RADIATION OF MICROWAVES AND THEIR ABSORPTION IN AIR

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP80-00809A000600340656-6
• Release Decision: RIPPUB
• Original Classification: S
• Document Page Count: 15
• Document Creation Date: December 22, 2016
• Document Release Date: August 16, 2011
• Sequence Number: 656
• Publication Date: September 18, 1950
• Content Type: REPORT

Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
CLASSIFICATION SECRET Rai.
CENTRAL INTELLIGENCE AGENCY
INFORMATION FROM
FOREIGN DOCUMENTS OR RADIO BROADCASTS
COUNTRY USSR
SUBJECT Scientific - Radio, microwaves
HOW
PUBLISHED Bimonthly periodical
WHERE
PUBL'ISHED' Moscow
DATE
PUBLISHED Mar/Apr 1947
LANGUAGE Russian
711 DOCUMENT CONTAINS INFORMATION A/FICTIII 7E NATIONAL DEFENSE
60
OF rN1
U. S. C.. E1 AMC X1, U 11110X0. T ITS 7 0011IINIOMEON THE 1001LAT)0I
ATE' WITMIN
U
01 IT% CONTEM7 IN Al? NAININ TO AN UNAUTM001X10 1110OM II 100?
MIUITXD IT LAW. REPRODUCTION OF TMII FORM II MONIIITOD.
CD NO.
',DATE OF
INFORMATION 1947
DATE DIST. lq Sep 1950
SUPPLEMENT TO
REPORT NO,
Izvestiyaa Akademii Nauk SSSR, Seriya Fizicheskaya, Vol XI, No 2, 1947,
RADIATION OF MICROWAVES AND THEIR ABSORPTION IN AIR
V. L. Ginzburg
Physical Inst imeni P. N. Lebedev
Acad Sci USSR and Gorlkiy U
As is known, the problem of building a sufficiently powerful and reliable
oscillator or microwaves of a given spectral composition in the wave length
range A 0.01-1 cm has not as yet been solved.
Klystrons make it possible to obtain waves with A ^- 1 cm, but further
decrease in wave length is limited by the necessity for still further decrease
in the dimensions of cavity resonators whose volume, generally speaking, is
,L.. 13 ; but this decrease also leads to a drop in power. However, combin-
ing klystrons with crystal detectors, which permit frequency multiplication
because of Their nonlinearity, does make it possible to work in the millimeter
range L'1], but not lower. Single-anode magnetrons permit propagat4ng
waves[2J with frequencies less than
H
6uo= "C -= 1.76?/0 ; x0= ; CH
Even in a very strong field (H = 3^104 gausses) A,,= 3.56 mm, so that it is
impossible to get wave lengths below a few millimeters by this method. Oscil-
lations with higher frequencies can be obtained in magnetrons with split anodes,
but even here the practical limit of the wave length is A = 1 mm. Thus, the
methods of generating radio waves now in use cannot go lower than the milli-
meter range or even, in general, the centimeter range.
Infrared spectroscopic methods can trace the wave-length spectrum of a
mercury-quartz lamp up to A - 0.04 cm X3-7. The ,spectrum of a mass radia-
tor extends from 1 = 0.01 cm to centimeter waves L k_7- However, in view of
the extremely small power and for many other reasons, the usefulness of a mass
radiator, even for laboratory purposes, is very limited. (There has, undoubt-
edly, been a lag in studying,the properties of substances in the micro-wave
50X1-HUM
I
- 1 -
eaaefT
CLASSIFICATION QRQRRT I.
STATE NAW NSRB DISTRISMEW7 EE11
ARMY AIR fBl
Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
 Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
field compared with the technical possibilities of generating these waves.
Research on the behavior of s erconductors in the millimeter wave range are
of great scientific interest 75-7. However, just recently [6J measure-
ments have been made with A = 3 cm. Meanwhile, it is obviously possible to
determine the coefficient of reflection of a superconductor for the whole
range which interests us by using a mass radiator; in any case, by employing
modern methods such measurements are feasible for A< 1 cm.)
The purpose of this article is to discuss certain unexploited possibili-
ties of generating microwaves ( ,1 . 1 cm). Section 1 discusses the prob-
lem of exciting molecular spectra lying in the microwave range. The study
of microwaves as a result of employing relativistic electrons is treated in
Sections 2 and 3. Sections 4 and 5 investigate the possibility of studying
microwaves by means of nonrelativistic electrons moving near a dielectric.
Section 1
The rotation spectra of the majority of molecules lie in the range of
microwaves; therefore, a "tube" in which the rotation level of'the molecules
is somehow excited will be a source of radio radiation. The energy radiated
by such a tube in a second equals __,w u.'4 21m 2 N8 v12.
,W (1)
3 A9-n M Nn z,
where lly~7n his the radiated frequency; ~by,ynl the square of the matrix
element of the dipole moment corresponding to the investigated transition
n -4 m; and Nn the number of excited molecules (holecules on the level n)
L7,.J.
The magnitude Iprn,l?--PO , where po is the constant electrical dipole
moment of a molecule, for instance in the case of a simple rotator
is equal to
2 ,
z JZ
--i y ,BaJ= T
- -2 7l
PJ,J-1Iz=P J{2J+t J,f1
where J-- is the moment of the amount of rotor movement, J = 0, 1, 2, 3, ???,
for the transition O J = ? 1, that is, n = J and m = J-1; I is the moment of
inertia of the rotator, Since ,0_10-fa,
u ?~ ~o'zaNn (2)
At room temperature, molecules are found on many rotation levels and
various transitions take place all the time. Of Bourse, in the presence of
thermal equilibrium, the tube with molecular gas radiates as much energy as
it absorbs from its surroundings. Therefore, Nn in (1) and (2) must be as-
sumed to be the excess number of molecules compared with the number during
thermal equilibrium with the surrounding medium. Because of the expansion
of the pressure lines in the tube, it is scarcely possible to obtain a higher-
than-atmospheric pressure; under these conditions, increasing the number of
molecules on excited levels by means of beating or any other method will
hardly permit making N greater than 101 per cubic centimeter. Thus, it is
reasonable to assume teat, for all tubes, N L 1018. In this event,
U ^-'10- erg/sec when2 2 = 1 cm; U ^-'1 erg/eec when a =1 mm and U ^' 104
ergs/sec when A = 1072 cm = 100 microns.
We can see that in the range which interests us, where k > 100 mic-
rons, radiation of microwaves through excitation of molecular spectra can-
not lead to powers greater than thousandths of a watt; we can speak only of
Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
 Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
015
microwatts, or less, when ~ N 1 mm. Such powers are, obviously, of in-
terest only for laboratory purposes. Radiation of microwaves by molecules,
in particular, can be used as an ideal frequency. standard for centimeter,
and shorter, waves where very exact measurements and stabilization of fre-
quencyr would be difficult. Frequency stability for molecular radiation, or,
more-precisely, for the frequencies of maximum intensity of a rotational
line possessing some width, is ideal. Of course, molecular spectra can be
used as a standard not on'- by "receiving" the radiation of molecules but
also by absorbing them; ,r example, if gas is placed in an "endovibrator"
cavity resonator ?_7, when the oscillation frequency coincides with v,,m
for gas molecules, the damping factor of the system increases abruptly.
Section 2
The most interesting and promising method of generating microwaves is
based on utilization of fast-moving radiators. As we know, if there is some
system (an atom, antenna, or oscillatory electron) radiating (in the system
of coordinate axes in which it rests) electrc-magnetic waves with a frequency
vo , a Doppler effect will occur when the system moves and it will radiate
frequencies
y' ) = v, /i-p 2
/-Pc~s0 (3)
where ~3 = c ; v is the velocity of the system; the angle between
and the direction of observation.
When the radiator speed is great, energy will be radiated chiefly in a
forward direction, and v > YO
Form the general viewpoint, the simplest method of using the Doppler ef-
fect for frequency "multiplication" is as follows: an electron, moving at
high speed, is placed in an electric field which forces it to oscillate in a
direction perpendicular to its velocity (we introduce the latter hypothesis
only for greater definiteness) [10 7. Let the electron move along the
y-axis with a velocity y = P0 ; moreover, in the direction of the x-axis,
l.c there be an electric field E = Eocos w. t, oscillating the electron in
the direction of x-axis (Figure 1). Then, by virtue of the Doppler effect,
at an angle & to the velocity, the electron will radiate frequencies (it is
easy to see that the factor VT_ _R 2 does not appear in this case):
w ?. - ~-p GoS e (1k)
~~ 1, we have
When B= O aid me a V 7-_ '
271'c
w(0) 2\Vv/
-3-
SECRET
NO FT
SECRET
Sanitized Copy Approved for Release 2011/08/17: : CIA-RDP80-00809A000600340656-6
- - -
 Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
Thus, an electron oscillating?in the described manner will "convert"
the frequency, increasing it 2 W times when B = 0 (for example, when
me _ 10, i.e. W 5 MeV, oscillations with 2 20 cm are converted into
radiation with A = 0.1 cm).
The intensity of radiation in the solid angle d -a equels
(6) ` ~'-'o Pee Z' (1- eos8)2-sin 2deos247(1 =.B2)) d1L
8 7,8 c I (I _(3 cos g) 5
where po = exo is the amplitude of the electrical moment produced by the f.eld
E and is equal to the electron charge multiplied by the amplitude of its dis-
placement along the x-axis; 2 the path traversed'by the electron; , the
angle shown in Figure ].
It is eagy to see that, providing
Ica (7)
practically all-the radiation Brill be concentrated in small angles, less than
Bo "' /2 ~1 f3) = arrow (8}
W
The total radiated enegy under condition (7) equals
P e
~`3 (z)a2 / 2 E oz , (9)
V'~ w364' 2 1 *0 2)11
(for the general case,
not Po 2 wo+Po 2 2 r W 4
~? -
3,3.'4(1-fi2)2- 3J3c11 ( 24 c2
where we find that by virtue of the equation of motion, when x V s13c.
the amplitude of the electron's-oscillation is
7C o'= PO Q ( ~G2/ (10)
e w,2 W /
-4-
SECRET
Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
SECRET
 Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
SEC OT
SEC
In order for the above formula to hold good, we must maintain the disparity
xa X (o) ,
that is, the field E mist not be too strong.
If we have not just one electron, but a cluster to which the current J
corresponds, during radiation by all the electrons, independently of each
ill be
other, the energy generated in one second w
u = . 1,82-10_1a/ GZ l Eo2?T. ergs/sec
in volts per cm, 2 in cm.
E
res
am
d i
o
p
,
n
where J is measure
When J = 10-2 amp, Eo = 10 v/cm, l= 102 cm, and mc2 -_ 10, u = 1.8?10-2
erg/sec; when J 10-3 amp, E = 103v/cm, U . 1.8.10-5 erg/sec.
It is c~ jr from the above examples and formulas that no great powers can
be obtained in the incoherent radiation of electrons. To obtain sufficient
power, electrons must be bunched. If the dimensions of the bunch d are less
that the radiated wave length A , the electrons forming part of the bunch
will radiate coherently, that is, the bunch will radiate as a whole, and the
multiple V will appear in (12); where V will be the number of electrons in
the bunch -- an obvious instance of this fact is pointed out by L. I. Schiff
f-87. (If d A ,the bunch radiates ke a "particle" wet v aNchargee
e P , and thus the multiple V Z will appear in (9). Since
N is the number of bunching* per second, the multiple Y will be added in (12).)
Utilizing bunching permits a limited increase in power with the same average
current J. Thus, if J = e 9V N - 10- amp and 'N = 107 bunches fly past in a
second, V will equal 6.10 and, in the above example, U = 1.8.10- ?6.109= 108
ergsj(aec : 10 w. When W = 50(W = 25 MeV), U will equal 250 W under the same
me
conditions; moreover, all radiation will be enclosed in the angle Oj ldegree..
As an example, let us demonstrate the dependence of the radiated energy
on the dimensions of the bunch, assuming that the latter is sharply limited
and has the form of a parallelepiped with the dimensions a, d, b along the
x, y, z axes. In this case, we shall have for the radiation in the direction
lying on the plane y, z
"sip X 0 Sing) . 2 C1
s in cos B (13)
7r6 9 cos
4) S1P1 A (e)
is the intensity of the bunch radiating as a whole, that is, when
where
j
l
a
7rb sin 9 < 1 and Trd_cos8
If individual bunches are found at greater accurately. fixed distances
than ??L , the radiation of these bunches will also Interfere. The intensity
in this case will be determined by the formula ?PoW the theory. of diffraction
for several slits. If k bunches are simultaneously present on the path I
at a distance n A (0) from each other,(n . 1, 2, 3,',--), the intensity will
increase k times compared with the case of incoherency in the radiation of
individual bunches. In practice, however, little is gained by deali~g with
several bunches simultaneously (for example, when N = 10 and 1-10 k - ZN = 1 ).
c 30
SECRET
Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
 Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
SECRET
Forming bunches is a special, and obviously, a difficult task. Co far
as fast electrons are concerned, the deve went of electron accelerator tech-
niques (the betatron, synchrotron, etc.flo permits us to hope that in the
near future obtaining particles with W-*-5 - 25 MeV will no longer be a prob-
lem.
Section 3
In the system examined above, an electron cluster traversing the path 1
must either be lost or turned back with the aid of a special device, for exam-
ple, a magnetic mirror. In many respects, it is more suitable to use, as a
source of radiation, an electron moving in a circle in a magnetic field H. In
this instance, there '' 48 no special need of producing oscillations in the elec-
tron about its trajectory, as its l,eripheral motion is in itself an accelerat-
ing process and hence the electron radiates. Recently, this radiation was
studied in connection with the theory of electron accelerators [8, 9_7.
Under condition (7), per unit of path the electron radiates energy
z 2H z //_W (lu)
~- 3 (c2 2~
where R = Wo == J4 _ - is the orbit radius.
If the strength of the current equals J, the energy radiated per second
will be a G. e 7, 7r R.e Tr 9ncz W
=3.S /D"`~?//(~G ~3J ergs par sac, (16)
where, in the last term J is measured in amperes (H, always in gausses).
When W - 10, H 10 gausses and J . 10-z amp, U = 350 ergs/sec.
me
The conversion frequency of electrons along the orbit equals
c?o= 7,+cal _ c = 7r c- 4( ) ? (17)
W J R to '
In the nonrelativistic case uto mH, and only this frequency icb is radiated.
But in a relativistic case all overtones (harmonics) of the frequency tuo are
also radiated. Moreover, the radiation is concentrated in the plane of the
orbit (in the region of the angles determined by formula f8_7, and the maxi-
mum intensity is converted [9 into an overtone of the order of )3,
i.e., to the frequency 13 J2. / (18)
When u,lw_ff the Intensity of radia-
tion increases slowly (from &o to iveff),
but when iu>ff, the intensity drops
sharply.
AB -0.
It is easy to understand formula
(18) and to obtain it on the basis
of the Doppler effect discussed in
Section 2. A fast moving electron,
moving in a magnetic field, is accel-
erated in a direction perpendicular to
47T ea W
~R 3 ? R 7nc2 ) (15)
uIuncI
Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
 Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
SECRET
the orbit and radiates in a narrow cone (9 < ?0 ) in the direction of its in-
stantaneous velocity; from the viewpoint of a motionless observer at the point P
(Figure 2), the electron radiates only as long as it is on tue segment of the
circumference AB, during which time the arc AB ~- 19o . Thus, the observer sees,
so to speak, a "flash" with a duration.,- R BQ = Bo- , which is repeated every
period, I.e., with a frequency wo/2 7r . Resolving such a field into a Fourier
series, we obtain all the vertones of h frequency bcrd, with the maximum atZa
frequency w eff" 2 N w r 3 where the multiple
~ o (mrc
appears by virtue of the fact that the radiator is in motion (see formula
(5)). In other words, the role of the frequency buo in (5) is played here by
the frequency ., where w, is determined in accordance with (17). Whe _ 10, IL:
Of
the maxiTum will take place at the 1,000th overtone; and when H = 104, . eff .
10-2cm. In the example discussed above with U = 350 ergs/sec, for the maximum
overtone the energy will reach Ueff ^ 0.3 erg/sec. Utilization of bunching
with V = 61109 gives U ti 200 kw and Ueff ^' 200 w, whereby we assume that a bunch
coherently radiates waves with A > eff.
The advantage of a system with a magnetic field is that, in principle, all
the energy transmitted by the electrons can be used for radiation. However, the
whole frequency spectrum is obtained, and if one is interested in quasi-mono=n
chromatic 'radiation, then in the example cited the efficiency is
14 1 000 6
rectilinear radiation (Section 2) the power of a cluster equals 10-~ ?5.10 ^
5.10 w, U:= 10 w; hence the efficiency 1 (the radiation is immediately
5,000
quasi-monochromatic). Further, in the case described in Section 2, all radia-
tion is in the forward direction and is distributed in the magnetic field inrthe
whole orbit plane; for - 10 in this connection we, obtain the multiple 2 7r-60.
Moreover, using a mirror (metal) can change the spectrum and the pattern of the
radiator's directivity in the magnetic field, but in the case of a rectilinear
radiator the efficiency can be made to rise sharply by turning the cluster back.
From preceding statements it is clear that It is Impossible to decide a priori
what type of radiator will be most advantageous -- the decision depends on both
technical and structural problems and on the objectives (whether a narrow or
wide spectrum is required, etc.).
Section 4
Generation of microwaves based on utilization of relativistic electrons is
closely connected with the problem of constructing electron accelerators and with
a number of other unsolved problems. But there is a possibility, at least in
principle, of obtaining these results with nonrelativistic electrons [lOJ. The
fact is that if an electron moves, not in a vacuum, but in a medium with a re-
fractive index n, then Pn will take the place of )3 in the formulas, i.e., the
relation of the velocity of an electron to the phase velocity of light in the
medium is c. Therefore, instead of (4), let us take the formula for the Doppler
T -T
effect in a medium (see, for exampke, I. M. Frank C11J ):
I)-~,r cos
Gv0 ; 0&~=; o) / fl i ) ,
i.e., the place of (=mlcl 2 is taken by 2)1 1 ; the case where )3n = 0.9;'5 or J3n = 1,005 (if J3n 7 1, then, beside the radiation discussed in this paragraph,
the Cherenkov effect will also take place (see Section 5)) corresponds to our
example with W = 10
me
SECRET
Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
 Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
1
Of course, if an electron moves in a medium, it is greatly retarded and
traverses only a very short path. For this reason, at first glance it would
seem impossible to make use of the Doppler effect in a medium to generate
microwaves. Thic, however, is not true, since, if an electron moves near a
medium, not in it, at a distance from it which is far less than the radiated
wave length, its radiation will be the same as in motion in the medium (this
observation was made by Academician L. I. Mande'-.'shtam; see his-calculations
(12]). Thus, an electron can move in a slit between two dielectrics or
simply over a dielectric plate, the surface of which coincides with the sur-
face xy in Figure 1. For an electron moving along the axis of a cylindrical
channel (in a vacuum channel) with e. radius r, when A = 0, the intensity of
radiation will be less than when the channel is absent by the multiple
P- IL 1, 27r ~? 1 r)~2
where Io (x) . Jo (ix); J0 is Bessel's zero function.
The relationship of p and 6 = 2 'lY /n2 - 1 r is evident from the fol-
lowing table: 1
5 0 0.`' 1 1.5 2 2.5 3 3.5 4 4.5
1 0.9 0.8 0.37 0.19 0.093 0.042 0.0185 0.0078 0.0033
It is clear from (21) and the table that, when r < 0.1 the channel
has pracically no effect on the intensity,
If an electron moves in a slit with a width h, the intensity of radiation
when h < 0.1 A is approximately the same as when there is no slit. When
there is one aYielectric surface and the electron moves at a distance h 4
from it, the intensity of radiation may be several times less than in motion in
a medium (a more precise definition of the proper coefficient requires special
calculation). Covering the surface with a semitransparent, conducting layer
(whose thickness is considerably less than that of a "skin layer") also weakens
the intensity, but the essentials and the order of magnitude do not change.
Let us give the formulas for the intensity relating to the case of move-
ment in a medium (practically the case of a narrow slit). Instead of (6); (9),
and (12) we shall have respectively (see I. M. Frank (11]): 1
?fj (e)= wo Po 2'N 201-finces B)2 sin Zi9cos2~p(/ ~6 2)Jd_a
8~ic4~319-f3 fl ocs Ole (22)
A, 2
3c4 fj (9. fi2,"r2)z 3 (,incz~ /~_(? z a12 ' (23)
= f.B2?~D-/z 1?_____ 2 (I-
0 W 2) ergs/seo / (24)
In (24), J is measured in amperes, Eo in volts per centimeter. Practically all
radiation is concentrated in angles less than.the angle
V ? -
Nowadays there are hirequency ceramic dielectrics with e up to 60 and small
losses (tg 5 < 0.0011)-3f. It is true that these figures pertain to much lower
frequencies than those in which we are interested. However, t is difficult to
.M. R F T
Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
 Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
anticipate fanomalou7 dispersion of E and tg 8 when k > 0.001 cm; thus,
we may assume, as we did above, that suitable materials exist with n =
E , 7-8, and without appreciable dispersion: If n a 7, j3n = 1 when
V - 5,000 eV for the energy of the electrons; if n - 4, ,pn = 1 when V = 16keV,
and if n = 2, fin . 1 when V . 100 keV. Thus, frequency conversion with the
aid of the suggested use of the Doppler effect in a medium may be feasible at
total voltages of some thousands or tens of thousands of volts, but market
relativistic effects would require a million volts. Morecver, the radiation
intensity in the case under (24) is considerably greater than that under (].2).
For instance, if as in the example given in Section 2, we assume that in (24)
Eo = 104, j = 10-2 and 2 = 102, and also assume that n ^ 4 and fin = 0.995
(for the same frequency change as when W . 10), U will equal 27 ergs/sec, or
me
1,500 times more than in formula (24). Nevertheless, as is clear from this
figure and from formula (24), the existing increase in power is possible only
if bunching is used, When Y - 6?109 in the example cited, U = 16 kw.
If we assume that Eo = 103 v/cm -- which is also necessary to satisfy con-
dition (11) -- U will equal ].60 w. Let us take it for granted that the presence
of a slit reduces the value of U to 16 w (this energy is obtained from the gen-
erator which produces the accelerating field E). Now, if we calculate the ef..
ficiency, relating the radiated energy to the energy of a cluster, in this exam-
ple the efficiency will be equal to 1/10, since the energy of a cluster equals
16.103. 0-2 - .160 w. It is evident that the efficienud could be made greater
than 10, and likewise, considerably greater than in the examples in Sections
2 and 3. The dependence of intensity on tFie'dimensions of a bunch was deter-
mined by' formula (13) (for a bunch in the form of a parallelepiped) substitut-
ing A. the wave length in a medium, for k . Thus, to enable a bunch to radiate
n
as a whole, when A = 0, it is necessary to satisfy the condition
fn >>
(26)
In practice, the radiation Pf bunched electrons is coherent for
cI (thus, when n . 2 d ,' = u.4 )
10 ci
DV must not exceed the following value of 4Vo;
To obtain sufficiently monochromatic radiation, these must be little change
in the magnitude of Pn along the trajectory of the electron and this magnitude
must be approximately identical for all electrons in the cluster. The requisite
conditions may prove extremely rigorous,: For example, if it be required that the
width of the spectrum A co should not exceed i~J, the dispersion of electron energy
dVo = r1-ftn1.
When pn a 0.995 and n = 4, a Vo = 15 eV, i.e., dVo . 10-3. Here we wish to em-
V
phasize the fact that when fin < 1, the electrons ure not retarded (in the direc-
tion y) by radiation, since the radiation reaction is directed along the x-axis
and works against the force of the field E. The application of dielectrics for
utilizing the Doppler effect in a medium is, of course, possible not only in the
analyzed rectilinear case but also in the case of an electron's motion in a mag-
netic field (motion on the circumference along a dielectric surface concentric
or parallel to the orbit).
During motion in a magnetic field, the nature of the radiation will be like
that described in Section 3. To obtain microwave radiation in this case, it is
possible, as in a magnetron, to create an "electron rotor" revolving with con-
siderable speed between two dielectric plates -- an electron cloud, modulated in
density [2].
SECRET
Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
 Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
1
SECRET
Section 5
The method of generating microwaves stated in Section 4 is based on uti-
lization of the Doppler effect in a medium. In this case an electron, in
order to radiate, must develop oscillations close to its trajectory. It is
also possitle to use Cherenkov's effect for microwave radiation which facili-
tates making use of uniformly moving electrons [14J.
Super-luminescent electron radiation, or Cherenkov's effect, as we know,
means that an electron moving with a constant velocity V in a medium with a
refractive index n will radiate electromagnetic waves if the following condi-
tion is satisfied:
> ryTT (28)
Moreover, waves with a frequency ware rauiated only in a direction which,
with ii , forms the angle B p, determined by the formula:
cos 90 = n (w) (29)
The energy radiated by an electron on the path Z equals [15J
Gz2 S W /I (30)
M>' 1- lez>2 2(C )j dlu,
where integration takes place in the, whole frequency range for which disparity
(28) holds good.
Radiation occurs when an electron moves near a medium in a channel or
slit, or simply above a dielectric; in practice, as stated in Section 4, the
distance between the electron and the dielectric must not be greater than the
wave length a . When there is a channel or slit, the radiation intensity can
be determined by formula 00) accurate to a multiple of the ordei.,f unity, if
the distancb between the electron and the dielectric is < 0.1 !~. (for a more
accurate determination, see article by 'V. L. G:nzburg and I. M. Frank [l2-7).
For ingtance, if n = 7, f3n = 1 when the electron energy V _ 5,000 eV and when
V = 10 eV, the radiated energy in the frequency interval d w is:
n 2
d = 2G~ wdw= 1.28-10_4" Z.w/~c..,= (31)
where, in converting to a numerical example, it may be assumed that 4/= 2.1012
(;k = lima), d w = 2?1011 and 2 . 20 cm-. (We may neglect the dispersion n
which is, obviously, perfectly admissible for many materials in the microwave
region. If dispersion is great, as a rule losses will increase, which is a
disadvantage.) The total energy radiated in the frequency range ti
x1212 is the sum of I x1212 for all degenerate sublevels of the beginning
and final levels C22J.
Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
 Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
SECRET
In (37) we assume, as we did in (34) that to >,A v andI v-V,I < y,.
At the maximum of the line when v e ya ,
where Z-is the time of the free path of a molecule and Z the number of its col-
lisions per second.
by Laurent's impact expansion mechanism and equals
7 __ Z
dv= 2.7rT 2?r
Formulas (34) -- (38) are directly related to the case of electric di-
pole transition. In large gas masses, generally speaking, less intensive
transitions may prove important. As we know, the ratio of the intensity of
quadripole (Ikd) and magnetic dipole (Imd) transitions to the intensity of
electrical dipole transition (Ied), according to the order of magnitude (as-
suming that the frequencies and other parameters of the transition are iden-
tical) are equal to
Ild ti(a 2:. l0-/6 Imd .~.1 % (39)
r d a L zed
where a is the radius of the atom; m the electron mass, and X the i length.
It is evident from (39) that quadripole radiation in the range ofra$io
frequencies is absolutely unimportant (this is also true of higher muitipoles)
and that only magnetic dipole radiation can play a part. In the latter case,
formulas (34) -- (38) are applicable after replacing ~:.
eel x,z l2 by fr',12 , (4o)
2
where I P122 is the square of the modulus of the matrix element of the mag-
netic momen.
x = X. (v, )= 87r2B2I xizl ? ", ' N,' k
3hc. g, Av
A4 w can see from (37) and (40), to find x (W) it is necessary to know
~/a,1x12I or I ? 1 2 1 2 , gl, and N1. Finding. all these values, except A V
is essentially a problem in molecular spectroscopy ('237? Theoretically,
this can also be done for rotational transitions by bringing in a few experi-
mental data. The value of A v for a gas which is not too dense is determined
Let us note that if 4 V ,.~1, the cited for;nulae will no longer be accurate and
must be somewhat modified [24).
Moreover, in (41) it is impossible to take, as is usually done, the value
of Z- used in gas kinetics. Instead we have to speak of a certain effective,
free. path time. The half-width of AV can, and must, be determined experimen-
tally. But, in accordance with its order of magnitude, 's' in (41) is equal to
its value of gas kinetics; for examp'e, in the case of a r, for a pressure of.
1 atm Z/ N 10-10 and 8 V n.i 2' 10 ( 4 V z,, 0.1 cm-1
c
LHere follows a,detailed discussion on the absorption of microwaves in air,
especially by water vapor and oxygen, summarized as follows:
LIf we disregard nonresonant absorption and factors which are of secondary
importance in absorption by ~sotopic molecules, we'may say that when ?t>0.2 cm
air absorbs waves at ,.... 0.5 cm and' 7 ..i 1.3 cm. Waves with A-0-5 cm are
absorbed e times (that is, their intensity drops e times) in a distance of ap-
proximately 300 m (at maximum absorption); waves with ?.- 1.3 cm are absorbed
by water vapor and their intensity diminishes e times in a distance of approxi-
mately ~0 '.:m when..the water content is about one percent by volume (at maximum
absorption, when )= 1.34 cm), Moreover, the absorption line of water is very
wide,. and the absorption connected with it is of great importance in a rather
wide range of wave lengths [26, 27]]
SECRET
Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
 Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
00
BIBLIOGRAPHY
1. R. Beringer, Phys Rev, 70, 53, 1946.
2. L. Brillouin, Phys Rev, 63, 127, 1943, Collected Works "Magnetron," Mos-
cow-Leningrad, 1946.
3. K. Shefe: and F. Matossi, "Infrared Spectra," Section 13, Moscow-Leningrad,
1935.
4. A. A. Glagoleva-Arkad'yeva, Trudy GEEI, Vol II, 1924. DAR SSSR, XX I,
540, 1941; xLV, 10, 1945.
5. V. L. Ginzburg, "Superconductivity," Sections 6, 11, 25, Iz Ak Nauk SSSR,
1946.
6. F. Bitter et. al., Phys Rev, 70, 97, 1946.
7. R. Kronig, "Band Spectra and Molecular Structure," Kharkov, 1935.
8. L. I. Schiff, Rev Sci Instruments, 17, 6, 1946.
9. L. Artsimovich and. I. Pomeranchuk, ZhETF, 16, 370, 1946.
10. V. L. Ginzburg, DAN SSSR, "Radiation of an Electron Moving near a Dielec-
tric" (in'printing).
11. I. M. Frank, Iz Ak Nauk SSSR, Seriya Fiz, 6, 3, 1942. .7
12. V. L. Ginzburg and I. M. Frank, DAN SSSR, "Radiation of an Electron and an
Atom Moving in a Channel" (in printing).
13. B. M. Vul and G. I. Skanavi, Iz Ak Nauk SSSR, Seriya Fiz, 8, 194, 1944.
14. V. L. Ginzburg, DAN SSSR, "Application of Cherenkov's Effect to Radio Wave
Radiation"(in printing).
15. I. Ye. Tamm and I. M. Frank, DAN SSSR, 14, 108, 1937.
16. W. E. Good, Phys Rev, 69, 539, 1946; 70, 213, 1946.
17. W. D. Hershberger, Journ Appl Phys, 17, 495, 1946.
18. R. M. Heiner, G. W. King, and P. C. Gross, Phys Rev, 70, 108, 1946.
19. V. L. Ginzburg, DAN SSSR, 35, 302, 1942.
20. V. L. Ginzburg, Iz Ak Nauk SSSR, Seriya Fiz, 7, 96, 1943.
21. E. M. Purcell, H. C. Torrey, and R. V. Pound, Pbys Rev, 69, 37, 1946.
22. V. Geytler, "Quantum Theory of Radiation," Sections 12, 13, GTTI, 1940.
23. G. Herzberg, "Molecular Spectra," 1939.
24. J. H. van Vleck and V. F. Weisskopf, Rev Mod Phys, 17, 227. 1945.
25. D. M. Dennison, Rev Mod Phys, 12, 175, 1940.
Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
 Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6
SECRET
26. R. H. Dicke, R. Berginer, R. L. Kyhl, and A. B. Vane, Phys Rev,70, 340,
1946; short report; Phys Revs 69, 694, 1946.
27, G. E. Becker Find S. H. Autler, Phys Rev, 70, 300, 1946; short report:
Phys Rev, 69, 694, 1946.
28. R. Schlapp, Phys Rev, 51, 342, 1937?
- 15 -
SECRET
Sanitized Copy Approved for Release 2011/08/17: CIA-RDP80-00809A000600340656-6