Topics in..
.
Chemical Instrumentation feature Edited by S. Z. LEWIN, New York University, New York, N.Y.
10003. P
These articles, most of which are to be contributed by guest authors, are intended to smx the readers of this JOURNAL by caldng ntten(ion to new developments i n the theory, design, or availability of clmnical laboratory instruntentation, or by presenting useful insights and ezplanalions of topics that arc of practical importance to those who use, or teach the use of, modern instrumentation and instrumental techniques.
Galen W. Ewing, Deportment of Chemistry, Seton Hall University South Orange, New jersey 07079 identification of rovalent compounds by the "fingerprint" approach. T h b is partly
not unreasonahle~figitre,this gives a tatal of 160,000 spaces or "channels" in the 8 to 40 Ghz region I(40 - 8) X 10°/2 X 10" 1.6 X XV]. By comparison, if we msume with Gordy ( 6 ) that an average resolution of 1 cm-' is attainable over the ent,ire infrared region from the viaihle I I to 50p, only 10,000 spaces are available. Of course there are many freqllenries which do not appear in m y know,, spect,ra, so these figures may he somewhat misleading; nevertheless, an extremely I s g e number of componnds ran in principle he examined in a microwave absorptiometer with little probability of overlapping. The chief restriction is that only gmeona samples can he studied, though the vapor pressure need not he high: 10.' to 10-8 t o n is the 11snsl pressure 1,ange. The quantitative ahsorption law (Lambert's law) can he expressed irl the form,
Po.lO-~%r lag ( P d P ) = ab
where Pa and P represent respectively the radiant (microwave) power incident upon and passing through the absorption cell, which has an effective length b. The ahsorption coefficient, a, corresponds to the absorbance, A , as usually employed in the optical region, taken per unit length of theabsorption cell. Ideally i t issfunction only of the number of absorbing molecules per unit path length, and hence should be related to the partial pressure of the subst,anee in a mixture of gases. The degree to which such a relation is valid depends on t,he way in which the absorption coefficient is measured and utilized. At low pressures and power levels, where saturation effects are not evident, the pressitre is proportional to , . , a the height of the absorption maximum. However a t higher pressures, the value of am., becomes constant, and an incresse in p1'e.sswe resulta in the broadening of the h e as measured a t half-height. Above this point it can he shown that the integrated absorption is very nearly proportional to the pressure:
XXX. Microwave Absorption Spectroscopy Microwave ahsorption is primarily a tool for ohsenring and measuring rotational transitions in molecules which pass e s a permanent electrical dipole moment. Absorptions can &o he observed tmder appropriate conditions in rnoleeules po.% sessing magnetic moments, such as 09, NO, NOa, CIO*, and in free radicals. I t thus supplements on the one hand farinfrared and Raman spectroscopy which can provide information about rnolee~rlar rotations, and on the other hand ronventiand magnetic susceptibility meanrements. The region of the elertrnmngnetir spertrum included in the term "mirrowave" has no sharp boundaries, and can be considered to overlap the far infrared. I t is often assigned the rough limits of 1 mrn to 100 cm wavelength, corresponding to s frequency range of 300 to 0.3 Ghz (1 gigahertz, G h , = los hz: 1 hz = 1 cycle per second). The region u-hich hns proved most fruitful for ahsorpbion rpertroscopy lies between approximately S m d 40 Ghz (7.5 mm to 3.75 em). I t should he noted that bhis indndes t,he f r e q~~encies useful in electron-spin magnetic resonance (ESR), so bhat the two method.: share some features in common. The phenomena underlying them, howeve?, are very different, and t,heir fields of utility therefore also differ. The microwave region is distinguished from the infrared primarily by ihe very different apparatus and experimental techniques required. Before discussion of the techniques, we will consider hriefly the thearetied implications.
=
~
where u is the frequency, p iq the pressure, and K is a proportionality constant. Fignre 1 shows the absorption curves of a typical band a t s. series of pressures. If a nonreaeting gas which does not a h s o h a t the same frequency is added to the sample, the value of the integrated absorption is found to be proportional to the mole fraction of the absorbing gas. Hence this integral (the area beneath the curve) is a valid qusntitabive analytical tool. I t s measurement is rather tricky, however. I t is necessary to he certain
(Continued a page A684)
Theory of Microwave Absorption There are no microwave absorptions which are ehclracteristie of specific bond types or functional groups, s~tchas we often find in optical and infrared spectra. Qualitative analysis can only he applied by means of comparison with known spertrs. The microwave region is excellent, as a complement to the inirared in the
FREQKNCY. " - Y o , MEGACYCLES
IFrom Re!. ( 7 ) bv mimiasion. John Wdcr rt: Sona.] Figure 1. Voriotion of microwove absorption curves with presrure chonge. Note that the peak intensily remains constant over a wide pressure range.
Volume 43, Number 9, September 1966
/
A683
Chemical instrumentation that the power Level is not too high, so that neither the absorbing sample nor the orystal rectifier deteotor becomes saturated. The temperature also must be cantrolled. With due care standard deviations of the order of 3-5% are to be expected (but remember that the sample may be only a few micromoles). I n this connection i t must be emphasized that both theory and techniques in miorowave absorptiometry are still young. There is every reason to believe that the signal-to-noise ratio cen be increased drastically over the best now available, so that the preeifiion as an analytical tool should be improved by a t least a fact,or of ten.
TOP
VIEW
[From Re/. (IS) by permission. V c G ~ o m - H i l tBook
Co.1
~ i g " r =3. Field configuration of the dominont mode in o rectangular waveguide. The symbol A. refen to the wovelength in the guide, which is slightly different from the wavelength in freerpmse.
Geometry of a rectangular wme-
Figwe 2. guide.
Microwave Techniques Microwave technology in its modern form originated with the development of radar during and immediately following World War 11. It was recognized early in
A684
/
Journal of Chemical Education
the game that the most convenient and effeotiveway to direct the radiation along desired paths within a. piece of laboratory apparatns is by means of waveguzdes. I n current practice waveguides most commonly consist of straight or cuwed metallic pipes or conduits oi rectangular crass section. I t can be shown, both theoretically and by experiment, that a waveguide of dimensions a and b (Fig. 2), where the ratio a / b is between 2.0 and 2.5 (the usual proportions) is capable of propagating an electromagnetic wwe
of wavelength not greater than 2a. The relative orientations of the electric and magnetic fields in the waveguide a t any instant in time take the farm suggested in Figure 3. The wave is moving along the guide, so that a t an instant one-half period later the fields will have assumed a pattern identical to that of Figure 3 except that all amow heads will be reversed. This pattern is referred to as the dominant mode of propagation, and i t is
(Continued on page A886)
Chemical Instrumentation
For absorption measurements, a spectraphatometer must be assembled from mirmwave components, and as in any
Table 1. Conventional Waveguide Bands Designation Inner dimensions Wavelength Frequency EIAa Bandb (in.) range (em) range (Gha) WR-137 J 1.372 X 0.622 3.66 -5.66 5.30 -8.20 X WR-90 0.900 X 0.400 2.4'2 -3.66 8.2&12.4 P WR-62 0.622 X 0.311 1.67 -2.42 12.4 -18.0 WR-42 K 0.420 X 0.170 1.13 -1.67 18.0 -26.5 W%28 R 0.280 X 0.140 0.749-1. 13 26.5 -40.0 a EIA = st,aodmdrdsadopted internationally by the Electronics Industries Association. The letter designations vary somewhat between manufacturers and in various fields of application; t,hasegiven are as used by Hewlett-Pnckard.
'
the only mode possible for wmdengths greater than a but less than 2u. Waves of lengths less than a will be propagated but various more complex field eonfigllrations (modes) will be possible and the measurements and interpretations of results may no longer he unambiguous. In practice the optimmn wavelength for a particular guide is specified as approximately two-thirds of the maximum wavelength, or about 1 . 3 ~ . The useful range is commonly taken as between 1.la and 1 . 7 ~ . This means, of course, that i t is impassible to cover the whole microwave region of interest with a single system of guides. Hence the region i3 divided into five bands with the properties given in Table 1. Many other bands have been established but are not included in the region of chemical interest. To equal the microwave ontput of a. klystron, the usual eqnations indicate that a black body would have to be aperated s t 10""C!
A686
/
Journal o f Chemical Education
spectrophotometer must include a. variable monochromatio source, an absorption cell (the eq~livalentof a cnvet), and a detector. Sources
Incandescent or black-body sources as u ~ e din the infrared emit far too little power in the microwwe region to be useful.' A radiofrequency are is a white source in this area, but i t also does not have snficient power in any small wavelength i n t e n d . Fortunat,ely we do not need to d e ~ e n don monochromatizine a
can be adjusted to any wavelength over s considerable range. The earliest of these to be used was the mawetm, a. vacuum tube in which the trajectories of electrons are affected by a (Contin7~edon page .4688)
Chemical Instrumentation constant magnetic field in such a way as to cause the electrons to give up part of their kinetic energy to oscillating electromagnetic fields in a cavity resonator. The magnetron is capable of generating high power a t frequencies up to about 26 Ghz, but i t is difficult to tone, and inconvenient to use because of the heavy magnet re. wired. Better suited to microwave spectmscopy is the reflez klystron, the basic structure of which i., shown in Fieme 4. Its action is best exohined h v as&mine. that oscillating fields dreadyexist in th;? metallic resonator cavity (whichalso acts as theanode).
[ F ~ o mRcf (IS)b y permission, MeG~oru-HillBook
C0.l
Figure 4. Schematic raprerentotion of a reRex klystron oscillator. As shown here it is connected to 0 cooxiol tranmirslon lins b u t p u t line"), but it could equolly well connect to a waveguide.
Electrons emitted by the cathode p m through a central bole in the walls of the resonator. The rapidly alternating electric field in the cavity causes a potential between the opposite walls which alternately accelerates and decelerates the electrons passing through. Hence some electrons move out of the gap with a higher velocity than their immediate predecessors, snd soon catch up with them, forming a bunch. All the electrons are repelled or reflected back through the gap again by the negatively charged repeller electrode. If the anode and repeller potentials are correctly chosen, each "bunch" af electrons will arrive hack a t the gap st just the right phase point in the cycle that they are slowed down by the field, and hence deliver energy to maintain the cavity in oscillation. The dc voltages required by the klystron are rather critical, and must be carefully reg~~lnted. To be more versatile, a klystron may be designed so that part of the resonating cavity is outeide the vacuum envelope. I t can then be tuned by means of a plunger which changes one dimension of the cavity. The repeller potential must be changed simultaneously to maintain proper phase relstions. The widest timing range sttainable is about 2:1, but in most commercial versions it is less than this, which constitutes the major limit,ation of the klystron in spectroscopy. Figure 5 shows a commercial klystron. The most recently developed microwave source is called the bacha~d-wave oseiG (Continued on page A690)
1
Chemical hstrumentation
(Courtesy 01 Ihria" Aasociatca. Figvre 5. A typical klystron. The woregulde connects on the bock. Note the threoded plunger for mechanic01 tuning.
lator (BWO). This is a highly specialized vacnnm tube x.hich can be made in a variety of physical forms, all depending on the same operating principle. Consider a. rectangular waveguide r e peatedly folded back on itself, as in Figure 6, with a beam of electrons passing from left to right through a series of holes in the guide. Aamme that a wave appropriat,e to the dimensions of the guide is passing through i t in the righLto-left net direction ("backward" with respect to the electron beam). The electronspassing through the first gap are alternately aided and retarded by the electric field, which it will be recalled is $ways directed across the narrow dimension of the w a v e guide. Yo just as in the klystron, t.he electrons tend to become bunched. Since i t takes the wave a longer time to get from one gap to the next than it doen the eleetron, when the electron gets to the second gap, i t finds a wave slightly lagging in
field relativeto the motion of the electron is effectively just short of one half cycle rather than of a, whole cycle. The result is that the electrons experience a slight deceleration in crossing the second gap, but the hunching effect i~increased.
[Redvown from Ref. ( 1 s ) by germiaaion. McarnulHill Book Co.] Figvre 6. Schernotic diogrorn of o BWO employing o folded woveguidestructure.
The process is repeated a t each gap (perhaps 50 in all), the electrons become mare highly hunched, and the phase lag becomes greater. The resulting repeated deceleration of the electrons causes the transfer of a considerable fraction of their kinetic energy to the oscillating fields. This power feedback is the source of the
A690
/
Journol o f Chemical Education
(Covilcay of I'arian Associalss.) Figure 7. Schematic representation of a helical BWO; (11 is the c~thode,(21 a magnetic shield, 13) the control grid, 141 0 pierced mods, (51low-voltage helix, 161 ceramic support rods. 1 7 1 the termination region, (81 a grounded element to collect stray ions, (PI woveguide Ronge, extending through the vacuum envelope, li01 hollow (annular) electron beam; electrkd cwmeclionr are mode through the prongs to the left.
oscillations. The frequency generated can be shown to be
where u is the frequency of the microwave (in Gb), n is the number of gaps in the folded waveguide, T,is the average time of transit of an electron from one gap to the next (in sees), and T, the corre..ponding time of propagation of the wave. If n >> I. then
n
s
1
2(T.
+ T,)
The values of n and T, are determined by the geometry of the tube, and hence are constant. . ' 7 can be varied by changing the accelerating potential applied to the electron gun. This means that the BWO can be tuned by control of a do voltage, s. major snlvant,age. One manufsctorer reports that he can cover the 8 to 40 Ghz span with four interchangeable BWO (Continued on page A898)
Chemical Instrumentation tnhes, where ten klystrons were formerly required. Figures 7 and S show a modern BWO tube, schemat,ically and photographically, I n t,his Blhe the folded waveguide is replaced by a helix findes nnmber 5 in Figure 7) of metallic tape. A microwave is conducted aloug the helis from right to left, and, provided that the circumference is between oneqr~arterand onehalf the wavelengt,h, the electrical potential hetween successive turns will slternste so H. ltr piny the s ~ rdr ~ n~ rhr w .~hlwsl\.r g.tp p v l ~ w t i a l . H I I I W i d d e d - g ! ~ ~,U,I c f t i ,o~p +~r ~~~ t w heel?^ t> Iron heam must be annular in cross section, pahung as close to the turns of the helix s+; possible. The electron gun, consisting of a cathode (11 with it* internal heater, a control grid (31, and it11 a m d e (4),is specially shaped to produce s w h an annular heam. The microwave energy from the helix is transferred to waveguide fields, and passes out through the gnide window shown a t the top (9).
(Counw of vonan Aaaociotsa.) Figure 8. Outward appearance of the B W O diagrammed in Figure 7. Note lhe waveguide Range.
There m e several semiconductor devices known which can he made to produce sustained oscillations in the microwave region. These include the tunnel (Esaki) diode, the Read diode, and the Gunneffect device. These have the same advantages compare3 to klystrons and BWO's that transistors have relative to their vacuum-tube counterparts: more efficient use of power, lack of appreciable heat dissipation, small size, and no need for vaenum-tight construction. They can be made to produce a5 much microwave power as a small klystron. So far as is known to the writer, none of these devices has yet been applied to spectroscopy, hut judging hy the history of other semicanduct,or devices, they probably will become the somcw of preference for this purpose. For an introdnetary discussion, particularly desoribing the Gunn device, see the article by Bowers cited in the hihliography (8). The output from either a. klystron or a BWO can readilv he introduced into a aaveguide system. Since the signal is essentiallv monochromatic. no comDonent
(Continued on page A698)
A692
/
Circlo Nr 190 on Readerr' Ssnice Card-+
Journal of Chemical Education
1
Chemical Instrumentation Detectors By far the most eommon type of detector for microwaves is a. simple silicon crystal diode. It is conveniently mounted on an insulating stitd, with a short, stiff wire projecting into the waveguide like an antenna, about one quarter-wavelength from a closed end. The diode is inherently a nonlinear device, and is selected to give an output proportional to t,he square of the electric field which it senses in the waveguide. Hence it.s indication is directly proportional to power. Occmionally a bolometer, also called a barretter, is employed as detector. This measures the heating effect of the field, which is proportional to the square of the cnrrent, hence also a measure of power. The silicon diode is much less expensive, and has a smaller time constant, so that it can respond qnicker to changes in micrawave level.
Spectrometers There are on the American market two fairly sophisticated microwave spectrometers. In addition, many investigators have built their own instruments, largely from cnrnmercidly available components. Thus the field is in a somewhat similar situation to that of infrared absorption a generation ago. Before dexribing the two commercid spectrometers, we will lead up to them gradually by discussing some simpler instruments which have been found useful. One of the simplest is diagrammed in Figure 9. I t consists of a section of w a v e guide acting as an absorption cell, connected through mien windows to s. klystron or BWO source at one end and to a rryxtd detector at the other. Linked to the sourre through a special waveguide section which withdraws some of the energy from the main guide is a precision cavity wavemeter (frequency meter) and a second crystal detector. The power supply for the microwave generator includes a mwtooth sweep oscillator. The sweep voltage is applied to the appropriate electrode of the oscillstor tube to vary its output frequency over as wide a range as desired or as feasible, and a t the same time to the hari,zontal deflection plate of a dual-beam oscilloscope. The rectified signal from the primary crystal, after amplification, is impressed upon one pair of vertical deflection plates. If no absorbing gas were present in the cell, the trace produced b y this signal would be a horizontal straight line,* hut if absorption occurs st one or more particular frequencies, a corresponding number of sharp dips will appear in the trace, as shown in the figure. The signal from the secondary crystal, similarly amplified, is applied to the other pair of vertical deflection plates of the dual oscillascope, to produce a trace which shows a sinele sham dim usuallv called a A c b d l v it will be somewhat curved
(Continued m page A700)
Figure 9. "Video'. spectrometer. Ichcmonr b reprerents me s o ~ r c e ; r.m. o trequency meter, P.S., Ine power r ~ p p l ywhich inrorporotss a sweep generator. Dl ond D., h o cryrtal detectors; A, ond A?, h e wide-bone amplBer~.each connected to one $el of vertical denenion plotes in the d ~ ooi ~ l l l o xops.
"wavemeter pip!' Manuel adjustment of the wavemeter cavity by means of its alibrated micrometer screw will muse the pip to move across the screen. When it lines up exrtotly with a dip in the primary trace, a reading of the wavemeter indicates the frequency of absorption in the sample. This simple system can give excellent frequency precisian, provided the absorption maximum iq a strong one, but it isnot very sensitive.
Modulation The sensitivity of a microwave spec-
A700 / journal of Chemical Education
advantages of this procedure are comparable to the advantages of chopping the beam of radiation in an optical spectrophotometer. Perhaps the most obvious way to modulate the wave is by applying an alternating potential in the khz region to the control element of the oscillator tube. Then the output of the crystal detectors em be amplL5ed by means of amplifiers tuned a t the modulating frequency. This has been done, with a resulting hundredfold improvement in sensitivity. (Continued on page A706)
Chemical instrumentation There are disadvantages which make i t quite inconvenient to search for absorp tion lines, and as this method is little used, we will not describe it iurlher.3 Another method of modulstion, which has newly pushed all othem out of the picture, is based on the Stark effect. Thin is the effect upon a rotational absorption line produced by the application of an intense, uniform, do, electric field parallel to the electric vectors of the rnicrowavit being absorbed. The line is shifted or spld into a number of component lines related in a known manner to the rotabianal an. magnetic quantum numbers, J end Me The number of lines produced by the Stark effect may be equal to J, to J 1, or to 2J 1, depending on the symmebry type of the molecule. The spacing of the Stark lines, or the shift if only one is produced, is dependent on the magnitude of the applied field.
+
+
Figure 10. Geometry of a woveguide, showing septum for Stark moduiation. The reptum is energized through the wire connection to the right.
The application of the Stark effect to modulation is accomplkhed by the use of an absorption cell waveguide wilh s. central metallic septum parallel tu it3 broader side (Fig. 10). The septum is electrically insulated from the walh of (he waveguide by inert spacers. Provision is made (not evident in Figure 10) for the circulation of the sample gas both above and below the septum. The septum is electrically connected to a. squarewave generator in the 10-100 khs frequency range, a t a potential high enough to produce 8. field of several hundred volts per em. The sqnare-wave is grounded through a diode ("clamped to ground," the electronic expert woidd say) so that the waveguide septum ia alternately a t ground snd a t a high positive potential. When the septnm is grounded, the normal absorption peak appears,
Figure 1 1 . Possible appearance of oscilloscope screen: at left, on absorption bond, un-modu lated; center, with Stark modulation but simple detection; at right, Stork modulotion with phoresemitive detector.
A similar, though not identical, system was reported recently by Johansson ( 8 ) who fomd it wluahle a- a detector in gas r.h~mmatugraphy. Ile (wed a klysllm at 9.7 Ghz, mod~tlnledat 50 ha.
(Continued on page A704)
A702 / journal of Chemical Education
Chemical Instrumentation .. while during the hall-cycles when the Stark field is ON, the unmodified absorption is replaced b y the Stark components. If the output of the absorption cell were detected and impressed directly on an oscilloscope screen, as in Figure 9, all components would be visible simultaneously, as neither the soreen nor the observer's eye could follow the rapid ON-OFF changes. This confusing situation can be simplified by means of a phase-sensitive detector. Thk is a network of diodes, resistors, and capacitors arranged to combine the signal from the waveguide crystal with a reference signal taken from the modulating generator, so that the signal coming from the guide while the Stark voltage is OFF will bedisplayed upward on the'scope, and that arriving when the Stark voltage
is ON will be displayed downward. Figure 11 shows the type of response to he expected for an absorbing molecule with J = 1. The elements of a simple Stark modulsted instrument sre shown in Figure 12. Note that the Stark absorption cell is of
somewhat larger size x-aveguide than the rest of the system, to reduce attenuation losses. The great advantage of modulation, by whatever technique, is the reduction in noise (or enhancement. of signal-tonoise ratio) which it makes possible. The
LEY Scope
ngure 12. Figure 9.
Elementary rpoctrometer with Stmk modulation and phore detection.
Symbols or in
key is the use of amplifiers responsive only to signals a t the modulation frequency. This increases the sensitivity greatly, but for intense absorptions, the selectivity of the unmodnlated spectrometer is still superior.
Double-Beam Spectrometers I t is possible to construct a microwave spectrometer utilizing the double-beam principle so well known in optical spectroohotomea. The Dower from the source is split between two identical sect,ions of waveguide, one of which contains the sample, then recombined. The reeombination is carried out in a microwave device called a. "magic T" because of its unexpected and unusual properties (not all of which are utilized here). The magic T accept.? the two waves through two of its arms, and emits through a third arm s. wave with amplitude given by the difference in amplitude of the two entering waves. This means that a signal will appear on111a t 8 point where an absorption is taking place in the sample. The signal is best measured by the superheterodyne method. It is combined with a wave from a secondary oscillator to give an intermediate beat frequency, for example a t 30 Mhs, whioh can be amplified in a tuned amplifier, then rectified and applied to an ooscilloscape or a recorder.
.~~
Zeeman Modulation The Zeeman effect is the magnetic analog of the (electric) Stark effect, and i t can be utilized in an analogous manner to modulate a microwave absorption process. I t is easilv a ~ ~ l i e a b to l e substances with Far other molecules, much stranger magnetic fields are necessary. Zeeman modulation is readily produced in the paramagnetic samples by winding a helix around a waveguide. The wavegu~de must be constructed of a nonmaenetic
A704
/
Journal of Chemiml Education
Chemical lsrstrumentation The helix is energized from an oscillator, just as in the case of Stark modnlatim~. If a square wave is required, the frequency mnst be comparatively low, because the inductive efloet of t,he helix makes difficult the production of a sharp square wave. The other part* of the spectrometer can he identical wit,h the Stark instniments. When st,rong magnetic fidds are required, the waveguide can he coiled between the poles of an electromagnet. Large magnetic fields cannot be made to akemate at freq~teneieshigh enongh to be efieobive as modnlat,ion, so the instrument, becomes, in effect,, an eleetron-spin resonance speotrometer. A novel speetromet,er has been described by Radford (ll),based on a form of Zeeman mod~rlalion. Radford has eliminated the use of a waveguide a a sample cell h y transmitling the microwaves as a beam in free space from a transmitting antenna to a receiving antenna, same 30 em or more distant. The sample to be studied ia held in, or flows through, a. glass container in the space between the antennas. Modulation is achieved by a technique which Radford calls "field spinning." The sample cell is placed within two pail5 of orthogonal Helmholtz coils. The two pails of coils are fed 90' out of ohase a t 50 khs. which produces a magnetic field rotat,irrg a t that frequency. So the direction of t,he ~~
~~
Figure 13.
H.-P. Model 8400 M t r o w o v e Spectron
field is varied periodicdly rather than its magnihde as ill conventional rnodidat,ion techniques. The instrument w ~ qtested by observing the absorption bands of the OH free radical. For further details, the original paper shodd be consulted.
~
Hewlett-Psckard, 94303.
The Hewlett-Packard Microwave Spectrometer
Palo Alto, Calif.
-....- ... .-- .......... ........... A706
/
Journol of Chemical Education
The H.-P. Model 84004 is a completely
integrated spectrometer system covering the 8 4 0 Ghz region with four interchangeable BWO sources. A photograph of the instrument is shown in Figure 13and a block diagram in Figure 14. The righe hand panel in the photograph contains the controls and inlet ports for the vacuum pumps and sample-handling system. Three glass stopcocks facilitate introdnction of sampler or flushing gas. I n the
(Conlinuerl on page -4708)
Chemical instrumentation section directly in front of the operator's table can be seen several seet,ions of waveguide with panel controls far power level and ealibrat,ion. To the operator's left are the controls for the microwave source and monitoring equipment together with the read-out osaillnscope and recorder. The operation of the instrument can be followed with the aid of the simplified block diagram of Figure 14. Microwave power is taken from the BWO through a. manually adjustable attenustor, A-1, to the Stark cell and detector. As in m y single-beam spectrophotometer, it is desirable to maintain thesource a t a constant power level. In the present instrument, this is accomplished by a leveling loop which consists of a thermistor detector feeding a power meter which in turn reg,,lstes the power output of the oscillrttor. The 6-ft Stark cell is modulated by a variable-voltage square wave s t s frequency of 33.333 khz, which also serves 8s reference for t,he phme-sensitive detector. A unique feature of this spectrometer is the "signal calibrator." Thii unit, shown a t the top of the diagram, takes some of the power from the BWO, modifies its form, and returns i t to the main line just ahead of the Stark cell. The side path ineludcq sn itttenuator, A-2, identical uith A-1, a modulator operating a t the same 33.333 khz frequency, and finally an adjustable phase shiitm. The modulator in this path does not operate an the Stark principle, but rather through a special silicon diode called a PIN diode, because
I A708 / Journal o f Chemical Educofion
Figure 14. H.-P. Model 8400 Spectrometer, rimpliRed block diagrom. "Mod" refers to a PIN modulator in the calibration section; A-1 and A-2 are identical ~Wenuoton;"Phase" is a manval phase shiher; Xtl is the detecting crystal; This a thermistor detector in the power-leveling loop. For detoils of operotion, see text.
its p- and n-type regions are separat,ed by a layer of intrinsic silicon. The PIN diode acts as anormal diode to frequencies below a b m t 100 Mhz, but above this, including the microwave region, it does not rectify, hilt acts as a high resistance. When the 33.333-khz modulating signal is applied, the diode (which is shunted
across the waveguide) slternately acts as though it were a short circuit and an open circuit, and the microwave is thereby complet,elymodulated. To use the ealibrator, the variable ilttenurttors and phase shift,er are adjusted so that the simnlated
(Continued on page A710)
Chemical instrumentation The operating frequency of the microwrwe source is continuously displayed on an electronic counter, the output of which can be used to place freqtleney markers directly on the recorder chart. The Tracerlab M i c r o w a v e Spectrometer Tracerlabs olTers three spectrometers with dc3ignr~ted model ni~mhers, 1001, 3001K, and 4001K, of s~messivelygreater degrees of sophistication. The respective frequency stabilitie.3, for example, are given ss lWJ, 2 X 10-5 and 1 X lo-' per hour ( d l are better if measured over ihort terms); attainable sensitivit,ies are a = 10-9, 10-9, snd 5 X 10-'Ocm-L. In addition they offer an extensive lineof components r h i c h can be assembled into any of it munher of configorations, to meet the requirements of a.part,icdar enstomer's application. Figure 15 shows the appearance of one such sssembly. Some of the available features are given in Table 2. Table
2. Tracerlab Options"
Sources:
(1) BWO's, 8-90 Ghz, (2) klystrons, 8-185 Ghz, (3) klystron with diode frequency multipliers, (4) quartz crystal oscillator with frequency moltipliers far fixed frequency operation in the range &GO Ghi;. Slabilizatim of Smrce: (1) unstabilized sweep, (2) high-Q cavity, (3) phaselockine to a hieh-stability reference oscilla~or. Frequency Read-out: ( 1 ) Wavemeter, (2) digital counter, (3) selectable marker tied to disolav unit.. (4) " . . freauencv . programming. Display: (1) oscillhscope, (2) stripchart recorder, (3) digital voltmeter. Receivers: (1) video crystal with short time-constant. (2) long time-comtant crystal, (3) s i erheterdyne system. Cells: (1) ~ i m p g (2) , Stark, with moduln,t,incfield mrallel ta or oemend~cular -~~~~~ to th; &et& vector, (3) single or dual beam, (4) in-line reference cell, ( 5 ) temperature controllable from -80' ta +300°C, (6) material: copper, steel, brass. alummum, sllver.. (7) . . len&h - up . to8ft. Vacuum and Sampling Syslem: (1) p r a sure stabilized, (2) static sampling, (3) dynamic (flow) sampling. Modulalion: (1) Stark, 100 khz square wave or sine wave up to 2000 volts, (2) dual-frequency Stark, (3) frequencymodulation.
.
Taken from the mmufaeturer's bulletin, withomission of a few sub-categories, which seem t o the author irrelevant to the present treatment. O t h e r Commercial Instruments The only other commercial microwave spectrometer known to the author is a product of Japan Electron Opbics. Their representative8 states that this has not, "racerlab Division of the Laboratory for Electronics, Ina., Waltham, Mass.
02154. a JEOLCO (U. S. A,), Inc., Medford, Mass. 02155; The parent company is Japan Electron Optirs Laboratory Co., Ine., Tokyo. (Calinued on page A718)
Circle NO. 161 an Readers' Service Card
A716
/
Journal of Chemical Education
-+
Chemjta/ lnsfrumenfafjon vet " -~ ham made available in the United States, and that they have no terhnirnl information concerning it. ~
~~
erenee* to many apectra may be found, especially in references ( 6 ) ,( I t ) , and (14).
~
The works listed helow are t,ho*e u.hich have heen most useful to the author in preparing this brief review. The list is not, intended to be exhaustive, hot will an entry into the field. ~~~
Typical Spectra A few representative mirrowave ahrorption spectra are shown in Figures 16-18. Details are given in the captions. Kc, atla- of absorption spectra, in the microwave region is s t all np-to-dat,e, hut ref-
~
(1) BARROW, G. M., "The Str.t~rttlreof llolecules," Benjamin, New- York, 1!)64. (Paperback: an elementary
Figurs 16. Stark-modulated spectra of formaldehyde vapor, rhowing lhe effect of varying the Stark voltage.
(Cmfinuedon page A 7 W )
A71 8
/
Journal o f Chemical Education
Chemical Instrumentation account which places microwave spectroscopy in a good perspective compared t,o other tools.) (2) Bowms, R., "A Solid-State Source of Microwaves," Scientific American, 215, (No. 2), 22 (Ang~lst 1966). (The Gonn effect and ot,her semiconductor devires compared to the klystron.) (3) L>AILEY, B. P., "Micr~waveSpertroseopy," in A. Weissberger, ed., "Physical MeLhods of Organic Chemistry," being Vol. I, Pert. IV, of A. Weissberger, ed., "Teehniqne of Organic Chemistry," chapter XL, p.
A720
/
Journol o f Chemical Education
2635; Interseience Div. of Wiley, New York, 1960. (A short discussion of the experimental and instrumental aspects of the subject.)
(4) GINZTON,E. L., "Micr~wave~," Science, 127, 841 (1958). (An excellent int,raductory survey of the field and of its applicat,ions, primarily nonehemical.) (5) GOLDSTEIN, J. H., "Microwave Spectrophotometry," in I. M. Kolt,hoff and P. J. Eluing, eds., "Trestise on Analytical Chemistry,'' Part I, Vol. 5, chapter 62, p. 3233; Interscience J h . of Wiley, New York, 1964. (A goad, brief, discnssion of theory; the experimental part reflects largely the anther's personal experience with partly home-built eqnipmenl.)
(6) Gomu, W., "Microwave and Radiofrequency Spectroscopy," in W. We..t, ed., "Chemical Applications of Spectroscopy," being Vol. I X of A. Weiusberger, ed., "Technique of Organic Chemktry," chapter 11, p. 71; Interscience Div. of Wiley, New York, 1956. (An excellent over-all treatment of theory, with no discussion of apparatus.) (7) G ~ R D YW., , SMITE, w. v., A X D T R A M B A R UR.~ , F., "Microwave Snectrascoov." Wilev. New Yark. 1653. background material; in spite of its age.) (8) IIIRVEY, A. F., "Microwave Engineering," Academic Press, London and New York, 1963. (Extensive, modern treatment of microwave devices and their operation; re-
aid
Chemical instrumentation ~
-
-
-
quired reading far anyone planning to build his own spectramet,er.) (9)JOHANSSON, G., Anal. Chem. 34, 914 1 2 . (Use of a sourcemoddated spectrometer as deteotnr in gas chrornat,agraphy.) (10) ]III)E, JR., D. R., "Microwave Spectroscopy," Annual Rev.9. Phys. Chem. 15, 226 (1964). (A "~tateof-thesrt? discussion, primarily concerning applications, very little on instnmentat~ion; bibliography of 171 references.)
..-.
TEMP
I II. . 1 lidic.nl hlicrowave Absorpt,ion Meter," Rev. Sci. Zmtr. 37, 790 (1966). (Describe* n homebuilt Zreman-modulntwl rcprctrometer.) (12) SUGDEN, T. hl. .