SuperSensitive Thermo and VaporSensing Elements and Their UsefulTemperature RangeHerbert A. Pohl Citation: Review of Scientific Instruments 23, 770 (1952); doi: 10.1063/1.1746170 View online: http://dx.doi.org/10.1063/1.1746170 View Table of Contents: http://scitation.aip.org/content/aip/journal/rsi/23/12?ver=pdfcov Published by the AIP Publishing Articles you may be interested in Thermo-mechanical characterization of glass at high temperature using the cylinder compression test.Part II: No-slip experiments, viscoelastic constants, and sensitivity J. Rheol. 57, 1391 (2013); 10.1122/1.4817435 Polymer-dispersed liquid crystal doped with carbon nanotubes for dimethyl methylphosphonate vapor-sensing application Appl. Phys. Lett. 102, 191912 (2013); 10.1063/1.4804297 AllFiber Modal Interferometer for Temperature Sensing with Negligible Strain Cross Sensitivity usingPMPCF AIP Conf. Proc. 1391, 344 (2011); 10.1063/1.3643544 Multi-layer aircraft structure inspection using super-sensitive Remote-Field Eddy-Current system AIP Conf. Proc. 557, 1906 (2001); 10.1063/1.1373985 Electronic Linearization of Temperature Using a Dual Element Sensing Technique Rev. Sci. Instrum. 43, 1161 (1972); 10.1063/1.1685866

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

128.248.155.225 On: Mon, 24 Nov 2014 02:30:42

770 LABORATORY AND SHOP NOTES

Table I also gives the Kallman and Furst gamma-efficiencies and those alpha-to-gamma efficiency ratios which could be computed from their data. These figures compare favorably with the neu­tron-to-gamma efficiency ratios reported here. It is seen that for neutron detection in a gamma-ray background the first four solu­tions are superior.

The only significant anomaly in the results is the poor agreement between our neutron efficiency and the Kallman and Furst gamma­efficiency for mesitylene. These authors attribute their low gamma­efficiency to possible contamination of the solution. It is difficult to evaluate the influence of contamination upon our results; however, three runs with independently prepared mesitylene solutions verified the relatively high neutron efficiency shown.

For better solutions we have also observed the differential pulse-height spectra (deriving from 2.55-Mev neutrons) with a sliding channel type pulse-height analyzer.3 Although resolution was inferior to that which was obtained with an anthracene crystal it was possible to observe a sharp cutoff. In a few instances a slight peak near cutoff was evident, which was due to plural scattering of the neutrons within the solutions. This peak was much more pronounced in the case of anthracene.

* Supported in part by U. S. Navy, Bureau of Ordnance Contract NOrd·7873.

1 Falk, Poss, and Yuan, Phys. Rev. 83, 176 (1951). 2 M. Furst and H. Kallman, Phys. Rev. 85. 816 (1952). References to

previous work by these authors appear here. a Francis, Bell, and Gundlach, Rev. Sci. lnstr. 22, 133 (1951).

Super-Sensitive Thermo- and Vapor-Sensing Elements and Their Useful Temperature

Range HERBERT A. POHL

Explosives Department, E. I. du Pont de Nemours and Company, Wilmington, Delaware

(Received June 25, 1952)

I N a previous publication,! thermo elements made from bi-films of polymers with selected molecular orientation were described.

They show a tenfold greater motion per degree change of tempera­ture than any known bi-metallic thermoelement. The following information is presented as an indication of the useful upper temp­erature range of the polymer bi-films.

The "second-order transition" temperatures (Table I) based on density studies give a measure of the upper temperature limit

TABLE I. "Second-order transition temperatures" for various orientable high polymers.

Material

Polythene (polyethylene) Saran (polyvinylidene chloride) Nylon (polyhexamethylene

adipamide or 66 Nylon) Polyethylene terephthalate

Polyacrylonitrile

Polytetrafiuoroethylene

"Heat distortion" temperature

TV.OC 46 to 50

65-82 74-76

(80-90) extrap.

(70-100) extrap.

130

I'Second .. order"

Tm.oC -90 to -70

-17 49-52

69-79

40-70

Nones

a Shows limited first-order change at 20°C. Complete first-order change at 327°C.

of bi-film operation if a very minimum of permanent distortion by heat or mechanical strain is desired. The heat distortion tempera­tures given are temperatures above which rather rapid distortion can be expected.

An interesting special case of thermal expansion can be created by selecting the degree of cold draw to which the orientable bi-film is subjected. At a "draw ratio" (the ratio of final to original length) of unity, for example, Nylon thread has a large positive coefficient of expansion along the thread (ca 150X 10-6 degrees-I). At a draw ratio of 5, it has a large negative coefficient of expansion of about

150X 10-6 degrees-I. At an intermediate draw ratio of about 2 to 3 which, incidentally, can be produced uniformly only with care,2 the coefficient of thermal expansion can be set close to zero. This property is useful, for example, in the case where a water-sensitive fiber such as Nylon in conjunction with a water-insensitive fiber such as Dacron* or Orlont is used as a temperature insensitive hygrometer element.

The bi-films studied have been made only on a limited experi­mental scale and are not being produced commercially in a finished form.

1 H. A. Pohl, Rev. Sci. lnstr. 22, 345 (1951). 2 Champetier and Bonnet, J. chim. phys. 40, 217 (1943). * Du Pont trademark for polyester fiber. t Du Pont trademark for acrylic fiber.

Stabilization of Photomultiplier Tubes R. SHERR AND J. B. GERHART

Palmer Physical Laboratory, Princeton University, Princeton, New Jersey (Received September 16, 1952)

By use of a circuit suggested by G. A. Morton,' two methods for stabilizing output pulses from electrostatically focused

photomultiplier tubes against supply voltage fluctuations have been investigated. Both are useful with such multipliers as the RCA 5819 and the RCA C7l51. The same stabilizing circuit is used in each method, the difference resulting from the point in the dynode system at which the circuit is introduced.

H.V. SUPPLY

1M 05

1M 04

1M 03

1M 02

2.2 M 01

DYNODE STRUCTURE RC A 5819

3.3 M PHOTOCATHODE

FIG. 1. Stabilizing circuit for interior dynode.

The first method,' illustrated in Fig. 1, consists of fixing the potential of an interior dynode of an RCA 5819 multiplier (e.g., dynode No. 7 in Fig. 1) with respect to the potential of the preceding dynode by means of a battery. The battery voltage V necessary for stabilization depends somewhat on the properties of the individual multiplier. Battery voltages of 90--135 volts were found to be satisfactory. A variable resistor R connects the two dynodes on either side of the interior dynode. The value of R required for stabilization can be set as follows. When R is varied for a convenient, fixed supply voltage (800 to 1100 volts) the output pulse height from the multiplier varies as shown in Fig. 2 (supply voltage 1100 volts, V = 90 volts). By observing the output pulses on an oscilloscope, R may easily be set at RA corresponding to the half-maximum point A. The output pulse is then unchanged for comparatively large changes in the supply voltage. A typical stabilization curve is shown in Fig. 3.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

128.248.155.225 On: Mon, 24 Nov 2014 02:30:42

LABORATORY AND SHOP NOTES 771

90

80

70

en ~ 60

i 50

~4O w

~30 20

10

O~~--~--~--~~--~--~ o 2 3 4 RA 5 6 7 R (MEGOHMS)

FIG. 2. Resistance characteristic of stabilizing circuit (first type).

The width of the plateau and its starting voltage depend only slightly on the value of RA. Curves such as that shown in Fig. 2 will depend on the initial voltage used; RA varied from 5.2 to 4.5 meg in going from a voltage of 8oo volts to 1100 volts. However, there was no important change in the final stabilization characteristics (Fig. 3), and therefore the value of R used for stabilization is not critical. Stabilization occurs only for high supply voltages (above lloo volts in the 5819 multiplier used here). The plateau is broad and fiat, 2oo to 3oo volts wide in this study. The normal gain of the mUltiplier is reduced by a factor of about! by the stabilizing circuit when operating near the beginning of the plateau. The statistical distribution in output pulses from a scintillation crystal is unaffected.

A second method for stabilization, shown schematically in Fig. 4 for the RCA 5819, was investigated in which the potential

FIG. 3. Stabilization curve (first type).

14

1-12 % !a10 1M %8

~6 i4

2

/

O~--~--~--~----~--1100 1200 1300 1400 1500

SUPPLY VOLTAGE (VOLTS)

of an exterior dynode (e.g., dynode No.6 in Fig. 4) is fixed with respect to the potential of the preceding dynode. As in the first method, the battery voltage V most suitable for stabilization depends on the individual multiplier. Battery voltages of 50 to loo volts were most satisfactory in this investigation. The variable resistance R is set as follows. First, the supply voltage is set approx­imately at the desired operating level. The variation in output pulse height with R is shown in Fig. 5. R can easily be set at RB corresponding to the point B by observing the output pulses on an oscilloscope. For this value of R, pulses are stable for changes in the supply voltage. Satisfactory stabilization at point A, as in the first method, is not obtained with this set of dynodes.

When the multiplier is stabilized at B, the voltage between dynodes 6 and 7 is about 10 volts, independent of the battery voltage V and the supply voltage. Typical stabilization curves are shown in Fig. 6. For large battery voltages the stabilization curve has a characteristic maximum and minimum. As the battery voltage is decreased, a fiat plateau is formed which disappears as V is lowered still further. The width of the stable region varies from

multiplier to multiplier and is affected by both the battery voltage and the supply voltage. Stabilization was obtained (using the above procedure for finding RB) for supply voltages as low as 600 volts, where the stable region was about 30 volts wide, and for supply voltages up to 1500 volts where the stable region was about 200 volts wide. The normal gain of the multiplier is reduced by a factor of ! to to by the stabilizing circuit. Like the first

FIG. 4. Stabilizing drcuit ~ for exterior dynode. ~ '1 r-

O~ 06 07

method, this stabilization does not affect the pulse-height distribution.

Both of these methods have the advantages of excellent stabi­lization and simple noncritical circuit adjustments. The stabilized multipliers are no more affected by external magnetic fields than

160

140

120

~IOO :.J g ~ 80 l-I

1:l60 I w ~ 40 ~

20

OL-----~~----~----~------~-o Ra 2 3 4 R (MEGOHMS)

FIG. 5. Resistan~e characteristic for stabilizing circuit (second type).

ordinary multipliers. Both methods, however, produce a con­siderable loss in multiplication as compared to unstabilized operation with the same over-all voltages. The first method is especially advantageous where a broad plateau at fairly high supply voltage is desired. The second method is most useful where stabilization over a wide range of supply voltages is desired, or when it is desirable to vary the gain of the multiplier. In the latter case, R is readjusted for the new operating point.

While the tubes described here refer to the RCA 5819 and the RCA C7151, it is likely that similar behavior will be found for the

~/ -V, 67 ~2 VOLTS

!1 20 ---V'90 VOLTS

it 0 8OO~------~900~-------~~--VO~~~JS--SUPPLY VOlTAGE-

FIG. 6. Stabilization curve (second type).

931A, IP21, and IP28 multipliers. The dynodes are numbered differently for the latter tubes (7 is an inside dynode in the 5819, while 6 an.d 8 are inside dynodes in the 931A). In employing either method, the proper dynodes must be used. The difference in

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

128.248.155.225 On: Mon, 24 Nov 2014 02:30:42

772 LABORATORY AND SHOP NOTES

operation of the two methods described depends on the peculiar focusing properties of the dynode structure. We have used the second type of stabilization for about a year and found it to be dependable. Since no current is drawn from the batteries, it has not been necessary to replace them. We are indebted to-Dr. G. A. Morton for stimulating discussions of these experiments.

This work was supported by the AEC and the Higgins Scientific Trust Fund.

1 G. A. Morton. RCA Rev. 10. 525 (1949).

Frequency Stabilization of Apparatus for the Measurement of Paramagnetic Resonance

Absorption * JACK M. HIRSHON, ROBERT L. WHITE. t AND GEORGE K. FRAENKEL

Department of Chemistry, Columbia University. New York. New York (Received August 29. 1952)

T HE determination of paramagnetic resonance absorption spectra in the microwave region is ordinarily accomplished

by measuring the small change which occurs in the Q of a resonant cavity containing the paramagnetic material as a function of the strength of an external magnetic field. The apparatus is therefore extremely sensitive to any deviation of the klystron frequency from the resonant frequency of this cavity, and consequently a high degree of stability of the klystron frequency is required. Furthermore, since the dispersion which accompanies the para­magnetic absorption alters the resonant frequency of the cavity, it is necessary to maintain the klystron and cavity frequencies in coincidence as the absorption line is traversed. If the spectrum is determined point by point, the retuning of the klystron or the cavity may be accomplished manually, but if a continuous record of the spectrum is to be obtained by mechanically scanning the magnetic field, automatic control of either the klystron or the cavity frequency must be provided. In the equipment to be de­scribed, a stabilization system is employed which adjusts the kyl­stron frequency so as to coincide with the resonant frequency of the cavity containing the paramagnetic sample. Compensation is, therefore, automatically provided for fluctuations in klystron frequency and for changes in the resonant frequency of the cavity resulting from magnetic dispersion.

A number of investigators have used a Pound stabilizer! with a separate cavity to provide the reference frequency for control of the klystron and have either retuned the reference cavity manually or have operated with a zero-absorption signal, large compared to the resonance signal.2 In the latter case, retuning of the frequency to compensate for the magnetic dispersion is not necessary, but it is difficult to use a high gain system without danger of overloading the amplifiers. Whitmer and Weidnerll oscillate the reference cavity mechanically and in this way obtain records from which it is possible to measure the absorption signal corresponding to proper tuning of the klystron frequency.

The stabilization scheme utilized in the present equipment is a modification of Pound's 30 mc/s IF system.! In addition to employing the cavity containing the paramagnetic sample as a reference, this equipment differs from Pound's in the ma~ner of obtaining the IF signal. A heterodyne system is used here mstead of a 30 mc/s oscillator and crystal modulator.

A TE!o2 rectangular reflection type cavity, made from X band (3 cm) wave guide, is used to hold the sample. The cavity is connected to the test arm of a magic tee bridge.4. 6 Two type 2K25 klystrons are used: a signal klystron to deliver power to the magic tee, and a local oscillator klystron to provide the 30 mc/s by hetero­dyning with the signal klystron. The power from the H arm of the magic tee bridge and about half the power from the local oscillator klystron are connected to the appropriate arms of a magic tee crystal mixer. The heterodyne signal from this crystal is led to a 30 mc/s IF amplifier (signal amplifier), and the reso­nance signal is obtained from a peak reading diode detector at its output.

In addition to the signal channel, a second channel is provided for control purposes. A small amount of power is coupled from the signal klystron, together with the remaining power from the local oscillator, into the appropriate arms of a second magic tee crystal mixer. The 30 mc/s signal from the second crystal mixer is led to another IF amplifier CAFC amplifier) the output of which is a Foster-Seeley type discriminator. The discriminator provides the controlling signal for automatic frequency control of the local oscillator klystron, maintaining the local oscillator at a frequency of 30 mc/s above that of the signal klystron. , The stabilization signal is obtained from a phase sensitive, or lock-in, detector! which is connected, as is the diode detector, to the 30 mc/s output of the signal amplifier. The plate of the phase sensitive detector, a type 6AS6 tube, is connected to the repeller of the signal klystron, and the reference signal for the suppressor grid of the 6AS6 is obtained from the 30 mc/s output of the AFC amplifier. The output of the 6AS6 can be made to have the charac­teristics of a discriminator output voltage by proper adjustment of the phase of the reference signal. This phase is varied by adjust­ing a squeeze section located in the line which couples power from the signal klystron to the second magic tee crystal mixer. To determine the condition of proper phase balance, the output of the phase sensitive detector is connected to an oscilloscope instead of to the repeller, and the sweep voltage of the oscilloscope is attached to the repeller. The signal klystron is then swept through a small frequency interval around the resonant frequency of the cavity, and the squeeze section is adjusted to obtain the proper discrimin­ator type pattern on the oscilloscope.

In order to avoid overloading and blocking of the high gain amplifiers, it is necessary to operate with a small amplifier input signal. A small input signal can be obtained if the magic tee bridge is balanced for minimum output by means of careful adjustment of the cavity coupling. The size of the coupling hole is varied to provide coarse adjustment, and fine control is accomplished by adjusting a precision slide-screw tuner placed between the magic tee and the cavity. The bridge is slightly unbalanced when an absorption measurement is made to insure that the cavity is either over- or undercoupled.

This equipment has been used to observe the resonance spec­trum from materials which exhibit as weak a resonance as that obtained from a 50-microliter sample of 0.008 molar aqueous solu­tion of manganous nitrate. Absorption lines as narrow as a few gauss and as broad as 1000 gauss have been recorded. The spectra are observed automatically on a Leeds and Northrup Speedomax recorder, while the magnetic field is varied by continuously chang­ing the current in the electromagnet. The magnet current is con­trolled with a rheostat driven by a synchronous motor.

The sensitivity of the equipment appears, at present, to be limited by instabilities in the microwave bridge and by a change in baseline as a function of magnetic field, even when no paramagnetic sample is present in the cavity. Frequency fluctuations do not present a limitation. The instabilities in the microwave bridge and slide-screw tuner are due to temperature changes and to lack of sufficient mechanical rigidity. The dependence of the baseline on the magnetic field has been shown to be the effect of a change either of the Q or of the matching to the line of the empty cavity. Experi­ments are being continued to eliminate these difficulties.

* Supported in part by the U. S. Army Signal Corps, Squier Signal Laboratories.

t Department of Physics. Columbia University. 1 R. V. Pound. Rev. Sci. Instr. 17. 490 (1946). 'Few detailed account of apparatus have been given; see the following,

however. in addition to references 3 and 5: Bleaney and Ingram. Proe. Roy. Soc. (London) A205. 336 (1951); Bagguley and Griffiths. Proc Roy. Soc. (London) A204. 188 (1950); Schneider and England. Physica 17. 221 (1951); Cummerow. Halliday. and Moore. Phys. Rev. 72. 1233 (1947); R. Arnold and M. Kip. Phys. Rev. 75. 1199 (1949); Tinkham. Weinstein. and Kip. Phys. Rev. 84.848 (1951); Yager. Galt. Merritt. and Wood. Phys. Rev. 80, 744 (1950); Hutchison. Pastor. and Kowalsky. J. Chern. Phys. 20, 534 (1952); Robert D. Malvano and Arthur F. Panetti. Phys. Rev. 78, 826 (1950); Cohen. Kikuchi. and Turkevich. Phys. Rev. 85. 379 (1952); Pake. Townsend. and Weissman. Phys. Rev. 85. 682 (1952).

3 R. T. Whitmer and C. A. Weidner. Rev. Sci. Instr. 23. 75 (1952). • C. C. Montgomery. Techniques of Microwave Measurements (McGraw­

Hill Book Company. Inc .. New York. 1947). p. 515. 'Whitmer. Weidner. Hsiang. and Weiss. Phys. Rev. 74. 1478 (1948).

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

128.248.155.225 On: Mon, 24 Nov 2014 02:30:42