 |
Description  |
|
|
BACKGROUND OF THE INVENTION
This invention relates to lasers and more particularly to low-powered dual
frequency gas lasers the output of which may be rapidly switched between
the two frequencies in a controlled manner.
Low-powered gaseous lasers are well known and are widely used in a number
of applications where their high degree of temporal and/or spatial
coherence is of advantage. For a number of these applications, there is a
need to be able to operate the laser system alternately at a pair of
frequencies.
For example, a common technique for assaying a sample for the presence of
or measuring the concentration of a wide variety of specific substances
makes use of the selective absorption of radiation of different optical
frequencies by the sample. In many instances, this technique may be
simplified to observing the transmission of the sample at a pair of
frequencies, one of which is strongly absorbed by the specific substance
of interest but is not appreciably absorbed by other substances which may
be present in the sample, and the other of which is absorbed by neither
the substance of interest nor the rest of the sample. Non-dispersive
selective absorption meters designed to make such observations are finding
increasing application in such areas as process control, pollution
monitoring, and the like. Such meters may employ lasers as the source of
radiation, thereby taking advantage of the laser's high spectral
resolution with a consequent surer discrimination between substances
having similar or complex absorption spectra. While it is possible to
devise an instrument to perform the required pair of observations
simultaneously, a simpler, less complex and more easily realizable
instrument results if the observations are made sequentially in rapid
succession. Further, while it is possible to use a pair of lasers to
provide the pair of frequencies, a single laser with dual frequency output
is often more desirable. Consequently, dual frequency lasers which may be
made to alternate between the two frequencies in a controlled manner are
of utility in non-dispersive absorption meters. An example of such an
apparatus is disclosed in U.S. Pat. No. 4,059,356.
Other potential uses for time-sharing dual frequency lasers include display
generation, information processing, and communications, where the two
frequencies could be used, for instance, to provide separate channels or
to permit frequency shift encoding.
A variety of techniques have been employed to vary the frequency of a
laser. To aid in understanding the principles of these various techniques,
and to better distinguish between them and the operation of the present
invention, a short exposition concerning laser operation may be helpful.
Briefly, lasers are amplifiers of radiant energy which function through the
emission, stimulated by photons, of like photons. (As used herein, the
term "photon" means a quantum of electromagnetic radiation more generally
referred to as light-i.e., ultraviolet, visible, and infrared radiation).
The gain or amplification factor of such an amplifier is highly frequency
dependent and depends on the physical properties of the active material of
the laser. The gain is greater than unity only for those photons for which
the rate of stimulated emission exceeds the rate of absorption. Such gain
is only possible if the active material of the laser (1) can exist in a
pair of energy states which differ from one another by the specific energy
associated with the photons of interest and (2) can more densely populate
the higher energy state than the lower. In the materials of interest in
the practice of the present invention, the differences in energy states
are those associated with transitions between energy levels by the valence
electrons of atoms or ions in a gas or plasma (i.e. those transitions
associated with atomic spectra).
In operation, any of a large number of means, such as d.c. or radio
frequency electrical discharge in, or optical excitation of, the active
gaseous material, may be used to supply the energy necessary to raise or
pump electrons to the higher energy level in order to satisfy the second
of these conditions.
It will be appreciated that it is possible under appropriate circumstances
for both of the conditions just enumerated to be met simultaneously by
more than a pair of energy levels in a single material. Further, a mix of
active materials may be used to produce greater than unity gain at two or
more frequencies. In such cases the gain vs. frequency curve will display
a number of peaks, each centered at a frequency for which the electronic
transition meets these conditions. The relative heights of the various
peaks will depend upon the corresponding transition probabilities, density
of the inverted populations, linewidths, etc. The profile of each
individual peak will, for the cases of interest, primarily reflect the
Doppler and pressure broadening of the emitted radiation (due respectively
to the random thermal motion of and collisions between the individual
atoms or ions of the active material).
When used as a light source, a laser is operated as an oscillator, optical
feedback being provided at at least one frequency for which the total gain
of the laser, including allowances for losses in the feedback loop, is
greater than unity. This feedback is achieved by enclosing the laser in an
optical system so arranged as to circulate photons through the laser such
that their round-trip transit time is commensurate with the reciprocal of
the desired frequency (i.e., the optical path length is made commensurate
with the desired wavelength). Thus, in the case of the Fabry-Perot
resonator, typically used for this purpose, the separation between end
mirrors is chosen to be an integral number of half wavelengths, since the
round-trip optical path involves two reflections. Inasmuch as any
practical sized optical resonator will have a round-trip optical path many
wavelengths long, the desired frequency will be a high order overtone of
the fundamental frequency of the resonator, which typically will be in the
gigahertz range. Thus, such a resonator at any instant is simultaneously
tuned to a number of optical frequencies, each differing from the next by
a single cycle per round-trip optical transit time within the resonator.
To a first approximation, laser oscillation is possible only for this comb
of longitudinal mode resonant frequencies, and then only for those
specific frequencies which coincide with a greater than unity gain (after
allowance for losses in the optical path of the resonator) of the active
material.
If the mode spacing is sufficiently small compared to the width of a gain
line, then oscillation may be possible at more than one of the mode
frequencies. However, when oscillation occurs at a given frequency, the
gain around that frequency is reduced because of the increased rate at
which atoms are removed from the upper state. If the line is homogeneously
broadened (i.e., if the collisional width is comparable to or larger than
the Doppler width), then this effect tends to suppress oscillation of all
modes except the one with highest gain; likewise, when several gain lines
share a common upper level (as is the case for the He-Ne laser),
oscillation will occur at the frequency having the highest gain, and
oscillation at the other line will be suppressed. If the lines are not
homogeneously broadened, then, in general, simultaneous oscillation at
several frequencies is possible.
With this brief outline of the operation of a gas laser, it will be
appreciated that switching between significantly different frequencies
(i.e. frequencies separated from one another by more than the Doppler and
pressure broadened linewidth) is such a laser involves operating the laser
so as to alternately oscillate at a pair of frequencies corresponding to
two different transitions. Two basic prior art approaches for
accomplishing this switching between gain peaks may be distinguished: (1)
those approaches in which the resonant cavity is fixedly tuned to
simultaneously resonate such that two of its resonant frequencies
correspond to a pair of greater-than-unity-gain transitions in the active
material, and the optical feedback is alternately spoiled at one or the
other of these frequencies; and (2) those approaches in which the resonant
cavity is so dimensioned that the comb of resonant frequencies is not
commensurate with the difference in frequency corresponding to such a pair
of transitions, and the tuning of the cavity is then varied so as to move
the comb of frequencies back and forth, thereby providing resonances
alternately at one or the other frequency corresponding to the pair of
transitions.
The first of the above enumerated approaches may be accomplished by varying
the intracavity absorption, thereby altering the overall system gain to
less than unity at the frequency absorbed, by such methods as alternately
introducing and withdrawing appropriate optical filters into and out of
the optical path, or changing the frequency of an interference filter (as
by rocking), within the cavity. Alternatively, a movable dispersing
element (such as a grating or a prism) may be placed within the cavity and
oscillated or rocked back and forth so as to alternately prevent optical
feedback at all but one or the other of the frequencies corresponding to
greater than unity gain. In addition to requiring additional components in
the optical path within the resonant cavity of the laser oscillator, such
apparatus accomplishes the switching from one gain peak to the other by
physically moving a component which often is highly position sensitive.
The alternative approach, that of varying the tuning of the cavity, is
accomplished by altering the optical path length (i.e. the product of the
geometrical length of the optical path and the index of refraction of the
medium) within the resonant cavity, thereby changing the transit time and
thus the comb of resonant frequencies. Such changes in path length may be
accomplished by physically altering the geometrical length of the path, as
by physically moving a mirror which in part defines the path, or by
changing the index of refraction within at least a portion of the cavity
as by altering the pressure of a gas or inducing birefringence in a
crystal. These designs require either physically moving a critical
component defining the cavity, the alternate pumping and evacuation of gas
into and out of a portion of the cavity, or the addition of an exotic
optical component to the optical train in the cavity.
OBJECTS OF THE INVENTION
Accordingly, it is an object of the present invention to provide an
improved method for the operation of a dual frequency gas laser
oscillator.
It is a further object of the present invention to provide a method of
time-shared operation of a dual frequency gas laser oscillator which does
not require optical components beyond those required to contain the active
gaseous material and define the optical cavity.
Yet another object of the present invention is to provide such a method of
time-shared dual frequency laser oscillator operation which does not
depend upon physically moving components of the laser, altering the
pressure of a gas within the laser cavity, nor the use of exotic
electro-optical materials.
SUMMARY OF THE INVENTION
These and other objects are met in the present invention of an apparatus
comprising a dual-frequency gas laser in a resonant cavity so tuned that
the comb of resonant frequencies is not commensurate with the frequency
difference between the two operating frequencies of the laser, and with
one of the resonant frequencies so tuned as to coincide with one of these
operating frequencies. The cavity length and the overall laser gain are so
selected that the spacing between resonant frequencies is greater than the
breadths above unity gain of the individual peaks of the gain vs.
frequency curve. The apparatus further comprises means for subjecting the
active material of the laser to a controlled variable magnetic field. In a
preferred embodiment, this field is of variable magnitude and is so
applied as to be normal to the optical axis of the resonant cavity, and at
least one of the optical elements within the resonant cavity is arranged
to act as a polarizer with a plane of polarization normal to the magnetic
field direction. This polarizing function may be performed by the
Brewster-angle window(s) normally employed to terminate the plasma tube of
the laser.
As the magnetic field strength is increased from zero, the energy levels of
the active material are perturbed, and as a consequence the emission
spectrum of the material (i.e., the gain vs. frequency curve) is modified.
In the simplest case (the transverse normal Zeeman effect) each spectral
peak is split into three components, with the higher- and lower-frequency
components of each triplet moving away from a stationary central peak. The
outer components are plane polarized normal to the magnetic field, while
the central peak is plane polarized parallel to the field. In this
embodiment, the setting of the polarizer relative to the field acts to
spoil the gain of the central non-displaced component while not affecting
the gain of the magnetically displaced components. Thus, by varying the
magnetic field amplitude the frequencies of the peaks of the gain curve of
the laser can be displaced in a selective manner such that alternately one
or another gain peak can be brought into coincidence with a (fixed) cavity
resonance frequency.
It will be readily recognized that, inasmuch as the plasma tube of a
typical gas laser is terminated with Brewster-angle windows, this method
of frequency modulation does not require additional optical components
within the laser system. Further, it will be appreciated that this method
of frequency switching requires neither physically moving any component of
the laser system, nor laborious and time consuming alteration of gas
pressure within the resonant cavity, nor use of exotic electro-optical
materials.
Other objects of the invention will in part be obvious and will in part
appear hereinafter. The invention accordingly comprises the method and the
apparatus possessing characteristic features exemplified in the following
detailed disclosure, the novel features of which are set forth in the
appended claims.
BRIEF DESCRIPTION OF THE DRAWINGS
For a fuller understanding of the nature and objects of the present
invention, reference should be had to the following detailed description
taken in connection with the accompanying drawings wherein:
FIG. 1 is a schematic diagram partly in longitudinal cross-section of an
apparatus suitable for use in realizing the process of the present
invention;
FIGS. 2A through 2C are plots of cavity resonant frequencies and of overall
laser system gain vs. frequency illustrative of various stages in the
operation of the process of the present invention for one mode of
operation of the invention;
FIGS. 3A through 3C are plots similar to those of FIGS. 2A through 2C
illustrative of another mode of operation of the present invention;
FIG. 4 is a schematic view of an alternative apparatus embodying the
principles of the present invention;
FIG. 5 is a view of yet another embodiment of the invention;
FIGS. 6A through 6C are plots similar to those of FIGS. 2 and 3
illustrative of a mode of operation of the embodiment of FIG. 5; and
FIG. 7 is a table summarizing design parameters for a variety of
embodiments of the invention.
Like reference characters on the drawings indicate like parts in the
several figures.
DETAILED DESCRIPTION
Turning first to FIG. 1, as a suitable form of resonant cavity laser there
will be seen a schematic cross-section of an apparatus for use in
realizing the process of the present invention comprising an elongated
hollow glass tube 22 having coaxially mounted at opposite ends thereof
mirrors 24 and 26 defining between them a Fabry-Perot resonator. The
latter are preferably confocal spherical mirrors of the type well known in
the art, or the equivalent. In the particular embodiment illustrated, it
is intended that the laser radiation be emitted axially through mirror 24
(i.e. toward the right in the figure), and therefore this mirror should be
partially transmitting at the frequencies of interest, and accordingly is
preferably finished with a slightly transmissive multi-layer dielectric
coating as is well known in the art. Mirror 26 is intended to be
non-transmitting, and accordingly may be provided with a high-reflectivity
multi-layer coating. Preferably, the separation between mirrors 24 and 26
is adjustable, as may be provided for by a precision nut 28 and screw 30
supporting mirror 26 from tube 22. It will be appreciated that this form
of adjustment is shown only for ease of exposition, and that in practice
the adjustment may be accomplished in any of a number of other ways, as by
mounting mirror 26 on a piezoelectric element or the like. To limit the
cavity resonance to the fundamental transverse mode, an aperture stop 31,
also well known in the art, is mounted adjacent mirror 26. It will be
appreciated that plural mode limiting apertures might also be used.
Tube 22 is divided into two portions, 32 and 34, by a window 36 and a
partition 38, which separate and seal the two portions from one another.
Portion 32 of tube 22, which contains movable mirror 26, may be either
sealed off or open to atmosphere or free to accept samples, depending upon
the intended use of the system. To this end, portion 32 may be provided
with inlet and outlet ports 33 and 35. Portion 34 constitutes a plasma
tube, which is filled with the active medium of the laser. Window 36 is
supported at Brewster's angle with respect to the longitudinal axis of
tube 22 by partition 38. In FIG. 1, window 36 is shown with its plane of
incidence normal to the plane of the figure and parallel to the
longitudinal axis of tube 22.
Extending through the wall of tube 22 into the interior of portion 34 are a
pair of spaced-apart electrodes, such as anode 40 and cathode 42. Anode 40
can be simply a stub of an electrically conductive lead-in. Similarly, the
cathode 42 includes a lead-in 44 that extends through the wall of tube 22.
Cathode 42 also includes an open-ended hollow elongate metallic tube 46
connected to lead-in 44 and disposed coaxially within portion 34 of tube
22 adjacent fixed mirror 24. An open-ended hollow elongate glass discharge
tube 48 is disposed coaxially within portion 34 of tube 22 extending from
within tube 46 of cathode 42 to a point adjacent window 36. Discharge tube
48 is supported by glass wall 50 which is attached to and extends radially
inward from the wall of tube 22. Wall 50 also serves as an insulating
barrier between cathode 42 and anode 40, preventing an electrical
discharge between them except through discharge tube 48. It will be
appreciated that cathode 42 could either be a cold cathode, as shown, or a
heated cathode, and further, the cathode could be situated in a separate
cathode bulb communicating with the interior of portion 34 of tube 22
rather than being disposed coaxially within tube 22.
The interior portion 34 is filled with the active material of the laser,
chosen, depending on the desired frequencies, from any of a large number
of laser gas mixtures known to provide greater than unity gain at more
than one frequency (e.g., Helium-Neon, Helium-Xenon, Argon, etc). For
reasons that will become apparent hereinafter, pairs of gain lines, to be
of interest, must be separated from one another by at least three halves
their line widths. To limit laser operation to only a selected pair of
frequencies in some mixtures, the gain of the laser oscillator may be
deliberately spoiled, as by selectively coating mirrors 24 or 26 or window
36, to increase the system losses at the undesired frequencies, as is well
known in the art.
It will be appreciated that the structure thus far described in detail is
that of a typical DC-excited gas laser, well known in the prior art,
having a Fabry-Perot resonant cavity in which the movable mirror is
outside of the plasma tube, being separated from the active material of
the laser by a Brewster angle window. It will be understood that other
forms of resonant cavity laser (e.g., double Brewster-windowed plasma
tubes, ring lasers, and the like) can also be used in the practice of the
present invention, provided the stimulated emission corresponds to
electronic transitions in individual atoms and that stimulated emission
can take place at a pair of frequencies separated from one another by at
least a few times the width of the widest gain line.
Disposed about the outside of tube 22, opposite discharge tube 48, is an
electromagnet, shown schematically as a pair of series connected solenoids
52 and 54, disposed with their magnetic axes substantially coaxial and
normal to both the axis of tube 22 and the plane of incidence of window
36. Solenoids 52 and 54 are preferably disposed along the length of plasma
tube portion 34 of tube 22 such that, when they are energized, the
resulting magnetic field is substantially uniform along the length of the
gaseous discharge within the plasma tube. The electromagnet comprising
solenoids 52 and 54 is designed, by methods well known in the art of
magnet design, to produce a substantially uniform magnetic field within
plasma tube 22 having a peak field strength dependent upon the spacing of
cavity resonant frequencies, as will be described hereinafter. It will be
understood that the electromagnet may be multi-coiled and further may be
provided with high permeability cores and yokes in the realization of the
design. Electromagnetic solenoids 52 and 54 may be energized in a
controlled way, as by closing switches 56 and 58, thereby connecting the
solenoids to a source of DC, such as that shown schematically by battery
60, respectively through a resistance 61 and directly. For a battery
voltage v and a solenoid resistance r.sub.s, the current through the
solenoid will be v/r.sub.s when switch 58 is closed and v/(r.sub.s
+r.sub.r), where r.sub.r is the resistance of resistor 61, when switch 58
is open and switch 56 is closed. It will be appreciated that switches 56
and 58 are shown merely as an aid to clarity in the following discription
of the operation of the system, and that in practice other forms of
control of the current in the electromagnets might be used, and that, for
instance, the current can be made to be variable in response to a
modulating signal, or can be made cyclic, as by use of a sawtooth, square
wave, or sinusiodal signal generator.
The electrical circuitry controlling the laser system of FIG. 1 is
completed by switch 62 and DC source 64, which provide power to anode 40
and cathode 42. It will be understood that source 64 incorporates the
necessary ballast to compensate for the negative resistance of the plasma
tube. It will also be appreciated that the plasma could equally well be
excited by other means, e.g., by high frequency radiation from external
electrodes.
The operation of the embodiment just described may be best understood by
reference to FIGS. 2 and 3 which are idealized graphs of laser gain
against frequency for various operating conditions of the laser system
with the frequency axes corresponding to unity system gain. As previously
mentioned, the optical resonator, in this case the Fabry-Perot cavity
formed by mirrors 24 and 26, is many wavelengths long, and consequently
any optical frequency for which it is tuned will be a high order overtone
of its fundamental frequency. The cavity is resonant at those frequencies
for which the net phase shift of a wave making a round trip is 0. That is,
resonant wavelengths .lambda..sub.m will satisfy the relation
.lambda..sub.m =2L/(m-.PHI./2.pi.),
where L is the optical path between mirrors 24 and 26, m is an integer, and
.PHI. is the phase shift in radians due to such effects as diffraction by
apertures within the cavity. Ordinarily .PHI. depends only weakly on
frequency, but differs significantly for different transverse modes of the
resonator. If frequency, .nu., is measured in wave numbers, then
successive resonant frequencies of the cavity are spaced apart by 1/(2L),
neglecting the frequency dependence of .PHI.. This mode spacing is
indicated in FIGS. 2 and 3 by the indicia marked x along the frequency
axes.
Also indicated in the figures are laser system gain lines, centered about
frequencies .nu..sub.A and .nu..sub.B, corresponding to a pair of
transitions for which the gain is greater than unity in the (magnetically
undisturbed) active laser material contained in plasma tube portion 34. As
mentioned hereinbefore, the emission (and absorption) and consequently,
the laser gain corresponding to a transition between energy states in the
active material is not all at a single frequency, but is spread over a
narrow range of frequencies about the nominal frequency. This broadening
is primarily due to the random thermal motion of, and collisions between,
individual atoms of the active material. As an example, for a 10:1
helium-neon mixture at 3.5 torr, the 2948 cm.sup.-1 lines reflect almost
equally the effect of Doppler and pressure broadening, and for a 5 cm long
plasma tube 4 mm in diameter operating at 5 ma current, the line widths
are typically 0.018 cm.sup.-1 wide at half maximum, the stronger line, at
2947.9 cm.sup.-1 having a width of 0.023 cm.sup.-1 for which gain exceeds
loss when used in a typical optical system.
In accordance with the present invention, the pair of gain lines must be
spaced apart a distance several times their widths, and the spacing
between successive resonant frequencies is at least the width of the
broadest gain line. Accordingly, for such a laser operating at the
previously mentioned He-Ne wavelengths, the maximum overall length of the
cavity defined by mirrors 24 and 26 should be no greater than that given
by 1/(2L.sub.max)=0.023 cm.sup.-1, or L.sub.max =21 cm. To make the
operation of the invention most readily apparent, and to illustrate a
feature of the invention which can be achieved with a greater mode
spacing, FIG. 2 has been drawn with the mode spacing twice the greater
than unity system gain of either of a pair of equally broad gain lines.
That is, for the figure L is given by 1/2L=2B, where B is the breadth of
one of the lines. For the He-Ne laser example cited above, this would
correspond to a mode spacing of 0.046 cm.sup.-1 and an L of 11 cm. As will
be described hereinafter, other ratios of line width to line spacing may
also be advantageously used in the practice of the present invention.
When operated in the manner illustrated in FIG. 2, the system is adjusted
so that one of the gain line frequencies (here shown as .nu..sub.B) is
coincident with a resonant frequency of the laser cavity. This is
accomplished by displacing mirror 26, as by manipulation of precision
screw 30. To a first approximation (neglecting phase shift .PHI.), the
separation between mirrors 24 and 26 is set so as to make the optical path
length between the mirrors commensurate with the reciprocal frequency.
More precisely, L is chosen such that .nu..sub.B =1/.lambda..sub.m
=(m-.PHI./2.pi.)/(2L). Since L>>.lambda..sub.m, a small fractional change
in L will shift the resonant frequencies by more than their spacing, while
the spacing itself will be altered by only a small increment. Thus, this
condition may be met independently of the conditions the present invention
sets on mode spacing.
Additionally, the optical path length L is also chosen such that the
product of twice the path length and the difference in the frequencies of
the gain peaks, 2L(.nu..sub.B -.nu..sub.A), is not an integer. That is,
the cavity is so tuned that the resonant frequency spacing is not
commensurate with the difference in gain peak frequencies. The simplest
case, illustrated in FIG. 2, establishes a value for L such that
2L(.nu..sub.B -.nu..sub.A) is an odd half-integer [i.e., 2L(.nu..sub.B
-.nu..sub.A)=1/2, 3/2, 5/2, . . . , or 2L(.nu..sub.B -.nu..sub.A)=n+1/2,
where n=0, 1, 2, 3, . . . ]. Inasmuch as the cavity has been tuned so that
one of its resonant modes corresponds to one of the gain peaks, this
condition places the other gain peak midway between a pair of resonant
frequencies, x.sub.1 and x.sub.2, as shown in FIG. 2A.
(In order to insure the greatest path length in the active material of the
laser, the largest value of L meeting the odd half-integer condition and
consistent with L.sub.max will ordinarily be used. Thus, L=(n+1/2)/2D,
where n is the largest integer such that L<L.sub.max and where
D=(.nu..sub.B -.nu..sub.A), is a preferred value for the optical path
length between mirrors 24 and 26. The length of the plasma column might be
maximized in this way, so as to obtain maximum power from the laser, for
example. Note, however, that FIG. 2 does not correspond to this choice of
L.)
If the laser of FIG. 1 is adjusted as described above (i.e., with
L=L.sub.max /2), the closing of switch 62 will activate the laser, and the
condition schematically illustrated in FIG. 2A will obtain. It will be
understood that, in this condition, oscillation will occur only at
frequency .nu..sub.B, as only for this frequency does a resonance of the
optical cavity coincide with a frequency of greater than unity gain in the
laser system. Further, as is well known, for low gain laser operation the
radiation produced by a system such as that shown in FIG. 1 will be plane
polarized, since the Brewster angle window 36 acts to spoil the system
gain for that radiation for which the electric vector is normal to the
plane of incidence of the window. Consequently, the solid curves centered
at .nu..sub.A and .nu..sub.B in FIG. 2 represent only the system gain for
radiation plane polarized parallel to the window's plane of incidence, the
system having less than unity gain for radiation which is polarized normal
to this.
FIGS. 2B and 2C represent the situation as solenoids 52 and 54 are
energized first to half strength (FIG. 2B) and then to full strength (FIG.
2C), by sequentially closing switches 56 and 58, thereby connecting the
solenoids to the power supply represented by battery 60 firstly through
resistance 61 and then directly. The applied magnetic field perturbs the
energy levels within the active material, and as a result the gain
spectrum of the laser is affected, each gain peak becoming multiple. In
the simplest case (the normal Zeeman effect) there is a splitting of the
central peak, with twin satellite peaks up- and down-frequency shifted
from the central frequency by an amount depending on the magnetic field
strength. The magnitude of this shift, in wavenumbers, is given by
.DELTA..nu.=4.67.times.10.sup.-5 H,
where .DELTA..nu. is the frequency difference between one of the satellite
peaks and the central frequency and H is the magnetic field strength in
Oersteds. In most cases of practical interest, the situation is somewhat
more complicated, in that the anomalous Zeeman effect prevails, and the
line shift is related to the magnetic field strength by a different
coefficient. Thus, for the 2948 cm.sup.-1 neon lines, the actual shift is
about 9% larger (i.e., .DELTA..nu.=5.09.times.10.sup.-5 H). In the
anomalous Zeeman effect, each component of the Zeeman-split line may
itself be further divided into a number of sub-components. For the
magnetic fields of interest, these components are unresolved, and merely
act to effectively broaden the lines.
The various components of the Zeeman-split gain spectrum are variously
polarized (i.e., the perturbed energy levels in the active material result
in various transitions which preferentially emit and absorb variously
polarized radiation). The nature of this polarization depends upon the
orientation of the magnetic field with respect to the radiation
propagation direction. In the preferred embodiment illustrated in FIG. 1,
the field formed by solenoids 52 and 54 is normal to the optical axis
defined by mirrors 24 and 26, and the observed effect is the transverse
Zeeman effect: each line of the spectrum of the active material is divided
into components plane polarized with the electric vector parallel to the
magnetic field direction and components plane polarized with the electric
vector normal to the magnetic field direction. In the normal transverse
Zeeman effect, each line becomes a triplet having a central component at
the original frequency which is polarized parallel to the magnetic field
and a pair of half-amplitude up- and down-shifted peaks each polarized
normal to the field. Since solenoids 52 and 54 are also aligned with their
axes normal to the plane of incidence of Brewster angle window 36, the
central peak of the triplet is polarized normal to the plane of incidence
of the window, and the system gain for this radiation is spoiled. The
outer components of the triplet are passed without loss. Consequently, the
gain spectrum of the preferred apparatus with solenoids 52 and 54
energized consists of doubled gain peaks, as represented by the solid
curves on either side of .nu..sub.A and .nu..sub.B in FIGS. 2B and 2C.
In accordance with the present invention, the maximum value of the magnetic
field strength is chosen so as to produce a frequency shift .DELTA..nu.
sufficient to displace one of the satellites of the normally
non-oscillating gain peak (i.e., as illustrated in FIG. 2A the peak at
frequency .nu..sub.A) to an adjacent cavity resonant frequency, as shown
schematically in FIG. 2C. As the resonant frequencies of FIG. 2 have been
set such that the normally non-oscillating gain peak at frequency
.nu..sub.A is midway between adjacent cavity resonances x.sub.1 and
x.sub.2, and since the mode spacing has been chosen so that x.sub.2
-x.sub.1 =2B, the required frequency shift, .DELTA..nu..sub.max is B, and
for the normal Zeeman effect the maximum magnetic field strength in
Oersteds, H.sub.max, is given by
H.sub.max =2.14.times.10.sup.4 B,
where B is in centimeters.sup.-1. For the He-Ne example previously cited,
the coefficient is some 9% smaller, or about 1.96.times.10.sup.4, and as B
is 0.023 cm.sup.-1, H.sub.max w | | |
