USF Libraries
USF Digital Collections

Ultrasound hardware setup for CMP pad characterization


Material Information

Ultrasound hardware setup for CMP pad characterization
Physical Description:
Tadi, Bhaskar Vijay Kumar Reddy
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:


Subjects / Keywords:
Dissertations, Academic -- Electrical Engineering -- Masters -- USF   ( lcsh )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


ABSTRACT: Chemical Mechanical Polishing, (CMP), pads made of polyurethane material are utilized in the Integrated Circuit, (IC), industry to planarize wafers between successive process steps. The properties of such pads and their behavior must be known in order to determine under what conditions and for how long they can be used efficiently. This research involved the development of a system to study the properties of such pads. The system developed during this research enabled the pads to be tested under varying physical conditions. The setup used a combination of several instruments to provide excitation to the pad and acquire a measure its response. A central computer controlled the instrumentation system employed. In this research the determination of the physical properties of CMP pads was accomplished through the use of Ultra Sound testing. Ultra sound methods offer a non-destructive method of characterizing pads to be used in the production of IC wafers. Ultra sound characterization is currently one of the most widely used techniques utilized for non-destructive inspection. This report provides a detailed account of the hardware instruments involved and the method of integration of those instruments into a system that could easily, rapidly and accurately characterize CMP pads. The pad response was measured in terms of the signal voltage transmitted through the pad to the ultrasound sensor. The software stored these readings for every set of testing conditions. Changing the temperature, humidity and depth from the pad's surface where measurements are made changed the test conditions. These data were analyzed statistically to determine the behavior of the pad. This research was part of a larger research effort that provided the statistical tool required to determine the uniformity of a CMP pad.
Thesis (M.S.E.E.)--University of South Florida, 2004.
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Bhaskar Vijay Kumar Reddy Tadi.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 58 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001469439
oclc - 55732468
notis - AJR1193
usfldc doi - E14-SFE0000340
usfldc handle - e14.340
System ID:

This item is only available as the following downloads:

Full Text
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 001469439
003 fts
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 040524s2004 flua sbm s000|0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0000340
b SE
1 100
Tadi, Bhaskar Vijay Kumar Reddy.
0 245
Ultrasound hardware setup for CMP pad characterization
h [electronic resource] /
by Bhaskar Vijay Kumar Reddy Tadi.
[Tampa, Fla.] :
University of South Florida,
Thesis (M.S.E.E.)--University of South Florida, 2004.
Includes bibliographical references.
Text (Electronic thesis) 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 58 pages.
ABSTRACT: Chemical Mechanical Polishing, (CMP), pads made of polyurethane material are utilized in the Integrated Circuit, (IC), industry to planarize wafers between successive process steps. The properties of such pads and their behavior must be known in order to determine under what conditions and for how long they can be used efficiently. This research involved the development of a system to study the properties of such pads. The system developed during this research enabled the pads to be tested under varying physical conditions. The setup used a combination of several instruments to provide excitation to the pad and acquire a measure its response. A central computer controlled the instrumentation system employed. In this research the determination of the physical properties of CMP pads was accomplished through the use of Ultra Sound testing. Ultra sound methods offer a non-destructive method of characterizing pads to be used in the production of IC wafers. Ultra sound characterization is currently one of the most widely used techniques utilized for non-destructive inspection. This report provides a detailed account of the hardware instruments involved and the method of integration of those instruments into a system that could easily, rapidly and accurately characterize CMP pads. The pad response was measured in terms of the signal voltage transmitted through the pad to the ultrasound sensor. The software stored these readings for every set of testing conditions. Changing the temperature, humidity and depth from the pad's surface where measurements are made changed the test conditions. These data were analyzed statistically to determine the behavior of the pad. This research was part of a larger research effort that provided the statistical tool required to determine the uniformity of a CMP pad.
Adviser: Moreno, Wilfrido A.
Dissertations, Academic
x Electrical Engineering
t USF Electronic Theses and Dissertations.
4 856


Ultrasound Hardware Setup For CMP Pad Characterization by Bhaskar Vijay Kumar Reddy Tadi A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Electrical Engineering Department of Electrical Engineering College of Engineering University of South Florida Major Pr ofessor: Wilfrido A. Moreno, Ph.D. James T. Leffew, Ph.D. Grisselle Centeno, Ph.D. Date of Approval: March 30, 2004 Keywords: PID, SSR, UST, PSD, SCPI. Copyright 2004 Bhaskar Vijay Tadi


Acknowledgements I would like to thank Dr. Moreno, my major professor, fo r his guidance and support of me financially throughout my master degree program. I would like to thank Dr. Centeno who helped me to formulate, understand and present the statistical analysis associated with this re search. Dr. Leffew’s review of this thesis is also greatly appr eciated. I also acknow ledge the contribution of journals, papers, organizations and books, to which I had occasion to refer and, which are listed in the references. I would like thank my parents and Geet ha for being with me all the time. Without their love, affecti on, motivation and support this thesis would not have been possible. Last but not the least I wo uld like to thank my friends Prashant Datar, Raj and Ravi Yalamanchili.


i Table of Contents List of Tables iv List of Figures v Abstract vi Chapter 1 Introduction 1 1.0 Research Objectives 1 1.1 Chemical Mechanical Planarization 1 1.2 Ultra Sound Testing 3 Chapter 2 Data Acquisition System Description 4 2.0 Basic Devices 4 2.1 Lock-in Amplifier 7260 4 2.2 Typical Lock-in Amplifier 6 2.2.1 Signal Channel 6 2.2.2 Reference Channel 7 2.2.3 Mixer or Phase Sensitive Detector (PSD) 7 2.2.4 Low-pass Filter and Output Filter 10 2.2.5 Output of Lock-in Amplifier 10 2.3 Operating Modes of Lock-in Amplifier 11 2.3.1 Signal Recovery / Vector Voltmeter 11


ii 2.3.2 Single Reference / Dual Reference 11 2.3.3 Single Harmonic / Dual Harmonic 12 2.3.4 Internal / Exte rnal Reference Mode 12 2.3.5 Virtual Reference Mode 13 2.4 Functionality of Lock-in Amplifier 13 2.5 VP-9000 Motor Controller 13 2.6 PS2520G Power Supply1 15 Chapter 3 Temperature and Humidity Control 17 3.0 Temperature Control 17 3.1 Types of Control 17 3.1.1 On/Off Control 17 3.1.2 Proportional Control 18 3.1.3 PID Control 19 3.2 Temperature Sensor 20 3.3 Solid State Relay (SSR) 21 3.4 Temperature C ontroller 22 3.5 Humidity Control 25 3.5.1 Humidification Apparatus 25 Set-Up 25 Operation 26 3.5.2 De-Humidificat ion Apparatus 28 Set-Up 28 Humidity Sensor 30


iii 3.6 Integration of Humidity Apparatus 30 Chapter 4 System Integration 32 4.0 System Integration Tasks 32 4.1 Instruments Employed 32 4.2 Hardware Integration 36 4.3 Steps Required in Ta king Measurements 36 Chapter 5 Conclusions and Results 38 5.0 Conclusions 38 5.1 Effects of Ambient Conditions 39 5.2 Future Work 43 References 45 Appendices 46 Appendix A Temperature and Humidity Data 47


iv List of Tables Table 4.1: Instruments Integr ated Into the System 33 Table 5.1: Test Parameters 40 Table 5.2: Test Values for Sector 0 (0 -60 ) 41 Table 5.3: Test Values for Sector 1 (120 -180 ) 41 Table 5.4: Test Values for Sector 2 (240 -300 ) 42 Table 5.5: Multifactor Analysis Results with PID Control 43


v List of Figures Figure 1.1: Set-Up for Chemical Me chanical Planarization Process 2 Figure 2.1: DSP Lock-in Amplifier 7260 5 Figure 2.2: Block Diagram of Typi cal Lock-in Amplifier 6 Figure 2.3: Signal In with Certain Phase 8 Figure 2.4: Signal In with Dela yed Phase of 90 Degrees 9 Figure 2.5: VP9000 VELMEX Motor Controller 14 Figure 2.6: PS2520G Program mable Power Supply 15 Figure 3.1: K-type Thermocouple 21 Figure 3.2: Solid State Relay (SSR) 22 Figure 3.3: AC Controlled SSR Us ed with Temperatur e Controller with Mechanica l Relay Output 23 Figure 3.4: Thermocouple Hookup and Wiring for RS232 to Temperature Controller 23 Figure 3.5: Power and SSR Wiring to Temperature Controller 24 Figure 3.6: I-Series CNi16D Te mperature Controller 24 Figure 3.7: Humidification Apparatus 27 Figure 3.8: De-Humidification Apparatus 29 Figure 3.9: HIH-3610-001 Humidity Sensor 30 Figure 4.1: CMP Pad Test Set-Up 34 Figure 4.2: Temperature and Hu midity Control 35 Figure 5.1 600 Test Sectors on Pad 40


vi UTRASOUND HARDWARE SETUP FOR CMP PAD CHARACTERIZATION Bhaskar Vijay Tadi ABSTRACT Chemical Mechanical Polishing, (C MP), pads made of polyurethane material are utilized in the Integrated Circu it, (IC), industry to planarize wafers between successive process steps. T he properties of such pads and their behavior must be known in order to determine under what conditions and for how long they can be used efficiently. This research involved t he development of a system to study the properties of such pads. The system developed during this research enabled the pads to be tested under varying physical conditions. The setup used a combination of severa l instruments to provide excitation to the pad and acquire a measure its res ponse. A central computer controlled the instrumentation system employed. In this research the determination of the physical properties of CMP pads was a ccomplished through the use of Ultra Sound testing. Ultra sound methods offer a non-destructive method of characterizing pads to be used in the production of IC wafers. Ultra sound characterization is currently one of the most widely us ed techniques utilized for non-destructive inspection.


vii This report provides a detailed account of the hardware instruments involved and the method of integrati on of those instru ments into a system that could easily, rapidly and accurately characterize CMP pads. The pad response was measured in terms of the signal volt age transmitted through the pad to the ultrasound sensor. The software stored th ese readings for every set of testing conditions. Changing the temperatur e, humidity and de pth from the pad’s surface were measurements are made chan ged the test conditions. These data were analyzed statistically to determine the behavior of the pad. This research was part of a larger research effort that provided the statistical tool required to determine the uniformity of a CMP pad.


1 Chapter 1 Introduction 1.0 Research Objectives The objective of this research was to build a system that could be used by engineers and researchers to study t he effects of Chemical Mechanical Planarization, (CMP), on pads used for po lishing wafers in the IC industry. Physical factors such as temperatur e and humidity play a major role in determining the characteristics of a CM P pad. The research required a system composed of an assembly of various dev ices and instruments. Each system component possessed a special functionality. A single host computer controlled all the components of the system. 1.1 Chemical Mechan ical Planarization Chemical mechanical polishing, (CMP), has been used by the wafer fabrication industry for many years. CM P is used to polish prime silicon wafer substrates and gave birth to chemical mechanical planarization as a means of


2 micromachining thin layers deposited or grown on the substrate. Figure 1.1 illustrates the mechanization for the chem ical mechanical planarization process. WAFER CARRIER POWER DRIVEN SPINDLE SILICON WAFER POLISHING PAD POLISHING TABLE Figure1.1: Set-Up for the Chemical Mechanical Planarization Process With the correct polishing pad slurry and planarizing machine tool, thin layers of insulting material can be remov ed at a rate as high as twenty wafers per tool per hour. CMP is ideal for polishing a layer when a relatively stiff polishing pad is used. During the polishing proces s the high spots or peak areas are polished first and brought down to a point w here eventually the layer is a plane surface. “Flat” is not a very meaningful term since the wafer will tend to conform to the surface upon which it is placed. Uniformity is al so desired so that the layer is equal in thickness, both locally and globally. Changes in physical factors such as temperature and humidity produce physical effects on the CMP pads. When this happens the surface-uniformity of wafers polished using the affected pads c annot be guaranteed. T herefore, it is


3 necessary to determine how the pads re spond to ambient changes and control the polishing process in such a way as to maintain wafer quality. 1.2 Ultrasound Testing Chemical mechanical planarization p ads were used primarily in the chemical mechanical polishing of optical devices in the earlier stages. Now CMP is used to polish IC wafers. In this research, determination of the mechanical properties of CMP pads was accomplished through the use of Ultrasound testing. Ultrasound methods offer a non-destructive means to characterize pads that are to be used in the production of IC wafers. Ultrasound Testing, (UST), is one of the most widely used techniques for non-destructive inspection. Ultrasound works on the principle of sound perm eability through an absor bing visco-elastic medium. The difference in ultrasound abso rption in areas of varying viscosity and density is used to determine any nonuniformity within a si ngle pad by giving an in depth idea of the physical c haracteristics of the given pads. During UST, an ultrasonic wave is sent through the pad and the change in the transmitted signal with time and dept h below the surface at specified frequencies, temperature and humidity are measured over the entire pad and reported. These readings allow the under lying pad structure to be determined. Material Sciences and Statistical Analys is tools are then employed to determine the various characteristics of the pad. Based on these studies, researchers can suggest improvements in the CMP process in order to enhance the quality of the pads and ultimately the polished wafers upon which they are used.


4 Chapter 2 Data Acquisition System Description 2.0 Basic Devices 2.1 Lock-in Amplifier 7260 A lock-in amplifier, like most AC i ndicating instruments, provides a DC output proportional to the AC signal under in vestigation. In modern units, the DC output may be presented as a reading on a digital panel meter or as a digital value communicated over a computer in terface, rather than a voltage at an output connector, but the principle remains the same. The trad itional rectifier, which is found in a typical AC voltmeter, makes no dist inction between signal and noise and produces errors due to rectif ied noise components. However, the noise at the input to a lock-in amplifier is not rectified but appears at the output as an AC fluctuation. This means that t he desired signal response, a DC level, can be separated from the noise ac companying it in the output by means of a simple low-pass filter. Hence, the fi nal output of a lock-in amplif ier is not affected by the presence of noise in the applied signal. T he front panel of the 7620 lock-in amplifier used in this research is presented in Figure 2.1.


5 Figure 2.1: DSP Lock-In Amplifier 7260 [9] A lock-in amplifier can be used for two basic purposes. To recover a signal in the presence of overwhelmi ng background noise or to provide high resolution measurements of relatively clean signals over several orders of magnitude and frequency. Modern instruments lik e the 7260 offer many additional features. These in struments are used in varied fields of research such as Optics, Electrical E ngineering, Fundamental Physics and Material Sciences. The 7260 Lock-In Amplifier prov ides the following functions: Precision Oscillator Vector Voltmeter Phase Meter AC Signal Recovery Frequency Meter Transient Recorder Spectrum Analyzer Noise Meter


6 2.2 Typical Lock-in Amplifier The lock-in amplifier was a crucial instrument utilized in this research. A block diagram of the instrument is presented in Figure 2.2. Figure 2.2: Block Diagram of a Typical Lock-in Amplifier 2.2.1 Signal Channel The signal channel amplifies the input signal, including noise, by using an adjustable-gain, AC-coupled, amplifier in or der to match it more closely to the optimum input si gnal range of the phase-sensitive detector, (PSD). Instruments are usually fitted with high impedance inputs for volt age measurements. Many also incorporate low impedance inputs fo r better noise matching to current sources. However, in some cases t he best results are obtained through the use of a separate external preamplifier. The performance of the PSD is usually improved if the bandwidth of t he noise voltages reaching it is reduced from that of the full frequency range of the instrument. To achieve this, the signal is passed through some form of filter, which may simp ly be a band rejecti on filter centered at the power line frequency and/ or its second harmonic. The band rejection filter would be used to reject line frequen cy pick-up. Alternatively a more


7 sophisticated tracking band-pass filter c entered at the reference frequency could be employed. 2.2.2 Reference Channel The reference channel provides a high-level, stable and noise-free reference input. A well-designed refer ence channel circuit is very important. Such circuits can be expensive and often account for a significant proportion of the total cost of the inst rument. The internally generated reference is passed through a phase shifter, which is used to compensate for phase differences that may have been introduced between the signal and reference inputs by the experiment, before being applied to the PSD. 2.2.3 Mixer or Phase Sensitive Detector (PSD) This special rectifier, called a phasesensitive detector, (PSD), or mixer, performs AC to DC conversion and forms the heart of the instrument. The PSD is special in the sense that it rectifies on ly the signal of inte rest while suppressing the effect of noise or interfering com ponents, which may accompany the signal. The detector operates by multiplying two si gnals together. T he following analysis indicates how this gives the required outputs.


8 Signal In Reference (Internally generated) Demodulator Output Figure 2.3: Signal In With Certain Phase Figure 2.3 indicates that the lock-in amplifier is detecting a noise-free sinusoid, which is identified in the diagram as “Signal In ”. The instrument also receives a reference signal. The refer ence signal is used to generate an internal sinusoidal reference, which is also s hown in the diagram. The demodulator operates by multiplying the two signals toge ther to yield the signal identified in the diagram as “Demodulator Output”. Since there is no relative phase-shift between the signal and reference phases the demodulator output takes the form of a sinusoid at twice the reference frequency with a mean or average level, which is positive.


9 Signal In Reference (Internally generated) Demodulator Output Figure 2.4: Signal-In With Delayed Phase of 90 Degrees Figure 2.4 illustrates a si milar situation where th e signal phase has been delayed by 90 with respect to the referenc e. Although the out put still contains a signal at twice the reference frequency, the mean level is shifted to zero. This analysis indicates that the mean level is: Proportional to the product of t he signal and reference frequency amplitudes Related to the phase angle bet ween the signal and reference. Therefore, if the referenc e signal amplitude is maintain ed at a fixed value and the reference phase is adjusted to ensure a relative phase-shift of zero degrees, the mean level the input signal amplitude can be determined. The mean level is, of course, the DC component of the demodulator output. It is a relatively simple task to isolate the DC component by using a low-pass filter. The filtered output is then measured using a conv entional DC voltmeter.


10 2.2.4 Low-pass Filter and Output Amplifier Practical instruments empl oy a wide range of output filter types, which are implemented either as analog circuits or in digital signal processors. Usually, these circuits are equivalent to one or more stages of single-pole “RC” type filters, which exhibit the classic 6-dB/octave roll-off with increasing frequency. Also, there is usually some form of out put amplifier, whic h may be either a DCcoupled analog circuit or a digital multiplier The use of an am plifier at the output in conjunction with the input amplifier al lows the unit to hand le a wide range of signal inputs. When there is little accom panying noise, the input amplifier can be operated at high gain without overloading the PSD and little gain is needed at the output. In the case of signals buried in very large noise voltages the reverse situation is required. 2.2.5 Output of Lock-in Amplifier The output from a lock-in amplifier wa s traditionally a DC voltage that was usually displayed on an analog panel mete r. Modern instruments, especially those instruments used under computer control, prov ide an output reading in the form of a digital number. The analog DC vo ltage signal is usually also provided. Lock-in amplifiers that employ an analog form of phase-sensitive detector use an analog-digital-converter, (ADC), to genera te their digital output. Digital multiplying lock-in amplifiers use a digita l to analog converter, (DAC), to generate the analog output.


11 2.3 Operating Modes of Lock-in Amplifier The model 7260 lock-in amplifier is a sophisticated instrument with many capabilities. It incorporates a nu mber of different operating modes. 2.3.1 Signal Recovery/Vector Voltmeter The model 7260 can be used for meas uring the phase of the applied signal with respect to the reference. T he accuracy of this measurement is not usually paramount. This operating mode is called the signal recovery mode and is the default mode at power-up. In case s where the applied signal is essentially free of noise some of the circuitry needed for best signal recovery performance may be bypassed, which provides an im provement in the accuracy of phase measurements. However, the accuracy ac hieved is accompanied by an increase in noise. Selective circuitry selection is available in the vector voltmeter mode. 2.3.2 Single Reference/Dual Reference The lock-in amplifier can measure both the signal magni tude and phase of the applied signal at a single reference fr equency. This is referred to as the single reference mode. The dual referenc e mode in the model 7260 allows the instrument to perform si multaneous measurements at two different reference frequencies. This flexibility incurs a few re strictions such as the requirement that one of the reference signals is external and the other is derived from the internal


12 oscillator. In addition, the maximum operating frequency is limited to 20 kHz and requires that both signals be passed th rough the same input signal channel. 2.3.3 Single Harmonic/Dual Harmonic In some applications such as auger spectroscopy and amplifier characterization, it is useful to be able to make measurements at some multiple “n” or harmonic of the reference frequency “F ”. The only restri ction is that the product “nF” cannot exceed 250 kHz. The dual harmonic mode allows the simultaneous measurement of two different harmonics of the input signal. As with the dual reference mode there are a few restrictions such as a maximum “nF” value of 20 kHz. Wh ile in the dual harmonic mode there is no advantage to the use of the vector voltmeter mode. Therefore, simultaneous use of the dual harmonic mode and vector voltme ter mode is not recommended. 2.3.4 Internal/External Reference Mode When the internal reference mode is selected the instrument’s reference frequency is derived from its internal oscilla tor and the oscillator signal is used to drive the experiment. When the exter nal reference mode is selected the experiment must include some device t hat can generate a reference frequency, which is applied to the lock-in amplif ier’s external reference input. The instrument’s reference channel “locks” to the external reference signal and uses it to measure the applied input signal.


13 2.3.5 Virtual Reference Mode In the virtual reference mode, th e Y channel output is used to make continuous adjustments to the internal oscillator frequency and phase in order to achieve phase-lock with the applied signa l. The adjustments are performed in such a manner that the X channel output is maximized and the Y channel output is zeroed. 2.4 Functionality of Lock-in Amplifier The Lock-In amplifier was central to th e proper functioning of the system. This design used its oscillator and voltmete r sections. The oscillator generated a precise 26 KHz signal with am plitude of 0.5 V, peak to peak, that was amplified and used to excite the pad under test. The response from the pad was detected by an ultrasonic transducer probe and applied to a chann el on the lock-in amplifier. The amplifier was set to disp lay and transmit the readings to the host computer. 2.5 VP-9000 Motor Controller The VP 9000 is a programmable steppe r motor controller, which is capable of running up to four motors. The controller uses a microprocessor, support circuitry and possesses 64 Kiloby tes of nonvolatile Random Access Memory, (RAM), for storing setup par ameters and programs. Commands and data can either be entered usi ng the RS-232 Serial Interfac e or by using the front


14 panel menu. An alphanumeric display prov ides visual access to motor positions and setup parameters. The capability of having a host computer send commands to the controller through its RS -232 serial interface is available. A unislide Velmex motor was used to move the sensor to different parts of the pad. A stepper motor was used for t he angular displacement of the pad. A vacuum was used to hold the pad tight at the time of data acquisition to insure that there was no airgap between the pad and the sensor. The VP 9000 is pictured in Figure 2.5 Figure 2.5: VP 9000 VELMEX Motor Controller Motion can be specified in absolute as well as relative indices. An absolute index is measured relative to t he absolute zero position. A relative index is measured in a spec ified direction for a specified distance from the present position. The in strument manual provided tables that allowed an estimate the number of steps required, by the motor, to cover a given distance. Of the four available lines on the VP 9000 only three were used. The instrument controlled the position of the sensor relative to the center of the pad, (X-axis), the


15 height of the sensor above the pad surfac e, (Z-axis), and the angle of the pad with respect to the sensor position, (Radial axis). 2.6 PS-2520G Power Supply The PS-2520G is a Programmable Power Supply from Tektro nix. It offers three power outputs. Two of these ar e supply voltages from 0 to 36V and currents from 0 to 1.5A. The third one has a higher current capability of 0 to 3A and supplies voltages from 0 to 6V. The instrument has a Light Emitting Diode, (LED), display to indicate the volt age and current levels. The PS-2520G Programmable Power Supply is pictured in Figure 2.6. Figure 2.6: PS-2520G Pr ogrammable Power Supply The PS2520G possesses a GPIB IEEE-488 interface that enables a host computer to control it through the us e of Standard Commands for Programmable


16 Instruments, (SCPI). T he PS2520G was provided a 12 V output voltage that was used to open a vacuum valve. Manipu lation of the vacuum valve enabled a vacuum pump to create enough suction for the pad under test to be grabbed and held in place while measurements were being performed


17 Chapter 3 Temperature and Humidity Control 3.0 Temperature Control Temperature plays a major role in the determination of the mechanical properties of CMP pads. A temperature controller was used in order to observe the effects of temperatur e on the CMP pads. Data was acquired while the pads were subjected to different levels of temperature. The temperature controller accepts an input from a temperature sensor and has an output that is connected to a cont rol element such as a heater or fan. There are various types of control. Some controllers are dedicated to a specific type of control while others offer a programmable type of control. 3.1 Types of Control 3.1.1 On/Off Control An on-off controller is the simplest fo rm of temperature control device. The output from the device is either on or off, with no middle state. An on-off


18 controller will switch the out put only when the temperatur e crosses a set point. For heating control, the output is on when the temperature is below the set point, and off above the set point. Since the temperature crosses the set point to change the output state, the process temper ature will cycle within a small band of temperatures that contains the set poin t. In cases when the temperature cycles rapidly an on-off differential or “hysteresis” is added to the controller, which will prevent damage to contactors and valves. The differential ensures that the temperature exceeds the set point by a certain amount before reversing direction. The on-off diffe rential prevents the output fr om “chattering” or making fast and continual switches, which occurs if the temperatur e cycles above and below the set point very rapidly. On-o ff control is usually used where precise control is not required, in systems that cannot handle having the energy turned on and off frequently, where the mass of the system is so great that temperatures change extremely slowly or for temperature alarms. A special type of on-off that is used for temperature alarms is a limit controller. The limit controller uses a latching relay, which must be manually reset. The latching relay is used to shut down a process when a certai n temperature is reached. 3.1.2 Proportional Control Proportional controls are designed to eliminate the cycling associated with on-off control. A proporti onal controller decreases the average power supplied to


19 the heater as the temperatur e approaches the set point. A proportional controller gradually reduces the energy delivered to the heater to keep the temperature from overshooting the set point. Instead of overshooting, t he temperature will reach and maintain a stabl e temperature at the se t point. The proportioning action can be accomplished by turning the output on and off for short intervals. This “time proportioning” varies the ratio of “on” time to “off” time in order to control the temperature. The proportioning action o ccurs within a “proportional band” around the set-point te mperature. Outside of the proportional band, the controller functions as an on-off unit wit h the output either fully on if the temperature is below the band or fully off if the temperature is above the band. However, within the band, the output is turned on and off as a ratio of the temperature value to the va lue of the set point. At the set point, which is the midpoint of the proportional band, the output on-off ratio is 1:1. If the temperature is further from the set point, the onand off-times vary in proportion to the temperature difference. 3.1.3 PID Control The third controller type provides propor tional with integral and derivative functions, which is termed PID control. This controller combines proportional control with two additional adjustments, which helps the unit automatically compensate for temperature changes in t he system. These adjustments, integral and derivative, are expressed in timebased units. The adj ustments are also


20 referred to by their reciprocals, which are termed RESET and RATE, respectively. The proportional, integral and derivative terms must be individually adjusted or “tuned” to a parti cular system through the use of trial and error. PID control provides the most accurate and stable control of the three controller types. It is most appropriately employed in systems that have a relatively small mass or those that react quickly to changes in the energy added to the process. PID control is recommended in systems where the load changes often and the controller is expected to compensate aut omatically due to frequent changes in the set point, the amount of energy ava ilable or the mass to be controlled. 3.2 Temperature Sensor Many varieties of temperature sensor s exist. The thermo couple was best suited to satisfy the requirements of the system used in this research. A thermocouple is a sensor for measuring tem perature. It consists of two dissimilar metals, which are joined together at one end. When the j unction of the two metals is heated or cooled a voltage is pr oduced that is directly correlated with the temperature. Thermoc ouples are available in di fferent combinations of metals and calibrations. The four mo st common thermocouple calibrations are termed J, K, T and E. Ea ch calibration has a diffe rent temperature range and accuracy. Although the thermocouple calib ration dictates the temperature range, the maximum range is also limited by the diameter of the thermocouple wire. The K-type thermocouple was utilized dur ing this research; see Figure 3.1.


21 Figure 3.1: K-Type Thermocouple 3.3 Solid State Relay (SSR) A solid-state relay, (SSR), is a switch that contains no moving parts. Solid State Relays are used to switch various loads such as heating elements or resistive loads, motors, transformers or in ductive loads and capa citive loads. A solid-state relay is often used when a li ne powered device needs to be turned on and off by a control circuit in order to prov ide isolation from the power line. Solidstate relays are SPST; normally open, swit ching devices with no moving parts, which are capable of millions of cycles of operation. By applying a control signal, an SSR switches “ON” the ac load current, which is an action similar to that of moving the contacts of a mechanical contacto r. The SSR utilized in this research is pictured in Figure 3.2.


22 Figure 3.2: Solid State Relay (SSR) 3.4 Temperature Controller The I-Series CNi16D temperature cont roller was utilized to control the temperature during the experiments. The I-Series CNi16D temperature controller provided an RS-232 interface, which enabled remote control by the use of commands, which were specified in its configuration manual and generated by a computer program. The instrument posse ssed both analog inputs and outputs. A thermocouple was connected to the anal og input in order to sense the temperature. The RS-232 wiring assignments for attachment of the thermocouple to the controller are presented in Figure 3.3.


23 INPUT RS232/485 1 2 3 4 5 6 7 RTN RX TX + THERMOCOUPLE Figure 3.3: RS232 Wiring for Hook up of the Thermocouple to the Temperature Controller The analog output operated a SSR that supplied power to the heater. The details of the hookup betw een the controller, SSR a nd heater are pr esented in Figure 3.4. LOAD SIDE HEATER Vac FAST BLOW FUSE Figure 3.4: AC Controlled SSR fo r a Temperature Controller With a Mechanical Relay Output The details related to t he application of power and h ookup of the SSR to the controller is presented in Figure 3.5. TEMPERATURE CONTROLLER AC INPUT SSR


24 POWER OUTPUT1 OUTPUT2 L (+) N (-) NO C NO C SSR Figure 3.5: Power and SSR Wiring for the Temperature Controller Depending upon the current temper ature, once the set point was specified, the instrument enabled/disabl ed a relay. A tem perature dead-band could be specified within which the controll er maintained the state of the relay. The optional analog output could be programmed within a range of 0-10 Vdc or 020 mA. The output was selectable as eit her a control output or as a calibrated process value for retransmission, which is a unique feature amo ng controllers. The type of control was also selectabl e. On/Off, Proportional-Integral, (PI), Proportional-Derivative, (PD) or Proportional-Integral-D erivative, (PID), can be selected. The I-Series CNi16D Temperature Controller is pictured in Figure 3.6. Figure 3.6: I-Series CNi16D Temperature Controller


25 3.5 Humidity Control Humidity was also found to be one of the factors that affected the mechanical properties of the pads. A chamber was designed for humidity control. The humidity control was acti vated whenever the hum idity changed from ambient conditions. The humidity contro l consisted of both humidification and dehumidification apparatus, which activate d when the moisture level deviated from the adjustable set point. 3.5.1 Humidification Apparatus Set-Up The humidification system consisted of a fan assembly with an absorbent wick inside the chamber and a water rese rvoir with a re-circulating pump outside the chamber. The computer controll ed the pump automatica lly through the DAQ card. A small, adjustable re-circulati ng fan was added to t he chamber to ensure the atmosphere was properly and complete ly mixed. The humidifier and fan assembly were connected. After atta ching the fan electrical connections and water supply tubes the assembly was di rected toward the fr ont center of the chamber. Water inlet and drain tubes were connected to the appropriate fittings of the reservoir and re-circula ting pump outside of the chamber. The reservoir was only filled with distill ed water to prevent the build up of contaminants and discoloration. The pum p should be placed as close to the reservoir as possible for optimum operation. The re-circulating fan was normally


26 placed in a rear corner, turned on high an d aimed at about 45 so that it would direct air toward the cent er front of the chamber. Operation Once the setup was complete and t he chamber well or ganized the power supply, humidification and dehumidification apparatus were connected to a relay board that was connected to the DAQ card. The DAQ card supplied the humidification or dehumidification apparat us proper activation signals. The pump, in the humidification ap paratus, activates and fills the small reservoir in the fan assembly during humidific ation. The wicking material soaks up the water, the fan blows air through the wate r and the re-circulating fan activates. This action rapidly increases the humid ity of the whole chamber. During the humidification process, the pump was capable of supplying more water than the wick could retain. The excess water flowed from the fan reservoir through the drain tube and back to the pump reservoir. It was important to keep the drain tube as straight and clear as possible so that the return flow would not become blocked. Improper drainage would eventually cause the fan reservoir to overflow onto the chamber floor. The humidific ation apparatus is illustrated in Figure 3.7.




28 3.5.2 De-Humidification Apparatus Set-Up The Dehumidification apparatus consis ted of a mounting bracket with two drying capsules mounted to the rear wa ll on the outside of the chamber. A vacuum pump was connected to the capsules on the rear wall via plastic tubing. Quick-connect tube fittings were designed to seal themselves when the tubing was detached. This design was required in order to insure that the chamber atmosphere would not be affected. Inside the chamber a small re-circula ting fan connected and placed in the left rear corner. The fan was directed at a 45 angle so that it would direct air to the front center of the chamber. The fan positioning produced a good mix of the atmosphere to ensure a uniform environmen t in the shortest possible time. Once the tubing was attached the computer turned on the system. Humidity control requirements were a func tion of the desired level, the beginning level, chamber size, airflow rate and t he chamber contents. Frequency of door activity and room humidity levels outsi de the chamber severely affected the systems ability to maintain the desired le vel of humidity insi de the chamber. The dehumidification apparatus is illustrated in Figure 3.8.




30 Humidity Sensor The HIH-36 10 monolithic Integrated Cir cuit, (IC), humidity sensor is designed specifically for high volume Or iginal Equipment Manufacturer, (OEM), users. Direct input to a controller or ot her device is facilitated by the sensor's linear voltage output. The HIH-3610 requires a current of only 200 A, which makes it ideally suited for low drain battery powered systems. The sensor requires a supply voltage of 4.0 to 5.8 V and operates over a temperature range of -40 F to 185 F. The HIH-3610-001 Humidity Sens or is pictured in Figure 3.9. Figure 3.9: HIH-3610-001 Humidity Sensor The Relative Humidity, (RH), is deri ved from the mathematical relation RH = (((Voltage/5.0)-0.016)/0.0062) 3.1 where the parameter “Voltage” is the voltage val ue obtained from the output of the sensor. The voltage output of the sensor was sensed by the analog input of the DAQ card. 3.6 Integration of the Humidity Apparatus The 6035E Data Acquisition Card, from National Instrum ents, features sixteen channels for 16-bit analog inputs two channels for 12-bit analog outputs, a 68-pin connector and eight lines for digital Input/Output.


31 The DAQ card was primarily used by the system to implement humidity control. A humidity sensor generated a signal, which wa s mathematically related to the actual humidity; see equation 3.1. The signal was sensed by the analog input of the DAQ. The signal was digitize d and the digital value was sent to the program, which mathematically converted it to the actual humidity value. Based upon a calculation of whet her the humidity level wa s to be increased or decreased, the program instructed the DAQ to output signals to the appropriate relays in order to enable the requir ed system. The DAQ functioned in conjunction with the Nationa l Instruments SC-2050 I/0 Board and the SC-2062 Relay Board.


32 Chapter 4 System Integration 4.0 System Integration Tasks System Integration for this research involved the tasks of integrating all the hardware and instruments together and bringing them under the control of a single software program. The tasks involved with this research for which the use of hardware became necessary were: Controlling the position of the sensor above the pad Controlling the pad position relative to the sensor Generating an excitation signal for the pad Measuring the response produced by the pad Controlling and tracking the humidity and temperature levels during the experiment. 4.1 Instruments Employed This sub-section provides informati on about the hardware employed by the system. Table 4.1 provides basic information on all devices utilized during the testing. The instruments were c onnected to the computer via the like IEEE


33 GPIB488 and RS232C interfaces. A VC++ program was written to control the equipment from a console. Table 4.1: Instruments In tegrated Into the System Instrument Manufacturer Function VP 9000 Motor Controller Velmex, Inc. Control of motors used to move sensors PS2520G Programmable Power Supply Tektronix Provided power to hold the pad while measurements were being carried out Lock-In Amplifier 7260 EG&G (AMETEK) Generated excitation signals. Measured responses Data Acquisition Card 6035E National Instruments Control of switches used for tracking humidity Temperature Controller and Sensor Newport Instruments Temperature control and measurement SERVO – 260 Studio Amplifier Samson Amplify Oscillator Output Humidity Sensor Honeywell Measure humidity


34 A model of the setup employed to ca rry out experiments on the pads is presented in Figure 4.1. In addition to the pad shape depicted in the Figure 4.1, pads can be semi-circular as well as circular without the gap in the center. UST Probe CMP Pad Transducer Radial Scan Rotary Scan Figure 4.1: CMP Pad Test Set-Up The assembly presented in Figure 4.1 was mounted on a finely polished aluminum table. The supports for the t able could be adjusted so that the pad was always perfectly horizontal. Computer Controlled Electronics


35 Temperature and the hum idity control were required in the vicinity of the pads in order to study the effects of these physical factors on the pads. Therefore, the test setup depicted in Figure 4.1 wa s enclosed inside a wooden insulating chamber. The tem perature and humidity control setup is illustrated in Figure 4.2. DAQ Thermocouple Figure 4.2: Temperatur e and Humidity Control Temp Controller Humidity Monitor Environmental Chamber Dehumidifier Heater Element Fan Water


36 4.2 Hardware Integration Device manufacturers outfit thei r instruments with communication capabilities in order to ena ble operators to control the instruments remotely. Ethernet, USB, Serial and Parallel interfaces are ex amples of communication connection capabilities that ar e routinely provided. M any instruments that use the GPIB IEEE-488 interface possess the capability of decoding the SCPI instruction set. Some instruments have very specific command formats, which are specified by their manufacturers. In order to control such devices from a host computer a communication link must first be established. Once the communication link is est ablished the software is simply required to send instructions to the instrument in the fo rm of bit strings. Manufacturers provide libraries that include commands specifica lly designed to make the instrument perform specific tasks. These library file s were linked into the host computer’s program, developed for this research, so that the commands offered could be used freely anywhere within the program. 4.3 Steps Required for Taking Measurements The following steps were required in order to insure accurate measurements with respect to the functi ons of the individual instruments. 1. The position of the co ntact point of the sensor on the pad was located.


37 2. The sensor was moved to the point where the reading was to be taken. The Velmex 9000 moved the horizontal motor and vertical motor to establish the sensor over the contact point. 3. The vacuum pump was activated using the Tektronix Power Supply. 4. When the pad was firml y held, the sensor wa s pushed down to the desired depth in the pad and the pad was stimulated. 5. The amplitude of the response was measured using the Lock-In Amplifier. 6. Steps 1 through 5 were repeated for other points. Measurements were taken only when th e temperature and humidity were at the required level. The temperatur e and the humidity sensors notified the VC++ program of the current values of these parameters. Depending on the required value, the program activated the humidifier/ dehumidifier using the DAQ and the heater using the Te mperature Controller.


38 Chapter 5 Conclusions and Results 5.0 Conclusions The main purpose of designing and build ing this software and hardware system was to characterize the CMP pad and study the response to varying physical conditions. The re sults can be used to find ways of making the pads more resistant to such changes, which will ensure a longer worki ng life. Pads of various dimensions and geometry were tested using the system. There are various types of scans that can be used to collect statistical data from the pads. This particular system was designed to enable operators to perform various scans such as random, linear, full and sector-wise scans. The research dealt with sector scans. The circular pad under study had a diam eter of thirty-two centimeters and a thickness of four millimeters. For t he purpose of analysis the pad was divided into three 120 sectors. A zero position was marked on the pad as a reference point for the radial motor. M easurements were taken within a 60 sector within each of the 120 sectors. The three measur ement sectors ranged from 0 -60


39 120 -180 and 240 -300 These three sectors were identified as Sectors 0, 1 and 2 respectively. There were four parameter s associated with every me asurement sequence. The values of these parameters determined the number of obser vations and the density of points in every sector The four parameters were: Position: Each sector on the pad wa s identified by an integer. The position value specified the sector over which the measurements were made. A value of “1” specifi ed sector 1, which ranged from 120 -240 Radial Increment: This value specified the separation between consecutive points on a line within the sector. A value of “8” would leave eight millimeters bet ween consecutive points. Angular Increment: Similar to the R adial Increment; Angular Increment was the angular separation between two neighboring lines within a sector. Depth: This parameter specified t he depth to which the sensor was to be pushed into the pad a desired response point. The data obtained from each of the r uns were stored in text files. 5.1 Effects of Ambient Conditions The pad response to changing conditi ons of temperature and humidity was the focus of this research. Th e parameters and their test ranges are presented in Table 5.1


40 Table 5.1: Test Parameters Position Radial Increment (millimeters) Angular Increment (degrees) Temperature ( Fahrenheit) Humidity (%) 0 18 12 80, 90, 100 50, 60, 70 1 18 12 80, 90, 100 50, 60, 70 2 18 12 80, 90, 100 50, 60, 70 Experiments using the parameter values in Table 5.1 yielded nine sets of readings for each sector. The locations of test sectors for a pad are depicted in Figure 5.1. Figure 5.1: 60 Test Sectors for a Pad In order to check for the effects of temperature and humidity on the pad’s response a Multifactor Analysis F-test was used. The results from this test are presented in Table 5.2, T able 5.3, and Table 5.4.


41 Table 5.2: Test Values for Sector 0 (0 -60 ) Variable F-Value P-Value Temperature 415.23 0.000 Humidity 94.13 0.000 Location 3.02 0.062 Temperature and Humidity 31.81 0.000 Table 5.3: Test Values for Sector1 (120 -180 ) Variable F-Value P-value Temperature 410.48 0.000 Humidity 9.33 0.000 Location 2.73 0.021 Temperature and Humidity 4.51 0.002


42 Table 5.4: Test Values for Sector2 (240 -300 ) Variable F-Value P-value Temperature 307.58 0.000 Humidity 19.99 0.000 Location 1.34 0.247 Temperature and Humidity 3.97 0.004 The p-values in Tables 5.2, 5.3 and 5.4 represent probability. The p-value tells whether the F-value is signif icant or not. A value of 0.05 was assumed, which is termed type1 error and denoted by alpha, ( ). If the p-va lue is less than value then the corresponding factor effected the pad characterist ics. If the p-value is greater than 0.05 then the corresponding factor had no effect on the pad characteristics. The results presented in Table 5.5 demonstrate that Temperature and Humidity affected the pad’s response in ever y sector. The third column indicates that these two factors, acting together also have an effect on the pad. Therefore, if these tw o factors change together the pad’s response would be different than the response obtained when one factor remains constant and the value of the other factor allowed to change.


43 Table 5.5: Multifactor Anal ysis Results with PID Control Factor F-test Significance Sector 0 F-test Significance Sector 1 F-test Significance Sector 2 Temperature Yes Yes Yes Humidity Yes Yes Yes Temperature*Humidity Yes Yes Yes A system was developed to characteri ze CMP pads using an ultrasound technique. Stepwise hardware setup and integration were developed for the UST. The effects of physical factors su ch as temperature and humidity on the pad response were obtained and presented. 5.2 Future Work This research demonstrat ed that the data acquired from this system could be used to analyze the behavio r of the pads under changi ng ambient conditions. This analysis is very important since it helps in studying and improving pad properties, which will ultimately lead to better yields in IC manufacturing. This system was designed to enable re searchers to test pads using ultrasound testing. This syst em should be modified to run tests on pads using laser beams. In laser testing the respons e of the pad at a point is measured in


44 terms of the extent to which a laser-beam aimed at a point is scattered. The modular nature of the software makes it very easy to integrate this type of testing into the system. All that would be required is the incl usion of the device drivers for the laser interferometer and the develop ment of a new class that controls the laser. Coding for the laser experiments should follow the same philosophy as that for ultrasound testing.


45 References [1] Franklin A. Diaz, “Test Automa tion and Design of Experiments for Microelectronics CMP Pads”, Masters Thes is, University of South Florida, December 2000 [2] [3] [4] VP 9000 Motor Controller, Co mmand Manual, VELMEX, INC. [5] Tektronix PS-2520G Programmer Manual [6] Model 7260 Lock-In amplifier instru ction manual, EG&G in struments, Inc., 1997 [7] SC-205X Series User Manual [8] SC-206X Series User Manual [9] OMEGA, iSeries Controller Manual [10] HIH-3610-001 Data Sheet, [11] Prashant Datar, “Syst em Integration and Testi ng using Object Oriented Programming Based Control”, Masters Thes is, University of South Florida, May 2003 [12] J.M. Steigerwald, S.P. Murarka and R.J. Gutmann, “Chemical Mechanical Planarization of Microele ctronic Materials”, WileyInterscience, New York, 1997 [13] D. G. Totzke,, NIST conference, 2001, pp. 259 [14] A.K. Sikder, et. al ., Materials Research Society Symposium Proceedings, Vol 671, pp M1.8.1, 2001


46 Appendices


47 Appendix A Temperature and Humidity Data Sector one data is presented: Data for te mperatures of 80, 90 and 100 degrees and humidity of 50, 60 and 70. The rows indicate the readings on one line. Temperature-80; humidity-50 8.836000 8.786000 8.032000 8.235000 8.033000 8.279000 8.263000 8.633000 8.262000 8.226000 8.678000 8.620000 8.696000 8.615000 8.652000 8.698000 7.923000 8.685000 7.687000 8.356000 8.635000 8.256000 8.096000 8.680000 8.230000 8.632000 8.259000 8.082000 8.212000 8.236000 Temperature-90; Humidity-50 7.237000 6.832000 6.802000 6.823000 6.768000 6.839000 6.302000 7.537000 6.833000 6.806000 7.236000 6.902000 7.635000 6.989000 6.863000 6.899000 6.691000 6.905000 7.182000 6.836000 6.930000 6.860000 6.903000 6.730000 6.939000 6.856000 6.798000 6.876000 6.936000 6.933000 Temperature-100; Humidity-50 6.560000 6.136000 6.767000 6.736000 6.617000 6.727000 6.678000 6.768000 6.835000 6.682000 6.735000 6.761000 6.791000 6.802000 6.703000 6.860000 6.816000 6.855000 6.873000 6.878000 6.860000 6.965000 6.957000 6.838000 6.976000 7.003000 7.067000 6.993000 6.188000 6.973000 Temperature-80; Humidity-60 7.203000 8.268000 8.620000 8.663000 8.095000 8.336000 8.360000 8.669000 8.306000 8.275000 8.683000 8.263000 8.633000 8.650000 8.090000 8.256000 8.330000 8.306000 8.382000 8.239000 8.386000 8.363000 8.363000 8.260000 8.386000 8.368000 8.365000 8.366000 8.368000 8.320000


48 Appendix A (Continued) Temperature-90; Humidity-60 7.893000 7.367000 7.661000 7.923000 7.793000 7.959000 7.788000 7.879000 7.986000 7.832000 7.915000 7.908000 7.823000 7.877000 7.832000 7.933000 7.973000 7.965000 7.959000 7.922000 7.980000 7.951000 7.068000 7.939000 7.933000 7.028000 7.633000 7.936000 7.905000 7.970000 Temperature-100; Humidity-60 7.161000 7.681000 7.032000 7.611000 7.033000 7.279000 7.263000 7.633000 7.272000 7.221000 7.678000 7.620000 7.691000 7.617000 7.652000 7.698000 7.223000 7.685000 7.687000 7.351000 7.635000 7.256000 7.091000 7.689000 7.230000 7.632000 7.259000 7.082000 7.222000 7.231000 Temperature-80; Humidity-70 7.287000 7.288000 7.693000 7.200000 7.630000 8.376000 7.330000 8.236000 7.355000 8.220000 7.232000 7.272000 8.222000 7.237000 8.672000 8.272000 7.352000 8.362000 8.389000 8.280000 8.339000 7.335000 7.360000 8.298000 8.350000 8.393000 7.357000 7.383000 8.370000 8.357000 Temperature-90 Humidity-70 7.630000 7.098000 7.083000 7.233000 7.600000 7.266000 7.627000 7.260000 7.369000 7.055000 7.263000 7.225000 7.665000 7.638000 7.027000 7.230000 7.362000 7.268000 7.277000 7.309000 7.295000 7.369000 7.360000 7.237000 7.353000 7.303000 7.507000 7.336000 7.296000 7.338000


49 Appendix A (Continued) Temperature-100; Humidity-70 6.869000 6.807000 6.796000 6.902000 6.863000 6.906000 6.792000 6.876000 6.996000 6.822000 6.900000 6.922000 6.862000 6.882000 6.829000 6.957000 7.001000 6.973000 6.931000 6.964000 7.019000 7.063000 7.081000 6.983000 7.050000 7.099000 7.191000 7.001000 6.965000 7.632000