IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 45, NO.

4, AUGUST 1996

Instrumentation for Faint Microwave Line Emission Measurement

Shuming T. Wang, T. Koryu Ishii, Life Senior Member, IEEE, and Thomas C. Ehlert
Absitruct-It is difficult to detect faint microwave line emission from a low-temperature source. In this paper, cost-effective methods of detecting these emissions are presented. A Dicketype receiver (DTR) with an analog lock-in amplifier has been constructed to detect 22 GHz spontaneous emission from water vapor. The receiver compares the brightness temperatures of two waveguide chambers, one containing water vapor and the other empty. The system is capable of detecting a brightness temperature difference as small as 0.2 K. The experimental results prove that the emission from water vapor is detected for the first time with water vapor at 18 mm Hg and room temperature. The observed brightness temperature difference is 0.41 K. A ,samplingdigital integration technique (SDIT) is also used to detect the emission from water vapor. The experimental results agree with the results obtained with the analog lock-in amplifier. The main advantage of SDIT over the DTR is the flexibility on the inl.egrationtime and the range. In both methods, cost-effective brass plates are used for the low-temperature backgrounds for the first time.

for the detection of microwave emission from a lowtemperature source. In this paper, as an example of a lowtemperature source, discrete but spontaneous microwave emissions from water vapor are investigated. To the authors knowledge, this is the first such attempt. In this study, a faint line emission must be extracted from the background and system noise. To do this, a special Dicke-type radiometer [ I ] with a new cold background together with an analog lockin amplifier or with a sampling digital integration technique (SDIT) is employed. In both methods, a low-emissivity and cost-effective material, a brass plate, at room temperature is used as a cold background for the first time instead of a refrigerated matched termination or open sky in a conventional approach [2], [ 3 ] . It is also the first time that a waveguide section which is short-circuited by a brass plate at one end is utilized as the reference for transmission measurements.. The detected line emission is specified as follows: 1) terrestrial: to distinguish the attempts from the accomplishments of radio astronomers who routinely detect such emissions from astronomical sources 141, [5];
Manuscript received February 23, 1993; revised July 5 , 1995. S. T. Wang was with Marquette University, Milwaukee, WI 53233 USA. He is currently with the Department of Electrical Engineering, Kaohsiung Polytechnic Institute, Kaohsiung, Taiwan, R.O.C. T. K , Ishii is with the Department of Electrical and Computer Engineering, Marquctte University, Milwaukee, WI 53233 USA. T. C. Ehlert was with the Department of Chemistry, Marquette University, Milwaukee, WI 53233 USA. He is now retired. Publisher Item Identifier S 0018-9456(96)02954-3.

I. INTRODUCTION HE objective of this paper is to present the techniques

microwave: to distinguish the attempts from the accomplishments of spectroscopists who routinely observe emissions in the infrared [6], [7], visible, and ultraviolet regions [8], [SI; spontaneous: to distinguish the attempts from the accomplishments of researchers who obtained discrete microwave emissions from laboratory sources through the application of electric fields [lo]; emission: to distinguish the experiments from microwave absorption spectroscopy, a well-developed methodology [ 111, [ 121; and discrete: to indicate that the observed emission is at certain frequencies rather than the continuum emitted by solids, liquids and dense gases with a continuum spectrum 1131, 1141. Using a locally situated metal plate cold background, the potential applications of this research may include remote detection of processes which produce high-temperature gases, e.g., jet engines, explosions or chimneys; determining the rotational properties of molecules which are difficult to study by absorpl ion spectroscopy; studies of intermolecular energy transfer priocesses; and studies of the role of molecular rotation in energy transfer to and from surfaces.

WITH

11. DICKE-TYPE RECEIVER (DTR) ANALOG LOCK-IN AMPLIFIER

A. The Sensitivity of DTR with Analog Lock-in Ampl$er In the detection of microwave signals using heterodyne techniques, the sensitivity of the receiving system is limited by the noise temperature and the bandwidth of the system [15]. To reduce the degradation in sensitivity due to the gain variation of a receiving system, a Dicke-type receiver (DTR) using a switching technique is introduced [I]. This DTR consists of a front-enld RF switch, a heterodyne receiver and an analog lock-in amplifier. The front-end switch switches alternatively between two input channels, one is the source channel and the other is the reference. The sensitivity of a DTR can be calculated approximately from [161

1) ( L I T ) is ~ the ~ sensitivity ~ of a DTR in kelvin. 2) TsyL is the noise temperature of the receiver in kelvin. 3) A U ~ I isF the bandwidth of predetection RF section in hertz.

0018-9456/96$05.00 0 1996 TEEE

114

IEEE TRANSACTIONS (IN INSTRUMENTATION AND MEASUREMENT, VOL. 45, NO. 4, AUGUST 1996

--Heterodyne Recerver
I

Data
Acquisn'on

Fig 1 A block diagram of a Dicke-type receiver for measuring emirrion from water vapor
TO SPDT Switch

Fig. 3. Channel switching.

Fig. 2. A block diagram of a waveguide channel.

4)

~ L F is

the time constant of the post detection lowpass section (s).

B. Experimental Setup of DTR with Analog Lock-In Ampl8er A block diagram of a specific DTR construction to measure the emission from water vapor at 22 GHz is shown in Fig. 1. The input to the receiver consists of two symmetrical channels, a source channel and a reference, followed by a singlepole double-throw (SPDT) RF switch, a low-noise amplifier (LNA), a Maxier and a local oscillator, an IF amplifier, a detector, a lock-in amplifier, and a data acquisition system. The SPDT switch and the multiplier in the lock-in amplifier are synchronously driven by a switch driver at 2.1 Hz. The LNA has a bandwidth of 400 MHz centered at 22.235 GHz , a power gain of 29 dB and a noise figure of 4.8 dB. The IF amplifier has a bandwidth of 1 MHz centered at 30 MHz and a noise figure of 1.5 dB. This IF amplifier and detector are included in one unit. The lock-in amplifier has a frequency range of 2 Hz to 100 kHz and a sensitivity of 1 pV. The output of the lock-in amplifier is collected by the data acquisition system. Each channel consists of an 8.8-m long XB-band waveguide (RGSlKJ) chamber, an XB-to-K band waveguide (RG53/U) adapter, a K-band waveguide E-H tuner, an SMA-to-K band waveguide adapter, and a short piece of semirigid coaxial cable as shown in Fig. 2. Each channel is terminated with a three-foot length of XBband waveguide short-circuited by a polished brass plate at the end of the waveguide. The other end of the channel is connected to the SPDT switch through sections of XB-to-K band waveguide adapter, a K-band E-H tuner, a K-band-toSMA adapter and a short section of semirigid cable as shown in Fig. 3. The source waveguide chamber is filled with water vapor, and the reference waveguide chamber is empty. Both waveguide chambers are gas tight using the UG52A/U choke joint with thin mylar sheet windows and rubber O-rings so that the pressures in the chambers can be controlled. Soft solder O-rings are used for flat waveguide joints of the gastight chambers. The waveguide chambers and the brass plates

are at room temperature. Since the brass plate has a very high microwave reflectivity, its emissivity is low. Therefore, it creates a low-brightness temperature background. From either the radiative transfer equation or the lossy transmission line model [17], the noise temperature of the source channel at the input of the receiver would be higher than that of the reference channel due to the emission from water vapor. It should be noted on the presence of E-H tuners in Figs. 2 and 3. By tuning these E-H tuners, it is possible to minimize the reflection from the receiver input which is a 50 R system. The VSWR is adjusted to less than 1.1 using a signal generator at 22 GHz and a reflection-free termination as explained in the next section.

C. Experimental Procedure of DTR with Analog Lock-in Amplifier

The step-by-step experimental procedure is as follows: 1) Set the local oscillator frequency at 22.37 GHz. 2) Evacuate both the source chamber and the reference chamber. 3) Set the E-H tuners of source and reference channels for minimum reflection with reflection-free termination [ 171. 4) Select the time constant and output smoothing filters of the lock-in amplifier for the longest practical integration for the highest practical sensitivity. 5) Null the output of a lock-in amplifier by adjusting the phase shifter on the lock-in amplifier. 6) Reduce the phase shift by 90" so that the output of the lock-in amplifier goes to the maximum. 7) Set the sensitivity of the lock-in amplifier so that the lock-in amplifier is operated in the linear region. 8) Increase the phase shift by 90" so that the output of the lock-in amplifier is back to zero. 9) Connect the output of a lock-in amplifier to the data acquisition system. Set the sampling rate at 2 samples/s and sample the output of the lock-in amplifier for 2 min. 10) Reduce the phase shift by 90" and sample the output of the lock-in amplifier for 5 min. 11) Send water vapor to the source chamber and adjust the pressure of the source chamber to 18 mm Hg.

WANG c z t al.: INSTRUMENTATJON FOR FAINT MICROWAVE LINE EMISSION MEASUREMENT

115

manufacturer's specification, has a power gain of 29 dB and a noise figure of 4.9 dB, the noise temperature of the receiver is dominated by this amplifier. Therefore, at room temperature, the system noise temperature of this DTR is calculated to be 896 K using the equations [15]

T, = 2 9 0 ( 8 ' ~ -~ 1) ~ = 290(10 49 - 1) = 606K

Legend T W 0 T d 2

T r U

n T & & T K a m ~ns~~stlm

Pnx, -H20 presSure

Tsys

= Tt, max

= 290

+ T, + 606
(3)

= 896K

500

1000

1500

2000

2500

Time (second)

Fig. 4. Lock-in amplifier output versus time/pressure.

12) Sample the lock-in amplifier output for 5 min. 13) Repeat step 12 for water vapor pressures of 14, 9, and 4 mm Hg, respectively. 14) Evacuate the source chamber and sample the output of the lock-in amplifier for 5 min. 15) Increase the phase shift 90" so that the output of the lock-in amplifier goes back to zero and sample the output for 3 min. 16) Repeat steps 9-15 four times.

D. Experimental Results of DTR with Analog Lock-in Ampl8er

The experimental results of lock-in amplifier output versus time ,Ire plotted in Fig. 4. In this figure, the first 120 s show that the lock-in amplifier is well balanced, and the sampled values are around zero volts. From time 120 s to 420 s, the phase shift of the lock-in amplifier is decreased by 90, and the output of the lock-in amplifier is no longer balanced. The sampled voltages are gradually rising to a certain value which is proportional to the original noise temperature difference between the source channel and the reference channel. From time 120 s to 720 s, the source chamber is filled with water vapor at 18 mm Hg. The magnitude of the output signal level increases because the emission from the water vapor increases the noise temperature of the source channel. From time 720 s to 1920 s, the magnitude of the output signal level decreases because of the decrease in water vapor pressure in the source chamber. From 1920 s to 2100 s, the phase shift is increased by 90, and the output of the lock-in amplifier is returned to zero. The same process is repeated four times. E. Remarks on DTR with Analog Lock-in Amplifier The DTR as designed is capable of detecting the noise temperature increment in the source channel due to the emission from the water vapor. To the authors' knowledge, this is the first time that 22 GHz water line emission has been detected by the use of the DTR. Since the LNA employed, according to the

where 1) T, iis the equivalent noise temperature at the input of the IITR in kelvin. 2) FLNA is the noise factor of the low noise amplifier. 3) Tsysis the noise temperature at the input of the DTR system in kelvin. 4) T,,max is the maximum available noise temperature to the receiver input under operating ambient temperature in kelvin. The IF amplifier has a bandwidth of 1 MHz, and the lock-in amplifier time constant measured is 72 s [ 171. Therefore, using (I), the sensitivity of the employed DTR is approximately 0.2 K. As will be shown in Section IV, the lock-in amplifier output, 5 V at 4 mm Hg of water vapor pressure, is equivalent to the noise temperature difference between two channels of 0.32 K as seen in Figs. 4 and 7.

111. SAMPLING DIGITAL INTEGRATION TECHNIQUE (SDIT)

A. Principles of SDIT

If the input signal level is below the sensitivity of a heterodyne receiver, this signal is masked by the system noise. In order to detect this faint signal, two signal-processing steps can be performed: 1 ) balance the detected voltage due to the system noise, and 2) remove unwanted frequency components. For the DTR presented earlier, these two signalprocessing steps use an analog approach. The system noise in the signal channel is balanced by the system noise in the reference channel through the synchronously driven frontend RF switch and the multiplier in the lock-in amplifier. The unwanted frequency components are removed by the low-pass RC filter section in the lock-in amplifier. These two signal-processing steps can also be done by the digital approach. To do this, the detected video signals from the source channel and the reference channel are sampled by a data acquisition system over a certain period of time, respectively. The unwanted frequency components can be removed by summing or averaging the sampled values from each channel, and the system noise can be balanced by subtracting the sum or average of sampled values of the reference channel from

776

IEEE TRANSACTIONS ON INSTRUMENTATIONAND MEASUREMENT, VOL. 45,NO. 4, AUGUST 1996

Legend

Fig. 5. Experimental setup for the measurement of emission from water

vapor by the SDIT technique.
0
Trial 1 Trial2 Trial 3

the sum or average of sampled values of the source channel. This technique is used in radio astronomy [20]. A block diagram of the experimental setup to measure the emission from water vapor using a DTR with the sampling digital integration technique is shown in Fig. 5 . This setup is similar to the one shown in Fig. 1, except that the lock-in amplifier is removed. This experimental conditions are the same as in Section 11. That is, the local oscillator (LO) frequency is 22.37 GHz, and E-H tuners are tuned to make the reflection coefficient of each channel a minimum with reflection-free termination. The data acquisition system, consisting of an A/D board, an IBM PC 386 and software, samples the switch driver signal and the detected video signal from the Sperry 61A1 at the same time. Since the switch driver signal gives information about which channel the receiver is sampling, the sum or average of the sampled values from the source channel and the reference channels can be extracted separately. The experimental procedure of the DTR with SDIT is as follows: 1) Evacuate the source and the reference chamber. 2) Set the sampling rate at 210 samples/s. 3) Sample the switch driver signal for 10 s and the detected video signal from the receiver for 100 s. 4) Send the water vapor to the source chamber and adjust the pressure of the source chamber to 18 mm Hg and repeat step 3. 5 ) Repeat step 3 with the pressures of the source chamber at 14, 9, and 4 mm Hg, respectively. 6) Repeat steps 1-4 four times.

0
A

Trial4 Trial5

10

15

20

Source Chamber Water Vapor Pressure (mmHg)

Fig. 6. (Sum of sampled values from the source channel)-(sum of sampled values from the reference channel) versus source chamber water vapor pressure.

obtains 100 samples, of which 50 samples are from the source channel and 50 samples are from the reference channel. Since there exists a transition time while the switch changes from one position to another, only the 30 samples in the middle of each 50 samples are taken.

C. Remarks on DTR with SDIT The DTR with the sampling digital integration technique has some advantages over the Dicke-type receiver with an analog lock-in amplifier: 1) Time Constant Control: The sensitivity of the DTR with analog lock-in amplifier is determined by a fixed time constant setting of the system. However, with the SDIT technique, it depends on the duration of the sampling desired and is controlled by the software. 2 ) Switch Transition Time: With the SDIT technique, it is easy to discard the data taken during the transition period. 3) Post Signal Processing: Since the sampled data is taken directly from the detected video output, it can be used for any other signal processing. With both techniques, switching between the source and reference channels rapidly assures that variations in the system gains at the receiver during the observation will not affect significantly the results of the experiment.

B. Experimental results of DTR with SDIT The experimental result which shows a relationship between the difference of the sum of converted digital sampled values from the source channel and the sum of converted digital sampled values from the reference channel and source chamber water vapor pressure is plotted in Fig. 6 where each data point is the difference of the sum of 4500 samples. The sampling rate of the data acquisition system is chosen at 210 Hz which is 100 times the switch driver's frequency. Therefore, for each cycle of the switch driver signal, the data acquisition system

I v . CORRELATION

BETWEENTHEORY AND EXPERIMENT

The noise temperatures of the source and the reference channels to the input of the receiver can be calculated using

WANG et al.: INSTRUMENTATION FOR FAINT MICROWAVE LINE EMISSION MEASUREMENT

I17

the equations [17]

100 -

0.63

90 -

0.567

between the source and reference channels is due to the emission From water vapor at 22.34 and 22.40 GHz. Using (4)-(7), the relations between theoretical values of the noise temperature difference of the two channels and the water vapor pressure of the source chamber are plotted in Fig. 7. am(.) = (3.24 x 10p4e-"4/T) For comparison, the averaged experimental data from Figs. 4 . (1 0.0147 P and 6 are plotted in Fig. 7. In Fig. 7, the vertical axis on the left is obtained by normalizing the vertical axis reading at 1 (cm--l) 18 mm Hg to 100, and the vertical axis on the right is the noise (v0 - V ) ~ ( A U(v0 ) ~ U)' (Ai/)2 temperature difference of the two channels calculated using AU (4)-(7). Fig. 7 shows also that the experimental data obtained f2.55 x 10-sp,2 T3/2 from the employed DTR with analog rock-in amplifier and from the receiver using the SDIT technique agree with each Ai/ = 2.58 x other very well. The calculated data and the experimental data shop the same tendency with respect to the water vapor P pressure of the source chamber. X 1 0 . 6 2 5 GHz The disagreement between the experimental and theoretical calculations based on (4)-(7) can be explained by the (Zi) following: where 1) The theoretically calculated noise temperature difference is based on the experimental results of reflectivity of the 1) U is the operating frequency, (GHz). brass plate and the attenuation constant of the empty 2) a,(v) are the absorption coefficients at frequency v, channel. (cm-l). 2) Equations (4) and ( 5 ) are approximations based on the 3) p,, is the water vapor density, (g/m3). survey fitting of the experimental results. 4) P is the total pressure, (millibar). 5 ) T is the water vapor temperature, (K). From Fig. 7, the calculated noise temperature differences 6) A Y is the half-line width, (GHz). between the source and the reference channels with water 7) I/, is the resonant frequency of water vapor which is vapor prtssures at 18 and 14 mm Hg are 0.63 and 0.59 22.235 GHz. K, respectively. Therefore, both the DTR with analog lockSince the LO is operated at 22.37 GHz and the IF amplifier in amplifier and the receiver using the SDIT technique are is centered at 30 MHz, the noise temperature difference capable of recognizing a difference in the noise temperatures

where 1) T, and T , are the noise temperatures of the source and reference channels at the receiver input, respectively in kelvin. 2) T is the physical temperature of the source and reference channels in kelvin. 3) ps and p, are the reflectivities of the source channel brass plate short circuit and the reference channel brass plate short circuit, respectively. 4) a,, and aTeare the power absorption coefficients of the source and reference channels with evacuated chambers, respectively, (m-'), a, is the power absorption coefficient of the water vapor, (m-I). 5 ) d is the physical length of the source and reference chambers (m). In ithe experiments, the power absorption coefficients of a s , and otre are assumed to be constant over the frequency range of 22.34-22.40 GHz and are measured to be 1.546E-3 and 1.543E-3 cm-' at 22.37 GHz, respectively [17]. The power reflection coefficients of ps and p, are measured to be 0.899 and 0.902 at 22.37 GHz, respectively [17]. The water vapor absorption coefficient at 22 GHz can be calculated from the equations [ 181

80 -

0.504

a
E
2

!
0.441

3 70R

E
c

3 ! 603
Legend

0.378

?2

E c
.I
0.315

40

Theory 0 DTR 0 SDIT

0.252

0.169
4

15 Source Chamber Water Vapor Pressure (mmHg)

10

Fig. 7. Theoretical noise temperature difference between the two channels and experimizntal results versus water vapor pressure of source chamber.

(a)
+

778

IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 45, NO. 4, AUGUST 1996

of the source and reference channels of 0.04 K conservatively. This means that the receivers are capable of sensing a noise temperature difference channels at least 0.04 K. V. DISCUSSION In this research, a low-emissivity brass plate is introduced so that a cold background can be achieved at room temperature instead of using a refrigerated matched termination. This approach will create a mismatch at the receiver input. The available noise temperature at the receiver input of a mismatched input termination follows the relations of (4) and (5) [17]. Using the radiative transfer equation, the noise temperature at the receiver input of a waveguide chamber against a refrigerated matched termination is calculated to be ~ 9 1 T, = T s e P w d T(1(8)

was simply compared to the theoretical 290 K calculation. Other parameters are measured, and procedures to produce the temperature scale are explained in Section IV. VI. CONCLUSIONS The weak emission from ambient-temperature low-pressure water vapor at 22 GHz, has been successfully detected by a specially designed Dicke-type receiver using a brass plate as the cold background for the first time. The experimental data obtained from the DTR with an analog lock-in amplifier and with the SDIT technique match with each other very well. Both experimental methods agree reasonably well with the theory. It is noted that the techniques employed in this paper need not be limited to water vapor, nor to the 22 GHz, nor to waveguide-confined samples, nor to a particular pressure. It is also noted that a low-temperature reference can be produced using a room-temperature brass plate. This is a cost-effective way to obtain an equivalent low-temperature reference.

where

1) T, is the noise temperature at the receiver input, (K). 2) TB is the physical temperature of the refrigerated matched termination, (K). 3) aw is the absorption coefficient of the waveguide chamber, (m-). 4) d is the physical length of waveguide chamber, (m). 5) T is the physical temperature of the waveguide chamber, (K). Comparing (8) with (4) and ( 5 ) , the mismatched input termination can be considered as a cold background with temperature at (I - p)T and the length of the waveguide chamber is doubled. Therefore, it is logical to use a lowemissivity material at room temperature to simulate a cold background as long as the reflection coefficients of p,. and p3 in (4) and (5) are accurately measured. Due to the nature of the mismatch, there exists a standing wave between the receiver input and the brass plate termination. The mismatch is created on purpose, and the effect of water vapor reduces the magnitude of the standing wave which results in the increasing of the available noise temperature at the receiver input as well as the detected voltage increment which can be explained as the effect of emission with water vapor. Regarding the standing wave created by the leakage of the local oscillator to the receiver input and the influence of the sensitivity of the receiver, as seen from Figs. 1 and 5, the local oscillator is injected after the LNA. Therefore, there is no chance of local oscillator power returning to the receiver input. The receiver output is influenced by the presence of the standing wave at the input as seen from (4) and (8). The use of the brass plate is much more cost effective than the use of the refrigerated matched termination. In a conventional radiometer, the data may be recorded automatically for prescribed integration periods after low-pass filtering [20]. In this SDIT technique the low-pass filtering is done by summing sampled data for a prescribed number of samples taken from the video output of the receiver. For calibration of the temperature scale in the Fig. 7, the reference temperature is the room temperature (290 K) of the brass plate which is used as the cold background. No other calibration was done except that the detected output

REFERENCES
R. H. Dicke, The measurement of thermal radiation at microwave frequencies, Rev. Sci. Instrum., vol. 17, pp. 268-275, 1946. T. Orhaug and W. Waltman, A switched load radiometer, Publ. Natl. Radio Astron., vol. 1, pp. 179-204, 1962. K. Fujimoto, On correlation radiometer technique, ZEEE Trans. Microwave Theory Tech., vol. MTT-12, p. 203, 1964. 0.E. H. Rydbeck and A. Hjalmarson, Radio observations of interstellar molecules, in Molecular Astrophysics, G. H. F. Diercksen, W. F. Huebner, and P. W. Langhoff, Ed. Boston, MA: Reidel, 1984, pp. 45- 175. T. L. Wilson and C. M. Walmsley, Sub-millimeter astronomy and astrophysics, in Molecular Astrophysics, G. H. F. Diercksen, W. F. Huebner, and P. W. Langhoff, Ed. Boston, MA: Reidel, 1984, pp. 177-199. J. H. Park and B. Carli, Analysis of far-infrared emission Fourier transform spectra, Appl. Opt., vol. 12, pp. 3419-3501, 1986. H. Oelfaf and H. Fischer, Measurements of minor constituents in the middle atmosphere from IR limb emission spectra: A feasibility study, Appl. Opr., vol. 22, pp. 2515-2518, 1983. J. W. Simons, B. A. Palmer, and D. E. Hof, Characterization of several ultraviolet-visible emission lines from a lead hollow-cathode lamp, Opt. Soc. America B, Opt. Phys. vol. 6, pp. 1097-1102, 1989. B. R. LaFreniere, R. S. Houk, and V. A. Fassel, Direct detection of vacuum ultraviolet radiation through an optical sampling orifice: Analytical figures of merit for the nonmetals, metalloids, and selected metals by inductively coupled plasma atomic emission spectrometry, Anal. Chem., vol. 59, pp. 2276-2282, 1987. S . C. Stinson, Fourier transform techniques enhance microwave spectroscopy, Chem. Eng. News, pp. 21-24, Aug. 1987. J. R. Eyre and H. M. Woolf, Transmittance of atmospheric gases in the microwave region: A fast model, Appl. Opt., vol. 27, pp. 3244-3249, 1988. E. R. Brown, Submillimeter wave absorption of n-type InSb at low temperatures, J. Appl. Phys., vol. 57, pp. 2361-2365, 1985. 0. Lillester, Simple radiometric method for measuring the thermal broadband emissivity of material samples, Appl. Opt., vol. 30, pp. 50865089, 1991. A. I. Baskakov, S. P. Gagarin, and A. A. Kalinkevitch, Simultaneous radiometric and radar altimetric measurements of sea microwave signatures, IEEE J. Oceanic Eng., vol. 9, pp. 325-328, 1984. R. Pettai, Noise in Receiving Systems. New York: Wiley, 1984. J. D. Kraus, Radio Astronomy. New York: McGraw-Hill, 1966. S. T. Wang, Detection of microwave radiation from a pseudo blackbody, Ph.D. dissertation on file at Memorial Library, Marquette University, Milwaukee, WI, Dec. 1992. D. Staelin, Microwave spectrum of the terrestrial atmosphere near 1-cm wavelength, J. Geophys. Res., vol. 71, pp. 2875-2881, 1966. M. L. Meeks, Methods of Experimental Physics. New York: Academic, 1976, vol. 12, pt. B. ~, Methods of Experimental Physics. New York: Academic, 1976, vol. 12, pt. C.

WANG et al.: INSTRUMENTATION FOR FAINT MICROWAVE LINE EMISSION MEASUREMENT

179

Shuming T. Wang was born in Taipei, Taiwan, on June 15, 1961. He received the B.S. degree in electrophysics from National Chiao Tung University, Hsinchu, Taiwan, R.O.C., in 1983, and the M.S. and Ph.D. degrees in electrical engineering in 1988 and 1992, respectively, from Marquette University, Milwaukee, WI. Since 1988, he has been a research assistant in the Department of Electrical Engineering and Computer Engineering, Marquette University. At present, he is an associate professor in the Department of Eleclrical Engineering, Kaohsiung Polytechnic Institute, Kaohsiung, Taiwan, R.O.C. His research areas include microwave circuit design and detection of faint microwave signals. Dr. Wang is a member of Sigma Xi.

Thomas C. Ehlert way born in Milwaukee, WI, in 1931. He received the B S ,M S , and Ph.D degrees
in chemistry in 1957, 1958, and 1963, respectively, from the University of Wisconsin, Madison Since 1964, he has been with Marquette University, Milwaukee, WI At present, he is a Professor of Chemistry and Adjunct Professor of Material Science. His research areas include chemistry related to water softening, thermodynamics, energy conversion, microwave emission spectroscopy, material science, and applications of differential scanning calorimetry. Dr. Ehlert is a member of the American Chemical Society, American Association For Advancement of Science, and Phi Lambda Upsilon.

T. Koryu Ishii (MS5-SM65-LS93) was horn in

Tokyo, Japan, on March 18, 1927. He received the B.S. degree in electrical engineering from 1% hon University, Tokyo, in 1950, and the M.S. and Ph.D. degrees in electrical engineering in 1957 and 1959, respectively. from the University of Wisconsin, Madison. He also received the Doctor of Engineering degree from Nihon University in 1961. Since 1959, he has been with Marquette University, Milwaukee, WI. At present, he is a Professor of Electrical Engineering. His research areas include millimeter waves and microwave ferrite devices; thermionic and solid-state devices; circuit components and transmission lines; application of microwaves; millimeter waves; and quantum electronics. He has published more than 300 research papers in the area of microwave and related electronics and has authored three books. He has five U.S. and foreign patents. Dr. Ishii is a member of Sigma Xi, Eta Kappa Nu, Sigma Phi Delta, Tau Beta Pi, ASEE, AAUP, PCM, WSPE, NSPE, and is a registered P.E. in the State of Wisconsin.