rayleigh signal Brillouin scattering offers considerably increased range beyond the limit of the spontaneous Raman based sensor with the possibility of 3 K resolution . 1.6.2 Remote Temperature Pyrometers/ Radiation Thermometry Radiation thermometry devices are typical ly used for extremely high temperature measurements where other devices would be destroyed. This method utilizes modern imaging optics and detectors to measure the fl ux created by a hot objec t or environment. Radiation thermometry makes use of the fact th at all objects emit radiation in the infrared
17 and visible parts of the spectrum, the intensity of which varies strongly with temperature. As the temperature of the perfect or id eal radiator known as a blackbody increases, radiation is emitted at all wavelengths in a continuous manner. The radiometer senses radiant flux from the target and generate s an output signal whic h, through a calibration algorithm, is used to provide a measure of target emissivity of the surface. The emissivity is a parameter that is complementary to the spectral reflectance of that particular surface. Radiometers have various origins that cause inaccurate temperature measurement readings. In practice, the ta rget seldom approximates a bl ackbody radiator and the target environment, comprised of its surroundings and the atmosphere in its line-of-sight, is not as well controlled as during the calibration process. Temperature measurements by these devices are affected by an object's emissiv ity. Most organic materials and painted surfaces have an emissivity of approximately 0.95 (the standard preset in fixed emissivity units) . Inaccurate readings result, however, when measuring shiny or polished metal surfaces. Radiance temperature tends to be less than the true temperature and a correction needs to be applied for sources with emittances that differ from 1.0 to achieve an accurate temperature measurement. The er ror associated with this method can be as much as 44% for = 0.1 . All materials with temperatures above absolute zero degree emit electromagnetic (thermal) radiation and the amount of th ermal radiation emitted increases with temperature. The measurement of the amount of thermal radiation emitted by a material can therefore be used as an indicator of its temperature. The basic operating principle of radiation thermometers is to measure part of thermal radiation emitted by an object and
18 relate it to the temperature of the objec t using a calibration curve that has been determined either experimentally or theore tically (from Plancks law) . Typical radiation thermometers measure temperature above 600 C and dominate the temperature measurement market for temperatures up to 2000 C. The main pr oblematic issues for these devices involve field of view issues. 1.6.3 Interferometric Temperature Sensors Interferometric fiber optic sensors are a large class of extremely sensitive fiber optic sensors and are applicable for measuring almost any physical quantity. They are typically used when ultra-high sensitivities are required and/or in applications of localized measurements, although sensor lengths longer than one meter are sometimes possible. This sensing technique is primarily based on detecting the optical phase change induced in the light as it propaga tes along the optical fiber. Fiber optic interferometers are generally intrinsic sensors in which light from a source is equally divided to follow two (or mo re) fiber-guided paths. The beams are then recombined to mix coherently and form a fri nge pattern that is directly related to the optical phase difference experienced between the different optical beams. The phase difference can be measured by counting the frin ges, and then transformed into a physical dimension change that gives information about the temperature. The most common configurations of the interferometers are the Mach-Zehnder and the Michelson fiber optic sensors . The Fabry-Perot (FPI) sensor published by Lee and Taylor in the early 1990s discussed an example of another fiber optic temperature sensor based on interferometry
19 . They have later descri bed improvements . This work described the use of a light emitting diode (LED) as a low coherence light source. The sensor uses two FabryPerot interferometers in series, one for sensi ng and one for reference. The optical output from the LED is spectrally modulated by reflection from the sensing FPI. Then, reflection or transmission by the referen ce FPI produces an interferometeric beat response similar to that observed when a lase r is used with the sensing interferometer alone. This particular design was only opera ble up to 575 K due to the degradation in fringe visibility with increasing optical path le ngth difference. Other literature reports a reflectively monitored Fabry-Perot temperature sensor characterized with the capacity to measure a maximum temperature of 1320 K . 1.6.4 Fiber-Optic Fluorescence Decay Rate (FDR) Temperature Sensors Although various devices may be used for the purpose of measuring temperature in science, sensors based on optical fibe rs have become widely recognized as a convenient and practical tool in remote, ina ccessible spaces, in harsh environments where electromagnetic/radiofrequency interference is relatively significant and where accurate measurements are required . Histori cally in many instances, temperature sensors based on glass fibers were employed. However, the limitations imposed by glass based temperature sensors lead to th e search for more robust and stable systems on other high temperature materials. Majority of the gla ss based thermometric systems were limited by their intrinsic material properties to temper atures less than 400 C, in general, while others based on silica fibers may be used intermittently up to approximately 1000 C. For many applications temperature meas urements are needed well beyond 1000 C,
20 making the conventional fiber optic sensors unsuitable for deployment. Hence, appropriate replacements have been sought for this importa nt temperature region and are presented in this work. Single crystal fibers have rather remarkab le properties which have led to their use in numerous optical and structural applicati ons. Although, the bulk of such applications have largely been in the areas of nonlinear optics, solid st ate laser devices, modulators, amplifiers, high power beam delivery and beam delivery in medicine . These fibers have been employed in specia lized fiber optic sensing appl ications where their optical and mechanical properties are superior to those of glass based materials [37-39]. Single crystal fibers exhibit few or almost no intern al defects, offering the possibility of high crystalline perfection and near-t heoretical strength . While early work on fiber optic rare earth based thermometry employed Nd 3+ doped glasses and crystals whose fluorescence decay lifetimes were measured as a function of temperature [38,39], these were not found suitable since such fibers exhibited non-monotonic decay times over the entire temperature range investigated. 18.104.22.168 Historical Foundations on Fluorescen ce Decay Type Temperature Sensors Suggestions on the use of phosphors for th ermometric applications by Neubert date back to as early as 1937 . During th e later 1940s and early 1950s Urbach contributed to developments and approaches pertaining to phosphor thermometry. In 1952 Bradley  presented one of the first th ermometric applications by measuring the temperature distribution on a flat wedge in a supersonic flow field using a phosphor
21 granted by Urbach. U. S. Radium Corpor ation later marketed this thermometric phosphor technology by work from these pioneers. The advent of the ruby laser by Maiman in 1960 used the efficient fluorescence of ruby, in which Cr 3+ is the activator in an Al 2 O 3 host. Its temperature dependent fluorescence properties have also been applie d to thermometry in numerous journals . Various trivalent ra re earth impurities are now used, particularly Yb 3+ Tb 3+ and Er 3+ in a variety of hosts. Th is led to continuing progres sion of technological tools during the 1970s for exploiti ng the thermal properties of phosphor emission in a broader range of applications. Examples include hi gh quality optical fibers capable with high transmission properties. By applying th e phosphor to the tip of an optical fiber, Wickersheim and Alves  produced the optical analog of a thermocouple. Their design featured prefiltration of UV excitation radiation by reflection from series of dichroic mirrors covering a temperature rang e of 9 C to 250 C. In 1979, several other methods were presented including the ex traction of temperature from phosphor time decay. Sholes and Small reported only the pr inciple of the fluorescent decay thermometer for biological applications usi ng a ruby crystal as th e active medium. The device relied upon single shot measurements using a tungsten source modulated by a mechanical camera shutter to excite the material at 10 ms intervals. The time constant in this particular sensor was a monotonically decreasing function of temperature over the range from about 200 K to 400 K with a reso lution of 0.3 K . For high temperature applications, above 575 K, conventional sensor binder and adhesives and the optical fiber buffer and jacket materials suffered from thermal degradation.
22 McCormack produced the first working fiber optic temperature sensor based on fluorescence in 1981, which again relied upon lamp excitation (250 W halogen lamp) with mechanical modulation using barium ch lorofluoride activated by divalent samarium as the active medium. The material produced a narrowband emission with high quantum efficiency, however the decay rate was not a m onotonic function of temperature . The first monolithic all crystalline fiber optic temperature sensor was reported in 1996 . Both the fiber and th e sensing element were made of crystals in this novel construction. This approach allows the sensor to operate to the melting point of the fiber, which is approximately 2200 K for YAG. Th e probe that was tested consisted of an Er:YAG transducer tip that was pumped with a tunable Ti:Sapphire laser with 10 ns output pulses at 790 nm. This system was ope rable in the temperature range of 300 K to 1000 K with a temperature uncertainty of approximately 3 K. One recent thesis by University of Sout h Florida graduate student, Dameon Henry, reported the first direct measurement of molte n aluminum with an all-crystalline fiber optic sensor consisting of a single-crystal YAG (Y 3 Al 5 O 12 ) lead fiber and an Er:YAG tip. The sensor described in this work was test ed previously to 1000 K  before being extended to the highest temperature ever record ed for a fiber optic te mperature sensor at 1520 K and successfully measuring the temper ature of a molten conductor . The sensor relied on the dependence of the fluorescence decay rate on multiphonon relaxation of the impurity ions within the host crys tal. This relaxation was dependent on temperature as well as the energy gap between the fluorescent and terminal levels. A theoretical curve was fit to th e data using physical parameters of the transducer material,
23 and was distinctive in that it revealed el ectron coupling to a s econd phonon mode at high temperatures for the first time in any material. Another recent thesis by University of South Florida graduate student, Russell Van Cleave, reported an all-crystalline fiber optic sensor consisti ng of a single-crystal YAG lead fiber and an Yb:YAG tip demonstr ating maximum temperature measurements of 1700 K with a sensitivity of 0.01 /K at 1300 K . This work re presented the highest temperature measured by a fiber optic transducer at that time, with the previous record at 1520 K. New reports have cited fiber optic te mperature sensors ba sed on fluorescence decay using monolithic crys talline construction using Er:YAG and Yb:YAG phosphors to reach as high as 1,900 K while addressing th e future envisioned developments of this type of sensor . The authors have al so reported results which suggest that the temperature dependence of phonon assisted en ergy transfer is well described by the electron coupling to optical phonon modes at the two extremities of the phonon spectrum .
24 CHAPTER 2 ELEMENTARY PRINCIPLES OF FLUORESCENCE DECAY RATE (FDR) THERMOMETRY It has long been generally known that the fluorescence of a material is temperature dependent. There are numerous ways in which this temperature dependence is manifested. Fluorescence typically corresponds to weakly allowed transitions between electronic levels of the phosphor The material emits light because there may be no other way for the electron, once excited by incident radiation, to give up its energy and return from the metastable fluorescent state to the ground state. A variety of competitive processes, some radiative and some nonradiative, exist. Hence all phosphor materials can be expected to exhibit temperature dependent decay time over some temperature interval. Since a number of determinants underlie the in trinsic nature of the fluorescence process, we begin with a concise overview of it. 2.1 The Fluorescence Process The fundamental aspects of the fluor escence of solid materials are well understood. Prior to excitation, an active ions electrons occupy only the thermally accessible states according to the Boltzmann di stribution. A means to deposit energy in the material is required in order to excite a higher electronic state. This may be attained by exposure to electromagnetic radiation (visible ultraviolet, x-ray be ams, etc.), particle beams (electrons, neutrons or ions) or electric al current as is the case for semiconductors.
25 The atomic configuration will not typically remain permanently excited but will either return to its ground state or assume an intermed iate level. Conserva tion principles dictate that the amount of energy absorbed must al so be released. This may be manifested by emission of a photon with energy equal to that of the energy level difference, by transfer of energy via quantized vibr ational (phonons) exchange in the material, or by more complicated energy exchange mechanisms involving numerous processes. Many of the materials found to fluores ce efficiently are those for which the fluorescence originates from a deliberately added impurity. As an example, a host material such as Yttrium Aluminum Garnet (YAG, Y 3 Al 5 O 12 ) is transparent and nonfluorescent until rare earth impur ity ions such as Ytterbium (Yb 3+ ) are added. It then fluoresces with properties that make it useful in the design of thermometry systems. The dopant concentrations are usually on the order of a few percent or greater. Fluorescence as referred to here is the emission occurring from electronic transitions and is usually in the visible, near infrared (NIR) as for ytterbium, or the ultraviolet (UV) spectrum. An impurity ion in a crystal may be represented by a configurational coordinate curve as in Fig 2.1. Let us assu me that the ion is in its ground electronic state and its lowest vibrational state; temper atures below 70 K generally produce such a situation. A transition upw ard from A to B may be induced by the absorption of a photon of energy (E B E A ). The impurity will then be brought to the excited electronic state and to an excited vibr ational state level of this state. As a consequence of the Franck-Condon principle, wh ich states that atoms do not change their position during an electronic tran sition, the coordinate of the center, immediately after the absorption has taken place, is the same as th e coordinate before th e absorption. After the
26 absorption, the center will tend to transfer its vibrational energy to the lattice and nonradiatively decay to the lowest vibrational level designated by C. From this level the center will further decay to level D of the ground electronic state, giving up the energy (E C E D ) as fluorescent radiation. From D th e center will then decay to its lowest vibrational state. The difference in ener gy between the absorbed photon and the energy of the emitted photon is called Stokes shift. The energy level diagram shown in Fig.2.2 can be used to represent a fluorescent ion in a host lattice. The incident light exci tes ions from the ground state 1 to an excited state represented by the absorption band 3. Th e ions then decay by a fast radiationless transition to the metastable state 2. From this fluorescent level the ions decay to the aaa a D C B A b 1 3 2 P32P21w Fig.2.1: Absorption and fluorescence transitions according to the configurational coordinate model: Fig.2.2: Three level system with one fluorescent level and quick decay processes between level 3 and level 2. (a) absorption, (b) fluorescence.
ground state by purely radiative process and by other processes that we may call secondary (nonradiative). It is typically valid that the fluorescen ce spectral properties of any material will alter with temperature. This is so in pa rt due to the Boltzmann distribution governing the partitioning of the populations in various par ticipating vibrational levels of the ground, excited, and emitting states. A change in the intensity distribution results since oscillator strengths vary in accordance with selection rules and the Franck-Condon principle. The temperature dependence of these processes can be striking when there is competition with states which contend between radiative and nonr adiative deexcitation pathways. The rate of change for the population of the emitting state, 2, to a ground state, 1, is the sum of constant purely radiative spontaneous emission A 1,2 and a nonradiative component, W 1,2 which is temperature depende nt. The measured lifetime (in s) is given by the following expression: 2,12,11 )( 1 WATR where R(T) is the measured decay rate in units per second. 2.2 Elements Altering the Fluorescence Pro cess of Rare Earth Impurity Ions in Crystals There are many influences that have b een found to alter the fluorescence for rare earth impurity ions in materials resulting in an enhancement or degradation in the fluorescent process of interest. One must take into account ca reful consideration of these elements to be able to achieve the optimum results when using the fluorescent decay time as the correlator for temperature measurements. 27
28 2.2.1 Rare Earth Impurity Ion Concentration The parameters determining the fluorescence characteristics of any given phosphor are dependent on the concentration of the activating rare ea rth impurity ion. Overall decay time, intensity, relative spectral distribution, rise time, and response to temperature are all affected to some degree by the rare earth impurity ion concentration. Typically the more rare earth impurity ions in a host material, the more it will fluoresce, up to a certain point. However, when the c oncentration levels reach a certain limit, another nonradiative deexcitation pathway beco mes important. As the activator density is increased, the probability that an exc ited activator will transfer its energy nonradiatively to a neighboring dopa nt ion increases. The onset of this process is usually referred to as concentration quenching. For other dopants, even when in the same host material, the optimum concentration that produces the maximum intensity will be different [56,57]. One typically seeks to maximize the in tensity of the phosphor for thermometry applications. However, there are other c onsiderations that may warrant the use of different strategies. For example, at hi gh concentration the fluorescence decay profile may not be one of a simple exponential form This is important for decay time based thermometry since multiexponential and nonexpone ntial decay profiles are difficult to interpret. Complex waveforms make calibratio n and data analysis extremely difficult and at times intractable. The spectral emission distribution also can be sensitive to impurity dopant concentrations. Literature has illustrated this spectral emission dependence on
29 concentration doping for Eu:Y 2 O 2 S . They indicated that the 5 D 2 emission of Eu is especially sensitive to the Eu dopant concentrations. 2.2.2 Effects of Saturation High incident fluxes from any exc itation source may lead to fluorescence saturation effects. Phosphor efficiency changes as a function of incident flux when this occurs. Above a certain threshold value, overall intensity decreases, but for a lower dopant concentration this threshold value itself is lower and the rate of decrease is faster than at higher dopant concentrations. Literature cites th e increasing of electron beam voltage, independently of beam current, wi ll degrade the saturation of the phosphor according to Yamamoto and Kano for Yb:Y 2 O 3 . Conversely the situation may improve for some phosphors as cited in lite rature by Douling and Sewell in the case of ZnS and Zn 2 SiO 4 . Saturation is itself a temperature dependent phenomenon. Diminishment in the fluorescence decay time with increasing laser fluence has also been described in literature . Imanaga et al. suggests several possible mechanisms that may cause fluorescence satura tion behavior in this instance. In many situations the dominant mechanism of th ese saturation effects is beam induced temperature rise. To determine whether beam related effects will be a problematic issue for any given situation, experimental results and studies would have to be pursued for each particular circumstance. Typically one does not have situations in which they deliver an excess of laser excitation due to complexities in transporting beams over long distances. Optical fibers customarily are the waveguides of choice for transporting the beams and have limited fluence handling capacity for guiding the beams.
30 2.2.3 Host Lattice Asso ciated Impurities Inevitable amounts of undesired species in a host material may be detrimental in phosphor efficiency for thermometric applications. These impurities may change the atomic electronic environment experienced by th e activators so as to either augment or hinder phosphor performance. Ty pical characteristics of impur ities include a decrease in the emission intensity at fluorescence wavele ngth of interest. This is due to the impurities absorbing at wavelengths similar to those of the typical, thus effectively stealing excitation energy and decreasing the number of excited fluorescent centers. Furthermore, nonradiative energy transfer from an excited activator to such impurities may be efficient, thus increasing the d ecay rate and quenching the emission. In fluorescence decay thermometry, increasing the rare earth impurity dopant concentration of the fluorescing ion is plausi ble to decrease lifetime, thus enhancing the sensitivity of the temperature sensor. 2.2.4 Interactions of Different Rare Earth Impurity Ions The quenching or enhancement of the fl uorescent output of a crystal containing a certain type of rare earth impurity ion is of ten observed when another type of impurity ion is added into the host crystal. If th is second impurity ion presents a relevant absorption spectrum in a region in which the pumping source is emitting strongly and the energy absorbed by it is transferred by some mechanism to the fluorescent ion, an enhancement of the fluorescence may result. The opposite effect may take place if, given a certain fluorescent ion, anot her ion is added to the crysta l with no releva nt absorption band but with one level coupled to the metast able level of the fl uorescent ion. If the
31 additional ion is not fluorescen t or is fluorescent with some low efficiency, a reduction of the fluorescence output will result. There are two basic mechanisms that may induce energy transfer between rare earth impurity ions in crystals. The firs t is the mechanism by which energy transfer between an ion S and an ion A takes place in a cascade type, namely by emission of photons by the ion S and reabsorption of photons by ion A. in this case the ion S must be, by itself, a good emitter of fluorescence in a re gion in which the ion A absorbs strongly. In this case the lifetim e of the fluorescence of the ion S is not affected by the presence for the ion A and the emission of fluorescence by S shows a decrease only in correspondence to those wavelengths at which ion A absorbs. Another possible energy transfer mechan ism may be the resonant type, which produces what is known as sensitized fl uorescence, wherein the additional doping ion that provides the enhancement of the absorp tion features of the phosphor is called the sensitizer and the codopant ion is called th e activator which may decay radiatively or nonradiatively. The activator will exhibit st rong absorption providing there is efficient spectral overlap between its absorption spectrum and the emission spectrum of the sensitizer ion. In the present case the lifetime of the sensitizer if measurable, is found to decrease in the presence of the activator. Subsequently all the fluorescence emission of the sensitizer originating in th e state participating in the en ergy exchange is quenched as a function of the activator doping concentration.
2.2.5 Rare Earth Impurity Ion Particle Size Effects Literature has noted a change in fluorescen ce lifetime with a change in the size of the dopant impurity ion of interest . An example of this has shown an increase from 436 s to 598 s for a decrease in pa rticle size from 0.42 m to 0.11 m, respectively. This suggests that the fluorescence decay time of any phosphor will be the weighted average composite of the individual decay lifetimes arising from the distribution of impurity ions particle sizes. The quantum theory of decay rates of electric-dipole emissions explains the observable deviations of decay rate as a functi on of particle size. It predicts that the spontane ous lifetime of fluorescence is inversely proportional to the index of refraction of a fluorescent mate rial . The index of refraction n c in the proximity of an ion isolated in a crystal wi ll differ from the index of refraction of the same ion n s with a different particle size at some other isolated site. The relationship between the indices and the lif etimes can be found to be sosccsnnn // where o csand , are the lifetime values for the ioni c emission in the small particle, crystal, and vacuum cases, respectively. 2.3 Temperature Effects on the Fluorescen ce Lineshift, Inte nsity and Absorption Emission spectra lines are characterized by a wavelength for which the intensity is largest. The emission spectra line value typical ly alters with changes in temperature and is referred as the lineshift. The associat ed spectral emission width at the half maximum line intensity is referred to as the emission li ne width. Lineshift a nd linewidth deviations as a function of temperature are generally diminutive and are not often used as the 32
correlator in fluorescence thermometry. Literature has presented this latter technique for cathode ray tube thermometry wh ere they observed a shift to the blue of about 0.2 nm in going from -15 C to 72 C . Generally it is observed that emission spectra lines of phosphors get weak, i.e. become less bright, as the temperature of the material is increased. There may be regions where the selected emission lines are no longer sensitive to temperature deviations. The opposite effect may take place where at a partic ular value, termed quenching temperature, the strength of the emission line intensity fall s dramatically. The quenching temperature and temperature dependencies vary for each type of phosphor and for each of the emission lines within the spectrum of a given phosphor. When a phosphor is excited by a pulsed source, the persistence of the resulting fluorescence can be observed providing the leng th of the excitation source is much shorter than the persistence time of the phosphors fl uorescence. The fluorescence intensity usually decays exponentia lly according to the relation / t oeII where is the value of I( t) at t =0, and oI (lifetime or 1/decay rate ) is the standard 1/e ( 37%) folding time for the fluorescence. The decay time is often an extremely sensitive function of temperature and, therefor e a determination of its value constitutes a very useful method for fluorescence decay thermometry. At high impurity dopant concentrations, double exponential and more complex decay behaviors may arise which are more strenuous to analyze. 33
34 CHAPTER 3 THEORIES OF MULTIPHONON RELA XATION AND PHONON ASSISTED ENERGY TRANSFER IN RARE EARTH DOPED CRYSTALS To have a thorough technical acquaintan ce of the operation of the fiber optic temperature sensor presented in this work, it is necessary to examine the theories of multiphonon relaxation and phonon-assi sted energy transfer in ra re earth doped crystals. The absorption of thermal energy by a material causes vibrations in the lattice, which can be treated as discrete packets of energy called phonons. The vibrations of a crystal lattice have a certain impact on physical processes proceeding in the solid body. For example, the vibrational motion of atoms is the main contribution to the heat capacity of a solid and serves as the primary mechanism for heat conduction. The mobility of electrons in a crystal, and consequently, its heat conductivity is determined to a considerable extent by lattice vibrations. Phonons can interact with impurity ions of a given host lattice, causing phonon-assisted energy exchange between ions or stimulating th e relaxation of the excited ion to a lower level via non-radiati ve decay. The latter process is termed multiphonon relaxation, and phonon-assisted en ergy transfer has successfully been observed for various combinations of donor a nd acceptor rare earth ions in yttrium oxide (Y 2 O 3 ) at low temperatures.
35 3.1 Theory of Multiphonon Relaxation In Rare Earth Doped Crystals It was early in spectroscopic studies of material s when the dilemma of multiphonon orbit-lattice relaxation of excited states of rare earth ions in crystals arose. These studies observed that fluorescence from the upper of two adjacent states can be significantly quenched if the energy separation be tween the two states is too small. For example, in LaCl 3 a rule of thumb has developed which st ates that a level which is 1000 cm -1 or less above the next level will not, in general, be sufficiently long lived to fluoresce . This phenomenon was explai ned as being a result of energy being dissipated to the latt ice non-radiatively as phonons. From this phenomenon and the low maximum phonon energies of the materials examined, it was apparent that the process was of a high order involving several phonons . Experimentally, the dependence of fl uorescence quenching on the transition energy gap was observed through excited st ate lifetime measurements [66,67]. The critical energy gap has been observed to be different in various crystal hosts. Experimental results and subsequent analysis on relaxation rates between various excited levels of rare earth ions in laser crystals have shown the process to depend exponentially on the energy gap between levels involved in the process . A greater separation of the excited state and the imme diate lower lying level will require a larger number of phonons to conserve energy, and relaxation via n onradiative decay is less probable. In addition to the critical energy gap depende nce, a dependency on temperature has been found to increase transition rates drama tically at elevated temperatures.
3.1.1 Orbit-Lattice Interactions The modulation of the crystalline electric field by lattice vibra tions instigates the interaction of an isolated ra re earth ion with its dynamical crystalline environment. Expanding the crystal field Hamiltonian H CF in a Taylor series about the equilibrium ion position yields [68,69] H CF = + CFV ji jiji i iiQQVQV, ,... 2 1 (3.1) where represents the th normal mode coordinate and are partial derivatives of the static crystalline field in the normal coordinate system. iQ i jiV... CFV i c iQ V V & ji c ijQQ V V 2 (3.2) Therefore a given term of the interaction Hamiltonian is a pr oduct of two terms, one of which operates on the atomic system and one on the lattice vibration system. The expansion in equation (3.1) is extremely complex for all but the simplest crystals. The occurrence of processes involv ing the emission of many phonons can be accounted for by considering the fi rst-order term in the expans ion and carrying this out to high orders through perturbation theory. In ti me dependent perturbation theory, a process involving the transition from electronic state a to electronic state b with the emission of phonons can occur in diverse ways. When only the contributions from the two extreme terms are regarded and summing over all possible intermediate states and phonon modes, the probability for multiphonon re laxation between electronic states p a and b involving phonons is given by p 36
W(p) = 2 1 2 ,1 211jjj iii mmjinQnnQnn 2 2 2 2) () (1 1 1 1ai m aij i m ajm miaE EE E V Vp n )) (()()(aj i bj iE Egg (3.3) + 2 2 2 21 1 12jjj iiiajib jinQnnQnV n )) (()()(aj i bj iE Egg In equation (3.3) represent phonon mode occupation numbers, jinn 1 1 nm m are intermediate virtual states, and )()(j igg are frequency densities of phonon states. The electronic levels are assumed to be infinitively narrow and energy conservation is guaranteed by the delta f unction. The summation is over all possible intermediate states and phonon modes. Although the evaluati on of equation (3.3) is possible in principle, in practice only order of magnitude estimates are feasible [ 69]. The computation of transition rates are not likely using equation (3.3) due to an n-fold convolu tion of the density of phonon states that would have to be performed. The greatest difficulty arises from a lack of detailed information regarding the frequenc y, polarization, and propagation properties of the vibrations which limits its use to order of magnitude calculations. In addition, there is not sufficient data pertaining to the a ssociated strength of the ion-phonon coupling coefficients However, these complexities lead to a simplified phenomenological approach. Due to the interaction of the large number of phonon jiVV, 37
modes and intermediate states, it is assu med that these modes and levels will be statistically averaged out fo r high order processes. In the first approximation, it is expected that only the lowest orde r process dominates so only the W(p) with the smallest p contributes. 3.1.2 Energy Gap Dependence The simple consequence of the converg ence of terms of the perturbation expansion given in equation (3.3) is the dependence of the multiphonon transition rate on the energy gap. The ratio of the pth-order transition to the (p -1)th-order transition rate in a specified crystal is given by [68,69] = )1( )( p pW W << 1 (3.4) where is a coupling constant char acteristic of the host crystal which is necessarily small to ensure swift convergence of the expansion. The pth-order process approximately tracks the equation W (p) = B p (3.5) where B represents some constant. If the lowest-order process permitted do minates the decay, then the order is approximately determined by the energy ga p according to the following expression: p i = p MPE E E max (3.6) 38
where max is the frequency cutoff of the phonon spectrum and is the phonon mode energy. Using equations (3.5) and (3.6), a nd disregarding any se lection rules, the dependence of relaxation ra te on energy gap is given by pE W MP = C (3.7) MPEe where C and are constants which are characteristic of the particular crystal. Equation (3.7) is known as the energy gap law or exponential law due to the dependence of W MP on the energy gap MPE between electronic levels. 3.1.3 Temperature Dependence A simple model for the temperature dependence of multiphonon relaxation is given by W MP (T) = W MP (0) i pp kT Ee )1( (3.8) where W MP (0) is the spontaneous transition rate at low temperatures given by equation (3.7) [i.e. W MP (T)=W MP (0) at T=0 and the phonons are all in their ground state]. The total relaxation rate, sum of radiative and non-radiative rates, for ions in a excited host crystal lattice w ith coupling to one phonon mode is given by W MP (T) = A + C Ee i pp kT Ee )1( (3.9) where A is the radiative rate (assumed to be temperature independent). In a two phonon model, the total rate becomes W MP (T) = A + C 1 Ee1 1 1)1(p kT Epe + C 2 Ee2 2 2)1(p kT Epe (3.10) 39
40 Equation (3.10) is simply a modification of equation (3.9) contai ning the appropriate constants for the second phonon mode. 3.2 Theory of Phonon-Assisted Energy Trans fer In Rare Earth Doped Crystals Energy transfer processes are very vita l in solid-state luminescent systems, because they may provide an enhancement or degradation of the luminescent emissions. This is usually achieved by the in troduction in the host material of an ion of different type called acceptor or activator, in addition to the donor or sensit izer ion. The ion responsible for the desired emission is the acceptor ion. The donor ion provides additional spectral bands that absorb the pum ping energy of the source in the system. The energy absorbed results in the excitation of the donor ion. In turn this excitation energy is transferred to the acceptor providi ng another means of non-radiative decay. This process leads to a greater probability fo r amplified decay rates. Of course, if the activator's decay is radiativ e, an enhancement of the emission may result; if it is nonradiative, quenching of the emission will be the consequence. The term phonon-assisted energy transfer generally means a non-resona nt energy transfer process in which the mismatch of energy between the levels of energy donor and acceptor ions is compensated by the simultaneous emission or absorption of one or more phonons. This energy transfer process is governed by electr on-phonon interaction as other vi bronic processes are. The probabilities of phonon-assisted energy tran sfer as well as multiphonon relaxation are decreased approximately obeying exponential func tions with the increase of energy gap. The parameters expressing the exponential de pendence in these two processes are related
with each other. A dependency on temperatur e increases transition rates at elevated temperatures for this process in addition to the energy gap dependency. 3.2.1 Energy Gap Dependence of Transfer Probability The probability of phonon-assisted energy transfer according to the MiyakawaDexter theory [70,71] is expressed by W PAT ( E ) = W PAT (0) (3.11) PATEe where is the energy gap between the leve ls of donor and acceptor ions and PATE is a parameter determined by the strength of the electron-phonon coupling as well as the nature of the phonon involved and is determined by the host crystal. Equation (3.11) has the same form as that for the energy ga p dependence of the multiphonon relaxation rate given in equation (3.7). The parameter for multiphonon relaxation and in equation (3.11) are related to each other by (3.12) and a bg g 1ln 1 (3.13) where and are electron-lattice coupling constant s for the acceptor and donor ions, respectively. The expression ag bg is the energy of the phonon mode which contributes dominantly to the process. 41
3.2.2 Temperature Dependence of Transfer Probability Only spontaneous emission of phonons o ccurs at low temperatures. As temperature is increased, phonon-assisted ener gy transfer rate grows as a result of stimulated emission of phonons becoming operative. The temperature dependence of phonon-assisted energy transfer ra te can be expressed as W PAT (T) = W PAT (0) N kT Epe )1( (3.14) if it is assumed that the phonons involved in the energy transf er are of equal energy. The number of phonons emitted in the processes is given by N p PATE E N (3.15) where is the difference between the tran sition energy of the donor and the transition energy of the acceptor, usually referred to as the energy gap. The parameter is the energy of the relevant phonon mode in the host crystal. PATE pE 42
43 CHAPTER 4 FABRICATION OF FIBER OPTIC FL UORESENCE DECAY TEMPERATURE SENSORS There has been interest on crystalline ma terials prepared in the fiber form for a sustained period of time. This is partially due to single crystal fibers possessing near ideal physical properties in crystallinity and in tensile strength. Earlier work was addressed mostly on metallic materials such as the successful pu lling of single crystal metallic filaments directly from melt by Von Gompers in 1922 . In the 1950s work began to be centered on the magnetic and mechanical properties of metallic filaments  although the size and composition of these fibers could not be contro lled with precision. One method which allows the growth of crystalline fibers in a semi-controlled manner is the Steponav method in which the melts are drawn through shapers and crystallization is made to occur after passage through a die . Another method that has enabled us to grow high quality single crystal fibers of a desired length and diameter with proper crystallographic orientation, compositi on and doping is the Laser Heated Pedestal Growth (LHPG) technique . The LH PG method is classified under the nonconservative crystal growth process. In this process the solid is melted into the molten zone while crystallizatio n is in progress. Laser heating in the floating zone technique was used to grow single crystals of high melting point oxides, e.g., Y 2 O 3 Al 2 O 3 and Nd 2 O 3 . The floating zone technique has been applied in conjunction wi th a four-beam laser he ating method to grow
44 fibers of high strength and light weight composites . Single crystal fibers of Nd:YAG and Nd:Y 2 O 3 were grown by Burrus and Stone  using a modified floating zone technique. The development of an ax ial-symmetric focusing system at Stanford University  for use with LHPG enable d for the first time uniform heating of the molten zone with a single laser beam. 4.1 The Laser Heated Pedestal Growth (LHPG) Process LHPG and the related float zone growth technique are micro-variants of the Czochralski growth method. A number of hea ting sources have been used to produce the molten zone, with the most common method by fa r has being laser heat ing with single or multiple focused beams. A seed is inserted into the molten zone; as it is pulled out, surface tension of the molten materials forms a pedestal around the seed, hence the name pedestal growth. The melt is kept in place so lely by surface tension, eliminating the need for crucibles and eliminating one high source of sample contamination. This type of container-less growth apparatus permits the synthesis of materials with extremely high melting points. Illustrations and schematic re presentations of the LHPG method for fiber growth showing various regions involved in the growth can be found in literature . Conventional LGPG fiber pulling syst ems incorporate a stabilized CO 2 continuous wave (CW) laser typically with an output of between 15 75 W as a heating source. The usual focusing and turning optics can be found in literature . The fiber pulling system has the capability to be en closed in vacuum-tight chamber allowing growth in controlled atmospheres . The source rods are usually cut out of polycrystalline ceramics produced by mixing host and activator materi als, sintering and
45 hot pressing the mixture into flat disks. Cr ystalline feedstocks and fibers are typically used as source materials as well. 4.2 Advantages of the Laser Heated Pe destal Growth (LHPG) Technique One of the most attractive features of th e LHPG method is the rapidity with which fibers can be grown. With our pulling speed s fibers can be grown within a relatively short time. This allows for informational f eedback for the rapid r eadjustment of stock compositions and growth conditions for op timized materials making this method a powerful tool in the synthesis and engineering of crystalline materials in general. Fiber configuration is ideal for experimentally conducting absorption, emission and other optical measurements. The source rod length and the melt volume in LHPG applications are typically small. The cost of the chemical compounds required for the grow th of single crystal fibers is relatively small as a consequence. This results in the possibility to grow fiber crystalline materials which would be prohib itively expensive to grow by traditional methods . With the dependence on surface tension to maintain the integrity of the melt, the LHPG eliminates the need for crucibles or walls heated to high temperatures causing thermal gradients and stresses as is the case in crystal growth furnaces. Both crucibles and furnaces are well understood primary s ources of unintentional contamination in normal crystal growth. The absences of thes e surfaces allow for the growth of very pure crystal materials that are practically stress free. The small volume of the growth area impedes in the introduction of external pertur bations during the synthesis of the crystal
46 growth process. Impurity levels found in LHPG fibers are ordinarily solely determined by the purity of the starting mate rials of the source rods. 4.3 The LHPG Station at the Un iversity of South Florida The LHPG station located at the University of South Florida has been in operation for nearly twenty years. The growth appa ratus is composed of a stabilized 40 watt CO 2 laser manufactured by MPB Technology. The output of this laser is sent through a power attenuator and a beam expander before enteri ng the growth chamber. Inside the growth chamber exists an optical system produced by Pneumo Precision Inc. consisting of a reflaxicon, a scraper mirror and a parabolic mi rror. A detailed illustration of apparatus can be found in literature [80-82]. The optical system employed in this particular LHPG growth station enables the 10.6 m radiation to focus axial-symmetrically on the molten zone. By lowering a seed crystal into the melt, crystal growth proceeds by simultaneous upward translation of the seed and source rods while maintaining a constant melt volume at the laser focus. The latter function is performed by a computer which controls the growth rate and the diameter reduction of th e crystal. It is imperative that the melt volume be held constant to obt ain uniform diameter fibers. The transport systems for the upper and lo wer translators are identical. Brushless encoder servo motors equipped with the appropriate gear reducti on ratios rotate a precision lead screw which in turn translat es a hollow steel cylinder vertically. The cylinder holds a rod that has a removable tip as sembly for a seed or feed source material to be secured. The growth chamber is made of heavy nickel-plated aluminum and is isolated from electric motors, mechanical st ages and other extrane ous objects which may
degrade the crystal growth pro cess. This growth chamber fe atures a contamination free environment for crystal growth which can be evacuated by a turbo molecular pump and back-filled with high purity gases to give an inert or reactive growth atmosphere. 4.4 Properties of Yttrium Aluminum Garnet (YAG) Yttrium aluminum garnet (Y 3 Al 5 O 12 ), or YAG, is one of the most important optical materials with a wide range of appli cations such as laser host material, optical lens, and thermal barrier coating [83,84]. This is primarily attributed to its outstanding mechanical stability, low thermal expansi on and conductivity, low acoustic losses, and excellent optical properties. YAG has a cubic garnet structure with space group Ia3d or consisting of interconnected and slightly distorted octahedrons, tetrahedrons, and dodecahedrons with shared oxygen (O) atoms at the corner . The Yttrium (Y) ion at the 24( c ) position is dodecahedrally coordinated to eight O ions, which occupy the 96( h) sites. There are two aluminum (Al) sites, the octahedr ally coordinated Al 10hO oct at the 16( a) site and the tetrahedrally coordinated Al tet at the 24( d) site. Different cation environments and the cubic structure are th e basis of many of its electro-optical and electro-ceramic applications [83,84]. Because of its comp lex crystal structure, the electronic structure of YAG has only been studied recently [86,87]. YAG is a mechanically strong, thermally r obust, and chemically resistant material. YAG is also a very valuable component in high temperature ceramic composites because of its well known resistance to creep [88-91]. Table 1 li sts some of the important mechanical, thermal, optical and chemical attributes of YAG. The mechanical and chemical stability is comparable to sa pphire crystal, but YAG is unique with non47
48 birefringence and available commercially with higher optical homogeneity and surface quality. Density 4.5 g/cm 3 Transmission Range 250 5000 nm Melting Point 1970 C Specific Heat 590 J/kg*K Thermal Conductivity 14 W/m/K Thermal Expansion 6.9x10 -6 /K Mohs Hardness 8.5 Refractive Index email@example.com m, firstname.lastname@example.org m Lattice Parameters a=12.004 Youngs Modulus 300 GPa Table 4.1: Selected Properties of YAG. 4.5 Fabrication of the Single Crystal YAG Leads for Temperature Sensor The single crystal YAG lead fiber served as the optical waveguide to transport the excitation radiation and fluores cence to and from the transducer tip, respectively. By lowering a YAG seed crystal oriented in the [1 11] direction into the molten zone, crystal growth proceeds by simultaneou sly upward translation of th e seed and source rod while maintaining the constant melt volume at the la ser focus as mentioned in section 4.3. All the lead fibers were grown first, measuring 450 m in diameter and approximately 25 cm in length at a growth rate of 2.5 mm per minute. All th e crystal growth runs were fabricated with 1 atmosphere (768 torr) of air in the growth chamber. To prepare the lead fibers for optical testing, the two ends of each of the YAG lead fibers were mounted securely inside a short precision capillary tube with wax and then assembly was cut by a low speed diamond wheel saw to provide a semi-smooth surface normal to the fiber axis for polis hing. Diamond lapping films were used
49 successively in finer grades to achieve a good optical surface finish to ensure optimal coupling of excitation source light into and out of the fibers. Transmission of each fiber was determined by measuring the ratio of th e output power to the input power using a HeNe laser centered at 632.8 m with a power meter. Minimum transmissions of 70% of the original signal were measured for all the l ead fibers used in this dissertation project. 4.6 Preparation and Fabrication of the Rare Ea rth Doped YAG Phosphor Tips The substitution of the rare earth ions into the YAG la ttice takes plac e in yttrium sites. The generic doping formula for all the transducer tips can be expressed by A x Yb y Y 3-x-y Al 5 O 12 where different concentrations of Yb 3+ and a codopant (if used) rare earth ion represented by A are substituted in for yttrium. Table 2 shows a list of the ionic radii of various rare Rare Earth Ion Radius () La 3+ 1.18 Ce 3+ 1.14 Pr 3+ 1.14 Nd 3+ 1.12 Pm 3+ 1.10 Sm 3+ 1.09 Eu 3+ 1.07 Tb 3+ 1.04 Dy 3+ 1.03 Y 3+ 1.02 Ho 3+ 1.02 Er 3+ 1.00 Tm 3+ 0.99 Yb 3+ 0.98 Lu 3+ 0.97 Table 4.2: Ionic radii of rare earth ions .
50 earth ions which can be substituted into the yttrium sites of the garnet host. The steady decrease in size of the ions with increasi ng atomic number illustrates the lanthanide contraction. It can be seen from the table th at Ho, Er, Tm and Lu can be substituted into the YAG lattice relatively easily due to their ionic radii being le ss than or equal to that of the yttrium ions. For larger ions such as Nd, only a limited amount can be introduced into the crystal. At high doping levels signifi cant lattice strain will result which will be detrimental to crystal quality. The starting materials for fabrication of all the transducer tips were Y 2 O 3 and Al 2 O 3 Dopants in the form of Er 2 O 3 Nd 2 O 3 and Yb 2 O 3 were added to the mix. For the Yb 3 Al 5 O 12 base transducer tips, no Y 2 O 3 was used since we substituted ytterbium ions in for all the yttrium ion sites. All the oxid e powders were bought from Alfa Aesar Inc. having purities of 99.99% with respect to othe r rare earth constituents. The rare earth oxides used to prepare the feed material for the transducer tips were baked at 1273 C for 8 hours overnight in order to ensure its sesquioxide (RE 2 O 3 ) composition. All of the powders were baked in separate ceramic alumina dishes. After the constituents were dried out to remove any moisture they were measured out to 0.1 mg using an electronic balance. The materials were thoroughly mixed together to ensure a uniform distribution of the elem ents. The mixed powder was then pressed into a pellet using a 32 mm diameter die (Spex Corporation) mounted in a Specac 25 ton uniaxial press. The amount of pressure n eeded to achieve adequate packing depends on the amount of powder being pressed. Extreme pr essures lead to cracking of the pellet in two halves due to the flow of the powder from the center of the die towards its walls. We found that a total pressure of about 20 tons was sufficient to achieve compaction with
51 minimum flaking and no cracking of our sample. The first step in the packing action for our sample involves applying a pressure of approximately 5 tons to get rid of the entrapped air in the die. After one hour the final pressure of 20 t ons is applied to the sample for 5 hours. Once the sample has been pressed at a pr essure of 20 tons for 5 hours it can be removed from the die and is ready for the sinter ing process. Sintering is a vital process to reduce the porosity in our sample and to aid in partially fusing the particles below their melting. The sintering process is typically accomplished by heating the sample to a temperature somewhere between and of th e melting point of the sample for a certain period of time. During this pr ocess, the sample shrinks, part icles join together, and much of the void volume, which resulted from the initial misfit of the powder particles, is eliminated. In the case of solid phase si ntering, it occurs throu gh the diffusion of the particles in the volume of the material. The sintering temperature should be chosen based on the lowest melting point of the constituents in the sample. All of our samples were sintered for 8 hours at 1500 C. A typical feed preparation for growing our rare earth doped YAG transducer for a phosphor with 2% Tb 3+ and 10%Yb 3+ is given here. The congruent melting formula for YAG is: Y 3 Al 5 O 12 Since Tb 3+ and Yb 3+ ions will occupy yttrium sites, the congruent melting formula for 2%Tb,10%Yb:YAG is: Tb .06 Yb .3 Y 2.64 Al 5 O 12 The chemically balanced equation for th is particular temperature sensor is:
32 32 74 322 0.5 2 3.0 4 06.0 2 64.2 OAl OYb OTb OY Using the formula weights in mg of the re spective oxides, we fi nd that 298.1 mg of Y 2 O 3 254.9 mg of Al 2 O 3 59.1 mg of Yb 2 O 3 and 11.2 mg of Tb 4 O 7 will form slightly greater than 600 mg of the congruent mix which w ould be costly. The constituents were weighed out in ratios such that 1.5 grams of Y 2 O 3 was used per disc. This conveniently provided packed powder discs 1 mm in thic kness which made feedstock cutting process less tedious when it was time to cut the samples into 1 mm x 1 mm strips. The amount of laser power required fo r crystal growth is dependent on the melting point temperature of the garnet and th e cross sectional size of the feed rod. The melting point of YAG is 1970 C. It re quires approximately 27 watts of CO 2 laser power to melt a 1x1 mm square feed rod. Since th e power requirement scales proportionately with the cross sectional area, we could not us e feed rods larger th an the 1x1 mm due to the lack of enough laser power. Single crystal YAG fibers can be grown from crystalline feed at the rate of several mm/min with stellar crystal properties in cont rast to ceramic or polycrystalline feeds. Rapid growth rates for polycrystalline and ceramic feeds are not feasible due to the entrance of air bubbles into molten zone being tr apped in the grown crystal. Fast growth rates also incorporate major compositional va riation along the growth axis. A growth rate of 0.1 0.2 mm/min was found to be approp riate for good crystal growth and quality. Oriented seeds were used to initia te the growth of the doped and codoped transducer tips. Two growth generations were required to attain optimal optical quality from the packed powder. Therefore the feed stock was first grown into a cylindrical rod with a diameter of 600 m. The transducer tip was then grown from this rod measuring 52
53 1.5 5 mm in length and 450 m in diameter directly onto the 24 cm YAG lead fiber resulting in a monolithic crys talline sensor measuring approximately 25.5 cm in length. Many papers have reported cited on the im plementation of fiber optic temperature sensors based on fluorescence excited in a tran sducer tip [93-95]. The typical method of attachment was by either adhesive or mechan ical means which limited their operation to about 1000 C. With our monolithic probe a ssembly construction of the temperature sensor, the sensor should be operable up the high melting point of the YAG crystal provided there is a thermal source operable to that temperature. All the temperature probes were an nealed at roughly 1615 C in a high temperature furnace with air inside to remove strain in the material and homogenize the dopant ions present in the crystal. We have found in our work with rare earth doped fiber optic temperature sensors that the annealing process is imperative in order to achieve reproducible and stable long te rm performance from our sensors. The sensors must be annealed at a temperature equal to or great er than the maximum temperature that the sensors will be used at for repr oducibility. A ramp rate of 1 C/min was used to heat and cool the fiber optic temperature sensor duri ng annealing process to ensure the avoidance of thermal shock fracture. The furnace environment was kept extremely clean during the annealing process so that contaminating part icles did not penetrat e the surface of the fibers. Fiber transmissions are particularly sensitive to any surface modifications of this sort.
54 4.7 Distribution of Rare Earth Ions in YAG Homogenous incorporation and distribution of the rare earth dopant ions in the transducer tips for the fiber optic temperat ure probes is a necessity. When the dopants come from the melt phase into the crystal ph ase, its concentration has a change if the distribution coefficient k m is not equal to 1. If k m > 1, the ion concentration in the crystal is larger than in the melt, and it is becoming lower from the top to the bottom of the crystal in the process of crys tal growth, so a concentrati on gradient appears. Such gradients are known to be detrimental to perf ormance in the field of laser physics. The distribution of rare ea rth ions such as Yb 3+ in YAG crystals can be calculated by the following formula: k m = c top /c o where c top is the Yb 3+ (or other rare earth ion to be substituted for yttrium in YAG) concentration at the growth star ting position in the crystal, and c o is the initial Yb 3+ concentration in the melt. Literature has c oncluded that the distri bution coefficient of Yb 3+ in YAG crystals grown by the floating zone technique is 1.08 0.001 . The yttrium (YAG) and ytterbium (YbAG) garn ets are isostructura l with only a 1.5% difference in unit cell size  resu lting in the easy substitution of Yb 3+ into the yttrium site in the dodecahedron for YAG. The distribution coefficient of ot her rare earth ions such as Nd 3+ and Tb 3+ might be slightly smaller in YAG than Yb 3+ ions. To ensure that ions with k m < 1 values in the melt have attained a value approaching c o /k m a length of the doped crystal measuring 20-30 mm will be grown first and pulled out of the molten zone. The cap on the feed now should have a dopant concentration of c o /k m thus the phosphor tip on the probe will have a dopant concentration c o immediately. This allows
55 us to have transducer tips as short as possible resulting in the obviation of concentration gradients.
56 CHAPTER 5 YTTERBIUM BASED FLUORESCEN CE DECAY RATE FIBER OPTIC TEMPERATURE SENSOR SYSTEMS Ytterbium doped materials have offered a wide range of appli cations over the last three decades from fiber lasers to amplifiers . They have the ability to provide amplification over the very broad waveleng th range from approximately 900 nm to 1200 nm which has generated substantial interest recently. Apart from their broad gain bandwidth, Yb doped fibers used in amplifie rs can offer high output power and excellent power conversion efficiency . There is a wide range of possible pump wavelengths, allowing a variety of pumping schemes, includi ng the use of diode lasers or even high power Nd lasers. Fibers doped with ytterbium have been investigated by some researchers for sensing applications. Yb based fluorescence was observed by Kimur et al.  at room temperature from Yb doped porous silicon laye rs prepared by an electrochemical method which was developed for Er doping of porous silicon layers. After rapid thermal annealing in a pure argon atmosphe re at high temperatures above 900 C, the samples showed a sharp fluorescence band at around 1.0 m. Work by Maurice et. al.  discussed an intensity base d fiber optic sensor usi ng specially developed and noncommercial Yb 3+ doped fiber of 40 m diameter and 2000 ppm doping level. The device operated over a temperature ra nge from room temperature to 600 C, with an accuracy of 1 C being reported. A particular advantage of this species is the facility to
57 use excitation sources which are cheap, high po wer, and readily coupled into a range of fibers, associated with a fl uorescence emission spectrum in the near infrared which again is well suited to use with sensitive detectors. 5.1 Spectroscopic Properties of Yb:YAG The spectroscopy of the Yb 3+ ion is relatively simple compared to that of other rare earth ions. For all optical wavelengths only two groups of leve ls are relevant: the 2 F 7/2 ground state manifold and the 2 F 5/2 excited state manifold. Figure 5.1 illustrates the relevant energy levels of ytterbium in YAG. Previous work has shown that absorption and emission of Yb 3+ in YAG occur at 940 nm and 1040 nm, respectively . The rare earth ion Yb 3+ has a very broad absorption ba nd across the infrared spectrum, resulting from the 2 F 5/2 2 F 7/2 transition shown in Fig. 5.1. Decay from the excited state 2 F 5/2 metastable level is predominantly radiative because nonradiative transitions, due to phonon coupling or energy transfer with other en ergy levels, are inhibited except at very high temperatures. The 940 nm absorption peak is essentially due to the transition from the top of the lowest energy sublevel of the 2 F 7/2 manifold to the middle of the upper 2 F 5/2 Stark manifold. Four absorption bands ex ist for Yb:YAG crystals between 850 nm and 1100 nm . The main absorption band centered at 940 nm has an FWHM of about 22nm. A wide FWHM means that the crystal can accommodate some thermal shift of the pump light wavelength. The othe r two absorption bands are centered at 913 nm and 968 nm, respectively. Peak shifts in th e absorption spectra for increasing doping concentrations of Yb 3+ in YAG have been observed due to lattice deformation. Lattice deformation increases the splitting of electroni c manifolds that cause these observances.
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 12000 13000 14000 15000 Yb3+EMP=9540cm-1 2F7/2 2F5/2 Energy (cm -1 ) Figure 5.1: Relevant energy level diagram of ytterbium pertaining to the operation of fluorescence decay rate temperature sensing. 58
59 As mentioned earlier in secti on 4.6, the effective radius of Yb 3+ is smaller than Y 3+ which is replaced by Yb 3+ leading to these slight deformations. Research by Paschotta et. al. has revealed the fact that the details of the absorption and emission spectra depend, to some extent, on the host composition for rare earth ions . The measured fl uorescence decay times of Yb 3+ are typically around 1.0 ms and also vary by about 30% for different mate rials in fiber form. For example, fibers with higher germanium content in the core (introduced to achieve a higher numerical aperture) tend to have shorter lifetimes while Yb 3+ in a pure silicate glass has a lifetime of approximately 1.5 ms. The emission spectra may vary to some extent with pump wavelength, indicating some inhomogeneous broadening, although the broadening is predominately homogeneous . 5.1.1 Experimental Details for Yb Doped YAG Sensors The optical arrangement for the acqui sition of the temperature dependent fluorescence lifetime measurements for th e ytterbium doped YAG sensors is shown schematically in Fig. 5.2. For excitation of the ytterbium based temperature sensors a high power laser diode was used centered at 940 nm with a FWHM of approximately 4 nm which well matches ytterbiums ma in absorption band exactly. The 50 s laser beam pulses at a repetition rate of 10 Hz were firs t collimated by a 1.5 cm focal length lens and reflected off a dichroic filte r (HR at 940 nm, HT at 1040 nm). The reflected excitation beam is then focused into a 1 m length fibe r patch cable with a 1.5 cm focal length planoconvex lens. The opposite end of the fiber patch cable was fixed to a stainless steel SMA to SMA mating sleeve that was used to c onnect to the ytterbium doped temperature
Thermocouple Tube Furnace SMA Connector Fiber-Optic Temperature Probe 940 nm Pulsed LD Si Detector Digital Scope PC Multimeter HR 940 nm HT 1040nm Patch Cord Figure 5.2: Schematic of the optical layout for the fluorescence decay rate measurements. 60
probes. The induced ytterbium fluorescence and scattered laser light propagate back through the YAG lead and fiber patch cable. Af ter exiting the patch cable the beams are collimated by the plano-convex lens and the scattered laser light is reflected by the dichroic filter. A Si detector was used to detect the fluorescence emission at 1040 nm from the ytterbium ions and was acquired by a di gital averaging oscilloscope for analysis. A type B thermocouple consisting of a pair of Pt/Rh wires was placed in close proximity of the phosphor to monitor the ambient temp erature in the tube furnace. A PC was connected to the thermocouple and digital oscilloscope in order to collect 12 hour fluorescence lifetime data from the sensors using a LabView program. 5.1.2 Experimental Results for the Yb Doped YAG Sensors Ytterbium doped YAG phosphor temperature probes with varying concentrations of ytterbium (5 at. %, 10 at. %, 20 at. % and 50 at. %) were fabricated as described in the former sections 4.6 and 4.7. Figures 5.3 and 5.4 display the fluorescence decay curves recorded for the 5 at. % Yb 3+ :YAG phosphor sensor after = 940 nm pumping at 300 K and 1714 K, respectively. Figures 5.5 and 5.6 present semilog plots of the fluorescence decay curves for the former graphs, respectivel y. It is clearly discerned in these figures that the fluorescence decay for this sensor consists of only one exponential component by the successfully fitted linear lines in th e figures 5.5 and 5.65. The fluorescence decay rate was determined to be 990 /s ( = 1.01 ms) at 300 K and 26,977 /s ( = 37.1 s) at 1714 K. Fig. 5.8 exhibits the average fluorescen ce decay rate versus temperature for the 5 at. % Yb:YAG sensor (squares) from room temperature to 1714 K. Each fluorescence decay rate data point up to 1042 K in Fig. 5.8 is the average of four separate 61
62 measurements consisting of 256 decay trac es. The time interval between adjacent consecutive measurements was on the orde r of 30 seconds. Beginning at 1149 K, fluorescence decay rate stability tests were performed for the 5 at. % Yb:YAG (squares) sensor over a 12 hour period for every other data point repres ented in Fig. 5.8. The acute break and abrupt proliferation in the fluorescence decay rates beginning at approximately 1500 K is attributed to the stimulated emission of phonons by thermally populated phonon modes. Fig. 5.7 presents an exploded view of the fluorescence decay rates at lower temperatures. The observed initial decr ease in the fluorescence decay rates are due to changes in thermal population distri butions of the Stark levels of the 2 F 5/2 manifold of Yb 3+ This inhibits this sensor from be ing a candidate for temperature sensing applications in regions below approximat ely 600 K due to the nonmonotonic nature of the fluorescence lifetime making the determination of the decay time ambiguous. As a comparison for the sensor just presented with the phosphor grown from packed powder, a second 5 at. % Yb:YAG sens or was fabricated us ing a single crystal feedstock purchased from Scie ntific Materials Inc. Fig. 5.8 shows the decay rate from the probe grown from the single crystalline feed (hollow circles) and packed powder feedstock (solid squares). Good agreement is reflected in the figure allowing for the continued use of the packed powder feedstoc ks for fabricating the phosphors opposed to purchasing the costly sing le crystalline sources. It is clearly important in any practical application th at a stable and reproducible fluorescence response is seen fo r fiber-optic based temperatur e sensors. Figures 5.9-5.15 represent the results from 12 hour stability tests performed in the effort to determine if any significant changes are apparent in the fluorescence characteristics when the 5 at. %
-0.003-0.002-0.0010.0000.0010.0020.0030.0040.0050.0060.0070.008 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 Fluorescence Intensity (Volts)Time (s) Figure 5.3: Temporal evolution of 1040 nm fluorescence of Yb 3+ ions ( 2 F 5/2 2 F 7/2 ) at 300 K after pulsed excitation of the Yb 3+ ions at 940 nm. 63
-0.00010.00000.00010.00020.00030.0004 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11 0.12 0.13 0.14 0.15 0.16 0.17 0.18 0.19 0.20 Fluorescence Intensity (V)Time (s) Figure 5.4: Temporal evolution of 1040 nm fluorescence of Yb 3+ ions ( 2 F 5/2 2 F 7/2 ) at 1,714 K after pulsed excitation of the Yb 3+ ions at 940 nm. 64
0.0000.0010.0020.0030.0040.005 2 3 4 5 6 7 8 Ln of Fluorescence Intensity (Volts)Time (s) Figure 5.5: Temporal evolution (in se mi-logarithmic scale) of 1040 nm fluorescence of Yb 3+ ions ( 2 F 5/2 2 F 7/2 ) at 300 K after pulsed excitation of the Yb 3+ ions at 940 nm. 65
0.000000.000020.000040.000060.000080.00010 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Ln of Fluorescence Intensity (Volts)Time (s) Figure 5.6: Temporal evolution (in se mi-logarithmic scale) of 1040 nm fluorescence of Yb 3+ ions ( 2 F 5/2 2 F 7/2 ) at 1,714 K after pulsed excitation of the Yb 3+ ions at 940 nm. 66
20030040050060070080090010001100 980 990 1000 1010 1020 1030 1040 1050 1060 1070 1080 1090 1100 Fluoresence Decay Rate (/s)Temperature (K) Figure 5.7: Low temperature expanded view of the experimental data on fluorescence decay rate versus temp erature for the 5% Yb:YAG sensor. The squares represent data taken on two separate occasions, two trials per occurrence, and the error bars represent the standard deviation of the four relaxation rate measurements. 67
20040060080010001200140016001800 0 2500 5000 7500 10000 12500 15000 17500 20000 22500 25000 27500 30000 Fluorescence Decay Rate (/s)Temperature (K) Figure 5.8: Experimental data on fluorescence decay rate versus temperature for the 5%Yb:YAG sensor. The squares represent data taken on two separate occasions, two trials pe r occurrence, and the error bars represent the standard deviation of the four relaxation rates measurements. Open circles repr esent excursion taken using single crystal 5%Yb:YAG sensor for initial feedstock. 68
012345678910111213 1130 1135 1140 1145 1150 1155 Average = 1,143 /s Standard Deviation = +/0.9 /s Fluorescence Decay Rate (/s)Time (Hours) Figure 5.9: 12 hour fluorescence decay rate data for 5%Yb:YAG phosphor at 1,149 K +/1.3 K. 69
012345678910111213 1240 1250 1260 1270 1280 1290 1300 Average = 1,274 /s Standard Deviation = +/6.1 /s Fluorescence Decay Rate (/s)Time (Hours) Figure 5.10: 12 hour fluorescence decay rate data for 5%Yb:YAG phosphor at 1,248 K +/0.68 K. 70
012345678910111213 1650 1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 Average = 1,741 /s Standard Deviation = +/17 /sFluorescence Decay Rate (/s)Time (Hours) Figure 5.11: 12 hour fluorescence decay rate data for 5%Yb:YAG phosphor at 1,344 K +/0.70 K. 71
012345678910111213 3250 3300 3350 3400 3450 3500 3550 3600 3650 3700 3750 3800 3850 3900 Average = 3,566 /s Standard Deviation = +/81 /sFluorescence Decay Rate (/s)Time (Hours) Figure 5.12: 12 hour fluorescence decay rate data for 5%Yb:YAG phosphor at 1,439 K +/0.84 K. 72
012345678910111213 9000 9100 9200 9300 9400 9500 9600 9700 9800 9900 10000 10100 10200 10300 10400 Average = 9,739 /s Standard Deviation = +/111 /sFluorescence Decay Rate (/s)Time (Hours) Figure 5.13: 12 hour fluorescence decay rate data for 5%Yb:YAG phosphor at 1,530 K +/0.46 K. 73
012345678910111213 20000 20500 21000 21500 22000 22500 Average = 21,389 /s Standard Deviation = +/194 /sFluorescence Decay Rate (/s)Time (Hours) Figure 5.14: 12 hour fluorescence decay rate data for 5%Yb:YAG phosphor at 1,622 K +/1.4 K. 74
012345678910111213 25500 26000 26500 27000 27500 28000 28500 Average = 26,977 /s Standard Deviation = +/226 /sFluorescence Decay Rate (/s)Time (Hours) Figure 5.15: 12 hour fluorescence decay ra te data for 5%Yb:YAG phosphor at 1,714 K +/0.6 K. 75
Yb:YAG was exposed to elevated temperatures. It can clearly be seen that the standard deviations increased from the Figs. 5.9-5.15 as temperature was increased. This is attributed to the higher internal loss in the lead fiber at elevated temperatures. The experimental data reveals that one can dete rmine the fluorescence decay rate to better than 0.08 % at the lowest temperature of 1149 K examined and to within 0.9 % at the highest temperature of 1714 K inspected. This increase is due to the reduction in the signal to noise ratio. The relative sensitivity of the all crysta lline fiber-optic temperature sensor probes used in this work can be given by T W TW 1 S where W(T) is the fluorescence decay rate and W and T are respectively increments of the decay rate and temperature. Fig. 5.16 displays a semi-log plot of the 5 at. % Yb:YAG temperature sensors re lative sensitivity for temper atures greater than 1100 K. The sensitivity has two linear regions, one being where the sensitivity progresses to a maximum value at a temperature of 1484 K and another region where it is declining from thereafter, with slopes of .0040 /K 2 (y-intercept of 3.44 /K) and -.0035 /K 2 (y-intercept of -38.07 /K) respectively. A temperature measurement accuracy of approximately 1 K is obtainable with the lifetime measurement be ing determined to within 1.1 % at 1484 K. 76 Figures 5.17.19 exhibit an expanded view of the decay rates at lower temperatures for the 10 at. % Yb:YAG, 20 at. % Yb:YAG and 50 at. % Yb:YAG, respectively. Thermal redist ribution of populated states am ong the Stark levels of the 2 F 5/2 manifold of Yb 3+ again is responsible for the initial decline in the fluorescence decay rates in these figures similar to the re sults shown for the 5 at. % phosphor sensor.
77 The nonmonotonic regions in the decay rates for Figs. 5.17-5.19 at the low temperature end is suppressed to smaller domains with the incorporation of higher doping concentrations of the Yb 3+ ions in the host lattice. Figure 5.20 depicts changes in the multiphonon relaxation behavior by the a dvancement of the response to lower temperatures with increasing Yb 3+ dopant concentration due to alterations in the phosphors crystalline environment and phonon distribution. In creased dopant concentrations also result in a smaller inte raction radius between luminescent sites and impurities or defects which are capable of extinguishing the excitation energy. Although the 5 at. % Yb:YAG phosphor had smalle st nonmonotonic region for the varying Yb 3+ dopant concentrations, close examination of th e phosphor portion of the sensor revealed poorer visual crystal quality than the other samples which may have contributed to this finding. For comparison, Fig. 5.21 presents the results illustrating the relative sensitivity for the 10 at. %, 20 at. %, and 50 at. % Yb:YAG phosphors which are represented by the dotted, dash-dotted, and solid lines respectively. This figure clearly demonstrates that one can increase the relative sensitivity values at lower temperatures via the increase of Yb 3+ dopant ions into the host lattice of YAG. Although the relative sensitivity values decrease at the high temperature end earlier w ith greater dopant concen trations, it will be offset with better signal to noise ratio s in concerns pertaining to temperature measurement accuracy. Stability data for these sensors are presented in Appendix. An ytterbium aluminum garnet (YbAG) crystal was purchased from Scientific Materials Inc. to fabricate a YbAG phosphor sensor for analysis. The YbAG phosphor sensor was calibrated in the temperature regi on 298 K to 1373 K. The results illustrating
11001200130014001500160017001800 1E-3 0.01 0.1 Relative Sensitivity 1/W(T)[dW/dT] (/K)Temperature (K) Figure 5.16: Relative sensitivity versus temper ature (in semi-logarithmic scale) for 5%Yb:YAG sensor. 78
200300400500600700800900100011001200 790 800 810 820 830 840 850 860 870 Flourscence Decay Rate (/s)Temperature (K) Figure 5.17: Low temperature expanded view of the experimental data on fluorescence decay rate versus temperature for the 10%Yb:YAG sensor. The circles represent da ta taken on two separate occasions, two trials per occurrence, and the error bars represent the standard deviation of the four re laxation rate measurements. 79
20030040050060070080090010001100 830 840 850 860 870 880 890 900 910 920 Fluorescence Decay Rate (/s)Temperature (K) Figure 5.18: Low temperature expanded view of the experimental data on fluorescence decay rate versus temperature for the 20%Yb:YAG sensor. The triangles represent da ta taken on two separate occasions, two trials per occurrence, and the error bars represent the standard deviation of the four re laxation rate measurements. 80
200300400500600700800 1500 1550 1600 1650 1700 1750 1800 Fluorescence Decay Rate (/s)Temperature (K) Figure 5.19: Low temperature expanded view of the experimental data on fluorescence decay rate versus temperature for the 50%Yb:YAG sensor. The inverted triangles re present data taken on two separate occasions, two trials per occurren ce and the error bars represent the standard deviation of the f our relaxation rate measurements. 81
20040060080010001200140016001800 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 22000 24000 26000 28000 30000 Fluorescence Decay Rate (/s)Temperature (K) Figure 5.20: Experimental data on fluorescence decay rate versus temperature for the 10%Yb:YAG (circles), 20% Yb:YAG (triangles), and 50%Yb:YAG (inverted triangles) sens ors. Each data set represents data taken on two separate occasions, two trials per occurrence, and the error bars represent the standa rd deviation of th e four relaxation rate measurements. 82
500600700800900100011001200130014001500160017001800 0.000 0.002 0.004 0.006 0.008 0.010 0.012 0.014 0.016 0.018 0.020 10% Yb:YAG 20% Yb:YAG 50% Yb:YAGRelative Sensitivity 1/W(T)[dW/dT] (/K)Temperature (K) Figure 5.21: Relative sensitivit ies versus temperature fo r 10%Yb:YAG (dotted line), 20%Yb:YAG (dot-dash line) and 50%Yb:YAG (solid line) sensors. 83
20030040050060070080090010001100120013001400 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 12000 13000 14000 15000 16000 Fluorescence Decay Rate (/s)Temperature (K) Figure 5.22: Experimental data on fluorescence decay rate versus temperature for the YbAG sensor. 84
the fluorescence decay rate as a function of temperature are shown in Fig. 5.22. It can be seen that the fluorescence lifetime increas es monotonically with the increase of temperature for temperatures exceeding a pproximately 675 K. For the first 200 K multiple data points where taken in order to obtain a good calibration curve for the low temperature region in order to deploy this sensor in the thermally compensated mode which will be presented in a later section. The temperature dependence of th e multiphonon transition rate from 2 F 5/2 to 2 F 7/2 in YbAG was inspected and is shown in Fig. 5.23. The observed reduction in the experimental data (open circle s) initially is due to thermal redistribution of populated states similar to the formerly presented fluorescence versus temperature plots for the various Yb 3+ concentrations in YAG. The sharp break and increase in the multiphonon transition rate at higher temperat ures is attributed to the stimulated emission of thermally populated phonon modes. The theoretical fit (b lack curve) corresponds to the emission of 12 phonons of energy 800 cm -1 to bridge the approximately 9,600 cm -1 energy gap of Yb 3+ The temperature dependent transition rate is given by W MP (T) = A + C MPEe i pp kT Ee )1( = + 1s200,3 3.16 112 kT 1 cm800s)e1( The fit with the experimental data for temperatures gr eater than 600 K to the pure 12 phonon decay is satisfactory. The decay invo lves the second highest reported optical phonon mode affiliated with Yb AG . When the highest optical mode and lower energetic modes were inspecte d, the fit got progressively wo rse with the experimental 85
86 data. The 800 cm -1 mode makes the dominant contri bution to multiphonon relaxation and conserves energy in the lowe st order process than the lower energetic modes. The YbAG phosphor fiber optic temperature sensor was used in the effort to demonstrate the phenomenon of radiation trapping on the measured fluorescence lifetime of the phosphor. The room temperature fluorescence lifetime of the YbAG phosphor in ambient air was found to be 290.9 s (3,436.8 /s). When the YbAG phosphor tip was immersed in distilled water the fluores cence lifetime was determined to be 281.4 s (3,552.5 /s), corresponding to a 3.35 % reduction in the excited state lifetime compared to results obtained for air. This finding is at tributed to radiation trapping in which emitted photons by the fluorescent Yb 3+ centers are subsequently absorbed by centers in the distilled water. When the YbAG phosphor is immersed in the distil led water (n = 1.33), there is a smaller difference between the index of refracti on with the phosphor (n = 1.82) compared to when the phosphor is surrounded by air (n = 1.00), promoting the fluorescent emission to escape easier. Wh en the phosphor is surrounded by media with indices closer to its internal index of refraction (i.e. i ndex-matched), the likelihood of excessive reabsorption and reemission proce sses are reduced considerably due a smaller critical angle for total internal reflections in the phosphor. The effects of repeated radiative energy transfer has b een reported in literature for the excited state lifetimes of Yb:YAG and Er:YLF samples where reducti ons up to 25% were observed [109,110]. Our findings suggest that detailed information must be obtained in reference to a medias index of refraction to make proper corrections in order to acquire accurate temperature measurements using phosphor based fiber optic sensors.
200400600800100012001400 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 22000 Multiphonon Transition Rate (/s)Temperature (K) Figure 5.23: The multiphonon transition rate from the 2 F 5/2 manifold of Yb 3+ in YbAG is shown as a function of temperature. The solid squares represent the theoretical curve corresponding to the stimulated emission of 12 phonons of energy 800 cm -1 whose sum is approximately 9600 cm -1 The circles represent the experimental data. 87
88 5.2 Spectroscopic Properties of Nd,Yb:YAG As indicated previously in section 6.1.2, the 10%Yb:YAG sensor does not have any significant response until above 1,550 K. We were only marginally successful in extending its response to lower temperat ures by codoping the 10%Yb:YAG sensing probes with Tb 3+ as reported in literatu re . However, the st udies of the effect of Tb 3+ on Yb:YAG fluorescence decay did lead to important new insights about phonon assisted energy transfer at very high temperatures Building on this new found knowledge, we investigated the potential of Nd 3+ in extending the low temperature response of the temperature probes for the Yb:YAG sensors. The relevant energy levels for the Yb-N d system are shown in figure 5.24. For simplicity the Stark components within each leve l have been suppressed. The transistion energy indicated for Yb 3+ is, as before, calculated from the middle of the upper manifold to the top of the bottom manifold. For Nd 3+ the Stark components are located at 0 cm -1 132 cm -1 200 cm -1 311 cm -1 and 852 cm -1 . Because the highest state lies well above the rest, it can be disr egarded. Therefore, the en ergy shown corresponds to that from the middle of the lower 4 states in the ground manifold to the top of the 4 I 15/2 manifold. 5.2.1 Experimental Results for the Nd,Yb:YAG Sensors In the effort to enhance the respons e of our Yb:YAG based FDR temperature sensor, we introduced Nd 3+ as the acceptor ion for ener gy transfer. The temperature dependence of the fluorescence decay rate fo r the phosphor containing 2 at. % Nd and 20 at. % Yb. are shown in Fig. 5.25. Fig. 5.26 presents an expanded scale at the low
Yb3+Nd3+ E = 9,540 cm-16,600 cm-1 E = 3,400 cm-1 4I15/2 4I13/2 4I11/2 4I9/2 2F7/2 2F5/2 Figure 5.24: Energy levels in ytterbium and ne odymium relevant to the operation of Nd,Yb:YAG sensors. 89
90 temperature end for magnified inspection. Si gnificant response is found for the phosphor at low temperatures (Fig. 5.26) extending fu rther into the low temperature region by 600 K in comparison with the Tb,Yb:YAG sensor investigated in ear lier work [1,30,37]. The dip prohibits the use of the Nd,Yb:YAG sensor below 520 K. The creation of the nonradiative pathway via energy transfer by th e addition of the neodymium ions extends the response into the low temperature re gion by nearly 100 K, 600 K, 450 K, 150 K, and 175 K in comparison to the varying 5, 10, 20, 50, and 100 at. % Yb 3+ :YAG phosphors presented in the previ ous section, respectively. The sensitivity of the 2 at. %Nd, 20 at. %Yb:YAG sensor is shown in Fig. 5.27. The fluorescence lifetime for the 20 at. %Yb:YAG phosphor could be determined to within .03 % and 1 % at approximately 500 K and 1690 K, respectively. It is not unreasonable to expect the same accuracy in the determination of the fluorescence lifetime for the 2 at. %Nd, 20 at. %Yb:YAG phosphor. This would correspond to a sensitivity of 1.2 K at 500 K and 1.1 K at 1690 K. The 2%Nd,20%Yb:YAG sensor was sacrificed to examine the dopant distribution along the growth axis. A multiline argon laser was used as the excitation source for the Nd 3+ ions at 514 nm which was modulated by using a variable frequency chopper (See Fig. 5.28). A prism made with 60 angles was us ed in the effort to disperse the multiline emission output of the argon ion laser. The dispersing properties of the material from which the prism is made allows for the spatial separation of different emission wavelengths of the argon laser. After bloc king the unwanted laser em ission lines, the 514 nm excitation light was modulated with a chopper and focused down onto the codoped fiber which was secured to a translation stag e. A longpass filter (LP 1,000) was placed in
20040060080010001200140016001800 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 Fluorescence Decay Rate (/s)Temperature (K) Figure 5.25: Experimental data on fluorescence decay rate versus temperature for the 2 %Nd,20%Yb:YAG sensor. The stars represent data taken on two separate occasions, two trials per oc currence, and the error bars represent the standard deviation of the four relaxation rate measurements. 91
20030040050060070080090010001100 975 1000 1025 1050 1075 1100 1125 1150 1175 1200 1225 1250 1275 1300 1325 Fluorescence Decay Rate (/s)T (K) Figure 5.26: Low temperature expanded view of the experimental data on fluorescence decay rate versus temperature for the 2%Nd,20%Yb:YAG sensor. The stars represent data taken on two se parate occasions, two trials per occurrence, and the error bars represen t the standard deviation of the four relaxation rate measurements. 92
20030040050060070080090010001100120013001400150016001700 0.000 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.010 Relative Sensitivity 1/W(T)[dW/dT] (/K)Temperature (K) Figure 5.27: Relative sensitivity versus temperature for 2%Nd,20%Yb:YAG sensor. 93
94 the path of the Nd emission to block any s cattered light from reaching the detector and permit the 1.06 m emission to pass. When a sec ond longpass filter (LP 665) was placed in front of the detector along with the LP 1,000, our Nd emission signal was not altered. This verified that the use of just the LP 1,000 filter sufficed in blocking out all of the scattered laser light. For veri fication that we were receivi ng a real signal, the fiber was replaced with microslide in wh ich no signal was apparent. The crystal was polished to expose a fl at crystal surface for analysis. The 2%Nd,20%Yb:YAG crystal analyzed was about 15 mm long and the entire length was analyzed. Fig. 5.29 shows the plot of the voltage (which represents the distribution of Nd dopant concentration) along the growth axis of the crystal. We would presume a uniform signal along the growth axis of the crysta l providing the following conditions: 1.) the atomic weight fraction C A of the element of interest (Nd) is uniform, 2.) the excitation beam is always kept normal to the crystal s surface, and 3.) the volume of the crystal from which the emission is generated is cons tant. Making certain th at conditions 2 and 3 are met, any variations in the emissi on signal would denote a variation in C A The concentration of Nd in the crystal is controll ed by the segregation coefficient of the ion in the crystal matrix. The first and the last data points were taken at positions close to the starting and ending points of the growth run, th erefore they might have been affected by the unstable thermodynamics of the molten z one. As it is seen in Fig. 5.29, the Nd dopant concentration in the crystal approach es a constant value fairly slowly. Nd concentration in this crystal increased from a very low value near the junction with the lead YAG fiber to a final plateau concentration over a length of approximately 10 mm. This gradient is attributed to the relati vely small distribution coefficient for the
95 incorporation of Nd into YAG, which results in the segregation of Nd ions in the grown crystal and the melt. This data implies that a length of at least 10-15 mm should be grown first to attain the desired feed concen tration of 2 % Nd uniformly along the growth axis. This will ensure the concentration of Nd in the melt has attained a value approaching c 0 /k m Once the Nd concentration has plateaued, we will be able to use the remaining feed to fabricate the fiber optic te mperature probes. Since the cap on the feed now has a Nd concentration of c 0 /k m the phosphor tip on the temperature probe will have a Nd concentration of c 0 immediately. The former desc ribed process was implemented on all the phosphor tips in this work.
Det. Scope Ar Laser Chopper Translation Stage Phosphor Filter Beam Blocker Figure 5.28: Experimental setup for side fluorescence measurements for the 2%Nd,20%Yb:YAG sensor. 96
0123456789101112131415 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 Voltage (mV)Position (mm) Figure 5.29: Experimental side fluorescence measurements for the 2%Nd,20%Yb:YAG sensor. 97
5.3 Spectroscopic Properties of Er,Yb:YAG and Er:YbAG The experimental results from the Nd ,Yb:YAG probe gave impetus to further investigate improvement at the lo w temperature end. Erbium (Er 3+ ) was chosen to be the candidate as the acceptor ion for energy tran sfer next. An interesting feature of Er,Yb:YAG system is the fact that the Yb 2 F 7/2 2 F transition is resonant with the Er 5/2 4 I 15/2 4 I transition as shown in Fig. 5.30. The energy transfer process which occurs between these two ions is consequently also of a resonant nature. Due to the close energy match characteristic of the two levels 11/2 2 F (Yb) and 5/2 4 I (Er), the two levels should always be in thermal equilibrium and behave as a single manifold kinetically. The 11/2 = 940 nm excitation energy, afte r being absorbed by the ytte rbium ions, is transferred to the erbium ion level by a resonant nonrad iative energy transfer process followed by fast nonradiative energy rela xation to the erbium level 4 I 13/2 The efficiency of this ytterbium-erbium system is determined by the extent of overlap between Yb 3+ emission spectrum and Er 3+ absorption spectrum. It is also im perative that the transfer rate from ytterbium to erbium ions to be faster th an ytterbiums fluorescence decay rate. 5.3.1 Experimental Results for the Er,Yb:YAG Sensors A 2%Er,20%Yb:YAG temperatur e probe was fabricated for investigation. Fig. 5.31 displays the fluorescence decay rate vers us temperature for this particular probe from 300 K to 1,071 K temperatures. From th ese graph it is dis tinctly shown that improvements in the response at the low temperature end have been made by the monotonic behavior of the experimental data in this region. Although we have made tremendous progress at achieving stronger resp onse at the low temperature range, an abrupt reduction in the decay rate at temper atures greater than approximately 1070 K was 98
E = 6,140 cm-1 E = 9,540 cm-1 4I15/2 4I11/2Yb3+Er3+ E = 3,400 cm-1 4I13/2 2F7/2 2F5/2 Figure 5.30: Energy levels in erbium and y tterbium relevant to the operation of Er,Yb:YAG sensors. 99
20030040050060070080090010001100 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 Fluorescence Decay Rate (/s)Temperature (K) Figure 5.31: Experimental data on fluorescence decay rate versus temperature for the 2%Er,20%Yb:YAG sensor. The triangles represent data taken on two separate occasions, two trials per oc currence, and the error bars represent the standard deviation of the four relaxation rate measurements. 100
detected. This observation is credited to th e excited ions decay process being in energy migration limited regime for the dopant c oncentrations examined. At elevated temperatures there may be a significant reduc tion in the extent of spectral overlap of the emission spectrum of Yb 3+ with the absorption spectrum of the Er 3+ due to the thermal shifts in their respective spectra. The number of excited Yb 3+ donors within the critical interaction range of acceptors may also be t oo small leading to th e decay process being energy migration limited. This results in a time duration which is too long for Yb 3+ to exchange energy with another Yb 3+ ion within the vicinity of an Er 3+ acceptor ion. Fig. 5.32 illustrates the Yb 3+ fluorescence in a semilogarithmic plot for the 2%Er,20%Yb:YAG probe at 300 K and 1145 K fo llowing pulsed excitation. Take note to the initial deviation fro m a simple exponential dependence for the plots at both temperatures. The initial none xponential portion of the decay is attributed to scattered pump light leaking through the filter and relaxation by direct Yb 3+ Er 3+ energy transfer. At lower Er 3+ concentrations a smaller fraction of the excited Yb 3+ ions would be within the effective interaction sphere of Er 3+ energy sinks. This would lead to a smaller contribution of the direct relaxation process to the over-all decay and the nonexponential portion of the decays in Fig. 5.32 would effec tively become smaller. However, lower codopant concentrations will also lead to a less pronounced variati on of the fluorescence decay rate as a function of temperature wh ich contradicts our objective to broaden the response range of the sensor. The final portions of both plots are expone ntial. Literature has reported the same initial nonexponential behavior for Cr 3+ doped Eu(PO 3 ) 3 in which the nonlinearity portion increases with Cr 3+ content . Notice that the 1145 K plot has a much shallower slope than the 300 K plot The slopes, which represent the decay 101
0.00000.00020.00040.00060.00080.00100.00120.00140.0016 0 1 2 3 4 5 6 1145 K 300 KLn Fluorescence Decay Rate (Volts)Time (s) Figure 5.32: Temporal evolution (in semi-l ogarithmic scale) of 1040 nm fluorescence of Yb 3+ ions ( 2 F 5/2 2 F 7/2 ) at 300 K and 1,145 K after pulsed excitation of the Yb 3+ ions at 940 nm for the 2%Er,20%Yb:YAG sensor. 102
103 rate, should increase with increasing temperat ure when decay system is in the superfast migration regime. This is apparently not the case for the 1145 K plot due to the decay process being energy migration limited. The super fast migration regime is desired for simple interpretation of the data for temperature sensing applications. In the effort to keep the fluorescence decay process in the superfast migration regime we decided to replace all of the yttrium ions with Yb 3+ to form the host Yb 3 Al 5 O 12 while still using Er 3+ as the acceptor candidate of choice. Literature has reported that YbAG and YAG are isostructural with less than 1.5 % di fference in unit cell size from lattice parameter data . Recent literature has reported that the absorption coefficient of YbAG is 90.8 cm -1 at 940 nm . This makes it very reasonable to expect essentially complete absorption of the pump radiatio n at this wavelength in a heavily doped Er:Yb 3 Al 5 O 12 probe. The temperature dependence of the fluorescence decay rate for probes containing 5%Er:YbAG (circles) and 10%Er:YbAG (tri angles) are shown in Fig. 5.33. The room temperature lifetime for the 5%Er:YbAG se nsor was found to be 0.31 ms which is reasonable considering reports from literatur e citing the lifetime of YbAG to be 0.27 ms . One notices from the exploded lo w temperature graph given in Fig. 5.34 the recurrence of the dip as seen for other probes makes the 5%Er:YbAG sensor impractical for operation from room temperature to approx imately 450 K. Due to the same feature in the fluorescence decay rate for the 10%Er: YbAG probe, this sensor is for deployment from room temperature to around 600 K. It can also clearly depicted for Figs. 5.33 and 5.34 that the fluorescence decay rate for the 5%Er:YbAG (circles) and 10%Er:YbAG (triangles) are comparable at temperatures gr eater than 900 K. This finding attests that
104 the additional Er 3+ concentration had no dramatic eff ect on the fluorescence decay rate for the concentrations investigated.
2004006008001000120014001600 0 2500 5000 7500 10000 12500 15000 17500 20000 22500 25000 27500 30000 32500 35000 Decay Rate (/s)Temperature (K) Figure 5.33: Experimental data on fluorescence decay rate versus temperature for the 5%ErYbAG (circles) and 10%ErYbAG (triangles) sensors. Each data set represents data taken on two separate occasions, two trials per occurrence, and the error bars represent the sta ndard deviation of th e four relaxation rate measurements. 105
2003004005006007008009001000 3000 3200 3400 3600 3800 4000 4200 4400 4600 4800 5000 5200 5400 5600 5800 6000 6200 6400 Decay Rate (/s)Temperature (K) Figure 5.34: Low temperature expanded view of the experimental data on fluorescence decay rate versus temperature for the 5%ErYbAG (circles) and 10%ErYbAG (triangles) sensors. Each data set represents data taken on two separate occasions, two trials per occurrence, and the error bars represent the standard deviation of the four relaxation rate measurements. 106
107 CHAPTER 6 THERMALLY COMPENSATED FLUORESCENCE DECAY RATE TEMPERATURE SENSOR The material science and semiconduc tor manufacturing industries require improved accuracy in measuring the temper ature of silicon wafers during processing because accurate temperature measurements are critical to product quality and device performance. In particular, there has been a long term established need for the enhancement in the accuracy of surface temper ature measurements, with emphasis in the area of rapid thermal processing (RTP) of semiconductors and applications requiring temperature measurements where heating is accomplished via microwave radiation. In order to measure the surface temperatur e of a heated object, the sensor must achieve adequate thermal contact with the samp le of interest. Once the sensor has made contact, heat will flow from the object to the sensor, raising the temperature of the sensor in the process. Eventually thermal equilibrium will be reached by the sample and sensor. The final equilibrium temperature of the sens or will rely on a number of factors such as (1) the degree of thermal isolation of th e sensor and the sample from the surrounding environment; (2) the surface area, quality and pr essure of the thermal contact, and; (3) the thermal mass of the sensor relative to the sample. Typical practices in surface temperature measurements involve the thermistor, thermocouple, or RTD to be secured to the sample via adhesives or epoxies. This technique provides secure attachment and a thermally insulating barrier over the sensor in
108 order to measure the sample temperature at th e location of interest, but is not an option for, e.g., monitoring the temperature of a wafe r being processed. In addition, removal of sensor may be tedious and cause damage to the sample of interest. If the sensor has poor contact or a large thermal mass relative to the sample, the measured equilibrium temperature will be below that of the true su rface temperature prior to contact. Samples which are feeble conductors of heat may result in long equilibrium times due to the slow rate of heat flow to the location where heat is drawn by the sensor. The most frequent challenge with surface temperature measurements with electrical sensors involves the sensor leads conducting he at away from the sample due to their being good thermal conductors, resulti ng in erroneous readings. In vacuum applications the above challenges tend to ma gnify since conduction by air is eliminated. In electrically hostile environments involvi ng strong radio frequency (RF) or microwave fields, RF plasma, and high voltage gradients, the electrical sensors experience alarming problems which inhibit accurate surface temperature measurements. In the infrared radiometry approach, un certainties about the emissivity of the surface of interest leads to liberal errors. Many finished samples contain a variety of materials which are optically not resolvable from one another so the average observed emissivity is typically unknown. Furthermore, complications arise from some of the observed infrared energy being transmitted through semiconductor materials from a source somewhere behind the surface material of interest. Finally, field of view complications are typically associated with this technique making it an undesirable means of measuring surface temperatures.
109 The thermally compensated FDR temperature se nsor technique is an entirely new approach. With the use of a second laser oper ated in the continuous wave (CW) mode as the internal heating mechanism, the phosphor tip will be heated up. First, the temperature of a phosphor probe at a series of CW laser pow er levels can be measured with its tip in close proximity to the point of interest on a su rface. Another set of measurements is then made at the same series of launched CW po wer laser levels with the tip of the probe touching the surface of interest. Initial hea ting temperatures below that of the actual surface temperature will result in an increase of the tips temperature, and those higher would lead to a decrease. The smallest change in phosphor tips temperature will be experienced for initial temperat ures that are closest to the actual surface temperature. When the temperature difference is plotted ag ainst temperature, where it crosses from negative to positive for the contacting and nonc ontacting conditions should then give the true surface temperature. This novel technique with the crossing principle should allow for the true surface temperature to be determin ed even without the true temperature being directly measured. 6.1 Thermally Compensated Fluorescence D ecay Rate Temperature Sensor Model A simple model for the thermal transport mechanisms of the thermally compensated fluorescence decay rate temperature sensor used in the experiment is provided in Fig. 6.1. The end of the 400 m diameter sensor probe in which the pulse train excitation source and CW heating s ource will be coupled (end opposite of the phosphor tip) is mounted in a st ainless steel SMA connector, wh ich is assumed to be at room temperature T 0 It is further assumed that the phosphor will be cooled along the length of the growth axis of the lead fiber while simultaneously being heated via CW
SMA Connector at T0TS400 m YAG Lead Fiber with Length L 1 mm length (), 400 m diameterYbAG Phosphor Tip at Tem p erature T d Figure 6.1: Simplified model for the thermal transport mechanisms of the thermally compensated FDR YbAG sensor. 110
laser radiation. In addition, heat exch ange between the phosphor with uniform temperature T and the surface of interest at temperature T s will be accomplished by conduction through the surrounding gas medium. Thermal transport between the tip and the sample of interest will be considered separately for the sidewall of the phosphor transducer and its end. The conductive medi um between the phosphor and the sample is considered to be in the shape of a conical frustum with areas (a+d) 2 for the top and a 2 for the bottom, where a is the radius of the probe and d the distance between the phosphor tip and the sample. The rate of h eat transfer between the end of the phosphor tip and the sample is given by a(1+a/d)(T s -T) where is the thermal conductivity of the gas. The rate of heat transfer between the sidewall and the sample is taken to be a (T s -T) where is the length of the exposed tip. The temperature of the tip can be found using d a aa L a T d a aa T L a P T2 2 s 2 0 2 (6.1) where L represents the length of the mono lithic crystalline temperature sensor, a its radius, and its thermal conductivity. P represents laser power dissipated in the phosphor which corresponds to the absorbed power multiplied by the fraction of the decay rate which is nonradiative. The dependence of the phosphor tips temp erature as a function of its distance from the sample is presented in Fig. 6.2 by use of equation 6.1. The plots were generated using the initia l conditions of a sample surface temperature of T s = 100 C, a 111
connector temperature of T 0 = 25 C for the sensor probe, a total fiber probe length of L = 10 cm, a phosphor tip length and radius of = 1 mm and a = 200 m respectively, a thermal conductivity of =13 W/mK for YAG, and finally a thermal conductivity of =0.03 W/mK for air at approximately 100 C. From each selected P equation 6.1 yields one of the curves s hown as the distance between th e phosphor tip and sample is varied from 0.5 mm to approximately 10 m in step sizes of 10 m. When the phosphor tip temperature is below the sample surface temperature, the phosphors temperature shifts upward to a value towards that of th e sample when approaching and touching the sample and inversely downward for phosphor temperatures above the sample surface temperature. This phenomenon is depicted in Fig. 6.2. The curves from the figure are anticipated to be symmetric about the true surface temperature of the sample for phosphor temperatures the same magnitude ab ove and below the actual temperature as exhibited in the model. One must keep in mi nd that factors such as the samples profile, thermal conductivities of the sample and gas medium, and the degree of contact pressure will determine the magnitude of the shifts in any real situation. Fig. 6.3 illustrates the temperature diffe rence between the gap distances of 10 m and 100 m for each initial heating condition versus the initial heating temperatures at the specific gap distances of 10 m (hollow circles) and 100 m (squares). The dotted line in the Fig. 6.3 represents the linear best fit line for the gap distance of 10 m and the dotdash line signifies th e best fit for the 100 m gap. As depicted in the figure, where the linear best fit lines cross from negative to positive, the true surface temperature is revealed for both the contacting and nonc ontacting scenarios. The corresponding temperatures for which the contacting a nd noncontacting conditions crossed from 112
113 negative to positive in Fig. 6.3 were found to be 100 C +/1.2x10 -8 C and 100 C +/7.5x10 -8 C from the best fit lines, respectively. A second technique to determine the actual surface temperature would entail drawing straight lines through the data points which are closest to the sample surface for each data set and extrapolating them to the left to find the corresponding common intersection temperature. Fig. 6.4 exhibits the temperature of the phosphor at the specific gap distances of 10 m and 50 m at various laser power di ssipation values which were extracted from Fig. 6.2. It can be seen in Fig. 6.4 that all the extrap olated lines intersect at a common point. Calculations were pe rformed using Matlab to determine the six crossing temperatures for all six possibl e pairs. They showed 100.0043 C and 99.9943 C as the highest and lowest intersec tion temperatures, respectively, while the total average temperature for all the intersections was found to be 99.9993 C +/.0036 C. Sample surface qualities and profil es vary to an assortment of degrees due to diverse fabrication techniques. As a consequence, the minimum gap length will be determined by the particular modus applied to create the sample su rface. It is not unreasonable to expect that a minimum gap length of 50 m may apply for a rough surface. The same numerical calcu lations for gap distances of 50 m and 100 m at the same laser power dissipation values from Fig. 6.2 were performed. Th e four extrapolated straight lines had a common intersection of 99.9996 C. Variations in the intersection temperature values for the four data sets for th is case were only evident for values after 8 significant digits. Error in this technique w ould be minimized by investigating data sets which are perfectly symmetric mirror images about a horizontal line going through the true surface temperature.
114 As can be inferred from Fig. 6.4, one doe s not need to make contact with the sample to obtain exceptional re sults for samples with adequa te surface profiles with the thermally compensated FDR sensor. This nove l technique has the cap acity of providing a noninvasive means of measurement of surf ace temperatures without perturbing the sample with extensive heat drain. This model will only hold for laser heating sources which have a high degree of stability.
0.00000.00010.00020.00030.00040.0005 85 90 95 100 105 110 115 Temperature (C)Distance (m) 0.5 mW 1.0 mW 1.5 mW 2.0 mW Figure 6.2: Phosphor temperature versus gap length using the theoretical equation 6.1 with various laser dissipation values. 115
9092949698100102104106108110 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 T (C)Temperature (C) Figure 6.3: Phosphor temperature differe nce between the gap distances of 10 m and 100 m for each initial heating condition in Fig. 6.2 versus the initial heating temperatures at th e specific gap distances of 10 m (hollow circles) and 100 m (squares). 116
-0.00002-0.000010.000000.000010.000020.000030.000040.00005 90 92 94 96 98 100 102 104 106 108 110 0.5 mW 1.0 mW 1.5 mW 2.0 mWTemperature (C)Distance (m) Figure 6.4: Phosphor temperature versus gap le ngth at various laser dissipation values using the theoretical equation 6.1 w ith the extractions of specific gap distances of 10 m and 50 m from Fig. 6.2. Extrapol ation plots (dotted lines) are applied to each data set. 117
118 6.2 Experimental Details and Results for the Thermally Compensated FDR Temperature Sensor In the effort to demonstrate the opera tion of the FDR sensor in the thermally compensated mode, the YbAG phosphor probe presented in the former section was chosen as the candidate for deployment in this endeavor. The YbAG phosphor was chosen specifically due to it exhibiting the maximum response (16 % change in decay rate for the first 100 C) for the temperature ra nge selected for this experimental test in comparison to the other phosphors investigate d. Fig. 6.5 illustrate s the experimental setup for the thermally compensated FDR test A pulsed laser diode centered at 940 nm was employed as the excitation source for the Yb 3+ ions. The 50 s excitation pulses at a repetition rate of 50 Hz where reflecte d off a dichroic filter (HR @ 940 nm, HT @ 1040 nm) and transmitted through a NIR polarizing beam splitter cube where the excitation pulses were parallel to the plane of incidence of the beam splitting face. The pulse train excitation source was kept at a duty cycle which prohi bited any significant heating effects to the YbAG phosphor. A CW laser diode centered at 940 nm with orthogonal polarization was used in order to provide an internal mechanism of heating the YbAG phosphor. The 940 nm CW emission wa s normal to the plane of incidence to the beam splitting face of the NIR polarizi ng beam splitter cube. The polarized beams from the pulse train excitation source and th e CW source are then coupled into a silica fiber which is mated to a 2 m patch cabl e via a SMA to SMA connector. The opposite end of the fiber patch cable was fixed to a SMA to SMA mating sleeve that was used to connect to the YbAG phosphor temperature se nsor. The temperature probe measured approximately 10 cm in total le ngth and had a diameter of 400 m. The YbAG phosphor length at the end of the YAG lead fiber measur ed 1 mm in length with a diameter of
119 PLD =940 nm CWLD =940 nm Detecto r NIR Cube Polarizer Silica Fiber SMA Connector SMA Connector X-Y-Z Translation Sta g e Fiber-Optic Temperature Probe T/C Hot Plate Plexiglass Isolation Chambe r X-Y-Z Translation Sta g e HR 940 nm HT 1040 nm Computer & DSP Multimeter Figure 6.5: Schematic of the experimental layout for fiber optic thermally compensated YbAG temperature sensor.
120 400 m. The tip of the YbAG phosphor was polishe d perpendicular to the growth axis to a good optical finish using diamond lapping f ilms. The induced 1040 nm ytterbium fluorescence propagates back through the YAG lead fiber of the temperature probe and through the fiber patch cable. After exiting the patch cord s ilica mating assembly at the opposite end of the temperature probe, the be am was collimated by the plano-convex lens, transmitted back through the cube polarizer, and finally any scattered light is reflected by the dichroic filter while the fluorescence is transmitted. The fluorescence is collected by a silicon detector and then acquired by a digita l signal averaging oscilloscope for analysis. A computer and digital signal processor wa s also employed for data acquisition and signal processing in order to obtain real time decay rate measurements. The YbAG phosphor temperature probe wa s mounted onto an X-Y-Z stage in order to vary the gap distance between the phos phor tip and the sample surface of the hot plate which was oriented perpendicular to the growth axis of the probe. The X-Y-Z stage also provided a means to position the tip in close proximity to the sample surface region of interest. A multimeter was connected to a type E nickel-chromium vs. copper-nickel thermocouple which was adhered to the hot plates surface via high conductivity silver paint to monitor the surface te mperature. The experiment was conducted in a plexiglass isolation chamber at 1 atmos phere (760 torr) of pressure in order to minimize any wind drift which would cause e rroneous readings by the 25 m diameter thermocouple used to monitor the surface temperature. The YbAG phosphor temperature probe syst em was calibrated against a type S platinum-10% rhodium versus platinum ther mocouple which was oriented horizontally in close proximity with the phosphor tip. The calibrations were performed as described in
121 the former sections for the other phosphor pr obes presented. A first order exponential fit curve (t = 44.6 C) was used to make an acceptable fit to the calib ration points for the first 140 C as divulged in Fig. 6.6. Each data point in the figure represents the average of five measurements. Each measurement c onsists of the average of 256 decay traces. All the data points were fitted by a single e xponential over more than two e-foldings to determine the decay rate. The sample surface temperature was found to be 79 C +/2 C using the thermocouple that was fixe d to the hot plates surface. Extensive experience of the decay rate measurements in this temperature region has shown typical errors of 0.2 %. Fig. 6.7 displays the re lative sensitivity of the YbAG phosphor for the low temperature end of interest for this experi ment with a first orde r exponentia l decay fit (t = 56.6 C) to the data points. At approxi mately 79 C, a sensitivity value of .0015 /K is obtained corresponding to a calibration accuracy of better than 1.5 C. The measured temperature versus ga p length for the YbAG phosphor sensor at various initial heating temperat ures is presented in Fig. 6.8. The CW laser diode current driver was varied in order to dissipate vari ous laser powers into th e phosphor tip. Power transmission measurements disclosed that approximately only 2.30 % of the launched laser power in the YbAG sensor was transmitted through the tip of the phosphor. The diode currents of 0 A, 0.6 A, 0.65 A, 0.70 A, 0.75 A and 0.85 A were calculated to correspond to 0 mW, 98 mW, 130 mW, 169 mW, 202 mW, and 274 mW of launched laser power into the phosphor tip, respectively. Due to ytte rbium ions being primarily radiative, it is expected that only about 10 % of the calculated launch powers will actually be converted to thermal energy. In any even t, the corresponding la unch powers should be sufficient for the temperature range to be inspec ted in this endeavor. Each data point in
2030405060708090100110120130140150 2500 2550 2600 2650 2700 2750 2800 2850 2900 2950 3000 3050 3100 3150 3200 3250 3300 3350 3400 3450 3500 Decay Rate (/s)Temperature (C) Figure 6.6: Low temperature fluorescence de cay rate versus temperature for the YbAG phosphor. 122
102030405060708090100110120130140150 0.0000 0.0005 0.0010 0.0015 0.0020 0.0025 0.0030 0.0035 0.0040 0.0045 0.0050 Relative Sensitivity 1/W(T)[dW/dT] (/C)Temperature (C) Figure 6.7: Relative sensitivity of the YbAG sensor between room temperature and 140 C. 123
124 the figure is the average of 20 consecutiv e decay rate measurements which were converted into the corres ponding temperature reading using the calibration chart presented in Fig. 6.6. It is seen that for initial heating temp eratures below the actual surface temperature, which reflects the 98 mW and 130 mW laser la unching powers, result in the proliferation of the YbAG phosphors temperature rose upon contact with the surface. Conversely, for the initial temperatures above the actual su rface temperature, corresponding to the 169 mW, 202 mW, and 274 mW laser launching powers, a decreas e was detected in the YbAG phosphors temperature upon contact. It is quite apparent the smallest change experienced by the phosphors temperature pert ained to the case wher e the initial heating temperature was closest to the true surface temperature. This is delineated by the 169 mW case in the figure. When the YbAG sensor was not operated in the thermally compensated mode, the phosphors temperatur e was found to be approximately 27 C below the surface temperature when contact was made. It is not unreasonable that the error bars can be abated provi ding that the surface temperatur e of interest remains stable for an extended period of time. The nominal surface temperature of 79 C for this project was found to vary by +/2 C within a 10 minute heating-cooling cycle. The experimental data presented in Fig. 6. 8 echoes the model presented in Fig. 6.2. The smallest two gap distances from each data set in Fig. 6.8 excluding the case when no internal heating was a pplied (0 mW launch power case) are presented in Fig. 6.9. The smallest gap distance for each data set is 0 m in which the phosphor tip is abutting the surface of the hot plate. In reality this would correspond to a gap distance of approximately 50 m or so due to the semi-nodular surface which the hot plates surface
0.00.10.20.22.214.171.124.70.80.91.030 40 50 60 70 80 90 100 110 120 I = 0.0 Amps I = .60 Amps I = .65 Amps I = .70 Amps I = .75 Amps I = .85 Amps Pin = 274 mW Pin = 98 mW Pin = 0 mW Pin = 130 mW Pin = 169 mW Pin = 202 mW TT/C = 79 oC Temperature (C)Distance (mm) Figure 6.8: Experimental data of the phosphor temperature versus gap distance for various launching laser dissip ation values of the YbAG phosphor. 125
126 is composed of. In any event, linear extr apolations were applied in the effort to investigate the temperature at which each data set intersected with the others similar to the exercise introduced in Fig. 6.4. These calculations divulged 72.21 C and 84.73 C as the highest and lowest intersection temperatures, respectively, while the total average of all the intersections amongst al l the data sets was determined to be 79.74 C +/4.35 C. This experimental value obtained agrees rema rkably with the 79 C average temperature reading obtained from the thermocouple fixed in the same proximity of the sample surface. It is inferred that the proper m odus to apply the extrapolation intersection method to find the true surface temperature is to use two curves which are mirror images about a horizontal line through the true surf ace temperature. This would provide two data sets with approximately equal slopes in magnitude. In Fig. 6.9, this would represent the 98 mW and 202 mW launching cases. The ex trapolation intersection temperature for this corresponding case is 79.15 C. The temperature difference between the gap distances of 0 m and 100 m for each internal heating data set in Fig. 6.8 vers us the initial heat temperatures for each specific gap distance is presented in Fi g. 6.10. Calculation of the intersection temperature from the linear best fit lines for the contacting condition (dotted line fitted to hollow circles) and noncontacting condition (dashdot line fitted to squares) disclosed a intersection temperature of 80.167 C. Th e crossing intersection from negative to positive in the figure for the contacting and noncontacting conditions was found to be 80.49 C +/1.29 C for both respective conditions As exemplified from the figure, this technique also provides a dexterous means in calculating accurate surface temperatures for a fluorescence decay phosphor sensor operate d in the thermally compensated mode.
-0.25-0.20-0.15-0.10-0.05 0.000.050.10 60 65 70 75 80 85 90 95 100 105 110 Pin = 98 mW Pin = 130 mW Pin = 169 mW Pin = 202 mW Pin = 274 mW I = 0.0 Amps I = .60 Amps I = .65 Amps I = .70 Amps I = .75 Amps I = .85 Amps Temperature (C)Distance (mm) Figure 6.9: Phosphor temperature versus gap length at various launching laser dissipation values extracte d from Fig. 6.8 for specific gap distances of 0 m and 100 m. Extrapolation plots (dotted lines) are applied to each data set. 127
6065707580859095100105110 -8 -6 -4 -2 0 2 4 6 8 10 12 14 T (C)Temperature (C) Figure 6.10: Phosphor temperature differen ces between the ga p distances of 0 m and 100 m for each internal heating data set in Fig. 6.8 versus the initial heat temperatures for each specific gap distance of 0 m (hollow circles) and 100 m (squares). 128
129 A few implications may be implemented to accomplish the task of minimizing the measurement uncertainty using this thermally compensation technique. First the surface of interest should have the most uniform te mperature profile with minimum transient and steady-state temperature disturbances as po ssible to reduce measurement error for any gap distance temperature reading by the sensor. This criterion is extremely substantial in the extrapolation exercise used to inspect the intersection temperatures amongst the various initial heating temperat ures of the phosphor. Secondl y, any fluctuations from the internal heating source must be minimized fo r steady-state heating of the phosphor tip. Finally, idealistically one would desire to operate the thermally compensated fluorescence decay rate sensor at initial hea ting temperatures with the same magnitude temperature offsets above and below the actual surface temperature to be inspected. This would provide the harvest of a series of symmetric curves above and below the true surface temperature for the extrapolation exer cise to be applied. For gaseous media which are more thermally conductive opposed to the conditions used in this experiment, one would experience a decrease in the immensity of the shift in the phosphor temperature as the gap length is contracted.
130 CHAPTER 7 CONCLUSION In the course of this research there was successful demonstration in the improvement of the sensitivity of Yb:YAG based fluorescence decay rate temperature sensors. Long term stability tests performed unveiled these sensors as premier candidates to long term deployment in applications su ch as coal gasifiers, plasma/microwave processing, and a multitude of diverse industries. Pairing analysis of various rare earth elements with ytterbium in different ratio s disclosed improved response all the way down to room temperature which was nonexistent in our previous studies. The various codoped phosphor studies provided insigh t on applicable combinations to suit the response needs for specific process temperature ranges that may be needed in various technologies that may exist currently and that may develop in the future. Accurate surface temperature measuremen ts were achieved by operating a sensor in the thermally compensated mode for the firs t time. This novel technique is envisioned to be welcomed in plasma processing, microwave processing, and rapid thermal processing industries. With the combination of their inertness to electrically hostile environments, durability for long periods of tim e at elevated temperatures, and temporal response, this new technique could have significant economic impact in downtimes plaguing many industries. The thermally co mpensated sensor should be suitable for critical temperature measurements in which th e actual surface temperat ure of a sample is sought, by virtually extinguishing thermal drain loss currently experienced by
131 thermocouples. Future work should be directed toward application of the th ermally compensated fluorescence decay rate sensor to elevated temp eratures to conform to the needs of rapid thermal processing and other industrie s requiring accurate surface temperature measurements. It would be in teresting to see this techniqu e implemented to other hosts and fluorescing ions as well in order to verify the applicability of this novel method to other hosts. The methodology outlined in th is dissertation is appropriate for the temperatures for which this work was applied.
132 REFERENCES 1. J. L. Kennedy and N. Djeu, Sensors and Actuators A 100, 187 (2002). 2. E. Udd, Fiber Optic Sensors: An Introduc tion for Engineers and Scientists John Wiley & Sons, Inc., 1991. 3. E. Magison, 2001, Temperature Measuremen t-Physical Principles Underlie the Four Common Methods, in: Temperature Measurement in Industry (Inst. Soc. of America, Research Triangle Park, NC) Vol. 1, p.39. 4. M. R. Werner and W. R. Gahrner, IEEE trans. Ind. Electron., 48, 249 (2001). 5. J. R. Leigh, Temperature Measurement and Control Chapter 5, Perter Peregrinus Ltd., London, United Kingdom, 1988. 6. G. B. Lu and H. C. Yan, Proc. The 18 th IEEE, 2, 1221 (2001). 7. J. Castrellon, G. Paez, and M. Stronjnik, Opt. Eng., 41, 1255 (2002). 8. K. M. Koo, J. H. Kim, S. B. Kim, H. D. Kim, and D. Y. Ko, Proc. IECIC, 2, 229 (2001). 9.. E. P. Hastie and D. Bonnell, Alkali Va por Transport in Coal Conversion and Coal Combustion Systems, ACS Symposium Series, Vol. 179, Metal Bonding and Interactions in High Temperature System s with Emphasis on Alkali Metals (1982). 10. T. Sun, Ceramic News, 2,1 (1991). 11. J Schlichting, Oxidation and Hot Corro sion Behavior of SiN and SiAlON, in: Nitrogen Ceramics ed. by F. L. Riley, (Noordhoff, Leyden), p. 627. 12. J. Smialik and N. Jacobson, J. Am. Ceram. Soc. 59, 741 (1986). 13. M. Feber, J. Ogle, V. Tennery and T. Henson, J. Am. Ceram. Soc. 68, 191 (1985). 14. M. Feber and V. Tennery, Ceram. Bull, 62, 236 (1983). 15. D. Mckee and D. Chatterji, J. Am. Ceram. Soc. 59, 441 (1976). 16. T. Sun, Masters Thesis, Virginia Polytechnic Institute, (1986).
133 17. M. V. Roode and J. Price, 1991, Corrosion Resistant Coatings for Ceramic Heat Exchanger Tubes Operating in High ly Corrosive Environments, in: Performance Ceramic Coatings and Films, (Elsevier Science Publishers, New York) p. 625. 18. G. R. Pickrell, Ph.D. Dissertation, Virginia Tech, (1994). 19. J. V. Nicholas and D. R. White, Traceable Temperatures Wiley, Chichester 2001. 20. R. Baierlein, Thermal Physics Cambridge University Press, 1999. 21. R. E. Bentley, 1998, Temperature and Humidity Measurement, in: Handbook of Temperature Measurement, (Singapore; New York: Springer) Vol. 1. 22. B. Lawton and G. Klingenberg, Transient Temperature in Enginering and Science Oxford, 1996. 23. L. C. Lynnworthand and E. H. Carnav ale, 1972, Ultrasonic Thermometry Using Pulse Techniques, in: Temperature: Its Measurement and Control in Science and Industry ed. H. H. Plumb, (Inst. Soc. America, Pittsburgh) Vol. 4, p. 715. 24. D. W. Varela, 1992, Temperature Measur ement in Industrial and Laboratory Furnace Using Ultrasonic Thermometry, in: Temperature: Its Measurement and Control in Science and Industry ed. J. F. Schooley, (Amer. Inst. Phys. New York ) Vol. 6, pt. 2, p. 1027. 25. S. F. Green, 8 th Int. Heat Transfer C onf., San Francisco, (1986). 26. S. M. Vaezi-Nejad, 2000, Selected Topics in Advanced Solid State and Fiber Optic Sensors, Published by the (Institute of Electrical Engineers, London, UK), IEE Circuits, Devices, and Systems Series 11. 27. O. Lida and T. Iwamura, 1992, A Fibe r Optic Distributed Sensor for High Tempeature Measurements, in: Temperature, Its Measurement and Control in Science and Industry ed. J. F. Schooley, (American Institute of Physics, New York) Vol. 6, pt. 2, p. 745. 28. P. C. Wait, K. DeSouza, and T. P. Newson, Optics Comm. 144, 17 (1997). 29. D. P. Dewitt and G. D. Nutler, Theory and Practice of Radiation Thermometry Wiley, 1988. 30. D. Henry, Masters Thesis, University of South Florida, (1999). 31. E. Hecht, Optics, Addison-Wesley, 1987. 32. C. Lee and H. Taylor, J. of Lightwave Tech. 9, 129 (1991).
134 33. C. Lee and H. Taylor, Optics Lett. 13, 1814 (1997). 34. R. Atkins, C. Lee, and H. Taylor, Optics Lett. 13, 1038 (1988). 35. K. T. V. Grattan, Meas. Control 20, 32 (1987). 36. H. C. Seat and J. H. Sharp, Proceedings of SPIE 4185, 54 (2000). 37. D. M. Henry, J. H. Herringer, and N. Djeu, App. Phys. Lett. 74, 3447 (1999). 38. K. T. V. Grattan, J. D, Manwell, S. M. L. Sim, and C. A. Willson, Opt. Commun. 62, 104 (1987). 39. K. T. V. Grattan, A. W. Palmer, and C. A. Willson, J. Phys. E: Sci. Instrum. 20, 1201 (1987). 40. R. S. Feigelson, Mater. Sci. Eng. B1, 67 (1988). 41. Neubert, US Patent No. 2,551,650 (8 May 1937). 42. L. C. Bradley III, Rev. Sci. Instrum. 24, 219 (1953). 43. T. Bosselmann, A. Reule, and J. Schroeder, Proc. SPIE 514, 151 (1984). 44. K. T. V. Grattan, R. K. Selli, and A. W. Palmer, Rev. Sci. Instrum. 59, 1328 (1988). 45. H. Aizawa, N. Ohishi S. Ogawa, A. Endo, A. Hakamada, T. Katsumata, S. Komuro, T. Morikawa, and E. Toba, Sensors and Actuators A 101, 42-48 (2002). 46. V. C. Fernicaola, A. Y. Zhang, and K. T. V. Grattan, Rev. Sci. Instrum. 68, 2418 (1997). 47. T. Sun, Z. Y. Zhang, K. T. V. Grattan, A. W. Palmer, and S. F. Collins, Rev. Sci. Instrum. 68, 3442 (1997). 48. K. Wickersheim and R. Alve s, Ind. Res. Dev. 21, 82 (1979). 49. K. A. James, W. H. Quick, and V. H. Strahan, Control Eng. 26, 30 (1979). 50. R. R. Sholes and J. G. Small, Rev. Sci. Instrum. 51, 882 (1980). 51. J. S. McCormack, Electron. Lett. 17, 630 (1980). 52. Z. Zhang, J. H. Herringer, and N. Djeu, Rev. Sci. Instrum. 68, 2068 (1997).
135 53. R. T. VanCleave, Masters Thesis University of South Florida (2000). 54. J. L. Kennedy, D. M. Henry, and N. Djeu, Proc. of SPIE, 4833, 166 (2002). 55. J. L. Kennedy, N. Djeu, J. of Luminescenc, 101, 147 (2002). 56. R. C. Roop, Luminscence and the Solid Stat e, Studies in Inorganic Chemistry Vol. 12 (Elsevier, Amsterdam, 1991). 57. J. L. Sommerdijk and A. Bril. J. Electrochem. Soc. 122, 952 (1975). 58. L. Ozawa and P. M. Jaffe, J. Electrochem. Soc. 118, 1678 (1971). 59. H. Yamamoto and T. Kano, J. Electrochem. Soc. 126, 305 (1979). 60. P. H. Dowling and J. R. Sewell, J. Electrochem. Soc. 100, 22 (1953). 61. S. Imanaga, S. Yokono, and T. Hosh ima, Jpn. J. Appl. Phys. 19, 41 (1980). 62. H. P. Christensen, D. R. Gabbe, and H. P. Jenssen, Phys. Rev. B 25, 1467 (1982). 63. H. Kusama, O. J. Sovers. And T. Yoshioka, Jpn. J. Appl. Phys. 15, 2349 (1976). 64. G. H. Dieke, 1963, Paramagnetic Resonance, ed. By W. Low (Academic Press Inc., New York) Vol. 2, p. 237. 65. L. A. Risenberg, W. B. Gandrud, a nd H. W. Moss, Phys. Rev, 159, 262 (1967). 66. G. E. Barasch and G. H. Dieke, J. Chem. Phys. 43, 988 (1965). 67. M. J. Weber, Physics of Quantum Electronics ed. by P. L. Kelley, B. Lax, and P. E. Tannenwald (McGraw-Hill Book Co., Inc., New York) p. 237. 68. L. A. Risenberg and H. W. Moss, Phys. Rev. 174, 429 (1968). 69. D. L. Dexter and T. Miya kawa, Phys. Rev. B1, 2961 (1970). 70. N. Yamada, S. Shinoya, and T. Kashida, J. Phys. Soc. Jpn. 32, 1577 (1972). 71. J. P. Hurrell, S. P. S. Porto, I. F. Chang, S. S. Mitra, and R. P. Bawman, Phys. Rev. 173, 851 (968). 72. E. VonGomperz., Z. Phys. 8, 184 (1922). 73. C. Herring, J. K. Glat, Phys. Rev. 85, 1060 (1952).
136 74. A. V. Stepanov, Soviet Phys. JETP 29, 339 (1959). 75. R. S. Feigelson, 1985, Growth of single crystal fibers, in: Crystal Growth of Electronic Materials ed. E. Kaldis, (Elsevier Sc ience Publishers, Amsterdam) Chapter 11. 76. D. G. Gasson and B. Cockayne J. Mater. Sci. 5, 100 (1970). 77. J. S. Haggerty and W. P. Menashi, NASA-CR-72811 (1971). 78. C. A. Burrus and J. Stone, Appl. Phys. Lett. 26, 328 (1975). 79. M. M. Fejer, Ph.D. Dissertation, Stan ford University, Stanford, CA (1986). 80. R. S. F. Chang, S. Sengupta, G. J. Dixon, L. B. Shaw, and N. Djeu, Proceedings of SPIE 1104, 244 (1989). 81. M. Fejer, J. Nightingale, G. Magel, R. L. Byer, Rev. Sci. Instrum., 55,1791 (1984). 82. S. M. Jacobsen, B. M. Tissue, W. M. Yen, J. Cryst. Growth 109, 329 (1991). 83. G. deWith, 1987, in High Technology Ceramics ed. P. Vincenzini, (Elseveier Science Publishers, Amsterdam) p. 2063. 84. R. C. Powell, Physics of Solid State Laser Materials, Springer Verlag 1998 85. R.D. Shannon and C. T. Prewitt, Acta Crystallogr. B25, 925 (1969). 86. Y. N. Xu and W. Y. Ch ing, Phys. Rev. B 59, 10530 (1999). 87. Y. N. Xu and W. Y. Ch ing, Phys. Rev. B 59, 12815 (1999). 88. J. D. French, J. Zhao, M. P. Harmer, H. M. Chan, and g. A. Miller, J. Am. Ceram. Soc. 77, 2857 (1994). 89. S. Deng, Mater. Sci., 31, 6077 (1996). 90. R. S. Hay, J. Am. Ceram. Soc. 77, 473 (1994). 91. W. R. Blumenthal and S. T. Taylor, Acta. Mater. 45, 3071 (1997). 92. C. M. Wang, G. S. Cargill, M. P. Harmer, H. M. Chan, and J. Cho, Acta. Mater. 47, 3411 (1999). 93. J. S. McCormack, Electron. Lett. 17, 630 (1981).
137 94. Zhiyi Zhang, K. T. V. Grattan, and A. W. Palmer, Rev. Sci. Instrum. 63, 3177 (1992). 95. K. T. V. Grattan and A. W. Palm er, Rev. Sci. Insturm. 56, 1784 (1985). 96. X. Xu, Z. Zhao, J. Xu, and P. Deng, J. Cryst. Growth 255, 338 (2003). 97. A. R. Reinberg, L. A. Riseberg, R. M. Brown, et al., Appl. Phys. Lett. 19, 11 (1991). 98. M. J. F. Digonnet, Rare Earth Doped Fiber Lasers and Amplifiers Dekker, 1993. 99. S. V. Chernikov, Y. Zhu, and J. R. Taylor, Opt. Lett. 22, 298 (1997). 100. R. Paschotta, J. Nilsson, P. R. Barber, J. E. Caplen, A. C. Tropper, and D. C. Hanna, Opt. Commun. 136,375 (1997). 101. T. Kimura, A. Yokoi, Y. Nishida, R. Saito, and S. Yugo, Appl. Phys. Lett. 67, 2687 (1995). 102. E. Maurice, S. A. Wade, S. F. Co llins, and G. Monnom, Appl. Opt. 36, 8264 (1997). 103. X. Xu, Z. Zhao, J. Xu, and P. Deng, J. of Crystal Growth, 257, 272 (2002). 104. Y. Chen, P. Lim, S. Lim, Y. Yang, L. Hu, and W. Tse, J. of Raman Spect., 34, 882 (2003). 105. A. A. Kaminskii, Crystalline Lasers CRC, Boca Raton, 1996. 106. M. J. Weber, Phys. Rev. B 4, 9 (1971). 107. X. Xu, Z. Zhao, J. Xo, and P. Deng, J. of Cyrstal Growth 257, 272 (2003). 108. X. Xu, Z. Zhao, J. Xo, and P. Deng, J. of Cyrstal Growth 225, 338 (2003). 109. M. Hehlen, J. Opt. Am. B 14, 1312 (1997). 110. M. Hehlen, 1996, in OSA TOPS on Advanced Solid-State Lasers, eds. S. Payne and C. Pollock, (Optical Society of America, Washington, D.C.) p. 530.
Appendix A: Fluorescence Stability 024681012 820 830 840 850 860 870 Average = 848 /s Standard Deviation = +/1.0 /sFluorescence Decay Rate (/s)Time (Hours) Figure A.1: 12 hour fluorescence decay ra te data for 10%Yb:YAG phosphor at 1,118 K +/1.8 K. 139
Appendix A: (Continued) 012345678910111213 860 865 870 875 880 885 890 895 900 Average = 882 /s Standard Deviation = +/2.0 /sFluorescence Decay Rate (/s)Time (Hours) Figure A.2: 12 hour fluorescence decay ra te data for 10%Yb:YAG phosphor at 1,215 K +/2.0 K. 140
Appendix A: (Continued) 012345678910111213 900 920 940 960 980 1000 1020 1040 Average = 985 /s Standard Deviation = +/9.5 /sFluorescence Decay Rate (/s)Time (Hours) Figure A.3: 12 hour fluorescence decay ra te data for 10%Yb:YAG phosphor at 1,311 K +/2.2 K. 141
Appendix A: (Continued) 012345678910111213 1200 1300 1400 1500 1600 Average = 1,402 /s Standard Deviation = +/34.7 /s Fluorescence Decay Rate (/s)Time (Hours) Figure A.4: 12 hour fluorescence decay ra te data for 10%Yb:YAG phosphor at 1,399 K +/1.6 K. 142
Appendix A: (Continued) 012345678910111213 2800 3000 3200 3400 3600 3800 4000 4200 Average = 3,647 /s Standard Deviation = +/106.5 /sFluorescence Decay Rate (/s)Time (Hours) Figure A.5: 12 hour fluorescence decay ra te data for 10%Yb:YAG phosphor at 1,483 K +/2.3 K. 143
Appendix A: (Continued) 012345678910111213 8000 8500 9000 9500 10000 10500 11000 11500 12000 12500 13000 Average = 10,652 /s Standard Deviation = +/797.9 /sFluorescence Decay Rate (/s)Time (Hours) Figure A.6: 12 hour fluorescence decay ra te data for 10%Yb:YAG phosphor at 1,575 K +/5.4 K. 144
Appendix A: (Continued) 012345678910111213 21000 21500 22000 22500 23000 23500 24000 24500 25000 Average = 23,100 /s Standard Deviation = +/297.8 /sFluorescence Decay Rate (/s)Time (Hours) Figure A.7: 12 hour fluorescence decay ra te data for 10%Yb:YAG phosphor at 1,665 K +/2.0 K. 145
Appendix A: (Continued) 012345678910111213 23000 23500 24000 24500 25000 25500 26000 26500 27000 Average = 25,179 /s Standard Deviation = +/322.4 /sFluorescence Decay Rate (/s)Time (Hours) Figure A.8: 12 hour fluorescence decay ra te data for 10%Yb:YAG phosphor at 1,686 K +/2.6 K. 146
Appendix A: (Continued) 012345678910111213 960 980 1000 1020 1040 1060 1080 1100 1120 Average = 1,042 /s Standard Deviation = +/13.3 /sFluorescence Decay Rate (/s)Time (Hours) Figure A.9: 12 hour fluorescence decay ra te data for 20%Yb:YAG phosphor at 1,118 K +/1.9 K. 147
Appendix A: (Continued) 012345678910111213 1140 1160 1180 1200 1220 1240 1260 1280 1300 1320 1340 1360 1380 Average = 1,288 /s Standard Deviation = +/16.1 /sFluorescence Decay Rate (/s)Time (Hours) Figure A.10: 12 hour fluorescence decay ra te data for 20%Yb:YAG phosphor at 1,221 K +/1.5 K. 148
Appendix A: (Continued) 012345678910111213 1500 1600 1700 1800 1900 2000 2100 2200 2300 2400 2500 Average = 1,996 /s Standard Deviation = +/89.4 /sFluorescence Decay Rate (/s)Time (Hours) Figure A.11: 12 hour fluorescence decay ra te data for 20%Yb:YAG phosphor at 1,317 K +/2.6 K. 149
Appendix A: (Continued) 012345678910111213 4400 4500 4600 4700 4800 4900 5000 5100 5200 5300 5400 5500 Average = 5,029 /s Standard Deviation = +/147.7 /sFluorescence Decay Rate (/s)Time (Hours) Figure A.12: 12 hour fluorescence decay ra te data for 20%Yb:YAG phosphor at 1,403 K +/2.9 K. 150
Appendix A: (Continued) 012345678910111213 11600 11800 12000 12200 12400 12600 12800 13000 13200 13400 13600 Average = 12,625 /s Standard Deviation = +/167.2 /sFluorescence Decay Rate (/s)Time (Hours) Figure A.13: 12 hour fluorescence decay ra te data for 20%Yb:YAG phosphor at 1,490 K +/2.0 K. 151
Appendix A: (Continued) 012345678910111213 23500 24000 24500 25000 25500 Average = 24,636 /s Standard Deviation = +/263.2 /sFluorescence Decay Rate (/s)Time (Hours) Figure A.14: 12 hour fluorescence decay ra te data for 20%Yb:YAG phosphor at 1,571 K +/2.8 K. 152
Appendix A: (Continued) 012345678910111213 26000 26200 26400 26600 26800 27000 27200 27400 27600 27800 Average = 26,907 /s Standard Deviation = +/188.14 /sFluorescence Decay Rate (/s)Time (Hours) Figure A.15: 12 hour fluorescence decay ra te data for 20%Yb:YAG phosphor at 1,661 K +/1.9 K. 153
Appendix A: (Continued) 012345678910111213 2500 2550 2600 2650 2700 2750 2800 2850 2900 2950 3000 3050 3100 3150 3200 Average = 2,877 /s St. Dev. = +/66 /s Fluorescence Decay Rate (/s)Temperature (K) Figure A.16: 12 hour fluorescence decay ra te data for 50%Yb:YAG phosphor at 1,124 K +/1.2 K. 154
Appendix A: (Continued) 012345678910111213 3900 4000 4100 4200 4300 4400 4500 4600 4700 4800 4900 Average = 4,394 /s St. Dev. = +/66 /s Fluorescence Decay Rate (/s)Temperature (K) Figure A.17: 12 hour fluorescence decay ra te data for 50%Yb:YAG phosphor at 1,221 K +/1.8 K. 155
Appendix A: (Continued) 012345678910111213 8000 8100 8200 8300 8400 8500 8600 8700 8800 8900 9000 9100 9200 9300 9400 9500 9600 9700 9800 9900 10000 Average = 9,145 /s St. Dev. = +/181 /s Fluorescence Decay Rate (/s)Temperature (K) Figure A.18: 12 hour fluorescence decay rate data for 50%Yb:YAG phosphor at 1,316 K +/2.2 K. 156
Appendix A: (Continued) 012345678910111213 16000 16200 16400 16600 16800 17000 17200 17400 17600 17800 Average = 17,200 /s St. Dev. = +/318 /s Fluorescence Decay Rate (/s)Temperature (K) Figure A.19: 12 hour fluorescence decay ra te data for 50%Yb:YAG phosphor at 1,406 K +/1.9 K. 157
Appendix A: (Continued) 012345678910111213 23000 23500 24000 24500 25000 25500 Average = 24,339 /s St. Dev. = +/287 /s Fluorescence Decay Rate (/s)Temperature (K) Figure A.20: 12 hour fluorescence decay ra te data for 50%Yb:YAG phosphor at 1,495 K +/2.7 K. 158
Appendix A: (Continued) 012345678910111213 27000 27100 27200 27300 27400 27500 27600 27700 27800 27900 28000 28100 28200 28300 28400 28500 28600 28700 28800 28900 Average = 28,030 /s St. Dev. = +/264 /s Fluorescence Decay Rate (/s)Temperature (K) Figure A.21: 12 hour fluorescence decay ra te data for 50%Yb:YAG phosphor at 1,581 K +/2.8 K. 159
Appendix B: Thermal Compensation Model 160 2a d a+d T T0TS 45 x L Figure B.1: Schematic model for thermal transport between th e phosphor tip and sample.
Appendix B: (Continued) The temperature profile of the phosphor depends upon the rate of its internally generated heat (via CW 940 nm ra diation), its capacity to stor e this heat, and the rate of thermal conduction to its boundaries (where heat is transferred to the surrounding environment). Thermal transport between the tip and sample can be demonstrated by calculating the approximate thermal resistance ( T.R. Tip ) from the formula dx A 1 T.R.Tip 2 d 0 2 Tipx 1 x 1 dx d :note,dx ax 1 T.R. a 1 ad 1 1 T.R.Tip ada d 1 T.R.Tip (B.1) where A is the cross-sectional area of the inverted frustum, a the radius of the inverted frustum at the base (at phosphor tip), d is the gap distance between tip and sample, (a+d) the radius of the frustum at th e top (at surface of interest) and is the thermal conductivity of the gas as shown in Fig. B.1. The reciprocal of the thermal resistance ( T.R. Tip ) yields the thermal conductance ( T.C. Tip ), thus d a 1 a T.C.Tip (B.2) In an effort to approximate the rate of heat transfer between the sample and the sidewalls of the phosphor tip, we assume the le ngth of the sidewall to participating in thermal transport from the end of the phosphor tip to be approximately The sidewall area is calculated by 2 161
Appendix B: (Continued) 2a d T T0TS 45 L x 2 d 2d Figure B.2: Schematic model for thermal transport between th e phosphor sidewalls and sample. 162
Appendix B: (Continued) aah 422r2 (B.3) where is the length of the phosphor tip referring to Fig. B.2. The area of the sample of interest consid ered to participate in thermal transport with the sidewall is that of a disk with radius (d+) and is intercepted at 45 with the sidewall surface The corresponding area is given by 163 (B.4) 2 2dr where d is the gap distance referring to Fig. B.2. The geometric mean area is found to be the square root of the product of Eq. B.3 and Eq. B.4, giving da da 2 42 2 (B.5) Thermal resistance ( T.R. Side ) for the sidewalls can be approximated by d 0 d 0 1/2 Side Sidedxx a2 1 dx xa2 1 T.R. dx A 1 T.R. yielding the semi-quan titative result of a2 1 T.R.Side (B.6) Taking the reciprocal to approxi mate the thermal conductance ( T.C. Side ) gives .a2 T.C.Side (B.7)
Appendix B: (Continued) The heat equation following from the conservation of energy is (B.8) Change Energy Internal Out Conducted Heat Within Generated Heat In Conducted Heat where the heat conducted in will be through th e phosphor end and the sidewall of length from Eq. B.2 and Eq. B.7 respectively, the heat generated within will be from a CW 2 940 nm heating source represented by P for our system, and the only mechanism of heat loss considered is to be along the leng th of the probe lead fiber given by 0 2TT L 'a (B.9) where L is the length of the lead fiber, T 0 the temperature of the heat sink at the end of the YAG lead fiber and the thermal conductivity of the YAG lead. We assume that the phosphor has a uniform temperature T which sets the change in internal energy to zero. Applying these assumptions to Eq. B.8 gene rates the heat equation pertaining to our system which is given by 0 2 S STT L 'a PTTaTT d a 1a (B.10) Solving Eq. B.10 for T, the temperature of the phosphor tip can be obtained from the formula L a T d aa T L P Ts 0 d a aa a a2 2 2 2 (B.11) 164
About the Author Jermaine L. Kennedy received a B.S. in Physics in 1998 and a B.S. in Secondary Education in 1999 from the State University of New York at Buffalo State. In 2001 he completed a M.S. in Physics from the Un iversity of South Florida in 2001.